EngmEIFIlk+r,u;

JAN P. NORBYE
ü

CHILTON BOOK COMPANY PHILADELPHIA NEW YORK LONDON

t /s?4)
 lntroduction

Ir wAs A LoNc rrME from the day

I first read about the Wankel rotary engine to the dry its potential
became clear to me. And it would not have become clear at all if it
had not been for the articles published by the technical press in the
meantime. For my own understanding of the Wankel engine, I am
indebted to the work of many, notably R. F. Ansdale, Harry Mundy
and Karl E. Ludvigsen.
It has been my good fortune to see and drive Wankel-powered cars of
all origins, some even in the prototype or experimental stage, and my
archivist instincts led me to collect notes on the subject.
Refecting on the material in these files led me to start writing this
book. It is not a definitive work. Things are happening much too fast,
lg7l and new chapters would have to be added every month iust to keep it
Copyright @ by )an p. Norby.
up-to-date. What I have tried to do is to present the full background
First Edition of the Wankel engine itself and the first and second generation of cars
All rights reserved powered by it. Of course, the Wankel engine is still too young to be
Published in Philadetpry, by Chilton Book Company assessed in proper historical perspective.
and sirnultalgo-usly inontaiio, Canada, The full importance of a new development is not rrormally realized at
by Thomas Nclson & Sons, Ltd. the moment of its invention. |ohn Ambrose Fleming could have had
rsBN 0-8019-5 59t-2 little idea of the value of his twin-electrode thermionic valve when he
invented it in 1904, but it was later to revolutionize the entire technique
Library of Congress Catalog Card Number ß-161624
of radiocommunication. The situation must have been similar when
Designed by Cypher Associates, Inc. Valdemar Poulsen invented the magnetic tape recorder in 1900, and
Manufactured in the united states of America when Chester F. Carlson developed electrophotography (xerography)
in 1942.

?",
 urus parcnts ln l'/ -rtr, tllat ltts ll'lvcntion was to replace the
r.,(r)ru 8,il5 rru but it is only on the thresholct of rts career as an automotlve PowtrrPrarrt,
rvSereas the reciprocating piston engine has a long developrnent
history
piston-type aviation engine for practically all commercial arrd military
aircraft in the space of only 30 years! behind it.
A discussion of the Wankel engine today, in terms of its importance A multitucle of minor problems with the piston engine have been
to the future of transportation, can hardly be more accurate than specu- solvecl over the course of the years, and there is a knowleclge
bank con-
lation on the subiect of jet-powerecl aeroplanes would have been in 1941. cerning t5e design and production of such engines that cloes not exist
In both cases, feasibility hacl been established. Beyond that, the rest is for t5e Wankel engine. The challenger, beirrg new, Iacks the aclvantegc
guesswork. Let me now come to the defense of guesswork. Nations, of pre-triecl solutions. It will take tirne bcfore it can reaclt the same level
armies, corporatiorrs, individuals-all are guidecl by guessrvork. The most of perfection as is currcntly to be founcl in the typical_ Detroit VB.
momentous decisions taken by world leaders depend on eclucated guesses, fn all comparisons between piston engines and Wankel engines, we
are comparing a type of power unit barely beyond the experimentl-l
stage
and I believe that the more educated the guesswork, the.higher the po-
tential quality of the clecisions. If we are to try and make up our minds u,ith an engi,ie type deväloped at the cost of many millions of dollars. It
about the role of the Wankel engine in tomorrow's transportation pic- can safely be assumecl that the Wankel engine can be improved sub-

ture, the first thing we neecl is information about the Wankel engine. stantially as we gain more experience with it and try out new ideas that
We must take full account of its rivals, ancl of all conclitions (legal, sccm promising in one area or anothcr.
environmental, economical, ancl supply-relatcd) that affect the choice of In other *oi,lr, it would be presumptuous to attempt to itrclge the
power generation equipment for future vehicular transport. Wankel engine in its prcsent stage of development. We can safely accept
What will be the requirements for future automotive powerplants? its proven advantrg.r. Whatever drawbacks it contains at thc moment
Briefy, it looks like this: mai be overcome by continued research. But back to basics. Wlty should
t5c Wankel engine be considered as a potential sotlrce of motive power
Minimum pollution Maximum freedom for the auto-
for tomorrow's automobile? For the answer, look at the alternatives:
Minimum depletion of resources motive designer
Minimum noise level Maximum adaptability to existing
I\{inimum cost traffic conditions Fuelcells Diesel engines
Nlaximum reliability N{aximum ease of handling and Atomic power Stirling engines.
Minimum maintenance and wear operation Electricity Free-piston engines
I\{aximum efficiency Gas turbines I lybrid-electric systems
Steam engines
Future requirements are no different from those of the past, except for
the new emphasis on cnvironmental issues (pollution and resources). All the above are energy-conversion systems, and it is axiomatic that
Otherwise, automobile engineers have always had the same goals, and you cannot get something for nothing by converting energy from one
the same priorities. Try to think back to the year 1900, and you'll realize form to another. On the contrary, there is always e loss of energy. For
that proponents of clifferent power sources were then engaged in as hard the anti-pollution league, it shoulcl be a sobering thought, too, that
a strugglc for donrination of the automotivc field as they are now. u,henever energy is .Är.rted, therc is a risk of the process
not being
Then the struggle raged between gasoline engines, battery-electric complete, te, sometimes of toxic matter, some-
systems, and steam power. The winner was the gasoline engine, for its tirnes not power plants are not necessarily
higher efficiency, loler cost, superior power-to-weight ratio, greater emission-f d that internal conrbustion engines
operational range, ancl minimum of fuss. The gasoline-burnirg internal will produce exhaust gases with pollutants.
combustion engine that has dominated the automobile industry for half
a century opcrates on the four-stroke cycle. De.spite its 80-year develop-
I. Furl Crrr,s
ntent history, it has manv drarvbacks: nrcchanical complexity, weight, A ftrel cell cannot actually drive an engine. Rather, it is an engine, but
wcar, arrcl a rrcccl for nrairrtcnancc. only in the scnse that it piocluces electric power. Fuel cells have been
 t

i
I
I

I
I
I
I

vi TNTIIoDUCTTON INTRODUCTION vii I

I

The Wankel engine is also a four-stroke internal combustion engine,

I

I
but it is only on the threshold of its career as an automotive powerplant,
whereas the reciprocating piston cngiue has a long clevelopnrent history
behind it.
A discussion of the Wankel engine today, in terms of its importance A multitude of minor problems with the piston engine have been
to the future of transportation, can hardly be more accurate tfian specu- solved over the course of the years, ancl there is a knowledge bank con-
Iation on the subiect of iet-powered aeroplanes woulcl have been in 1941. cerning the clesign and production of such engirres that cloes not exist
In both cases, feasibility had been established. Beyond that, the rest is for the Wankel engine. The challen Eet, being new, Iacks the advantage
guesswork. Let me now come to the defense of guesswork. Nations, of pre-tried solutions. It will take time before it can reach the same level
armies, corporations, individuals-all are guided by guesswork. The most of perfection as is currently to be founcl in the typical Detroit V8.
In all comparisons between piston engines and Wankel engines, we ,.ii

are comparing a type of power unit barely beyond the experimental stage
with an engine type developed at the cost of many millions of dollars. It
can safely be assumed that the Wankel engine can be improved sub-
stantially as we gain more experience with it ancl try out new ideas that
seem promising in one area or another.
In other words, it would be presumptuous to attempt to iudge the
power generation equipment for future vehicu]ar transport. Wankel engine in its present stage of development. We can safely accept §{
4
What will be the requirements for future automoiiu. powerplants? its proven advantages. Whatever drawbacks it contains at the moment
r.I
'r,
l'i
Briefly, it looks Iike this: may be overcome by continued research. But back to basics.'Why should
iT
'.:l
t,,
,l
ir
Minimum pollution
,l

Maximum freedom for the auto- the Wankel engine be considered as a potential source of motive power '"1

Minimum depletion of resources motive designer

')
for tomorrow's automobile? For the answer, look at the alternatives: :r
:.;
,I
Nlinirnum noise level Maximum adaptability to existing
ll
:,

Irzlinirnum cost traffic conditions Fuel cells f)iesel engines i

reliability
.{
tl

Maximtrm Maximum ease of handling and

,4
'j
Atomic power Stirling engines
I\{inimum maintenance and wear
t.i
a

operation Elcctricity Frec-piston engines

Maximum efficiency Gas turbincs I Iybrid-clcctric systcrns
I

Stcaur engines ,,i

;

:.i
AII the above are energy-conversion systems, and it is axiomatic that
't

.t
you cannot get something for nothing by converting energy from one '..,(
),

form to another. On the contrary, there is always a loss of energy. For ;,1
:{
the anti-pollution league, it should be a sobering thought, too, that
l{

ii
.l

wheucver encrgy is cor)vcrtccl, thcrc is a risk of the process not being )!

:,
con)plcte, which in turn urcans wastc, sortrctirncs of toxic rrtatter, solnc-
i,1
,!

t'irncs not. Thcse "unconvcrrtioual" powcr plants are rtot necessarily

.l
,l
i1

crnission-frce. Nor nrust it be a.ssurncd that intcrnal conrbustion engiucs t:

.. ,|
it
will procluce exhaust gases with pollutants. iJ
i'i
!'t
,ir
.,)

l. Fuur Cmrs ;j,l

i,i
f. '!

A fuel cell cannot actually drive an engine. Rather, it fs an engine, but :,,
i.:

orrlv irr the scnsc tlrtt it nrocluces elcctric l)ower. liuel cclls have been '.j
 vi INTRODUCTTON INTRODUCTION vii
Frank Whittle probably did not even suspect, when taking out his
The Wankel engine is also a four-stroke internal combustion engine,
basic gas turbine Patents in 1930, that his invention was to ,.plr..
the but it is only on the threshold of its career as an automotive powerplant,
piston-tyPe aviation engin_e for practically all commercial ,,rd
military whereas the reciprocating piston engine has a long development history
aircraft in the space of only 30 yearsl
behind it.
A discussion of the Wankel engine today, in terms of its importance
to the future of transportation, can hardly be more accurate than specu-
A multitude of minor problems with the piston engine have been
solved over the course of the years, and there is a knowledge bank con-
Iation on the subiect of iet-powered aeroplanes would have been in 1941.
cerning the design and production of such engines that does not exist
In both cases, feasibility had been established. Beyond that, the rest is
for the Wankel engine. The challen Eet, being new, Iacks the advantage
guesswork. Let me now come to the defense of guesswork.
Nations, of pre-tried solutions. It will take time before it can reach the same level
armies' corporations, individuals-all are guided by guesswork.
The most of perfection as is currently to be founcl in the typical Detroit V8.
momentous decisions taken by world leaders depend on educated guesses,
In all comparisons between piston engines and Wankel engines, we
and I believe that the more educated the gu.ri*ork, the.higher -tt
. po- are comparing a type of power unit barely beyond the experimental stage
tential quality of the decisions. If we are tä try and make uf, ou, minds
with an engine type developed at the cost of many millions of dollars. It
about the role of the Wankel engine in tomorrow's transportation
pic- can safely be assumed that the Wankel engine can be improved sub-
ture, the first thing we neecl is information about the Wänkel
engine. stantially as we gain more experience with it and try out new ideas that
We must take full account of its rivals, and of all conditions (legal,
seem promising in one area or another.
environmental, economical, and supply-related) that affect the
choice of In other words, it would be presumptuous to attempt to iuclge the
Power generatjon equipment for future vehicular transport. Wankel engine in its prcsent stage of clevclopnrent. We can safely accept
What will be the requirements for future automotive powerplants?
its proven aclvarrtages. Whatever clrawbacks it contains at the moment
Briefly, it looks like this may be overcome by continuccl research. Rut back to basics. Why should
Minimum pollution Maximum freedom for the auto- thc Wankel engine bc considerecl as a potcntial source of motive power
Minimum depletion of resources motive designer for tomorrow's automobile? For the answer, Iook at the alternatives:
Minimum noise Ievel Maximum adaptability to existing
Minimurn cost traffic conditions F uel cells Diesel engines
Maximum reliability Maximum ease of handling and Atomic power Stirling engines
Minimum maintenance and wear operation Electricity Free-piston engines
Maximum efficiency
Gas turbines I{ ybrid-electric systems
Future requirements are no different from those of the past, except Steam engines
for
the new emphasis on environmenfal issues (pollution and resources).
All the above are energy-conversion systems, and it is axiomatic that
Otherwise, automobile engineers have always'had the same goals,
a,d you cannot get something for nothing by converting energy from one
the same priorities. Try to think back to the year 1900, and you,ll
realize form to another. On the contrary, there is always a loss of energy. For
that ProPouents of different power sources were then .ngrg.d
in as hard the arrti-pollution league, it should be a sobering thought, too, that
a struggle for domination of the automotive field as they
are now. whenever energy is converted, there is a risk of the process not being
Then the struggle raged between gasoline engines, battery-electric
complete, which in turn n'reans waste, sonretimes of toxic matter, some-
systems, and steam power. Tlte winner was the gasoline
engine, for its times not. These "unconventional" power plants are not necessarily
higher efficiency, Iower cost, superior power-to-weight ratio,
greater emission-free. Nor must it be assumed that internal combustion engines
oPerational range, and minimum of fuss. The gasolinä-burning
iriternal u,ill produce exhaust gases with pollutants.
combustion engine that has dominated the automobile
industr"y fo, half
a centttry operates on the four-stroke cycle. Despite
its 80-yeai develop- l. Funr Crus
ment history, it has many drawbacks: mechanical complexity,
weight, A fuel cell cannot actually drive an engine. Rather, it is an engine, but
wear, and a need for maintenance.
only in the sense that it produces electric power. ITuel cells have been
 vüi INTRODUCTTON INTRODUCTION ix
successfully develoPed (rt extremely high cost) for spacecraft and sub- frorn tomorrow's traffic picture is that the best batteries available today
marine.§' and many organizations are now investigating the possibilities
are inadequate to meet present performance requirements. Electric
of fuel cell Power for use in road vehicles. The fuel cellproduces electric prototypes I have driven have Iow top speed and restricted range.
energy without moving parts and with very low emissions. It has many Ratteries are heavy, bulky, and costly. The bases for judging batteries
other claimed advantages, such as highest efficiency at Iow loads, quigt are energy density (watt-hours per pound of battery); power density
operation and no loss of energy at idle. It promises close to I00 percent (watts per pourrd of battery); and life (measured in time and cycles).
efficiency. On the other hand, there may bä starting difficulties, and the High energy density is required for high mileage per battery charge,
Power-to-weight ratio is low. High utilization time will be necess ary to and high power density is needed for rapid acceleration. Different types
offset the high cost. of battery may be combined within the same system, so as to comple-
For a fuel cell to be able to propel a car, it would also have to be ment each other's characteristics. It applies to all of them that they
bulky. Its fuels, oxidants and catalysts can be both dangerous and ex- tend to run hot under sustained periods of discharge, so a cooling system
pensive. The automotive fuel cell is strictly experimental. Future will be needed. One obvious way to stretch the interval between stops
ad-
vances in fuel cell technology are needed beforC we can consider for recharging or change of batteries is to add a regenerative device
even
specialized vehicle applications. In any event, research and development which charges the batteries when the vehicle is coasting downhill or
costs would run into millions of dollars. decelerating. The most direct method would be a generator driven by
2. Arourc Porvrn the road wheels.
There are no batteries available today that fulfill the requirements of
We are not concerned with atomic power generated in stationary a practical electric vehicle. They are either prohibitively heavy, costly,
nuclear reactors (such as now in use by some ,titity companies) ,nä or inefficient. The common lead-acid battery has low energy density.
nlade available to the consumer as eJectricity. At the vehicle end, such
The sodium-sulphur battery runs hot. Zinc-air batteries require a com-
power sources would be dependent on electric vehicle technology which
pressor to feed oxygen to the cells, and a spent-air separator to remove
is dealt with belorv. I{ere we are talking about an on-board -nu.lear the gas conrponents that are purnped through the cathode without par-
reactor, feeding electric batteries to drive an electric motor. Possible ticipating in the reaction. TJre Iithium-halide battery has a long charging
heat sources are plutonium, curium, thalium, promethium, and uranium.
time and short life. Short life also excludes the lithium-chlorine battery.
Of these, only uranium U7 i is mass-produced. The others are in very The nickel-cadmium battery has low energy density and high cost. Silver-
short supply. zinc batteries have short life and expensive conrponents.
Development is so far advanced that a reactor could be made as light The creation of a modern electric car is contingent upon a break-
as 800 Pouncls, which would be considered "portable" in the automotive
through in battery technology. Given an adequate storage battery, the
industry. Ifowever, such a reactor would require massive concrete shield- car would still require a new type of motor. Present motors are not suit-
ing to keep radiation below hazardous levels, in the order of 15,000 to able for automobile propulsion. Experiments with homopolar motors,
25,000 pounds. The size and weight of such a package remove the alternating current irrduction motors and permanent magnet motors
on-board nuclear reactor from consicleration for prrr.nger cars. have so far proved inconclu.sive. Spokesmen for some of the companies
7. EmcrRrc Vrsrcr.Es that have been most active in the electric vehicle field have admitted that
the electric car is, at best, a distant-future possibility.
The electric car promises silent operation as well as absence of emis-
sions. Electric motors deliver full torque at all speeds, giving a higher 4. Gns TunsrNEs
rate of acceleration frorn a clead stop. At standstill, no fuel is used. The gas turbine is attractive to automobile engineers because of its
Because the electric motor is such a simple piece of machip€ry, main- low emission levels, vibrationless operation, excellent torque character-
tenance and repair costs can be expected to stay minimal. The electric istics, high powcr-to-weight ratio, multi-fuel capabilities, and non-stalling
car has no rreed for the costly automatic transmission of today's pas- characteristics. Cost has bcen the k.y problern about developing gas
senger car and needs no differential, engine cooling system, generator, turbines for cars. But today wc Are approaching the point where the gas
exhaust system or nltlffier. W]rat threatens to exclucle the electric car turbine can con'rpete effectively on a cost-per-horsepower basis with the
x INTRODUCTION INTRODUCTION
a reservoir for the working fluicl (water or Freon) in addition
to the fuel
Iargest truck diesel engines now in production. The gas turbine, during
tarrk for the burner. Accässory clrives are more complicated than on a
some 20 years of development for road vehicles, has overcome most of
of
its problenrs. The worst operational drawbacks were excessive fuel con- Vg, and t5e steam engine n.Ldr an oil separator to reduce the risk
steam cutoff is more expensive than
sumptiorr at part load, Iagging throttle response, a low pressure ratio water contamination. The manual
ceiling, ancl lack of retardation on closed throttle. the throttle linkage in today's car'
Concurrently with new advances in fiberglass, glass ceramic, and other It is clear that the steam car has disadvantages in both cost and weight
compared with present V8s. It may have the same ease of
driving ,t'9
nraterials, new heat exchangers have been designecl, giving over 70 per-
offer simplified maintenance, horvever. Even if modern steam
technol-
cent effectiveness combined with a pressure drop of only 8 percent. The
pressure ratio gf a gas turbine is limited by the heat-resisting properties ow can overcome the traditional drawbacks, such as a starting delay,
t5e freezing risk and the explosion risk, the thermal efficiency of the
of the turbine material. New alloys have been developed to withstand
steam engine remains inferior to the gasoline engine's. The steam
tur-
gas temperatures up to 1,800oF., and the pressure ratio is no Ionger felt
to be a limiting factor in raising gas turbine output. bine is a promising alternative to the reciprocating steam engine. It is

The acceleration lag was caused by the necessity of speeding up the more compact, witl higher power-to-weight ratios, but also considerably
compressor before the power turbine could deliver the required torque. more costly.
Like the gas turbine, it requires a torque converter and reduction
gear-
Throttle response in the best gas turbines is still not instantaneous as in
gasoline-driven piston engines, but the acceleration lag has been reduced i,g. It strouta be easier to maintain than conventional steam engines,

to an acceptable 1.0 second (against 0.3 second in the modern V8) by .nä has proved its excellent reliability in ships and locomotives.
greater precision of fuel control and reducing rotor inertia to a mini-
murn. The obiections regarding lack of retardation on the overrun have 6. Drcsnr. ENcn'IEs

been conquered by installing a variable-pitch nozzle system in the power Common in trucks and buses, with a bad reputation for producitg
dark, dense, and smelly exhaust grs, the diesel engine
may turn out to
turbine. This feature also improves part-load efficiency. Automotive gas
be a real sle.p., in th; long-term clean air car race. The
diesel engine
turbines down to 360 horsepower are on the way. Truck and bus applica-
the today for railway, marine,
tions are under study, but no concrete plans for turbine-driven passen ger is the main pä*., plant thräughout worlcl
and there are some car apPli-
cars are in process. heavy truck ancl inclustrial uses, Passenger
cations (Mercedes-Benz, peugeot, Perkins and Austin ) . Keep in mind
,. Srrerr ENctNns that the Wankel engine does have diesel-fuel capability.
Although the basic steam engine is less expensive than a modern V8 A compression-ignition engine, with high airlfuel ratios and extremely
a complete steam power system would cost three times morel However, higS compression iatios, thetiesel's exhrust is actually healthier
(if this
gasoline_engine. The cliesel
since tlte steam engine procluces maximum torque at zero r.p.m. it can can be saicl) tSan tSat of the spark-ignition
exhaust is u.ry low in carbon *ono*id. content. The
diesel engine is
use a cheaper and less complex transmission. Emission controls are
of vibration
simpler and the exhaust system for a steam engine woulcl cost about half more expensive, has a poor pou,er-to-weight ratio, is not free
gives off loud combustion noise. Great pro$ess has been made in
of that on a V8 gasoline engine. The steam engine does not require a and
cooling system wiih a radiat;, thermostat, fan and fan drive, and water the past ten years, but a total solution is not yet in sight.
pump. It has multi-fuel capability and offers reverse torque operation
for reta rcla tion .
7. Srtnrruc Eucluns
In contrast with the gasoline engine, the steam engine does not suffer The Stirling engine is a "hot air" engine, ätr external combustion
fluid' It
an efficiency loss uncler part load. No steam is wasted at standstill, while engine rtrnnin'"g on-k.rosene fuel, with hydrogen as its working
levels, and is practi_cally vibra-
the V8 continues to use gasoline while idling. Impressive advantages? has multi-fuellapability, very low noise
is usually thought of as a power source in hybrid-el9cj1ic
cars'
Yes, but there's the other side of the coin. The steam engine requires tionless. It
an expensive burner assembly, with an electronic brain, a steam gen- It for constant-speed operation, although it could be pro
is best suited
vided with gearing to drive the wheels direct. In the hybrid
app_lication,
erator, plus complex sensors and controls. The condenser costs about
twice as much as a radiator in a conventional car. The steam car needs t5e Stirlinf engine plus the electric batteries, alternator and motor,
 xü TNTRODUCTTON
xiü
INTRODUCTION
controls and reservoir, form an extremely bulky installation. It is also
heavy and costly. But thermal efficiency is high (39 percent) and its Independently conducted tests show the following results:
specific outptrt beats both the diesel and the gasoline engine. Engine CH CO NO,
Only two comPanies are devoting any great amount of attention to NSU Ro-80 1.54 19.7 -
the Stirling engine: Philips Gloeilampenfabriek in Eindhoven, Holland, Mazda R-l l0 1.4 I 1.9
arrd General Motors. |udging by reports on their experimental units, we Curtiss-Wriglrt I .l I I7.8 1.14
can only conclude that a hybrid car with Stirling engine is not a prac- l97l limits 2.2 27,0 4.0
tical automobile. 1971 limits 1.5 23.0 1.3
8. Fnrn-PrsroN ENcrNrs
The free-piston engine is not capable of running a mechanical drive
trairr. It is an external combustion engine and a gas mover. The best The emission characteristics of the Wankel engine are fully discussed
proposals for its use in automobiles involve its application as a feeder in Section II, Chapter 15.
for a gas turbine. Combining a free piston enginä with a gas turbine The advantages of the Wankel engine over the reciprocating piston
gives the same low emission levels that the gas turbine delivers, with engine are dealt with in Section I, Chapter 7. There is iust one aspect
even greater multi-fuel capabilities. With a gas thermal efficiency of of the Wankel, for use in passenger cars, that I would like to enlarge
44 percent and a turbine efficiency of 83 percent, the overall thermal on here. It concerns the greater design freedom offered to product
efficiency at the shaft is 36., percent, which is comparable to a modern planners, engineers and stylists. In other words, when Detroit starts
diesel engine. The free-piston engine is inherently balanced and virtually manufacturing Wankel engines, we should look not for detail modifi-
vibrationless. Torque clraracteristics, thanks'to the turbine, are excelleni. cations in transmission systems or chassis engineering, but for basic
What are the drarvbacks? Since the free-piston engine has no revolv- changes in vehicle architecture.
ing slraf t, outside units are required to provide ,...rrory drives, for How will adoption of the Wankel engine affect future passenger car
starting, oil pumps, coolant pumps, and fuel pumps. Starting is usually design? It will have immediate effects in three areas: safety, comfort and
accomplished by compressed air fed into the bounce chamüers at 400 cost.
Psi. The normal automobile has no high pressure air source, but it is Sarnrv
claimed that a standard truck-type brake system air compressor can do
tlre iob. The turbine can conveniently be equipped to drive pumps and The Wankel engine will enable significant gains to be made in two
other accessories. areas of accident avoidance: visibility, and handling precision. Visibility
The added weight and complexity of running both a free-piston can be improved because the small engine will allow a lower hood line
engine and a gas turbine in the same vehicle has led engineers io the and a sizeable gain in glass area. A driver who can see more is a safer
Iogical conclusion that they are better off with the gas lurbine alone, driver. Handling precisiorr will be improved because the Wankel engine,
built as a complete unit to undertake all operating firnctions for itself being smaller, will have a lower center of gravity, and its lightness means
rather than receiving exhaust gas flow from a separate engine. a reduced engine mass, which translates into irnproved weight distribu-
tion. A lower center of gravity in combination with improved weight
CoNcr,usIoNs distribution means increased controllability, nrore predictable steering
There are many indications that the best short- and medium-terrn response, and reduced risk of rollover.
solution lies with the internal combustion engine (piston-type or The Wankel engine will have little influence on the effects of a col-
Wankel). With the use of positive crankcase ventilatiär, pr.-h.ated lision. But there is this point: future cars are expected to have bunrpers
intake air, transmission-controlled spark advance, afterburn.i, with or good for l5-mph impacts without dam age to the car. Some form of side
without supplementary air iniection, exhaust gas recirculation, and cata- impact protection is also likely to becon're a legal requirement. All this
Iytic converters, it is possible to meet current standards with a comfort- will add to the weight of the car. With retention of the V8, the car
able margin. This is true of both Wankel and piston-type engines. would gain weight. The Wankel engine will allow these safety features
to be installed without making the car hcavier than prcsent cars.
 xiv TNTRODUCTTON
CorurroRT
How can the Wankel engine possibly contribute toward
greater crea-
ture comfort? Since it is practically free of vibration,
all ocäupa,ts will
be more comfortable. Next, Iet us suppose that the
designer'of a new
nredium-size car had to work on two pärallel
versions of the same basic
vehicle, one with a six-cylincler reciproärting piston
engine and the other
with a twin-rotor Wankel engine. He woulJ ioon find it rt
the smallness
of the Wankel engine woulcl allow a roomier interior.
A net increase in
Iegroom, for instance, promises more comfort.
A roomier interior also Contents
contains more air, which means more efficient
ventilation. It also plays
a part in sa fety, since efficient ventilation
can help combat driver fatigue.
Cosr
Certain savings can be had in the Wankel-powered Introduction
car as a direct
result of the engine's compact size and light SECTION I-DssIcN
weight. With a front-
engined car of medium size, the need for power
reduced' Since the engine's space requirement
steering would be l. Principles of Operation 3
is smaller, it will take Iess 2. Geometry l+
ntaterial to- package it. The space saving therefore
also indicates that a 3. Displacement z6
saving can be made in_the frame, the .rlgin. craclle,
and t5e sheet metal 4. Sealing
around it' The Wankel engine's compact size invites 30

stallation for front-whcel-drive or midsliips engine

transverse in- 5. The Conrbustion Process 12
installation, thereby 6. Housing and Rotor Cooling 5l
elinrinating the propeller shaft and avoiäing other
costly transmission 7. Advantages of the Wankel Engine 59
complications.
Iior some examPles of how this works out in practice, SECTION I I-DnvELoPIvrENr
take a look at
the cars I have described in Section III of the book.
8. Felix Wankel 69
No doubt a brief study of these automobiles will raise
a number of 9. NSU Develops the Wankel 87
questions in your mind. It is my hope that
I have provided the answers 10. Curtiss-Wright r58
elsewhere in the book. If I havän't, I would like
t; know about it, and ll. Toyo Kogyo 246
invite your correspondence.
12.. Daimler-Benz 276
13. Citroön 303
Io, P. Norbye 14. Other Wankel License Holders 310
15. Exhaust Enrissions 331

SECTION III-APPLICATIONS
16. Introduction )53
L7. The NSU Wankel Spider 357
18. The NSU Ro-80 368
lg. The Mazda Cosmo 387
20. The Mazda R-100 )97
?.1. New Mazda Rotary Cars 406
22. The Citroön M-35 415
z). The Mercedes-Benz C-l I I 427
24.The Mustang RC2-60 452
25. Conclusions 460

Appendix 168
Other Rotary Engines 468
Bibliograplry 515
Articles, SAE Papers, Books 5r5

Secti on I

DESIGN
 1

Principles of
0peration

IrB WeNxm nNctNn is an irrternal

combustion engine operating on the four-stroke cycle: intake, compres-
sion, combustiorr ancl exhaust. These four strokes serve to clraw in the
air-fuel mixture, compress it to intensify the charge so that a greater force
is released during combustion, and clean the combustion chamber in
preparation for the cycle to begin again. Keeping this in mind, an inter-
nal combustion engine can be thought of as a device for harnessing an
expansion force and converting it into motion. In this sense, the Wankel
engine can be thought of as a halfway point between the conventional
automotive piston engine and the ground transport gas turbine (rr op-
posed to aircraft turbines for jet planes).
AII operational cycles occur in the sanre area in the piston engine,
while each phase of the cycle occurs in an area specifically designed for
it in the Wankel and turbine engines. The Wankel an«l turbine engine
functions are spaced out according to the gds flow path, while in the
piston engine, everything takes place in the same area and the various
phases are separated only by time. But, even though the Wankel engine
appears to be more closely related to the turbine than to the piston en-
gine (because of their common rotary motion), there is a great difference
-the Wankel engine (as well as the piston engine) is an intcrnal com-
bustion engine while the turbine u'orks by external combustion.
A better unclerstancling of the Wankel may be gained by re','ierving
briefly thc opcratirrg principlcs of both thc four-strokc cyclc 1>iston cr)-
gine and the gas turbine. A single-cylinqler piston engine takes four
strokes (two crankshaft revolutions) to cornplete its operating c1'cle.
Starting at top deacl center, between cycles, the first stroke is a down
4 PRINCIPLES OF OPERATION
stroke-intake. Fresh gas mixture is drawn into the cylinder by the
vacuum created by the downward motion of the piston. Tfte piston then
moves upward in the compressiou stroke and squeezes the mixture into a
small space at the top of the cylinder (combustion chamber).The com-
bustion stroke (often called the power shoke) is the downward working
stroke that produces the power. The compressed gas is ignited by a spark
plug, and the force of expansion created by the burning of the air-fuel
mixture drives the piston downward under very high pressure. The ex-
haust stroke is an up stroke, following the power stroke. The motion of
the piston helps push the burned gases out of the cylinder. At the end
of the exhaust stroke, the cylinder is emptied of burned gases and ready
for another intake of air-fuel mixture. Then the cycle is repeated.
The intake phase of the gas turbine takes place through an annular
chamber which leads to a compressor. A compressor is an air pump
which, for gas turbines, consists usually of radial-flow or axial-flow vane
wheels working inside closed chambers to build up air pressure. There
is continuous flow throughout the compressor and turbine; therefore,
instead of having a compression ratio, the gas turbine has a pressure
ratio. This pressure ratio (4.5:l is considered higlr) is far lower than the
compression ratio of an internal combustion engine (where 8: I is norrnal
and l1:l is high). Diesel engines, which rely on compression pressure
ratlrer than spark for ignition, have compression ratios as high as 22.5:1.
The combustion phase of the turbine takes place in the burner sec- a
tion. Ignition is performed by a continuously glowing plug which func- In the four-stroke Wankel, all
tions as a spark plug. Gas turbines do not have carburetors to mix air and three working chambers are in
continuous action. While rotor
fuel; they compress clean air and inject fuel into the combustion section. face A is sweeping out the re-
The same principle is used in piston errgines and Wankel engines rvith maining exhaust gases and pre-
fuel in jection, but the point is that gas turbin es depend on fuel injection paring to begin a new intake
phase, chamber B is beginning
for their efficient operation while internal combustion engines work well compression and chamber C is
with either carburetors or fuel injection. about to complete its expansion
phase. In the second sketch,
|ust as the piston engine harnesses an expansion force, so does the chamber A with its
goes ahead
gas turbine. A turbine is like a compressor in reverse-a vane wheel gets intake, while chamber B is ap-
its momentum from gas flow ancl converts the gas thrust into mechanical proaching maximum compres-
sion. Chamber C has iust started
torque. The automotive gas turbine actually is two turbines rvhich have its exhaust phase. In the third
no mechanical connection-the first drives the compressor and the sketch, ignition takes place in
engine accessories and the seconcl drives the output shaft. The main chamber B, chamber A is about
to complete its intake phase, and
difference between the gas turbine and the Wankel engine is that chamber C is in the middle of
the turbine's operational ftrnctions are spaced out axially along the tur- the exhaust phase. The fourth
bine shafts, while the Wankel's areas of activity are spaced out radially sketch shows expansion in cham-
ber B, completion of intake in
from the rotor. chanrber A, and continued scav-
Mechanically, the W'ankel engine differs significantly from the piston enging in chamber C. (Rotor
revolves CCW.)

N Etr
engine. T'he cylinders and pistons are replaced by working chambers and
ffifl,lnffi AUSSC HIE 8E N
 COMPRE SSED AIR FUEL NOZZLE
FUEL LOIY PRESSURE COMBUSIOR
SPARXING PLUG EXHAUST GAS
REGENERATOR

HIGH PRESSURE & HIGH TEMP. GAS

TO EXHAUST
RAOIAL COMPRESSOR
REGENERAIOR
ORIVE

LOIY PRESSURE
HOT GASES OUTPUT SHAFT
-+
VARlABLE -+
AIR INLET
GENERATQR - NOZZLE
-+
BLADES

-'} COMPRE SS 0R-

R TURBINE ROTOR
TURBINE RO TOR

SHAFT

FUEL OIL
PUMP PUMP
INTAKE
AIR
The automotive gas turbine does deliver shaft horsepower because a second turbine
COMPRESSOR is addecl behincl 1tr. first one that drives the compressor. The second turbine is
AIR COMP'IESSR I
HIGH TEMPERATURE
ruRBtNE
POWER
TURBINE
integral with the outl>ut shaft an«l can bc couplecl to a nornral transmission.
HIGH PRE5SURE GASES

shows how a simPle PrinciPle of

in n lot of vital accessories and a COOLING FAN
WATER PUMP

END COVER
HYORAULIC TOROUE
CONVERTER

COUPRESSED AIR FUEL ilOZZLE

COMBUSTOR

HIGH PRESSURE & HIGH TEMP. GAS

RADIAL COUPRESSOR

HIGH PRESSURE GAS

+
+
AIR IIILET
-+
+ END COVER

TURBIilE ROTOR OIL FILTER

----+
SEPARATING WALLS

This is the twin-rotor Wankel engine that powers the NSU Ro-80. Its simpli-city
The iet aircraft gas turbine does not deliver shaft horsepower but thrust. The is immediately apparent rvhen youlompare it with the other power units described
turbine merely drives the compressor, and the exhaust gas provides the thrust. and illustrated here.
8 PRINCIPLES OF OPERATION PRINCIPLES OF OPIiRATION 9
rotors, encased in a stationary housing.As a result, phases replace strokes, (strokes) with its own area. In other rvorcls, each area along the working
rotor seals do the iob of piston rings, ancl ports take the place of valves. surface corresponds to a part of the familiar four-stroke cycle.
Instead of a crankshaft, tlrere is a mainshaft, the iob of the crankpins The rotor leaves a certain arnount of free space between its faces
being performed by eccentrics. and the working surface throughout its movement. This space is continu-
The housing is not macle in one piece, but is built up around the work- ously changing in size, shape, and position. These volumetric changes
ing chamber. This chamber, shaped like a figure-eight rvith a fat waist- provide the pumping action for gas intake, compression, combustion and
Iine, contains a three-cornered rotor which moves around the working exhaust. Minimum volume in the working chambers is achieved at the
surface in an orbiting motion, all three corners always being in contact minor axis, which corresponds to the end of the exhaust phase and the
at some point. The working chamber is provided with ports for intake beginning of the intake phase on one sicle of the housing. On the other
and exhaust and a spark plug for ignition of the mixture. side, minimum volume corresponds to the end of compression and the
Wankel volumetric efficiency is Iargely determined by the position, start of combustion. Conversely, maximum volume is achieved along the
size, and shape of the intake port. The rotor, along its directed travel,
major axis. At the top end, this corresponds to the end of intake and
moves into positions which provide each of the four operational phases
the beginning of compression, while at the bottonl it signals the end of
the expansion resulting from combustion and the beginning of the ex-
The cross-section of the thrce-rotor Mercedes-Benz C-lll engine shows very clearly
the rotor phasing gcars, with 24 teeth on the stationary reaction gcar and 36 teeth haust phase.
on the inner ring gear in the rotor.
This four-cylinder, four-stroke Citroön DS-21 engine typifies design trends of the
Sixties. It has a multitudc of moving parts that are lacking in the Wankel engine.
The two-stroke engine has intake, compression, combustion and exhaust periods, from a separate crankcase conlpartment. Othenvise, all pre-compression would be
but the four stages are compressed into one crankshaft revolution, or two piston lost.
strokes. Thc four-strokc engine takes two revolutions. In a one-cylinder, two-stroke The two-stroke engine has twice as many power strokes in a given number of
engine, as shon,n, every rcvolution brings a power stroke. The plug fires every time revolutions as the four-stroke engine. It would seem logical that a two-stroke engine
the piston approaches top deacl ccnter. The explosion of the burned gases and the with an equal number of cylinders would produce twice the power of its four-stioke
intake of fresh mixture is squeezecl into the short time between the end of the c-ounterpart. In practice, that is not so. Because the intake and exhaust phases of
po$'er stroke and thc begirrrriug of the next compression stroke. Naturally, the the two'stroke are not clcarly defined and overlap with each other and with the
power atrd cornprcssion strokcs ilre orrly partial strokes. The two-stroke engine has
no valves. f'he pistons open and close ports as they move up and down. This
obviates the necd for the complicatcd valve gear and eliminates a potential source
of wear and noise. At the start of the two-stroke cycle, the rising piston uncovers
the intake port and draws a fresh charge of mixture into the crankcase, below the
piston. On top, at the same time, the piston is compressing the charge from the
1>revious cycle. Thc compressecl chargc is ignited and the combustion pressure forces
the piston down. During the power stroke, the piston uncovers the exhaust port,
allorving burned gases to escape. A transfer port in the block connects the crankcase
with a port placed bclow the exhaust port.
When the piston uncovers the transfer port, the partly compressed mixture in the
crankcase rushes to the combustion chamber. The piston then starts its upward
travel and compresses the charge, while fresh mixture is drawn into the crankcase.
Each nrain bearing is sealed so that each cylinder in a two-stroke engine is fed

This picture is considerably complicated by the fact that all three is considerable overlap all around, but Without detrirnental effects be-
faces of the rotor are going through part of the operational cycle simul- cause gas flow is controlled by rotor motion. The rotor opens and closes
taneously, whereas irr a piston engine all the action takes place on top- the ports and compresses and scavenges the mixture.
one "stroke" at a time. The result is that when one chamber is in the The rotor can be better understood by comparing it, again, to the
middle of the compression phase, the next is beginning the exhaust piston engine. Picture tlte rotor as being the pistons and connecting rocls
phase, and the one behind is well into the intake phase. This triple all rolled into one-if mounted on the crankpin, it would revolve rvith
action is possible because of the radial location of the various functions. the crankpin. If it were al.so to rotate on the crankpin, its centcr woulcl
Iior example, fresh mixture enters at "10 o'clock"; compression begins still describe a circle, but its corners would clescribe other t1,pcs of cun,es.
bcfore "12 o'clock," and reaches maximum at ") o'clock"; from ") If the rotor rotation were carefully timcd to the crankpin rotation in
o'clock" to between "7 and 8 o'clock" is pure combustion, until the specific ratio, the ctlr\/e could be preclicted ancl, if the ratio werc kept
ä't

rotor uncovers the exhaust port and the exhaust phase begins. The constant, the curve would be repeated for every orbit and every revolu-
exhaust phase overlaps both combustion and intake phases. In fact, there tion.
 PRINCIPLES OII OPERATION
t3
rvith the nrainshaf t. This reaction
ar on the rotor. If the reaction gear
gear must have 54 teeth, providing a
of the motion is performed by the
follows:
Rotor 54 tceth 3 revolutions
r
Gca )6 tceth 0 rcvol utions
Shaft 0 teeth I rcvolu tion
When the rotor advances 90 degrees, the eccentric advances
30 degrees.
Each time a rotor apex passes the intake port, the
mainshaft starts
another complete revolution, thus there is a power
irnpulse for every
one-third turn of the rotor, giving one power phrr.
for each complete
mainshaft revolution.

;.

!:

:'i
t,

1
.rt

..:

;)
,;

,l

..:

This cutaway shows the high-performance version of the three-cylinder,

Saab engine. Although it has f,ar fewer moving parts than a four-stroke t
:

is less efficient.
l

In actual operation, the pressure of expanding gases acts on one lobe

of the rotor to produce rotary motion. That explains why the rotor
revolves around itself. Gas pressure on one face turns the rotor, which
brings up another face, and the process is repeated. For the Wankel's I

rotating piston to exert Ieverage on the mainsh aft, it must act on a point i

away from the shaft centerline. The rotor is thus mounted on eccenhic
bearings to provide this leverage. Rotor rotation is kept in phase with .l

the rotation of the eccentric bearing by use of phasing gears. One part
of the phasing gears is a station ary reaction gear fixed to the end cover
 ROTOR
HOUSING
t
\- t\f oroR
The epitrochoidal chamber has
v min. maximum volume on one rotor
face when one apex coincides
Geometry with the rnaior axis (white rotor
with firm outline). The dotted'
outline rotor with one apex co-
iuciding with the minor axis
gives minimum volume on the
opposing face.

Tnn WeNxsr ENGINE has no direct

dimen'sions of
counterparts to the piston engine's t ore and stroke. The the resultant shape is callecl a hl,potrochoicl. If tltc point is actually lo-
the working charnblr are clictatecl by rotor wiclth, rotor radius,
ancl
cated on the peripltery of the gcneratiug circle, thc trochoicl thercby
rotor ....rrlricity. Rotor width is a rirright line, drawn radially
across
rotor center created is clllecl a cycloid. A cycloid is the cun'ecl path traced by, point
the rotor face, rotor radius is a straight line drawn from the
on the periphery of the generating circle as it rolls along a straight line.
to the rotor apex, and eccentricity is the distance from the mainshaft If tfue base is changecl from a straight ]ine to a circle and the generating
center to the rotor center (corresponding to crank throw
in a piston
by the rotor circle rolls around the outside of the base circle, the curve generated
engine). The shape of the workini chamb.r is determined
is called an epicycloid. The base circle must remain statiorrary and its
radius and the rotor eccentricity.
in a piston center must not be displaced-no rotation can take place. If, on the
T5e working surface can be compared to the cylinder wall other hand, the generating circle rolls around the inside of the base
inner surface of the
engine so far as function is .or,..inecl-it is ttre
circle, the curve generated is called a hypocycloid. But, if the point
the rotor, mounted
5ousing. As was nrentioned irr the previous chapter, chosen for a curve is movecl from the periphery of the generating circle
on eccentrics, both rotates on its axis and orbits around the
mainshaft
permanent contact with closer to its center, the curves generated will be quite clifferent. T'hey
center while the three rotor apices maintain will be, in this case, trochoicls instead of cycloids. A variation of the
is defined
t5e chamber working surface. The shape of the working surface trochoid, the epitrochoid, is formed if the point is not on the circum-
as an epitrochoid, but, in orcler to understand the
generative process,
A trochoiclal ference ancl it never meets the periphery of the stationary base circle
t5e evolution of the .pitro.hoicl begins with the trochoid. while the generating circle rolls arouncl the base.
curve is t5e path described by a .l-r.n point on or within
a circle as
Throughout these generative processes it must be kept in n'rind that
tSis circle ,oilr, without slip, around the perip-hery of another
circle
the center the relative size of the circles does not matter except for one condi-
until it returns to its starting position. During this rotation, tion-if the circles are not of eclual size, one must have a raclitrs that
points de-
of the circle merely clescribäs another circle, but all other is a whole multiple of the other's radius. In other words, frerctiortal
scribe other curves. It is these other curves that make
the Wankel en-
variations between the base circle ancl generating circle radii cannot
gine possible from a theoretical point of- view' exist. Ary atter:rpt to use a ratio of l:1.5 or l:2.2, for example, would
the
An epitrochoicl is the curve dÄcribecl by the chosen point when result in an irregular curve in which the chosen poirrt rvoulcl fail to re-
aty (base)
rolling (generating) circle is positionecl outside the station turn to its point of cleparture after one complete orbit.
circle. If the g.n.irting circle iolls arouncl the inside of the base circle,
 GEOMETRY 17

But, in order to create the geonretry of a true Wankel engine, a ratio

The drawing on top shows the between the two circles must be established.
curve described by a chosen point Make the generating circle radius one-half that of the base circle, and
on the periphery of a circle as
itrolls along e flat surface. Re- choose a point on its periphery. Roll the generating circle around the
move the flat surface and sub- base circle, arrd the path of the chosen point describes a figure-eight
stitute another shaped epitrochoid. The minor axis, in this case, is exactly twice the
circle) and rol
circle around it, base circle radius because the point touches the periphery twice on its
The chosen p tour; the major axis is four times the inner circle radius. The ratios
pincer movement. resulting are as follows:
Base circle raclius 1.0 unit
Gencrating circle raclius 0.5 unit
This family of epitrochoids has Half nraior axis 2.0 units
been obtained by changing the ITull major axis 4.0 units
position of the chosen point on
the generating circle. The inner-
most epitrochoid was created
If the chosen point is moved from the periphery of the generating
with the chosen point at the circle to a position at one-half the radius, and the generating circle is
periphery of the generating circle. rolled around the base circle, the curve generated is once again a figure-
The outermost epitrochoid was eight epitrochoid, but this time it has a fatter "waistline." The minor
crcated by a point close to the
center of the generating circle. axis is 7.5 times the base circle radius, and the maior axis is 3., times
The distance between the chosen the base circle radius. The chosen point never gets closer to the base
point and the center of the circle periphery than one-quarter of the base circle radius, and never
generating circle determines rotor
eccentricity for the resulting gets farther away than three-quarters of the base circle radius. It is this
Wankel engine. basic two-lobe epitrochoidal shape that represents the shape of the work-
i.,g chamber in a practical Wankel engine. Of course, the number of
If the radii of the generating and base circles are equal, and the usable trochoid shapes created using a generating circle radius half as
point chosen to describe the curve is located on the periph ery, a kidney- Iong as the base circle radius is practically unlimited. The chosen
shaped epicycloid is generated. A straight Iine drawn through the nar- point on the generating circle can be placecl any distance from the
rowest part of the kidney is called the minor axis, while a similar line center, and moving it toward the center of the generating circle reduces
drawn through the widest part is called the maior axis. In this case, the eccentricity of the rotor motion. Briefly stated, eccentricity in the
both circles being the same size, a ):l relationship exists between the generating circle corresponds to rotor eccentricity.
minor axis and the center of the base circle. The chosen point on the The ratio between the circle diameters is analogous to rotor gearing.
circumference of the generating circle touches the periphery of the base A two-lobe chamber demands a three-lobe rotor, w]rich, in turn, ne-
circle at only one point, and the rest of the way it moves farther and cessitates a 3:2 gearing relationship between the rotor annular gear
farther away from the center of the base circle until it reaches its and the reaction gear. If the inner gear in the rotor has 72 teetlr, the
apogee at three times the radius. T'he place where this occurs is diametri- reaction gear must have 48 teeth. Witltout this phasing the engine
cally opposed to the spot where the chosen point coincided with the cannot run.
base circle periph ery. Variations iu the geomctry for the two-lobe working charnber can be
If the two circles are the same size and the chosen point is placed created by shortening the raclius of the gcnerating circlc to one-thircl
at the halfway rnark on the radius of the generating circle, the resultant of the base circle raclius. Choosing a point on the periphery of the
curve is an epicycloid, kidney-slraped but with a longer minor axis than generating circle, roll the generating circle around out.side of the base
when the poirrt was on the periphery of the generating circle. This is circle. The chosen point will touch the base circle periphcry, not twice
because, in this instance, the chosen point never touches the base circle. but three times, on its tour ancl thc rcsultarrt epitrochoid will resemble
 This two'lobe ePitrochoid results
f rom using a generating
circle
with one'half the radius of the
base circle.

-l
\
\
\
I
,
I
I
/
/ --
/
l
I

Tlir epitrochoid results from \

using a generating circle of the
exact same radius as the base \
circle. \--

a three-leafed clover. A Wankel engine could be designed with a work-

ing surface of this shape, but it would require a four-lobe rotor.
Iror a three-lobe chamber with a four-lobe rotor, a gear ratio of 4:3
is necess ary (i.e., the reaction geu must have Zr% fewer teeth than
\-

-t'c/

the rotor gear) . If the inner gear in the rotor has 64 teeth, the reaction
gear must have 48 teeth to keep the rotor in phase. To continue this
geometric experimentation, set the radius of the generating circle at De6nition of terms: l, 2 and
one-quarter the base circle radiu§. Choose a point on its circumference This 7 = Rotor aPex. A, B, and C =
ated Workingchamber.a-Rotor
and the shape generated will be an epicycloid resembling B four-leaf one- face. b : Trochoidal surface. c
clover. Placing the chosen point closer to the center of the generating and d = Phasing gear§. e, f and
circle will bring the four leaves together nearer their tips, and the tips g = Rotor normal. h = Eccen'
will be brought closer to the stem. A Wankel engine designed in this lric shaft. i = Mainshaft center.
manner would need a five-lobe rotor. As the generating circle radius is
M =, Center for the rotor nor'
mals and phasing gear contact
shortened with relatign to the base circle radius, the number of rotor point.
lobes required to produce a workable Wankel engine becomes greater. c
Unless the rotor has one more apex than the working surface has cham-
bers, the rotor cannot be made to assume the duties of gas flow direction
and conhol.
The limits of practicality are soon reached, however. AII companies
tlrat are now engaged in development of the Wankel engine on an in-
dustrial scale have adopted the two-lobe epitrochoid.
The efficiency of the Wankel engine depends on compression ratio,
iust as in a conventional piston engine. The compression ratio in a
piston engine, however, is virtually unlimited-to raise it, the designer
 GEOMETRY
S
2I

has only to build up thc piston crown. Rut, in thc Wankel engine, the
compression ratio is linrited by the rotor radius and eccentricity. Once
these two dimensions are selectecl, the maxinrum possible compression
ratio for the engine is fixed. More specifically, the compression ratio
is not restricted by either the radius or the eccentricity, but by the

\r'l-
radius-to-eccentricity ratio. This is usually expressed as the R./e ratio
(engineers often refer to it as the K factor). Because rotor size and
- orbital path cletermine the shape of the workirrg surface, it follows that

t l'hD the spaces between the rotor faces and the working surface are dictated
by the same considerations. The combustion chamber is formed by the
trochoidal surface, the rotor face, and the end covers. This means that
the maximum and minimum volumes of these spaces are Iimited once
the basic design parameters have been established. Rotor width has no
theoretical limit, but if an extremely wide rotor was adopted, the result
would be slow and incomplete combustion. On the other hand, an ex-
tremely narrow rotor could bring the maximum volume of the working
. 25c chamber so low that the engine would not run. Rotor width can be
rU -j-\ The leaning angle of the apex
.ß seals is minimum at the maior

e) i-{l'.t \
\
and minor axes, and reaches its
maximum halfway between the
axes.R: rotorradius;e =
eccentricity; T = rotor apex epi-
+{:b trochoidal path; Ta = seal tip
ti epitrochoidal path.
/

\==L=-- -*z
.50n
.49
.10

2-LO8E EPITROCHOID WITH INN€R ENYELPF

EFtrECf OF nROCHO\D-,SHAqE Ay
Oh€RALL DÜ'I€NSION§ OF EN91NES
COI,,PR I§INO IOENTTCA L .51F1 EP| YOLUI''E

differences in maximum R. EPITROCHOID GENERATING RADIUS

o . EQUIDISTANT
'Hä'äLill'x*,xl',i; r . ECCENTRICITY
CURVE RAOIUS

e means eccentricity. The sliding

airs highest potentiai .r*prrrrioi Apex seal leaning angle is determined by rotor eccentricity and the
equidistant curve radius.
 GIIOMETRY 23
Jrt I

engine. With 5igher K factors, Wankel engines can be designed with

1e I
coäpression ratios higher than is permissible with present-day fuels.
l-
I

F or instance, a K factor of I I .5 allows a maximum compression

ratio
of 30:1, a7,l K factor allows 18:1, and a 3.9 I( factor l0:1. The
2. engine
§'
I

f
«loes not always run, of course, with the full comPression ratio per-
mittecl by its clesign (bccause of combustion charactcristics), and the
actual compression ratio clepends on design of the cavity in the rotor
face. The äesigne, aims for the highest compression ratio comPatible
,22
with fuel anti-knock ProPerties.
§
ü
grnar'h|,e o p nar'56tto
§

R
§
§

a
I Eptrxhoiden rnlt

6
Felix Wankel has made a complete study of possible rotary engine
It configurations. These four are described as epitrochoids with inner
.nu.iop.r. In each case, the rotor has one lobe more than the
? envelope. The I zZ configuration (left) permits compres-si^o1 rltios
'leaning
up to äbout 140: I with äpe* seal angles. up to 19.5o. The
& 4 z6 28
o 2i 3 configuration (sccond fronr lef t) allows a maximum compressiorr
ratio of i S.5 t aira lcaning anglcs arc liuritcd to 30". The 7zl
lEANtNo ANGLE t,I configuration (third fronr lef$ cannot reach higher qomp_lession ratios
'
--> 41.8o. The 125 con-
' -loware very high:potential
than-7.5:l ,tia leaning angies
figuratiorr (right) has uery -(6.0: )
compression I and
2- tog€ EP|TROCHOL0 uir.c.ptabiy1riätr leaning angles with a maximum of 56.1o .
AAtoLE $I
il VELY
USEO ON AN €N6INE ,ttZE OF fOonwt fue fHE
tlA"oR Arß
Thlr qr3nh shows the relationship between maximum
- leaning angles, R/e ratios
and highest potential compression ratios.

compared to cylinder bore-it is manipulated to gain the most

reasonable
combustion chamber. Most clesigners choose rotor width approximately
equal to one-half the rotor radius. These configurations
- are characterized as epitrochoids with outer
A low R/e ratio (ot K factor) gives the highest compression ratio, .rr.räp.i. t[ri menns thc rotor is an epitiochoid as well as the
which means a small rotor radius ancl high ...intricity. The Iower the ;;;i;;.. Aff of thenr permit conrprcssion ratios in excess of 100:1.
The l:Z.onfrgriation'(left) has ; low maximum leaning angle of
K factor, the shorter the minor axis of tlre epitrochoid. A longer radius 19.5o. The oit".rr, reading fro* left to right, have leaning angles of
means a Iarger rotor, and bigger overall äi-ensions for t[e 30o, 1I.8o and 56.4'.
whole
24 DESTGN GEOMETRY 25
Just as the R/e ratio- sets a high compression ratio limit, it also sets
a ]ow ]imit. Building the engine down io this Iimit rimental to sealing effectiveness as well as to seal durability. A low K
will gi,ve t6e most factor does bring certain advantages, however, such as low bearing
compact housing dimensions. But, Iow R/e ratios give rise
to another Ioads, a low surface-to-volume ratio, and large bearing surfaces in Pro-
problem-because the "waistline" of the workini chamber
will be portion to rotor dimensions. High K factors also have their strong points,
sharply marked and the overall dimensions small, the apex
seals will
be subiect to fat greater changes in angularity (relative among which are greater rotor cooling capacity, more available sPace
to the working for side seals and oil seals, lower oil seal rubbing speeds, smaller aPex
surface) than would be the case with higtr.r R/e ratios.
This is det- seal swing angles, and reduced sensitivity to exhaust/intake phasing
overlap. With radially disposed apex seals, the maximum swing an-
c
gularity of the seal on the working surface is 15 degrees in an engine

NNW%
?mor t Q§ ?oer.lg|o gmar.g0c gmor- 410c

with an I 1.5 K factor, 25 degrees in an engine with a 7 ,l K factor, and as

high as 50 degrees in an engine with a 3.9 K factor.

€ttq(
llypolruiotdm itt tnmr Hi)ttftgur
These configurations are hypot
h-ypotrochoid curve is traced ty
slipping, inside the base circie
compression potential. The l:Z c

co_rnpression as hi8h as
?'. The 4:5 configuration
-ratio (righ$
l0'4: l, but iti leanirg'ailglä
goes up to *5,

l,Z I,J ilbcrsclzung J + 1,5

The.se. sh-apes
used in |re hypotrochoids with outer envelopes.
these four e.x3rnprgs correspona .-r.tty'i"
- The rotors
the four hypotrochoids with inner är,u.iop;; ,h;;
all of them. The
:aning angle limi
 DISPLACEMENT 27

who maintained that the unit volume of one chamber must be multiplied
by the number of faces on each rotor because each face goes through
the same full operational cycle that takes place in each cylinder of a
reciprocating piston engine. This led NSU to cornPare the Wankel
engine with a single-cylinder, two-stroke piston engine, which has the
same power-stroke frequency in relation to crankshaft revolutions. Both
have one power stroke for each revolution of the output shaft.
Indeed, a single-rotor Wankel engine does share certain operational
.characteristics with the single-cylinder, two-stroke piston engine. These
include mean effective pressure, torque, horsepower and r.P.m. How-
Displacement ever, it is wrong to rate them on a sinrilar displacement basis, because
the Wankel engine operates on the four-stroke cycle. It is fairer to
compare the Wankel with a two-cylinder, four-stroke piston engine,
which also has one power stroke per crankshaft revolution. A single-
rotor Wankel engine goes through two complete operational cycles in
two mainshaft revolutions. A single-cylirtcler piston engine requires two
crankshaft revolutions to go through the four strokes that nrake uP its
Tnnnn ARE THREE pRrNcrpAL REASoNS
*hy it is important to calculate the displacement of an engipe: operational cycle. Two-cylinder engines do not complete their cycles
l. Displacement provides a basis of comparison with äther engine any faster, but do producc twice as many power strokes, therefore two'
types (specific output). Better methods probably exist, such as cylinder engines deliver one power inrpulse for every revolution
rating horsepower against weight or horsepower against fuel con-
sumption Per hour, but the industry is conditioned to its own ocating piston engines and Wankel
definition of specific power output; horsepower per liter (or per n engine at left is 250 cc. Two such
with comparable firing and torque'
cubic inch) di.splacement. -rotor Wankel engine shown in the
2. International motor racing rules include engine displacement
Iimits, therefore an equivalency formula for the Waniel engine B
"11'ä:lii, :'. *:,i:li1 #fl?:,1"f. Ifl
I 5: l, and the nraxirnum seal leaning angle
is necessary to enable Wankel-powered cars to compete. This has is 2 5.5' .

Iong traditions, although formulae based on weighl, piston area,

and fuel consumption have been hied at various times.
3. Several countries have imposed a tax on motor vehicles with en-
gine displacement as its basis.
When NSU started production of a Wankel-powered car, the German
authorities made an attempt to tax it under the existing cylinder dis-
placement system. Because each combustion chamber held SCjO cc., NSU
said the engine should be rated as a 500 cc. engine. The German tax
authorities said the engine had to be taxed as a I,500 cc. unit, because
the rotor has three faces and there are three power impulses for each
rotor revolution. The same thing happened in the rr.ing world. In
1962, NSU submitted proposals to the FIA and FIM ,rguing in favor
of rating the Wankel engine on the basis of the volum. of , single work-
rll (ü
,O'A fE

ing chamber. They received the support of the German and British , Ctl SofalJ I t lOC ctst 2 I 250cal
delegates but \t'ere opposed by the representatives of France and Italy D/5-t ?-'Ztl'i ß'OJ i7.4l'o E- 'Jt5'i ß.N,Ir(t r
sfi - Jll ta §lY-ltJr{ trY .{tr{
rrqE .-, I-
- -

-qF
 DISPLACEMENT 29

A comparison between the single-rotor Wankel engine and the two-

cylinder piston engine is as follows: a power stroke in cylinder number
one corresponds to a power impulse against number one rotor face.
At the same time, cylinder number two is taking in a fresh charge and
rotor face number two is compressing fresh gas while face number three
has iust completed an exhaust phase. When face number two enters the
power phase, cylinder number two in the piston engine begins its power
stroke. When firing occurs on face number three, firing again occurs in
cylinder number one. The torque fluctuations and volume variations in
the two engines are identical, and completely in phase with regard to
degrees of crankshaft ancl mainshaft rotation. Ilven with these questions
settled, äD equitable displacement racing formula for Wankel engines
eluded the sanctioning bodies.
From time to time, moves were made in racing circles to circumvent
the problem by abolishing the displacement limits. It has been sug-
gested that a combination formula, stipulating both minimum weight
and maximum fuel tank capacity, could be the solution. This would
allow cars powered by piston engines, Wankel engines, gas turbines,
and steam engines to compete on realistic terms, with variations in
fuel tank capacity allowed according to the type of fuel usecl.
Gas turbine cars have completecl in the Z4-hour race at Le Mans and
the Indy 500 race, ancl equivalcncy forurulae, based on air consunlptiott,
have been worked out to produce a rclative cylindcr displucernent rating
for them. The Commission Sportive Intemationale of the Federation In-
ternationale de I'Automobile (the international governing body for
motor racing) then looked into ways of applying the same rule to
Wankel-powered cars. With a single rotor and a chamber volume of
500 cc., the intake air volume of a Wankel engine is only 500 cc. per
mainshaft revolution. That means it is equivalent to a one-liter (1,000 cc.)
four-stroke piston engine. A two-cylinder, one-liter engine will complete
Others argued that the Wankel engine represents a three-cylinder, an intake stroke during one crankshaft revolution or, an equivalent of
four-shoke engine, maintaining that positive torque is produced over 500 cc. A four-cylinder engine, with 250 cc. per cylinder, has only two
240 degrees of mainshaft rotation, just as a three-cylinder engine pro- intake strokes during one crankshaft revolution, and again there is
duces torque over two-thirds of each crankshaft revolution (it goes 500 cc. intake air volume. These facts influenced the CSI to adopt a
through three cycles in two revolutions). This argument was effectively formula which rated the Wankel engine's displacement at twice the
Put to rest when the leading independent expert on Wankel engines, combustion chamber yolume multiplied by the number of rotors, This
Richard F. Ansdale, delivered a paper entitled "Rotary Engine De- formula is still in effect.
velopment and its Effect on Transport" to the Society of Automotive
Engineers of Australasia (October 15, 1968), in which he proved con-
clusively that the torque characteristics of the single-rotor Wankel
engine correspond to those of a two-cylinder, four-stroke piston engine
and not to those of a three-cylinder engine.
 SEALING 3l

Piston rings also aid heat flow fronr the piston and Prevent excess lu-
bricating oil frorn reaching the combustion cltanrber. All rings contact
the cylinder wall around their entire periphery. They also have circum-
ferential sicle contact with the piston grooves they are in (except under
transient conditions). Most modern pistons have three comPression
rings and one oil control ring. The rings act as radially exPanding springs
against the cylinder bore. The most common piston ring material is
cast-iron-it's sprirrgy ancl weärs well. Cltronre-platcd rings redttce bore
wear, but chrome plating is expensive. I{owever, many engines use rings
with a certain chrome content because chrome has two important ProP-
erties: it's immune to corrosion, and far harder than iron-chrome
Sealing melts at 3,407oF. It also has a, beneficial burnishing action on the cyl-
inder walls, because it smoothes down peaks arrd high spots instead of
tearing them off. Chrome is itself Iong Iasting and also gives longer bore
Iife.
Most present-clay engines have piston rings with a high molybdenum
content. Molybdenum is a tough, slippery, silver-white metal that has
Tnn Mosr cRUcrAL of all the prob- a lrigh melting point (4,770oF,) and offers high resistance to scuffing. In
Iems that beset the Wankel engine during the early stages of its cle-
adclition, it resists corrosion and has a high break-in or seating potential.
velopment concerned sealing. To obtain efficient operation, the working
Its abrasive-wear resistance is rnoderate (between iron and chrome), it
chambers must have seals that prevent gas leakage between them. There
has good scuff resistance (its greatest aclvantage), and is also somewhat
are several paths that must be blocked-across the apices between the
porour and capable of maintaining an oil film better than chrome-plated
rotor faces, and around the sides of the rotor faces.
or plain cast-iron rings.
The sealing performance of the Wankel engine is determined by
many variables of engine design, such as basic configuration, material
engine is considerably
and dimensions of sealing elements, Iubricating conditions, rotor and
housing cooling conditions, and precision of machining and finish. When lTiJ',x,lfllrJ.":'"j
compared with the reciprocating piston engine, the Wankel epgine fias
lems are assravated. At right, the seating line fi :tnis; ffl:i'grtt?l:
handicaps in its gas sealing mechanism. The length of the gas leakage Wankel engine.
paths, tJre size of the oil seal, the number of clearances, the configura-
tion of the seals and their operatirg conditions constitute serious dis-
aclvantages. It is also undeniable that the gas sealing efficiency of tfie
Wankel engine is unfavorable, especially at low speed because t5e
angle of each phase of the Wankel engine is 1.5 times Iarger than tftat
of the four-stroke reciprocating piston engine.
Sealing consists of providing uninterrupted contact over tfie whole
primary ancl secondary sealing areas. The prinrary areas are between a
.scal ancl a statioltary surface, the seal being canied in a moving com-
Ponent. Secondary sealing areas are between the seal and its slot, groove,
or bore. Sufficient force must be applied to the sealing element to
maintain contact. Broaclly speaking, the rotor seals in a Wankel engine
perform the same duties as the piston rings in a reciprocating engine
because they provicle a seal for the conrbustion gases.
HKM KKM KKM
)x500 1x500 )x500
 RECIPROCAflNO EN6.
SEALING 33

NCIUOINO §€AI IPOUND ruRrS seals were eventually devised. Sealing the side seal corners was easier.
Sirrce three of them are carried on thesame surface, they intersect

t
\ KKI,I
near each apex. The problem was solved in clifferent ways by the
various manufacturers, but all solutions were based on the same prin-
ciple of blocking the gas leakage path by interlocking the sicle seals to-
gcther at al} corners.
§
{ Both piston rings and Wankel engirre rotor seals nrust be Iubricatecl
to prevent sticking. Both rely on gas pressure to form a seal rvith the
{§ opposing surface, and both are also spring-loaded to ensure contact at
all times. If the rotor apex seals were not spring-loadecl, resistance to
§ cranking would be so high that a special high-capacity starter motor
PECIPR, EMO. PISTON ClrcUr,E ONIr would be necess ary. At the operational speeds of the Wankel engine,
§ the spring-loading is of no value-it is only to help start the engine.
§ The top compression ring in a piston engine is backed up by another
ring which is intended to trap blowby gases that manage to leak past.
§
{ 6 Even the oil control ring nrakes a contribution to gas sealing. As a
\a result, the piston engine has the benefit of a three-stage gas sealing
N I system, which is not feasible in the Wankel engine. The Wankel can
R have only one seal; neither apex seals nor side seals can be backed up
by r second line of seals. This is one of the key reasons why the sealing
2
I problem was so serious, and why the solution had to be found in
specific areas-at or near the apices and edges of the rotor. However,
even assuming satisfactory sealing systems have been devisecl, there
Pü,9 35 always will be the risk of gas leakage across the exhaust and intake
-> tEtNtNo tNotE ["1 ports, as in a two-stroke piston engine. (This corresponds to valve over-
Irp in a four-stroke piston engine.) At very high rotor speeds, a gas
FOR *"

Sealing line length comparison between a single-cylinder lr., cubic

inch piston engine and a single-rotor Wankel ängine KKM.250. The
dotted- Iine repre-sents the sealing line for the piston engine including 0r
the valve seats; the solid line represents the piiton ring iircumfer.n.ä l.
only.

Because of the triangular shape of the Wankel engine rotor, it cannot \

use seals of the ring tyPe. Rotor seals are of the strip type, some straight
and some curvecl. Apex seals are straight strips inieried in raclial slots These are the forces that act
at eacll rotor aPex. Side seals are curved, to follow the curvature of the on the apex in a Wankel
seals
engine. C : coriolis forces. F :
rotor flanks, and are inserted in segmental grooves in the rotor sides, as centrifugal force. G
close to the egge as possible. However, each seal strip has two ends, sure. GR : - gas pres-
radial gas pressure.
facing Perpendicularly from the direction of seal loading, and special Gr. : lateral gas pressure. R -
corner seals to block the Ieakage path around the sideJ of the apex
friction. rd : gAS pressure in
percent of atmospheric.
74 DESTGN
Ieakage path also may be created at the constriction on the working
sttrface due to reversed centrifugal loads on the apex seals. This form
of leakage is equivalent to the effect of valve bounce in a four-stroke
piston engine.

30
20
t0
0
travel ancl are subiect to very high acceleration. The sliding velocity of t0
20
tJre aPex seals varies with rotor position, even at constant mainshaft 30
r.P.m., because of the duplex rotation of the rotor-it rotates on its
own axis while orbiting around the mainshaft. For example, at 5,000

This graph shows the magnitude of the forces acting on t!. 1p9x
seals. F: centrifugal foäe. C : tangential acceleratiqn (coriolis
ioi.C1. yo = leanirlg angle. (O"q a-pef seal _strip in this example
*.,gt r änly 2 grams.-Rotör radius is fixed at 65 mm. and eccentricity
is 915 rn*. Rötational spced is a steady 10,000 r.p.m.)

K -769
lr*t.212 ;lb Pmar.- 23o
€n. ' 197
€ '8,5
Vp - Q597xVc
.,,, \ Pmox.' 6070 kg (50 ot)
d5 '1lmm
/ 3l mß.ßOOO
ß*r1 -Üt^in
V min.- 13,6 m/st-

\
)---+
,l K - 566'
/ 9r?)O J3
moJ.
t

€n, ' 11§

e -45
Vp - 0,117 xVc
Pmoxi 5110 kg (50 ot)
ds - 15mm
ßr;. 226 m/s16O0O
J
I
Wnif, Q5 m/s tlmin
,CKHN
g.tu|ß yEtnntEs
cd|Pllfrttü fi,rnfü{o corra. Etü. A^tO This comparison illustrates the influence of the K factor (R/e -ratio),-.leaning angle,
lgcrPßocAlfiüt cilc. 4,f,Ft AcEt ENT ße.b wür, compressiän ratio, eccentricity and main bearing diarneter on the sliding velocity of
itr ip.* seals. K : R/e ratiä. R - rotor radiui. e - rotor eccentricit_y. Q = com'
A simple comparison between sliding velocities in a reciprocating pi.ssiän ratio. Gtr - potential compression ratio. /o,rrx : maximum lea-ning angle.
piston engine and a Wankel engine. V, : maximum chaÄber displaceÄent. P-.' - combustion pressure. d. = main
V. = apex ieal sliding velocity (min and max).
bearing diameter.
36 DESIGN SEALING )7

compatibility-with a harcl working surface and soft seals, the

seals
r.p.m., the apex seals in Curtiss-Wright's RCZ-60 have sliding velocities
however, wear the working surface- Both
witlrin a band stretching from 42.5-108 ft./sec. Compare this with the wear out quickly. Harder seals,
ma-
figures for some piston engines, all at 5,000 r.p.m. NSU and Toyo Kogyo were conficlent they were using comPatible
contact-materials that required a mini'
terials for surfaces in rubbing
Engine Minimum Maximum
mum of lubrication.
I-Ionda C-92 0 66.8 f t./sec. It was clear to the test engineers in both Germany and |apan that
Chrysler V8 (31 8) 0 $.5 ft./r... the chatter marks occurred because, under certain conditions,
the apex
Mercedes-Benz 220 0 62.6 ft./sec.
seal configuration, movement, and construction broke down the oil film
contact. Ob-
The forces that work on the apex seals are centrifugal (positive and on the rubbing surface, resulting in direct metal-to-metal
viously, reasor,rUt. seal life ca
of
negative), gas pressure (both directions), and friction against the work-
the hijf,-speed metal-to-metal
in
ing surface. Negative centrifugal force is formed because during the
three ways: first theY reduced t . €s'
approach of the rotor apex toward the minor axis (the waistline of the
figure-eight), the apex is moving closer to the mainshaft center-not to dampen the rocking motion of the seals without causing seizure;
then they inserted flat springs of beryllium bronze under the seals
to
radially, but in an easy curve. This motion is opposed to the normal
plated the
centrifugal loads provided by rotor rotation, and thus constitutes a brace them against the working surface; finally they chrome
negative force. working surface to reduce the friction that was a partial cause
of the
The apex seal is perpendicular to the working surface only when at problem.
the maior and minor äxes; at all other rotor positions, it is at an angle.
This "leaning" angle of the apex seals against the working surface These are three rotor oil seal versions cleveloped !v NSU,
presented
varies during the motion of the rotor. On its way from minor to major in consecutive order. Their featurcs are described in ttre text.
axis the seal adopts an increasing leading angle up to the halfway point,
then the angle is gradually reduced until the apex seal returns to its
perpendicular position, relative to the working surface, ät the major
axis. On its way from the major axis to the minor axis, the apex seal Druckholle'
runs at a trailing angle throughout. Maximum angularity is reached at ventll _ _<
the halfrvay point between the axes. The maximum "leaning" angle in
a particular engine is dependent on the raclius/eccentricity ratio. The
angle gets higher with lower R/e ratios. Extreme angles (from per-
pendicular) are detrimental to proper sealirg, ancl therefore are to be
avoidecl. It is generally helcl that 30 degrees is a practical limit.
The process of lubricating the gas sealing elements of the Wankel
engine is fundamentally clifferent from lubricating the piston rings in
.TI

a reciprocating engine, and is considerably more complicated. During

the initial stages of development, both NSU in Germany and Toyo I

Kogyo in |apan tested Wankel engines using oil mixed with the fuel in
proportions of 50:l or 100:1. Actually, there was no need to mix oil
u,ith tlre fuel, because tJre rotors were cooled by internal oil flow and
there was considerable oil leakage frorn the rotor. This amount of oil
Ieakage seemed to be sufficient for lubricating the seals and the working Ausfithrung C
Ausführung A Ausführung B
surface. Despite this presence of lubricating oil, asymmetrical wear of
the apex seal tips and chatter marks on the chamber working surface
were problems. Seal tip wear is, first and foremost, a matter of material
 38 DESIGN
The rocking motion of the apex seals is a phenomenon sinrilar to
piston ring flutter under extreme piston acceleration. It is brought
about by a combination of seal tilting during reversals of gas pressure Assembled version of the 1966
ancl radial movements of the seals caused by the difference in radial ex- sealing system develoPed bY
NSU.
pansion of the working surface. Freedom of radial movement is, of
course, necessary to prevent the formation of deposits. The apex seals
have a clearance of about 0.001 inch in their slots. When the pressure
in the working chamber is highest, gas pressure on the apex seals works
Dld*trlo
to increase the cbntact pressure of the seals against the working surface, Ecrtrl -<
while friction tends to produce a tilted position of the apex seal under lrltrilrdrr
low-pressure conditions. The apex seal tip is held back by its rubbing Dld*8obrt
Daoa*Eoarotrdrr
on the working surface, while the seal base is carried ahead by the slot Dlol*Surlhnhdl

in the rotor.
The apex seal has three consecutive positions during the pressure
reversal that occurs when combustion pressure from the chamber lead- Sealing system finally adoPted
irg it overcomes the compression pressure from the chamber trailing it. for the production version of
the KKM-612 (Ro-80). Dcht'
On its way towarcls the minor axis, the apex seal is standing up against Leiste : Sealing strip. Eckteil =
the leading wall in its groove, and is held in that position by compres- Corner triangle. Leistenfeder =
Seal spring. Dicht'Bolzen =
sion pressure. When combustion begins ahead of the apex seal, gas Trunnion. Dicht-Bolzenfeder =
pressure acting against the exposed part of the seal tilts it, forcing the Trunnion spring. Dicht-Streifen'
tip back towards the trailing edge of the groove. Continued combustion feder : Side seal spring. Dicht'
pressure brings the bottom of the seal over to the trailing wall of the
Streifen = Side seal.

Evolution of the sealing system. These examples are based on NSU practice. In the
1959 version (left), the apex seals had fragile corners because the number of Grenze unbekonnl
components was excessive. In the 1963 version (center), the apex seal itself had 600
adequate durability but did not provide proper corner sealing. In the 1966 version
(riglit), the floating triangle corner of the apex seal provided requisite corner sealing §roo
without involving a degree of complexity that would create other problems. q,

§§ eoo
{
0

/..RE|ANE? Plil t0

azSPRNO R 20
-\
§a
30
G
§vl 10
§ s0
SEAt
60
Juli 62 Jon.63 Juli 63 Jon.61 Juli 61 Jon.65 Juli 65

Life of NSU carbon seals improved in sensational leaps over a three'

'Wear is measured in microns per hour and test duration
v.ri span.
ii*. ii in hours. The engines were running at a constant 5,000 t.p.m.
at a mean effective pressure of I02.1 psi.
40 DESIGN SEALING 1I
groove and holds it there until the exhaust phase begins in the Ieading usually showed up as sharp pressure fluctuations, followed by low-
chamber and compression pressure again drives the seal to the leading frequency pressure waves within the intake ports.
wall of the groove, where it stays throughout the combustion process. Toyo Koryo decided against a metal-to-metal interface, for which
Control of this rocking motion of the apex seals was achieved by cutting Iubricatiorr would be a vital requirement. They adopted a specially de-
gas Pressure relief slots on the lower edge of the apex seal to reduce veloped carbon compound as a new material for the apex seal becatlse
the lag before the higher pressure can reach the base of the groove and carbon has self-lubricating properties and the presence of an oil film
assist in the sealing. This modification was also effective against a phe- is not absolutely necess ary. Flowever, because the Iubricant in the in-
nomenon called "spitback" that occurred across the apex seals under ternal combustion engine also serves as a gas sealing agent, it became
conditions of very rapid pressure rise in the working chamber. Spitback necessary to study the lubrication of the apex seal in terms of gas sealing
efficiency in acldition to the problem of friction and wear.
In addition to gas sealing, there are oil sealing duties. Because the NSU pointed out that carbon seals must be at least 5 mm. (0.197 I

rotor is oil-cooled in nrost Wankel engines, the oil has to be sealed inch) thick to have sufficient strengtl'r. Metallic apex seals need be only i

inside the rotor and prevented from escaping into the working cham-
1.6 mm. (0 .063 inch) thick to be effective. Tests showed that the pres-
bers. The seal has to be positioned at the btaring edges because the
rotor bearing itself needs an oil film for lubrication pürposes. ence of a lubricant still had a favorable effect on the gas sealing Per- I

formance, although the carbon apex seals formed a geometrically linear

ROTOR COOLINO OIL SCHEUATIC
contact with the working surface.
There are three methods of supplying lubricating oil to the apex seals
of the Wankel engine: one is to mix oil with the fuel mixture (tt was
done in early tests); the second is automatic metering of lubricating
oil from the rotor side; and the third is an independent feed system by
which oil is introduced into the intake ports. The oil-fuel mixture
method would have worked if the necessary amount of oil had not been
in constant proportion to the fuel flow. This method has been reiected
as unsuitable for modern automobile engines. The second method also
was reiected, because it is almost impossible to accurately control the
amount of oil leakage from the side surfaces of the rotor under all en-
gine operating conditions. Independent feed systerns have been success-
fully developed, however, and they supply the necessary amount of oil
to the intake port according to engine operating conditions.

EUPPLV COOL r6
0lL rC
f,ri
El
.d

H
#l
ri
 THE COMBUSTION PROCESS 43

4 times the initial pressure). Combustion pressures are high at wicle

open throttle and low during low-speed cruising or idling conditions.
Under high pressure, combustion is speeded up. A controlled swirl or
turbulence in the combustion chamber will also speed up the flame
front. AII types of combustion chambers can be given some amount of
turbulence, but too much turbulence is undesirable because it increases
the rate of heat loss to the cylinder walls. But the faster the flame front
travels, the smaller the risk of abnormal combustion (knock, ping and
rumble).
Knock occurs when end gas ignites spontarreously towards the end
The Combustion of the power stroke. The flame front s1>reacls out from the plug ancl
the mixture burns properly and evenly. Then, heat and pressure fronr
Process the beginning combustion work together to heat and compress the last
portion of mixture before the flame front reaches it. When tempera-
ture and pressure become too great for the unburned mixture, it ignites
but doesn't burn smoothly. This explosive burning may increase pres-
Tnn \MeNxnr ENcTNE RUNs on the sure 200-300 psi above the pressure existing in the rest of the combus-
same gasoline that is used in most reciprocating piston engines, and its tion chamber. The sudden pressure rise spreads throughout the cham-
combustion Process is fundamentally the same. Air/fuel ratios are simi- ber, and causes high-frequency pressure fluctuations which produce a
Iar in both Wankel ancl piston engines, ancl compression ratios are knocking noise, The intcnsity of thc knock clluse.s the conrbustion chartt-
closely comParable. To understand the differences, a comparison of the ber walls to go into sympathetic vibration, aud this gives rise to the
basic combustion process as applicable to both types of ängine is nec- familiar "ping" sound.
essary. Knock is often referred to as "detonation." It is unlikely that even the
The air-fuel mixture of the internal combustion engine is ignited in highest intensity of knock occurring in an engine in service reaches the
the combustion chamber by a spark plug firing at a pre-set point, start- level of a true detonation wave as found in laboratory tube-type ap-
ing a wave of flame spreading out from the sprrk plug. This flame paratus, so knock is technically a more correct term. The pressure fluc-
front continues to move across the combustion chambär until it reaches tuations also cause a heat loss from the combustion chamber to the
the other side and the cornpressed mixture is burned smootSly and surrounding engine parts which may result in a power loss. When they
evenly. Flame front speed varies from 20 ft./sec. to over 150 ft.1sec. combine with surface ignition, the pressure fluctuations produce sharp
This speed depends mainly on air/fuel ratio, compression ratio, mixture cracking noises and the phenomenon is knorvn as "wild ping." In addi-
turbuJence, and combustion chamber design. Flame front travel is quite tion to noise and power loss, knock leads to overheating of valves,
slow when the mixture is too rich or too lean. When the same amount spark plugs and pistons. Overheating prornotes further knock and may
of fresh mixture is compressed into a smaller space combustion will be also Iead to pre-ignition-the two phenomena are mutually provocative.
quicker and more of the heat value in the fuel charge will be utilized, Overheating also severely shortens the life expectancy of valves and
thus the comPression ratio is a direct indication of efficiency. The pres- spark plugs.
sure gain with a 5:l comPression ratio is about 300 psi. T[e pressure Surfacc ignition is causecl by one or more hot spots irr thc cornbustion
gain with a l0:I compression ratio is about 700 psi. More *oik must chamber that ignite a portion of the mixture before the gas is properly
be done to comPress [t,. mixture when the compression ratio is hig5, ignited by the flame front. Such hot spots may be glowing combustiorr
but there is a large gain in thermal efficiency ancl power because less chamber deposits, poorly cooled areas, sharp edges on the head or
heat is lost to the cooling system and more of the fuel energy is put to block, an overheated spark plug electrode or an exposed spark plug
useful work. If the initial pressure at the moment of firing is fOO psi, thread. Surf-ace ignition can occur either before or after the plug fires. If
the final pressure will be almost 600 psi (normal pressure rise is 3.5 to it occurs before, it is called pre-ignition and corresponds to random over-
 ++ DESTGN - ?0ouO.T
600 v.0.7 -...- 20on.el-l
10onl.Tf
advance of the ignition timing. Pre-ignition opposes piston travel on the -.-----.--104.o.7
60on0.[--
The top graph shows gas velocity -"'-"'-21ouO.T.
comPression stroke. The force of too-early combustion may burn in e working chamber of sym- ---8Pn.O.l
through the piston crown, melt the cylinder walls, sever valve heads or metrical design at 6,000 r.p.ffi.,
crush valve measured in meters per second. -O.I
_springs. When surface ignition occurs after the firing of (One m. sec.: 3.06 ft. sec.)
the spark PIug, total combustion time is reduced and the ,erult is The line code indicates rotor
Power loss. This form of combustion is called post-ignition, which position. 80" v OT : 39" BTC
(before top dead center) . 20" n
often leads to "rumble." Rumble (or "pounding;') is unique to re- OT
ciprocating engines and probably has no counterpart in Wankel en- - 20" ATC (after top dead
center). The lower graph shows
gines. It's a deep rattle that emanates from the crankshaft. The noise the same curves for e working
chamber of non-symmetrical de-
is due to vibration of connecting rods and crankshaft at a cycle time sign developed to increase gas
aPProaching bottom dead center, set off by abnormal combustion. velocities.
Rumble occurs with unclean engines of medium compression ratios as
-----Eot o.r-
a result of secondary f ame fronts during combustion. Once begun,
_
itrv.Of -
rumble tends to go on. If caused by surface ignition, it can disappear on -.lPu Of --
full-throttle acc€leration because the additiänal gas fow and
-ir.rrur. 2ey.O.f -_

will eliminate glow and speed up flame front tiavel. Rumbl; is not - 2trn.O.f -,
confined to any one tyPe of combustion chamber-it can exist with all -Ol
----10"n0^tJ
current designs. -- --6 0o nO.l l-
-r
---$0ono.f

These graphs compare the volumetric r'xTffn:"r"ff

,H;'r§,r.::fll.,#,,$u:li:i
: ignition. Auslass :
;
engine. Zündung
:'-:'1";'ä,#,t:§l;fol
tr",ffi,;
g. Kammervolum Vk
- chambei
nt. oT : toP dead öenter. uT :
mpression chamber volume.

When an engine continues to run after the ignition has been turned
a-t'
off, the engine is "running-on" or "dieseling." Running-on is an erratic
process which may lead to sudden reversals of rotation and consequent
mechanical damage. It can be caused by the idling speed being too high,
air/fuel mixture too lean, ignition timing too late, or fuel octane rating
too low. Some sources believe running-on is a case of surface ignition,
others maintain it is a case of compression-ignition. The condition can-
not be cured by changing to a spark plug with a colder heat range.
Gas transfer should help speecl up the flame front travel. One would
expect that the gases ignited first would ride the crest of the gas trans-
fer wave, and that thc flanre front woulcl travcl ahcad of the gas trarts-
fer. Papers relcasecl lry Curtiss-Wright, howevcr, inclicate that irtstcncl
of being additive to the high gas transfer velocitics, the flame front
I

I never qüite catches up with the gas transfer wave, and never reaches
all the combustible mixture. There is, therefore, an "end gas" problem
 TFIE CON,IBUSTION PROCESS 17

This chart shows the time taken continuous. Because of the rotor's motion and the ever-changing shape
from the moment of ignition and volume of the combustion chamber, the combustion process is
to the beginning of ionization
measured in milliseconds, meas- quite different.
ured over a 2,000 to 6,000 r.p.m. Wankel engine research has given rise to a new term- gds transfer
range. The dark "A" field shows
the spread obtained between po-
velocity. The combustion chamber itself travels along the working surface,
sitions 2 and 1, i.e., iust before pushed from behind by the rotor. The fresh charge travels irr "pockets"
the minor axis and about 60 between the rotor face and the working surface. 'Vv'hen a pocket widens
degreesafter the minor axis. The
dark "B" field shows the spread during the power phase, the gases, in full process of combustion, spread
obtained in the tight area be- towards the next apex with incredible speed and gas transfer velocity is
tween positions 2 and 7, i.e,, iust extremely high. Combustion spreads just as in a piston engine, with the
before the minor axis and about
25 degrees after the minor axis. same kind of flame front. In a low-turbulence piston-engine combustion
The lower lines represent flame chamber, the flamc front advanccs fronr the spark-likc ripples spreacl
velocity. The " A" Iine represents from a stone dropped in a pool of water. In the Warrkel engine, the
the distance from position 2 to
position 1, while the "8" Iine flame front cannot keep up with the gas transfer. It is this phenomenon
represents the distance from posi- that makes the placement ancl location of the spark plugs very critical
tion Z to position 7-a much factors in achieving effective combustion. The spark plug must ignite
shorter way.
<.5 the bulk of the mixture containecl in front of the rotor face, not in the
E.
>./
ls* narrow sections at either end of the working chamber. A spark plug
8.
€I

positioned well beyond the minor axis is presumed to be in a better posi-

J,
l-

gJ 100
=o
'ü tion for successful flame propagation because the gas transfer is more
50e violent on the Ieading side than on the trailing side, and because the
5 45 leading side increases its volume as the rotor rotates. Flame propagation
Shoft sPd nxl0lRpm. to the trailing side is difficult because of violent gas flow and the non-
uniformity of the air-fuel mixture. On the other hand, a spark plug posi-
with Wankel engines iust as with many piston engines. Gas near the tioned at, or slightly aheacl of, the minor axis is in a better position with
rotor's trailing apex is quenched between the rotor trailing and working respect to flame propagation and air-fuel rnixture uniformity; but the
surface, Preventing the advance of the flame front towards the trailing
aPex. There is also some flame front travel in the direction opposite
to the gas transfer, but it is very slow. However, the spark is spread
"backwards" to some extent-especially in the early portion of the
Power phase. The flame front facing the rotor rotation iust sits and
waits for the fresh mixture to be brought to it, rather than going out
to meet it. This produces relatively slow combustion because the en-
tire combustion chamber is being displaced during the combustion
phase.
This graph indicates results of
Another factor adds to the complexity of the combustion process. experimentation with flame prop-
In the vicinity of the minor axis, the combustion chamber is divided agation on different air-fuel ratios
into front and rear sections. The trailing section undergoes compres- in an NSU engine. The "A" field
0.9
above shows the tinre taken by
sion at the same timc expansion is taking place in the leading section. the flame front to travcl f rom 2
I
o'8r
While the gases in a reciprocating piston engine always are contained position 2 to position 1, while E 0'7 E&
within a cylinder area between the valves ancl spark plug on one side and the "8" field represents the dis- CN
0.6<
tance between -Bosition 2 and E,
-
the piston crown on the other, the gas flow in the Wankel engine is position 7.
0.5

Air rotio,l
48 DESIGN TI]E COMBUSTION PROCESS 49

flame front becomes stationary owing to the quenching action occurring Various types of ports havc bccn devclopccl. Tlte NSU-tyPc Wankcl
on the trailing side, which leaves a certain volume of unburned encl Bäs, engine has an intake port on the periphery of the rotor housing ancl
swept into the peripheral exhaust port by the trailing apex. By adopting Curtiss-Wright uses the side port type, in rvhich the intake port is
dual spark plugs, it is possible to maintain positive ignition and mini- Iocated on the siclewall. At Toyo Kogyo, only the peripheral port was
mize the combustion duration. Dual plugs also ensure more complete tested at first, but later the side port and tlre combi-port (which is a
combustion, which results in higher power output and reduced exhaust combination of the two) were tested extensively. Later, Toyo Kogyo
emissions. developed a system with dual side ports.
As in piston engines, both centrifugal and vacuum spark advance can Each port arrangement has its merits ancl drawbacks. The peripheral
help acceleration. With single-plug engines, more spark advance is ad- port meets the requirements of high speed and high output. With
visable because it provides more reliable ignition with highly variable iia. ports, combuition is stable under low-load conditions at low
air/fuel ratios. More power could be developed by Iater ignition (after speed, including idling, ancl it is easy to determine the port . timing
the minor axis), although some loss of smoothness might result. Curtiss- most suitable for raising volumetric efficiency at low speed. However,
Wright found that a spark advance of 4-5 degrees was needed (that cor-
responds to 27-30 degrees in a piston engine) for maximum power at
5,000 r.p.m. In a dual-plug ignition system with the ignition advance of Variations in mean effective pressure (top chart), air-fuel ratio,
the upper spark plug set at the minor axis and that of the lower spark specific fuel consnmption caused by changes in spark Plug _position.
plug variable, ignition advance of the lower spark plug has a great influ-
The line rcprcserrts spark plug position 20 mm. before the minor
axis. The solid iin. repreients a spark plug position 28 mm' before
ence on flame front travel on the leading side, but has hardly any bear- the minor axis. The -. -. -. line represents dual ignition with spark
ing on the flame propagation on the trailing side. Combustion on the plug positions 56 mm. before, and 7.0 mm. after, the minor axis.
F-.- : Mean effective Pressure (kg/cm'). I.0 kg/cm' - 11.2235 p-si
trailing end is delayed, and, consequently, this is the area most prone (pounds per squirrc inch). Luftzahl = air-fuel ratio g/qqh :_ s_pecific
to cause knocking. iuel consuurption, granrs per horsepower-hour. Drehzahl U/Min =
Combustion problems arise any time rotor speecl and gas velocity are r.P.m.
increased. Combustion is affected by the manifold and port configura-
tion, combustion chamber shape, gas transfer veJocity during the minor t0
axis transition, and spark plug location. 9

Iilame front position at 10", 20' and 40" after minor axis in an Mittlerer Druck P^,
experimcntal NSU engine. These sketches show very graphically the
problems of a traveling combustion chamber.
I
lr0
n' EW n.OT
0,9

20' EW n.OT

60 Zundzeitpunkt vor OT in oEW

10

2000 1000 6000

Drehzohl n [UtminJ
50 DESIGN
even with side ports an adequate time-port area for higfu-speed intake
requirenrents is attainable. This is a kef feature of the Wankel
engine,
and a nraior difference bctween it and the reciprocating engine (which
has a mechanical restriction between the valve timing and valve
Iift). This advantage for the Wankel engine is due to the fact that
its rotor moves through a larger angle during each operational p6ase
than the crankshaft in a reciprocating engine nroves cluring one stroke.
Wankel engines with side ports have not only the best timing for low-
speed operation, but also can provide the required intake portlime-area
to raise the useable r.p.m, range. Another distinction between the side
port and the peripheral port lies in the turbulence of intake mixture.
Intake gas furbulence obtained with side ports is more violent than
that
obtained with peripheral ports. This contributes to better atonrization
and air-fuel mixture uniformity. The pattern of the movement of intake
mixture also varies with the shape of the rotor face cavity. The clouble
sicle port engine makes it possible to use the two ports ,.pärrtely,
one as
a low-speed intake port and the other as a higli-speed lntake'iort. In
this case, the intake gas velocity can be kept at äptimum level at all Tnn WaNKsr. ENcTNE can be
times, depending on the operating conditions. adapted to either air or liquid cooling, iust as a reciprocating piston
engine, but the cooling problems are strikingly different. The basic duty
of the housing cooli.g system is to lower the temperature of the areas
exposed to the highest heat input and to minimize temperature differ-
ences throughout the housing.
The same station ary area on the working surface is always exposed to
the same phase of the operational cycle, therefore the heat distribution
in the housing is uneven. A cooling problem exists only in the area where
combustion and expansion take place-the area inrmediately surround-
irg thc spark plugs; thc re.st of the workirrg surfacc rcquires no cooling.
This uneven heating can cause housing distortion which, in turn, can
prevent the proper functioning of the gas and oil sealing elements. The
time during which the combustion chamber is coole«l by fresh gas is
fairly short, therefore the wall temperature of the conrbustion chamber
is high and sensitive to changes in load.
The maximum temperature of the working surface is much higher
than that of the sidewalls; Iocal overheating can destroy the oil fihn on
the surface. Sudden acceleration with a cold engine, especially in winter
or when auto ignition occurs during high-speed drivilg, exposes the rotor
housirrg combustion chanrber wall to a repcatecl sudclert aud very largc
thermal loacl, As a result, therrnal fatigue or thermal shock cracks can
occur around the spark plug holes. In general, cracks occur nrost fre-
quently across the spark plug holes in the axial direction, where there
is a high stress concentration, although in extreme cases cracks can even
reach the water iacket. There is also the danger that thermal distortion
 This chart shows rotor housing
temperatures as measured with
the_rmocouples under varied speed
and power output conditions.
The - line represents an out-
put of I horsepower at 1,600
r.p.m. The - . - . - . Iine represents
ln ggtput of 1.65 horsepower at
2,100 r.p.m. The solid line rep-
resents an output of ).3 horse-
p-ower at 4,000 r.p.m. The spark
plug is positioned about 20; be-
fore the minor axis, and it is in
the area behind the plug that
the highest ternperatürei were
recorded.

r111ar
ta.rrrtrrlr --loo'F The water circulation systern developed by NSU for the single-rotor
I hedtlOO ipr KKM-400. f)urchlass ftir Luft u. Dampfblasen : Passage for
steam ancl air. Wasser-Einlass : Water entry. Rückfluss zum Mit-
LCJ hp6pZ4oo R
{oooF
iPI telgehäuse _ retunr liuc to center casing. I'Iauptstrom : Main cool-
5.1hPo.ooo

ll1j.l,l\"ft!:x,:litff:tlll:::'};,',1'" -
coorant now to the end

housing, and rotor seals must be kept in mind. New high-conductivity

materials for the rotor housing have been developed in an effort to reduce
the maximum temperature of the trochoidal surface, and coolant circu-
lation methods that match the cooling capacity of each location with
the distribution of heat in the housing have been devised. An example of
The intake portion of the hous. 5@
ing remains relatively cool while
cpm
vrt o6a trroilb this is the multi-pass forced-flow cooling system developecl by Curtiss-
ternpcratures rise to great heights Wright. The coolant flows back and forth through the rotor housing,
in the expansion portion. Hous- from one end cover to the other, and back again. The flow passages are
ing material had a great influence
in this connection. The solid line parallel to the mainshaft axis and internal ribbing of the end covers
represents temperatures recorded redirects the coolant flow. The flow area and wetted perimeter (and
in a cast-iron housing, while the thus the hydraulic diameter) vary from pass to pass, in accordance with
dotted Iine represents tempera-
ture distribution with an alumi- the heat input from the combustion chamber. In hot areas of the hous-
nurn housing, recorded under ing, coolant flow velocity is high and vice-versa. This system assures even
iclentical conditions.
heat dissipation and permits balanced cooling with large, clog-free, and
easily manufactured passages which prevent vapor accumulation. Ther-
mal distortions are small because the design allows uniformly thin heat
transfer walls-the outer walls run cool and maintain their design pro-
file. The zones of maximurn heat input correspond to maximum stress
of the side covers will cause deterioration in t'e gas zones, so that the addition of ribs serves to irnprove structural strength
and oil sealing of the as well as heat transfer.
rotor side surfaces.
The housing assembly must be designed to limit Cooling the rotor in the Wankel engine cannot be compared to cool-
this tendency toward irg a piston in a reciprocating piston engine. The piston is largely air-
thermal distortion, ancl the relationrhip, between
the cooling system, cooled, hav]ng a voluminous chamber (cylinders plus crankcase) to draw
 FFlOtr 9(J|AP
HOUSING AND ROTOR COOLTNG 55

on for circulating air, while the rotor, encapsulated inside the housing,
requires higher-intensity cooling. The very fact that the rotor moves at
one-thircl mainshaft speed magnifies the cooling problern. If rotor cool-
ing is not sufficient, various troubles will occur, such as sticking of the
sealing elements to the sealing grooves or to the sliding surface of the
housing, and pre-ignition resulting from local overheating on the com-
bustion face.
txr{Ar5Y 9(n7 There is an important relationship between the cooling characteristics
,INTEFMIO,AI'
(orr rlrcr)
H$O
of the rotor and its material, whether cast-iron or aluminum. Both cast-
iron and aluminum rotors have their merits and drawbacks. Most manu-
facturers use cast-iron because their test results indicate that such a rotor
3xO xso is safer in practical use. An aluminum rotor has the benefit of Iow inertia
This is the cooling system developed by curtiss-wright and high thermal conductivity, but it demands more cooling. Air-cooling
U 5 engine.
for the twin-rotor RC2-60 is hardly adequate for an automotive Wankel rotor and water cooling is
feasible but would cause enormous mechanical complications. The
Wankel rotor, however, has a very convenient structure for being inter-
of the I' su KKM- s-oz-engine. The oil pump nally cooled by oil circulation.
t draws oil from
through a 6lter and then-into a heat &.hrng.r. Lubricating oil of the same type that lubricates the main and eccen-
ntinues straight into the hub oi. ltr.- äto-r]"circütrt.,Still under
the hub sides into^the ru-p, lnria. it, tric bearings circulates inside the rotor and, after cooling the rotor
Water; OEL = Oil.
L' *t.i. trr.-purp can pick it up efficiently, all the oil flows out of the rotor and fresh oil is pumped in.

This schematic shows the oil circulation system of the Curtiss-Wright RC2-60 U5
englne.
a,e6arilge
a,l @
o/t
coo,
F/L

-L I =x=q

ol **61f*-.-
--_t
TANK I r1t Il
2 gt/,, 55t߀
-">o I o Ful
FILTER
.<-
tT 4f-
SCAnENQ€
PUfilP
EAT ExcxenicER
O /Nt€r t/N€ cap/4c/r/ rat AccapoaNc€
H

,v/TH o/L Flory cue tt€,

PUMP @ ouTl.€r LtN€ rausT HlNo/€ TaT/4L o/L FlAt| aLUS
EvrpA/NEp //< oF 2 ctt, Ft /Mttt/, M/|xtMuM,
@ o/L THEtztttosfAf MUSr co,vreot a/l tV TE^lp€F4fUeEptMp
fo
1-,)wesscr /zo'Ft'' lNo L/Mtf aAC( PR€djt/R€ oN sqv€NiE
To
@ bru oy'srgrt ptusT gE Ar tE4sr ldtrlont
OEt

-f, sHA ltrS r/41//Nq lt N&AEO 70 /55u€E

AQ4NSf DP/4/UAic( 7lqau64 P/HP.' aN srlurDowN,
,IIAIN SHAFr HOUSING AND ROTOR COOLING 57

NSU's first rolor_ cooling system Detail design of the inside of the rotor is important and, at the same
.ur.g pressurelfed oil frAti, ;; time, the amount of the oil jetted from the oil hole in the eccentric
intricate circulation prth' inside
bearing to the inside of the rotor must be carefully controlled. The design
Ih. rotor. It was giu.n- up in
favor of using rotor motion to of the rotor cooling system affects lubrication as well as engine balance.
force oil to tt,. t äti.ri- areas. Oil entry temperature, temperature rise, ancl period of contact rvith
the metal are controlled to avoid coking or other lubricant deterioration.
Oil flow is continuous, but velocity and oil pressure are dependent on
the engine's rotational speed. Shaft and rotor inertia pumping super-
COOL- OIL §UPPLY impose a roughly sinusoidal pressure increment on the oil feed pressure.
The rotor cavities are maintained full or near-full at all times, oil being
metered at the pump exit port.
The rotor bearings have oil scals that permit a certain controlled
Ieakage so as to assure lubrication of any high-friction areas on the rotor
flanks. In contrast with the nretcrccl apex scal oil, thc Iubricant for the
side seals orr the rotor flanks scrves not only as gas scaling agcnt but also
as antifriction material. The metering oil pump in the incluctiorr systenr
supplies oil for the sole purpose of gas sealirrg. Maximum flow character-
9raphic illustration of the vector istics of this pump are determined through experiments, according to
field that determin.r-r...i.ration
forces in the rotor. gas sealing performance rcquirements at full throttle. [ust as the top ring
of the reciprocating pistorr cngine is aclcquatcly lubricatccl by a vcry
small quantity of oil filnr which is left aftcr bcing scrapcd by the sccond
ring ancl the oil control riug, the sicle seals of the rotary combustion
engine are continuously lul>ricatccl lry thc oil filrn which is left on the
sidewalls being scrapccl bv thc cccentric bcaring oil sculs.
OF ACCELERATION
While thc piston rings of the reciprocatirrg cngirrc nrust stop at top
and bottonr clcacl ccntcr ancl rvithstarr«l cxtrenrcly high acceleru tiorr

NSU developed a system of guide

vanes to expel the hot oil from
!LIilE the rotor.

This sketch shows how the ac-

celeration forces in the ,oio, .rn
be utilized for driving oif
'to
EFFECTIVE IilERTIA OF the
OIL PARTICLES extremities of the inner cavities
to cool the rotor.
58 DESIGN
Ioacls, the side seals of the rotary piston engine have the advantage of
sliding on the sidewalls continuously-in one direction only. However,
the movement of the side seals is not necessarily simple, therefore the
7
clesign, machining and material of the side seals had to be studied in
combination with the sliding surface of the side housing. When oil is
used for cooling the rotor, and the rotor bearing is lubricated under
pressure, considerable amounts of oil would be ejected onto the sliding
surfaces of the sidewalls unless effective oil seals for preventing oil leak-
age into the combustiorr chamber were provided. During the early stages
of development, the oil seals were relatively primitive and oil consump-
tion was far from satisfactory. From the point of view of mechanical
engineering, the oil sealing provided on the side surfaces of the rotor
seerned to be simple at a glance, but the actual development of the oil
seal turned out to be an unexpected difficulty.
In addition to the mechanical and thermal distortions inevitably oc-
curring on the sliding surfaces of the side housing, the operational con-
ditions and other requirements of the oil seal itself were by no means ONn oF THE MAIN ADvANTAGEs of the
the size ancl weight of a piston
simple. During the rotor's rotation, one part of the oil seal rides on the wankel engine is that it is about hatf
oil film and separates it from the sidewall surface. At the same time, the engine of comparat le powgr _9ujput.
curtiss'wright 'rade a direct conr-
other part of the seal is in contact with the surface of the sidewall. parison betweän their hcz-oo u 5
wankel engi,e and a chevrolet 28)
rated at 185
Obviously, the rotor side seals cannot also serve as oil seals. That would cubic inch vg piston engine. The wankel ängine yts
mean a presence of oil on the full rotor sides throughout the rotor's horsepower; the chevroteiat l9s.
The wankel wäighed only 87 pounds
travel-except for the very small areas between the side seals and the while the vg was over twice as heavy
at 607 poun-di. Rtgardless of maxi-
has a far superior weight Per
rotor edges. It is not practical to keep such a large volume of oil on the mum power output, the wankel engine engines'
rotor flanks during the rotor's complete cycle, nor can the existing side horsepower ratio whe, compared
with reciprocating
seals do an acceptable iob of preventing leakage-the oil seal must be
between the eccentric bearing and the rotor. All constructors now build POUNDS PER HORSEPOWER
their Wankel engines in accordance with this concept.
Horsepover \Y ankel Engine Piston Engine

I0 3 tt-l 5

20 l l2'5
30 ) ,11'l
100 l-1.8 4-9 ''
ZSO 0.8-1.5 2J4
500 0.6-l .2 l ''-)

to the sanle scale:

Dimensional differences are approximately
RC2-60 u5 283 V8

Height ?'l '' inches 3l ' 5 inches

Width 7?"1 inches 28'0 inches
Length l8'0 inchcs ?9 '5 inches

feet
accessories, occupies z)'7 cubic
The chevrolet vg, with its essential
60 DESIGN
of volume. The Wankel engine requires only 5.1 cubic feet for its instal-
Iation. The next maior advantage is a reduction of parts.
The Wankel engine is far less complicated than a reciprocatirg piston
engine. The block is replaced by a housing and two end covers. The
housing is built up of sections containing trochoidal chambers, separated
by sidewalls, and the cylinder head is eliminated. No valve train is neces-
sary, because valves are replaced by ports that open and close automati- -
G

o
cally by rotor motion. The difference in the total number of parts in a J

piston engine versus a Wankel engine is enormous. The 287 cubic inch I

Clrevrolet V8 has a total of 1,029 parts, the Curtiss-Wright RC?.-60 U5

-
]-
r9

Iras only 633. There are 388 moving parts in the Chevrolet V8, while the ,
lrl

(J
RC2-60 U5 has only 151, IL
6
A reduction in the number of parts provides greater simplicity and H
o
Iower cost. Studies have shown that a typical American V8 costs about
The Wankel e
$2 per horsepower. The cost of a Wankel engine probably could be tating
reducecl to half that figure, or $l per horsepower. This rough estimate is weigh
class
supported by an independent cost comparison study. A 90 horsepower, bines, .-
- .

four-cylinder piston engine of modern design, using cast-iron for both io.riitg piston aircraft .engines'
ere
block and head, and requiring the lowest-cost tooling and machinery, Automotive Piston engines
far heavier.
was compared with a ll0 horsepower twin-rotor Wankel engine with an HORSEFOWEß
aluminum housing ancl cast-iron end covers. Both engines were calcu-
lated for the same production rate, and the rates were broken down data
practically the same. The detailed
according to accessories, material, and production time. The results of 6 horsepower the total cost is
tables:
indicated that material cost for the reciprocating engine was 7% higher can be studied in the following
than that for the Wankel rotary engine. At the same time, Iabor costs
cosT COMPARISON (IN U'S' DOLLARS)
were 57% higher for the reciprocating piston engine, and quality was 'W'ankel
'Wdnkel ReciP. ReciP'
Iower.
More recent experiences by the |apanese manufacturer of Wankel Engine Engine Engine
h.p. 6 h.p.
Engine 6 h'p'
engines, Toyo Kogyo Company, and the German firms, NSU and Fich- 115 h,P. Ioo
tel & Sachs, show that, for 100 horsepower engines, manufacturing costs Accessories $ 80 '00 $87.00 $35.00 $29'oo
of the Wankel are 26% Iower than those of the reciprocating engine Material 5o'oo 70.00
5.00 7 '50 10 '00
when considered at the same skill of production, while for small engines Labor 17' 50 2
44.00 l) .2' 17 ' 50
Overhead 3l '00
Conrparison of a twin-rotor Wankel engine (Curtiss-Wright's RCz.
60 U 5) with a 283 cubic inch Chevrolet V8 piston engine. All Total $ 178' 50 §zzo.o, ffi $5650
relevant data are given in the text.
cosT COMPARISON (%)
Wankel ReciP.
Wankel ReciP.
Engine Engine Engine
Engine
v-8 \ 115 h.P. r 00 h.p. 6 h.p. 6 h.p.
104 r00 6l
Accessories r00
r00 r40
RCz-60-U5 Ir{aterial 100 r33
t00 r42
Labor
142 r00 ß3
Overhead 100

Total 100 IN r00 100

62 DESIGN HOUSINGANDRoToI{COOLING6S
The Wankel engine requires no exotic materials and its components I i,rbalu,cc, tfiis is casily corrccted
can be manufactured on conventional machine tools (with certain spe- i,s[aft a,cl tlrc Wa,kcl c,gittc crlll
cialized adaptations), Production methods are of a conventional nature,
re sult in bcnt ancl burnecl
valves,
and most of the operations, in assembly as well as in manufacturing, are
familiar to the average automobile factory worker. f)ue to the snrall l>r«lkcn colrtlccting rocls ancl cratrk'
clamagecl valve gear, brokcn 1>istorrs, its
number of moving parts, the maintenance required is limited, mechani- shafts. Letting a wankel eirgine spin
at rotatiorral speeds beyond
brcakclown or clam The
cal reliability is higher and service life extended. These factors, in com- design linrit cannot result in any ,n..hrnical ^gq
of rvcar ancl increasecl
only cleterioration that will ,.rült is a higher
bination, result in low operational costs. rate
The reduction in the number of parts also indicates less powcr loss «lil colrstllttPtiott.
due to internal friction. Measurements of the various friction Iosses by The wankel
d'trr\vr - \'r rbrr'v encourages
engine v-' high volumctric efficierrcy bccause
Curtiss-Wright show that seal friction is the biggest single friction factor ga btrsti,'rrr cluttrbcr «locs ttot Proccccl
-about 50% higher than friction in the bearings and gcars. Thc scconcl thbutirrilslnootlrswcep.Moretlvcr,
greatest friction in the Wankel engine occurs on the interface between thxplosivechargeover?T}degreesofstroke clttration of 90 cle'
,i,-, intrkc
the rotor and the eccentric bearing. This friction is comparable to fric- rn This gives the wankel
tion between cylinder walls and pistons, and although it reaches approxi- grees of crankshaft rotation in a pisto, engi,e.
mately the same magnitude so far as Ioads are concerned the temperature
problems of the piston engine do not exist. The piston engine's valve
gear absorbs a Iot of power, while it costs nothing to open and close the
ports in a Wankel engine because the rotor movement that performs
this function is part of the basic power-generating process.
The piston engine is subject to vibrations in several planes. The first-
ancl second-order forces of inertia at play in a piston engine never can iil
be equal ized in a single-cylinder engine. Iiven an inline six, or a Vl?., ll;9
which has all primary forces in perfect balance, also has unbalanced
,fl

-:fu
i{
K A NG LE iii.i
seconclary forces. Prirnary balance is achieved by using the motion of one CH A M B
5TT,T1'.
-"'äEölpnöclrtno ii,.§S§
ENGINE
iii
il
piston to counteract the opposite motion of another. But the secondary ;I
rJ
forces, clue to the swinging nrotion of the connecting rods on their crank- J,

600
1l
j.i

pins, Iack synrnretry and the result is engine vibration. The Wankel ,ii
'i,

cngine has no such problerns because it is free of reciprocating motion-

:l

'.t
;i
.too ;i

all motion is strictly rotary. Although the Iarge angle of each operational -: ,i
U'
phase in the Wankel engine poses a hanclicap to its low-speed perform- c ir.
',

ance, it is advantageous for torque fluctuation. The torque fluctuation

I 200
lrl
rate of the twin-rotor Wankel engine is almost equal to that of a four-
G,
=
o GINE
stroke, six-cylinder reciprocating piston engine. o
lrl
o SING I'

By eliminating reciprocating motion in the working parts of the G,

o. 600 i;
{
s

engine, the Wankel avoids all problems of alternating inertia stresses

!:
G,
ur
o i
bccause there are no unbalarrced inertia forces. In a piston engine, all - 400 7

reciprocating parts must be designed with adequate material strength -(J

to witlrstand these inertia ]oads. Reinforcements often have the disad-
t

200 :
a

vantage of addirg to the weight of the parts in question, thereby further

I

COIB.
increasing the inertia stresses. Inertia loads also impose a Iimit on the o''l .I REY.
CRANKSHAFT ANGLE
rotational speed of the engine. Iiven though rotor eccentricity is respon- -.)l
 64 DESTcN
engine a substantial volumetric efficiency
aclvantage. The cSange from .20
ntinimum to maximum volume in the working ;hrmber takä
place r?cx vtl.ocrTv ooilSrAnY Foi
cluring three-quarters of a nrains]raft revolution,"
compared to one-5alf aoo l- t RoroRs ALL ttzEt
crankshaft revolution in a piston engine. O BXP
Consequentiy, it is Iogical to cooo
expect at Ieast 50% better breathinf in a
Wankel engine. It could be Iro
I

even more, because the intake port

BRAXE
RPT
tlre 270 degrees intake phase, *trit. HORSEFOUER T.',"
560
opens and closes more gradually. T
the conrbustion chamber at the end rooo
l--a norOnS
has no quench areas or other corners 5ao
where inert gases may remain. In I

addition, centrifugal force assists exhaust gas T"'

evacuation, which in turn
contributes to higher volunretric efficiency'--th.r, 520 O RPH

' is no similar effect in

a reciprocating piston engine .
smoother than in a piston engi
a Wankel engine is 300
I

gine produces positive

torque for about trvo-thirds o1 o
le as opposed to one_ ao 50
quarter or Iess of the cycle in a
engine. OISPLACEI{E'{T / ROTOR

In this chart the to For higher power requirements, additional rotors and housings can
engrne are compargd be stackccl on a basic \Vankcl cngine. Adding rotors adds horsepower
with thosc , ,ä
"f
twin'rotor wankel \-torque outpu[ of a and specds trp thc cngiuc irr an altnost lincar curve.
output, while in a % of mean torque
between 200 ,na J ot'U torque vaües
i.
The seals in a Wankel engine can be compared to the piston rings
900 in a piston engine. At excessive crankshaft speeds, which produce ex-
treme piston accelerations, piston rings begin to flutter. There is no
800
corresponding phenomenon in the Wankel engine because the seals

Lt slicle unidirectionally. There nre accelerations and decelerations even at

600 l_I RTCIPßI)CAIII{G T

t-
G[{ts
steady output shaft speed, but no conrplete stops ancl no reversals of
sliding direction.
Gas pressure loads on the bearings arc higher in a Wankel than in a
Lu piston engine, but centrifugal loads are lower; therefore, the risk of bear-
irg failure is far snraller. One reason is that rotor eccentricity is smaller
l_lt than the crank throw in a piston engine. Negative torque loads, such as

L\IV
400
are present in a piston engine during the compression stroke, do not

f
300 exist because the Wankel has no reversal of rotor rotation at any time
during its operational cycle. Moreover, the Wankel engine has no parts
Y x
3
that must be accelerated, stopped, and returned in linear, reciprocating
b
200

motion. Also, the output shaft makes three revolutions for each rotor
\ - orbit, which means the Wankel can be thought of as having a slow-
ß c tt{Gtr{ts
moving piston, which immediately introduces certain advantages:
468t0
l. Lighter inertia loads. . .
r{uM8tn 0t cyut{t)tßs/ß0I0R
2. Reduced need for lubrication. . .
BAr{xs
3. More freedom in seal design.
66
DESIGN
is clearly superior to trre reciprocating

:'ff.l^| lH,T'I:t,H'l' JJ,'J HI, *

Section ll

DEVELt)PMENT
 I

Felix Wankel

Tur vrLLAcE oF Lann in Germany's

"Black Forest," is situated in scenic countryside in southwestern Ger-
many, close to the Swiss border and not far from F-rance's Alsace prov-
ince. Lahr is near Donaueschingen, the town where the confluents Breg
and Brigach form the beginning of the river Danube. The region is
called Swabia, and forms the heartland of what is now the German
state of Baden-Wi.irttemburg. The land is hilly, but not savage, and,
although close to the AIps, the altitude of the Black Forest mountains
never approaches the tree line. The hilltops are woocled, the valleys
dark, the climate extreme. Winters are bitterly cold and summers very
hot. The people of Swabia are known to be taciturn, stubborn,
unapproachable, industrious, stolid, unimaginative, slow, practical, non-
intellectual, shrewd, subtle, settled in their ways, and of an unimpeach-
able morality. This is where felix Ifeinrich Wankel was born on August
I 3, t g0z.
A boy growing up in this area could become a farmer, blacksnrith,
lumberiack, wine-grower, butcher, baker, or go into related activities.
Land, vegetation, botany and life were close to young Felix Wankel.
His father, Rudolf Wankel, was a forest commissioner. FIis mother,
Gerty (nde Heidlautr) was frorn a local village family. For a boy with
this background to beconre a major inventor of mechanical devices is
as unlikely as the story of a medical doctor in the American midwest
building a machine which was to beconre a principal factor in the forma-
tion of the automobile inclustry. Yet Elwood Flaynes, a general practi-
tioner in Kokomo, Indiana, built one of this country's first practical
automobiles. And, Itelix Wankel did become an inventor.
70 DEVELOPMENT
I{is father was killed in the beginning of World War I, when Felix
was only 12.lle completed grade sclrool locally anc{ stayed in high school
until l92l . At this point in life, it was customary for a young man with
a mincl of mechanical bent to seek an apprenticeship with an industrial
enterprise, usually in a foreign countty, but frequently in another Prov'
ince in his own nar:ion. This would lead to a series of subordinate
positions, which in conrbination made up the learnirg of a trade, and
qualified the man for his chosen profession.
Gottlieb Daimler, Carl Benz, August Horch, Emil Hermann Nacke,
Heinrich Kleyer, Richard Stoewer, Heinrich Büssing, Nikolaus Dürkopp
and other founders of the German automobile industry had launched
their careers through engineering experience obtainecl in the shops of a
variety of far-away industries. This, too, would have been the logical
procedure for Felix Wankel, when he graduated from high school at the
age of 19. But economic conditions did not allow him to follow the
customary course. Postwar inflation in Germany hit hard a widow with
a son in high school, and Felix had no choice but to go to work and earn
a living. He moved to Heidelberg and found employment as a salesman
for a publisher of scientific books. fn this ancient university town, close
to such industrial centers as Karlsruhe, Darmstadt, Frankfurt and Stutt-
gart, young Wankel was not devoted solely to learning the skills of sales-
manship. He was forever experimenting in his mind with new ideas for
machin ery of various types.
He had always worked with things and insisted on trying out every-
thing in practice-theoretical answers were not good enough for him.
It was this frarne of mind, his basic mental makeup, that drove him to
Felix Wankel with the inner rotor of the DKN'I-51.
take one of the most important steps in his career. At the age of ?2,
Irelix Wankel opened his own mechanical workshop in Heidelberg. His
first rnoney-making contracts involved such prosaic machine work as was working as an independent consultant on high-pressure lubrication
grinding cylinders. But the importance of this stage of Wankel's work apparatus, he started to make suggestions for irnprovements in the seal-
in terms of his later achievements lry in the fact that in his own shop ing area. This work made trim very conscious of the problerns of sealing
he learned the key factors of machining and production, the meaning It
the gaps between two or more surfaces in direct contact. became
of precision, and thought about improvements to the internal combus- obvious to him that the same principles could be applied to obtain
tion engine. As early as 1924, he began to make sketches for a rotary better sealing in internal combustion engines. Soon, he devoted all his
piston engine. He had, even at this stage, a keen sense of its shortcom- time to the task of improving cylinder sealing in piston engines, and, in
ings. He evaluated his own ideas without prejudice and concluded that 1928, he set up a test establishment for motorcycle engines.
his designs were not suitable for further development. Felix Wankel took out his first patent (DRP ,07 584) in lg}g-it cov-
Throughout this period, Felix continued to study. He went to night ered an engine with a reciprocating piston inside a horizontal cylinder.
school and took correspondence courses, and the idea of the rotary The piston had a crown at each end, and the cylinder had a combustion
engine would not leave him. Although he wanted desperately to build a chamber at each end. The piston's wrist pin was mounted eccentrically
rotary engine, he didn't know where to start. He began by experimenting in relation to the mainshaft, so that piston movement produced main-
with rotary valves on conventional motorcycle engines. In 1926, when he shaft rotation. Thcre was nö connecting rod-the wrist pin doubled as
72 DEVELOPMENT FELIX WANKEL 73

crankpin for the mainshaft. This engine called for no new sealing tech- for sealing p.urposes alone. Any wide packing surface produces a consider-
niques, and enabled him to use standard piston rings. He soon realized, able power loss in high-speed machines because of friction. It is also
however, that this was sidestepping the issue, and that without icleal difficult to cause a wide packing surface to bear uniformly tightly be-
sealing his rotary engine could never become a reali§. He recogn ized cause the small heat warpings of the material, which are unavoidable in
that many earlier types of rotary engines could not work with high practical operation, have a particularly disturbing effect. Certain earlier
internal gas pressures because the seals were not of sufficiently robust seal types tried to circumvent these difficulties by not allowing any gas
construction. He also realized that sealing elements were dependent on pressure to act on the underside of the packing bodies. The pressure on
the presence of an oil film to fulfill their intended function. the sealed interface was obtained solely by mechanical spring pressure.
From 1929 to l93l Wankel built and cleveloped twenty devices to test Wankel saw that a pressure-tight seal permittirrg heat expansion while
his ideas concerning rotary valve seals. He made a disc valve engine and working under low friction conditions could not be obtained solely by
installed it on a motorcycle that ran satisfactorily. He also succeeded in spring pressure because the gas pressure, which tends to force its way
running a test engine with a cylindrical rotary valve. The next two years between the sliding surfaces, is capable of exerting a considerable power
were spent evaluating the results and applying for patents. In his search action in a direction which separates the sliding surfaces. A spring force
for the basic concept of a rotary engine, Wankel defined the heart of sufficiently strong to counteract this gas pressure would cause too high
the matter. FIe needed a rotary piston that would also rotate the shaft. a pressure load on the sliding surfaces, resulting in wear and considerable
But all attempts, all his tests, failed for the same reason-he was unable friction when the gas pressure ceased. Wankel's idea was to direct gas
to provide proper sealing of the working chamber. In 1933 he was still far pressure so that it would improve seal tightness and act in the direction
removed from a successful engine. of any necessary spring loading rather than opposing it. His new packing
Then began a new and bright period. In 1933 he was engaged by body obviated the disadvantages of too wicle a seal edge and clid not
Daimler-Benz for research in sealing components, rotary valves and require any high spring force. It simply drew upon the operating gas to
rotary engines. The engagement ended within a year when he fell out apply pressure, and spring-loading was necessary only to a suitably adiust-
with the management of Daimler-Benz, but he soon signed a contract able degree for supporting the seal in its position. According to Wankel's
with BMW for similar work. His principal assignment at BMW was the invention, the seal was effected by a packing body which surrounded a
development of a piston engine with rotary valves. The result of his work cavity or opening with a iointless sealing edge. For the purpose of reduc-
in the area of seaiirg finally bore fruit. irg friction, the packing surface was limited to a narrow strip. The pack-
In his patent (DRP $7 701), dated November 6, 19)6, Wankel said ing body was found suitable for both disc-type and barrel-type rotary
that sealing is not ideal if the seals are formed by the sliding surfaces valves.
themselves, no matter how high the precision of machining and assem- In the case of the disc valve, the lower face covers the entire cylinder
bly because this only results in high friction and rapid wear. Sealing bore and the packing body takes the form of a ring with an inverted I-
duties should be performed, Wankel concluded, by special sealing ele- shaped cross-section. A spur on the base of the L provides the sealing
ments designed for that purpose only. Wankel applied for this patent edge. The disc valve does not rub against the top of the cylincler block,
on September 24, 1934, covering certain new improvements in gas sealing but bears against the packing body which is mounted in the block. The
techniques. He called his sealing element a "packing body." base of the L is spring-loaded against the sliding surface and the mast
The packing body is defined as being closely fitted in its supporting of the L is guided in a slot which is sealed by two small piston rings.
body and pressed with its end surface against the machine part which Gas pressure admitted from the cylinder to a chamber below the mast of
slides over them. Earlier attempts at gas sealing, Wankel pointed out, the L-shaped packing body will act irr the clirection of the spring-loading
had failed because gas or liquid pressure could act on the whole under- and add to sealing tightness. The amount of gas pressure coulcl be regu-
side of the packing body in question. This offered a very large exposed Iated by an adjusting screw controlling orifice size. The principle is the
surface, making it necessary to dimension the tip surface on the friction same in draw-type rotary valve applications but the plcking bocly tlkes
side of the packing bocly of correspondingly large surface area in order a different shape.
to prevent overloading the working surfaces. Sealing surfaces or packing Itdoes not matter whether the machine is a compressor or a heat
edgcs around the operrings were therefore much wider than was necessary energy engine, the sealing elenrents havc to be urritecl so that thcy fornl a
74 DEVELOPMENT
Wankel developed this sealing
system that completes a full
sealing path, end to end. The various seal- grid for a rectangular oPening.
ing elements must be interlocked to block leakage through corners or A : seal strips. B - ioint trun'
ioints, In many earlier engines a sealing path could not be completed nion. This invention was the
basis for the side sealing of the
because of certain design or constructional characteristics, but the rotor in the Wankel engine.
Wankel patent covered a complete sealing path. It specified possible
applications to both compressors and internal combustion engines.
Wankel claimed his seals would withstand gas pressure loads of 30 atü
(1.0 atü - L{J psi) with a seal surface contact of 5 mm. with no greater
Ieakage loss than exists in a reciprocating piston engine.
The sealing grid effectively covers an entire annular chamber of rec-
tangular cross-section; the top cover revolves while the sides and bottom
are stationary. The top cover is sealed by complete seal strip rings, inside Two views of a split-ring seal
patented by Wankel. The insert
and outside the annular chamber. The seals are backed up by spring- almost blocks the gap. A _ sPlit
Ioaded seal carriers and the sealing path continues around the bottom ring. B - L-shaped sealing in'
in a U-profile-this is the direct basis for the rotor side seal system in the sert.C-leakagegap.
Wankel engine. |oints in the seal strips are made by trunnion pieces
and the seals are lightly spring-loaded so as to provide effective sealing
despite variations in gap size. Such variations can occur as a result of
axial thrust or thermal expansion. The narrow seal contact wiclth of ,
mm. was chosen to give small friction losses. There are no centrifugal
loads on the seal strips because they are located in the stationary parts.
Between 1934 and 1936 Wankel succeeded in taking out sealing Wankel invented a dual seal
patents that gave him blanket coverage of all possibJe rotary valve, com- for disc valve applications. Both
rings have conical profiles, and
pressor and internal cornbustion engine applications. One type of seal the upper ring is solid while the
Wankel designed for use in engines with rotary valves consists of two lower ring is split. The two rings
rings, placed one on top of the other. One is soJid, the other is split. Both are interlocked by a dowel Pin
to prevent their rotating relative
have a conical profile, so that their interface is inclinecl. It is a face-type to each other. A = solid ring.
seal, sealing a cylindrical cavity against its own recess walls as well as B : dowel pin. C = split ring.
against the lower face of the head or lid. Gas pressure tends to force the Below, plan view of the split
ring.
rings apart, because the inclined interface produces opposite reactions
in the two rings creating a more effective seal. Relative rotational move-
ment in the two rings is prevented by a small dowel pin inserted at one
point on the interface.
Wankel invented another type of split-ring seal where the gap is not
straight-cut, but stepped. The steps run in opposite directions, so that In case of overlapping circular
the gap is T-shaped. The crossbar on the T is filled by an insert which openings Wankel had a sealing
effectively blocks the leakage path, except for the small slot formcd by system, too. Gas pressure from
inside the chamber was used to
the stem of the T. A seal for the sides of interacting disc valves was also make the seals more effective,
developed where two seal ring segments are combined to form a figure- pressing the inside members at A
and D against the outside mem-
eight perimeter. The joints are sealed by gas pressure, forcing the open bers. B and C represent seg'
ends of the segments together. ments of sealing rings.
Wankel's first rotary engine, developed during this period, was a far
 Zu &r Prrcnrrbrifr 637 701
Zu d:r Prtcarrbat, 657 701 Zo &r Puarr<ltirr 637 70 1 Kl. 46 rr Gr. 3
il.{6rr Gr.8' rr
Kl. {5 Gr. 3
Blau ll.
Abb. z

Atü.t
*Lain A-D

r;;
1__

+
i.

H
:1 CORNER SEAL

Felix Wankel's concentric rotary engine with rotary valves at two different levels
Ied to the developnrent of a disc valve engine for BMW in 1931.

SEAI SPRINGS
cry from the refined machinery now being produced by NSU and Toyo
Kogyo. Application for patent was filed on September 24, 1934 for a Central parts of the Wankel concentric rotary engine'
two-rotor rotary engine, in which all rotation was concentric. Two rotors
were necessary to make the engine run on the four-stroke cycle. The two
annular working chambers had a rectangular cross-section. Three sides of ROTARY VALVE Atb.l ROTARY VALVE

the working chamber were made up of a cast-iron housing with cooling LEAO]NG ßOTOR EDGE

water iackets. These three sides were the bottom, plus the inner and
outer peripheries. The fourth side was a revolving cover-to which the
rotor was fastened (or with which it was integral).
The rotor filled less than half the working chamber and had concave
i:o r-l_.
faces with a rectangular cross-section nearly 180 degrees apart. The
rotors were concentrically mounted on the rotor shaft, one above the
other. Engine timing was provided by counter-rotors, spaced 120 degrees
apart, revolving in a direction opposite to the unidirectional engine rotors.
The counter-rotors overlapped with the engine rotors and formed a com-
bustion chamber. The counter-rotors were kept in phase by gearing (to .\l

make sure they did not touch) in a I : I ratio. It is the relative movement
of the engine rotors and the counter-rotors that provide the necessary
pumping action for engine operation.
Plan view of the concentric 'ot^i
engine' ,
TRAILING ROTOR TDGE
78 DEVELOPMENT FELIX WANKEL 79

The lower of the two chambers charged and compressed while expan- Hitler's assumption of power, Wankel had helped uncover an embez-
sion and scavenging took place in the upper chamber. The operational zlement case that went against the NSDAP (National-Sozialistiche
principles of this engine involved using the rotor in the lower chamber Deutsche Arbeiter Partei). When he became State Chancellor, [Iitler
as the compressor rotor and the rotor in the upper chamber as the sought revenge and Wankel was branded a traitor to the party and im-
power rotor, Both were mounted on the mainshaft and were concentric prisoned for months.
with it-the end covers were integral with the rotors and revolved with Free again in 1935, he moved his workshop from Heidelberg south to
them. TJre lower chamber had an intake port, while the exhaust port his native Lahr. In 1936, Hermann Goering's air ministry heard about
was in the upper chamber. The two chambers were linked by a com- Wankel's work and invited him to conduct his experiments at the DVL
municating port, and the opening and closing of all the ports were con- (Deutschen Versuchsanstalt Ftir Luftfahrt), in Berlin's Adlershof. Wan-
trolled by rotor movement. kel declined because he did not want to live in Berlin and Persuaded
With this design, the counter-rotors provide differences in chamber Goering to set up a separate institute for him: WVW (Wankel Ver-
volume. One counter-rotor intrudes into a part of the compression cham- suchswerkstätten) in Liudau on Lake Constance. The ministry had
ber that has iust been vacated by the compressor rotor and sweeps high expectations and invested several million rnarks in this research
i

institl,rte and its experimental shops. Here, Wankel was to build a

§
t

through it in the opposite direction of travel. That movement creates a

rapid expansion against the trailing faces of both the compressor rotor rotary valve engine for the Luftwaffe. Dr. Bensin get, the key man be-
hind the Wankel engine that powers the Mercedes-Benz C-lll, was one
I

and the counter-rotor, which results in the aspiration of a fresh charge

I

from the intake port. As the compressor rotor continues its rotation, its of his assistants. Itescarch was dirccted towarcl rotary engiues during
Ieading face meets the incoming charge, shuts off the intake port, and 1938, but 1939 found Wankel occupied with more routine tasks. He
begins to compress the fresh charge against the leading face of the op- invented a systcm for improved pi.ston cooling in an aircraft engine and,
posite counter-rotor. At the point of maximum compression, the move- in association with Dainrler-llenz, he de v'eloped a rotary disc vah'c,
ment of the compressor rotor opens the communicating port, and the which went into production on the DB60l Vlz aircraft engine in 1942.
compressed gas is admitted into the power chamber, where it is ignited The reasons for using the disc valve insteacl of the rotarl' tubular valve
as soon as the port closes. (which Wankel favored) were not technical-in 1910 it was decided that
Gas pressure from the combustion propels the power rotor. Expan- the simple design and ease of manufacture of this type valve over the
much more complicated tubular type was to take priority over all tech-
:

sion is allowed to take place due to the relative movement of the power :

rotor and the counter-rotors. For every revolution, the compressor rotor nical considerations. The government was confident that Wankel could
admits a fresh charge on its hailing face while its leading face is busy solve the problerns that had croppcd up during the early development
compressing the preceding charge. For every revolution the power rotor of the disc valve in 1936. The Wankel disc valve was a plate with one of
is driven by gas pressure against its trailing face while its leading face two open sectors placed between the cylinder head and the block, cover-
expels the burned gases, ing the cylinder bore. As it revolvecl, the open sectors workecl as Ports to
The engine had many drawbacks, such as complicated machinery, adnrit fresh rnixturc ancl allow burned gas to be expellccl. Wankel hrrcl
excessive bulk and weight, and questionable efficiency. Its significance been discouraged by problems with overheating-cooling the large-
was mainly in the sealing system developed for it. Here, Wankel devel- diameter disc valve was extremely difficult because it had little direct
oped a brilliant solution to the persistent problem. Complete sealing I
contact with the cylincler head for heat dissipation, and it had a large
grids were formed around the working chambers, comprised of a exposed area. He knerv that overheating would certainly lead to distor-
number of curved and straight sealing elements arranged end-to-end and tion and leakage and a serious loss in volumehic efficiency.
in series, interlocked by cylindrical plungers, with freedom for the entire The Daimler-Benz type DB60l Vl2 aircraft engine hacl been origi-
sealing grid to move axially and exert pressure against a sliding surface nally designecl in 1935 as the prototype for a whole new family of power-
from its mounting in the stationary housing. This sealing system was to plants for the aviation industry. It had trvelve cylinders in an inverted
have a great effect on future developments.
I
V-formation, was water-cooled, and had a light-alloy block. Cylinder
These patents were a triumph for Wankel, but he had moments of displacement was 2,020 cubic inches (33.4 liters), and output was 1,050
despair during the same period because of the political situation. Prior to horsepower at 2,400 r.p.m. The initial prototype had carburetors, but
80 DEVELOPMENT

direct fuel iniection was developed for subsequent versions. In 1939, a

DB60l with turbocharging (rated at 2,770 horsepower developed at3,l00
r.p.m.) was installed in a Messerschmitt Me 209 fighter plane for speed
record attempts. The advent of the war prevented any international
records from being set, and its performance in test flights became a mili-
tary secret.
Using this basic engine, Wankel went to work. Special cylinder heacls
I

were designed in which a nitralloy steel disc with two ports was cen-
trally mounted on a hollow trunnion in Iine with the cylinder bore.
l'hese discs themselves performed double duty as wheels. Each disc
meshed with the adjoining discs to form a train of gears. The drive was
taken from the crankshaft via a pinion mounted in place of the cam-
shaft ancl the discs were geared to rotate at one-quarter crankshaft speed.
Fuel injection was retained for the disc valve version, the injection noz-
zle passing through the center of the trunnion that carried the disc. The
decision to place the disc valve model DB60l in production for the Luft-
waffe's fighter planes and light bombers was made only after the experi-
mental engine had completed a 200-hour, full-load test and it was proved
that the disc valves were still in satisfactory condition.
'I'he sealing of the disc valve was effected by an assembly of circular
piston rings for the circular cylinder head ports and the disc carried an
expansion sleeve. The two small piston rings were assembled outside the
expansion sleeve and pegs on the sleeve fitted into the gaps of the rings
to prevent the gaps from moving into line. This assembly was mounted
in circular recesses surrounding the cylinder head ports. The flat upper |unkers |umo KM8 engine installed in ,,, torpedo. This cross-section shows very
well why there was no space for conventional valve gear.
face of the valve disc bore against the flat face of one of the piston rings
and the inner expansion rings. End pressure on the disc was caused only
by the part covering the valve port, the remainder of the disc being ex- of 100 IUMO KM8 engines designed for the special purpose of propel-
posed to combustion pressure on both sides, therefore being in balance.
Iing torpedoes. This engine was considered by experts to represent prog-
A small metering valve, driven from the cylinder head side of the disc, ress in automotive development. High output, coupled with light weight,
regulated the supply of oil to the disc surfaces.
in the smallest package space is an important obiective in automobile
Daimler-Benz discontinued production and development work on the as well as torpedo design. The KM8 engine was required to perform
DB60l in l944,but another company was still actively engaged in experi- only for a very short time under abnormal operating conditions. There
ments on the Wankel clisc valve-|unkers. The engineering business is no doubt that with suitable changes in design, and under less adverse
started by Hugo funkers in 1893 had become a major factor in aviation operatirg conditions, Ionger life could be expected.
when |unkers invented the all-metal airplane in 1910. In 1937 the Nazi Design obiectives for the funkers KM8 engine called for capability to
government reorganized the |unkers industries for a variety of military
drive the torpedo at 40 knots, which translated into a power requirement
production needs. |unkers was not interested in the Wankel disc valve of 275 horsepower at 3,650 r.p.m. Minimum weight was essential, and
for an aircraft engine, but for a torpedo engine. therefore there was a strong leaning towards ]unkers standard aircraft
Between l%7 and 191, ]unkers built ancl testecl a number of experi- practice in design. The engine had to operate under full load immedi-
mental engines using Wankel's rotary valves. When World War II ately after start-up and accelerate to maximum speed in the shortest
came to an end a production order had been issued for the manufacture possible time. It was started by a compressed-air started motor and had
 FELIX WANKEL 83

Each head was divided into an upper and lower section to accom-
t
I modate the flat clisc valves. T'he heads were attached to the cylinder
block by means of stucls which passecl through both sections. Valve
)
I
i
I
I
I
cooling was recognized as a particular difficulty, therefore the lower
I

VALVE
I
cylinder heads contained water passages to allow the coolant (sea water)
to flow from the cylinder blocks to the upper head section and the
upper cylinder heads had inlet and exhaust ports corresponding with
VALVE AXIS :
those in the lower sections. The cylinder barrels were screwed into the
underside of the lower cylinder head section. This type of ioint required
.l

I
I
I
I

t
t
no further sealing arrangement. The lubrication system was conventional
t
I
for the most part, although the disc valves were lubricated by a separate
I
multi-piston pump which delivered a fixed quantity of oil to each disc
to each disc bearing.
t.

PISTON
I

t
and.
The disc valves themselves had teeth cut on their peripheries similar to
the disc valves on the DB60l engine, and formed two gear trains without
any need of gear wheels or idlers. The valve discs were located between
I
the upper and lower cylinder head sections. Cavities were machined in
', the upper sections to provide suitable clearances and two inlet and
I
I

two exhaust ports per cylinder were cast into the lower cylinder head.
AII ports were circular ancl each hacl a stcel se:rling inscrt, hcl«l ilgainst
the disc by the gas pressure in the cylincler. The outside diameter of
I

the seal insert was made gas tight by trvo ordirrary piston-type conr-
Cross-section of |unkers KM8 torpeflo engine with Wankel disc
valve. Plan view öf the valve disc and its ports (lower left). :
I

I
i

Detailed section of the Wankel disc valve showing how critical the clearances were.
I

to reach maximum speed and power within one to two seconds from :

cold.
5.9
too2
3 t.5 1.7 0.20
The fuel includecl a mixture of exhaust gas and oxygen, which meant +O;O5 +O.OS -O,OZ O.lä
that the engine had to be capable of withstanding higher temperatures
and pressures than exist in gasoline engines. Because the required life
I

of the engine was only a few hours, hard wear of individual parts could
be tolerated and, because the torpedo was only Zl inches in internal
diameter, the overall cross-section of the engine had to be extremely
compact. Wankel's disc valve was chosen for this engine because the
planned cylinder displacement of 260 cubic inches (43 liters) precluded I

the use of normal overhead valve gear in the available space.

The resulting KM8 was a liquid-cooled, four-stroke, V8 engine with
magneto ignition, 90 mm. bore and 85 mm. stroke. Each cylinder had
a displacement of 32.6 cubic inches (0,542 liters), resulting in 26l cubic Max. Min.
inches (4.34 liters) displacement for the whole engine. The comPres-
Clearance at A. r'7 - l'5:o'zr.55- ro.95 r'68 -I'55: 01t 3.
,, ,, B. 6.35 +3.o5 + and 6'g+3* r.5- ro.8 and ro.B
sion ratio was 6.6: I and the cylinder heads were silicon light-alloy - ro95-(B.Z+r.g):o35 - (B'8 *z) - e. and
castings. The same material was used for the integral cylinder block t ,rc. 6'r7-o'I3:6.o4 and 6.o4- 6.rS-o.z-5.95 5.95
S.78 -o2& -5.82 == O. I 3.
and crankcase.
81 DEVELOPMBNT FELIX WANKEL 85

pression rings. The upper sides of the lower cylinder head sections were Wankel built an adiustable rotary compressor in 1944, but the end
machined out above the center of each cylinder to make room for the of the war interrupted his work. From 194, to 1946 he was imprisoned
valve bearings, which were pegged to prevent rotation. A set of needle by the French occupation forces. The Wankel Versuchswerkstätten
rolters reduced friction between the inner diameter of the disc and the was dismantled in 1945 and, even after his release, he was forbidden
outer diameter of the bearing. The circular opening through the sealing to conduct research and experimental work. This turned out to bc sonre-
insert and the narrow section that canied the compression rings mini- thing of a blessing in disguise, for it gave him time to review his work,
mized the area under pressure and reduced friction. There were two port write several papers, and produce initial designs for a rotary piston
apertures in each disc, shaped to provide quick opening and closing. The engine with his own sealing system.
gas pressures acting on the disc valves were counterbalanced by means The event that was to enable Wankel to make a fresh start was the
of thrust bearing rings installed in the upper cylinder head sections. release of a man named Wilhelm Keppler from the War Criminal's
The disc trains were driven from the crankshaft via pinion gears at Prison at Landsberg in 1950. Keppler had been Hitler's business con-
the rear of the engine and bevel gears. The clisc valve spindles were Pro' sultant, State Secretary without portfolio in the Nazi heyday. Keppler
vided with splined couplings which allowed precise valve timing. Shear had.known Wankel since they were both young men. When they first
pins were fitted to the valve train gears to prevent damage of the discs met in 1928, Wankel had voiced his ambition to enslrre the greatness
and gears in case of disc seizure. The target output of 275 horsepower of the German nation by building a superior type of engine. He wa.s,
at 31650 r.p.m. was obtained when running oh a mixture of oxygen of course, refening to his rotary engine ideas. It was Keppler who had
and exhaust gas with a supercharge pressure of 1.5 atmospheres. Run- arranged, at the instigation of I'Ians Nibel, chief engineer of Daimler-
ning the engine on gasoline fuel without supercharging, the engine de- Benz, to rescue Warrkel from prison after Hitler had had him iailed.
liveied I l0 to 120 horsepower, which represented a specific output of Keppler also protected Wankel in all ways during the years that fol-
only 0,127-0.461 horsepower per crrbic inch (?r,1-27.8 horsepower Per Iowed. Thus, again aided by Keppler, Wankel opened his new estab-
liter). lishment in l95l-the Technische Entwicklungsstelle in Lindau on Lake
When running on the oxygen mixture, there were problems with the
,

Constance, not far from the site of the original W\A)y.

disc valve bearings and the gas tightness of the sealing inserts. When Borsig, in Berlin, was anrong the first to give the new company a
running the engine on "free air," however, practically no sealing diffi- contract. The contract specifically coverecl sealirrg systcnrs for air com-
culties were encountered-the engine was run as Iong as 50 hours with- pressors. Wankel was also to do some work for Go etze, of Burscheid, in
out any trouble. An experimental |unkers 210-S single-cylinder-engine connection with improved piston ring sealing. Keppler, still extremely
had similar disc valves and was reported to have run trouble-free for well connected within Germany industry, steered Wankel to NSU. They
more than 200 hours. were then doing well with racing motorcycles, and in l95l NSU took
During the tests several design changes were introduced. Two bronze notice of Wankel's work on rotary valves, with the idea of using them
thrust rings with provision for pressure lubrication were mounted in on racing motorcycle engines. This resulted in an initial agreement
the upper cylinder head section to avoid wear on the discs because of between Wankel and NSU that was to be followed by a long Iine of
excessive pressure loads directed against the lower surface of the uPPer contracts.
cylinder head. As a precautionary measure, the discs were made thicker The technical aspects of the development of the workable Wankel
to strengthen the bearings. engine are covered in more detail in the chapter dealing with NSU. It
About this time, Wankel's sliding seals also got into production. The is interesting to note that Felix Wankel never went on the NSU pay-
Mercedes-Benz DB60l disc valve engine was used until 1944 and led roll as an employ€€, but negotiated contracts with NSU for the rights
to ioint contracts between DVL, WVW and such companies as Auto to his engine and future research work. These contracts have since
Union, Hanomag and NSU. By 194r, Wankel had firm contracts also brought him considerable revenue. He gets a substantial share of NSU's
with Daimler-Benz and Borsig, Germany's largest manufacturer of income from the selling of licenses to manufacture and market Wankel
railway rolling stock and other railroad equipment. Wankel himself engines, and he retains the right to engage in work for other clients as
spent the war years 1940-1 944 in Berlin-Adlershof, engaged on a re- he wishes-he heads his own small establishment in Lindau.
search program for DVL. Wankel always has been rernoved from the day-to-dry testing of the
86
I
DEVELOPMENT
engine he invented, and not at all concerned with the tooling required
for its production. But, he has had the satisfaction of having the en-
gineers from NSU come to him with their problems. He continues to
experiment with new tlpes of seals and with new geometrical shapes in
his own laboratory. He has also done considerable research work on
adapting his engine to diesel fuel (compression-ignition) operation. Un-
der his iesearch contract with Borsig in Berlin, Felix Wankel was still
engaged on development of large rotary compressors as late as 1960,
and, in December, 1969, he was awarded an honorary doctorate at the
Technical Institute of Munich-late recognition of a lifetime devoted to
the unwavering pursuit of one basic idea. NSU Develops
th e Wanl«el

As wes MENTToNED PREvrousLY, Fe-

Iix Waqkelts rotary valve designs created an interest at NSU. This, in
turn, led to experiments with the Wankel compressor-the rnanagement
felt that the device had potential as a supercharger for motorpycle en-
gines. fn 1954, several test units of various sizes were built, and proved
remarkably effective. One test compressor delivered air to the engine
at pressures up to 8 atü (120 psi), with arl overall adiabatic efficiency
exceeding 70%. A small NSU engirre equipped with a Wankel qompres-
sor was installed in a motorcycle designed especially for setting speed
records. The basic powerplant had quite humble origins-it was the
50 cc. unit that normally propelled the NSU Quickly Moped. The
compressor was belt driven and delivered air at 1C-.15 psi to a pres-
surized carburetor, thus increasing the engine's power outpqt to 13
horsepower from its stock 1.6-a power gain of 8127o.
It was in these experimental comprpssors that Wankel's complete
seal grid, the result of his long-time work on sealing systems, 4nd the
two-lobed epitrochoidal working chamber and triangular rotor were
united.
The Wankel compressor consisted of two rotors, one inside the
other, both located within a common stationary outer casing. The
outer rotor had 4n epitrochoidal inner surface, which providod a work-
ing interface with the inner rotor. The inner rotor was shaped like a
triangle with curved gides, with a port in each face, halfway between
the apices. The central shaft did not rotate, but was hoJlow and served
as an air intake duct. Using a hollow mainshaft was not Felix Wankel's
idea; it was invented by ongof his closest associates, Dipl. Ing. Ernst
 HAIRPIN.TYPE
88 DEVELOPMENT VALVE SPBING
OO
Hoeppner, chief draftsman of Wankel's technical development center.
The mainshaft carried a control cylinder which, as it turned through a
certain angle, timed the opening and closing of the ports in the inner
rotor. The shaft also carried an adiustable control shell piece which
separated the atmospheric-pressure intake duct from the high-pressure
outlet duct and effectively blocked off the port during compression. Air
entered the intake duct from the end cover, was fed into the inner
rotor and escaped through the face port into the epitrochoidal working
chamber. Continued retation brought the inner rotor to the point
where the port was blocked by the adjustable control sleeve, signaling
the end of the intake phase and the beginning of the compression
CAMSHA

phase. After maximum compression, the port was aligned with the out-
let duct and the compressed air was released. To ensure adequate out- CAMSHAFT .
DRIVESHAFT
tt.

Basic patent drawing covering the Wankel rotary compressor. The

circle R lies inside the circle K. The numeral 2 is the radius of
circle R and the numeral 3 denotes the radius of circle K. Wankel EXHAUST
proposed to let circle K remain stationary and letting circle R
revolve inside it. Allowing points D' and D', oD extended radii, to
roll with the circle R, produces the three-lobe hypotrochoid T. If EE.BEAR ING
BUILT-UP TH
the circle K is allowed to rotate around circle R, different curves CUTAWAY FL\AruHEEL
CRANKSHAFT
R

result: a two-part hypotrochoid consisting of an outer lobe H. and

an inner lobe Hr.
The 500 cc. two-cylinder engine used in the record-breaking motorcycle had a rotary
Wankel compressor. Bore was 63 mm. and stroke 80 nlm. giving 199 cc. displace-
Ha ment. Power output was I l0 net horsepower at 8,000 r.p.m.

Iet phase duration, the central control cylinder would swing along with
the inner rotor for about l0 degrees, then return to its normal position
in anticipation of the next intake phase.
With this design, one phase was in progress at all times on each of
the inner rotor's three faces. When one face completed an exhaust
phase, the next was about to complete its intake phase, and the third
was about two-thirds into its compression phase. The inner and outer
rotors revolved in the same direction. Both rotors had simple rotation
around the mainshaft center; however, their rotational speed differed
to produce the volumetric changes that make compression possible. The
speed ratio was 3:2, the inner rotor being controllecl by gears in the
outer end cover. Every time the outer rotor made three revolutions, the
inner rotor revolved twice.
Of course, Wankel realized almost from the outset that the epitro-
choidal chamber and the triangular rotor invented in March of l9r1
was adaptable to operation on the four-stroke principle. The Wankel
compressor was, after all, only an air pump and, like all air pumps, the
addition of fuel and spark converted it into an engine. The develop-
 !
I
t
t

The DKM-54 conrplctely assembled, but partially cut away to show both rotors.

ment of such an engine was the next step that Felix Wankel and NSU Cross-sectional aspect of the DKM-54.
were to take.
Experiments with port locations showed that the four-stroke cycle
was, indeed, the most suitable mode of operation for this engine. A by the rotor apices during rotation. It was Professor Othmar Baier of the
decision to concentrate engine clevelopment based on the four'stroke Technical College in Stuttgart who demonstrated that the shape was,
cycle was made by April 13, 1954. Wankel spent the following 7y, in fact, äD epitrochoid. This was of more than acadenric interest, because
years on intensive research and development of his engine. At this time it was to facilitate mathematical analysis of the engine and simplify the
he was assisted by , staff of NSU engineers, Ied by Dr. Walter F roede. working out of practical methods for machirring the working surface.
Drehkolbenmotor (usually shortened to DKM) is a German worcl
which simply means "rotary piston engine." It was the name given to Inner rotor of the DKM-54 carried spark plugs in its faces and had
the motorized version of the Wankel rotary compressor, which was ports in its sides.
first tested in the NSU factory on February l, 1957 . At first, it could
not be made to run under its own power. After some carburetor adjust-
ments, fhe engine ran for several minutes-long enough to get a torque
output reading. The Wankel engine was a reality.
In the course of his geometric studies, Irelix Wankel came to recog-
nize that epitrochoids with inner envelopes were the only type that
promised a combination of realistic displaccnrent for autonrotive en-
gines and useful compression ratios. FIe arrived at the shape of the
working chamber empirically and did not identify the curve described
92 DEVELOPMENT NSU DEVELOPS TI.IE WANKEL 93
Geometrically, the DKM Wankel engine was identical to the rotary cessitated slip-ring electrical connections, and made cleaning or replac-
compressor, but there were notable differences in their mechanical ing the plugs a major operation.
make-up. For comparison with current Wankel engines, it is important
One maior difference between the compressor and the engine was that
to note that there was no eccentricity in the rotor motion of the I)KM. the rotor faces had no ports, the intake charge was Ied through side
The volume variations were due only to the relatively flat faces of the ports in the inner rotor into channels in the outer rotor. Adrnission was
inner rotor. The output shaft was connected to the outer rotor. The controlled by the relative movement of the two rotors and the exhaust
outer rotor transferred torque-the inner rotor was incapable of such port was located in the periphery of the outer rotor. Slicling speeds of
duty because it did not run eccentrically on its shaft. the apex seals in the DI(M-54 varied betwcen 3,700 and 9,600 feet per
The first developrnent engine, DKM -54, was a very complicated de- minute at 15,000 r.p.nl. Sliding speeds in such an engirre are highest
sign, built to test the trochoidal concept and a variety of proposed
around the major axis and lowest around the minor axis-the averdge
sealing systems. The number 54 was simply an indication of its cham- sliding speecl in the DKN{-54 rvas 6,700 feet per minute.
ber volume, measured in cubic centimeters. At this stage, no thought Tests of thc DKM-54 proved fcasibility be1,s114 a shaclol of doubt-
was given to production methods and costs-the object was to establish the validity of the fundamental principles had been established. While
feasibility and obtain confirmation of the basic theory. Wankel continued to stucly all possible gcomctrical configurations in an
The fresh mixture entered through a carburetor at the side of the endless search for the ideal shape for both rotor and working chamber,
engine and was admitted to the working chamber via the hollow shaft the NSU engineers under Walter Froede concentrated on developing
of the inner rotor, intake ports irr the inner rotor, and ports in the the existing concept.
inner rotor faces. The spark plugs were mounted centrally in the faces Among the early Wankel engine configurations exarnined was a de-
of the inner rotor and revolvecl with it. The spark plug location ne- sign with a thrce-lobe working chambcr arrcl a four-lobe rotor. This ne-

Elevation of the first Wankel engine, the DKM-51.

NSU Wankel KKIU-lZ5

94 DEVELOPMENT NSU DtrVELOPS TIIE WANKEL 95

cessitated a six-phase operational cycle, and the idea was rejected. De- for a given charnl>cr clisl>laccnlent than otlrer configurations lraving nlore
velopment concentrated on the two-lobe chamber with the triangular Iobes. This proved to be a great advantage in reducing the size and
rotor because the two-lobe design provided shorter overall sealing length number of Ieakage paths.
In mid-l 957, a new and Iarger Wankel engine was designed and
built-the DKM-125. Its charnber volume was 125 cc. (7.6 cubic inches).
Ultimately, its power output was raised to 28.4 horsepower at 17,000
r.p.m. (outer rotor and output shaft), which equallecl I1,300 r.p.m. at
the inner rotor. The inner rotor had a 65 mm. radius and a width of
38 mm. The R/e ratio was 6.85:1, which allowed compression ratios up
to 17 .6:1.
Spring-loading was not required for the apex seals, because the apex
seal swing angle was limited to 26 degrees. Th. DKM-125 was fully
balanced and did not have any mechanical vibration. Bearing loacls re-
sulted from gas pressure only, because both the inner and outer rotors
revolved around their own centers of gravity. Both rotors revolved in
the same direction and were connected by phasing gears that positioned
the two rotors but had nothing to do rvith transmissiorr of the engine
torque. The outer rotor was made of nitrided steel but, despite the
use of high-grade metal, it tended to distort from true epitrochoidal
shape under centrifugal force. The rotating inertia in the outer rotor was
so high that it had a detrimental effect on the engine's ability to ac-
celerate-it wanted to run at a constant speed. The inner rotor, instead
of being one complete casting, was built up from a number of com-
ponents, including separate side platcs and a variety of seals and seal
parts. Initially, the inner rotor hacl no provision for cooling (the outer
rotor was water-cooled). Later, water cooling was added to the inner
rotor-with considerable cornplication.
.--\ Port sealing was a big problenr because of the high rubbing velocity
of the outer rotor. It was found to be almost iurpossible to make a
-r
satisfactory gas seal for the exhaust port, where high gas ternperature
aggravated the already unfavorable conditions. The problem also was
aggravated by the fact that friction losses in transmissions increase with
a rise in gear rotational speed, and the output shaft revolved at outer
rotor speed. The stationary nature of the nrainshaft adclcd to the clif'
ficulty of provicling a suitablc transnrission. Irffort.s were rnacle to Iinrit
the operational speed without engendering unacceptable sacrifices in
power output.
This sketch shows how the use of circles R and K, alternating in Volunretric efEciency was fourrcl to bc a8% at 7,000 r.p.rn. ancl 70%
thc roles of base circle and generating circle, can create both at 16,000 r.p.m. The DKN,{ was testecl at speeds up to 25,000 r.p.rn. in
two-lobe and three-lobe epitrochoids.
Belou,: Proposals for inner rotors revolving in the same direction order to prove the mechanical safety of the engine. Peak torque \vas
as the outer rotor with the inner rotor doing two-thirds the speed reached at 8,000 r.p.m. and peak mean effective pressure was 120 psi.
of the outer rotor at left and three-quarters the speed of the outer
rotor at right. T'he engine ran most ecoilomically at 12,000 to 14,000 r.p,m.
96 DEVELOPMENT NSU DEVELOPS TFIE WANKEL 97

The DKM-125 had been designed without regard to installation prob- DKM with a stationary outer housing. In other words, he decided to
Iems and transmission requirements. The dual rotation made a stationary prevent the outer rotor from rotating, and to confine rotation to the
outer casing necessary if the engine was to be installed in a velticle. inner rotor and the shaft.
This would have added substantially to the weight and cost, and the This line of thought led him to the invention of what he called
obvious course was to redesign the power unit with vehicle installation "kinematic inversion" and to the introduction of eccentric rotor move-
and power transmission in mind. This work was carried out by Dr. ment. These principles were first combined in a new engine called the
Froede and the research staff at NSU, the result of which was a concept KKM. KKM stands for Kreiskolbenmotor, which means "circuitous
that made the Wankel engine truly practical. Dr. Froede discarded the piston engine."
idea of adding an extra casing and concentrated on redesigning the
tln KKM' shown in his basic patent
Walter Froede shows off the KKM'125. .oJ:,,',: "-''ä:'J:,,?l ,IJ3,',ffi.

ou3' 2''T,
. cEilrilrr* .X^ror' m:^"r=*rustto* *.?,lln,'2go
?lLd llot. l!, lSOt ll !üootr'ttoot ll

tlwENTOß
ÜALTEN 6. FNOCDE
98 DEVELOPMENT NSU DEVELOPS THE WANKEL 99

The kinematic inversion consisted of transforming the outer rotor of very slight friction loads and efiects of random variations in engine
the DKM into a stationary housing and mountirg the inner (and only) ,.p.*. Tlr. annular gear on the rotor's inner surface had 45 teeth. It was
rotor on an eccentric. The mainshaft also became the output shaft, and, meshed with a stationary 30 tooth reaction gear locked into the end cover
due to the use of a stationary block, the installation problems asso- to give a 2:3 ratio with the rotor.
ciated with the DKM were eliminated. To keep the rotor in phase, a re- Dr. Froede's first KKM was a 12, cc. single-rotor unit, built and tested
action ger was shrunk into the end cover. Its function was merely to in lg57 . The KKM- LZ| weighed 37 .4 pounds with a cast-iron housing,
keep the rotor in orbit. Reaction loads were low, since the gears carried
no torque. Theoretically, no forces acted on the phasing gears except
Froede's patent covered multi-lobe rotors and epitrochoids with up
to six lobes.
One of the configurations covered by Dr. Froede's patent is this Aug. 2, lg00 w. o. FR,EDE lgl7l,,g9
DKM with a two-lobe inner rotor and an almost circular outer iorARr tntERNtL c'EtlnttT.
rotor, eccentrically mounted in relation to each other. ,frI*tßttllro T.:I-.-r rl
,*sr ror.
.
Au3. 2, 1060 uv. o. FioEDE 2rg472g0
HEIT CEHERAIINC ROTAü IilTENfiAL MUSilIOII EIIOIIIE

e
?tlod llor. lt, leot ll Stotr-Soot l0

§
\0
@\
,lp lo

N
@\
b@
,rfl

tlwENTor
YALTER O, FNOEOE lllvENrOr
UALTEN A FfrOEÖE
 s
1r'

Au3. 2, 1900 w, o. FRoED: 2rg47r2g0 ;l

:.
IGIT GENEilTNO NOTART TTTENTAL OOIE'§rlOil EIOTNE
TthC lloY, t!, lClt It lürotl.tLrrt ,

Closeup of the KKM-125, with one end cover cut away to show the totor.

and only 23.2 pounds with an aluminum housing. This engine repre-
I},Il/ENTOR sented the first combination of Felix Wankel's sealing grid, the epitro-
UALTER A FNOEDE choidal workirg chamber, and eccentric rotor movement, and was the
basic Wankel engine as we now know it.
The invention of the KKM led to simplified breathing and ignition
systems. The mixture was fed into the workirrg chamber through a rirani-
fold and ports in the housing, instead of through the shaft as in the
DKM. The spark plugs were rerrioved from the rotor faces and installed
An elaborate version of a KKM with side ports, patented by Dr. in the housing (made stationary) which had great operational advantages
Froede in 1960.
for the ignition system.
 NSU DEVELOPS Tt{E WANKEL IO3

seals on the eccentric were relievecl of pressure, serious leakage stopped.

Improving the oil seals themselves was not regarded as a high-priority
matter. One of the biggest parasites to power output was found to be
the turbulence of the-oit inside the rotor. NSU instituted a new line
of research in order to achieve better control of the oil flow. This led
to the development of a phased oil extractor inside the rotor, extencling
to each apex.
AnotSei problem, rotor bearing seizure, was cured by adopting needle
roller bearings,in place of the plain bearings. This led to a complex
assembly, *hi.h was in turn replaced by two tracks of double rollers,

The housing for the KKM'125

D trf. r.a a.... Lt. ttr
had water passages contoured
for heat distribution differences.
Both ports were peripheral. The
spark plug was positioned almost
llere all the parts of the KKM.125 are placed in proper relationship to each other, eiactly at the minor axis. (@
spaced out to show each individual part more clearly. (@ lliffe Technical Publi- lliffe T echnical Publications)
cations)

At first, Felix Wankel strongly opposed the KKM because he felt

that the introduction of eccentricity to the movement of the rotor
placed his entire invention in jeopardy. He insisted that his sealing This exploded view gives clear
details of the rotor sealing system
system, äs applied to the DKM-125, was the basis of the success of the used on the KKM'125. The
early prototypes, and he did not approve of the sealing modifications rotor's inner gear was a seParate
necessary for KKM operation. The eccentricity of rotor movement gave piccc which bolted in, meshing
with a stationary reaction geet.
rise to fluctuations in apex seal loadings, including negative centrifugal (@ Ilifre Technical Publications)
force loading every time an apex approached the minor axis, while the
apex seals in the DKM engine operated under stable conditions with
unidirectional centrifugal loads. When it turned out that there were
no fundamental problems with the sealing system of the KKM-175,
Wankel was mollified and began ioint experimental work with NSU
on new KKM designs.
The sealing grid which Dr. Froede developed for the KKM-125 was
both simple and effective. The corner seals brought the apex seals in
direct contact with the side seals. The trunnion which carried the apex
seal had two small segments cut from its end to allow access to the apex
seal for the side seal ends. The trunnions provided a longer leakage
path for gases trying to escape and thereby slowed down the leakage The rotor runs on a roller bearing carSled on the sculptured. ec-ce1t!,.,
gas flow. Apex seal angular swing was restricted to between 20 and 30 which was inielral with the nra-in shift. (@ llifte Technical Publi-
degrees per side. This cured 90% of the oil problem. When the oil cations)
r04 DEVELOPMENT
Elevation of the KKM-Z50
staggered in relation to each other. Later, NSU reverted to plain bear- gtne.
ings, after it had been established that the type of rotor bearing used
had not been the principal cause of the failures. The bearings froze
not because of defective bearing operation but because of centrifugal
distortions in the rotor. The rotors that suffered from this trouble were
made of light alloy. At that time, Iight alloy was considered preferable
to cast-iron because of its light weight and high conductivity. It had
been adopted because it was thought that rotor ternperature distribu-
tion would be very even and no distortion problems would crop up.
The rotor was cooled not only by the oil inside, but also by the in-
coming air-fuel charge, and to a small extent by the apex seal tips
through their contact with the water-cooled working surface. The spark
plug hole in the housing presented an interruption in the surface swept
by the apex seals. This was no problem in terms of seal durability and
sealing effectiveness, but it did mean a high risk of local overheating
of the horrsing metal around the spark plug hole. That in turn led to
increased risk of cracking in that part of the housing. The problem
ultimately was solved by developing a cooling system with higher ef-
fectiveness in high-temperature areas and lower effectiveness where little
cooling was needed.
When Dr. Froede was satisfied with the state of development of the
KKM-125, he designed a new engine with twice the displacement, the

Test data on the I«M-I25

engine. N= horsepower. Dreh-
zahl - r.p.m. Verbrauch : fuel
consumption.

of the KKM-250
 Test data for the KKM-2 50 en- / -tl \ss_, .\r
grne. N
= horsepo*et: P" = / /
2-
\ \
mean effective pressure. Ver-
brauch : fuel-consumption. §§ f- -\,
\ ,9
Drehzahl : r.p.m. \
ä
\
§
a

i
I l§
--|> RPN
KKH 25oE
BnEi veaffieeapusnaL NqAKE
I
PORTS OF YARIOUS SIZE
t
lt 22rttt DlAll 2.lex 27nvn ßI€CfANOUI-AQ

Mean effective pressure graph for the KKIU-250 rvith two clifferetrt size intake ports,
throughout l2,000 to 9,500 r.p.nl. spccd rlngc. Port I rvus circulrtr with n diittttctcr
of 22 mm. Port 2 was rectangular, l8 x 27 mm.
Eii&mn
il
t.,

t
I(KM-250, drawing fully on all the lessons learnefl from experirnental
i'
i

work with tlt'e earlier model. It was de.signecl as a test enginä to estab-
Iish the best spark plug position ancl to ex1>lore ncw.o*b,rstiop cSarn-
ber configttrations. Tlre nrainshaft ran'in one plain bearing on eacS
side of the rotor, while the rotor ran on plain bearing strätts on its
eccentric. (The initial design carried the rotor on a needle roller bearing, 16
but was later modified to use a four-race roller bearing.) Peripheral ports
were used, and lubricating oil was brought to the bearings lhrougl, the
partly hollow mainshaft. Radial channels in the eccentric
then fäa the
oil into the rotor for cooling. Fuel consumption plotted against
The KKM-250 was about 9,5 inches in outside diameter, with mean effective pressure in the
a
r'r'idth of 7 inches. The two peripheral ports were positioned fairly close KKM-2 50 at a constant speed of
5,500 r.p.ffi., with a peripheral
together near the minor axis, which produced higü phasing overläp. An intake port of, 22 mm. diameter.
aluminurn rotor was used, and the compression ratio was 8.0:1. The
spprk plug position was 20 mm. behincl the minor axis. The engine &t'tEPttbßgN
was
equipped with a 32 mm. Solex carburetor and the flywheels *är. fixed
to the output shaft entirely by the friction of the cone clusters set up
by end thrust loads. An R/e ratio of 7.77:l was used, which offered the
KKtl 25oß
possibility of ntnning with compression ratios up to 19.5:1. The
maxi- 8.SFC AT PART LOAD
mum apex seal swing angle was 23 degrees, rotor radius was g5.5 mm. COt'lSfAN T SPEED n. SOoRPt'l
PERIPHER/IL INTAKE
2ZrrntAAf,
IO8 DEVELOPMENT NSU DEVELOPS THE WANKEL IO9

and eccentricity was I I mm. As in all NSU The I«<M-2rO had an installation weight of 48.4 pounds. Power out-
engines, rotor width was
about four times the eccentricity. put rose in an almost straight line-I0 horsepower at 2,700 r.P'T', 70
The KKM-250 underwent endurance testing in luly, 1959. These irorrrpower at 4,300 r.p.m; 30 horsepower at 5,500 r.p.T., 10 horse-
tests confirmed the basic soundness of both concept and design, and po*ri at 7,600 r.p.m. .r,d a peak of 4l horsepower at 9,000 r.p.m. Peak
successful 1,000-hour tests were completed by 1960. Another series of iorqu. (zg foot pounds) *ri developed at 6,000 r.p.m. Mean effective
tests were made to find ways to reduce material and machining costs, pressure rose in waves fro* 80 psi at 2,500 r.p.m.. Above 6,000 r.P.m.
pressures remained high, off slightly to 12, Psi at 9,000 r.p.m.
with particular reference to the trochoidal bore and the finish of the -*mfalling
lOO7, or better at atl speeds above 5,500
end-covers. Volumetric efficien.y

This graph shows the effect of variations in air-fuel ratio on specific This graph shows the effect of chan-ges. in ignition advance on
fuel consumption and mean effective pressure in the KKM-250 rp..inä iüel consumption and mean effective pressure in the KKM'
englne. 250 engine.
t,

§§
§
ä
\J(
§
{
t.t

u
$
Q,,
I <> t6N trpN 1nEhNA.5 OF CÜ1HRAE1E req R SNq

I
I t
qtz qp q8 qü qo
KKH zsol§
-> FAR

BNEP AND ESFC YERSUS

FUEL.AIR RATIO
IIO DEVELOPMENT NSU DEVELOPS TIIE WANKEL III
r.p.rn., the low point was 82% at 4,500 r.p.m. At 3,500 r.P.m., the en-
gine breathed at 90% of capacity. Specific fuel consumption was almost
constant at .675 pounds per horsepower hour from 3,500 r.p.m. to
7,r00 r.p.m. Above 7,500 r.p.m., however, fuel consumption rose sharply
to 0.78 pounds per horsepower-hour. irg surfaceof great harclness wore the seals rapidly. NSU clecided in
NSU used the KKM-250 to test hundreds of different shapes and favor of softer ieals and harcler working surfaces, reasoning that it was
materials for both single-piece and cornplex apex seal assemblies. With better to fit new apex seals than to replace the much more expensive
some systems, complete failure occurred after only lr-20 minutes and complicatecl housing which containcd thc cpitrochoiclal surface.
running; other systems that were identical geometrically, but had dif-
ferent material combinations, survived endurance tests of 1,000 to 2,000
This graph shows horv heat transfer to the cooling media (ryater
and oä1 urried according to r.p.m., relative to the heat value of the
KKM-Z 50's outPut.
sas temperature in the I«M'z 50
IH f,',x'1"'Hä:,lJä ::l.H:' I

llt
t
1,

t
;
F
I
I
I
:'
t'
§
F §§ WATER

t
\I
§§
H
i. §§).
3 I

R§
* I
R{
HR
§§
J Txtot 20

RPt'l §§

t JI Tttot
RPH
EXIUUST TENP AT TULL
THROTTLE ßßM 250-7
HEAT DISTRIBWION
FULL THROTTLE
 NSU I)EVIILOPS 'f'I IIq \VANKIr]L l 13

Froecle concludecl that flame speed was superimposed on the tangential

gas velocity as the mixture left the trailing part of thc working clunrbcr
past the minor axis ancl entcrcd the exparrcling, lcading portiou. Gas
velocity was almost invariably greater than the speecl at which the flame
could spread. The travel time for the flame front varied consiclerably
clepending on the air-fuel ratio.

In this endurance test, the KKM lost po\\'er as the apex seals wore
out. Whcn ncw apcx sclrls wcrc instlllccl, full power was restorcd,

I +ljJD
-ß12A9-r {D oATl
K$t1 2s0 -fl.ül!o--fl,t2Q-t
END OF A'OOHRS ENOUfuAIVCE TEST
55OO RPH FUL THruTTLE
This greph shows how mean effective pressure and specific fuel consumption patterns -l
evolved during a 100 hour endurance test of the KKM-Z50 engine. §ü
{ -( t{ -t11
s
§-i t \§
§t
The magnitude of the forces that worked to circulate the oil inside §_§
the rotor was first established on a KKM-}i}. In one instance, a small § §- üü
foreign body was left within the rotor. It wore away so much of the q§. 3
§_§ § §§
§
rotor wall near an apex seal that a failure occurred and a considerable
amount of oil was lost. NSU then began to investigate how this natural § §
(,
§
L
e.
L§
§j
circulation could be best utilized to cause oil to flow to and from the l{
t)
rotor without the assistance of pressure-feed or scavenge pumps.
§ §
NSU also developed a way to study flame propagation. The method
was based upon conveniently spaced ionization detection plugs. A spe-
}-
q §_i
cial KKM-250, with five detection plug positions, was prepared for this
investigation. The detection plugs were interchangeable with the spark t §-§
N
t2
§s
5 ttt
plugs. Position I was 56 mm. before the minor axis, position 2 was
24.5 mm, before the minor axis, and position ) was 45.5 mm. after the PERATINq Tfi-fE tUASl
minor axis. The fourth position was 77 mm, ancl the fifth position 154 -F KKl,r 250 l9
mm,, after the minor axis. Separate tests confirmed that the beginning of
ionization coincided with the arrival of the flame front. Quartz windows LOSS OF PC)WER DURINO
with photocells behind them were installed to observe the combustion
process, then tests were made at various speeds and with various compres-
AN ENOURANCE-RUN
sion ratios with the spark plug mounted in one of the five optional NO PAPT§ EXCHAIIaED
positions. With rising rotor speed, gas velocity increased, as did flame
CONS fAN T .JPEED. 5OA0 R Pt|
front velocity. Consequently, the combustion process was by no means
as slow as the elongated combustion chamber shape indicated. Dr. C0N,ITTANT LOAD.tOZ tbs/sg rn
 t
IT4 DEVELOPMENT I
t
t
Many tests were performed to find the best spark plug position t
Ir
and the highest practical compression ratio. Compression ratios varl,ing t
betwccn 6.8: I arrcl 12.6:l were used rvith a symmetrical combustion
I
I
t

chamber in each rotor face. The effects of placing the spark plug in two t
I

alternative positions in front of the minor axis, and similarly, in two

t
I,,
l
ä
ta

l
Wear of a1>ex seals on the KKM-250 (thousandths of an inch) in a I
300-hour enclurance tcst. Only the scal tip in contact with the I
i
epitrochoidal surface has significant wear. The sides of the apex seals ;-
wear as wcll as the rotor side seals, on completion of an initial 200- I
hour pcriod of breaking-in. t..
§'
l,
?
I
I
i:

!l',I ff
i

HY#;' :1 il 1:l tl];;:,täl' ?1.,äi [i,';i" ;113'äi']

i "iff
study of flanre propätgation. Tlre dctection plugs are thrcaded in iust
like spark plugs, ancl arc interchangeable with the spark plug. llere
the spark plug is in position Z (just ahead of the minor axis).
I

positions after the minor axis, were tested. Until the onset of knock both
spark plugs were fired simultaneously, ancl power output, efficiency and
§
\ fuel consumption curves were plottcd for the rvhole specd range and
a
§ I for lean mixtures. Maxinrum rnean cffective pressure was obtained u'ith
compression ratios betrveen 9: I and I I : I , u,hile mirt iruum fuel con-
s sumption was obtained witJr a compression ratio of 9.5:1. The highest
q pressure at the lower compression ratio was obtained with two spark
plugs firing simultaneously, while at higher compression ratios the best
§ performance was achieved with a single spark plug in front of the rninor
axis. Dr. Froede concluded that a single spark plug placed after the
rninor axis allou,ed lorver pressure and ignited air-fuel mixtures of varying
I ratios with greater certainty.
2
Tests with multiple spark plugs failed to shorv conclusive evidence
, P OF §IOE ,.EAL§
that the combustion process could be accelerated by multiple ignition,
0 but it seemed that power output coulcl be raised by about 5% by install-
0 50 t00 150 fro ing two spark plugs spaced at certain intervals. Extensive experirnental
--*PERATN6 NHE work was necessary before satisfactory low-speecl opcration and idling
were attained. Finally, it was possible to maintain 600 r.p.m. output shaft
speed, at a corresponding rotor speed of 200 r.p.m., for many hours with-
KKI'? 250/6 out irregular ignition ancl combustion. \
\IEAR 0F §ElL- CO|rPOlIEilfS The experimental departnrcnt of Goetzeu,erke, thc lcacling pistorr ring
manufacturer in Germany, played an important part in the clevelopnrent
OURINO ENDURANCE RUN
CONSL r SPEEO 5000RPn
of the various oil sealing principles. One tyl>e of oil seal was so effective

C0NS7. LOAD r
that even after 200 hours runnirg the amouut of oil Iost \l,as too small to
,O? lbs/s?i? measure. At the end of 500 hours, the combinecl losses of the lubrication
116 DEVELOPMENT
and rotor cooling systems amountecl to Iess than 0.00441 pounds per
Test data on the KKM-150 en-
h orsepower-hour. gine. PS : horsepower. g/PS-hr a
o- I

Cooling oil circulating througlr the rotor also wets the encl-cover sur- -- grams per horsepower-hour. .7 s.
'.4
Spec. fuel cons. - specific fuel
.6<
faces, and thus lubricates the side sealing strips andjoint trunnions. Until €
consumption. lb./in.2 - pounds o.l
1963, 1-2% Iubricating oil was added to the fuel to ensure lubrication of per square inch (psi).
the apex seals. Then NSU realized the advantage of metering the oil
separately into the cylinder and arranging for the metering control to
I
respond to variations in load and speed. One KKM-250 engine with a t'
i-t
t.
cast-iron housing and nitrided working surface underwent a 500-hour
t,.

t,.

endurance run at a constant 5,500 r.p.m. with the fairly high mean effec- I

tive pressure of 130.8 psi. For this test, 2% Iubricant was added to the
I
1:

t,

n.0t-R
fuel, A test of 2,000 hours was successfully completed by, similar engine. i
Sholt rpecd p m.

Shaft speed was maintained at 5,500 r.p.m., with a mean effective pres- signed to run with a lubricant added to the fuel in a l:40 ratio. This
sure of 105.8 psi. Although chatter marks appeared on the working sur- Iubricarit proportion proved aclequate for all bearings and sealing ele-
I

face, to a depth of 0.0031 to 0.0035 inch, regrinding the surface restored ments. The engine developed about 5 horsepower at 6,000 t.p.m. and its
I
I
b
I
the original output characteristics. The longest full-load test before 1963, j axial fan supplied sufficient air for adequate cooling of the end covers
at a mean effective pressure as high as 147 psi, Iasted for 700 hours. By ;
., and central casing. Another, the KKM-I50W engine, was water-cooled
reducing the mean effective pressirre to 105.8 psi, an endurance test of and capable of developing between IZ and 20 horsepower according to
i.
L

1,400 hours was successfully completed. By then, the cast-iron seals were the cross-sectional area and the timing of the ports. The rotor of this
worn out, but the side-sealing elements and joint trunnions remained engine was, of course, oil-cooled. Another version of this engine was
good for twice this periocl. developed specifically for marine applications.
fu a passenger car manufacturer, NSU naturally pursued Wankel en- The shape ancl proportions of the epitrochoidal surface of the KKM-
gine research and development with automobile applications in mind, l50L engine were identical to those of the KKM-I 50W engine. This
but the company remained sensitive to other potential areas of applica- engine, however, was air-cooled. After preliminary stuclies, pr'eference was
tion, and designed and tested many units far below practical motor given to circumferential cooling fins and a radial blou,er. TIte engine
vehicle size. For example, there was the air-cooled KKM-60 unit, de- proved capable of a continuous power output of about 12 horsepower
with housing temperature linrited to 200oC. The weiglrt allowance of
the KKM-I50 was as follows:
[Iousing light alloy 2.38 lbs.
Rotor, complete u'ith seals 2.27 lbs.
End cover-pump takc-off end-cornplcte
with oil pump, race and connectors 2.27 lbs.
Eccentric and bearings 1.87 lbs.
End cover-dynastart cnd-rvith pinion and
pipe con ncctors 4. I 9 lbs.
Irlyrvheel-output end +.96 lbs.
Miscellancous snrall parts \ 1.32 lbs.
Total 19.26 lbs.
D1'nastart, blo\\,cr artd bllancc n'cight I +.7\ lbs.

It was fourrcl that friction losscs variecl alrno.st lincarly bettvcen 2,000
r.p.m. and 6,000 r.p.rrr.; the corresponding nreall effective pressures for
Elevation and cross-section of the KKM-150. tlrese speeds were 22.05 psi arrd 76J, psi. 'flte change in thc ratc of oil
 NSU DEVELOPS TTIIT \VANKIiL I 19

circulatiorr was virtually linear-bctu,een 7..64 ancl 6.16 pounds pcr mitt-
ute. Power output exceecl ccl 24 horsel>o\\'cr at 8,000 r.p.nr. IIou,ever, this
potential power output ne\/er was fully exploitecl becausc the port sizes
and timing were altered, which, in conjunction u,ith the irrstallation of
an air filter and an exhaust muffier, considerably reclucecl the effective
power output. A version of this unit for marine propulsion developed
over l8 horsepower at 7,000 r.p.m. Between 3,000 and 6,500 shaft r.p.m.
the mean effective pressure exceecled I 17 psi and the fuel consumption
was 0.691 pounds per horsepower-hour.

Test rcsults obtainecl u,ith thc KKM-100. lbs./HPll = pounds per

horsepower-hour. BSI'-C = specific fuel consumption.

of the KKM-400 50

+5

i0
Elevation of the KKM-100 cngine.
t5

§, \N
\
p0
§" §
lro (,
fq) R
fE 90
t0 80 §
§
I

o , 2 3 ta 5
+ RPN
KKn t00-l
rESTSf ANO PE RFORT'ITßE
,flTHq'T HUFFLER
 I2O DEVELOPMENT NSU DIi\/EL,OPS 'l'l lli \\/ANKLlt, l2l
Nornral high-temperature spark plugs were used extensively for the cally constant at about 7%. Within the nornl:ll perfclrrnance range of
development of the engine. Special plugs, with long central electrodes the engine , l37o of the heat was lost to the cooling rvater and )6% ex-
that reached close to the workirrg surface, were also very successful. These pellecl *itt the exhaust gases. The amount of heat lost by convection
wcre su1>erior for starting, idling, arrd running under no-load or negative- I ancl racliatiorr was not cletennirrecJ rcadily; Itou'cvcr, it was thought to be
Ioacl c«»clitions. A variarrt of this typc plug was dcvcloped in collal>ora- about ?. to 5"Ä,
tiorr with llobert Bosch G.nr.b.I{. and it gave good starting and idling Tests witS the KKI\{- 750 ancl KKI\{-400 cngincs showe«l that thcre
',i
I

I
clraractcristics, in addition to having a life of more than 200 hours under :
were notablc cliffercnces in tlrc relation.ship llctween coolirrg watcr tcnr-
full Ioad. 'l'his experirnental plug was choscn for developmcnt irrt«l a .L

pcraturc apcl cxlurust gas tcrnpcraturc. J'ltc tctttl>crittttrcs \l'clc Ycly cl«lsc
production design. Ignition by a standard contact breaker assembly was i, engines with high n/. ratios, such as tltc KKM-?,r0, but thc coolant
used, but experinrents also were conductecl with transistorized electronic tenrperature remainecl about trvo-thircls the exhaust gas temperature in
ancl piezoelectric systems. .ngin., wit5 low R/e ratios, such as the KKI\4-400. Both engines were
But it was the KKM-400 that was to Iead NSU directly to the first tested under full loacl at 4,000 r.p.m. shaft speed, with similar air-fuel
Wankel-powerecl autonrobile in the world. The KKM-400 engine had a ratiog. Dr. Froede concluclecl that the KKM-400 requirecl a smaller radia-
single chamber with a displacement of 400 cc. (24.4 cubic inches). Its tor, when installed in a car, than dicl the KKM-250. Other differences
gencral configuration was identical to that of the KKM-I 50 unit. The were tSat the KKI\{-400 ran with slightly higher oil temperature but
KKM-400 was shown to have a more favorable surface-to-volume ratio derived more useful energy from the heat value in the fueJ.
tlran its preclecessor, the KKM-750. It had a :t 28 degree angle of obliq- Some experirnental .ngi,r.r sufferecl rotor ancl gear failures. Several
uity, as opposecl to the 23 clegree angle of obliquity of the KKM-250. research teams measurecl the clestructive forces present ancl attributed
Power output varied between 40 and 50 horsepower. f)ifferences in power tSem mainly to shock loads created when the pressure rose durirrg com-
output and fuel consunrption coulcl be obtained by simple changes in the bustion and to s|ock waves set up whcn the exhattst port was uncoverccl.
carburetor settings to give either maximunl power output or minirnum Other possible causes were gas pressure acting on the rotor, and rota-
fuel consumption. The lowest fuel consumption, 0.487 pounds per horse- tional forces. None of the early NSU Wankel engüres had any of the
power-hour, was recorcled at 4,000 r.p.nr-this increased to 0 .ß6 pouncls cold-starting problcnls that the experts had predicted. Dr. Iiroede reported
per horsepower-lrour at nraximum output. irr 1961 "Contrary to popular prediction wlren the errgine was first irltro-
By 1962, the engine hacl reached a high level of efficiency. If fuel heat cluced, there has never been any starting problenr either at nornlrll anr-
inptrt is clefined as lA0%, it was found that up to 28% of this energy was bielt temperaturcs or cluring very colcl u,eather. Slightly higher rotational
convcrtecl to useful u,ork. I-Ieat lost to the cooling oil renrained practi- speeds may have been requirecl for starting, but, on the other ltand, the
breaka,,r,ali torque is clecicleclly lorvcr than that of the colrventional recip-
"l-hese heat balance charts compare the KKM-250 and the KKM-100. The larger
rocating pisto,i engine. Conscquently alrclut the same Power is required
unit had lrigher hcat losses to the exhaust, but lo*'er heat losses to the housing for stariing the twä tvpes of cngines. A KKM-400 engine starts at 130 to
cooling systcnr, yet it had more of the fuel's heat energy available for useful power. 150 shaft i.p.*. and a 0.9 l',.p. stirrter motor fitted to an experimental in-
stallation in a car has proved entirely satisfactory under all conditions en-
Hcot -prr ccnt
t0 20 50 40 50 60 80 t00 countered over a very long period."
TSere was only one problcnr to rvhich NSU had no possible solution
in sight lry lg62-seal tip u,ear. T'ip wcar qr the apcx seals was so critical
tlrat in Iuly of 1962 NSU coulcl not run an errgine for more than 100
hours at'5,Ö00 r.p.m. without replacing the apcx seals. Between that clate
e
+
and |anuary, ß$, a serics of new matcrials was tested-and reiectcd.
e

Urcful heot 0it Wotr Erhougt

Unburncd fucl
in c r houst KKM 4OO In May and |une new carbon compouncls were developed, and solne ran
over ?50 hours at 5,000 r.p.m. Wear was within 10-15% of the previous
o ond rmrduoI
o
o
v
Unburned fucl year's engines. Iiurther cJevelopment work led to seals that wore Z-7%
Urrtul hrot 0ir Y/otrr Erhourt in arhoust KKM 25O
ond r:riduol of the 1962 wear rate, before production of the KKI\4-502 engine began.
IZ2 DEVELOPMENT NSU DLIVELOPS T'l IIr \\/ANKEL 1Z)

This was confirmed in repeated tests exceeding 200-220 hours; after Irn- The protot),pe single-rotor 600 cc. engine had an R/e ratio of 7.68:1,
uarlr, 1965, tcsts up to 700 hours shorved even more wcar resistance, ancl which permitted compression ratios up to 19.7:l and ga\/e a maxinrulll
seal tip wear was firrally judged "satisfactory." apex seal swing angle of 23 clegrees. Actual cornpression ratio ufas 8.5:I,
Up to 1961, all NSU experimental engines hacl been built on the and seal tip sliding vclocitl,r,aried betrveen 44.7 ancl 108 ft./sec. at 6,000
sirrgle-rotor principle. When plans were formulated for a 1962 production r.p.m. A slightly smaller design had a chanrber volume of 447 cc. This
car with a new Wankel engine, it was felt that a charnber volume of 500 design's 5.66: I lt/e ratio gave a more pronourtced waistline and a maxi-
cc. wou]cl be sufficient to nreet the performance goals of a special sports mum apex seal swing angle of 32 degrees. IVlaximunl possible compres-
version of the Prinz. But there \lras considerable discussion as to rvhether sion ratio was 14.6:1, and seal tip slicling velocity variecl between 28 and
the 500 cc. should be all in one chamber or split between trvo rvorking 90.5 ft./sec. at 6,000 r.p.m.
chambers in a twin-rotor arrangement. NSU built its first twin-rotor The engine destined for the NSU Wankel Spider was the KKN4-i02,
Warrkel engine in 1961 . It was based on the KKI\it-400, having the same an enlarged ancl improved version of the KKI\4-400. NSU Sport Prinz
geomet y, but hacl narrower chanrbcrs of 300 cc. displacement each. Tl-ris cars equipped with the KKN,{-400 engine were test clriven for nrore than
KKM-2 x 300 was designed to develop about 60 horsepower at 5,000 500,000 rniles whilc the KKIVI-502 was on the clrawing board. The initial
r.p.m. With a clisplacement equivalent to l.Zliters (ß.5 cubic inchcs), it version', the KKN4-500, was the very first unit clesigned and clcvelopecl as
actually developecl 66 horsepower. The engine had been originally de- an autonrobile engirre. Its power output was Iimited to about 44 horse-
signecl for automobiles ancl nrarine propulsion, but was soon superseded power ancl particular attention wils paid to high torque at low specd-
l>y nrore advanced projects. obtained by the appropriate choice of port opening periocls, spark plug
Witlr this experience to go on, the engineering office began a thorough position, combustion chanrber shape ancl compression ratio-witltout
stuclv of both types. Basic geometry was identical, with an R/e ratio of straying front the principle of the peripheral intake port.
7.15:1, a maximunl swing angle of 25.5 degrees, ancl rotor widths fixed The KKIVI- 502 engine was designed in differcnt versions, rvith both
at 4.8 tirnes eccentricity. The single-rotor design was preferred, and ulti- peripheral and dual-side intake ports. Pou,er output was 30% greater
nrately selected, for a number of reasons. It
offered a considerably u,ith peripheral ports at 6,000 r.p.ur. Because the KKI\l-502 \\'as irttended
shortcr sealing path, had il nlore favorallle surfacc-to-volume ratio, ancl for use in a sports crlr, u,lrcrc rrraxinltur-l power was requirccl, NSU coln-
cost lcss to produce. mittccl itsclf to thc pcril>lrcral intakc 1lort.'['hc rcsu]tarrt lirck of sttrtlotlt-

Iload test rnileage shows fuel economy obtained with thrce different NSU cars using HPsr
the KKI\{-250 and KKM-100 cngines. (w perpmA ilürptl
.f5
ml -
üd-tith Hü

, PRINZ rrfi MA
60
| -o-
--o-- PRTNZ KKtt tM
t5 t -.-o- - sFnnfPHNr rrx+a/g
-

B
§
t
t Comparison of power output
curves for the KKM-502 engine
to §0 6t a0 gr§ with peripheral and dual
intake ports. The solid
side
line
->nilcs/b represents the peripheral intake
R0l0 fESf |llLEAeE (AeeUmtXlPr, ffi.1UtTSt port, the broken linc rcprcsents
AYERAOE PESULT,S ON IEYEI -POTO EONI OREOTOTUS dual sicle ports.
1@ 5(m
üür perd (c1m.)
 NSU DE\/ELOPS TI IE \\/ANKEL 17.5

intake port so positioned that tlte rotor closed it off and began the com-
pression phase 50 degrees after the major axis-earlier or later closing
caused a loss of power at lou,er shaft speeds.
, Both cast-iron and aluminum housings were tried. The aluminum
/, housing was far less susceptible to tcmperature variations between various
areas, keeping between l00oc. and l50oC. over about 120 degrees of
trochoidal surface, from the middle part of the compression phase to the
( beginning of the exhaust phase. Tenrperature never exceeded l50oC.,
-o e\/en cluring combustion. The cast-iron housing tenrpcrature exceedecl
I
I 300oC. cluring the combustion phase, with fairly steep build-up arrd fall-
0, lto
td
I off. From 50 degrees beforc the minor axis on the port side uP to 50
o
degrees before the minor axis on the ignition side, tentperatures in the
cast-iron housing closely paralleled those recorcled with the alumirlum
housing. It was founcl that housing temperatures in excess of 200'C. Ied
I

MAX. PERFORMANCE
to a breakdown of the oil film on the trochoidal surface. The aluminum
MAX. PERFORMANCE EXTRAPOI.ATED
housing was preferrecl, clcspite its higher cost, on the basis of its superior
soo txTeKE ct-oar STANDARo TtMtNc heat dissipative qualities. In the production engine, both the end covers
lOotXTeKE CLO§E and the housing \ /ere light alloy parts designed for low-pressure die-
casting, and the coolant passages were fornrecl l>y welding thin sheet metal
plates to the cncl flangcs ancl llort stulls. J'ltc rulll>irrg surface of eaclt cncl
-' cover was sprayed with a molybdcnum ancl steel coating by the Ferral
2000 aooo

ENGINE 3PEEO - RPM slc This graph shows spccific fuel consunrption curves for a number
1,5
of \\rarrkel cngines, and the averagc curve for reciprocating piston
This graph shows the effect of intake port timing on mean effective engines.
pressure, throughout the r.p.m. range of the KKIVI.502. Earlier intake
closing irnproves low-speed perfonnance, while later intake closing
raises high-speed 1>ower output.

1,2
ness at idle and under part-throttle conditions was not considered objec-
tionable in a sports car which was intended for operation mainly at l,l
higher r.p.m. The ports were cast-in, and their positions were chosen to
provicle a symmetrical port-opening diagram. Port sizes, opening periods lr0

ancl carburetor throat size were chosen to obtain adequate torque at rela-
qe
tively low shaft speecls. The maximum torque was 57 footpounds at 3,500
r.p.m., arrcl the corresponding mean effective pressure was 147 psi. At qa
Irigher speeds, the torque dropped to 47 foot pouncls at 6,000 r.p.m.,
wlrich indicated a maximum of ,).) horsepower. Fuel consumption at 07 Tloyo Kogyo 2t100
speeds between 2,000 and 5,000 r.p.m. was below 0.58 pounds per horse- SU KKM 502
q6 cw Rc 2-60
Power-ltour and the Iowest figure, measured at 2,500 r.p.m., was 0.503
1>clurtcls per horsepower-hour with the carburetor set for maximum power. :-NSIJ t«M 507 RcciProaling agines
I{ighest mean effective pressure was obtained at 4,000 r.p.m. with the
QS
q1 -' - -
btTry
nn
126 DEVELOPMENT NSU DEVELOPS TIIE WANKEL IZ7

process. Later, it was establishecl that no molybclenum coating was re- the side seal got into the rotor segment ncxt to the oil seal and threat-
quired. Instead, tlre inner surfaces of the end covers tvent through an ened to cause further oil consumption problems. The solution, worked
induction hardening process which made them even more wear-resistant out theoretically before a prototype was built, was to mount the metallic
than dicl the molybclenurn spray. The exact housing configuration was ring and its attendant rubber O-ring in a seal carrier designed to equalize
gas pressure on its leading and trailing sides, '*,ithout reducing the spring
clictated by production considerations, although circumferential ribs were
aclclecl to stiffen the assembly.
pressure on the oil seal. This meant providing a gas leakage path behind
The apex seals on the KKM- 502 were made of carbon; the side seals the seal carrier, opening when pressure built up ahead of it and closing
were made of cast-iron. The apex seals were radially disposed, and the when equal pressure had been established. This oil seal rvorked effec-
sicle sealing strips were arranged to overlap their Ieading joint trunnion tively when ncw but clcteriorated rapidly to cause high oil consumption
tangerrtially and abut normally against their trailirg link block. This at an unusrrally Iow mileage in the life of the car. A nerv system was de-
arrangement prevented danrage to the sealing surfaces even when wear velopecl ancl soon replaced the original clesign in production-the solu-
devcloped. I{owever, it clid open up a new leakage path, and this con- tion was to seal off the rotor from its eccentric bearing. The rotor was
figuration was discontinued when better methods were developed. The +
modifiect to carry a slip ring, matching up with a conical neck on the
failure of many durability tests during 1963 stemmed fronr burnecl or acljacent bearing carrier. Two seal rings, both in radial tension, were
broken apcx seals (carbon type). Careful selection of intpregnating Ina- nrountecl in the slip ring and kept the cooling oil front Ieaking out.
terials, closc attention to precision ntachining and finish ancl stricter The NSU cngineers wcre still coucerncd aborrt gas leakage past the
inspection ancl quality control resulted in an improvement in seal life side seals ancl its effect on oil sealing. Tlte new rotor hacl an oil scal in-
fronr the bcginning of 1965. By I968, engines ran routine tests of over serted in a circular slot irrside of the sidc seal grooves, tapered to block
1,000 lrour.s cluration without any further apex seal failure. raclial movernent of any oil that might escape that far. 'Ihis two-piece oil
.l"he
KKM-502 rotor was shaped to transfer a minimum of heat frorn seal was spring-loaclecl against the encl cover to telieve gas pressure build-
its faces to the hub bearirrg. The hub bearing ring part of the rotor was up ahead of it by provicling a leakage path around the oil sealing inter-
connectccl to thc inner walls of the rotor faces by a series of bridges face. In aclclition, thcrc was a gas prcssurc rclicf valvc, crtrricd in t]rc errcl
spanning the flow of the cooling oil. A roller bearing with hardened steel cover, provicling an escapc route to thc cxterior for gases trapl:ccl bctrvccn
outcr races was pressecl into each end cover ancl carried tlte eccentric, the rotor-bearing oil .scals ancl thc rotor-cncl covcr oil scal.
which was locatecl axially by a separate ball bearing. Spur tceth were f'Jris was not thc final vcrsion of thc KKI\/I-507. oil scaling systcnt. Ottc
machined on the external periphery of one of these races, to mesh with remaining problcm was oil leakage through the dual radial seals rvhen
tccth nrachinecl in the annulus of the rotor. An external-internal spur gear the engine was at stand.still. The outer oil seal helped keep oil frorn
oil pump was accommodatecl in one end cover. Because of fractures entering the combustion chan-rber. However, this was an expensive seal-
experiencecl with earlier designs, this annular gear was made an integral ing system that had to be simplified. In the last version, used on the
part of the rotor, a mallcable cast-iron part with more than the required KKN{-502 engine, the piston ring-type oil seal in the bearing carrier neck
w,As rcplaccd lly a carbou-typc unit sinrilar to those usccl in watcr pttntps,
fatigue strcrrgth.
'Ihis scal wäls sprirrg-loaclccl aucl fully inclcpcuclcttt of gas pressllrc. Its
Oil consrrnrptiorr had been the second largest problem with the KKM-
weakness was its dependence on a baclup seal in the form of a rubber
500. The main cause of oil burning was leakage of cooling oil from the
rotor. This oil could be sealecl off frorn the conrbustion chamber at two diaphragm. The rubber cliaphragm rvas vulnerable to clistortion clue to
places-either between the rotor and the eccentric, or betrveen the rotor local overheating, but this design effectively cured oil leakage even at
ancl the sideu,all. When the Spider rvent into production, the rotor had 8,000 r.p.m.
an oil seal l>etrveen the rotor and the end cover, carried in the rotor flank The cooling systern usecl a circunrferential-path cooJant florv pattern.
in a nrachinecl groove insicle the gas-sealing side seals. The ring face had Cool water enterecl the top of the housing ncar the rnajor axis, oll the
a 6 degree taper, as do some oil control rings used in modern piston ignition side, florvecl clon,n through channels in the clircction of rotor
engines. The metallic ring had a rubber O-ring inserted into a channel movement, then florved up again on the port side to an outlet imrnecli-
cln its irrsicle circunrference, rvhich blockecl the oil frorn passing under the ately opposite the intake pipe. Tu,o side-flow patterns clivertecl cool \t'atcr
rirrg and onto the end cover wall. The gas pressure that leaked bcyoncl to the end co\,'ers via a T-junction in thc passages, at the bottonr of the
 NSU DEVELOPS TI-IE \VANKEL rz9
(N) ,o prevent high-tension leakagc as a rcsult of plug fouling. The plug was
positioned before the nrinor axis and its points \\,cre contained in a
.fl small antechamber connectecl to the combustion chanrber by a snrall-
diameter "shoot hole."
By 1968, NSU felt that tlre engine coulcl be aclapted to norn-ral cool
4) ignition. Advancing the spark plug position fronr Z0 to ZB nrm. before
Performance data for the KKM- tlre minor axis gave a l0% power increasc throughout thc useful spced
,02 as installed in the Wankel
-o Spider. BMEP : mean effective range (2,500 to 6,000 r.p.m.). Iiuel econorny, holevcr, hacl been 20-25%,
7nw, pressure. \: volttrttetriceffi' better with thc spark plug positionecl 20 rnnr. bcfore thc rniuor axis. 'l'o
A 0,9 ciency. SFC _ specific fuel con- offset this drop in econorn)/, NSU tricd clual ignitiorr on thc KKI\{- r0?.,
0,6
-n sumption. HP : horsepower.
0,a0 with one plug ,6 mm . before, ancl the .sccorrd plug 20 nrm . ufter, thc
,gJtrut
@tür ntinor axis. TIris grlvc higlrcr l)o\vcr «lutllut nncl lolcr fucl consurrrllti«lrr
(1.) lp than hacl any sirrgle-plug svstem.
7 0,75 The KKI\{-502 conrbustion chanrbcr hacl an elongatccl shape ancl it
-S es
Ht was feared the engine woulcl bc pronc to abrronnal conrbustion phcnonl-
o,{5
1(m ena. The flarne front hacl a Iong way to travcl, arrcl tcnrpcrature risc ratc.s
oagne tared lrg.m
ancl prcssurc \\,äl\/c fornrations \\,crc cx1:cctccl to cilusc sl)orrt:rneous conl-

housing near the major axis. This portion of the coolant ran uprvards in
Irr this clurrt, friction l)rcssrlr(: iu thc KKIU- 502 eugine is sholn,
thc cncl covcrs nnd was routccl back to the outlct 1>ipc through a'l'-iunc- llrokcu tklrvrr tu tlrc \';rriolrs sourccs, frlr tlic full sllcctl ruugc. Ir{Ulrl'
tion in the top of the housing. : n)c;ln cffcctivc frictiorr l)rcssurc.
A two-stage, two-barrel Solex carburetor rvas clevelopcd espccially for
t
this Spiclcr enginc. It was built as a unit with the intakc nrrlnif«lld.'l'hc
prinrary throat was 18 nlm. in clianrcter and containecl the idle speed i
r
,.
t

fuel feccl system as well as the choke mechanism. The primary tlrroat
t2
hacl its throttle valve about an inch beJrirrcl the choke. Thc seconclary
n
throat was larger, 32 mrn. in dianreter, with a throttle valve very close to
tlre port. It wa.s positionecl to prevent exhaust gas from loclging in the
port arca when the secondary throat was closecl. Uncler part-loacl concli-
,/l- le--.rrcrg. lü,rldc q;r tttoltlo
tions only the primary throat was active; at a certain point the seconclary tr.lrüne oad eeoln orddod I
// lt ,*or
throat opened.
Another item that was designed especially for the Wankel engine was ,e
the ignition systenr. The spark plug itself receivecl most of the attention. t6
In a Wankel engine, the spark plug never gets the cooling effect of ,1
freshly aspirated, cold air-fuel mixture. This makes it necessary to use ,2
ru cool plug-a plug that dissipates heat faster than others. This need for
,0
a cool plug in the Wankel engine caused some ignition irregularities 0
(difficult starting and uneven running at part throttle and on the over- c ',(tolt vithou, tl t
run). The KKNd-502 usecl a single-surface discharge 12 mrn. spark plug a
with an attenclant costly Bosch high-tension condenser-type discharge 2
systenr. NSU found this sl,stem essential with a single spark plug and 0
peripheral intake port because a high voltage rise rate was required to tpod lcem)
 NSU DEVELOPS TI-IE \VANKEL I3I

3
The mostly aluminurn KKM- 509 / 506 weighc«l orrly lr4 pottncls;
including the transmission, its weight was only 210 pouncls. Displace-
*§
ui ment was equivalent to 2.0 liters, or IZZ cubic inches, while output was
§ ll0 horsepower at 6,000 r.p.nr. with an 8.7:1 compression ratio. TIre
apex seals in these new engines were allowed lcss maximum angularitl'
to thc working surface. Experinrental engines hacl used 28.2 dcgrees-
: 24.8 degrees was the limit in the KKM-i}Z and KKI\{-509/506. One
twin-rotor Wankel errgine was exhibitcd at the Intcrnational Auto Shorv
ar1 in Frankfurt in Septcmber, I965. It was rcclcsigrrecl cluring thc follou'ing
winter and became the KKI\'I-612. By April, 1966, units were running on
§::; the test bcd ancl irr a DKW l\,{urrga Iight military velticle.
!o,to
s 465 Thc KKN,I-612 was thc basis for thc Ro-80 autonrobile. This \\'As
ä oeo unusual irr that the engine was not installed in any existing vehicle-tlte
0,5,
car \4/AS designed around thc engine. Util izing the extra frccclont pro-
vided by the engine's small bulk and lol u,eight, the designers achievecl
exceptional vehicle architecture with great success. The car was the end

orgino ryod I r*rn)

WOf PEßFoRruNCE DA|A Cutaway view of thc KKIvI-622 (6lZ) from the spark plug sicle. Thc rotors are 180
r---.. SlillP,X Pl/.ß lOCAll(N 20rm A l,t)lgl§lE degrees out of phase.
Or<tl, 2errm? r .
.-.-... S.P L. 56 rnln 8,ftU5 NmmA . .
Tests with the KKIVI-502 engine showed the eftect of moving the
slrark plug forward fronr 20 mm. after the minor axis to ?.8 mm.
before the minor axis, ancl using dual plugs, one at 56 mm. before
tlre minor axis, the other Z0 mrn. aftcr it. One plug, 20 mm. after
thc utirror axis, gave Iowest power and highest specific fuel con-
sunrption. Dual plugs gave best pou'er throughout the speed range,
and lorvcst specific fuel consumption. A single plug at 28 mm. before
thc nrinor axis shows a parallel curve at slightly less favorable levels.

bustion under high-load conditions (such as during acceleration or climb-

ing a gradient). But, early in the program, NSU engineers were pleas-
antly surprisecl to find the knock resistance of the Wankcl cnginc quitc
high and its octane requirements very reasonable.
After the Wankel Spider went into procluction, research into sealing,
production nrethocls ancl tooling for twin-rotor errgincs wxs irrtcnsificcl.
In 19(t4, NSU built a hvin-rotor engirre with the rotors from the produc-
tion KKN4-502. This experimental engine was called the KKM-509 /106.
It incluclecl an integral water ancl oil heat exchanger instead of the ex-
ternal unit usecl with the single-rotor procluction engine. AII the accessory
clrives were placecl at the front, ancl a cutalay section on the flywhcel,
plus a semi-circular weight on the mainslraft nosc, proviclecl gcnerous
counterbalancing outside the engine casing.

I
 NSU DEVELOPS TTIE WANKEL I))
product of many years of concerted engineering effort by an outstauding
team. Ewald Praxl was the overall coordirrator of the total vehicle con-
cept, while the power unit was designecl and developed by Walter
Froecle and Georg |ungbluth. The key chassis engincers were Rudolf
Strobel and Herbert Brockhaus. Developnrent was the responsibility of
Hans Georg Wenderoth.
Chassis testing with trvin-rotor engines for the Ro-80 began early in
1965. The KKM-612 put out 136 horscpower at 5,500 r.p.rn. ancl pro-
ducecl a peak torque of 117 foot pouncl.s at 4,500 r.p.nr. It wus esscntially
\ t
two KKM-502 units placed cnd-to-encl on :l conlnlon ecccntric shaft. 'I'he
I
I
KKM-612 had the same chamber dimensions as the Spider: radius 100
I
I
mm., eccentricity 14 mm., and u,iclth 67 mm. Its R/e ratio was 7.14:1.
This engine was the first NSU Wankel engine with cast-iron apex seals,
replacing the KKM-502's carbon compouncl seals which hacl a tendency
to break up under detonatiorr. The carbon seals worked out fine in
Iaboratory tests, but on the road, they did not hold up because at ex-
tremely high rotational speeds the centrifugal,advance rnechanism rvould
advance timing to the point of detonation, and this form of abnormal
combustion immecliately cracked the brittle carbon seals. The inrproved
KKM-612 nrctallic apcx scal \l,rrs a Iirrk-typc scal. Urrlikc the KKI\l-502
seal, the split between the center and corner pieces was on the tip rather
than on the axial face of the seal strip. It was partially self-adjusting to
wear on both axial faces, and to conform to any convex or concave de-

ROTOR BEARING CENTER SEPARATION ROTOR GEAR titAlN BEARIi{O

1550 H HOtJtri
MAIN ECCENTRIC
BEARING SHAFT
HOURS HOURS

Reliability and durability tests

with the KKM-61 2 were going
on when the author visited NSU
in the summer of 1967. A num-
ber of engines were tested at
85% of full load at various
up to 5,000 r.p.m. The
speeds
figures indicated on the drarving
are averages for the n)any test
END COV
engines. Under such conditions, 33OO HOURS U@ HOURS IP@ HOURS
3,000 hours is roughly equivalent RrttrueR ptn SEALS
to 200,000 road miles. w HouRs 3r0o HouRs 3lq) HoURS
Cutarvay drawing of the KKIU-612. Dotted arrows represent fresh mixture; black
arrows represent exhaust gas.
134 DEVELOPMENT

fornration of the seal strip groove. The seal was installed oversize and In later versions of the KKM-
612, a new ty'pe of apex seal rvas
had a break-in period of 30 hours. Snrall slots in the leading face of the used. In the course of approxi-
apex seal ensured rapid pressure build-up beneath the seal to prevent matcly 20 hours of normal wear,
"spitback" (firing across the apex seal from the leading combustion this seal woulcl be worn down
to thc lcvcl of thc corner seals,
chamber into the following one, i.e., the chamber under comPression). and thereby close off the gas
It was founcl that the cast-iron apex seal was incompatible witlt the passagc in thc two corners.

chromed working surface used with the previous carbon compouncl

seals; therefore NSU developed an entirely new coating rnaterial for
t5e working surface called EINSIL. This material, a nickel plate con-
taining fine particles of silicon-carbide suspended in an electrolyte, is
integratecl with the surface during the plating process. The plating layer hydrou'lic pu^b e lon
is 200 microns thick when the working surface is rnachined. This coating t ign. lood on ollernotor.
I

is superior to chrome in terms of cost, quality, consistency, ancl relia- relererie curle lor I

perlormonce
biliti. With this courbination, NSU achieved 3,000 to 9,000 ltours seal meosuremenl s
I

tip iif.. Under deliberate test bed knocking, the old-type carbon seals
stäyecl together for only l0 seconds; under identical conditions the iron Pump f xrc. Pump O In the KKM-612, mean friction
olternotor lilli4g ) \1 pressures on all parts risc sharply
seals ]ived for l0 hours. rvhcn r.p.nr. excccd 3,500, ap-
Largely as a result of these changes, the KKM-6LZ had 2.5% less inter- proxinrately the sanrc point at
nal frictiop tfian the KKM- 5OZ. The new apex seals reduced friction by n,hich torque becornes truly sig-
nificant for acceleration.
17% under wide-open throttle conditions, and new oil seals cut friction
Ioss at the side ,urir..s by 59%. Parasitic losses in main bearings and oil
circulation were reclucecl by 46.7To. Still, the NSU engineers were sur-

ffi{as 'äg/fs
(Y
o.'cesJrrres
for the
On the KKM-612 prototype, the triang_qlar corner seals
,.rt äia not ,.r.h the *oiiing surface. They allowed a small passage,
apex
§ \
parallel to the rotc r sides, rt bottr apex .iotrrets. In the course of
at4 \
otor+rott beoring t cooling oil \
nor*ri ir.ri, oniv- irrauri deterioration in seal effectiveness could
-o\ opex seo/s
+ gos
\ - t-
\ occessorres
be expected. moin beorings I oil pump 5
t4 exchonge
t4 2s
60p0
RPM \a.
o

-i-- o og1 cleoronco

md c
-i-.
A2tO §o(, slde seo/s
+ bolt s
\

\
- § ,\
t
\ C c§ ,5 \\,
li ,30 o oil seol \
a
\, E \
o
0.
tU This graph compares the mean t ,
I effective friction pressure of the aP
o
KKN{-502 engine rvith the twin- \, rolort rotor
becr ing +
rotor KKN,I-61 2, at a steady a-
tr cocltng otl
speed of5,000 r.p.ffi., on wide T
5

open throttle. I{ighest gains were a

in the oil seal, main bearing, and
oil pump areas. 0

KKM 502 KKM 612

u964 ) u967 )
 L0 rslocä iomotrk mirlurc slronglh (0. gS . 5 clo rich I

095
0.95

t,0

t.0

0.95

w
This graph of the exhaust gas temperatures in the KKM-612 shows
a high general level, which is favorable for complete combustion and
This graph shows the differenccs that result in mean effective -pres'
Iow emissions. sure witti variations irr air/fuel ratios and r.p.m. You can read the
graph as a topographical nrap showing altitude differences.

Q(o/d KKM 502 KKM 513

,001 Unburned exhwsl

gos ond residue

Distribution of housing skin tem-

perature in the KKIU-612, meas- cor'- Co,,,Orrison of heat balance be-
ured across the leading spark plug tween thc KKfvI-502 and the
at 6,000 r.p.rn. on full throttle, Eräousl gos KKI\ t-(r I ? cugirrcs :rt {,000
shows that tcrnpcratures in the r',1),lrr., slron, higltcr cllcctivc llcr'
irnnrccliate conrcr of the shooting Iorrrrirrrcc in tlrc llc',r/cr urrit itttd
hole and thc e pitrochoidal sur- f:tr lcss rrrrllrrrrrcd cxltitrtst gn§.
face may beconre so high that Al - otlot Lcss hclt is dissipltcd by the
the stresses may cause cracks in wotcr oil, arrd ulore by the wrlter. NSU
the casing. The copper alloy I says that the 4,000 r.p.rn. curves
bushing around the plug takes oll are representative for the full
away the tenrperature peak from speed range.
the alurninunr.
lN -JN effeclive
,70 pcrlor monce
IN
1g ttli
a 2(n
nO
to 4000 RPM V!,',or
t9 I
t.0 =115
oF These graphs show lrow horscpowcr and specific fuel coltsumption
tn are affected by a step-by-step leaning out of the air-fuel mixture.
Curves are shorvn for a constant 2,000 r.p.m. under variable load.
tX)
,r0
,@
{.
SFC lbs/HP hr
f

ffill ffi
t
t
lr

These sketches show the circumferential and axial distribution of

rotor housing skin tenrperature in the KKM-612 engine, at 6,000 t:
r.p.rn., on wide opcn throttle. The cross-section on the left shows t
the intake and exhaust ports, irr addition to the location of the two I

spark plugs. !'

I
 hal .t5 I lbs/HPhr NSU DII\/ELOPS TI IE \\/ANKEI, 139
ovlpul

prised at the amount of wcAr founcl in the cngincs of cars that had bccrr
t0
t,
running on a stop-start tcst cycle. T'o cure this problent, they tu'icc
changed the kind of cast-iron used for thc rotor tip scals, each tinre
ohlood gaining a significant improvemcnt in cold-start wcrlr, ancl they are at
Cloh lood present satisfiecl that the trouble lras been cured.
5l ch lood There were also ignition difficulties at first, due to sen'icing troublcs
ancl poor electrical dcsign, but thcse \\,ere curecl by charrrging fronr trvo
lucl consumplion spark plugs Per rotor to one, ancl from conventiorral cotttact-breakcrs to
lbslhr those of the transistor-assisted type.The dual-plug ignitiolr systcnr of the
KKN,I-612 had tn,o spark plugs, one locatecl 56 mm. before, and the
'l'he u,i«lc ol)cu throttlc curvcs shon,a drop in porver with rich mix- otlrer 24 mm. after, the minor axis.Thel, wcre fircd sinruJtaneously. J'he
turcs. l)uzzlirrg? N{isliring occurrccl soouer rvheu running rich. Curves use of clual igrrition gave a nunrber of aclvatrtagcs in vicrv of tlte long
rrrc shou'rt for a cortstaut 5,000 r.p.nr. WOT: wide open throttle.
ancl flat combustion chambcr:
I . Steeper pressure rise.
6MRPH 2. Iiaster combustion.
This graph conrpares tcmpera - tc0o''
3. Higher thermal efficierlcy.
-20mn f
+ltmm
ll 4. Lower exhaust gas tcmperature. r!
-senmll 5. Better cold starting ability.
--F-
I plug potiliotr cllor Thc change to singlc-1>lug igrrition \\':.ls rruclc rtrainlv hecausc ts'itt
lnnot ortt spark plugs necclecl separirtc igrrition s\'.stclns.'.['§,o 1>lugs cotr]cl l>c u'irccl li

n in series onlf if one \\,'as conrplctcll, insulatccl frottt the cnginc; if they
I

a-,
-- tt wcrc rvired in parallcl, thc one u,ith thc srnallest gap rvoulcl fire first ancl
short out the second. Also, trvo c«lils wcrc nccclccl, and thc spccial clis-
igailion odtoaco
l.8fc) tributor hacl tu,o sets of breaker points ancl a cloublc-\\,ipcr r«rtor u,ith
resistance.
t0
O@RPH
tn,o opposed electrode arrns. Thc single-ignition systcnt substru'rtiallt,
rcduced procluction costs, a nd thc transistors providccl i.l strong sltirrk
despite plug fouling, in adclition to extencling the slteccl range of the
'l'his gral>h shows thc fucl consumption of the KKM-612 (complete
Cornparisort of spark plug and
rvith acccssorics) in a lorr«J vs. specd diagranr. NSU points out that
seal configuration betrveen KKM-
thc engiuc rcquires a fairly rich nrixture to assure driveability, even
502 and KKM-612. The tapered
u'ith äln accelcrator punrp in the carburctor.
seat is required to assure identi-
cal ancl reasonably srnall breath- KKM 6'3
EHEP
ttÖe/in'l ing volumes. It n,as rnovecl from
bottorn to top, because the ( t967 )
carlier t)'pe brought a clanger of

/,
a6 /
I\ \-) 0.5 tbs lHPh

;7
cleforrnation of the housing sur-
facc as a result of overtightening.
The new plug has a longer
055
threacl, w'hich aids heat transfer. @othing br:erithirg
iolume iolume
nß -- - On the KKI\{-502 t}re plug
scrcn,cd into
was
thc alurninum hous-
0.7 ing; on the KKIvI-612 it is
- lodgecl in a copper alloy bush-
t0 j1

ing. The ncw plug is of the sur- topered

ilÖ§lHPht face gap type, the old one con- bottom lop seot
ventional. seol
RPT'I
l

i
F
ü.
 NSU DEVELOPS'|IIT{ WANKIIL I4I
6000 R PH
Performance curves for the KKM- engine by several hundrcd revolutions. Tlte prcsent Ro-80 KKN{-612 plugs
612 prototype with five stages havc tapcrccl seats, u,lrir-h contrcll plug Po.sitiou lllorc tlcctrnttcly'tltittt tlo
of engine modifications during
its development. The upper compressible gaskets. The tapered seat is locatecl outside of the threacled
curves show effective pressure in section because overtightening could producc clistortion of the trochoidal
psi (pounds per square inch). The
lou,er curves show specific fuel
consumption in pounds per horse-
power-hour. (l) = Spark plug po- cooting thickness 250 microns
s eol ' Dluo inloke
sitioned 28 mm. before the mi-
=
thickn.l ireo-, monil.
nor axis, apex seal thickness 2 §
-29 2 mmi bo,,o^ long
ffiffi., plug seat at the bottom, \(,
o
- 561o21 a aa
with long intake manifold. (2) = E
u a
Dual spark plugs, 56 mm. be- \o
a fore and 24 mm. after the minor t
a
shorl axis; seals, plug seat, and mani- I
§t
ller
rrs
fold unchanged. (3) = Plugs as C

q- (2), but 5 mm. thick apex seal; o

o
t plug seat and manifold un-
I

a
ta changed. (1) : PIug seat changed
rcod ruaalng öilonco ln
(,
to top, otherwise as (3) . ( 5) =
t.t- Short intake manifold, otherwise rtf OcA riiitag llmo la hn.
t4 as (4).
2W tml 1@O W0 ,@@PPH The Einisil coating on the u,orking surface of the KKM-612 yielded
average rvearing qualities in the range of 50 microns per 80,000
nriles, in actual rorrcl use. 'I'he test bcd field in the chart includes
many diffcrent corrclitions of opcratiorr, inclucling wide open throttle
at 6,000 r.p.m. These results wcre obtained with the latest forrn of
apex seal (slotted on the leading face).
This sketch shows how gas pressure is utilized to improve apex seal
tightness in the KKM-612. IIad the corner seal edge (l) been solid,
it would hal,e blocke«l the high-pressure combustion gases, because
of the tilt forccd upon it by the sudden buildup of pressure in the
lcading chanrbcr. This phenornenon was cured by machining a
number of srnall slots along the inner leading edge of the seal. This
allowed combustion gases free entry into the volume behind the
seal, whcre they would force the seal strip radially outward, as well
as against the trailing wall in the seal slot.

lttl bd ruaning limt la fut

Wear tests on the bcnch and with KKN,I-612 engines installed in

cars, shorved such lorv wear on the new apex seals against the
Einisil surface of the epitrochoidal surface, that the functioning of
the apex seal is never in danger. Even 3 mm. wear cannot influence
performance characteristics.
 NSU DE\/EL,OPS 'I'I IIi \\/ANKEL r43
DC morking

Ncssuto diqmm
the combustiorr chambers, rcccsscd in thc rotor faccs, arc svntlnctricirl.
combuslion cttombcr. N,{allcable cast-iron is onc of the bcst nrateriels availablc for colrtrolling
prrssuru dbgrcm expansion uncl clistortiorr at high tcurpcraturc; thcrcforc thc rotrlr rulls
&tlom ol sr,l shl rvith aclcquate clearance for thc clifferencc in cxpansion ratcs bettvcen
thc rotor ancl housing 1>arts.
'flrc enginc housins is \\'iltcr-coolccl n'itlr r:ircurnfcrcntial flou,. Thc
rvorking clranrllcr Irousings urc nurclc «lf ;rlunrirrunt rtllol,frlr goocl lrcnt
Nstssun
dillerencc clissipatiort, §,hilc tlrc ccr)tcr scparlting \\,all lrnd tlrc crrcl co1,g1s ilrc lllrtclc
of cast-iron. Ccrtnin slnall urcr.s on thc irrrrcr s^rrrfltccs of tlrc cttd co\Icl.s
'f'he apex seal is depenclcnt upon the presence of gas pressure be-
havc a thin coilting of nroll'bclcnunr spravccl on t«l givc adclccl \ /crlr
hind it to cffcct a propcr sealing facc. These curves are rln example resistarlce. AII fir,c housing parts :rre boltccl togcthcr \\,ith long tcrrsiorr
of rcta rclccl prcssure buildup, such as that which existed with the bolts parallcl to thc cnginc uris. The \\,iltcr punrp clcliver.s 29 gallons pcr
non-slottcd scal. minutc at 6,000 r.p.nl., ancl is positionecl at thc uppcr front encl of thc
engine. Coolant flows circ:unrfcrcntiallv in parallcl strcanrs through all
thc hou.sirtg parts^, an«l tlrrouglr thc oil arrcl \\'rltcr lrcrtt cxcltangcr cnl-
lime pressure pickup locolton, bccldecl in thc srunp. Coolant rcturns througlr a passagc at tltc top of thc
errgine-coollrrt capnc'itv is 2,7 gt llrlris. TIrc cooling .st'stcrn \\,a.s nroclificd

jlllslae monilold The marine vcrsiou of thc I\o-80 crrgirrc fornrs :l very colnpact polcr unit for an
- inborrrd/outbolrd instrrllrrtiou.
BDC (cold lobe )
--i71loke
DC
-f
The old-type apex seal suffered
no sDit - bock
f rom the spitback phenomena. (ign.ciav. 35o B\DC )
'l'lre rcsults are shorvrt in the
Iower graph here. The uPPer
graph shows the results obtained
with the new sealing formation,
which eliminated spitback. intoke monilold

BDC (cold lobe )

-inloke
DC

sDit . bock
( ign odv 40o BIDC )

track if thc seat wcrc positionecl too closc to thc working .sttrface. The
spark plugs are I8 mm. units because of their improved resistance to
fouling. 'Ihey are not threaded clirectly into the ltousing but into a coP-
pcr alloy bushing, rvhich recluces the ternperature peak and the risk of
local cracks in the housing.
In the procluction Ro-80, about rvhich nrore will be said later, thc
cccentric slraft is supportccl at each cncl b1,ir plain tri-metal l>earing. In
the interest of sirlple clesign and assernbll,, there is no center bcaring
betrvcen the rotors. The tu,o malleable cast-iron rotors are identical, and
 5
o
=

Tlre Ro-I35 is ratcd at 135 horscpo\+'er at 6,000 r.p.m. It has a heavy-cluty cooling
systenr and a w'ater1>roof elcctrical systcm. The clrive trairr is made by ZF and the
propeller is geared down in a I .51 : I ratio. T'he drive can be turned through 26
dcgrees each wäy, which offers exccllent rnaneuverability.
In 1968 NSU installed an Ro-80 engine in this small motorboat, for test and
dcmonstration purposes.
in production, ancl now incorporates a rcdesigned thermostat, an im-
proved heater vah/e, and a clifferent fan.
This racing boat has distinguished itself in many events. It is powered by a twin-
rotor NSU Ro-135 developed frorn the KKIvI-612, which drives the Ro-80 passenger TIre oil purnp in the KKNI-617 cngine delivers 5.2 gallons pcr minute
car. In seven starts in the two-liter class of international boat races, the Swedish at 6,000 r.p.nr. The rotor bearings, which havc no ccnter groove to in-
pilot Zctterstriim took three first places, one second and three thirds. In the 1969 crease the oil holding capacity of the bcaring shells, are fed through the
Rouen (I'rance) 21-hour race, it finished third overall.
partly hollorv mainshaft. Oil from tlre cccentrics i.s also squirted directly
into the hollou/ rotors to provicle cooling. An adclitional small oil punrp
in line rvith the engine's nrain lubrication pump supplies pressllre
through an external pipe to the-torque convcrter, rvhich also uses engine
oil. The return flow combines rvith the retunr engine oil in a passage at
the bottorn of the engine, thcn is clrainecl to the sun'rp. 'l'otal oil capac-
ity is 1.7 gallons. Lubrication of the lto-80 enginc is, thcrcfore, just a
matter of putting in a little oil now ancl then to replncc the amount of
lubricant metcrccl into tlrc carburctor to ]ullricatc tlrc rotor and tlrc
housing. Bccan.sc tlrc sutl)p oil irr a \Vankcl cngirrc is rrot sullicctccl to
blou,by or contanrination it \\,u.s phnnccl nt the outsct to ubarrclorr
periodic oil changcs, but thc cälr hacl to pass high-nlilcagc tests bcfore
NSU could be sure that thcre \\,cre no ill effects. The decision to actually
abandon oil clranges rvas nracle after nrorc than 3 nrillion rniles of test
clrivirrg over a four-year periocl, prececlcd ancl acconrpaniecl by bcnch
testing to evaluate the roacl test results. NSU has been tr1,i1,t to ntake
the production \\/ankel crrginc a uraintcrutncc-frcc powcrplaut.
In contrast to thc 1>iston cngine, in which tlrc piston ring perfornrs
closely rclatccl furrctir)ns as a grls scal ancl oil control clentcnt arrcl n'hcre
146 DEVELOPMENT NSU DE\/ELOPS TIIIi \VANKE[, 147

thc 1>iston rnovcnreut irrvolvcs blowing cotnbustiorr gascs past thc rittgs, pcoplc b«lught rnotclrcl,'clcs, arrcl tlrc I )rc.sdrrcr lhuk, rvhicll controllccl
the Warrkel engine provides a sharp distirrction between the furrctions NSLI, rvas not interested in world recorcls and racing successes, but in the
ancl timing of lubrication and gas flow. NSU has designecl u,ltat they company's profit-and-loss statements. While Dr. liroede and Felix Wan-
call fl secondary compression charnber which retards passage of com- kel had made great progress on rotary engine development, NStI's in-
bustion gases by allo'iating marginal overpressure through a relief chan- come had dwindled as a result of the motorcycle boom's coming to an
nel controllecl by a ball-opcratecl pressure valve. This provides a neutral end. It was doubtful whether NSU had enough capital strength to bring
zone which impeclcs the mixing of end gas (volatilc combustion resiclue) both the new engine and the Prinz automobile to a state of readiness
with the oil. for procluctiorr. To concluct large-scale experirnentation with an uncon-
Two dual-stage sicledraft carburetors arc tuned to the breathing llat- ventional powerplant concurrently w,ith the design, testing ancl develop-
tcnr of the KKM-612. The two intake ports combine in a single circular ment of a conrpletely new car required investmeuts beyond the scale on
port orr the trochoidal surface of the central wall between tJre trvo work- which von Ileydekanrpf ancl his directors coulcl operate. I{owever, the
ing chanrbers to give earlier intake port closing and thereby raise lorv- NSU Prinz did go into production, and Wankel engine development
spcecl torque. There is a complex of muffiers to silence the exhaust, was never interrupted. How did NSU manage both? lielix Wankel sirn-
rvhich is not easy on a Wankel engine that, just like a two-stroke has ply went out in search of new sources of financing.
ports rather than poppct valves ancl is sensitive to back pressure. The Wankel associated himself with a post-war millionaire, Ernst Flutzen-
two chambers have indcpenclent systems to prer,'ent interference between laub. Trained as an architect, turnccl inventor, I-Iutzcnlaub lrad createcl
tlrc trains of powcr impulses. Separate pipes from each housing lead to a a small conglomeratc by uniting srnall companies in the plastics and
pair of sntall nruffiers that lrandle the high-frequency compotrents. These rubber field.s, rvhich wcre cxploiting his invcntious. I-Ie solcl licenscs
are followecl by, series of two partitiortecl units of golfbag proportions u,orth $500,000 outsicle Gcrmarrl,, llncl s:l\\, a lcgal opporturrity to rcducc
that eliminate the lows. his taxable income through such operrltiorts. This did not fall into the
category of tax evasion because, at that tinre, thc Gcrman Govcrnmcnt
wanted to encourirge exports, ancl rcvcrruc fr«»n invoiccs carning foreign
currency were only taxed at the special lol rate of 25%. The experi-
Fronr the dry thc first experimcntal Wankel engine ran in tlte NSU enced Flutzenlaub stakccl $250,000 of his own money on the Wankel
Ial>orat«lry until thc clay when the first KKM-502 procluction enginc was cngine, then he and Wankel forrtrccl a conlpany kuorvn as Wankel
asscnrblccl, ovcr scven ycars hacl passed. You have read all account of G.m.b.H., rvhich was to act as a sort of cash collecting office for future
what sorne of the technological problems were and how they were Iicensees. Wankel ancl Hutzenlaub each helcl a 50% intcrest in the conl-
solved. Insuperable as they must have seemed to many, they were over- pan/, which guaranteeclj}l, of its income to each. As the next step,
conte. It is a snrall miracle that so much was accomplished in those Wankcl and I-Iutzenlaub tricd to 1>ersuacle clozens of inclustrial finns to
seven years, ancl it would seem a fantastic achievement if it had all come buy a license for the "engine of the future." Tlleir activity impressecl the
about in an environment free from financial worries and stress. The fact management of NSU. Von I{el,dekampf of NSU felt encouraged; he saw
is that throughout this initial research and developnrent periocl, NSU hope for frcsh financing through the proposed liccnsing agreemeuts.
was in a state of perpetual financial turmoil. The setle of liccnses was held up by nlan), factors. Onc was Iack of faith
T5c clestiny of the company seerned to depend solely on the reactions in the engine. That was natural-investors are traclitionally skeptical
of the financial world ancl public opinion as to the worth of the Wankel about inventions. But NSU's involvemcnt with the \Varrkel engine
engine. NSU stock fluctuated in spectacular Ieaps and bounds on the tended to be nrore of a hindrance than a help. NSU was knorvn as a
stock market, depcncling on the latest news from the laboratory. small firm with goocl proclucts but a dwindling nrarket. Its managcnrent
When I)r. Iiroede was given the task of collaborating with Felix was untested in that kincl of situation. At that time, no leader of indus-
Wankcl on the conversion of tris rotary compressor to an internal com- try could imagine that NSU, rvith its 7,000 employees and an annual
}ustiop cngine, NSU still callecl itself the "Nunrber One Motorcycle gross turnover of less than $ 50,000,000, \l,as capable of producing a
Maker in the World," and new racing successes and international class technical revolution. It was felt that such a revolution could only corle
record attempts were maintaining its reputation at top level. But fewer about with the financial aid of, and the prestigious association rvith, a
 DEVELOPMENT NSU DEVELOPS TI-Itr WANKEL r49
148
other sizes and versions of the Wankel engine \\,ere to be rnanufactured
worlcl famous conrpany. One of the key men in the Dresclner Bank,
-outspokenly by other liccnse takers.
Hernrann llicfiter, aclvised volr Hey'clekampf to clo some-
What profit potential was there? A friend and advisor to Irelix Wankel
t5ing to improve NSU;s reputation, something to make the Wankel figured that if in a ferv )'ears only l0% of all engines rvorked on the
engine more palatable to the cautious inclustrialists approached by Wan- Wankel principle, it rvould mean a production of between 7 and 8 mil-
kel and Flutzenlaub. lion Wankel errgincs per year in the Western rvorld. If $18.50 per
T5e NSU Prinz was von Heydekampf's answer, but the question of engine could be countecl on as the combined Iicense and royalty fee,
whether that woulcl be enough to entice Wankel engine backing from NSU ancl Wankel rvoulcl collect about $l50 million annually. Of this
outside remained unanswered. amount, NSU woulcl retain $90 million, and after taxes, woulcl still have
Wankel tried to ncgotiate a deal rvith Krupp in 1957 and 1958, but
came from Curtiss-
a surplus of $50 million-a figure almost as high as its entire capital
Krupp did not want any part of it. The first bite
stock.
Wright wSose presicleni, noy T. Hurley, thought they coulcl take the
it. Up to this point, Dot much was known about the Wankel engine and
.*p.ii*ental engine, refine it, and mass-produce Curtiss-\\rright
the agrecment betlveen NSU N,fotorcnrverke AG ancl Irelix Wankel, but
signecl an agreemer)t on October 21,1958 that they would openly share rvhen the contract u,ith Curtiss-Wright was announcccl, an avalanche of
aü additional patents and design clranges with NSU. The agrcement publicity follorved. The press reaction was overwhelmingly favorable,
also gave Curtiss-Wright all rights for Wankel engine manufacture, in and both speculators ancl ald established brokers began to bu1, up NSU
all sizes and for all ,ppti.rtions, in the United States. NSU is reportecl stock. Even conservative bankers ancl critical investors fell for the "en-
to Save received $2.1 million from Curtiss-Wright for the license, rvhich gine of tomorrow" promise.
also includecl a 5% comrnission (or the selling price) to NSU on all
The Dresdner Bank held 5l% of the NSU stock and controlled the
Wankel errgines that Curtiss-Wright might build and sell. company, the other 49'/o was helcl by 5,000 individual shareholders.
As soon är Flutzerrlaub heard of the agreement betrveen NSU and Nominal value of the capital stock was equivalent to only $4,500,000-
Curtiss-Wrigfit, |e went to Neckarsulm (West Germany) with _strong NSU stock prices had renrained sadly lou, eveu in the period when NSU
demands. He warrted 40% of all NSU's income fronr Curtiss-Wright, I
motorcycles dominated the market. A new lorv was reached in 1957,
and a contract stipulating the same percen tage on revenue from all rvhen NSU stock was quoted at 124.r% of par. (lluropean firrns often
future license takeri. The NSU directors put up a tough fight. They said I issue stock in lots rvith different par values. The stock exchange listings
Wankel hacl beep on the payroll since 1961, with a rnonthly salary of refer to percent of par value, whether the share has a face value of $25
DM 3,000, ancl fiis acconrplishments at NSU were the propertl' of the t)
or $1,000.) Suddenly, anonymous aclvertisenrents bcgan to appear in
NSU company. Flutzenlaub explained that Wankel had brought thenr ä
I
German newspapers: "§SU shares bouglü at toP prices." Ob.servers
5is life's rvork, unique experiencc arrd ability, the essence of lifelong t guessed that they were inserted by sonreone rvho was anxious to unload
experinrentatiou-in fact, assets that could not be overestimated at 407". a package of NSU stock and wantecl to first clrive up the price. Von
r
NSU gave in, mainly to avoicl the unfavorable publicity of a court Process
Heydekarnpf got u,orriecl ancl tried to firrcl out who was behincl the ad-
E

II.
against the inventor, ancl Wankel ancl Hutzenlaub received their 40%
t,

vertisements. They were tracecl as far as a larvyer irr Düsscldorf rvho

l.

ui t6e initial payment from Curtiss-Wright. Roy T. Flurley paid little

r:
i

would not reveal the identity of his client. Ansrvcrs to tlre aclvertiscrnerrts
attention to tliis last deal. To him, the money he paicl was a small stake r
f were refused or unanswerecl. NSU stock corrtirrued its upwarcl trend ancl,
on the fantastic fortune to be reaped through Wankel engine applica- l,:
I

at the end of L959, it stood at an even 1,000.

tions for aircraft, boats and automobiles, as well as for stationarv pur-
tr
I
h
r
This rise was completely against tlre expectations and calculations of
DOSeS. E
the Dresdner Bank. They were unrvilling to face the possibility of spend-
Because of its unique position relative to Wankel G.m.b.H., NSU
I.
F
!,
t
t:
ing year after unproductive year with this revolutionary new powerplant
was not regarclecl as ,- regul*r licensee but was allowed to manufacture It
on speculation that the \\/ankel engine nright one cla1, replacc the exist-
and sell Wankel engines against a fee specified in a seParate agreement.
ing piston engine. While NSU was gcttir)g grcat publicity thanks to the
NSU's planning *ri cautious. At first, NSU was to manufacturc station- I Wankel engine, and morale among the workers as rvell as managcment
ary Wapkel engines, to use as pumps for instance, then they would
was enormously strengthened, the Dresdner Bank officials decicled it was
{

pädu.e the 40O cc. Wankel engine destined for the NSU Prinz car. All
JI

l
{
i
L

Ii
ti
t
:;.
 NSU DE\/ELOPS 'I'I IE \\/ANKIiL I5I
I5O DEVELOPMENT
that the bank might hrrve solcl blocks of somc size to Curtiss-\Vright or
time to epd their long-established mesalliance with NSU. Because of the
to Mr. Flick, who could have changccl his nrincl about the Warrkel clr-
Sigh stock market quotations they saw a chance to sell their NSU stock
gine since 1958. That started a riot, ancl llichtcr calmecl clorvn the
at a Sigh profit and free themselves of the small, but embarrassingly
audience only by showing them the experinrental Wankcl engirre arrd
busy, inauitrial operation. The bank first offered its NSU holdings to
explaining that more work was needecl bcfore it could be exploitcd.
Irriedrich Iilick, chairrnan of Darimler-Benz AG, who refusecl. I-Ie was
By mid-year, 1960, the stock hit the 3,000 mark. Thc recorcl \\'as
at t5e time busy with the Auto Union acquisition and was merking reached on fune 70, 1960, at the I\{unich stock exchange, r,i,ith quota-
arrangements to sever the DKW motorcycle clivision to concentrate
on
tions of 7,200! The end of the NSU stock rnarket romance came u,hen
passenger cars ancl light commercial vehicles. Then, feelers were put to
tlre Dresdner Bank sold part of its rcnraining 26%. T'he result was a
Roy'f. Hurley of Curtiss-Wright.
sheer drop in the price of NSU stock, which hacl the cffect of clestrof ing
Hurley was interestecl in taking over the entire 5l% at the current
the high hopes of the speculators. The1, in turn began to sell their own
value, but tSis was not to be. Holcling 5l% would mean that Curtiss-
NSU holclings, bringing prices further clorvn. By the end of 1960, the
Wright rvould, in effect, acquire control of NSU Motorenwerke AG'
stock held at about 1,300 and cluring 1961, it nroved only n,ithin the
GerÄany 5as no restrictions regarding foreign ownership of its industry,
1,300-1,600 range, continually up ancl down.
but tSere were more strbtle complications. With control of NSU, Curtiss-
The principal reason for the inhercntly speculative naturc of all cleal-
Wright also would be in a position to clictate to whorn future licenses
ings in NSU stock was the uncertainty of thc Wankel engine's poterrtial.
*ouLl be grarrtecl. It was thä shrewd Hutzenlaub rvho put a stop to the
The public had few facts, if any, upon rvlrich to base its opinion and
cleal. T5e Dresdner Bank gave him the cold shoulder, but went along
judgment. This was partly rcnrcclicd whcn tlrc technical prcss was in-
witS 5is protest, pretenclirrg that th ,y were acting in order to Prevent an
vitecl to a senrinar on the invcntiorr, clcsign an«l clcvclopntcnt of tlre
AmericAp capital invasion. There were no further attempts oll thc
Wankcl cnginc on ]anurlr)' 19, I960.
part of the bänk to fincl a new parent conlpany for NSU' 'l.hc nrecting \\,as lrclcl irr thc I)cutsclrcs I\luscurrr irr I\Iuniclr rtttcl or-
Insteacl, the Dresclner Bank began feeding its NSU stock to the cx-
ganized by the V. D. I. (Vcrein I)cutscher Iugcnicure) to cliscuss the
changc during the week of |anuary 8 to lr, 1960. Thc average Germatr
invention, dcsign and developnrcnt of thc NStl Wankcl Ilotary Conr-
industrials thän lay around 370-the NSU stock sold at 1,0251 I)irector
bustion Enginc. T'he urccting wils attcnclccl lry I ,3 50 scicrttists, cn-
Iiritz Andre of thä l)rcsclrrer Bank defendecl the action by sayirrg that
giueers, teachcrs ancl stuclcrrts. Profcssor I)r. Ing. Iirrrst Schnriclt, of
the Sank wantccl to act agairrst the nracl price pattern of NSU stock.
the Technical Institute of I\4unich, gavc a lccture on tlrc Iristoric;al dc-
This corrsicleration was fclt to be merely an alilli against possible accusa-
velopmeut of internal conrbustion crrgincs iu Gcrurany, frortr N. S. Otto
tions of stock manipulation durirrg 1959,The bank admitted that it sold
and Rudolf Dicscl tdFclix Warrkcl. f)r. lng. Iilugcrr Wilhclut IIubcr,
about Salf of its NSU holclings in |arruary, I960, rvhich brought its irrter-
heacl of the I{uber Pou,er Plant Rescarch Institute, discussccl his own
est in NSU clown to 76%. Jihe bank received some $12,300,000 in tlte
work on the Wankel enginc artcl sholecl l souncl filnr of the cngine
trarrsactiop, for stock with a par valuc of A nlcrc $1,150,000. Tltc followirrg
runuing on a clynarnornctcr. Profcssrlr I)r. Irrg. li. Ilaicr, of tlrc'l'cclrtri-
nront5 NSU Motorenwerke AG was recapitalized. New stock was issued
cal Institute of Stuttgart, dealt rvith the geometry of the epitrochoid.
to raise its capital from $4,500,000 to $6,700,000. |trst how the new stock
He demonstratecl that the working chanrbcr ancl rotor can havc up to at
was to bc clislributecl was a rnatter of cliscussion at thc general asscrnbly
Ieast seven cliffcrcnt shapc.s in working charnl>crs «lf varyirrg curv:.lttrrc.
or1 luly ZZ, 1960-a meeting at which the Wankel engine played an im-
Felix Wankel briefcd the auclicnce on thc historical cleveloprnerrt of
portant role.
rotary engines fronr tlrc l>cginrring of tlre ccntur1,, rutcl prescrrtccl rl
I-lernrarrn Richtcr ancl Erich Vierhub of the f)resdner Bank faced
lengthy discussion on scaliug for gas tightrtcss at thc slidirtg surfacc.s.
350 stockholclers. As a rule, only about a clozen stockholders showed uP
Then I)r. Walter [irocclc outlincd hi.s KKI\t clcsigrr rtrtcl clcvcl«lpn)ent
for t5c nreetings, and business was usually concluded in a quarter of an
work at NSU. [Ie shorved that u'ith the three-lollc rotor clesign the conr-
hour. But not ttrir time. Iironr the way their questions were answered,
pression ratio of the e ngine \\,as limitcd to I 5.6: I . A four-lobe rotor
the stockholclers reached the conclusion that they were not going to be
u,ould allorv a I00:l colnpression ratio and a fir,e-lobe rotor woulcl allow
invitecl to share in the new stock issue. They noted that holders of 27%
:

a compression ratio up to infinity. Dr. ITroecle u,ent on to sa), that, in

of NSU stock were not represented at the meeting, and they guessed
,q ':

ii
,{
I r1'
t* I5Z DEVELOPMENT NSU DEVELOPS TTIE WANKEL I53
i"*{

NSU stock prices towards the end of 1960. All money institutions in
g-'. i.

5is opinion, the engine could be used for a multittrde of power applica-
Ai .'

tii?
1
L\'.'r

r
.

tions. I-Ie preclicted that the engine woulcl first be used as a small con- Germany are taxed for the market value of their stocks and bond hold-
S,?ir
ings. This enabled the bank to list a very low value for its NSU stock
starrt speed stationary unit of under 50 horsepower, for such duties as
j.
e.
ii. i
in its annual tax declaration. A Iawyer in Düsseldorf, named Engler,
,[J

pumping water or generating electricity. Froede gave credit to Goetze

.j j.

l*
)t Werte For its notable contributions to thc developmcnt of the sealing who representecl a group of snrall NSU stockholclcrs, accused the
Dresdner Bank of willfully forcing NSU stock prices down, after the
r. system, and pointed out that five of the principal oil conrpanies in
.''-.

bank itself had sold its NSU holdings. The most important of these
i=a
:-f I
ö.r*rny hacl run test engincs in their laboratories to cletermine fuel
f*I octane requiren'lents and select the most suitable lubricants.
minority stockholders was a medical doctor frorn Stuttgart, Dr. Diehich
ii Albers. His holdings had e par value of $20,500, which gave him iust
I.1 i
Soon after, NSU announced that, after spending $A rnillion on its
under yr% control of NSU. Dr. Albers was instrumental in the forma-
ü,:" :
ä'
[{ development, the Wankel engine was still not ready for Production.
l :,
t'iJ
.
.

This did not affcct the stock price, but stockholclers and other interested
tion of an NSU stockholder association, which published advertise-
ments with the following text in various daily papers: "NSU stock-
parties were anxious for some word frorn NSU as to when procluction
t*-, L .

holders, keep your shares. The stock price fall is directed by high
ti.: :.

i6 likely to begin. In desperation, one small stockholder wrote a letter

!-j

ii'"
f.t

'ir
'7.:-'.
finance, so that they can bry ),our stock as cheap as possible. Be
to NSU. In tfie guise of a potential custonrer, he asked innocently rvhen
patient. You own one of the most valuable German stocks.-Schutz-
'i. i
-,i . t
A:.'
nr. 5e could Sope t; buy a Wankel-type mälrine engine. TIte answer fronr
r..
gemeinschaft der NSU Aktionäre e.V."
NSU was aatea August 22, 1960, and was signed by one Dr. Hirsch.
t. ,

i
:. , I
Through this association, Dr. Albers formed a united front of 600
i.'l r' It said tSat NSU piannecl to start procluction by micl-1961. The en- minority stockholders. IIe valiarrtly pursued all clues to what was going
gine was to be a 400 cc. unit, clevcloping 45 to 5, horsepower at 5,000
f i"
ti:

on with the Wankel engine, During the first part of 1961, he discoverecl
'"1t'
it' i.p.*. Not prinrarily a marine engine, a marinc version would no doubt that Daimler-Berrz had a batch of NSU Warrkel engines on test.
follorv. Woirlcl tfte customer plcase write to NSU agairr in the miclcllc
; .)
l,'

Daimler-Benz was not yet reacly to bry a license though. liirst, in typical
of t5c follorving ycar? TIre stockholcler ga\/c the lettcr maxinltlm pull-
I
'i: :

fashion, they wanted to evaluate the Wankel engine as a practical auto-

\'

licity and NSU stock went uP again, almost 200 points.

ri motive engine. NSU releasecl the test engines to thenr, having full con-
r
Dr. I-lirscfi lvas only An assistant attached to the boarcl of directors,
':.
i-*l:
t.r fidence that they would act honorably, Do matter what their conclusions
but it did rnanngenrcnt no goocl to explain that f)r. I{irsch hacl vastly
,
i,t"r. :

in the technological fielcl. f)r. Albers sounclecl the :rlarnr artcl denratrde«l
oycrsteppcd his authority. N,{anagement nraintained that thc timetablc
ii'

an explanation from NSU, whose management made no comment.

:.l ,'-, ' for the start of procluction indicated 1962 or 1963 rather than mid-year
Then he started to spread rumors about a payment of $5 million from
1961. But speculators as well as stockholders preferred to believe the
i,

Daimler-Ben{to NSU in the hope that confirmation or denial would

t;
nlore optimistic Dr. Hirsch rather than the conservative management.
be forthcoming. Actually, the deal did not involve any money at all; it
:. .:,
.,

T5e net result was the rnuzzling of all of NSU's management men, in-
1

. .,:'
was simply a verbal agreement between the directors of the two com-
clucling Dr. Gercl Stieler von Fleydekampf.
.l I

panies that all Daimler-Benz's contributions to the development of the

...i Vori I{eyclckanrpf clebatecl with hirnself whether or not to resign,
Wankel engine, including any patents they might take out, were to be
but 5e stayecl on. There was a conrplete ne\4's blackout on Wankel en-
ti
1

:r .\
made available to NSU at no cost. Dr. Wolf-I)ieter Bensinger of
' :1.
gipe clevclopment-no interviervs were givett, nobody visitecl the factory,
Daimler-Benz's engineering department translated the NSU power unit
no information was release«l. Stock prices fell to the 1,800 mark as a
\, I
;lE

result of the lack of press co\/erage of NSU activities. Then, it No-

:f i\
:1.-+
into proper Mercedes-like dimensions and began building test engines
of his own. But the acquisition of a Iicense was postponed indefinitely.
vernber of 1960, the Drcsclner Bank be gan to aclvise its customers
:'

On February 25,1961, NSU issued a Wankel engine license to Yanmar

againsf buying NSU stock. On December 12, 1960, the announcement
l,j

.i

Diesel CompäDy, Ltd. of Osaka, |apan, covering gasoline engines fronr

.
...).

tlrat Ficlrtel & Sachs AG of Schweinfurt hacl signed a license agree-

^,.::.,
,?

,
I to 100 horsepower and diescl engines from I to 300 horsepower for all
ment with NSU for procluction of Wankel engirres in the 0.5 to l?
:,
"i.

l:;
applications other than motorcycles, passenger cars and aircraft. Two
5orsepower range procluced no rise in tlte price of NSU stock, rvhiclt
't ,tr
days later, NSU signed a contract with another |apanese firm. Toyo
.,*'
on Decentbe r 7 was I istecl at | ,7 55 .
.

Kogyo bought a license to manufacture Wankel engines for gasoline

l, -l'
,'{
The f)resclner Bank gaiued dcfirrite tax advantages frorn the fall of
,'.,1..

'.{, '

i.

;.
' ir'l
{
t"{'
,4
".,'i;l
154 DEVELOPMENT NSU DEVELOPS TIIE WANKEL I55
fucl with power output frorn I to 200 horsepo\t/er, for application within AII these license agreements antedate the actual production of Wan-
thc framework of 'foyo Kogyo's product line as of 1961. From these kel-powered cars by NSU. It was in mid-1963 that von Heydekarnpf said
arrangements, NSU was said to have collected cash payments of NSU hacl plans to nrake 5,000 Wankel-Spiders in 1964. The car was, in
$ I ,2 50,000. fact, shown at the F rankfurt Auto Show during the week of September
About this time, NSU decided to send a letter to its stockholders, 12-22, 1963. It
was a white hardtop coupe designed by Bertone, re- a

warnirg them ag^inst bcing overly optimistic about tlre Wankel engine.
I

sembling the Sport-Prinz, t

This sent the stock down to 1,200, but it u,as soon to go up again. NSU
rl

With an actual vehicle beirrg readied for production, a ferv other i

i
:

invited 700 of its clealers to Neckarsulm to show them some Prinz proto- companies became interested. i
types powerecl by 400 cc. Wankel engines. These cars were strictly Rheinstahl-Hanonlag signed a license contract for Wankel engines t.
).

cxperimental, huniedly cobbled up with Volkswagen transmissions and of 40 to 200 horsepower on December 19, 1963. On April 15, 1964, i
Ford radiators, but they would reach 100 m.p.h. on 50-55 horsepower. Alfa Romeo, which had been experimenting with Wankel engines of l-

There was corrsiderable doubt about their longevity, but the net result their own design for several years, signed an agreement covering gasoline- :
of the experience was to send NSU stock up about 100 points. fuel Wankel engines from 50 to 300 horsepower. Alfa's interest dated i:
In 196l , Hutzenlaub sent drafts for license agreements to Citroön back to May, l959,when there were rumors that a merger between AIfa
ancl Renault, Krr,rpp, M.A.N. and Klöckner-I{umboldt-Deutz. Both Rolls- Romeo and NSU was planned.
Royce ancl General lVIotors sent represcntatives to NSU to examine Rolls-Roycc Ltd. of Derby bought an extensive license covering diesel i
and evaluate the engine. Renault hacl sent a tcam of engineers to NSU
,,n

ancl hybricl Wankcl cngincs fronr I00 to 8 50 horscpowcr on licbruary

irr 1960, where they werc permittecl hunclrecls of nrarr-hours to study 17, 1965. The following n)onring, NSU sold il Wnnkcl engirre liccnse ih
ir
;i.r
).,
.1

the Wankel engine. The Irrenclrnren prctenclecl that RcnanJt lracl carlier to a firnt irr Conttlluttist lllast Gcrnrarry (Dl)lt)-Vcrcinigtc Volkscigcncr
.".1
-:1
,;l
t!
cleveloped a rotary engine of its own-their stucly of the Wankel engine Betriebe Autorttobilbau. 'I-lte agrccntent specifiecl gasolinc-fucl cngincs :l
was purportedly a strictly acadernic irrvestigation. in two ranges, frorn 0.5 to 25 horsepower and fronr 50 to 150 horse-
:l
1
,l't

Their attention, holever, centered on the Wankel engine's sealing POwer.

I

system. As soon as the Renault team returned to Billancourt, they filed Dr. Ing. h.c. F. Porsche K.G. of Stuttgart-Zuffenhausen acquired the :{1
i'i
for Iirench patents for the same sealing systenr! Renault then brought rights to manufacture gasoline-fuel Wankel engines from 50 to 1,000
.4'T
:l a
"t
.rl
forth the olcl Cooley engine and pretended they wantecl to try to ap- horsepower by its contract of l\{arch 2, 1965. .f
,x
4

ply Warrkel's sealing systent to the American design.

r '1,

When the Warrkcl engine was still in the cxpcrinrerrtal stage at NSU,
:i
it

Of the other interested parties, Citroön was not to enter any formal
1,4

an Israeli group acquired 5% of the NSU stock. As progress continued, I

agreenlent for some years, and the story of its joint ventures with NSU
,.:1

tlreir holdings-increased to 12% by the tinre of the \/olksg,agenrverk .t

.)
(Comobil and Comotor) is told in the chapter about Citroön's work
d

takeover in 1969. This group fornted a company incorporated as Savkel, ,J

I
with Wankel-powered cars. Klöckner-Humboldt-Deutz AG of Cologne
'I
.1
under Israeli laws with headquarters in Tel Aviv, and \+,as able to acquire li
,l

acquired a license covering diesel engines of all sizes, for all applica-
rl

a Iicense for manufacturing Wankel engines in Israel. NSIJ/Wankel 'I )

..t
tions, on October 4, 1961. f)aimler-Benz ÄG formally took out a license granted to Savkel the right to manufacture gasoline-fuel engines of 0.5
'l
:,It
for Wankel engines from 50 horsepower up on October 26, 1961. Four to 30 horsepower for all applications, except for land vehicles, aircraft .
I
,.1

clays later, M.A.N. Maschinenfabrik Augsburg-Nurnberg AG signed an

.,1

and marine installations. .i

I

agreement with the exact same stipulations contained in the Deutz

.l

General Motors invited the technical press to its "Progress of Power"

,l

v.t

contract. On November 2, 1961, Friedrich Krupp of Essen signed an

..t

Presentation at the GM Technical Center in Warren, N{ichigan, on , .,1

,2
identical agreement. These agreements with German firms were es- Mry 7, 1969. A number of unconventional power units were demon- .,
.f .l

tinrated to have brought NSU a gross license fee income of about

_t

stratecl, including steam cars, electric cars, a gas turbine bus, apcl :,.1

$1,500,000. F. Perkins, Ltd. of Peterborough, England, a subsidiary of

...,1

various hybrid power systems. The \Vankel engine, how,eyer, \r/as miss- )
I

Massey-Ferguson of Canada, acquired a Wankel license for gasoline

,i
ing, but runtors of GI\,I's interest in the Wapkel engipe remaiped '.t_-!
'l
and diesel engines up to 250 horsepower on August 8, 1961. This deal strorrg. On fune l, 1970, Gencral l\{otors confinnecl that it was fioldirrg
.,t
'.
,,
.J

brought NSU a sum of about $37 5,000.

I
't
discussions with Wankel G.m.b.H. According to a report from Frank-
I
r,i
,l
.rl
-(l
'.,
t

I
I
 d.

It'
It DEVELOPMENT NSU DEVELOPS T'I-IE WANKEL t57
t16
to make
I
t.
a engines run efficiently on non-lead gasoline, Nissan also is
furt, GM was interested in obtaining 40% of the of Wankel
t
! shares
engaged in irltensive experintents with electric cars, steam epgines and
G.m.b.H. Industry sources in Europe said GM was conducting the
I,

rt gas turbine engines, in parallel with improvements to the conventional

r$ talks through its West German subsidiary, Adam Opel AG. These internal conrbustion engine. As part of its program, Nissan had kept
i§: sources indicated that GM was offering the equivalent of about $27.5
#$' tabs on the rotary-piston engine and had bcen negotiating witfi the
ä[
million for the 40% interest. This would Iet GM in on the ground German conrpanies for quite some time.
ilt:
floor of all present and future licensing arrangements. A GM spokes- The other l^p^nese contract was signed in |apan on November ?6,
man said only that "discussions are being held with Wankel G.m.b.H. 1970 by representatives of Wankel G.m.b.H. and Audi-NSU Auto-
as part of our stated policy of investigating all possible sources auto- Union AG and the chief executive officers of Suzuki Motor Company
#fi
rü motive." The outcome of the GM negotiations will be fully discussed Ltd., of Hamamatstt. The licensing agreement covers manufacture and
f;i: in a later chapter. distribution of Wankel engines of 20 to 60 horsepower for motorcycles.
ft Before the end of 1970, two other |apanese companies signed agree-
iI
$i; ments for the rights to Wankel engines in certain power ranges. The
Suzuki produces some 500,000 motorcycles a year plus two lines of small
n. d'
Passenger cars: the rear-engined Fronte 360 and the front-rv[eel-drive
first was Nissan Motor Company of Tokyo, manufacturer of Datsun cars
11. f,,
.).
Fronte 800. Suzuki has a labor forcc 10,000 strong, and annual turnover
ia.:

ss
ifr and Nissan cars and trucks and parent company of an industrial group for the 1970 fiscal year will approach $300,000,000.
$fi including Ituii Heavy Industries and Aichi Machinery Company. Nissan
,. i: How many of these licensed companies have developed the Wankel
i. tJ
is placed fourth among ]apanese inclustrial establishments. The corn-
engine concept for their own applications is covered in the following
;'; ll
t
ii pany began to develop into a large-scale enterprise after World War II
i:
j,t, chapters.
i,t
'l t' when Nissan resumed automobile production in cooperation with Austin
+l
t'$'-
t'ä
Motor Co. Ltd., Longbridge, Birmingham, England. Mergers with
Prince Ltd., and Isuzu Motors Ltd., Iaunched a steep rise in economic
il
[: ä:

ir growth about five years ago.Strengthened by an ever-growing procluc-

:* tion capacity, the company now procluces a wide range of moclels, start-
it:
tr'

il'
X' 1., ing with the small Datsun Cherry, Sunny, and 510, including the Dat-
l'i
t,r'
:E z-
sun 2102 sports car, the medium-sized Nissan Cedric and the Nissan
9{
1i :Ir
lr President with its 4-liter engine. Today, Nissan Motor Co. Ltd. has
;
t: about 48,000 employees in different factories on the coast and in the ad-
I'
t.
I jacent hinterland of Tokyo Bay.A total of 1,230,000 vehicles came off
).
;:
t:
the production lines in 1969. With the extension of the Tochigi works
completed, an annual capacity of 2 million cars is planned for L972. The
company had a turnover of $1,850 million for the 1969 fiscal year.
Nissan's license from Wankel G.m.b.H. and Audi-NSU Auto-Union
AG entitles the |apanese company to produce Wankel engines from
i 80 to 120 horsepower for passenger cars. That does not mean there will
It
r,
ta.
be Wankel-powered Datsun and Nissan cars in 1972. Mr. Katsuji Kawa-
mata, President of Nissan Company, signed the agreement at the NSU
T
:i

head office in Neckarsulm on the afternoon of October l, 1970. Nissan

t

I
has no inrmediate plan to produce the engine on a commercial basis,
i
I
;
but will instead push ahead with research and development in the in-
terest of exhaust emission control. President Kawamata, however, com-
t
I
;
I
f mented that the rotary engine would play the part of a reserve unit
i
,, rather than the mainstay in the fight against air pollution.
Faced with the prospect of tighter emission standards and the need
I
t

.
 I

CURTISS.WRIGI{T 159

rapidly to meet the demand, expanding still further after the United
10 States entered the struggle. Curtiss concentrated on airframes while
Wright devoted most of its capacity to engines. A subsidiary, Wright-
Martin Corporation, built the famous Flispano-Suiza V8 aircraft en-
gine under license. After the war, Curtiss planes an«l Wright engines
continuecl to set recorcls. It wils a Wright "Whirlwirrd" engirrc that
i

ti powered Lindbergh's plane, "The Spirit of St. Louis," on tlrc first non- :i
r1

Ifi stop flight from New York to Paris in 1927. After the 1929 merger, the :l,!.1

li:
I !r Curtiss-Wright Corporation became an integrated aviation manufactur- t
I
,!
{,
ing concern, bringing about a new era in air transport.
Curtiss-Wright
I
E
tl
a,
t.
I
{:
In 1936 Curtiss-Wright president Guy Vaughn dissolved all existing .i
t

xi
F
il
subsidiaries except Wright Aeronautical Corp., and operations were .l

taken over by the parent company. Wright Aeronautical Corp., 98%

$i
r.
Ei
l|
I,; owned, was merged into the parent firm on Oct. 31, 1951, the business
f.
üi
ii being continued as Wright Aeronautical D,ivision. During World War ,

II, Curtiss-Wright was a prime manufacturer of pursuit planes, dive

fit
t.
I:
t;
Ir
{: KNowN MATNLY FoR Irs contribu- bombers, scout observation planes, transports and of aircraft engines
I

i
I tions in the world of aviation rather than land transportation, the and propellers. For a number of years, the company existed mainly on
f
iL Curtiss-Wright Corporation is a relatively young organization, having defense contractsfrom the U.S. government. It let itself be bypassed
been established on August 9, 1929 by a merger between the Aero-
L
t,

when the turbo-jet gas turbine engines came in, and its turnover sank
§'
,1

a
plane & Motor Co., the Curtiss-Wright Aeronautical Corp. and their in one year from $599,000,000 to $389,000,000-a )5% drop.
relatecl and subsicliary companies engaged branches of the
in various
,i'

{i
1
aviation industry. The Iatter group included Curtiss-Wright Airplane
I
Co., Curtiss-Wright Airports Corp., Curtiss-Wright Export Corp., i
&,'

i
{
Basic components of the IRC-6.
i,
i
Curtiss-Wright Caproni Corp., Devon Corp., Keystone Aircraft Corp., 1

s,
l,
Moth Aircraft Corp., N.Y. Air Terminals Inc., and N.Y. & Southern \
7

l.
{
+
Airlines, Inc. I
I

i
i.; Both the Wright and the Curtiss interests had traditions starting from t

|:
i
the very beginning of controlled flight. It is proper to say that Curtiss- I
I

Wright's history began when the Wright brothers made the first
1

;
l
f
powered flight at Kittyhawk, North Carolina, on December 17, 1903.
I

"ä-f
-w-r
I
I (
I
I
t In 1909 the Wright brothers formed the Wright Company of Dayton, 1

l,
t.
Ohio, to manufacture airplanes. In I.911, another intrepid aviation l
i

pioneer, Glenn H. Curtiss, sold the Army its second airplane and the
t

gb 6
.1

i' Navy its first. In 1916, he formed the Curtiss Aeroplane arrcl N{otor ;
1

Corporation.
In 1907, Glenn Curtiss, together with Dr. Alexander Graharn

O
i ;

Bell, I. A. D. McCurdy, F. 'VV'. Baldwin and Lieut. Thomas Selfridge,

O
formed the Aerial Experiment Association which built a number of Ilr.-\s

C)
I
''l
I

'
airplanes and performed various aerial experiments. t
FTF\

The outbreak of the war in Europe in 1914 gave a decided irnpetus

to the demand for airplanes, and the Curtiss and Wright plants grew
-
)
160 DEVELOPMENT CURTISS-WRIGI{T 161

Bucking the trend, Curtiss-Wright continued to manufacture piston- Irorsepower Cyclon e 7 .It never really got into competition rvith Pratt &
powered aircraft engines long after the turbo-prop gas turbine had Whitney and General Electric for the lucrative contracts that were
proved its clear superiority for applications where turbo-iets were not being offered in the late F'ifties and early Sixties to supply gas turbine
considered suitable. Radial piston-powered aircraft engines still in pro- engiies for the new-generation civil aviation airplanes. Irrstead, the
duction in 1965 were the 3,700 horsepower Turbo-Compound, the 2,800 .oiporation started to look at diversification moves as a possible rvay out
Irorsepower Cyclone 18, the 1,625 horsepower Cyclone 9 and the 800 of its troubles.
Curtiss-Wright president Roy T. Flurley, a former Ford Motor Com-
pany executive, tried hard to generate some enthusiasm for the com-
Elevation and cross section of the IRC-6. a : axial width of chamber. R - rotor prry'r activities by backing technical innovations. The first time was
radius. e= eccentricity. b : maximum clearance along maior axis.
1956, when Curtiss-Wright built an experimental air-cushion vehicle. It
was an automobile without wheels, supported on a layer of air forced
underneath the car by fans and ducts and kept frorn escaping by skirts
around the entire vehicle perinreter. It remained purely experimental,
but it received favorable press reaction, and quotations for Curtiss-
Wright stock went uP.

Full throttle performance of the IRC-6 with side intake port. The
engine developed 100 horsepower at 5,400 r.P.T. f!. torque curve
stayed within a 20 foot-pound range from 2,000 to 6,000 r.p.m.
120
TORQU E

./ too
\ o
/ to
J
t20 q
_g oj-
A lr.
#t t
lrl
/
Ja'
!00 Oe
o
G
I. R(lI(lR WITH IIITTRIIAI R(lT(lR GEAR
o
2. STATI(II{ARY GEAR BRAK E
3. R0I0R HoUSHG 80 H CRSE PO WER
oF
I. EI}IAUST P(IRT E,
5. SPART PIUC lrJ
6. S|DE H0USllG - DRIYE SIDI
o 60
= G,
-
\ .9
,. sltl[ H0us[{G - A}ITI.I)RIYE SItlE o- !
E. IilTAXT P(IRT lr, o-
g. rAtlt BtARrilc (tliltEn) 3t,
.6 -to
,y
o 40
E,
t0. MAril BEARHG ((lUTEnl
:E / \o
tt. BArAlrct }IflGltT
J
t2. trYwHEtt
lrJ
v,20 \ S.EC.
.5 I
13. tGlilIttlr{ cot{IAcT }lAxtR
E, / \ I Y- o
I - AIIAI. WIllIH OT CHAMBER trl IT
R
- GTNENAIIIIG RADIUS o
t - J.4 to
a

tccEr{rR!üTY
ob rooo 2000 3000 4000 5000 6000
k
- R/c
! _ llAllllut tßEt0il 0f CHATBER EN GINE SPEED - RPM
 \. CUI\T'ISS-WRIGIIT 163
ög
u roo ,/
nouncing a new technical miraclc to boost the stock price so as to sell
Y/ /
TL VOL UMET IC
lrJ
() EFF ICIE N personal stock at a huge profit-before annorrncing the cut in diviclen«ls.
G 90 No action was taken against Curtiss-Wright, l>ccausc [Iurley not only
F had acted in good faith but had also kept lris Curtiss-Wright holclings.
trJ / As if to adcl further proof of their belief in the Wankel engine, Curtiss-
= 80
J=
o
/ Wright began an intensive research program under the direction of
70 Max Bentele. Bentele was the expert who hacl sold Hurley on tlre idea
of the Wankel license in the first place.
I20 0 This Gernran-born cngincer \rras 49 years old when Curtiss-Wright

II
signecl the contract with NSU. Bentele had stucliecl at the Technical Iu-
loo o
,/ stitute of Stuttgart and got his engineering cliploma in 1972. Post-
graduate studies earned him the title of Doctor of Engineering in 1938.
/

7
AIF FL OVl
800 Bentele left Germany in 1956 to accept a position as staff engineer with
G,
Curtiss-Wright. He worked on the design of the single spool gas turbine
\-to 600 U-65) and reciprocating aircraft engines, then hc becarne manager of
J
I
400 [- /
o
= ,/ Wide open throttle performance of the IRC-6, u'ith peripheral in.

I
J take port, shorvs Iower intakc Iosses with correspondingly higher air
TL flow and outl>ut.
g 200
r20
o IUE \ o
o rooo 2000 5000 4000 5000 6000 7000
TOR
!oo @
J
ENGI NE SPEED- R P M ./:
,/ t
F
80 IL
/ I
Volunretric efficicncy easily exceeded 100% with the peripheral ld
intake 1>ort. I Iowever, the peripheral intake port proved slightly I40 60 :)
o
inferior with regard to air utilization (compared with the side port E,
version), despite air flow characteristics allowing an intake of almost Z o
1,100 1>oun«ls pcr hour at 6,500 r.p.m. 120 40 F

The second time was Jate in 1958, when Curtiss-Wright purchased roo / ,/
the [.J.S. rights to the Wanke] engine frorn NSU Motorenwerke AG.
Again, the press reacted favorably, and the stock market responded in E BRA E /
very positive fashion. The $l shares had been selling for about $30
lrJ 80 t- ORSEF

for a long timc. After thc Wankel engine deal, the price rose to $40.75.
o
=
o-
All buyers hoped that through manufacture of the new type power- ut 60
plant, Curtiss-Wright woulcl soon be able to pay fabulous dividends.
U,
G /
o /
Then the conrpany announccd that its quarterly dividend was being E 40
lrJ
cut in half because of other changes in the profit picture, and the stock Y
price fell again. This brought trouble from the govenrment: I Iurley
/
E 20
was callcd up to answer to the Securities and Exchange Cornmission, @

who suspected stock manipulation. It looked like a classic case of an- oo'
I OOO2000 3000 4000 5000 6000 7000
ENGINE SPEED- RPltl
16+ DEVELOPMENT
IRC6 TEST PERFORMANCE +

engineering projects in 1958 and embraced the Wankel engine with FULL THROTTLE
i.

great enthusiasm. I{e was in charge of all rotating combustion engine S MALL VS. LARGE SIDE INTAKE PO RT

projects from 1958 through lg$ and he left Curtiss-Wright in 1965. t,

Bentele's place was taken by his former assistant, Charles |ones. t20
U'
]ones was chief project engineer of Curtiss-Wright with
responsibility o
- lllr.e
s )RT
J
design and all rotating combustion engines. He /
-'S_

foi reciprocating engine too

-Curtiss-Wright
.f
I

ioined in 1950 as a test engineer and in 1955 he was \

F
lr.
o*
made section head for charge stress and applied mechanics, then he
worked as a proiect engineer on the l-65 engine in 1957 . The followirg P 80
f
I
lrl
60 o
E
The side port gave high volumetric efficiency over a wide speed o
range. The peak cfficiency was 90%, while the peripheral-port engine F
exceedcd 100./o. Mechanical efficiency was low at low speeds, indi-
40
catirrg low rnechanical friction. Friction losses increased moderately
with rising speeds. The thermal efficiency curye for the side intake tooo
port version of the IRC-6 shows )4% to be optimum for maximum
econorlry over a wide r.p.m. range. E, ,/
{t, -\ 800
./
l oo
ll
E
v0L JMETRIC EFFICIEN }Y
o
J
@
-,
a /
UJ
() 600 /
^l

/ / //
I
G
F 80 o 400
=
lrJ

=
J
IL / ./
f
J
o / E
60 a 200

o
o tooo 2000 3000 4000 5000 6000 7000
.ft oo ENGINE SPEED - RPM
t§
t!
MEC HANICAL EFF ICI EN ]Y
UJ
J
Port size proved to tave a great influence on engine power and speed
characteristics. The small port showed higher torque at lower r.p.m.,
9 80 \ air flow being considerably higher in the critical 3,000 to 5,500
z r.P.m. range.
T
o
.
lrj
60 yer he got his first assignment connected with the Wankel engine.
For the next four years, he designed experimental Wankel engines of
all sizes. In 1962 he was given responsibility for all rotating combustion
40 tN0rcAl IAL EFFI( IENCY engine design, including basic engine development applications and
aI
I rST ECON OMY advance design. Hurl.y, who had made the decision to back the Wankel
lr
L
lr,l engine, Ieft Curtiss-Wright only about a year after signing the agree-
J 30 ment with NSU. His place was taken not by an engineer, but by a
= Iawyer, T. Roland Berner.
Apparently it was never Curtiss-Wright's idea to start producing
G,
lrJ
I
F 2oo' 3000 4000 5000 6 000 Wankel engines of any kind in its own factories. They planned to study
I OOO 2000
applications, develop basic units, and sell licenses to other companies.
ENGI NE SPEED - RPM
 166 CURTISS.WRIGI-IT
(

DEVELOPMENT
r)
)
167
:

This has not, as yet, resultecl in any contracts of major importance. To c) In North America on marrufacture by Curtiss-Wright-Nsu
date, Curtiss-Wright's revenue from the Wankel engine rights is pa- ).
Wankel lyr% for automotive, 3% for other
tlretically small, hr 1969, the company earned $1,143,000 on the Wankel d) For subJicensing entrance fees and royalties received by Curtiss-
engine. That means about l?.0 per share was earned from Wankel royal- Wright from sub,licensing-Curtiss-Wright, 60%,, NSU Z4oÄ,
Wankel 16%. Flowevcr, if errgines arc cxportccl to Gcrmany-
:

ties, against total Curtiss-Wright earnings of $ 1,36 per share. The over-
all financial picture at Curtiss-Wright is so complex that it is difficult Curtiss-Wright 257o, NSU 457r, Wankel )0%
to sum this up as "good" or "bad." I

When you read about the Wankel-powered cars built by NSU,

I

: Experiments with dual side intake ?tRtPHERilL ttllAtE

ports on the RC-60 (the renamed t'\.
Toyo Kogyo, ancl Daimler-Benz, the question of what Curtiss-Wright i - - - OUIL SIOE lltlAXES oo

has accomplished in the Wankel engine department since taking out

k

i Iater versions of the IRC-6) show - \tr

I

:
I almost equal airflow character- uJ

its Iicense more than 17 years ago is inevitable. The engineering develop- !
istics to the peripheral-port en- 8oE
(,
: grne up to 3,500 r.p.m. The pe- F
ments will be covered later in some depth. On the commercial side, the .
td
!
ripheral-port engine has better
achievements have been more visible to the public. Three things have
f
volumetric efficiency above that
,

:
!
, ! ooo 609
happened: I
speed.
l. Curtiss-Wright succeeded in re-negotiating its license agreement
t
I
a

!oo
with NSU in 196+.
i
I

,l

Z. Curtiss-Wright contracted with Westinghouse to develop Wankel- I

ft
J
ß
powered generator sets in 1965. Il 2 aoo
3, Curtiss-Wright sold a sub-license to Outboard Marine Corporation
I
I ]I

oJ
in 1966. I
I
tr aoo
E
e
The new contract with NSU called for revenue distribution on Ii- I

cense fees and royalties according to the following table: aoo

a) Outside North America-Nsu 547o, Wankel )6%, Curtiss-Wright
r0% o
b) However, if engines are imported to U.S.-Curtiss-Wright 757o,
o t@o 2000 )ooo
CRANXSHAFT SPEED. RPTI
.ooo 5000

NSU 157o, Wankel l0%

The peripheral-port engine was

PIug voltage required for the IRC-6 was far lower than that re- fitted rvith an experimental type
-f,Y
2q
quired for a large automotive V8 piston engine. of self-adiusting apex seals. It is
comparcd rvith a sicle-port engine
22 ßEPRESE}ITATIVE
running with solid apex seals, in
vE AUTOUOTTVE EXolllE
zo coxPRESgtON-RArrO tO't 9 terms of peak compression pres-
Ir cUt sure. A side-port engine will de-
I velop its theorctical pressure rise
lc IJ
at lower r.p.m., because the effec-
3
D
J' l4 UI
UI tive low-speed compression ratio
F
J
o l2
bi
t
a
is higher, due to earlier intake
o Z closiug with side ports. But tests
lo
xJ
o
that with the self-adiust-
a \ \ r.:
a
o
L,
G
showed
ing apex seals, the peripheral-port
6
A
I engine had considerably higher
o
I pressures from cranking specd to
I 3 idlirrg specd. The difference at
t
a
opera tional speeds is less than
onc point (in relation to an 8.5: I
sTCeDll conrpression ratio).
I
XILO IPAßXS PER II}'UTE
ao ao to r@ 20ü .00 a')C t'.'J r](^/ lc,oO IOOO
I cRArilslrlfl sPEeo - rpm
I
I
tI
I
t
168 DEVELOPMENT Nov. 7, 196l M. BENTELE ET AL 3,007,460
COOLING ARRANGEIIENT FOR ROTARY UECHANISIIS
The deal with Westinghouse was announced on December 3, 1965, Filed llarch 29, 1960 5 Sheets-Sheet I
and concerned the development of a Wankel engine to provide basic
power for a lightweight electric generator set developed by the Aero-
space Electrical Division of Westinghouse Electric Corporation.
Being offered initially to the military services, the Westinghouse
generator set is seen as meeting the need for a transportable unit to
supply precision power to such forward combat area systems as radar,
communications networks and mobile missile Iaunchers. Military 124
t21
equipment of this type requires a close'tolerance power supply for elec- t4
ur,,
tronic components, but the power generator equipment must be capable t24
97
of being moved on short notice along with the tactical gear. When
mobility and transportability become paramount, weight of the equip- ?A
ment ancl fuel consumption, because re-supply can be difficult, are
major considerations in design. The Curtiss-Wright RC engine was

*ffi ä:f 'fru:'l::o'l:"1;i';t"",'il1':i'ü1"H:il;*:t;f

r.p.m. The torque curve stayed at I l0 foot-pounds from 5,000 to ,ii, il24o6 -1
5,500 r.p.ffi., and volumetric efficiency exceeded 100% between
1,250 and 7,000 r.p.m. "5'd3

8HP-
\zr-' \\
t50
/
llO /
E,
7
/
lrJ
I
I
tri
ro
an ta
E,
o
Z, / 58

r
trJ 90 t{o
x -i
r>
tr g
/
@
I
70
- 20 rrt
f,
o
E
o ,^,t'H[8ft.e,
trHAFILE:t JONE,3
jl \ )-
r EROINANO P. SEILLINßE,R
il* A.ß,h
/ /_
/tt"* \ EY
ATTOFINry
/ co
s
E.
I

l!
LJ
j Basic patent for the rvate r cooling system developed at Curtiss-
o Wrigh-t, shows passages of varyirrg size according to heat distribution.
60
.ooo 3@o 6@0 7000 aooo
CRANKSHAFT SPEED - rpm
Nov. 7, 1961 M. BENTELE ET AL 3,007,460 Nov. 7, f 961 M. BENTELE ET AL 3,007,460
COOLING ARRANGB,IENT FOR ROTAßY rEcluNrsus COOLING ARRANGET'ENT FOR ROTAßY uEct{ANlslls

Flled Xarch 29, 1960 3 Sheets-Sheet 2 Flled llarch 29, f960 5 Slrcctg-Sheot t

t
,allz, t'- \ n?r-llLLfiä

\\
\\ -
\-
\-

-tt-?'-/

l-1 '1 r-l

|.fl +t-1

bi;- --!4 -
?rf.LI[) t
,2c ttsa
92

lNyü,tT0R's
#Afr,ß'fraff 7 MAX EENT ELE,
E,HAFILEII JONE,s
TE,RDINANO P. SDLLIN6E,R
)YfuA. The upper sketch shou's the hert clistribution cun'|e (ZOl as estab-
ATTtrRNEY tirtr.a'bvcxpcriurcnts at Curtiss-\\'right. The lorver sketch sho§'s
how ll{ax Bentclc propr:scd to cool a trvin-rotor ltousing, usin-g dis-
tribution to both riniti frorn a cornn)on pipe, individual circulation
Proposed coolant guidc-vane configuration for the end covers and ihr"rgh four encl covers, ancl return flow through separate outlets
housing. into a common PiPe.
M"y 26, 1964 M. BENTELE ETAL 31134,537 J,,ly 23, tgGl M. BENTELE ETAL
COOLING STRUCTI.'RE FOR ROTARY UECHANISIIS COOLING AilD LT'BRICATION SYSl'fl FOR ROTART
Flled Feb. 15, f962 2 Sheete-Sheet I Fltcd Iay n. 1960

,I

l-r1 r-l _1
-TTITIE
t-H
l-J

--l-J
F11
löH
r-'l

---
l-r
I
.l
4ErJ
INVENTORS
l-J MAX ETENTELE
MAfä?ffi+t,-e CHARLES J()NES
E,HARLE 5 JtrNEs ALEXANDER H. RAYE
BY

tlryrfu7t,Op,btfu
ATTORNEY
ATTORNEYS
One of Curtiss-Wright's first patents for supplying and draining
Radial separating rihs between the coolant passages were claimed to rotor cooling oil via the eccentric shaft. e : eccentricity. 2 - urinor
reduce the tcndcncy to produce chatter marks on the working sur- axis. l0 - rotor. 14 : eccentric axis. 16 - nrainshaft axis. 18 =
f ace. The anglc of the ribs u,as intendcd to add strength to the working surface. 34 : ring gear. 36 - rotor gcar. 54 = central
housing. passage. 56 = radial passage. 58
- annulus inlet.
 July 23, 1963 M, BENTELE ETAL 310981605 July 23, 1963 M. BENTELE ETAL 3,098,605
CCOLINC AND LUBRICATION SYSTEU FOR ROTART UECIIA}II§IS CCOLING AND LUBRICATION STSIEil FOR ROTARY uEcr{ANISTS

Ft lod llay 21, 1950 5 Sheete-Sheet { Flled Iay 27, 1960 § Sheetg-Shect 5

,\ FlG.,5,

\
§
s
r)

7. G. 6.
ld

INVENTORS
MA,I*XEN+EIE MAX BEN'TELE
JONES C,HARLES JONES
=HARLESH. RAYE
ALEXANDER ALEXANDER H. RAYE

,/';;, fialp*, .fr,hx, fu

BY BY

,forrli, fu o,r, 1uln t /,ät

ATTOFINEYS ATTORNEYS

Detail of the rotor cooling system in cross-section (above) and

elevation (below). Arrows mark the direction of cooling oil flow to and from the rotor.
 3,081,753
Müdl 19, 1063 R. T. HuRLsr ETAL 3,0811753 March 19, 1963 R. T,HURLEY ETAL
ROTAßY COUBUSTION ENGINES
ROTART COIIBI'§fIOT ENGINES
Ftlcd larcb 10, 1960 { Sbcota-Sbcct {
Fllcd laroh 10, f960 4 §bcota-Sbcot §

\ \
)
)

§l § I

\
§

INVENTOR.' -
\I I}WENTOU
\b RoY T:-HüirLev i?X,JäEHibEE

ü+'4:Y'A
MAX EIENTELE -JoHN-Ä- KIMBERLEY
BY
J()HN A. KIMBERLEY
,/r/-";,fu,,*.?A,e
/
nTToGtYEYs
frrr'".erv€rs

Elevation of the Ilurley rotor

g'it[ raised apex lands'
Cross-section of the Hurley rotor with raised apex lands.
 Müü 19, 1963 R. T. HuRLEy ErAL H. IENTELE ETAL 3'l7l590
3,0E11753
ROTART COIIBI'IITION EICI}TES coxsTnücTlotl ron RoIAßr cqßuttloll DlclllEti

fll.d Itrob 10, 1960 0 Sbrotr-tbrot I

{ Smtt§D..t 2

Fl G. 3.

F lG- 4- /,\
\,--2
-t6
,44J
I

FI G.

F lG- 6_

INYENTO&'
wtLLlAM
--'-M§( J- Dq!\IER
E}ENTELE
INYEI.ITOR
ROY T. HURLEY
MAX BENTELE
JOHN A. }<IME'ERLEY
?ä.,{.a7^ ?*(....r&
j.lsrg Bentele's Patent for irnProved oil
' ,€tra
elevation. T'lte principle is trvo'sta
56 and the
Details of the raised apex land and seal configuration. The top ing gears, i
drawing shows how exhaust gases are to be prevented from flowing the point o
beyond the exhaust port. two seal ri
,ätun . and force clependeut on rotor acceleration loads'
 Doe 3, 1903 M. EENTELE Srllzrt?O
H. BENTEL STII2ITO
§R COOI.ED ROTOR FOR ROTTil EtüIlIS
ln coqrD RoroR roR R0IAnI lG}tArtt[ 0 $..t -$rot 8
?11.d Juao E. t96f 6 sbcct6§[i3t I

-L

{,
't§
§

I /,§'

\l' §
,fivülmß.
tWex Beurete
N
,ry,,E,.,rr.-9,rr/*rrßa
-(
/ ' I )1^;'fu*?-L,Eu
DY

ATTot?NEYo,. N A'rroRNEYlt'

Air cooling experiments began with patents covering air cooling of

the rotor rather than the housing. 'This .ro5.i..[ion sho*, "tt . Bentele'
multiple ribs inside the rotor struciure. The inner makeup of the air-cooled rotor, patented by Max
lhs 3, 1003 It. lENrEr-E 3rt taE70 Dcc. 3, 1963 M. BENTET-E 3rl I2r870
Iti cuD Rolu Foi Rmrnil Elltfifl AIR COOL@ ROTOR FOR ROTARI ECIüIIIS
0, 1e0l 6 $ootr+Strit t Fll.d Juac 6, 1961 6 §bcctg-Sbjit §

/z
QA' scAL

g.
,RE' FRgE FRON
ENCtO.cHnExt oF
&ttJ ott tt Qat SSALE
y,
40

-
;Wf, r-frn
ORD'7 OF
otl sen,.

f)§ ?ERIFERY

K ancroR - 6

ARCA FREE FRON ENCßOAC,+

,9. MENT OF DOIH OIL

q qes §eeLs.

u t,,

K racroR t 9

7DY ilax
"!;'i!!!
I,NVENm,K
TI,TX EENTELE

ry,ry*%;k ry,%%v&.
DV

Y ATTaRNEYS.

Air flow tlrrough the air-coolcd rotor went from end cover to end Raising the R/e ratio (I( factor) can consiclerably increase the
cover, through ports in both rotor flanks. rotor profile area frce fronr any kind of sealing elements.
April 19, 1966 M. BENTELE 31248,630 April 1g, 1966 M. BENTELE 312461336
ROTARY COilBUSTION ENOINE AND I{ETHOD OF
OPERATINO SAXE
NOTARY COXDUSTIOII ENOIilE ATD IETHOD OF OPERATINO SAIIE
1 Shcctc-Shcct I
Orlglnal Fttcd llay 1, 1962 z sheets-shcct o Orlgtnal Ftled llaY 1 , 1962
t4?
FIATE

s11c
2l
II
rt'
,t'

'-L

26'

'h='J

..lEl
l-t r
_IE'

t*H JrJ
l-J

NVEI\IQR. -
MAxBEflTeur MAX.-EENTELE.-

//*
rblHIS ATTtrRNEY
DY

Il-) 'JLJ
-hrr
--
-l -tJ

jector and spark plug configuration for the heary-fuel wankel

Iniection _pump cross sectiorr, and f uel iniection rafe for Bentele's
patcntecl heavy-fucl engine.
In
-
.ngin., patented in 1966 by Irlax Bentele'
July 23, 1963 M, BENTELE E1AL 31098,605 July 23, 1963 M. BENTELE ErAL 310081605
CCOLING AND LI'BRICATION SYSTEI FOR ROTARY IIECHANI§IS! CCOLING AND LUBRICATION SYSTEil FOR ROTARY UECI{I}IISIII;

Flled Iay tf, 1960 § Sheetp§[;st 2 Flled Uay 27, 1960 5 Sbcetq-Shect U

F lG- 2
20

F G 3-

TNVENTORS
MAX BENTELE
CHARLES JONES INVENTORS
ALE}ANDER H. RAYE
I- € BY .*.H=3=]J§EE
,rtyi ft"'f-, ?til*,t fr't ALEXANDER H. RAYE
f4.,,,firbt/*
BY
ATTORNEYS rtloy*,
ATTORNEYS
Elevation of the rotor cooling system. Inside the rotor are three sets
of finnecl passages (64), locatecl adjacent to, but radially inward from
the apex portions. All passages are connected to an inlet manifold
and an outlet nranifold. Exploded view of the rotor cooling systcm.
 May 8, 1962 3,033,180
May 8, 1962 M. BENTELE 31033'180 M. BENTELE
RCTAT.ING COUBUSTIC}I ENGINE SEAL CONSTRUCTION
ROTAIING OOUEUSTION ENGINE SETL COhSIRUSrIOII

Ftlod JaD. n, f9'60 5 Süccts-Sboct I Fllod JaD. 29, f96O I Sibccts-Sbrot 2

E _3
MAX,B'flI$t,UE Tct
lNvü,tTor.
MAX AENTE LE

ATTORNE.Y

Max Bentcle applied for a patent for this sealing configuration in

|anuary 1960. 'I.hc solicl, full-width at)cx seals had an intricate in-
ternal spring-loading. 5? - apex seal. 56 : ioint trunnion. 54 = Eulirrged cletail of thc apcx scirl corncr irr Bentclc's sealing system.
spring. 62 : sidc seals. 5Z - apcx sertl. 54 : sPring. 56 - joint trutrniou. 67 - side seal.
May 8, 1962 r.. BENTELE 310331180 rManü 19, 1063 R. T. HURLEY ETAL 3,0811753
ßofArlrlc cq{Eu§tloll EilGINE SEIL CO|SIETIIO| ROTARY COUBT'SI:ON ENCINES
Pllod Jro. 29, t96O 5 5örotj-§i!311 ! Ftlod larob 10, 1960 { Stort;'§[11t I

60
a'?
,/
u
v-2/

l-r 1 r-l -E
l-H
l-J

,'\1e'i

tMvü{lok RoY *PYIUFrtte"

MAX- EENT ELE MAX BENTEIE
JOH]-i Ä. }<IMBERLEY
,{z^ä
/
/.*1,-/-,4Lr/L
ATTtrRNEY rA 77O'€t!Ef I

Hurley's pate,t covered a raised larr«l near each apex, to aid

sealing
Apex and side seal assembly according to Bentele's patent. and gas flow characteristics.
Mü'ch 2, 1965 M. BENTELE ETAL SrIZlr^590 Illerch 2, 1965 M. BENTELE ErAL 3,171590
OIL SDTL CONSTßI'CTION rOR ROTARY COIIBUSTIO]I EICI}IES
orL sEtL colls8ßlJc?roN roR RoTART colBusuor ElrcrxEs
Fthd D.o. 16, tg6o ! sbcctlagbcet z Fllod Dco. 16. 1900 0 §brtlrsü..3 I
FI F lG- 5-

,0
'lZ
.50
I}WENTO&' N.I\/SNTOR'
WILLIAM J. DERNER \^/ILLIAM
--'-M§( J. DERNER
MAJ( BENTELE BENTELE
nt
ll, n, F,'*7u, lrrl* A
U
l;( 75ö
72
BY
iloyn, I ?

ATTORNEYS ATTORNEYs
Closeup of the basic «lual oil seal configuration patentcd- by- -Mt*
alrcl_ are axially .flexible to
Be,tele. The .rrn,rlar legs cxtcrrcl inn,:rrcl,'l'he
Thcse two sketcllcs show the nrovcnrcnt of the oil I
seal vent hole i
insure contarct with thc enJ c'over walls. legs of the inner seals
during rotor rotation.
are stiffer than iir. i.gi of the outer seals and do not ad_apt !g-
changes
I
'l'he outer
I in tolerance, since tlrey operate with oil on both sides.
seals will confornr to tficnnal clistortions between the rotor and the
end covers.
f,lrr,ü 2, 1965 M. EENTELE ErAL J,IZI§90
OIL sE.TL COIISTRT'CTIOfl ?OR ßOTAßr COTBI,'Sf,IOr EilGIIIE§
lMaretr Z, 1965 M. BENTELE ETAL 3rl7lr590
OIL SETL CONT'TßUCf,ION FOR ROTARY COßIJSJrIO}I EIEI}IES
?ltod D.o, 16, 1900 6 SbootrSürot { E sbccti-§[333 o
Frld Dco. lg. 1960
F lG-6_ FIG- ll.

F IG. 9.
24-

t1-

F lG- lo-

F lG. 12.

I}tIl/ENTOR'
\/VILLIAM J. DERNER I}\I\/ENTOR§
MAX BENTELE
w'L1äXX{EE#[E
','Fr;rur, /trlnrr/-'fr(
W^rt,A.
It'yi,'F;
ATTORNEYS
'il,'1i''
J
rn'y', ?'r lx t /;
ATTORNEYS

Variations on the dual oil seal.

Variations otl the dual oil seal'
 CUITTISS.\\/RIC]IIT I9I
said that their comparries hacl signcd a Iiccnse agrccntcnt coverirrg clc-
velopment and manufacture of thc ncw-type po\r'erplant by OI\'IC, irr-
volvirrg thc payurcut to Curtiss-\Vright of fccs itttcl gurlritrttcccl r{r1,i1ltics
in excess of $1,000,000. ON4C, the world's largcst manufacturer of out-
board motors and a leading proclucer of stern-clrive marine engirtes artcl
fibcrglass boats, concluclccl ilgrccnrcttts rvith NSU i\ufotorcnwcrke AG,
Neckarsulm, Germfltry, ancl Wankcl G.m.b.lI., Litt«lau, Gcnnany, as well
as with Curtiss-Wright. These agreernents grartt OMC non-exclusive
licenses to manufacture and sell Wankel engines as powerplants in its
marine products throughout the world, ancl for certain nolt-marine pur-
poses in North America. ON4C also has options for rvorldwicle licenses
for a limited number of non-marine applications.
In addition to Evinrude and Iohnson outboard motors and stern-
drive engines, OMC produces several other prodttcts powered by gasoline
t
I
engines. They include Cushman golf carts ancl inclustrial velticles, Latt,n
Boy power lawn rnowers, Pionecr chairr sAws, ancl lilvinrtr«le arrd |ohn-
son snowrnobilcs. The negotiation of the liccnsc agrcctrtettt l>ctrvecu
Curtiss-Wright and OMC, Berner and Scott saicl, marked the culnrina-
tion of a co-operative developmcrrt effort carriccl ottt o\/cr a pcriocl of
several years. During this time, OMC had extctrsively tcstcd ancl eval-
I

I Installation sketch for the RC2-90 Y2 engine.

;
curtiss-wright's 4 RC-l g0-2 air.cooled engine.
I
chosen because of its relative simplicity
(only two moving parts in the
engine itself) and because of its weight,
size and fuel consumption acl-
vantages.
The moclel chosen for the Westinghouse unit
curtiss-wright RC2-60-NS. curtiss-wriitrt engineers
is clesignated t5e rah'
say the Ng engine
can be operated at a constant speed of
+,SOO r+.*. ancl is aclaptable to
generating equipnlent in the 4, to
60 kilowatt range. The Westing-
house genera tor set, inc]ucling tJre RC
engine, rveigSs under I ,l 00
pounds, comparecl to about 4,-700 pouncls
foi a upit ivit6 a cliesel en-
gine' Iruel consunlption with the Cürtiss-Wright
RC engine is expectccl
to be about 9., gallon.s per hour, comparecl to a
gas turbi,e-powered
unr-t of equivalent wcight rvhich would
i,r. 17 gallo,l, per 5our.
Curtiss-Wright Corporation ancl Outboard turrine I eoYftxoi to. c@!lt3 lli oucl It. OtL Croq-lf
Corporation ar- t OIL TILTEI il. tufL ltLSt tO DY.iT}OCTL IOUIT AIOI
nounced on March 2, 1966, that Outboarcl lr. r.l.?.
N{arine had obtained Ii- l. SrAiYfi tr. TUIL OUlLlr
It 7U3L ?UI, z. tro?3Lllt tLlLc
censes uncler patents to develop, nranufacture a orL ?ltltult tfxtot
and market Wankel en- t. lrfrrxEt I'. IOUIT äO !1 oll. Yltr
gines' [n a ioint statement, T. äo]and a. otL ltLEr lfrt tfxtot rt olL tul? t+ cltarrtCtc -clclra ,ltovtllct
Berner, Chairman and president rACXortrtt la. tlYDiluLlc rr,I, !a qL rilr 0rAlr
of Curtiss-Wright, ancl 'vV'. C. Scott, President of 7.
l. tmtr I'. YACUUI ?UI' !a @ol.rro lrLgll I oullol' tvtltr
Outboarcl Mari,e, a.
tLue
r^arclo ra. lLllttrrol
 192 DEVELOPMENT CUIT'|ISS.WRIGHT 193

so that
uated models of the rotating combustion engine for marine ancl other were relatively thick and stayed at low and uniform tenrPerature,
quite
USCS. they helped iestrict deformation of the inner walls, rvhich were
The starting point for the design of Curtiss-Wright's first experi- thin.
mental Wankel engine was the NSU DKM-54. It was redesigned on a Tle coolant passages were designed as a nrulti-pass forced flow system,
vastly enlargecl scale, and received the benefit of Dr. Iiroede'r kin.rnatic
-di.trt.,l by heat transfer requirements. In othcr
flow velocity being
inversion. It was called the IRC-6 and operatecl on the KKM (Kreis- worcls, the äoolani *m speedecl up in thc combustiorr
zolres to carry
kolbenmotor) principles. It had a clisplacäment of 60 cubic inches, or away as much heat as possible, ,,id slowed clown in other areas with
almost one liter, and it was eight times bigger in displacement than the patcnted by
lower heat input. The Äulti-pass water cooling system was
NSU KKM'125. The design was not started until an evaluation pro- Max RclteJc, Charlcs |oncs ancl Fcr«linartcl P. Sollirrger on Novenrbet 7,
gratn of dimensions, R/c ratios, porting configuration, rotor and fuouiing 1961. It took full accou,,t of thc fuct that tlrc hcat input to thc ltotts-
cooling, carburetion and ignition hacl been concludecl. ing is not unifornr arouncl its periphcry becausc cach of the operatiorral
Bentele had decided to stick with the three-cornerecl rotor ancl trvo- pliases always takes place acljacentto thc srure portion of thc housirrg
surface. This was met by clesigning channels of different size,
Iobe epitrochoidal housing, as developed by NSU. The design pattern giving
of the Curtiss-Wright Wankel engine was closely related to ihe KKM- different coolant flow veiocity according to requiretnents. The patent
l2r, but it was designed with the same geometrical relationship as the also covered the repeated return of the coolant through different
pas-
DKM-54. In the IRC-6, rotor radius was l1).1 mm., eccentricity was sages between the two end covers before exiting from the housing to the
:

16.6 mm., and the R/e ratio was 6.85:1. radiator.

The IIIC-6 nrainshaft ran in trvo sleeve bearings, one on each sicle Wit5 this clesign, the coolant flow pilssilgcs, plrticularly in rcgi«ltrs
of the eccentric, with two nrore bearirrgs outsicle the balance weights. of hig5 heat inpu1, have srn«loth hyclroclynarnic contours, that is, they
Save no abrupl changes in clirection or Areil. This fcltttrc serves
to
The flywheel was mounted separately and was easily detacfiable to
facilitate experiments to study its effect on engine operation and tor- preverrt the presence äf «Jcacl spots in the coolant flow Passagcs. Dead
spots are ,r.ä, in the flow passages having littlc or no flow vclocity
sional vibration cltaracteristics. An aclvanced liquid cooling systern was of
devised for the housing. It worked on a principle they called "multi- tLe coolant. By avoiding dead spots, any vapor produced in the pas-
z.
I
t
i

sages is instantly carriecl away by the coolant flow, which in

pass." The housing structure was made up of double walls intercon- turn Pre- I
I

'l'he serially con-

nected by ribs. The ribs formed passages for the coolant. T[e outer walls u.nt, hot spots resulting from vapor accuntulation.

Air cooling system configuration designed for the twin-rotor RC2-90 Y housing.
Cooling air flow through the RC2'90 YZ rotor housing.
RoroR HousrNG rNTErffolrDltJn / srDE

COOLI NG
AIR IN

HOT SPOT
ACCELERATOR

, .1-r I
COOLI NG
AIR OUT lva
AIR OUT AIR IN
+- <-

I
tt
DOWNSTREAM GRADUAL a

IN
I

F CROPPING :

{
 195
194 DEVELOPMENT CUI\TISS-WIIIGIIT
seal sweeps over the trochoidal track and
strikes the reinforced portions
nected groups of passages are circumferentially spaced about the axis
of the rotary mechanism ancl the passage groups are so connected that of the *rtt in raPid succession')
clecicling on an oil-
the passage group to which the liquid coolant is first supplied is Iocated curtiss-wright followed in NSU's footsteps by
coolecl rotor for the IRC-6. The first type
of rotor was an aluminum
acljacent to one cucl of a region of relatively high input to the outer alloy proviclecl the
body. As the coolant florvs through these serially connected groups of forgi,g wit5 forced-florv oil cooling. T'hä alurninurn
the Prirrcipal rnoving part
passages, it progresses arouncl the axis of the outer body toward the important benefits of great weighisavi,,g in
thcnlta-l cotrductivity' A
other end of the high heat input region. of the engi,e ancl a rotor structure witli high
in prevelrting the
The nrulti-pass, forced-florv cooling system is clog-free, prevents vapor high heat clissipation rate was particularly beneficial
lightu'eight construction
accumulation, and is casily nranufactured. The block ancl end housings formation of hot spots rvithin tire rotor, u'hile
greatly reduced the energy losses that resultecl
fronr rotor inertia forces'
are locatecl by dowel.s ancl clamped together by 2l bolts, spaced witlr however, clenrands
regard to gas loads and located within the area of the coolant passages A rotor constructcd of a lightrveight rnetal alloy,
alloys fail from overheating at a
ancl hollow clowcls. T'hc next step was a water-coolecl housing with a adequate and efficient cooling; sucli
such as cast-iron or steel'
very thin inner u,all ancl non-axially positioned, passärge-forming ribs considerably ]ower temperature than nratcrials
In the IRC-6, oil was tld into the rotor shaft oil intake uncler pressure'
linking the inner and outcr walls of the housirrg. (If the ribs run axially of the oil going insicle
they tencl to produce chatter marks on the workirrg surface as each apex The flow separated at the rotor bearing-most

I of trre Rc z-g0 irnprovecl nronth by month,

ligh-spcccr 1>crfornrance
Evolution «rf RC2-90 pcrfornrancc cluring 1966, at a constant 5,000
althoug5 tnrgct ontl>ut wrs ncvcr achievccl. 'I'lte curvcs sllow otrtput
r.p.rn. using IP 5 fucl. fuel'
rclatecl to f*el flot' at I conrt,,rii O,OOO r'p'ru', trsing JP
'
BsFc. ug/xP-xn
520
520 / JUNE
/ te66
lltlTlAL
./
TARGET J'2o

200 (EOUAL TO SCALEO

--- I
2CO
RCI-60 PERFORHANCE) /

210
2.O

?oo "oo
3
tr, d
l! 3
r60
I Iu
I
t^l i
o c
tr
o
)
E t?o
I 20 u
L, I
v I
c
tr o
o ro

ao

o I
I

-ao -ao
160
)

r2O eoo
t20 t60 ?oo ?.o 20o.
- LB/Hß
FUEL FLOW
FUEL FLOIV - LB/HR
 201
M.y 26, 1964 M. BENTELE ETAL 31134,537 CURTISS-WRIGHT'
Instead of a big
to cool the rotor while the rest lubricatecl the bearing.
COOLING STRUCTI'RE F'OR ROTARY UECHANISIIs

Fllcd Feb. 15, 2 Sheets-Sheet 2 separate chambers, each back-

hollow inside the rotor, there were three
1962

ing up the rotor face cavities which were subiect to the highest tem-

perature inPuts.

A mechanic adiusts the throttle control of the 4 RC-60 on the test bench'

I-t
t---ra-

l-J
I
'1 r-l
--l
-
b

-t.J

ATTtrRNEY

Diagonal fins for cooling the rotor housing were patented by Bentele
and fones in 1964.
 703
CUIlTISS.WIIIGHT
of the rotor. There also were minor co,tributory
problerns associated

with the machining of cast-iron, after having usecl alunrinum.

The next step was to clesign a cast-iron rotor u'ith iet
irnpingement
the thermal gradients, and also
cooling. That *r, calculatecl to nrirr imize
clistortion
to give astructure which coulcl better rcsist thc
heat threats
ribbing of the rotor
that remaincd. Sonre moclifications irr the intcnral
differclrccs bctween thc theoreti-
were macle aftcr tests hacl shown son'lc
cal oil flow, as calculated on a basis of the accelerative forces in the
a high-speed canrera' The
rotor, and the actual oil flow as observed by
difference was founcl to bc clue to inertia
in the oil lnass' Tlte end re-
l. l0r0l TIYUIItll CUT IIIIXCT sult was the I-beam rotor of 1964 '
WIIGHI rervorked and carried
'. By that time, the IRc-6 had br.,, thoroughly
2. t?ltl Slrltorllr GIrl lUlm rr lx tlllltG 10. lxllll rlxt;01D construction of the rotor was
the designation RCI-60. The internal
t. t0I0t [0u5txG ll. tI[AulI tlxtr0lD designecl to ensure complete circtrlation
of the oil in an entire pass
l. TIHIUSI ?OII t2. rcctslorv Gilt !0I ltou$xc tSrough all cavities prior to eiection into the scupPer in the lateral
5. ltDI H0uilr0 - Dllul $DI It. tGxtIt0x c0xIrcI rlrltt
6. llDI lt oustxG - l xlt-Dilul $Dt H.0t1 PtISSUrt ?Ur?l
of the 4 RC-60 engine'
l. mnrt PonI(Dul I txrrrtl t5. 0t1 scrulxct tul? - ,toxl Test results showing the performance
t. rltx !trlrxG tö. 0t1 tcrulxGt ?ut? . llll

Elcvation and end view of the 1 RC-6 engine.

\
{00 ,/
The first IRC-6 was tested in March of 1959. Extensive endurance ,/
v
r
tests at continuous high power resultecl in a gradual pounding of the -E 300
apex seal slot irr the light alloy rotor. Naturalll,, this belhnouth distor-
a
F

tion of the sealing surface caused a power loss. Bentele concluded tfiat =
A-
200
/
F

if they were to continue development work on light alloy rotors, an o

J

irtsert of higher wear resistance than the basic aluminum was needed. 100 //
This initiatecl the developnrent of the cast-iron rotor.
A noclular cast-iron rotor was designecl and used experimentally. It 0 rrl
pcrfornrecl satisfactorily, under less stringent conditions, with no internal IO ROUE E
f00 J
rotor cooling at all. A new cast-iron rotor was clesigned for splasfi
It
a
iil- tr a
t

rr
o F
a i E
t
coolirg with scavenge oil. This was a considerably less complex ancl a
a §
t,-
t
300 t{

costly systcnr than that which had been regarclecl as necessary for tfie gl =
C'
aluminunl rotor. Going to a cast-iron rotor solved the apex seal slot - I
.l
200
e
C'
F
-E
G
a
wear llut introclttcecl other problenrs. One of them was inconsistent ald L

T
unpreclictablc peak Power output. Thc cngineers went back to rework- E .ö
ing their stress ancl Iteat transfer analysis calculations, made new cooling
rJ

\,
\ \
I
a z.J

tests and temperature nteasurements, ancl came up with the answer tfiat l-
a
.5
a

.ta
thc problent centered on thermal distortions within the elastic range E
a

I
IHOUSTilD IPN
 31052,435
ZO4 DEVELOPMENT Nov. 6, 1962 M. BENTELE
uI.ILrI-UNIT ßOTARY ENGINE
members. Oil motion inside the rotor was ensured by the fact that ac- Fllod NoY. 17, 1959 I Süreta-§liit I
celeration in the rotor changed according to rotor position in a cyclical
pattern. The I-beam cast-iron rotor was so good that tests were macle
rvith higher nrean effective pressures and at higher rotational speeds.
Tltis brought on a lrew series of problerrs, such as molybclenum cle- §
n
,(ll
terioration and increased rate of seal wear. \:
§N cr
The k.y problem was not gas leakage, but seal tip wear. Naturally, Gr
\a \D

\/
\t
such wear was detrimental to effective sealing; therefore parallel develop-
\a
nrettt in both areas was begun. TIte emphasis, however, was placecl on
durability rather than efficiency. Apex seal end Ieakage was most signi-
ficant at low speeds, they found, because of rotor width variations ac-
cording to operating tentperatures. New self-adiusting apex seals were
\t
developed, giving a considerable reduction in minimum starting speed,
)
and also permitting low-speed idling without excessive flywheel inertia. §
s
\s
\5
Rare cast-iron housings soon gave way to cast-iron housings with a
soft nitriding on the working surface. Nitriding gave an improvement in 6
seal tip wear, but warped the housing surface and prevented proper seal- l'
irg under high mean effective pressure conditions. Molybclenum spray
on the working surface solved the warping problenr by permitting a
grinding oPeration on the surface af ter spraying. Th is feature, copl- »
binccl with a switch to an aluminum housing, brought development to
a new peak in 1960. N
s
Torqrre is procluced far rnore steadily in a four-rotor engine than in
a two'rotor unit. The curve for the single-rotor has a ströng negative
segmen t.

\o
§N §
r000 §l§I

§
3000

rrl
E e nlroR
2000 MAX EENTELE
*fu\ßrr,*'
J
I
l-
g.

I ./
ga

o
J
ATTORNEY
r
o
)-
1000
2r-Ff

Patent drawing for the folr-rotor Curtiss-Wright \\'rrnkel engitre, shon'ing the main-
shaft arrangement and rotor phasing.
tl , rrrr iorout
-10 00
CRATT I}IGI[. DtGRTtS
Nov. 6, 196e M. BENTELE lNov. 6, 1962 M. BENTELE
UrLTI-LrltIT nm^ax EtcIlIE TT,LTI{,rI,T BOTAST EITII|E

fllcd llry. tl, 1959 ftlod llot. 17, f959

ilI
t

fl

{' § r',+

il t

[n il.'y'ENTO8
HAX AT,NT E,LE
I}.,I/E}JTOß
gv MAX EENTELE,
I

/fu,^-ßJ*' l
I
I

I
I
I
I
l
I
I

Mainshaft for the four.rotor engine.

'
Cross-section of the fotrr-rotor engirte'
a
I

I
l
i
I
 208 DEVELOPMENT CURTISS-\\/RIGHT 209
(tteus rN oorrEo LrNEs Nor FURNTsHEo wrTH excrrre) patented in 1965 by Max Rentele and William T. Derner arrd was
based orr the fact that just as the orbital nrovernent of the rotor subjects
r
I
FUEL SUPPLY
---
L_
I
==-i I
-1
I
the oil seal to variations in centrifugal force cluring the course of every
rotor revolution, the oil seal is also subjected to temperature variations
TANK I

as the rotor completes its orbit. These factors tend to cause distortions
I I
I
I
I I
I

of the oil seals on the rotor sidcs. Adding a nunrber of seals of the
I
I
I I
D.E A.0.8.
I
EANK EANK
-F-
I

L -'l I I
tf r"' same type arouncl the eccentric bearing does not solve thc 1>roblern
rl/ \I,'
I I
I

I
I
\ll because oil which leaks past one seal rvill build up a pressure head be-
POSITIVE I
.-- FUEL IN.'ECTION
tween the seals-causing the eventu:rl breaktlrrough of the oil irrto thc
I
INLET HEADD1 !
' NozzLEs
chamber. The new type oil seal lvas designecl to maintain inrprovecl
I
( 2OOO PSI OPENING
L_+__ PRESSURE }
sealing contact with sidewalls aud encl covers.
FUEL
INJECTION
The invention consisted of using two seal rings-one inner and one
PUMP outer. Both were supported by annular legs, extending inward into the
PUSH-PULL CABLE
rotor, axially flexible to ensure sealing effectiveness. The annular legs
FUEL
FILTER AUTOMATIC
-..<F
.,. FoETonffoTr--rilE
-lD. of the outer seal were rnore flexible than the legs of the inner seal;
FUEL FLOW
AOVANCE
MECHANISI'
therefore the outer seal was better able to take up clearance ancl con-
MIN. MAX.
-4^urn FROM ENGTNE GOVE R NOR

AT II3 ENGINE SPEED

Schematic of the fuel iniection system developed for the RC2-60 ,r:.
Estimated fuel consunqrtion of the RC2-60 Ul0 engine througltout its speed range.

Curtiss-Wright then adopted chromium plating of the working sur-

6Oc F, ?9.92"
face in the alurninum housing. This chrome sürface, with cast-iron
Hg.

alloy aPex permitted long-duration cyclic tests with averuge wear

1eals,
rates of 0.0015 inch per I00 hours for the apex seals and leis than
0.0005 inch on the working surface. The first iyp. of seal used for oil
control at the eccentric rotor bearing was a cantilevered sheet metal t
c
scraPer. This tyPe seal was difficult to make and difficult to maintain. I
o-
T
Worse still, the sealing surface would not remain perfectly level dur-
3
ing minute axial displacements of tlre rotor. As reiult, th; seal failed I

^
to give consistent oil control. NSU hacl made better progress and was = .90
9
using a pivoted Belleville washer type cliaphragm ,.r1. Curtiss-Wright F

adopted it, but found that, although it *ri a definite improvement,

.:)
o-

it UI
z .80
still had some of the same shortcomings as the cantilevered seal. It o
()
also had an additional problem: materiali with suitable
wear character- J
istics proved to lack the necessary strength to withstand the high tr.I
hoop D
IL .70
stresses at tlte sealing edge. Conversely, Laterials that could handle
the (J
stress Ievels at the edge did not have adequate durability. Curtiss- IL
6
Wright found an effective and relatively iimple solution, evolved l!
o-
.60
through closer definition of the seal angle at th; scraping edge, closer o
Iirnits for flatness and surface finish between the seal än0 the housing,
and more precise definition of the optimum unit loading. It was .50

80 roo r 20 t40 r80 200

HOR SEPOW E R
 CUI\TISS-WRIGHT ZII
210 DEVELOPMENT
form to the end cover wall. The inner seal had the iob of presenting a
barrier to the main body of oil inside the rotor, although it was work-
ing rvitlr oil on both sides. Included was a drain gutter of annular shapc
between the trvo seals, into which any oil that leaked past the first seal
u,oulcl flo\r,. This florv was forced by the rotation of the rotor, which
actecl as a rotary pump. This invention has never come into use, how- perfornrance of the RC2-60 Ul0 uncler wide opelr throttle (full load).
ever, and we can assume that the rotor face cavity performs a certain
percentage of the sanre rvork. The question of compatibility between
quenclr areas ancl exhaust emissions seems to be the crucial point here.
- 60. UIO HEAVY FUEL
RC 2
While NSU remainecl faithful to the peripheral intake port, Curtiss- ESTTMATED i,rAX IMUM LOAD PERFORMANCE
Wright designed the IRC-6 with a side intake port. Bentele and |ones 600F, 29.92tt Hg
werc well aware that side ports reduce overlap between exhaust and
200
intake phases, and they chose side ports because they permit earlier
intake closing u,ithout reduced mixture filling. A power peak increase
of Z0% was possible by switching from side to peripheral intake ports, 180
F
IL
Iones aclmitted in November, 1966. This is analogous to changing valve
overlap fronr c.ssentially z,ero to over 100 reciprocating-engine degrees. 160 @
J
Periphcral intake ports also increase the power loss due to exhaust I
back pressnre, in similar measure to a high-overlap piston engine. There 140 lrl
200 3
o
E
Schematic of the oil circulation system of the RC2-60 Ul0 engine. t80 120 o
l-

160
!oo
FUEL INJ.
PUi,P
orL
TE MP
t70.F t 5. t40
-.> ENGINE G,
l^l
,
o
120
o.
BREATHER
lrt
FILTER
a loo
E
o
I, 80

TANK
2 GAL. RI

STRAINER
SCAVENGE
PUMP

I. INLET LINE CAPACITY IN ACCORDANCE WITH OIL FLOW CURVE.

Z OUTLET LINE MUST HANDLE TOTAL OIL FLOW PLUS ENTRAINED AIR OF 2 CU.FT. /MIN.
MAXIMUM.
5.OIL THERIIOSTAT MUST CONTROL OL IN TEHPERATURE TO !7OOFt5' AND LIMIT BACK
PRESSURE ON SCAVENGE PUIIP TO 50 PSI MAXIMUM. looo 000
2000 3
4. OIL LEVEL IN EXTERNAL SYSTEH MUST BE AT LEAST C' BEI-OW SHAFT CENTERLINE
UNLESS VALVING !S PROVIOEO TO ASSURE AGAINST DRAINBACK THROUGH PUMPS ON
S H UT DOWN.
CRANKSHAFT SPEED - RPM
 CUIT'IISS-WITIGI_IT 213

reliability. Some plug fouling then was encountered. An auxiliary air

gap at the spark plug reducccl the fouling, but the periods between plug
changes due to fouling ancl gap erosion still wcre rrot acccptable. A ca-
pacitive discharge systcnr clcsignccl by Curtiss-Wright producccl a voltage
rise rate four times faster tharr that of the starrclarcl autonrcltive systcm.
This system reducecl gap crosion and sensitivity to plug fouling, and
plug change intervals were incrcased four to ten tinres.
Curtiss-Wriglrt's intcrcst irr uutomotive lpplicutious grcw in 1962.
Iiurthcr rcscarch on thc Waukcl powcr unit was dircctccl towitrrls low-
spced operating characteristics, ancl progranrs of part-throttlc and variccl-
throttle operation were instituted. The high-speed potential of the RCI-
60 was investigated in 1963. This test engine had Iarge peripheral ports
for botlr intake ancl cxhaust, ancl characteristically high phase overlap.
It developecl a peak power of I 54 horscpower at 7,000 r.p.rn.; torque was
alrovc 120 foot pouncls fronr 4,000 to 6,600 r.p.ru.
Irollowirrg the redesign of thc llcl-60 to iurl>rove low-spccd perf«rrnr-
ance, Charles |ones directed the design of a twin-rotor unit incorporating
all the improvements. The designing of the RC2-60 U5, the powerplant
that was to be usecl in autonrotivc applicatiorts, was startc«l in |anuary
of 1963, and it first ran on the tcst bench in Sc1>tcurbcr of that ycilr.

This graph shorvs nozz,le florv rate in relation to pump shaft rotation. The sketch
on the right shou,s how the tcst was accortrplishcd.

DI,RATION ff |NJECTTO]ü
USING STROEOTAC DL70s2t2 ilozzLE
S/N I.O EI PUHP O@O RPI
DELIVERY VALVE - SIilGLE 6EAL
The RC2.60 Ul0 heavy-fuel engine. omnrl RETRACTIOI{
Lt NE LENGTX ( lNCt{ESl 20 l.D.( r{CxES) .120
Fu)w trJ umslsrnoxE AT Sooo RPf,
exPeriments have been made in fapan and will be dealt with in a later
chapter. lrJ
lri Cltrrtl tttlv
The surface wear on end covers and center separations was a fairly E
(, llrttcrlc ot
2rCr-u'
tr,
easy problem to solve. They tried molybdenum spray on the surfaces in o
;i.
1959 and found that it worked-so well that they continued to use it for I
I
all subsequent engines. For moderate loads and speeds, Curtiss-Wright lrl
F , I)
3
E
saicl, uncoated iron and aluminum parts could be used. When the I
IL IL

o J
IRC-6 engine was first tested in 1959 it developed 100 horsepower at )o

TT
F
o
F
F
o
F
J'
o
lL
5,500 r.p.m. t! E
b o ö
Initial engine tests used standard automotive ignition systems, wit6 .t l! at
spark plugs of normal heat range. With an increase in power output U) o
rr)
n
t\
N

ancl rotational speeds, the resulting temperature increase made it nec-

essary to tlse spark plugs having very cold characteristics to ensure plug 60 80 loo 120 l/to 160 f80 200 ??o 2{O 260 e00 500 520 5{O
PUIP ROTATION OEGREES AFTER START Of I}IJECTION
 7T4 DEVELOPMENT CURTISS.WRIGIIT' 2I'
This was the version that was two years later installed in the Ford Mus- 43 r.p.m. to 445 r.p.m. in 0.15 second. Sustairrccl starts were obtainecl
tang that is clescribed in Chapter 9 of the next section. dorvn to 60 r.p.m. s]raft speed, and even below. On clynanlonleter tests,
TIrc durability of the first RC2-60 had reachecl, in 1962, the point the oil consumption of the RC2-60 U 5 was a mere 0.09 pouncls per hour,
where a 1,500 hour cyclic endurance test was run without adding any oil on test cycles simulating high-spcccl road-loacl conclitions. (This corre-
to thc fuel and using regular pump gasoline. Some 40% of this test was sponds roughly to 1,000 miles per quart.)
rtrn at the maximunr rated engine speed of 5,000 r.p.m.; Zr% of the time As of |une, 196+, the RC2-60 U5 had completed 2,000 hours of bench
tlrc cnginc was run at wide open throttle (WOT). The engine recorded a testing. Performance and reliability were regardccl as favorable. No in-
nrean effective pressure of Il7-l 30 psi for 600 of the 1,500 hours, and herent problems carne to light, and no new problerns were encoulttered.
ovcr 100 psi for the remaining 900 hours. But many of the olcl problcrtrs pcrsistccl: scal tip wcllr, chatter ntarks,
Cranking speed investigations led to an interesting discovery-the fan- and cold starting. The overall sealirrg problern of the Wankel engine,
tastic acceleration from first fire to idle speed. The engine went from fones recogn ized, did not end with the clevelopment
of effective rotor
sealing element configurations arrd the selection of uraterials providing a
This ch;trt shorvs horv fucl florv variecl rccording to engine speed and compatible rotor to housing interface. It rvas eclually inrportant, he felt,
canr rotation. to achieve adequate rotor and housing cooling lry sirnplc engirreering
lrJ 2Uo
means, on the basis that effective cooliug \vils neccssary to pr$'ent clis-
Y
o NOZZLE OPENING PRESS. 2OOO PSIG tortion of the vital parts and their seals. Leakage past thc sicle seals gen-
G. LENGTH OF LINE = 2I3IN.X.060IN. DIA,
ts
a PUMP INLET PRESS. r 55 PSIG
tBO
NOZZLE NO. r OL 7OS 212
RC2.(r0 tll0 rvoul«l ruu ou u vuricty of fuels, llut bcst restrlts
E
trJ
o- OELIVERY VALVE r PVE 6TS?2,6mm-Omml T[c
o RAUK SETTING r .362-REF. '.625 wcrc obtrrincd u'ith JP 5 ic't fucl (:rftcr tlrc: crrgittc hacl l:ecn modifiecl
tr
lrJ
t60 FUEL r FUSUS OIL r sPGRr.806@ ge.r t«l run u'itlr tlr:rt Irrt'l).'l'llc irrjt'r'trlr lrrttl sllrtrk lllrrg t'ttttligttt':ttitltt is
F slton'tt in thc ttlll skctch.
lrJ

=
3 t40 352 RACX

t
. SETTING NJECTOR SPARK
!
-() \
PLUG

#
6 .20 PvtP tLtrtxr scxfrlYlc
3
o SECONDARY INJECTIOT'I
I

o
=
J
toO
lr.
J
Lr.l
f,
t! 80 E
5OOO rpm (EXCEPT AS NOTED)
I 85: I CR
o ilcr
o.
E RCt-60
o
pORI CLOSUn[ SPILL
3
?o i o.90
I
(L
ci
II
40 =
f,
a o.80
z
PREM I UM
GASOL INE
z o
()
o
tr DIESEL NO. 2
60 td
F f o.70
o - - -SPILL PIOI NT L
E,
9
- 80 5
(J JP4 3000
C) LrJ
(L o.60 rpm JP5
a
too I poRT cLosuRE
( I MPROVED)

- trJ
Y
E.
o.50
40 60 80 too
(D
t?o ?o t20 140
tooo 2000 5000 4000 äooo BRAKE MEAN EFFECTIVE PRESSURE, pSi
PUUP SPEEO . RPII
 + .aooo RPr oPEI, BTOC cot{Ftc. zr7
l.o - +- oPEN, ATOC cot{Flc. CURTISS-WIIIGIIT'
.-... PRE CHAIBER ATOC CONFIG.
C PRE CHAXSER TDC CONFIO.
Housing wear was found to occur in a characteristic wave pattern.
This wear was severe enough, in early engines, to reduce aPex seal life
.9 \i

.0
as well as sealing effectiveness. Wear of the working surface was not
I SrC regarded as a bariier to effective operation. ]orres felt that the nature and
.7 päprgation of such wear must remain a subiect for continued study.
buiirg 1965, |ones came up with a new type of coating for the working
'6 surfacä. It *.r a Wolfram-carbide alloy layer which gave an excellent
hard-rvearing, but very costly, surface. Several thousand hours of testing
30 60 70 to 90 loo showed virtüally no wear on the surface, even thouglt 307" of the test
IIEP cycle was run ai wide open throttle between 3,600 and 5,000 r.P.m. After
SPARK S:28 hours on a r.u.r.-duty cycle the engine showed only 0.0000421
PLUG
incS of wear on the working surface. This wear rate is iust over Z%
of the wear recorded in a 1,500 hour test two ycars earlier. Wear of thc
INJECTOR working surface had been eliminated-
During 1965, new rotor oil seals also were developed. Tests with the
new seali showed leakage rates down to a few hunclredths of a pound
___+-_ per hour under roacl-load operation, and only slightly rnore at maximum
cox f r o utlTt o x 'l' ipeed. This was followed by a program to determine the minimunl rate
OPET{ COHBUSTIOTI CHAXBER
BTDC COHFIOURATIO}I
ATDC PRE. CHATBER COilFIOURATIOI{
ilot{- gwt R L- tloil-s c AYE ilG E oi .oltrolled leakage necessary to maintain acceptable \I'ear rates on the
SPARK eccentric seals and rotor side seals.
INJECTOR
PLUG

Rotor spray pattern inclicates changes in intensity and direction at

SPARX various points during the conrbustion process.
PLUG

coxFtoußlTtor "!'
-.-+.- ao*lr,nrrrTrox 'o"
_t _
oPEt{ COXBUST|Oil CHATIER TOC PRE. CHATEER CONFIOURATIOiI
ATDC CONFIOURATIOX NOil-SWtRL - nOil-3ClYEN6E

Four different fuel inicctor ancl spark plug configurations were

The curves in the graph above show that configuration "A,"
tcstecl. .100 AToc

u,ith its oy>cn combustion chamber,injector arrd spark plug slightly '
aheacl of the nrinor axis, give highest nrean effective pressures with
zOE ATDC frt
lou,cst specific fucl consumption, throughout the useful speed range. T0c
zOE BTDC

.to.
erally was small, and in the RCZ-60 U5, gases that escaped this way \l,ere BTDC

60'BTDC
automatically vented to the side intake ports. Leakage past the trailing

A
apex seal was considerecl sinrilarly unirnportant, as urrburnecl gases
passed into the chanrber behind to be burned on the next cycle. Un-
burned gases that passed the leading apex seal, however, went dir ectly ^
into the exhaust. Development work therelore was concentrated on
inrprovemcnts in lcading seal efficiency uttclcr prcssure build-up and
reversal.
^
 \

'-{:
) ic.r,
.a
H ta,ae
a *aa - .U'4., t--
a r<aa, ' 4a Jflaa,
t.*.aac ütu --4.
n.-rtt'atu -a),
l:.Da

' \ -E;--
, \-. r,.-
J
u-
} E
NO!i{E WEIOXT UE

THE CURTISS-WRIGHT MOOEL RCE-60.UIO HEAVY FUEL ROTATINO COTBUSTIOI{

Et{olNE wEloHT, lN THE CONFIoURAT|OI{ OEptcrEo ABovEf 8UT W|T}|OUT THE ENOINE INCLINATION LIMITS
FLYWHEELI IS 325 LBS. THE WEIOHT OF THE FLYWHEEL SHOWN IS 5O.2 L8S.
HAXIMUil TILT FORE AftlO AFT- laolo (2901
HAXIMUU ROLL - aLo/o l?tol
Installation diagrarn for thc RC2-60 UI0.
Installation diagram for the RC2-60 U t 0.
The conclusions drawn from their original 1963 experiments with tfie design, research and experimental work in a number of other directions
single-rotor RCI-60 led Curtiss-Wright to conclude that Wankel engine as multi-rotor engines, air-cooling, and multi-fuel operation. After
-such
design features had to be tailored to the application.They later reversecl the evaluation of the IRC-6 test results, two new engines had gone on
this opinion, but before they got to that point they had conducted the drawing board. One, a single-rotor unit, was eventually developed
 i-
Laaa -I-
rct D

axü 4?
R,,..,JA TET

.,trl )rA or
Je *, dtlL''''t
dro ac*u '

il
,a- Ü ..tr-
t-ia * r
-r

-tfl 'F.Hm,n,,al
J---ܧItE,
ar--daD
. J, G.,-,Y
,-E
-Ft
JÜJAI'

ENOIT{E WEIOHT
a
THE CURTISS-WNIOKT XODEL RGz-CO-UIO HEAYY FUEL ROTATII{O COTBU§TIO}I
.1
EN0|NE WEtoHTr til THE CoNFtoURATtOit oEPTCTEo AIOVE, tUT WTTHOUT THE
a

FLYWHEEL,19 !29 Lts. YHE UEIOHT OF TXE FLnTHEEL sHOUil tg CO.z LBg.

.nu,nlTlnor,o, umrrs
Installation diagram for the RC2.60 Ul0.
HAxttuM TILT FoRE Allo AFT -og% (e9el
)rAxtuux RoLL -szoh(elol

Installation diagram for the RC2-60 U l0'

into the RCI-60, while the other, a twin-rotor unit, was used to evaluate There are only two ways to put more than three rotors on a rotor
the basic twin-rotor configuration. With both engines, chamber climen- shaft in a Wankäl engine. Either the bearings and reaction gears must be
sions were identical and shaft and bearing arrangements were simplified. split or the shaft must be a built-up assembly. Bentele decided to use
The work on these engines encouraged Bentele to investigate the possi- ; integral one-piece mainshaft, which meant that the bearings and
bilities of a four-rotor engine. stationaiy gears had to be split along the plane of the engine axis. (A
 April 19, 1966 M. EENTELE
r,24A,030
April 19, 1966 M. BENTELE 3,246,630
NOTARY COTDU§TTON ENOINE AND TETHOD
OF OPERATIRO SAIE
notARY colBusrtor EHOIIIE AND IIETHOD OF OFENATI}IO SAIE
0rl3tnal Fllcd IaY t, 1962 ? §hcctr-Shmt t
Orlglnal Fllod llay 1, 1962 7 §hcctrSfii3t a

.a_
\\ IFV:
\ P\
7?--\

tl \--
'l

Itrqryl-r
I9!!!IEI-c
4A'A.TE.
4A'Ar.E..

[ 'J. lJ -1E
d-

br

l-) -l
MAx%äffTeur
I-r'1
t-H
F1 MAx%EffirLE EY
l-J
%Z/."r" % //€-
-H EY

%Z)*z /A- HI5 ATTtrRNEY

- I.II3 ATTtrRNEY
rvith ntaxitnum
Iniection sl)rily 1::tttcrn prol>osccl for a rotor cavity having maximum Ipjcctiop sl)ril), prtttcrrl l)Ioposctl. f«lr :l rotor cavity
dcptlr in tlrc Ic;rrlirrg portion of the conrbustiorr chamber. clc,th irr ttrc ii,rili,g 1,,,riiori .f tlrc co,rbusti.rr clra,rbcr.
 226 DEVELoPMENT
split-rotor
shaft would introcluce structural weakness ancl make e,gine
assembly difficult.) The plan was to use a one-piece
rotor shaft with four
eccentrics' Bentel atent for a four-rotor Wankel engine
on November 6, shaft was designed Iike a one-plane
crankshaft for a gine, with eccentrics #?. and
.

clianretrically opP #4
novel form of split bearing.
#t ancl #). Bentele also devised a
The engine had an integral gear-ancl-bearing unit, so that
each gear
half also carried a beating half. Tr9 complete rotoi shaft and
beari,g
alsembly was held in place within the housing by bolts,
ancl t6e rotor
shaft bearings were split into two semi-circularlriu.,
on the line of the
Iowest bearing loads. The ioint faces between the bearing
Salves Sacl
interlifting ribs for accuracy of alignment. Each rotor had
its own reac-
tion gear, and the four gears in the four-rotor engine also were
split into
two semi-circular pieces. The sidewalls \r,ere däsigned with
big holes
near their centers to allow the engine to be assenrblecl
by threading the
rotor shaft through a succession of separating walls and
working cSaur-
ber housings.
The four-rotor engine was assemblecl by nrating the mainsSaft
ancl the
front end cover with its accessory drives. tlr.,r, one rotor
ancl rotor Sous- *
t
The wankel-powered Bertram boat at I
speecl.
i
I a,
i
Installation of the RC2-60 M4 in a Bertram 22 boat.

ing followed, the rotor fitting onto its bearing on the first eccentric.
After onb separating wall, with its port, the next bearing ancl gear set
follorved. They were piloted in the side separating rvall, and the split
plane was serrated as a further precaution against dislocation. Ttris
process was repeated until the final end cover ancl flyrvheel assenrbly
could be bolted in place. The eccentric arrangenrcnt proviclccl po\\rcr
strokes at 90 clegree intervals, with a firirrg order of l-4-2-3.'l'here wAS,
however, a small unbalanced couple which nraclc it neccssary to add a
balance weight at both ends of the shaft.
This four-rotor engine had a nraximunl output of 425 horsepou,cr at
6,500 r.p.m. and showed fairly flat torque arrd fuel consumption curves.
It was the first multi-rotor Wankel engine clesignccl and built anywhere.
But the |apanese were not far behincl, and werc soou to surpilss the
achievenrents of Curtiss-Wright in severitl areas.
Automobile engineers used to argue among thernselves about how
many cylinders a 200 horsepower engine shoulcl have, clue to the fact
that breathing and torque characteristics, the conrbustion process itself,
7ZB DEVELOpMENT CURTISS-WRIGHT 729

engine balance and transmission requirements change greatly accordirrg it was too early to say, in ]anuary, 1961, that this was in fact the case.
to the number of cylinders. An engine having too many cylinders tends The preliminary tests gave readings of 872 horsepower at 1,525 r.P.m.,
to be excessively complex and often runs with extremely high friction with a brake mean effective pressure as low as 106 Psi.
Iosses, while to«r few cylinders leads to enormous bore sizes ancl ineffi- In a continued study of scalc cffects, Bentele went to the opposite
cient conrbustion because the flame front fails to reach the far corners extreme and designed the smallcst Wankel engine yet-the RCI-43.
in time to produce power. AII these arguments were applied to t5e Design studies indicated that a basic single-unit, die-cast aluminum
Wankel engine during its design. engine, rated at l0 horsepower for high speed, could be provided in the
In one experintent, the linear dimensions of the RC-6 epgine were Z5-pound class and within an envelope of .66 cubic feet. This package
Iiterally scaled up about 30 times, to give a displacement of \qZO cubic
inches. Adiustments were made on dimensions and features which af-
fected stress, heat transfer and thermal gradients, but the basic design Side uew of the RC2-60 M4.
remained the same. With its dual side ports, the aspiration capacity of
the RC-6 proved more than ample. The preliminary tests showed some
of the Power potential in this engine. However, it suffered from one
phenomenon that had not occurred in the smaller Rc-6-detonation.
This happened under highJoad conditions, iust as in piston engines.
It seems safe to conclude that the size of the cornbustion chamber
provided opportunities for spontaneous ignition or surface ignition, be-
cause flame front velocity failed to increase on the same scale as the
physical dimensions of the combustion charnber. Bentele insisted that

Ftrel ggn_sqmption related to engine speed and vessel speed, for the
RC2-60 M4 installation in the Bertram boat.

t5
t2
RC2-60 tN t7 FooT BOAT
lr e4.q5 POUNO GROSS weroxr 5000
t4tt. x 20-3 at-. pnöC'
ro I.76 : I INBOARD - OUTEOARO DRIVE RATIO
6.03 SECONOS ACCELERATION
E FRou o To t so reCf
4500
o
a
tr
lrJ
o,
vt 4000
z
o
J
(
J
I
RPM
FUEL FLOW
I 5500
o
J
fL
J
l!
f
tr
3000

2500

t5 20 ?5 30 35
BOAT SPEEo, MTLES PER HOUR
 231
CURTISS.WRIGHT'
engine configurations were thoroughly defined,
and test rig capability
worked out for a wider scope of äuriurtion.
Test instrumentation in-
cluded thermocouple heat flux measurenrcnts in
both air- and water-
that rotor housing cooling
coole,c engines. These comparisons showecl
test engine was des-
was the biggest problem. Tüe first totally air-cooled
ignated the RCI-60 Jl.
test hours. That
By MarcS of l963,the RCI-60 Jl hacl run about 500
month the engine exceedecl its target power output
of 105 horsePower by
overheating. Success
a margin of l0 horsepower, or almost loTo, rvithout
was achieved mainly by the anall,tical approach
prior to testing, although
a considerable number of significant tethnical
refinements had to be re-
heat inputs
designed and evaluated in oider to match the experimental
theoretically
(which did not conform exactly to those that had been
however, were drawn from this test Pro-
predicted). No final concrusions,
gram.
T5e air-cooled Wankel engirre is attractive for use in
light aircraft, fixed
throttle resPonse, and in-
and rotary wing, because of its acceleration,
at or above the levels of attain-
herent ease of control-all of which are

The Owens Concorde boat headipg for open waters'

Latest version of the RC2-60 IU4. Two of these were installed side by side in an
Owens Concorde boat.

includecl all accessories except the fuel tank. The basic single-unit engine
could, of course, be extended by the addition of one, two or three power
sections providing 20,30 and 40 horsepower engines-each rotor addition
wotrld add some l0 pounds. The RCI-4.3 was air-cooled, which led to
an investigation of the possibility of air-cooling for the larger Wankel
engines.
It was chiefly an interest in air-cooled aircraft engines that led to an
analytical feasibility investigation of air-cooling for Wankel engines,
small and large, early in 1961. The analysis led to the conclusion that
the Wankel engine could be ccoled adequately by air within the power
output requirements for aircraft applications. However, this conclusion
was challengecl because some engineers were worried about the localized
heat flux peaks of the Wankel engine (much higher than in the four-
stroke pistou engine becanse of the frequency of ignition).
A more detailed feasibility study then was started. Ultimate aircraft
 87
232 DEVELOPMENT CUI(f ISS-\\/I{IGI]'T
weight ratio comparable to currcnt turboprop engines. The
RC2-75 led
nrent for a reciprocating engine. hl addition, the Wankel offers low Wankel
directly to t5e clesign of a sornewhat largei air-coolccl, tlin-rotor
inertia for torsional nratching, relatively low cost, low fuel consumption, is de-
engirltlre RCZ-90 YZ.It wrrs clcsignccl to burtr cliescl fuel and
ancl light u,eight. Missi«rn studies in this area gcncrally shorv the engine
plus fuel weight to be less than that for a comparable gas turbine, even scribed later.
though the bare Wankcl engine weight is higher. This conclusion has Thc I1CZ-75 dicl not have thc clisaclvantagcs of the ttrrboprop, strch as
held true for all but short duration flight missions. tSrottle rcspousc lag, lrigh sturter rnotor powcr rc(lLliretltents arrd high
What actually resultcd from the air-cooling experiments was the crea- fucl .or,runrption Ä a bi,rus, it' prcscrrtccl no scvcrc lteecl for sllccializecl
overSaul ancl nraintcnallce pcrsorlrrcl.'l'lrc ttcxt.stcp wtls to clesign
a f«lur-
tion of a definable systenr for rninimization of air pressure drop and irr 1962
rotor, air-coolccl aircraft .,rgi,',.. T'lris crrgirrc \\'rls c()lltPlctccl
artcl
quantity florv. This was achievecl through close matching of air-sicle
requirements to gas-side inputs, based on a desired metal temperature given the designation RC4-60 IZ.
clistribution. ITollowing re-evaluation and re-clefinition of objectives, The RC4-6 O lZ usecl the same t),pe l-bcam rotor that had been devel-
Curtiss-Wright rvent on to de.sign a larger air-cooled aircraft engine, the op.a io, the RCI-60. Rotor cooli,rg rurr by the salne inertia-actuated oil
circulation methocl, ancl seerling clcmcuts \\'erc of sirnilar design. Air
to
RC2-75.
This engine was gasoline fueled, with a propeller recluction gear for cool the Sousing was suppliecl bv a gcar-clrivctr, t$'o-sttrge cast alumirlum
Iiglrt aircraft propulsion, and was rated at 27 5 horscpower. That is equiv- axial conrpres.sor operating at a l.?-:l llrcssttrc ratio artcl clclivering 1.75
for
alertt to about onc horsepower per cubic inch, giving it a power-to- cubic feet of air pci r..o,rJ. Unfortunatcly, thc nrilitary requircnlctlts
which tSis engiire \\,as clesignecl did not nraterialize atrcl it was llevcr
developed.
Initial test results obtained rvith the passenger car version of
RC2.60 U 5.
the
Iironr 196?. onwarcls, thc Preclorninant dircction at Ctrrtiss-Wright was
r{o
the cleyelopment of a basic urrit, aclrrpt:rl>lc ttl n vrtrictv of irpplicrttiotts
ratSer than tfie exploration of various siz,cs utrcl spccializ.ccl t'ersiotrs of thc

t4
BIF?4ra| l-a 6
engirre for specific applicatiorrs. I)o'clo1>nrcttt work ccnterecl on thc
RCz-
reo t
o,
lr.l 60 ancl the yRC-180-2. The RC2-60 is a I85 horsepowcr, licluid-coolccl,
too
= twin-rotor engine cleyeloped for vchicular, marinc, grouncl sullport and
-
@
220 00 otSer nilitary an,l conlnrcrcial applications.
'fltc othcr etrgittc, the air-
I torl*
;
iF

n r00
I
lr,
l_J coolecl ynC-t80-2, being clevelopcd uncler a U.S. Navy contract, rveighs
ZT1oounfls apcl was designed to producc 310 horsepower at 6,000 r.p.m.
Tästin/ with the IICZ-60 has bcen conccntrrrtccl irr the area of dernon-
stratipg the feasibility of various configuratiorrs.'['lte first tests \l/crc
nlacle
f
o 60.E. 29.92 ' Hc.
Comparisorr betu'cen the RC2-60 arrd a V8 piston engine of com-
E, BARE ENGI NE
o t40

r
F paralile perfornlallcc, shou'ing clifferences in fuel consunrption char-
HOR SE / actcristics.

G
too
,§ ui
_l__l---l_
2OOO RPH

I
trJ lrl I
I
o 5-
=G
/ \-- Htct
fr60
(n
.r{
E:.so
3-
to
-T/t L.

- i\
2-6,
GI NI

L
G,
'8o NEi
o ä9.?o
L-
- (/ ldrL .eo
vEt illol
\ \-
?o äa.50 ..1-
8,2
tg
rooo 2000 3000 {ooo 5000
.€
'rc,ä t5
t-t-
20 25 50 {o 50 60 70 80 loo
CRANKSHAFT SPEEO - RPM BRAXE TEAI{ EFFECTIVE PRESSURE- 9rl
 23+ DEVELOPMENT
irr 1966 with a 17 foot Bertram boat,
CUR'I'ISS-\VII.IGI.IT 85
using a, inboard-outboard installa-
tiorr. Next, trvo Rc2-60 engines were
inställecl sicle uv riae i, a, ou,ens
Irr 1969, Curtiss-Wright reached an agreemcnt rvith Lockheed air-
concorcle boat' Rcsults *.Ir. rnost encouraging. craft for clevclopment of the RC2-60 engirre for installation iu thc O-star
tiss-Wright's managentent: "The future fones rcportccl to cur- civilian version of thc U.S. Antry's Q112 rcconuaissance pl:luc.'l'ltc
is pronrising i, tlre ,rarinc -a
Wankel errgine has 85''1, n)orc powcr than thc origirral YO 3, as tlrc
Powerplant areäl bccause of the outl ut smootliness,
high specific outpnt
ancl' as comparecl to two stroke .yät.
engines, improvecl fuel consump-
Army calls an improved version of the Lclckhced plarrc. 'l'hc YO-3 is
tion on the orcler of 20% nrore, freecloir basically a Schweizer, two-place, high-performance sailplane rn,ith a 100
lower noise ]evel.,,
frorn ;
fuel-oil mix, and a
horsepower piston engine aucl slow-moving, six-blaclc propellcr. Installrr-
workirrg in collaboration with westinghouse tion of thc RC2-60 could be acconrplishccl at a cost of only 6% irr wcight.
Aerospacc Iilectrical
Divisiott, cttrtiss-\vriglrt suPl>liccl a-protot.rp.
cngine for an Itc2-60 ge,- Irr ordcr to explorc thc po.ssibilities of l rnilitary ntarket for the Witn-
erator sct of 60 kilowatt_capacity. It kel engine, Curtiss-Wright startccl a progranr to clctcrurinc its nrtrlti-ftrcl
liad , biushlcis 400 cycle AC gcrera-
tor cantileverecl fronr tlre engine aclapter capal>ility. Thc fucl in qucstion was IP5 fncl, n,hich is usccl currclrtly for
housing ancl engine flyu,6eel
rvithout aclclitional bearirtgs. ivesti,rgliouse girs turbine jet planes. A study conrplctccl in l9(rl inclicatecl that the basic
it conrparecl with existin[ units, uring bothrelerrä information on how
gasolinc and cliesel pisto, Wankel engine woulcl operatc rvith spark ignition, ancl without ally
engine's and gas turbin.r. O,, a'wciglit
ba.sis, the RCZ-60 i*l-; sligSt change in compression ratio, on IP fuel. Curtiss-Wright dicl not clcsign
aclvantage over the gas turbine
(90d pouncls against 1,000), wSile t5e or builcl cliesel-typc high-conrprcssiorr cngincs that operatcd u'ithout l

gasolirte piston engiric wcighecl

i,4l0^pouncls *a the clie.sel carriecl spark ignition. Charlcs fones fclt that conrprcs.sion-ignition woulcl re- il

terrific wcight pc,alty.ll a

rluirc a clcpnrturc fronr tlrc basic gcourctry artrl proportious which Ctrr-
l

pouncls. 'l'1,. Itcz-(r0 took .p lc.ss

.t
'1

!,sso
tlrarr tlre gas turbine (37.6 culric s,,.cc
il

j."t against 45). The gasoli,e engirre tiss-Wright lracl clcvcl«;pccl througlr thcir cxl)cricttcc. Sprtrk igrrition tvils ,i
I

clictatcd by thc basic cnginc clcsign, u,hich linritccl conrprcssi«rtt ratio t«r
rl
84 cubic fcct, ancl the bulky clicsel
tt'sccl
idl cp5ic fcct. ,I/

t'ltlrrcs Irclotv tlrosr: n('('cssrtry frn' ('()nll)rcssirlrr iSirtit iott, ,

J
-l
I

illl utlrrlttrttitlrr
.,1

Basic compo'ents of the RC2-60 u5 engine. I'he first vcrsion of thc lreavr,-fucl Warrkcl cnginc wäls ..t

of the RCl-60 with spark ignition, and fucl iniection

rt

water-coolirrg, I
clirectly into the conrbustion chirml>cr. J'his nrcant using clicscl-type lloz- ,,;

zles ancl a high-pressure injection punrp, consc(lucntly thc nrulti-fuel

Wankel engine carriccl the cost penalty of the high-prcssure injection
)

equipulent. I?ucl injcctiou is not the sanre in the heavy-fuel cttgine as in :I

a gasolirywngine. The gasolinc cngirte has an air intake valt,c, rvorking

,,

as a throttle, to control engine spced. hl thc dicscl engine, crarrkshaft

B coo r.p.m. is controlled by the anrount of fucl injcctcd into the colnprcssccl ;

intake air.
Iiuel injection was clictated b1, thc type of fucl, rvltich has a Iow octane
rating and very low volatility. IP fuel is a h1,61ro.urbon in the kerosctre
family, having a specific gravity ancl volatility comparable to paraffin.
When burnecl, it leavcs conrbustion products sin'rilar to those of gasolinc
plus the nitrogen fronr the air; how-
-carbon dioxide and rvater \/apor,
ever, it contains no carbon nlonoxicle. With fucl injection, thc combus-
tion process is quite diffcrent, because the nrixing of the particles of
fuel and air takes place insiclc the combustion chamber. The fuel is
injected in a fine spray to facilitate this mixing. Because the fuel is I

bound to travel in straight lines, äs directed by the nozzle, thc air must
be made to swirl in order to ensure conrpletc vaporization. .

In the gasoline engine, combustion takes place so quickly that it ahnost .

 CURl'ISS.W}\IGTIT 87
creasecl. Instead of a Pressure rise, the result is relatively constant
pressure
with increasing volume. In practice, the pressure rise rate is almost a
matter of chance.
With cliesel engines, the rate at which fuel is iniected into the com-
bustion chamb., ?o.r not necessarily control the rate of burning, nor
can the engine or its injection equipnrent ensure sufficient time for a
degree of äi*ing that lvill guarantee clean ancl efficient combustion.
The moment burning starts in some area of the combustion chamber,
another area may contairr a bocly of partly vaporized fuel with vaPor
envelopes .pproäching the fuel's self-ignition temperature. For this
reason, the fuel throu[hout the combustion chamber may ignite either
because it is reached lV the progressive spread of the flame from the
initial nucleus, or because a bocly of fuel vapor reaches its self-ignition
temperature, or a combination of the two. The diesel engine has three
pl-,aies of combustion: it starts with a delay period, which corresPonds
to about 15 degrees BTC to about 3 or 5 degrees BTC. In this delay
period, ignition is initiated but there is no rneasurable change in Pres-
iur. rise rate. Pressure continues to rise under continued upward piston
travel iust as if no injection had taken place. NIext comes the peak
pressure period, wlrich corresponds to ) or 5 clegrees BTC to approxi-
hately l0 clegrees ATC. During this period, tltere is a sharp Pressure

The world's first \Ä'ankel-porvered airplane first flew in 1969. Built by Lockheed, it
In 1966 Curtiss-Wright installed an RC2-60 U5 engine in this Reo Zy2 ton 6 x 6 is powered by Curtiss-Wright's 185 horsepower RC2'60 engine.
military truck.

can be said to be completed at constant volume. That is, the combus-

tion chamber displacement changes relatively little during the combus-
tion Process-tlte burning of the fuel nrixture is over in a flash. The
result is an extremely high pressure rise rate, followed by an almost
equally fast drop in pressure as cylinder displacement is rapidly increasecl
during the power phase. In the diesel engine, there is no mixture in t5e
first place; there is no spark and no flame front. No such combustion
phenomena as pre-ignition, knock or rumble exist in a diesel engine.
Fuel iniection determines the start of combustion. A fine spray of fuel
is iniected into the compressed air, and atomization occurs sirnul-
taneously in several areas of the combustion chamber. However, it takes
a while for all the fuel to be iniected and atornizecl, ancl consequently
the combustion Process must be stretchecl out in time. fu opposecl to
the constant volunre contbustion of gasoline engines, the aieiel engine
gives what is often called constant pressure combustion. Pressures rise
steadily during initial expansion; but displacement is also steadily in-
 B8 DEVELoPMENT
rise from about 450 to over 700 psi. This period also represents a pSase
of rapid ignition and combustion of the whole of tfie fuel present in
the c1'lindcr. The pressure rise rate, usually some 30 psi per clegree of
crankshaft rotation, is a cletermining factor in causini diäsel kÄck or
rough running. (Diesels are noisier than gasoline.ngi,r., because the
Pressure rise during the power stroke induces transient vibrations in the
engine structure and causes the outer surface to give ofi noise. T6e noise
Ievel is determined by the characteristics of the exciting force, by struc-
ture resPonse to vibrations and by the ability of the engine ,urir.., to
radiate sound.) The rising volume period is the last phase of combus-
tion. It has a cluration from about l0 degrees ATC to about 60 degrees
BBC. Pressure falls off as volume increases, caused by the piston com-
pleting its Power stroke. Most of the fuel is mixecl *itt, air during this
period, injection is completed, and combustion is concluded. '

By running with moderate compression ratios and spark ignition,

Charles |ones felt that the problems usually associaterd wiitr heriy fuels
ancl their slow combustion process could be circumvented. Early devel-
opnrent centered on fincling the best location for the nozzle, the iornr
of
The RC2'60 retains its water'cooled housing- and oil-cooled rotor,
aircraft. This is the power u,it for the rocfih.eä - ----' even when used in Curtiss-Wright proposed a four-rotor air-coolcd Wankel engine for military planes,
Q.strr. and a twin-rotor air-cooled unit for helicopters.

tlre nozzle spray pattern, the location of the spark plug, ignition and
iniection timing. As the results of these tests began to roll in, patterns
started to emerge.
Four different types of combustiorr chamber were testecl. The first
was an "open" combustion chamber rvith the injector installed very
close to thelspark plug angle so as to iniect fuel against the rotation of
the rotor, while the spark plug \l,as angled the other way. The ignition
point was about I0 degrees before the minor axis. The secorrd t),pe was
similar in the iniector/spark plug relationship, but both wcre moved
back to a position about l0 degrees after the minor axis. This engine
demonstratecl capability of operation on both lP4 and IP5 fuels, as well
as on diesel fuel #Z and high-octane gasoline. The third type of com-
bustion chamber was a pre-chamber design rvith non-swirl and non-
scavenge characteristics. Both injector and plug were mounted well after
the minor axis. The fourth combustion chamber was also a non-s\+,irl,
non-scavenge type, with spark plug and iniector advanced to about 6
degrees after the minor axis.
The essential element of all these experinrents was coordinated igni-
tion and injection for burning of the injected fuel at a controlled rate
to avoid unduly high burning rates with low octane fuels and insensi-
 CUR]'ISS-\\/ITIGI I'[' Z4I
tivity to cornbustion lrg witlr Irln,cctanc fucl.s. Tlrc succcss of thc
proccss str«»rgly «lcllcrrclcrrt ulx)n controllcd u,ct'tirtg ittt«l cvltlx)til-
\\/äl.s
tion frorn thc rotor facc, conrlliuccl rvith tlrc transfcr of air pust thc rotor
cavity ancl nrinor axis of thc troclroicl aucl combnstion chrutrber surfacc
tcnrpcrirtures.
'I'he encl rcsult a1>pcarccl to llc a rratural outgrowtlr of tlris cngillc's
funclantcntal gcolllctric charactcristics. It wu:r^ clcsignccl t«l .s\\'ccp rrll «tf
the charge air, in a rcälsonably prcclictablc urrcl collsi.stcttt flrslritln, llitst
the stationary iniector-igniter conrbination. Superinrposed turlluleltce
inclucccl by rotor "squish" action was anothcr clcsign ob jcctive. Irull
aclvantage was takcn clf its rcl:rtivcly loctlizccl lrcrtt input distrilltttiort. In
effect, "repetitive turbulcnce" was built into thc cngine witlrout rcsort-
irg to shrouded intake ports or swirl pockcts. I{etaining the lorv K
curtiss-wright also inst,lrc«l rrn I(c2.60 factor (R/e ratio) of thc RCI-60 prcscrvccl thc aclvantagcs of high swcpt
sonnel Carrier in 1966. It
u5 c,gine i, A, F.V. 4)Z Arnrored per- volume pcr unit of frontal area plus u rcasonallly flcxiblc corrrllustiort
hacl .or.iä'.al,nost-iti,ö'OO miles by August 1969.
chamber shape, as opposecl to the situation u,ith high K valucs requirecl
for compression-ignition operation.
Curtiss-Wright's spark-ignition, heavy-fuel in jection enginc en joyecl
the benefits of unthrottled part-load opcration at low fuel-air ratios with
the attendant savings of: (.) Iorver purnpirrg losses; (b) recluction of

These \r,ere the arguments used by Curtiss-Wright in its rregotiations with the U.S. {

military in ordcr to promote the usc of their Wankel cngincs in the Drone Heli- I

copter.
l

Cost Comparison
,R,C2-90-Y2 VS BOEING T-50BIO AND ALLISON 250-C-14
:

mN tN THE OH-sOD DRONE HEUCOPTER l

RC2 -90- Y2 DESIGN SIMPLICITY RESULTS IN:

I 30.75'h LOTYER UNIT SELLING PRICE
I 3I - 64 % LOWER INVESTM ENT COST
I 26.47'h LOIYER MISSION FUEL CONSUMPTION
I I6'h LOWER (BIO} AXO 4'/O HIGHER (C-!4} ENGINE PLUS
FUEL.WEIGHT
I 83'h LOIYER TRAINING COST
I 6O'h FEIYER BILL OF MATERIAL LINE ITEUS
I g7-9!'h FEWER GRITICAL ALLOYING ELEMEI{TS
I CONVENTIONAL IIANUFAOTURING TECHNIQUES AND EOUIPMENT
I SIMPLIFIED }TAINTENANCE ANO HIGH DEGREE OF RELIABILITY
A Ferret Scout Car was equil>pecl rvith a I INHERENT STABILITY
in 1967 . By Augusr 1969 it i äa1är.ita 9l,r^ttrt.\\/right RC2.60 U5 Wanket engine
16,800 troublc.free miles. I RAPID RESPONSE AND SI}IPLIOITY OF CONTROL

;
äää i lEääEäEäääääIää-ääääääää ää IiäääägEäägä
aä $' i§s
äaää iläu§ää
[ä rääii §ä$I
ä§ä[igääeiä's äE *ääääg
ää $ü äi läiärlä
: r :l aä
ää [ä

ää
g; ääsE ;Fä ä iälääääEääE ä[äääEääääälä ää1äääää E

o $"ü o
Hcr c
1

.j'? ua = v
=F"\JA

GG"J)-
'.r=r-fa
8Pq €
ätic
1AY,:
i-- (-]1;.
rr-{
=

AIR-COOLED FAMILY
ö. ='o 1L
rDGi,.-
^:.E=
-i'tc
IIJ 1rrh i

Llodcl RCr-60 RC2-60 RC4-60 RC2-90 RC3-90 RC4-90 RC5-90 RC6-90 = oä |t)
a'r. J

206 lrz 465 620 77 5 oa

HP (aircraft T/O) r03 310
6.000 6,000
930
6,000 6,000 oot7 -
RPI{ (aircraft T/O 6,000 6,000 6,000 6,000 '/a-\

HP (surface equip.)
)
80 160 )20 z+0 360 480 600 720 rt
,J-'-
^1
-
a-f
',t)
,J

5,000 5,000
:.,i-r t

5,000 5.000 5,000 5,000 5,000 ?{€

v-J

RPil1 (surface equip. 5,000

)
766 417 3r7 +10 510 605 710 tüOct

3E
Dr)' \r'eight-lbs. 19z
r.96 r.z9 l.0t 1.02 0.8s2 0.882 0.781 0.764 Er-F
-1 ct ,,i
Specific \\,cight-lbs. /hP.
s.9 45.4 5I .9
X' ).i
Lcngth-in. 24.4 29.4 )9.4 32.4 (Dö H

zz.0
3

20.7 27 .3 8.3 ?.5.2 27'8.4.Z .^ -t

-
\\ridth-in. 20.7
zr.0 Eöä
a-r F\
20.7
i
-

18.0 18.0 18.0 19.9 zz.0 ,)

r 5.9
1S.
Hcight-in. r 8.0
Volume-cu. ft. 5.26 6.35 8.5 7.0 9.06 rz.2 Z0.z Y-3ü
-)-t'
A.
-) h{

Spccific volume<u. ft./hP. 0.0 52 0.031 0.021 0.02i 0.020 0.020 0.020 0.022 rFt
*.o
§ Ff-r
v
=.
-.r - C-
r)
=v)
FI

EJ
.J
t-. 1

;ö
)
r.
)rH
ai

D=.N +
T 'ö \^,
='oa
241 DEVELOPMENT CUI{TISS.WRIGI_IT 245

The ratirrgs and estimated size and weights for the liquid-coolecl In tests witS this engiue, Curtiss-Wright was able to show a power
which
engines include the distributor, fuel boost pump, fuel injection pump, increase at somewhat ,i.1,., nrixturcs tharr originirlly anticipated,
errgirtes' Be-
starter, oil pulnps (dty sump .system), coolant pump and rnanifolds. were assunrecl to be sinrilar to those of cortvcntional cliesel
cause the maximurn usable fuel flow was lirnitecl by the rate at which
air, Itowever, cunrbustion at
LIQUID-COOLED FAMILY the injected fuel could fincl the nccessary
Model RC2-90 RC3-90 RC4-90 RC5-90 RC6-90 high fuel flow was uncloubtedly less than complete.
into
llJre specific output of engiies running on IP5 fu_el trarrslated
l

I.lP (w/o fan) 285 428 570 7r3 855 IlC cngittcs. Ilighcr
I-IP(rvlfan) ZIS 4r+ 552 690 828 sonre Z0-ZS%, lcss tSan Curliss-Wright's gasolirrc
specific air ao,rru,nlltiorr was, «lf coursc, rtcccssilry cluc to
the cxcess-air
IIPM 5,000 5,000 5,000 5,000 5,000 to attain
Dry weight-lbs. 34, 480 613 740 860 syste,n. It is interesting that the fuel-irrjectecl engine appeared
a higher efficiency in the lorv- ancl micl-power regions (i.e., loY fuel
Specific rveight-lbs ./hp. I .21 | .lZ I .08 1.04 I .01
flo,,u)-, rv5ic5 may have l>cen part of the theoretically
preclicted gain
Length-in . 28 34.5 41 47 .5 54.0
Width-in , Z).8 2).8 27.8 2).8 23.8 over t6e carburciccl enginc. Bascd on other "hybricl" crrgirrc
tcst re-
I-leight-in, 22.0 22.0 22.0 27,§ 22.0 sults, there were still further irnprovemcnts possible in
specific f ucl
Volume-cu.ft. 8.5 10.5 12.4 14.4 16.4 consumption. The RC2-60 Ul0 progranr proved that tlte high-speed
Specific volume- characteristics of the RC engine, togethcr with its inltcrcrrt compact
cu. ft./hp. 0.030 0.02 5 0.022 0.020 0.019
clime,sions apd Jorv wcight, cäulcl bc rctainccl cvct't u'lteu burnirrg heavy

fucls. T5e resulting attractivc powcr to wcight r:rtio_put this cnginc irr
burning tlrese fttels, particu'
a conrpetitive posilion with otircr errgirres
A separate progranr on nrulti-fuel Wankel engines was unclertaken larly i; light äf the RC engine's lorv specific fuel consumption
ancl
on behalf of Rolls-Royce Ltd. under corrtract with the British Govern- low noise level.
ment's Fighting Yehicle Research and Development Establishment in Development of the Curtiss-\\/right Rotating Combustion errgine
Iingland. This joint venture resulted in the creation of the liquid-cooled continues. Over 35,000 hours of engiue operation have been
accumu-
RC2-60 UlO engine in 1965. It differs fronr all other Curtiss-Wright Iatecl to date in test ancl evaluation programs. The con'lPany says it is
Wankel engines in its use of pcripheral ports, but has all the typical ready for mass Production.
design features evolved in the multi-fucl capability investigation, such
as direct fuel iniection and spark ignition. Two different housing designs
were prepared, with notable clifferences in port timing. ,/

INITIAL ROTOR HOUSING DESIGN

Irrtakc port opens 4+3" ATC
lntakc port closcs 840" ATC Intakc port
Iixlraust port opcns 215" ATC arca-2.00
lixhaust port closcs 609" ATC square inches
Ovcrlap 166"

ALTERNATE ROTOR HOUSING DESIGN

Intakc port opcns 4+1" ATC
hr takc port closcs 945" ATC Intake port
lixhaust port opcns zl5" ATC 2.400
Exhaust port closcs 609' ATC square inches
Ovcrlap I 68'
 TOYO KOGYO 247

agreement and a delegation was invitecl to

Neckarsuhn. On Octobcr 3,
presi,Cent 1\{atsuda and a group of five technical Illell went to visit
NSU'
11
Thcy were s5or,v. rcsurts of Lc,ch tcsts o, thc KKI\'I-125,
KKM-250
a slight lack of stability
aucl KKIyI-400 single-rotor cngines. Although
surprisecl to see that
was seen at idling s1:ced, the |apa,ese perty was
be
at high speecls was so snrall tt',rt a coin coulcl balancecl
'ibration
on the eugine. At that timc, KKN'I- z5o and KKN'I-400
engines urotttttecl
over
i, NSU prirrz cars unclcrgoing fielcl tcsts hacl lrlrcacly covered
25,000 miles. Negotiations for an a
t[at sucfi an agreelueltt wou]cl not
Toyo l(ogyo C

e
il

thc cngirrccring staff went to NSU'

t why the wankcl engine hacl not
group lcft for home after arrange-
Iu DncnuBER, 1959, news of the
Wankcl rotary combusti«ln engine reachecl Hiroshima. At that time,
the engineerirrg staff at To1,6 Kog1,o collsiderecl it to be notlting more Kenichi yamamoto, Manager of the Rotary Engine
Development Division of Toyo
than a pipe dreanr. Skeptical criticisnr was voicecl from various parts Kogyo ComPanY.
of the world, and the argur)rents of thc engineering comnrunity were
dividecl for arrd against thc cngine. Hou,e!'er, the rotary engine synr-
posiunr sponsorccl lry the V.D.I. (\/ercin Deutscher Ingcnieure) irr
fanuärry, I960 nraclc it clcar that thi.s cnginc 1>ossessccl grcat futurc
possil>ilities.
By carly 1960, Toyo Kogy'o's cngirrecrirrg staff hecl rcachccl a clcfinitc
opinion on thc rotary cuginc. Thel' fclt that it rvoulcl bc possiblc to
put tlre \Vankcl enginc into practical autonrotive use. To1,s Kogl'o
inrmediatcly sourrdccl out NSU on thcir vicws conccnring a liccnse
agrcenreut, but the rcply,\\/a.s discouraging. A few mouths latcr, on
I\{ay 21, 1960, the anrbassaclor of the l?ecleral Rcpublic of Germanl,
to |apan, Dr. \\/ilhclrn IIaas, paicl a visit to I-liroslrima. Toyo Kogl,cr
receivecl hinr, ancl cluring thc lunchcon held after lr is tour of tlrc
plant, Dr. I Iaas expressccl hi.s apprcciatiorr for thc wanrl rcception
ancl askecl if he coulcl clo any,thing to bc of assistance to the conrpan\,.
Thc topic of conversation soon turrrccl to tlrc Wankel crrgine ancl T\Ir.
il'suneji I\,latsucla, Prcsiclcnt of Toyo Kogyo, rcvcalcd tlrlt thc conl-
pany was attxious to sigrr a liccnse agrccnrcnt u,itlr NSU. up to thrrt
tirnc, NSU hacl been approaclrccl lry apl)roxinratcly 100 othcr conr-
1>anics, inclucling sonlc irr |itplur, rvith sirnilar suggcstiorrs.
Irr lu]y of 1960, T''oyo Kogt'o rcccivccl uncxpcctccl u,orcl fronr NS[J,
through Antbassaclor Ilaas, that NSU was prcparecl to conclucle a license
 Yamamoto attributes the speedy results obtained in his department to the extensive Experimental Mazda four-rotor Wankel engine.
use of-computers. Here, t_est readings are continuously fed-in from four dynamome-
ters, placed in a memory bank, and are instantly available for analysis .
-

Experimental Mazda three-rotor wankel engine.

The first twin'rotor production engine carried the designation 0813. It was de-
veloped for use in the Cosmo car.

TTYIN OISTRIBUTOß8
250 DEVELOPMENT
ments had been made for shipment of technical infonnation, draw-
ings, and tcst engines from NSU. Upon their return, a development :
conlmittee, composed of members frorn the company's design, rna-
terial rescarch, production engincering, manufacturing, ancl test di-
visions was organ ized for the purpose of carrying out full-scale research
and development work on the Wankel engine. Toyo Kogyo decided
to commence research rvork on the KKM-400 engine.
The company spent $750,000 to build and equip a special rotary-
engine testing center. The test chambers were monitored by closed-
circuit television ancl all test data was immediately reduced to digital
fornr for computer storagc ancl Iater analysis. A substantial crew, with
up to 180 members, was assigned to the project, under chief engineer
Yoshio Kono and his rotary engine development chief, Kenichi Yama-
ntoto.

Temperature distribution in the water-cooled rotor housing, under

wide opcn throttle, at 2,000, 3,000, 4,000 and 6,000 r.P.m. runnln; condttloor
tull lord
u.ß.r te4.= gotC (tnlat)
o ?000 ß7Ä
with c, Jlu R?tl
end cover,
Temperature distribution arottnd the side wall and artt
different curves for 2,000, 3,000, 4,000 and 6,000
r.P.m. A R?rl

x Soco np{

I
I
in the rotor (with oil cooling), at 5,000 r.P.m. and wide
.t
I

Temperature distribution -.
1
t

open throttle. (
I

5000 r.P.m. FULL THR0TTLE 1

i

rl
//
219" 251" 223' 200'
;F-',c
i

192" 'l
'I

;
I

,,1

.,1
)

.
I

:l

I
. runnlnS cotrd I I lonr
- lutl loed
u.tGr t?lP.-90'c (tnlet) :

O ?e0 37s1

d Jooo R?H

A .ooo R?H

x (ooo R?H I

I
 TOYO KOGYO 253

The first NSU KKN{-400 enginc and its acccssories arrivecl at Toyo
Kogyo in November, I961. Thc engine was inrmediately dismantled,
examined, reassemblcd and placcd on the test bcnch. The tcst showed
an output of 4).8 horsepower at 9,000 r.p.rn. Toyo Kogyo built their
prototype engine No. I from KKM-400 drarvings supplied by NSU.
Tlre test results with prototype engine No. I wcre extremely discourag-
a-

.s
irg. The engine showed excessive vibration at iclling speed, emitted
F

E a

-, large amounts of white srnoke, ancl its oil consrrrnption put it be1,on6
F

-F.s all practical use. When thc cnginc had mn for 200 hours power output
@ suddenly dropped. Upon tearing dorvn the engine, it was found that
R chatter marks on the trochoidal surface hacl caused the electroplating
d
to fall
Y

off.
During the first year, To1,s Kogyo's stuclics concentrated on the I

fundamental phenolrcna of thc rotary conrbustiou cnginc arrcl its proll-

I
I

lems. In the following trvo )'ears, enrphasis was placed on mapping

tlre basic problems of the Wankcl rotary engine. Ilarly in 196?, parallel
work was accelerated with the clarification of various problerns on the
-rffi rprn W.O.T bench. A program was instituted to mount the engine on a test car
so that its adaptability as an automotive engine coulcl be tested in its
Cas_ pressure differences between the true environment. The prototype engine was installed in a snrall test car,
three working chambers, meas-
ured in pounds per square inch (p"). but new problems were cliscovcrecl whcn the car was put into opera-
tion. The engine ran very smoothll,at high r.p.nl., but it becanre un-
steady at slow speeds-when the throttle was closed, strong vibrations
followed. Considering the type of passenger cars planned by the com-
pany for future production, a decision was made to develop a trvin-
rotor engine.It must be noted that this was at a time when the single-
rotop engine was not yet consiclercd satisfactorl,.
During the seconcl and thircl yerrs, all efforts hacl been directed to-
wards the solution of the basic prol>lems. In the fourth and fifth )'ears,
Toyo Kogyo studiecl the various factors affecting cngine pcrformance
and concentrated on inrproving the rotary engine for autornotive use.

N:,::::,
EX

This sketch, and the three in- ..r (N\\ -uQil rol rpring
r'
sets, illustrate the variations in
apex seal leaning angles. When
:,|\§Ps L':ir,,,,r-.lL
iiü

Gas and oil

seal configurations
one_ apex seal has zero leaning
for the Cosmo engine.
f,nSle .,(toryalong minor axis)]
the other two are at maximum
Ieaning angles, one negative and /*,,ro,
one positive.

,i
:-a
;i
254 DEVELOPMENT
Test results rvith the original prototype engine illustrated many un-
100

90
T rl
TO UNBURNED G AS, ETC.

solved problems. Because of this, the necessity of accumulating new

data on the basic operational phenonrena of the rotary engine was
real ized, and a thorough basic research program was started to solve 80
the fundamental problems-Toyo Kogyo wisely saw little or no value TO EXHAUS GAS
in random trial-and-error tests. 70
Next came the clcvelopmerrt of special test methods and measuring
instruments for this purposc. Irull-scale engine tests were devised to
verify the results of the basic research. Despite outside criticisms and 60

l-\
doubts about the rotary engine, Toyo Kogyo felt that evaluation stand-
8
ards for the rotary engine should be as severe as for the piston engine,
UJ
(,
z 50
t-
which enjoyed the benefits of over 70 years of development. In April, UJ
(J

1963, the rotary engine development division was organized. With

E.
ul
G

I\4r. Kenichi Yamamoto as its nlan aget, the division started off with 40
four departments-research, design, testing, and materials research. At
first, the clivision had only five test benches to work on and the build- TO COOLING
ing was old and gloomy. The facilities and surroundings could not be 30
called at all suitable for this t),pe of research. Then the company took a
decisive step ancl built a new test laboratory. The construction of the T-
basic test cells in the new laboratory was completed in fanuary, 1964.
20
.l-his
was followed by the construction of the endurance test cells, which TO NET HP
were put into operation at once. 10
The new facilities were equipped with the most up-to-date eJectronic I
conrputers ancl industrial television, which were to play important roles
I
o 1o3
Oulcr oil tcol
'ä:''
IIeat balance chart ror the ,"t"'ir'j-:ä:
Oil rcot rpring showing the
gas aud unbumed
perccutage of lic:rt gi'c'- off to wrrte r, oil, exhaust
gas, throughout its sPeed range'

Front Bonkl
The housing incorPorates water
cooling by a urulti-Pass sYstem,
Stotionory gcor Details of the oil seal. designEd to
Provide-. maximum
heat" radiation according to the
heat Ioad distribution. Water
oler I
florv is axial so as to give reciPro'
trtnp .A, catirrg balance. The radiator is
of the sealed, high-Pressure tYPe
and is connected to a generous
reserve tank.
" Sidc houring !99'"191-
216 DEVELOPMENT
N
in expediting research ancl development. Test benches with centralized E10
T
t)
controls, capable of running endurance tests 21 hours a dry, were in- UJ
E
stalled. The control roonl was locatecl apart from the test benches, and
E t
all the tcst results were automatically recorded and typed by an electric o
L t
t
typewriter. By August, 1964, when construction on the laboratory was
ut
I
completed, there $/cre approxirnately thirty benches ready for operation. E5
lr
lt
trl
Yamamoto designed the experimental Toyo Kogyo Wankel engines EXHAUST INTAKE

with ligllt-alloy rotors. The housing and end-covers were made of cast-
iron, and the trochoidal surface was given a hard chromium plating.
'l'he Wankel engine for the Toyo Kogyo I\{azda car contrasts with the
BDC TDC BOC

NSU units in having apex seals made of carbon rather than cast-iron. ECCENTßIC SHATT ANGLE

Toyo Kogyo now believes in the wear-resistance and self-lubricating

This tirnc-area cliagrarn conlparcs thc cffcctive port areas of side
ports to 1>crilrhcral ports, throttgh onc full cyclc.

Lubrication and oil circulation system in the 0820 engine. Oil is cooled in a water- rtdc port
oil heat exchanger. Oil and water flow is controlled thermostatically. Oil for lubricat-
ing the apcx seals and side seals is supplied by a small metering PumP, which de-
livcrs minute amounts of oil into the rvorking chamber in proportion to engine load. 15
Oil is conrpressed by the gear-driven oil valves and fed via the thermovalve to the
heat exchanger and pressure regulator. Constant-pressure cool oil is passed through
a full-flow filter before it enters the nrain gallery of the housing. )a

,\ 10
u
c
!
a
C'
o
l.
tta

/
l
t This chart compares
fleoc tr.vGl tlrc (nr)
the fluctuations in i_on speed for side polts
(;;lid fi*), .o*'ti.ports with exhaust he
-
broken line),. and combi'
irorti t itiiä'.,t cxhairst lrcat vulve (irrcful The test engine rln at B
äorrstant 1,000-r.p.nr., with a ,r',.ri, cffcc 86 psi. 'l'lte skctch shows
s1>ark lrlug location arrd firing lroirrts.

:-'- -
g o Swch

Tq d
i
g onCo
tll I ll ,JLoiling Drrlrrbulor
Ey
Dual coil four-plug ignition sys-
tem for the 081 3 engine.
258 DEVELOPMENT TOYO KOGYO 259

propertics of carbon, ancl urinirnizes the risk of apex seal breakage by hacl two longituclilal ancl several crossing
pcrPenclicular holcs, the ell-
de tonation through thc use of two sparrk plugs for each rotor, with uiüt.iio,,t ancl elinrination of
gineers clairnecl reclucccl high-frcquency
slightly staggercd ignition tirning. the chatter mark phenorrr.non. I;gcnious, and no doubt exPcllsive to
As was the case rvith NSU ancl Dairnler-Benz (about which more rvill the .rorr-lrollow rnetal seal was soon replacecl by a bctter ma-
produce,
terial-a carbon compouncl. A spccial carbon compou,cl
im1>rcgnatccl
be saicl in a later chaptcr), nruch clevelopment work at Toyo Kogyo
went into fincling compatible materials for apcx seals and the trochoidal witS aluminum was developecl inr the apex seal rnaterial'
It had a
qualities and high durability'
surface, so as to reclucc wcar on the seal tips and eliminate chatter low coefficient of friction, ,.jf-lubricating
that of ordinary
marks on the working surface. Whcn compared with a piston ring, The strength of this nraterial was nrr.l, higher than
the harcl chro-
which always sliclcs back ancl forth ovcr the silnre contact surface, the carborr ancl its friction clraractcristics clid not clarnage
problcnr of wear
apex seal älway.5 changes its sliding surface; this gives it an added ad- mium plating. The use of this new seal elirninated the
Tlte rcstrlts of a 60,000-
vantagc agairrst wear. N{orcover, its motion is uniclircctional, and the ancl chatter marks on the trochoiclal surface.
of 0'0+ inch in
leaning anglc of thc scal strip changcs cyclically along the trochoidal mile car test shor,r,ccl \\'ear on thc carb«rtl tlPcx sc:lls
path. The rnetal apex seals used in the early experimental engines
producccl chatter marks on tJre trochoiclal surface and wore rapidly. cavity designs with respect to
cornparisorr of gas
I ht.r but higher peak
Retaining iron as the ilpcx scal strip rnaterial, the Toyo Kogyo cn- velocity. 'r.hec !rrn,l,r..s
girrecrs devcloped tlre cross-hollorv seal. Their proposed method was to
drill a hole crosswisc ncar the tip of the apex seal and then drill another i:l:i'fl,,3'iilt
section showecl mucS nore unifornl
,:,rü!':,,:r*{':1.:X','T",H;
gls-vclocrty, .with a-peak near
initial firing, then
holc lcrrgthlisc to intcrscct u,itlr the first hole. The results of tests on the minor axis reaching its "ir*i*urn"b.fott the
this new cross-hollclrv scal proved to be satisfactory. Because the seal graduallY dirninishrng'

tnillng
rfrl ilJg

llnl
llrrt qttio ftwr
lniling sf,fi plug FIame propagation along the
Itrnr poprlin lrqn trochoidal surface (looking di-
ldhe rgalt plug rectly at the rotor face).
 TOYO I(OGYO ?61

height and of 0.004 inch in width. A power loss of 15% was recorded
due to wear. In this case, the wear of tlre chronrium plating of the
rotor housing was under 0.0001 nlm. and was considered to be
negligible.
Lubricating oil was supplied to the working surface on which the
apex seals slide by mixing oil into the fuel. This was done by metering
the oil with a small, variable-stroke, plunger-type metering pump. The
oil was diverted frorn the main passage into the float chamber of the
carburetor in proportion to engine speecl and load. Racing engines run
continuously at 10,000 r.p.m.To1,6 Kogyo tcchnicians say that in these
high-speed Wankels, the bearing matcrial is a biggcr problem than apex
seal material and construction.
Recent reports sa),that, de.spite tlteir stlccess with carbon seals and
a chromium plated working surface, Toyo Kogyo rvill soon go to metal
seals and a nickel/silicon carbide surface. The reasons Are believed to
Arrangenrent of the dual spark plugs.
be overwhelmingly econontic.
The inner and outer dual side seals, arrangecl on both flanks of the
rotor, were made of the same material as ordinary piston rings. Pin-
shaped side seals also were provicled at the joint portions of the apex
seals and side seals. These seals \r,cre pushed against the sliriing surface
Performance curves for the 08 r 3 cosmo engine.
of the housing by spring force and gas pressure behind the seal. The
lZo rotor seal constructiorr prevented lubricating oil florving through the
engine from entering the combustion chanrber. Near the center of the
rotor flanks there were double concentric grooves which housed wave
E
I
@
t Performance curves for the 0820
t, (R-100) engine.
l-
CT ,/
o
L)
A
ta
t
v
q
l- d
o a
6o
I c
U
B
A 00 ß
o
to
.l
lr a E
so ,l CL
a
o I
xta, 40
a
I ä
l. 6 D
E
a
i
cs
3 ut I

t T
\r

A ? 2

ri
v, u,
A (,
o0 T lL'
.o 6
U
a
ci
a
l&a

t
a

looo Zooo Sooo Aooo Sooo &oo t23.5 6 ?1,

tml
englnr apeed (rpm) ENGI}€ REVOIUIION r.R,n
262 DEVELOPMENT TOYO KOGYO 26)

springs ancl oil seal rings. Thc tapered lip of the oil seal ring was pushed tainecl regardless of a small overlap. Tlrc rcsult wtls otttstatrcling ltigh-
against the encl cover of the side housing and this scraped off exccssivc speed p.ifnrmAncc without any sncrificc of lorv-spcccl flcxibility. T'hc
oil. Experinrcnts with clifferent rnatcrials for thc crrcl covcrs enclccl with intake ports wcrc of thc cloulllc siclc llort tvpc, u'itll lllixtttrc intlrkc
Toyo Kogyo choosing the same material for the newest engines that taking p1... from both sicJes to each rotor. 'Ihc ports oll the intcr-
hacl been used on the earliest experimental units. The Mazda Cosmo mediäte housing are called the "printary" stage ports arrd those oll
had aluminum end covers with a molybclenum-based coating; on the the side housinf tn. "seconclary" stage 1>orts. The ports were connectecl
Iater R-100, R-I30 ancl RX-z, the encl covers were incluction-harclened to a two-sta Ee, four-barrel carburetor, and each port was fed fronr a
cast-iron. completely ina.p.nclent incluction s1'stenr. The prinrary stage nratti-
Although NSU and l)aimler-Benz are firm believers in peripheral folcli *.r. prefieate«l by cxhaust gas in orclcr to spcccl tU) crrgirlc wälrtll-
ports, Toyo Kogyo, along with Curtiss-Wright, chose side ports. The ttp, as well as to improve "clriveability" and fuel econolny.
penalty for NSU was irregular running at light loads and low speed
(ovenun).Sicle ports gave slower port closing and opening, which nreant
lack of peak power but better low-speed and light load scavenging. On Cross section of the 0820 engine.
the earliest test engines, Toyo Kogyo used a peripheral intake port. This
induction system was largely responsible for the unsatisfactory per-
formance at Iow speeds, the shaking irr the vehicle on the overrun, and
the unsatisfactory idle which was experienced on prototype car No. I .
In the fall of 1963, Toyo Kogyo began to design an engine with a side
port induction system. This new system was an induction systern with
the intake ports locatecl on the side housings, at right arrgles to the
ports of the pcripheral system. The reduced overlap lessened the mixing
of combusted gas with nerv mixture, ancl combustion at low shaft speeds
was stabilized. The vehicle with the side port engine showed clear
superiority to all earlier versions.
The cycle time of the rotary combustion engine is 1.5 times that of
a reciprocating engine (1,080 degrees shaft angle); therefore the Toyo
Kogyo engineers felt sure that port-opening time dred could be ob-

General layout of the 0813 Cosmo engine. Corburetlor

(Pl (s)

Mclcring Oil Pump'

Dirlributor
Rodiotor-
-+
-+ Prelture Swilch
-.Rclief Volve
--+

Moin Beoring
+
€
+ Oil Pump
Rolor Beoring
Oil coolor
--+
Regulolor Volve

Thermovolve
Oil Pon
Oil Stroincr
264 DEVELOPMENT
TOYO KOGYO 765
The Cosmo engine had a combination of sicle and peripheral intake
ports. The induction system began with a special triile-tSroat Strom- proachecl wide-open throttle corrclitions. The two seconclary throats de-
berg carburetor. The central primary throat, which rr.i in operation livered fresh mixture to peripheral intake ports. Tlte cornbi-port was
at
all times, fed a channel in the wall that separated the two rotor cham- almost equal in volumetric efficiency to the peripheral port, and corn-
bers. The channel was split in two, giving a sicle port to each chamber. bustion in the case of the side port was more stable than was the case
The other two throats went into action änly when the engine load ap- in the combi-port.
When the 0820 engine for the R-100 was designed, Toyo Kogyo
abandoned the compromise combi-port system in favor of dual side
Elevation of the 0820 engine.
ports-one primary and one secondary for each rotor chamber. The
carburetor was of the downdraft type with clual throats, one per
rotor chamber, and the gas passage frorn each throat was split in two
right below the carburetor mounting flange. One charrnel led into the
dividing wall, the other into the end cover. Each rotor chamber had
intake ports on both sicles, arrd the two central ports belonged to
separate gas flow systems. It was an important feature of the side ports
that their outer edges could not be located very far outboard without
causing the apex seal "bolt"-which must have axial freedom for proper
sealing-to fall through the aperture, rn,hich u,oulcl be producing, or at
best traversing, a wear-inducing clepression. This meant that a sicle inlet
port could not open before top clead center, and that a side exhaust port
could not remain open after top dead center. Consequently, there could
be no overlap between the two side ports. It was also less easy to design
side ports with an adequate capacity (which is why Toyo Kogyo re-
verted to double side intake ports), ancl the gases they carriecl hacl to

Complete R-100 (0820) engine with clutch and gearbox.

o
 TOYO KOGYO 267

obtained by moving the two ports as close as p«lssible to the nrinor axis
of t6e trochoid, so that both simultaneously remaitl open for a long
periocl as the rotor swceps through the top clead center Position.

0820 rotor with all seals in Place.

)
-t/

Explocled vierv of 0820 rotor and its seals.

Front view of the R- 100 cngitte, complete with accessories.

be turned through at least one right angle. Agrinst these disadvantages

must be balanced tlre inrproved lorv-speed torque that followed from
the absence of overlap between side ports.
It is interesting to note that peripheral ports suffer from few of the
disadvantages of side ports. Although their cross-sectional area may be
restrictecl by the width of the rotor ancl its housing they have ample
capacity because they stay open longer, and the gases which flow
through thcm need not be turrred through a right angle. I\4oreover, a I
I
i
l

peripheral intake port can open before top cleacl center and a periph-
I
I

eral exhaust port can close after top cleacl center, giving better timing
than can be obtained from side ports. Overlap is important to ensure
o
complete emptying of the exhaust gases and gqocl filling with fresh ,t
ntixture. Overlap is theoretically detennined by the positions of the
rotor apex seals, wh ich open and close the peripheral ports as they
srveep over them. Maximum overlap for a rotor of afiy given size is
268 DEVELOPMENT 1'OYO KOGYO ?69

Stable combustion at part throttle is desirable for the Wankel auto- firecl by two coils arrd two clistributors. This systenr was intenclecl to
motive engine, especially at low speecls. The stability of combustion at proclucä more complcte conrbustiorr regardless of flanle front trirvel
part tlrrottle is dictatcd by a large number of irrterdepenclent factors, pattells.'I'wo spark plugs wcrc irrstallcd, ollc olt cach siclc tlf tlrc tttitttlr
such as induction system design, intake and exhaust port timing, com-
bustion chamber shape, spark plug location, and transfer port con-
figuration. One of Toyo Kogyo's solutions, on the 0813 engine, was
the use of two plugs per chamber, 5 degrees apart, one above the other,

Rotors 6tted on the rnainshaft.

Rotor housings with exhaust ports and spark plugs.

The two rotors and their phasing gears, placed in proper relative
positions.

Mainshaft of the 0820 engine.

I
I
a
I
t
I
 270 DEVELOPMENT TOYO KOGYO ?71
Tlte plug on the Ieacling sicle in the direction of rotation was Io-
axis.
cated 18 nrnl. from the axis; thc one on the trailing side 30 mm. from transfer holes connccting the trochcliclal surface with the clect'roclc
the axis. In this arrangement of dual spark plugs, the sizes of the chambers were not the sanre on the trailing and lcading siclcs; thc onc
on the trailing side was 4.2 nlrn., ancl the one on the lcading sicle
lZ.4 mm., in cliarnetcr. 'l'hc two sizes wcre ncccssary bccause of thc clif-
fercrtce in gas pressurcs bctrvecn the acl jaccrrt chantbcrs scllaratccl by,
the apex seal. It was not 1>ossiblc to nrakc a largc lrolc on tlrc troch«licltl
surface at thc position of thc trailirrg spark plug; thcrcf«rrc thc trailing
sllark plug wl.s nlollntc«l irr:r ptlc'kr:t sliglrtly rct:cssccl frorrr tltc:
troclroidal surfacc, n,it'lr rr srrurll tnrrrsfcr lrolc c'outrcc'titrg it n,illr tlrc
combustion chanrl)cr. At tlrc positior r of tlrc lcld ing s1l:rrk pltrg, t'lrr:
diffcrcnce in gas prcssurc wäls rrhnost zcro.'l'llc clcctr'«lrlc crlrrkl thcrcfrlrc
be brought close to the trochoidal surfacc, rturking ignitiou rtncl rc-
sistance to fouling at lorv speed much rnore stablc as conrparecl to that
at the trailing side. The spark plugs werc iclentical for both thc leacling
and trailing sides, and were specially clcvelopccl for the 0813 engine rvith
due attcntion to their resistancc to foulirrg at lorv speecl ancl rcsistance
to 1>rcignitiorr ut'lriglr spccd.
Another problcur \\,as c«roliug.'l'lrc Warrkcl crrgirrc has n "lrot lobc,"
which is nevcr srvcpt lry cool urrburuecl gase.s, äls is the courbusti«rrr
Errd covers and ccntcr sidc wall with intake ports.
c]rarrrbcr of a piston cngirrc, ancl tlrc tcrnpcru turc riscs rapidlv lls tlrc
enginc loacl is incrcasc«I. In thc I\'lazclu cnginc, tlrc irrcorning rnixturc
Tr'o vie's of the fully assernbrcd 0gz0 engine. for snrall throttle operring.s \\'rts llrclrcirtcd so rl.s to wiu nl thc cool siclc
of the engine and minim ize thc tcnrperuturc cliffcrcnces :rcross it, as
well as to improve atomization. Tlrc proccss coulcl be carriccl a stagc
further by blcecling exhaust gilscs^ througlr passilgcs in tlrc cool 1>arts of
the rotor end housings.
The IVIazda engine housing was coolecl by rvater that flowccl fronr
the water pump, installed at the uppcr front encl of the enginc, through
channels in the housing (in thc ariirl clircctiorr) to thc rcur sirlc housing
ancl back into the recirculation cyclc. Ilibs in the clrannels were so
designecl that the watcr florved at a rate proportional to the thermal
load distribution. After circulating through the engine, the cooling water
passed through the bottorn bypass-typc thcnnostat to an alunrinunr
radiator. The use of this type thermostat ensurccl a sufficient amount
of coolant circulation insiclc the cngine \r,hen thc themrostat u,as closecl.
This made it possible to holcl the rnaximunr u,all tcnrperature below
4l0oF. The cooling fan was drivcn lry a fluid coupling thirt rcduced
noise and power Ioss at high spcecl. The oil clclivercd fronr the oil
pump passed through the tlrernrovalve to the oil cooler, rvlrich \\'as air-
cooled. The oil therr n ent throtrgh the oil filter arrcl entcrecl the en-
I gine. After lubricating each of thc bearings, thc oil was ejectecl into
the chanrbers insicle the rotor, where it coolcd thc rotor. Incrcasccl rotor
I

i
277, DEVELOPMENT TOYO KOGYO 773
housing temperatures-by up to 70"IT. comparecl to those of the housing spends several expensive hours irr the plating bath) and
formecl in the original manner-were a penalty of 'Io1,s Kogyo's basic
at the same
time to make use of the nruclt cheaper die-crrrting pro..rr, clcspite
design. This was due to thc relatively lclw thcrnral concluctivity of the t5e
consequently poor bond bctrvecrt chrornium ancl aiuininunr.
chrome steel layer on the working surface. Tcsts conclucted on the two
Thlir solu-
tion was to rnake a Jicensing agreenrent with the Dochl er-larvis Di-
kinds of rotor housing, however, showed no difference in durability or vision of the National Lead Conrpany of Anrerica to use a new
wear resistance. metal
spraying and casting technique known as TcP-Transplapt
Thermal fatigue, 0s a result of alternate heating ancl cooling of a
Coating
Process.
given area, was also a problenr in early Mazda engirres ancl caused crack- In this Process, s thin layer of steel (which aclheres strongly to alu,ri-
ing at strcss corrccntration poirrts arouncl the spark plug holes. Tltese num) is sprayed on the inner trochoiclal core of the .li; rvhicS
cracks, in extrenre cases, penetrated to the water iacket. The first ap- preheated before the aluminum is pourecl in. f'he result
is

proach of the Toyo Kogyo engineers was to improve cooling to cope is a steel-
coated, die-cast rotor housing that is machined before being
plated
with the most critical condition-rapicl accelcration of a cold engine. a very thin Iaver of chrome. After experience with the p'"ro..r,
In winter the combustion chanrber walls could sucldenly rise from a low -with it
became possible to hold the thickness of tile steel layer
ambient ternperature to 450oF. This was the reason for adopting the to 0.015-0.040
inclt, after
most modern type of bottom-opening thermostat; one that continuously to only 0.002llchining, and to recluce the thickness of the chrome Iayer
inch, thus saving plating time. Because the wear of
controllccl bypass flow as well as flow through thc racliator. A further chronre plating is negligiblc, thcre was no problcm about thc
cause of the cracking trouble was lack of flexibility in the trochoiclal life of
the rotor housing. The tuore inrportant proüI.,r, concerlipg t5e rotor
wall ancl concluction of heirt from it to the colcl end housings through housing was how to obtain an accurate ancl thin chrom.
the bolts holding thcm together. Improved flexibility and reduced heat ptrting on
transfer were sirnultaneously achieved by separating the side-bolt bosses
Cutaway drawing of the 0820 engine and transnrission.
from the trochoiclal wall, and housings macle in this way proved corn-
plctely resistant to thcrmal fatigue.
As happened with other manufacturers, chatter marks on the trochoid
surface of the rotor housing wcre experiencecl by 'l'oyo Kogyo cluring
the early stages of cleveloprnent. It was believed that the chatter marks
wcre the rcsult of higlr-frequency vibrations, ancl basic research was per-
formecl on various subjects such as the vibration characteristics of the
apex seal and the rotor housing, friction clraracteristics as an exciting
force, the influence of lubrication, the sliding speed of the apex seal
and its relationship with contact pressure and the lcaning angle of the
apex seal. It was necessary to test arrd try every possible material avail-
able. Great numbers of rotor housings, all with chatter nrarks, were
soon pilcd high in the laboratory. T'hen a new method was corrceived
for measuring vibrations. Aftcr six nronths a proposal was madc to alter
the scal configuration.
In the course of cleveloprnent, several cffcctivc rncthods of reducing
the chattcr marks werc found, but any methods that woulcl have irn-
posed a limit to the durability of the rotor housing were rejected as
unclesirable for practical use. At first they usccl a conventionally cast
aluminunr alloy rotor housing rvith hard chromiunr plating approxi- I

mately 0.006 inch thick on its trochoiclal surface.'l'heir next objectivcs I

were to recluce the thickness of this chromium layer (because thc r«ltor
I
I
I
 I

274 DEVELOPMENT TOYO KOGYO 275

the trochoidal surface. T'here is a fair prospect that chrome-plating with the spark plug was not the answer with the Wankel engine. The work-
a thickness of several tcns of microns can be practically applied, requir- irg conditions of the spark plugs for the Wankel are determined by
ing no grincling, in procluction by using special equipment. the relationship of such factors as the construction and heat value of
T'he conventional ordinary carbon seal material could not endure the the spark plug itself, the cooling cffcct of the plug seat, tlrc location
scvere operating conditions required for the apex seal, especially in the of the plug, the volume of the chambcr arourtd the clcctroclc and thc
areas of strcngth and wear. F or scveral years, Toyo Kogyo's research size of the shooting space between the electrode and the combustion
group made a stucly of carbon in close cooperation with a special de- chamber. In conjunction with their studies of engirre clesign, the Toyo
velopment teanr organ izccl by a carbon company. The development of Kogyo engineers invented netv types of spark plugs in cooperation rvith
a non-destructivc inspcction mcthod (NDI) and a carbon which was the enginecrs of lcacling sperk plug mukcr.s. Spnrk 1>lugs thlt lastccl bc-
strong enough to endure the severe operating conditions encountered tw,een 7,500 and 8,100 rniles on the road, ancl plugs tltat could ettdure
allowcd 'foyo Kogyo, by 1967, to run routine tests up to 8,000- more than 100 hours of running at full throttle on the test bench, were
8,500 r,p.nl. Further improvement irr the seal enabled them to raise cleveloped as a result.
engine speed further. After five years' expcrimcntal work it was apparent that in orcler to
The wear of the sicle seal was negligible, but the wear on the design an engine that would fully utilize the best of the inhererrt
permanent coutact section of the oil seal caused an increase in oil characteristics of the rotary principle, a twin-rotor sicle-intake-port en-
consumption and therefore could not be neglectecl. Toyo Kogyo had gine would be the most desirable for automotive application. In 196+,
adopted the aluminum sidewall for its light weight and high cooling an engine with 30 cubic inches (191 cc.) chamber clisplacenrent was
efficiency. Flowever, the slicling snrface of the aluminum sidewall, with dcsigned. This engine, with an output of ll0 horsepower at 7,000 r.p.Ilt.,
no surface treatmcnt appliecl, was subject to serious wear and posed macle its first appeararrce on the rtrarkct in 1967, ntoturted irr a Mazda
problcms in practical use. There were several possible methods of sur- ll0 S sports car.
face treatment. Thc sliding surface of the sidervall in the production
I\4azda engine was sprayed with high-carbon stcel. Flowever, because all
of the seals which slid on the surface of the sidewall were macle of
metal, sonrc arnount of wcar naturally occurred on the hardened srlr-
face of the housing even with lubrication. The wear in the cylincler
bore of the reciprocating piston engine is greater at top dead center.
Similarly, the wcar on the sidewalls of the Wankel due to side seal
frictiotl occurrecl mainly at a position near top dead center. On the
otlrer hancl, the \\,ear on the siclewalls due to oil seal friction occurred
within the circular envelope clrawn by the outer edge of the oil seal.
After considerabJe milcage, wear on the sidewalls sometimcs became so
great as to rccluce powcr ancl increase oil consurnption. Because the
surface of the side housing was flat and the sprayed-on coating thick
enouglt, the worn out housing could be used again by grinding the
surfacc-the fact that the siclewalls were slightly unclersize gave rise
to no r)ew problerns.
The spark plug in the Wankel engine is exposed to more severe heat
conditions than in either the two-stroke or the four-stroke reciprocating
piston enginc. The problem of high heat input can be solved by
using spark plugs with a high heat value. But, taking into consideration
cold-starting characteristics and the risk of electrode fouling during
continuous running at part throttle, irnproving only the heat value of

T
 DAIMLIiI{.RIINZ ?77

l2 Although Hitzinger was convincecl of the engine's goocl qualities, he

did not want to pay too much for the right to nrake it. The Dairnler-Benz
attorneys rewrote contract after contract, ancl the NSU management
found Hitzinger to be a real hawk of a negotiator. I{e also kept finding
new reasons for delaying the signing of the contract. This is specula-
tion, but it's possible that Ilitzingcr stalled bccause Itlick was buying
more and more NSU stock, which in turn rvould havc meant that
Daimler-Benz, once it owned NSU, could nrirke Wankel engines with-
Daimler-Benz out paying a Iicense fee or royalties. I-Iowevcr, Hitzinger soon real ized
that ownership of NSU was a long way off, ancl a contract with NSU
was signed in October, 196l . The monctary arrangements involvecl
payment of a flat fee of $750,000 in three annual installments. But,
there was also a clause giving a minimum pa)/merrt to Wankel G.m.b.FI.
and NSU each year. This was bitterly fought by l{itzinger. The reason
this clause was in the NSU contract was to guarantee that license
takers would not unduly delay development and production of the
NSU-nurLT WaNxur rEsT ENcINEs Wankel engine for business reasons. In other words, they were forced
were running at Daimler-Benz as early as 1959. Wolf-Dieter Bensing€r, to push the Wankel engine forward as quickly as possible in their own
formerly head of the passenger car engine design office, was in charge interest.
of the program. Bensinger had joined Daimler-Benz in 1943, after IZ Bensinger began design work on a Mercedes-Bcnz Warrkel ertgirre in
years at DVL, where he u,orked mainly on control units for aircraft en-
1960. His plan was to test and develop both a single-rotor aud a twirt-
gines. In 1945, he was assigned to a new passenger car engine program. rotor version of the same basic engine simultaneously. The single-rotor
His work led to the development of the new line of engines with over- version, with a chamber displacement of 42.5 cubic inches (700 ...),
head camshafts actuating slightly inclined valves via finger followers. was tested first. This engine lost its vibratiotrless running after 400 to
These were the engines that went into production in the l95l Mercedes-
500 hours under high loads, which corresponded to 25,000 to 30,000
Benz 220 and 300 series. miles on the road. After that period, the seals began to flutter and the
Bensinger went to Lindau to see Felix Wankel in 1960. They had casing showed pigns of pressure waves ancl vibrations that robbed the
known each other through DVL contact since 1934, and Bensinger engine of performance. As with the other Wankel developers, chatter
wanted to consult Wankel himself about his ideas and proiects. How marks showed up on the trochoidal surface. Lingine life \4'as considerecl
did Bensinger propose to use the Wankel engine? His first plan was to good enough only for minicars and motorcycles, the engines of u'hich
develop a 1.4 liter Wankel to replace the piston engine in the ZZ0 SE. usually were overhauled within 25,000-30,000 miles anyway. liollow-
Felix Wankel was, of course, delighted. Bensinger began an intensive ing NSU practice, Bensinger then used peripheral intake ports, rvater-
test program, and during its first year of testing, Daimler-Benz spent cooled housings, and oil-cooled rotors. Poler output from these Ilew
about $750,000 on Wankel engine experiments. It is significant that engines was remarkably good, but performance fell off at an early stage.
Daimler-Benz had not yet acquired a 'Wankel engine license. There was, Tests with an 85 cubic inch, tn,in-rotor urtit in "as-Ilew" conclition
holever, a tacit understanding between the engineering staffs of the showed it to produce 160-170 horsepowcr at 5,000 r.p.Ilr.
two conrpanies that NSU would be given the benefit of Daimler-Benz Chatter marks on the working surface was the worst problenr cluring
test reports and any improvements made, in return for furnishing test the early experimental period. I)r. Bensinger gave this explanation of
engines to Stuttgart free of charge. Bensinger's report to the manage- the phenornenon: "The apcx seals push oil, carbotr particles, cotnbus-
ment of Daimler-Benz, concerning Wankel engine feasibility, was tion by-products and even fuel ahead of them. When the resistance
favorable, and the president, Walter Hitzinget, decided to negotiate a of these substances becomes too high, the seals tend to lift off the
contract for Warrkel engine rights. trochoidal surface. When this happens, the trailing edge of the seal
 rr8.t.
DAIMLER.BIiNZ 279

tips tencl to break off minute particlcs from the rvorking surface."
When serious work on the Wankcl engine began at Daimler-lJcnz in
]960, chatter nrarks coulcl be observecl aftcr r)o nrorc thau fir'c hours'
running. Since that tinre, there ha.s bccn nruch iurprovcmertt. Scal
c]rattcr r)evcr nlity bc conrplctcl),cliurirr:rtccl, slirl I)r. ]Jcrrsingcr, but tlrc
F;g 3 F;g 4
consequences have, at prescnt, bccn rcducecl to levels u,here thcy no
longer have arly influence olr the life of thc trochoiclal surface.'f'he solu-
tion to the problem was Nickasil.'l'ltis is rl lltl,sv of nickcl, crcatccl Ir),
IS 22 galvanic process, lluriccl in a vcr1, firrc silicorr cnrhirlc llc:itrirrg arrcl usccl
t6
26 for the entirc rvorkirrg srrrfncc. It rva.s origirut]lv clcvclollc«l irr c(x)l)cratirlrt
2t betu,een Daimler-Benz ancl l\{ahle Kolbcn\r,crk, although NSU took it
up later, aucl thc proccss has bccn uraclc avr.lilablc to \Vaukcl cngirrc
nl Iicensees all over the rvorld.

f;g S r'9 6 The most serious problern, rrext to chatter nrarks, \\'äls scal tip \\,c:lr.
The solution \\/ils not to bc founcl by stuclying nratcri:rls alrtuc. It \\'as
necessary to analyze thc combu.stion proccss, prcssurc levels ancl clircc-
tion, and apcx seal bchavior uncler thc cver-\rrlrving conclitions alortg thc
trochoidal path. 'f'hc courl>u.stion proccss in thc \Varrkcl cngirrc llcuc-
fits from one outstanding Aclvantagc in tcrnrs of scaling, cluc to tlrc fact
38
that gas velocity is far higher in the s1>ztrk 1>lug arca at thc nronrent of
ignition than in the othcr parts of thc conrbnstion chanrl>er. Rcn-
singer's tests showecl that the gas vclocitv ncar thc minor axis wils
Drawings from the apex seal carrier patent taken out by Goetzewerke.
Fig. I shou,s spring loaclirrg of the seal strip rvith a convex spring, a multiple of the flanre front vclocity. Irr his test engirrc, flanre front
without sprirrg-loacling of thc scal carrier. Irig. 2 i.s a cross scction of trat'cl took 1>llcc orrlv irr tlrc lcn«lirrq «lircc'tiorr. 'l'lr(' ul)lx'r' fllrrrrc f rtlrrt
thc silnre clcsigrr. Irig. 3 uscs il concnvc slrring to load an inclirred \\ras stationary; ignition occurrccl orrl,v u,lrcrr courprcsscrl gns \\,rts llrorrglrt
seal strip (rs shown in lrig. 4). In Irig. 5 and 6 clouble-concave
springs, both one-;>icce and two-piece, are shown in applications in-
cluding sprirrg-loading of thc seal carrier.
The loss,of somc frcsh lir iuto thc cxh:rust rlrca, against the clircc-
tion of /rtot rot:rtion, is lln u«lvant:tgc irt clisguisc. Tlrir air helps
purge thc flrcu of cxhnust glrscs thlt uright otlrerrvisc hnvc bceu
Irlanrc front forrnations at various points during the cornbustion rccirculatccl. Thc air itsclf «loc.s uot csc:r1:c through the exlurust 1:ort,
but is su,cpt aloug into thc conrirrg corrrprcssion phasc by the trail-
1>hasc arc shorvn in thc ccnter schcrnatic (looking directly at the
rotor facc). Thc skctch at left shou,s rotor positions before and after ing a1>ex, The curve on thc riglrt shows the progrcssivity of variations
thc monrcnt «rf ignitiorr. Thc cliagrarn at right illustrates gas velocify in chamber vohlrne.
at the mirror axis, clurirrg onc full rotor rcvolution. Einlossvorgong

nitlbr Goryorchwlndlgk.lt Vdumcnvcrloul

Enlorr rcH e rrt-

l§

\trlo'rt drr Gorgoschrnhdrghit ir dr o

Exrcntorunltr
kblnrn Achr
2?Oo- +
)Otbonstrllung |!r F lorvnr nt rontvürtou, (bor
rnd noch ZT drr Lodbohn
280 DEVELOPMENT DAIMLER-BENZ 28I
forward to the flame front. By the time the trailing apex seal reached advantages were claimecl for the new seal configuration. The first
the flame front, temperature and pressure in the combustion chamber stemmed from the fact that the strip could be slimmer and nrore flexi-
were considerably diminished. A leaking apex seal at the trailing end ble than was previously possible, with the result that slight distortions
did not show burns-it only resulted in a loss of performance. Ahead in the piston or the track could be more reaclily accommodated. The
of the spark plug, the combustion chamber walls were relatively cold. second related to the cold clearances between the sidewalls of the work-
Because the flame front was chasing the leading apex tlrroughout the irg chanrber and the ends of the seal strips and seal carriers. This system
combustion process, temperature ancl pressure both declined from their was not adopted, however, as results failed to warrant the additional
peak values reached earlier in the process. Since the apex seals oper- cost and complexity.
ated in a relatively .low temperature belt, Bensinger concluded they Nothirg has been found that has giverr superior results to the rela-
could be made of materials having low heat resistance, such as carbon tively thick apex seal strip located directly in a groove in the rotor,
or aluminum. spring-loaded raclially and provicled with slots along the bottom Ieading
The experimental seals were clesigned to provide fuJI, or almost full, edge to ensure sealing action by gas pressure, without delay followirg
sealing until 2.5 mm. wear allowecl the seals to escape from their pressure reversals. For instance, the C-l I I apex seal strips are 5 mrn.
grooves. Seals made with carbon materials showed wear characteristics thick and 7.5 mm. high. The trunnions are l0 mm. in diameter and 6
in the range of 0.002 mm.per hour (or l00 kilometers), which would mm. long, while the side seals are 1.5 mm. wide and 2.5 mm. deep.
give them a lifespan equal to more than 100,000 kilometers (rr about Current engines use cast-iron apex seals, but they are not considered a
62,500 miles) on the road. These wear characteristics failed to hold true final solution. Experiments with ceramic edges on the apex seals are
whenever fuel deposited on the working surface washed off the oil film. being conducted right now. Daimler-Benz continues its search for better
Frequent stopping and starting was a severe barrier to long seal life. apex seal materials, so that with further clevelopment they will not be
The problems were somewhat alleviated by the fact that in case of the limiting factor in engine life between overhauls, even under the
apex seal failure, such as when they stuck in their grooves or the seal worst operating conditions.
tips were too worn to provicle a proper seal, there was no harm in con-
tirrued operation of the engine. Starting difficulties resultecl, but the
engine woulcl not stall. Apex seal strip, corner seals, and seal spring, as used in the C-t I I
Some engines were testecl with a novel type of seal configuration de- engines . (Photo: Ludvigsen)
veloped and patented by Goetzewerke in Burscheicl. Instead of the
simple seal strip, a more contplex arrangement was inventecl. The strip
was not seated in a groove cut directly into the rotor, but in a channel-
section seal carrier that seated in the groove. The strip was spring-loaded
radia]ly away from the carrier, and the carrier from the rotor. Two main

uo{ItErsl E Nntr I ELtErL1 I5T$IFEO€R

Final apex seal configuration de-

UCHTTE6IENECXE
veloped by Daimler-Benz.

orclrItoLtEN rarl
otorlDou€lr FCOT

OCHISTR€fEN
 DAIMLER.BENZ 28)
effect. The stroke lirniting pins also could be made eccentric, so that
turning them would alter the stroke of the plunger during running.
The wick device was discarded in favor of a more advanced metering
system, using a special drip-feed oil pump that delivered minute quan-
tities of oil to sleeves in the intake ports. The sleeves had a cavity that
held an oil film over a large sector. The incoming air carried part of
the oil film in with it; not in the middle of the airstream where it woulcl
have formed part of the mixture and been wasted, but along the metal
surface from where it was depositecl on the apex seals. The cavities
prevented the oil from getting into the center of the airstream and
kept it close to the surface so that the apex seals swept the oil film with
them as they passed the intake port. The oil to fuel ratio was estirnated
to be I :150, thus the experimental C-l I I used no more oil than a piston
engine of similar output.

Intake port sleeve with oil filnr cavity, as used on the C-l ll engines.
(Photo: Ludvigsen)
Wick-type metering of lubricating oil into the intake port for apex seal lubrication,
was patented by Daimler-Renz in 1965.
'I
,i,,4

il

The apex seals depend on the presence of an oil filnr for satisfactory
operation. F-or a Iong time, NSU simply hacl rnixecl lubricating oil witfi
the gasolinc, so that the fresh charge would provicle the ,re..irrry seal
lubrication. Although Bensinger adoptecl this nrethocl for the earliest
test engines, it was unreliable and suffered from wicle variations in oil/
gasoline ratios. In acldition, it is unsuitable for diesel engines. Daimler-
,i

Benz began to study methods to mix accurately metered quaptities of 1

oil with the incoming charge. In a patented design, oil was fed from
a rnetering device to a metal-fabric wick projecting into the inlet port.
The amount of oil supplied was directly proportional to the speed of
the rotor. The metering device was a simple flow control unit housed ip .i

a shaft revolving at rotor speed. A short plunger contained in a diametral

hole in the shaft had its travel limited by small-diameter pins fitted into ' ,i

drillings that intersected the first bore. The shaft housing 5ad two
diametrically opposed ports-one connected to the oil pump and t5e l.

other to the wick. As the diametral hole in the shaft swept past the
port leading fronr the pump, the plunger was forced away iowärds t5e
outlet port, After an irrterval of 180 degrees of rotation the plunger was
forced in the other direction, to allow a small quantity of oil through {'
the outlet port. This action was repeated trvice per revolution, and the I't
,r. .

qtlantity of oil clelivered at eaclt stroke was dependent solely on the

travel and diameter of the plunger-fluctuations in oil pressure had no
 DAIMLER.BENZ 285

In the course of his devcloprncnt rvork on the rotor cooling systenr,

Bensinger founcl that agitation of the oil within the rotor resulted in
Pattern of oil flow inside the foaming. Foaming was undesirable because it recluced the oil's cooling
C-l I I rotor. effect. It was also found that some of the oil, particularly in the apex
cavities, did not circulate ancl therefore was incapable of providing the
intendecl cooling. One proposed solution was to block off the apex
cavities from thc oil circulation systcnr ur«l fill thcnr rvith soclittttt-rt
high-concluctivity rnaterial Iong used in the exhaust valve stcnrs of
Mercedes-Benz piston engines to increase heat clissipltiort. This has sub-
scqucntly bccn founcl to bc unncc'cssrlry.
The C-l1l engines usc a rotor cooling systern cluitc similar to that
devcloped by Curtiss-Wright for the RC2-60. Oil sealing of the rotor
bearing was a clifHcult problcnr for a long tirttc. A portion of thc siltnc
clil tlrat cools thc rotor hacl to lubricatc thc rotor bcarittg. Bccause
Cutaway sketch showing oil flow Dairnlcr-Bcnz clrosc lllairr bc:urings rrttlrcr llutrt lurll or rollcr bc:ltrirrgs
towards the rotor face and along for thc output shaft, llcnring lulrricution wts Iriglrly criticrrl. Au utnl>lc
the direction of rotation, until
arrestedat the leading apex and rccirculatory oil supply rrrrrlcr prcssurc u'oulcl bc iclcnl, llut tlrcrc is nln'rtys
drained back to the eccentric
bearing.

53
\,,,\
6
F,g.2
>19
I >7
:/4
:8
u- 13 I

o Schmctzstitl
300
o Schmclzrtilt nicht ousgeschmolzen
t PcJ
KASTENKOLB
OHNE öUNÜCXUAUF -2t
200

260

240

Patent drawings covering the

220 drainage of cooling oil from the
rotor, by opening outlets when
centrifugal loads force oil in their
200 direction. A scoop is used for
trapping excess oil.
r80
 DAIMLER-BENZ 287

Fig.2 the risk of excess oil leaking through to the combustion chambers. Oil
would accumulate behincl the seals ancl remain pressed against the
seals by centrifugal forces of increasing magnitude at increasecl r.p.nr.,
until the oil rvoulcl force its way through to thc sicles of the rvorking
chambers. Daimler-Benz found a \/ery elegant solution by fornrirrg a
charnber insicle the rotor. The rotor turns rnore slowll, than the ec-
centric shaft, centrifugal forccs act in the clirection of the ecccntric, and
the oil contairted in the rotor is kcpt in constant circulation. Oil inflow
ancl outflorv arc controllcd by ccntrifugal forcc action corubined with a
new dual rotor bearing seal. As thc rotor rcvolvcs, oil can florv into the
1 rotor cavity towarcls one apex fronr both sicles of the ecccntric. It can
only florv out on the oppositc sicle (i.e., the rotor face o1>posite the apex
Drawings from Daimler-Benz's that is receiving the oil). If too large a proportion of the hot oil fronr
patent for sodium-filled holes in
thc rotor casting, behind each the rotor cavity fails to return to the mainshaft bearing, ancl insteacl
apex. recirculates back into the rotor, A shield on the eccentric catches it and
throrvs it into the outlets. Because the shield revolves faster than the
rotor, it easily scoops up excess oil. This is backed up by a dual rotor
bearing seal, u,hich also has to function as a gas seal to keep combustion
fumes from contaminating the lubricatirrg and cooling oil. These seals
are fittecl on both sides of the eccentrics that act as rotor bearings. Each
has an inner ring ancl an outcr ring. The oil that passes the inner one
is then containecl in a chaml>er fomrecl bv raclial ribs betlecn the tn,o
seals. The ribs have a slight clearance from the resilient lip of the outer
Oil sealing method, äs applied ring. This lip extends radially inrvard and effectivcly prevents the oil
to the C-l I I engines. fronr escaping into the combustion space, whilc stayirrg at surface Jevel
aud providing lubrication whcre nccclccl.
I)1,the usclof pcriphcral intirkc ports, f)airnlcr-Bcnz. is ablc to obtrrin
yolunletric efljciency l;etq,ccp ll0 arrcl ll5%, at 2,000 r.p.rn. 'Ilris is
possible bccarrsc gas flon, continucs cvcu uft'cr t]re onsct of compres-
sion, ancl gas vclocity nlr1, cvcr) irtcrcasc a.s t]rc port arca is rcstrictcd
by the passing of thc rotor's trailing apcx. Gas flow is split by this
rlpex as it travels rlcross thc port arcA, ancl fresh gns bcgins to fill thc
next u,orking charnl>er, which is not )'et halfw,ay tlrrough its cxhaust
phase. This reversecl florv helps exlraust gas evacuation and cloes not
cause any notablc loss of fresh nrixture, sinrply because the charnber
aclvances so quickh' into the intake arrcl con'rprcssion arcas. Tests with
side intake ports shon,ecl that thcy coulcl not give cqual volumetric effi-
ciency for tlo reasons. ITirst, the overlap belween intake and exhaust
was missing and seconcl, the absolute tinre available for intake was
smaller.
Detail of the oil seal configura- The first Wankel engines dcsigned ancl built at I)ainrler-Bcttz uscd
tiorr uscd on the C-l I I engirres. carburetors. There were multiple reasons why fuel irrjection was evcntu-
 288 DEVELOPMENT .DAIMLER.BENZ 289

ally adopted, and, again, the change was only made after exhaustive oxygen-it is this ignition lag that produces a bucking sensatioD, com-
examination of the full combustion process. Combustion is, of course, monly called the "snatch" phenomenon. Fuel injection solves this prob-
slower in a Wankel errgine than in a piston engine. This is unfavorable Iem, because fuel delivery is shut off whenever the throttle is closed.
in terms of thermal efficiency but does result in favorable part-load Only when the throttle is re-openecl suddenly will abnormal combustion
operational characteristics. T'he use of twin spark plugs to shorten flame take place. Similar problerns occur during idling, although they are
front travel distance was rejected by Daimler-Benz, first of all because easily overcome because of the narrow r.p.Itl. band.
it didn't help much ancl second because of the faster pressure rise and
related problems. The engineers chose to use one surface-gap spark plug June 14, 1966 w. spRtNGER ErAL 3345r?BS
per chanrber. The Wankel enginc clocs uot develop the same compres- RO'Af,T.IISTOT IHTEBNAL COIBUSTIOII EIIOIf,E
sion pressures as does a reciprocating piston engine because more heat Plhd llarcb 19, fg6l,
is given off to the surfaces of the combustion chamber. This disadvan-
tage does, however, diminish as rotational speed rises. Fuel consump-
tion of the Wankel engine tends to be higher than that of piston en-
gines of similar output, particularly at part load and low r.p.m. Daimler-
Benz ultimately chose fuel in jection for the C-l I I for these reasons, in
addition to the following particular disadvantage of carburetors as used
on Wankel engines. Condensed fuel droplets in the intake manifold of
a piston engine usually are vaporized ancl mixed with the air due to the
warming effect of the hot valve head and the fluctuating gas velocity in
the induction system. In the Wankel engine, however, everything is
cold and the areas are so Iarge that gas velocity tends to be low. With i

side intake ports, some turbulence can be created, but with disappoint-
ing results. Fuel iniection overcomes all these objections. On a closed FIG I I
I

throttle, the carburetor-equipped Wankel engine continues to aspirate

!
t

fresh mixture and, because the ports are ahvays open, the engine runs i

on exhaust gäs, or unburned gas mixed with fresh air. Depending on

r.p.m., there will be a time lag betrveen the closing of the throttle and
the first ignition of these gases, rvhich have had time to mix with fresh I

The use of direct fuel iniection modifies the operating cycle of the \Vankel engine
I

Ir
to some extent. Irrjection begins rvell before the intake port is closed off and the
beginning of compression. it
l:
lr
llow thc Rotary Plrton Englnc Worlr
ljIr
Ir
t;
!i

lr

+ + 4 4 Il

F F F F
t'
Srvirl injection in the intake port, as patented by Springer and Lamm. ir
l-a Arrrtlr A-tO \brÜlrllrE tt-12 A.:.bß.n
I
I

i
i
Jrurc 9, 1964 F. K. H. NALLINGER ETAL
ROTARY PISTON ENGINE
3,136,302 DATMLER-BENZ z9r

Pllcd larch 20, 196l Several important patents were taken out by Daimler-B enz cngittcers
in the course of developing an advarrced gasoline injectiorr systcnr for
the Wankel engine. I{einz Lamm ancl Willi Springcr of Dairnlcr-Bcnr
AG received German patent D 38,450 on March 23, 1962, covcring port

ffi
iniection. The method of injection was to clircct thc iet ag^inst the
cavit), in the rotor face. They claimecl this woulcl givc thc effcct of dircct
iniection without any loss of conrprcssion whcrt the raclial apcx scals
passed the nozzle bores in the working surface. Iiuel was in jected into
the area of the highest air velocity, rvhich promoted better mixing
through turbulence. The intake port diameter was narrower at the rvork-
ing surface than back at the elbow wherc the in jectiorr nozzle was posi-
tioned. The spray was so directecl that the fuel did not hit the port wall,
but progressed into the working chanrbcr without chrtrtging course.
Lamm and Springer also patentccl a hclically wourrd shect nrctal guide
to fit inside the intake port. This woulcl have the effcct of starting tur-
bulence well back in the port area while still confirring the spread of the
injection spray to a small area on the rotor face. This invention was
intended to improve the mixture preparation, to proviclc adclitional cool-
I
ing of the rotor faces, ancl allow stratification of the clirtrge within the
combustion space. Howcvcr, port-typc injcction was soorl rejcctccl iu

N
favor of direct injection, which pron'risecl nrore controllable stratification
and combustion characteristics.
-'2''-'
\t5 Nallinger and Bensinger took out U.S. patent number 3,136,702 on
./' |une 9, 1964, concerning fuel injection ancl the stratified chargc. The
( injection nozzle was described as iniecting in a circunrferential clirection
tr/ of the rotor directly into the working chamber. The patent covercd
( \l :
yi
methods to ensure aclvantageous stratification of both rich ancl Iean
mixtures within the combustion space cluring part-loacl operatiorr of the
.\\\
\
\:
'tta
>J l, engine. Stratification was achieved by the nozzle arnurgcment and posi-
tioning. The basic idea u,as to usc two spark plugs pcr chanrber. The
rich nrixture was to be ignitecl b), the plug-thc learr portion of the
mixture \r,as to be ignited by the heat in the previously ignitcd mixture.
INyENTORS The first spark plug then coulcl bc placecl only rvith regarcl to initial
ignition to provide proper and safe ignitiorr of rich mixtures such as
would exist during full-throttle operation. The seconcl plug would be
positioned so as to take care of end gas-it woulcl be fired fairly late in
FIG.2
the combustion phase. The injection nozzle was to bc so inclinecl that
its spray would be aimed at tlre plugs in the conrbustion space, u,here
compression and gas transfer take place.
Nallinger ancl Bensinger reasonecl that the rich fucl nrixture rvould
Stratificd chargc fucl injcction ancl d_u3l- ignition, according
to a
fill the heacl of the combustion spacc, right up against thc plugs, at thc
patcnt by Nallingcr arrd llcnsinger in t961.
moment of firing. The remaindcr of thc chanrbcr then would contairr
 March 23, 1965 H. O. SCHERENBERG 311741466 March 23, 1965 H. O. SCHERENBERG 311741466
INTEßNAL COUBUSTION ENCINE INTEßNAL COUBUSTION ENGINE
tllod Flb. 2, 196l 2 Shcets-Shcct z Flled Feb. 2, 196l 2 ShectrShoct I

FIG 9 FIG IO FIG II

FIG 12 FIG 14

FIG I3 FIG I5
?ßQel
l2
t,
'!-"
t\ \

FI G6

FIG 8
Ir{vEirToR
II{VENTOR
HAilS O. SCHERET{BERG
HAIIS O SCHERTNEERG

FIG I6

Various suggestions by Scherenberg for use of dual iniection nozzles in each cham- Scherenberg also proposed stratified-charge in jection, with single iniectors and a
ber. spark plug positioned iust after the minor axrs.
79+ DEVELOPMENT DATMLER-BENZ 295

a relatively lean mixture, to be ignited by the burning of the rich mix- 27 cubic inches (450 cc.) chamber displacement designed in 1965 and
ture ahead of it. Rotor rotation would, of course, bring the lean mixture 1966. This was the first Mercecles-Benz Wankel engine designed sPe-
to the area where combustion had been started. They worked on the cifically with passen gü car installation in mind. A three-rotor version
assumption that the engine would operate without turbulence in the followäd almost immediately. Bensinger wanted to explore the multi'
combustion space. They also envisaged the flame front travelling against rotor concept because more rotors mean more firings Per output shaft
rotor rotatioD, ä concept which later tests proved to be wrong. revolution, which results in reduced torque fluctuations. That, in turn,
Progress came fast when Hans Scherenberg began to take an active cuts noise and vibration. Beyond that, it was soon found in testing that
interest in the development of the Wankel engine and applied his inven- t5e hexane content of the exhaust gas was greater wlten the clisplace-
tive talent to the fuel iniection system. His work was directed towards ment of each combustion chamber was greater. With more rotors, each
high-speed operation with minimized fuel losses. Another main goal chamber is smaller, and exhaust emissiorrs are reduced.
was better partJoad operation. The result was a multi-nozzle fuel iniec-
tion system, which would render stratification possible and improve
operating characteristics. Scherenberg received U.S. patent number
3,171,466 on March 23, 1965. His attention had focused on the use of evation of the three-rotor C-lll engine.

two nozzles per chamber, arranged in various configurations to produce

best results under a variety of conditions. Scherenberg was first attracted
to fuel iniection because it would not involve the flow losses of fresh
fuel that he regarded as inevitable with carburetors. The twin-injector
idea came to him when he looked at the elongated shape of the com-
bustion chamber. He pointed out that twin nozzles could be arranged
in such a fashion that satisfactory mixtures could be obtained regardless
of the unfavorable shape of the combustion space. Depending on spa-
tial requirements, the nozzles could be positioned either one behind
the other, as viewed in the circumferential direction of rotor rotation,
or axially opposite each other in the sidewalls.
A staggered arrangement was recommended to achieve the best dis-
tribution of fuel throughout the entire working air. It was felt that if
one or both of the nozzles were arranged in the end walls, mixing would
be good because the jets from the individual nozzles would cross each
other or be directed against each other so as to impinge against each
other. The use of two nozzles could also improve the operating charac-
teristics during partial loads. In order to ensure proper atomization at
the nozzle.s under part-throttle conditions, Scherenberg ProPosed to shut
off one nozzle. Then the relatively small fuel quantity to be iniected
would have to pass through a single nozzle and thereby maintain pres-
sure and stratification characteristics. Stratification of the charge during
partial load operation could be achieved using two iniectors by so arrang-
ing and constructirrg the nozzles that the one which is turned off during
part-throttle running iniects the fuel charge into that region of working
air which moves closest to the spark plug(s). This would make it possible
to operate the engine on very lean air/fuel ratios.
The geometry of the C-t I I engine evolved fronr a twin-rotor unit with
 t DAIMLER.BENZ 297
I
Irr 1967, a ncw thrcc-rot«rr enginc wils «lcsigrrcd to nrcct U.S. exlutust
I emission control stanclarcls. It had 560 cc. cli.splacerncnt pcr chanrber,
I and its chief geonretrical features were:
Radius 103 mm.
Eccentricity l5 mm.
I Width 70 mm.
R/e ratio 6.87: I
R+e ll8mm.
This experimental unit was the direct basis for the C-l I I clesign, which
has widcr rotors llut retairrs thc sjanrc raclius ancl ccce ntricity. Rotor
width went from 70 mrn. to 75 nrrn., and displacernent per chanrber rose
to 600 cc. The transistorized ignition system uses one surface gap spark
plug pcr chanrbcr, which gives a strorrg spark at extrenrely high r.p.nl.
and at very high operatirg temperatures, even with deposits on the
engine. electrodes. The concentric electrode plugs were developed by Beru for
Dainrler-Benz. They have a heat value of 320 and are especially resistant
to high temperature and are able to fire despite fouling or lead deposits.
The front cover on the Gl I I engine hides the rotor shaft sprocket
I Iorsepower,
and chain drive to the accessories. On the other side, three sprockets are
torque, and mean effective pressure cuwes for the three-
rotor C-lll engine. stacked, the chain threading its way on a slalom course through from
top to bottom. The top sprocket is carried on a shaft that clrives both
N. tPS OINI
eoo the fuel iniection pump and the lubrication oil metering pump. The
280
200
210 Comparison of three and four-rotor C-lll engines.

220
200
t80
t00
ra0
t20
too
sol
I

.ol
.ol
,ol lmkpt
30
20
22
t0

o rooo 20OO 30OO .loo0 60CO OOOO ZOOO

n lU/minl
 DAIMLER-BENZ 299

a geu which has a worm gear attachment to the

same shaft also carries
ignition distributor shaft, which is mounted vertically. The center
sprocket drives the shaft for the cooling oil punlp, and the bottom
sprocket drives the water punrp shaft. The nrain oil pump supl>lies oil for
rotor cooling, shaft bearings ancl bearing seals.
The Bosch mechanical injection pump is driven by a single chain and
runs at one-half mainshaft speed. The iniection pump is fed by two
electric fuel pumps mounted outsicle the tank. Exhaust noise and tem-
perature created a muffier design problem. The exhaust gas flow is noisy
simply because the exhaust port opens sudclenly. The C-l I I engine nor-
mally is run on 95-100 octane European premium gasoline. Tests with
regular gasoline have shown that the engine is renrarkably uncritical of
fuel grade. Drop in performance is insignificant, and no cases of ab-
normal combustion have been observed.
The big question today is: will f)aimler-Bcnz procluce the C-lll en-
gines or derivatives of them in quantity? Uhlenhaut said in September,
1969, that he did not want to go on recorcl as saying that all problems
of the Wankel engine had been solvecl. f)evelopment work at Daimler-
Ignition side of the four-rotor C-lll engine.
End cover for the C-l ll engine. (Plroto: Ludvigsen)

Iniection side of the four-rotor C-lll engine.

 I
t
.T

t
I
t
I

--#fl I
r---ffi
k'g\
I
T

Author's face registers amazement wherr confronted with the torn'down three-rotor
C-lll. Flousingi side walls, and end covers are pla_ced in their proper relative posi-
tions, but sepaiated by wooden blocks. (Photo: Ludvigsen)

Mainshafts for the three-rotor C- l t I engine. Tlre eccentrics are not equidistant,
while on the four-rotor unit they are. (Photo: Ludtigsen)

Accessory drive at the front of the C-lll engine. (Photo: Ludvigsen)

 r3

Citroän

h' l,rAv IIAvu s'l'RucK yoII irr rcvicw-

ing the licensc agrecnrents nradc by NSU for thc Wankcl crtgirre rights
that therc was no urcnti«rn of Citroön, ancl yct tlrc conlpillly is lluilding
Rotor installccl in its trochoiclal chamber on the C-l I I engine. (Photo: Ludvigsen) the Wankel-p«lwcrcd M-35 (coverccl irr a latcr chapter). Citroön fcluncl
a different \r'Ay to gain expericnce rvith the \\/ankcl engine. In explora-
tory talks with NSU in 1962 ancl 1963, the Gernran conlpany proved
Benz Iracl been successful in that the engine was far, far better than it willing to consider the formation of ioint subsidiaries in partnership with
rvas only a few years before. But only when he (Uhlenhaut) had a fleet Citroön to be active in areas that pronrised to bc mutually bcneficial. In
of a few thousancl cars in service rvith Wankel engines, running satis- |une, 1964, representatir,'cs of Citroön ancl NSU nret in Geneva to forrn
factorily, woulcJ he fecl that all problems were, in effect, solved. One is a local company-Societe d'litucle Comobil. Conrobil was capitirlizccl at
1le\,,er sure until production is a fact. The worst problem today is ex- $300,000, half of which was held by NSU ancl half by Citroön. The
haust emissiorrs-there is no imnrediate ancl satisfactory solution to that. purpose of the company was to prepare, and evaluate the marketing of,
Uhlenhaut thinks the Wankel errgine can be produced more cheaply an automobile equipped rvith an NSU \\/ankel rotating combustion
tharr conventional piston engines, on a dollar per horsepower basis, when engine. The technical and commercial staffs of the two companies went
it i.s "fully cleveloped." The machining of the epitrochoidal working to work to complete the assignment,
snrface was a big heirclache, but the time involvecl has now been reduced It rvas planned that this work would take a pcriod of years. Conrobil
to an cconomically acceptable level. Daimler-B enz C-l I I test engines was not only to make proposals for future Wankel-powered cars, but also
run ovcr 60,000 miles or the equivalent between overhauls, which means was to produce designs for one such vehicle. On l\4arch )0, 1965, Como-
tlat durability is fully comparable with piston engines. The C-l1I en- bil began a program for the preparation of manufacturing and sales for
gines now are running in a variety of Mercedes-Benz passenger cars and a new car equipped with an NSU Wankel engine. Accorcling to the
trucks, around the clock, to amass tlre data needed for Dr . Zaltn and his understanding betu,een Citroön and NSU, the car was to be made lr)'
colleagues to iustify quantity procluction of Wankel-porvered cars. How Citroön in France, except for the power unit which woulcl be nlanu-
soon? Iiactory spokesrnen clecline to mention any number of years or factured by NSU in Gernran),. Assenrbly plants wcre to operate in both
months, but my impression from numerous interviews is that a Wankel- Iirance ancl Gentrany.
powerecl I\4ercedes-Benz production car can be no more than two years The next phase of collaboration between Citroön arrd NSU was the
away. creation of another joint subsidiary, Conrotor SA, witlt hcaclquarters in
304 DEVELOPMENT CITROEN 305

Luxembourg. Its purpose was much more ambitious than that of Como- cided to increase its capital to $ I ,620,000 ancl in Iicbruary, 1969, a
bil. Conrotor actually was to manufacture and distribute Wankel engines further increase was rnacle, raising ca1>italization to $2,020,000. In Iuly,
for all applications ancl also make ancl sell all the accessories required for 1969, the board agreed to double its capitalization-to $4,040,000.
these purposes. The Compagnie Europeenne de Construction de Comotor was at that time buying a future engine plant site in the Saar,
Moteurs Automobilcs (Comotor S.A.) was organized on Mry 9, 1967 which was scheduled for completion by the end of 1970. At the end of
and was capitalizecl at $1,000,000. Comotor was authorized by its found- 1969, the Citroön M-15 was announced. The vehicle is built by Citroön,
ers to participate in other corporations to conclude any commercial, and the engine is made by NSU in Neckarsulrn.
inclustrial, technical or financial transactions, inclucling real and personal This NSU engine used the latest technology developed during the first
property operations, directly or indirectly related to the purpose of its two years of Ro-80 production. For instance, the working surface in the
charter. The six menrbers of Comotor's board of directors are Piene old single rotor KKM- 50?. engine was chrome plated, but the working
Bercot of Citroön, Gerd Stieler von Heydekampf of NSU, M. A. Bun- surface for the M-35 engine has a special nickel-silicon alloy coating
ford, A. Noel, M. Geoffroy, and I. I. Baumann. During its first meet- known under the trade narne Einisil. The NSU Spider engine often
ing, the board elected Pierre Bercot as its chairman and president of developed cracks around the spark plug holes, but on the M-35 engine
Comotor S.A. On October 22, 1968, Comotor's board of directors de- the housing has a copper sleeve into which the spark plugs are screwed.
TIre KKM-502 carbon seals had a short life, which led to high fuel and
oil consumption. The seals in the M-35 engine are made frorn cast-irou
I?lyrvheel cncl of the M-35 cngine, showing the two spark plugs.
and have provcd to have good clurability. '['hc corner segnrents of thc
apex seals have a new configuration. The curved side seals, pressed
against the sidewalls by spriDBS, are no longer a problem. T'he oil tem-

Front view of the M-35 engine, complete with accessories.

 I
306 DEVELOPMENT CITROIiN )07

pcrature irr the NSU Spider engine norurally hovered around the 360oF. loacls.This conclition comes about because of exltaust gas backpressurc
Ievel, which put great stress on the bearings. Through the use of a heat when the throttle is closed. (Jnclcr such corrclitions there is of ten a
exchangcr, the normal temperature has been Iowerecl to around 250oF. partial vacusm in one or two of the working chambers, and only small
It was not until fune, 1969, that NSU installed the M-35 type heat quantities of ex|aust gas arc let in. As a result, large cluantitics of ex-
exclrangcrs orl tlrc Ro-80. In th is system, the oil runs in two circuits. Saust gas are mixecl *ittr snrall quantitie-s of frcsh tnixture-thc result-
Thc first feecls the rotor bearings and cools the rotor internally; the oil ing *i*ture will not burn. As the rctor follows its course, this mixture
filter ancl the heat exchanger are incorporated in this circuit. The second flows out of the exhaust port, at which point it finds little or no back-
circuit takes oil fronr the sump to lubricate the seal tips and mixes oil pressure. With more freih mixture added as the chamber Passes the
with the fuel through a special metering clevice built into the fuel pump. intake port, the mixture finally becomes combustible and is ignited after
Lubricating the scals by this nrethod means that a certain amount of compression. The combustion exerts a force on the rotor that is detected
oil consunrption is inevitable, but it is generally no higher than one as ',r-ocking." Because there are three ignitions per rotor revolution, the
qlrart per 600-650 miles. engine loses three power phases every second revolution. Side ports
NSU fourrcl out that both single- and twin-rotor Wankel engines had rvould eliminate th; probläm, but NSU relies on peripheral po-rts be-
cause of the higher uälun,etric efficiency. On the Ro-80, the bucking
is
a problem rvith bucking or "snatching" at low speeds and under light
cus5ionecl by tlie hydraulic torque converter. Witll the M-35's convcn-
tional clutcli ancl transnrission, the technically best solution was to go
Iixploclccl vicrv of the I\'f-35 engine.
to fuel-inicction.

Vicw of t[c rotor urotion irr tlrc M- j5 crrgrrrc.

a
 CITROIiN 309

One enging with peripheral ports that does not have this problem is
the C-l I l. As has been nrcntionecl, Mercecles-Benz completcly shuts off
the fuel supply wherr the throttle is closccl. But injection is expensive,
itnd Citroün rv:rntccl sonrcthing sinrplcr. Citroörr has triccl t«l ovcrcornc
the problem with a reviscd ignition svstem. I{owever, this does not
entirely eliminate bucking, although it is not noticeable insicle the car.
The ignition system includes two coils, two breakers, and two spark
plugs to ensure continuecl operation in case of failure in any one unit.
A special vacuum advance mechanism also is incorporated, which
severely retards the spark on the overrun. The resulting Iate firing effec-
tively Iimits the effects of the conditions that were such a problem with
the original NSU Spider engine.

I\,f-35 rotor, complcte u,ith seals.

M'35 rotor, with eccentric ancl mainshaft in place.

 OTHEIT WANKTL LICENSE FIOLDIIIIS 31I

AIfa llomco, t[c farnous Italiarr builclcr of high-pcrformance cars,

ß4 was among t5e first to recognizc the potential of the Wankel cngine.
As early,r t95g, it was rumored that Alfa \\/rls rtegotiating a take-over
of NSU. Alt|ough tftis never transpired (V«llks\ /agcll evetttualll' bought
NSU), Alfa was among the first to conclude a liccnse agrecmetrt. Set'eral
prototype engine, *.rä built, having 500 cc. chaurber displaceurent. Both
iingle--änd hvin-rotor engines werc investigated, and stuclies were con-
cluctecl on porting, basic geometry, ancl combustion, tls well as oll sealing
ald lubrication. No test reports havc been nracle at'tlilablc, Ilo\\'et'er, so
no real idea of their progrcss in the field can be cleterminecl. If Alfa
Wankel Romeo were to rnake a wholesale corrversion to Wankel engines in its
cars at some future clate, they coulcl revolutionize the Italiatt autonrobilc
Holders inclustry, now clomirratecl by lTiat. Iiiat's attituclc towards tlte Warrkel
engine was skeptical until reports frorn Citroön were studiccl in Torino,
but they now seem prepared to admit the enginc's potential. Alfa
or r,r o'r'Irlrn wankel licensc holcl- Ilonreo is locatccl irr t'lrc riverl city of Milarro. ßtrt cl iffcrctrces clo not
ers, antollg w]rich are N4.4.N., Krupp, IJanomag-I-IenscSel, encl there. Iriat is strictly il 1>rivatc cntcrPrisc (ptrblicll' orvttccl) u'hilc
Deutz,
Porsclte, AIfa ltonreo,_Rolls-Royce, Perkins, and f ichtel Alfa Romeo is part of a govcrnmetrt-orvttccl sttl>sicliary.
& Sachs, only
one has clottc any rcally serious work in cleveloping the powerplant
for Of t[c two British (.ornpanics thnt havc couclttctc«l \A/arrkcl crtgittc
commercial application-Iiichtel & sachs. researcS, only Rolls-Royce has concluctcrl aIl itttcttsivc ittt'cstigrttion.
I-Ianornrlg hacl, in 1961, ä Wankel test program, but the Perkins, t|e worlcl's Iargest builcler of cliescl eugil'lcs, has bccn n'lorc coll-
entire project
has not anlotllrted to vcrv nruch. At any rate, Hanonlag cernecl with the developrnent of their differential diesel engine, rvl'rile for
is now closely
allied rvith Dainrler-Il cnz, ancl if Daimler-Benz ever deäides to go
into years Rolls-Royce has been engaged in research ancl cleveJopnrent of
1>rocluction of a \Vankel-1:ou,crecl truck they may well utilize Hanämag,s diesel and multi-fuel versions of the Wankel. The reason llolls-Royce
Iiccnse to c]o so. clecided to obtain a Wankel license is that the conrpanl' had the prontisc
- KruPp, the orgarrization that controls, among other thipgs, Europe,s
Iargest iron arrcl stecl works, was at one time interested
of a research contract for a diescl-fuel rotating contbustiorr cnginc fronl
in dleveloping a the Unitecl Kingclom N{inistry of Defence. Britain's military was look-
Wankel errgine in cliesel fornt for its line of trucks. Krupp has sirce ing for a new powerplant for tanks and armorecl r'ehicles that rvoulcl use
solcl
its truck intcrests to Dainrler-B enz, ancl although it still retains
its Wan- Iess flammable fuel than high-octane gasolinc. The contract \t/es dtrll'
kel ]iccnse llo research ancl clevelopment work is now taking place. signed, and the Rolls-Ro),ce techrricians rvent to u,ork. The technical
I\4'A'N., äDother German truck manufacturer, was also interested details of their experirnental engines have been classified information
in t5e
clicsel Warl:1. Iiarly tests were, however, disappointing tuncler the terms of British governnrent contracts. The main design
apd researcS,
although still in progress, is not intensive. features of the prototype engine have leakecl out, Itolevcr.
Ilotlr Porsche anc] Dcutz have not done too much work on the Wan- The Rolls-Royce unit cliffers from all otlrer Wankel engines in that it
kel. Porschc has hclcl a license since March 2, 1965, but works on the conrpression-ignitiou 1>rinciple. Curtiss-Wright's hear')'-fttcl
so far 5as clonc
Iittle l>ut evaluatc existirrg engines to determine their suitability
for errgines still rely on spark ignition, but l(olls-llo1,ss clccidccl to rlisc
variotts aPPlications. It is interesting to note that neither porsche compression to the point rvlrere an electricallv triggered spark no longcr
nor any
of its cngirtecring staff ltas taken out any patents ip connection with rvoulcl be required. This, of course, became something of a geometrical
t5e
\\rankcl. No clata olr Deutz test enginer irrr been made available problem in that the maximum conrpression ratio is dict:rtccl bt' tlte sizc
up to
this point, btrt it is w'orth rroting that the company has the of the rotor and its ecccntricitr, (R/e ratio ancl c). A rotor uncl conrbus-
industrial
'strcngth llecessary to l>ecome a maior factor in \Vankel procluction tiop chambcr configuration tltat pernrits a high ctt«luglt conlpre.ssitttr
sltotrlcl a firnt decision I>c macle to produce the engine. ratio to ignite the air-fuel rnixture tends to have very poor burnirrg char-
I
3IZ DEVELoPMENT OTI]ER WANKEL LICENSII I_IOLDERS 313

_study of all possibilities led Rolls-Royce to decide on a

acteristics.
{
stlperclrarged Wankel engine, using a Wankel-type compressor.
the spark plug would be normally-a few degrees before the minor axis.
The combustion process is similar to that obtaincd in hea1,y-fuel Curtiss-
Acttralty, compressor is bigger than the.ngin.. It här r larger rotor Wright Wankel engines, but sornething new is aclclcd. When the burtrccl
_tlre
ancl Iargcr cllanrber volttme. The compressor ]rai an outlet port"near gas leaves the enginc through the cxhaust port, it is rotrted back into the
its
minor axis that fcccls into the engine's intake port, ancl t6e air is comprcssor, entcring through il separatc port, to corrrl>lctc its cxptllt.sion
furt5er
colnpre.ssccl. Iiuel is iniectecl from a nozzle pläce«l approximately
wSere and cornbustion against thc contpre.ssor rotor. The conlpressor is thtrs
both gear-driven arrd gas-clriven.
O1>era rc Rolls-Royce engine. The r The Rolls-Royce Wankel engine is saicl to havc a pcak operttting speccl
mtio the sarnc clircctiän. I _ In
clr:unl rcl:rry cornprcssion charnber. of 4,400 r.p.m., which is considcrably below that of sPark'igrrition Wan-
l>cr. I froi,, .,,gine to conrprcssor kel engines, but about twice as fast as reciprocating piston-tyPe diesel
harrst
cngincs. Thc Wankel is also rc1>ortccl to be onc-qttnrtcr the sizc of :l
piston-type clicscl crrgirrc of conrparalllc p«lu,cr-rltrcl firr Iightcr. It is llso
half the size of a gas turbine of the sanre power, says onc report.
Irichtel & Sachs A.G. i.s GcrmAnv'.s lrrrgest pr«lclucer of small tu'o-stroke
errgincs for outcloor po\l/cr c(luillrrrcrrt, clutclrcs uttrl ltttttlttuttic t rlttts'
missions, ancl ball and roller bcarings. It was also olre «lf tlre first Gcrntan
irrtlrrstrics to lluy l Wlrrrkr:l c:rrgirrc liccrrsc. ''l 'ltis llßr(:ctltcttt c(l\'(:r('tl il
range of power units far srnaller in size than any consiclcred lly the otlter
license takers. The Fichtel & Sachs Wankel engines are too small to be
consiclcrecl for passcngcr cilr appliclrtions, but thcv arc usecl in some
snowmobiles and all terrain vehicles (ATV's).
Procluction of the Fichtcl & Sachs single-rotor KM37 engine began in
1965. The design contrastccl with NSU practicc irt sevcral arcns, notaltll'
in cooling of the housing ancl rotor. Ttre housing was air-coolecl ancl thc
rotor had no provision for oil cooling. Iiichtel & Sachs relicd on nrixing

Dimensional sketches of the Rolls-Royce 2-R6 engine.

653mm

I
3r4 DEVELOPMENT 8I EP -o
60+
oil with the gasoline in normal two-stroke fashion for apex seal lubrica- / 50
o.
lrt
tion-the spccifiecl oil quantity was ?% by volume. The KM37 com- -
parecl most favorably rvith conventional single-cylinder, two- and four- // {o@
stroke piston engines within the same performance range. The KM37
/
specifications were: /
G,
Displacement 108 cc. trJ

Compression ratio 8.5: I o

=
o-
Power output 6.5 horsepower trJ l.{
U'
at 5,500 r.p.rt. E,
o G
Weight 30.5 lbs. - \ / 1.2
T

\o-
d-

X
T'he installation dirnensions were 16., inches Iong by 8.66 inches wide
by 11 .22 inches high. This small size was directly attributable to the air- J
cooling ancl thc abscnce of an oil cooler. The KM37 unit was ertremely t l.O I
()
smooth, ancl the vibrations were estimatecl to be of the same frequency /\ 83 FC IL
o
range as those of a 200 cc. single-cylinder two-stroke unit. The single-
rotor Wankel engine is, of course, the equivalent of a twin-cylinder / .E

four-stroke cycle engine, and the resulting torque fluctuations are con- /
siclerably lorver than those of single-cylinder piston engines tuned to
o
deliver comparable specific power. The vibration amplitudes and the OOO 2000 5000 {000
resulting mass accelerations and inertia forces in the KM37 were only
I

CRANKSHAFT SPEED - RPM

15% of those for a 200 cc. two-stroke engine.
Iingines u,ith slightly greater displacement soon followed. They were Irerfor,rance cur'es for t[e KM.37 engine.
the 9.8 cubic inch (160 cc.) KM48, and the I8.5 cubic inch (300 cc.)
KM9l4. But before discussing the existing engines in detail, w€ shoulcl specific fucl consumption of the KM- 37 engine is plotted against
look at the overall approach to the Wankel engine taken by Fichtel & ni.rn effcctive Pressure (Psi).
Sachs and follorv the clevelopment work performed by the Schweinfurt
engineers. 2.OO

Irichtel & Sachs ran into the same difficulties that plagued NSU-
chatter rnarks on the trochoidal track and high seal tip wear. At first r.30
the surface was electroplated with hard chrome. After a longer running
pcriod, cltatter nrarks on the chromium layer developed to the point
where the Iifespan of the engine was impaired. Chatter phenomena
E,
t
\
were reduced considerably by increasing the thickness of the apex seals Ä r.oo RPI'
I
o
N \\
frorn 1.5 to 3.0 mnr. ancl tests with carbon-type sealing elements com- .90 5000
pletely eliminated tltenr, even on the chromium layer. Unfortunately,
\
cD .80
trlflfl
J tr@O -z -/
* /
the \\,ear of the carbon-type seals provecl to be too high, especially at (J
a
.70
ESTITATED
high speecls and whcn the engine ran at high load. In addition to this lr.
o
co
.60
problem, the sealing elements tencled to break, and erosion occurred at
the edges. When the output shaft speed rvas held down to 3,000 r.p.r., 50
the unit rAn for 1,500 hours before the appearance of chatter marks on
tlre bore. At a steacly speed of 4,500 r.p.m., however, they appeared after ro
?3 30 «)
BMEP. PSI
 )16 DEVELOPMENT FIOLDERS 317
OTFIER WANKEL LICENSE
5caring, cornbirrccl with phasirrg gears ancl il slecve
1,000 hours. Iiichtel & Sachs found that the matching of high quality
grily cast-iron scalirrg clcntertts with a chronriuur-platccl working surfacc of thc ecccutric
arrangement to ensure iotor alignment. Sleeves on
the nrainshaft ex-
cxtcnclecl the servicc life an additional 1,000 to 1,200 hours between
te.cled into the rotor from both sides ancl linecl uP with
internal abut-
ovcrhauls. The cltromiunr layer now has been replaced lry a mixture of by four contact
steel and bronze. With this surface, which has a very long lifespan, no ments on the rotor. The rotor was guided on its course
surfaces, plus the phasing gears to some extent.
The roller bearing on the
chatter marks are proclucecl. The surface wear amounts to slightly more
tlutn 0.001 nrnr.pcr hour; the raclial length of the apex seal is only worn
off about 0.008 mm. per hour. June 15, 1965 E. ANsoRG 3,1891263
PACKINO SYSTEI POR ROTARY COIBUSTION ENOINE AND THE LII(E
Irichtel & Sachs engineers established economically acceptable toler-
Fllcd Jan. 21, 1965 2 Shrotr.Shmt I
altccs ancl clearances for the individual elernents of the sealing grid. l'he
grid of the KM37 diffcred in detail rather than in concept from the F/GI
original Wankel arrangement. Light spring-loading behind ever)/ seal /6
ensurecl aclequate functioning under normal operating conditions, with
gas Pressure util ized to effect sealing. Experinrents with other types of
seals did not procluce any improvenrent. One Irichtel & Sachs engineer,
Iirnst Ansorg, took out a patent for a V-shaped and T-shaped apex seal.
I-le claimed that his seal system woulcl ensure a permanent joint be-
tw'een the rotor and the housing, regardless of their relative movements,
ancl rvould not allow any undesirable gas flow in the working chamber.
'fhc V-shapecl Ansorg seal was lodged in a dual slot and provided witlr
spring-loacling against the rvorking surface in the radial direction by a
helical spring positioned on the apex radius. The seal could not be in-
stallecl if it were of one-piece construction, so Ansorg devisecl a built-up
scal consisting of trvo narrow elongated plates integrally connected at
right angles to form a sealing edge. The seal plates were adapted to F/GP
engage certain portions of the seal groove walls and thereby direct gas
Pressure behind the seal plates when desirable. Because of the rounded
shape of the insicle edges of the seal plates and the flat shape of the
engaging portions of the groove wall, there was to be line contact be-
tween them when in sealing engagement. A variant of this system fias
individual seal strips, each with its own sealing edge, both held in con-
tact with the working surface by pressure from a spring-loaded ball
located in the rotor. Ansorg cJaimed that his seals offered larger available
surfaces to receive gas pressure ancl that this would give greater §ealing
effectiveness.
Another engineer, ITranz Rottmann, proposed a different approacfi to
elirninating the chatter marks on the working surface. He took out a Nvsuroa
patent for more secure mounting of the rotor on its eccentric bearing. |rn sl O' torg
Rottnrann claimed that an engine built to his specifications would oper-
ate with reduced vibration at lower noise levels. In addition, there would 'fui
ß.r^-,
' urU
be no rotor tilt that could cause wear marks on the housing. The heart ofr
of his device was a huge roller bearing imbedded in the cylindrical face
v-shapcd apex seal, patentecl by Ernst Ansorg,
 318 DEVELopMENT OTTIEIT WANKEL LICENSIi }IOLDIiI{S )19
eccentric was concerned only rvith transnritting the rotary
motion to t5e and to the 5igfiest heat inptrt in thc Area slrrrotlnding the spark plug.
shaft.
FicStel & Sactrs managecl to keel; tcntpcrature cliffcrcnces within reason-
ßut, the true solution to the chatter mark problem did not
Iie simply ablc bouncls by positioning . lrri,r,ary intakc port within the hot rcgion.
in thc cltoice of conlPatible materials, apex ieal configuration,
or rotor The decision io aclopt this arrangement \vas trtacle because it was felt
guiclance. The k.y was temperature conirol, to
reduce heat distortions tlrat for industrial engines thc aclvant age of siruplicity was nlore im-
of both rotor ancl housing. There is an unavoiclable temperature
varia- portant tfian the higher performance poteutial of an engine witlr an oil-
tion lletw'een the various areas around the rotor housing.
Tlr. housi,g is cooled rotor, especiäIl1, when the cosl ancl conrplication of arl oil cooler
subiect to the least hcat input in the area surrounding
the i,take port, ancl oil circulation pumP were consiclered.
June 15, 1965 E. ANSoRG g,lg91263
PACXINO SYSTEI FOR ROTARY COIBUSTIOX
E}IOINE A}{D THE LIIG
Fllsd Jaa. Zl, 196I Feb' ,^o;
2 SäcrtrsDrct

r/G,a
2
,,-T?, E ARBANGE v E N r ;^ 1"^'.; :l I I,, o*
^,
. o, u,, *: ::i-t"
Flled llarch 15, 1966

i
F/G,4 I

I
I

II
{
I
I

t
luveuroR
Srntf Oasorg

i
T-shapcd al)ex seal, as patented by A.nsorg (above).
tion on thc \/-sha1>ccl seal, rvith a spring-loaded ball Below, a varia-
at the intersec-
tion of thc tn,o blaclcs. Rottnrantt's paterrt for rotor guiclance'
 DOWEL
ü wasxEn
BOLT
FAN HSG
OT,l IEI( WANKLIL LICI'INSL: I IOLDIIIIS 3Zl
ROTOR HSG
OOWEL
L'tvlsxER
Surface temperature in the trochoidal chamber generally was kept
belorv 400olr. The u'orking surface wils clrronrc-platcd art«l grourt«l to a
DOWEL
BOLT SPßING
SPRING
fine fini.sh and the cncl covers werc sprayccl rvith a special bronze coat-
APEX SEAL
APEX SEAL PIN SPRING
ü wasxen
BOLT
APEX SEAL PIiT
ROTOR
ing. Earlier test units usecl a molybdettum coating, wltich wits very
expensive. This led to the cleveloprtrcnt of tlre brotrze coating, rvhich
RET RING RET. RING
I wasxrn SPACER
BOLT
c'n,ErGxr
BEARING offered similar wear characteristics art a nruch rcclucecl cost. The housing
SEAL
MAGNETO
SEAL
BRG. RACE and end covers \\rere light alloy castings with a ttutnber of cooling ribs
RET. RING
PIN BOLT on their external surfaces. Air-cooling was proviclcd by an axial-flow fan
working through full clucting. Thc light alloy axial-florv fan, u'ith its
SPRING c'wEtGHT
SLEEVE

airfoil-section blacles, absorbecl vcr1, Iittle horsepo\tler.

HOUSING
XEY
NUT
WASHER
BOLT

STARTER CLUTCH _.---.

BEARING
GEAR
Pcrfomrance curves for thc KIU-914 cngine.
PIN [AI}{ SHAFT
SEAL KEY
üwesxeR
26
SEAL
BOLT @
24rF
RETAITIER
BOLT BOLT

/ TORQUE
\
WASHER SEAL
FAN
L
SPRING
COVER BEARIilb 22 ,;
END HSG
ü wnsxeR
LIXER
SEAL / \ ?og
:)
BOLT
BOLT
SPRING
I o F
ENO HSG
WASHER !8
Elcvation of tlre KM.gl4 engine.
22

/
r8
-
E,
rt't 16
/
o
=
ft14
/
o SAE
gt2
tr, / HP

ro /
STID
/
I
TUEE
GASXET
BOLT
ADAPTER t
E,
\ x
STUO
FERRULE
COYER
STAI{D-OFF
8iI
SPARX PLI,G
\ )
(D

\ \- ./ rH
-...!-l

BSFC o
c)
t

I
7
ll
F

Cross-scction of tlrc KfvI-914 i

engirre.
i tooo 2000 3000 4000 5000 6000
CRANKSHAFT SPEED . RPM
I
 ,^
o
\ sio
9 / TORO E
8t
IL

lrt

I
I
I ./
7Do
bo G
r800

,/
IL-
F

// 600
5 I

7
/ ,/
,/
l_r / r {oo
G
)
ld
6
L] /xp .
l! ,/
8r L_l I I ? ?oo
llJ
o
!'
\ I 0

I
I'crfornrancc crrrves for G L__l o
=
o
tlrc KIU-48 errgine. a \ - E
E
J
ooo
v
ol
L_l
'rG
- a ,/
L_l A. @
J
z
f, Boo
,/
t_l I \ \ BSFC
.El
(, (o

L_l I L
I
lr
o
.o
3
o
z
Goo
,/
L__t o
L__t 400
,/
LI
II ?oo
,/
tooo 2000 3000 4000 5000
CRANXSHAFT SPEED - RPM
o
r ooo 2000 5000 4000 5000
ENGINE RPM

SPccific frrcl consunrption of the KI\{-914 cngine, plottcd against r.p.m. (mainshaft Cooling fan air flow (cubic fcct 1>cr nrinutc) plottccl against rnainshaft r.p.rn. for the
spccd) ancl nlcrtn cffcctive l)rcssurc (psi). KI\'l-g14 errgine.

Carburetor air flow in the KI\I-914 cnginc.

-
lr.
40
(J
I

äro
J
E,
CU. FTl TIN.
a
G
o20
F
UJ
(ts
:)
E)
tr
3ro

o
rooo 3000 .ooo 5 000
ENGINE RPM
 , )- J-- o
sio
/ TORQ UE\ 1
8t
TL

t,
td
,o
7f r800
/
l./
) ooFE
f ,
t_ r 3 t600
,/
/ ,/
I {oo

L
r

E
§
6
/^, -
lr- ,/
o
o
Gr L_l I I zoo

I
I
l!e
o
t_l \
I o
I'crfornurr)cc crrrvcs tor tr =
o
t'lrc KIU-48 crrgine. a t_l \ -
G
llJ ooo
J .e G
L] r
d. a ,/
]_t A. @
z
fl
,/
I \ \
J Boo
.8
t_l
I
BSFC (, (,
l_l I I
lr.
o
@
z
3
o
ooo
,/
t_l (J

t_l ,/
t_l ,/
tooo 2000 3000 4000 5000
CRANXSHAFT SPEED - RPM

I OOO 2000 5000 .tooo 5000

ENGINE RPM

SPccific ftrel consrrnrPtion of the KM-911 u ' plotted

engine, against r.p.m. (mainshaft Cooling fan air florv (cubic feet 1>cr rninutc) plotted against mainshaft r.p.m. for the
spcccl) ancl nrciln effcctive prcssurc (psi) ''
. KN,I-914 engine.

ÜIEP

Carburctor air flow in the KI\I-914 cngine.

- 40
IL
()
I

äro
J cu. FTl mlil.
G,
a
o20
G,

F
lrJ
ü,
f,
o
G
1

3ro

o
rooo 2000 5000 aooo 5000
ENGINE RPM
 )24 DEVELOPMENT
In May, 1968, I-Ielmut Keller delivered a paper to the Society of Auto-
motive Errgineers on his cooling system clevelopment work at Irichtel &
Sachs ancl other features of the Wankel engines from Schrveinfurt. The
cooling ribs on the latest engines were longer on particularly hot areas
of the housing than they were on the cooler parts. The power expendi-
ture for the blower rernained within the normal limits for air-cooled,
reciprocating piston engines, using axial fans made of light alloy. The
cooling system was based on the use of pressure fans, by which cold air
was taken in and blown over the cooling ribs of the side housings and
trochoid housing. The maximum speed of the engine was automatically
Iinritecl to 6,000 r.p.m.by using a simple throttle control device sensitive
to gas velocity.

Cutaway nrodel of the KM-914 engine.

-_2

APEX SEAI§

CooLINq FAN

EXHAUST
ROTOR

SPARK PLUG
Closeup of the combustion chanrber in the KM-914 engine.

RECOIL STARTER
Ilxhaust gas tcnrpcraturcs arc sorncwhat highcr in thc I?iclrtcl & Slchs
rotating combustion engines than in con\rerrtiorral errgirres. T'lle incluc-
§ tion system on these engines is complicated and rro doubt derived from
hvo-stroke piston engine experience; for, although the KI\/t37 operated
on the four-stroke cycle, the induction systern relied on two-stroke prin-
ciples. Fresh mixture from thc carburetor was lecl to a Passage in the
eccentric. At certain rotor positions, ports in the eccetttric Passage lined
up rvith ports insicle the rotor, while ports in the rotor sides, one in-
OUTPUT SHAFT
boarcl of each apex, lined up with transfer ports that cltannelecl thc
§TARTER HANDLE
inconring charge to the nornral intake port po.sition. This wtts clone to
obtain prc-heating of tlrc frcsh chargc ancl to irtrprovc tctttperxtttrc tttti-
formity in the housing. Gas tenrperature incrcased by abclut 120"F. oll
the \\'ill'through the rotor.'l'his rrrelrrt 1loss in volttttlctric cllrcicrlc!',
but did provide sonrc co«lling cffcct for the rotor.
COOLING FINS Rotor bearing Iubrication ncvcr was a problcnr, becausc a gas mixture
contairring oil flou,ecl though thc eccentric bcariug always in thc salnc
direction, and it was the experiencc of Irichtcl & Sachs that this always
 /r\, CAREURETOR
I AIR FILTER
)

\\ ]ARBI

\ E7r COOLING AIR

T OUCT

-. EHGINE
ijr. EXHAUST PtPE
MUFFLER
4+,
iJ'.t a,

COOLING AIR
-

One of Anrerica's largest manufacturers of sno\ttlllobilcs, Polaris, added a Wankel

enginc option in 1968. T'lrc urrit is Fichtel & S:tclts'RCl-18.5.

IIere is Curtiss-\Vright's suggcstccl irrstallation of the 18.5 cubic inch Fichtel & Sachs
single-rotor, air-coolcd Wankcl crrgirrc in :lll :tll-tcrrrrin vchiclc.

A HOYER
COOLING AIR
INLET DUCT

CARBURETOR
AIR FILTER

CARBURETOR
AIR DUCT COOLING AIR

v
Dual Irichtel & Sachs Wankel engines installed in an experimental Delta-Hover air CAREURETOR
cusltion vehicle. The upper engine provides propulsion by driving an airscrew, while
thc lower provides a ground-effect lift force by pumping air under the vehicle.
)78 DEVELOPMENT OTI.IER WANKIIL LI CIiNSII I IOI.,DIJI(S )29

supplied a sufficient lubricating fihn. This applied also to both main KNI48 SPIICIFICATIONS
I>carings, which were joinecl with the adrnission charnber by means of a
Direction of rotatiorr Countcrclockwise (whcn looking at drive end)
channel and thercby reccivecl sufficient oil. Nevertheless, it was eventu-
Displacement 9.8 cu. in. ( t 60 cc. )
ally clecicled to introduce a srnall oil pump to feed the requisite minute Air-coolcd (blorver)
Cooling
quantities of lubricant (varying with engine speed) into the induction 8:l
Compression ratio
port. This required separate tanks for oil and gasoline. I\,Ietering the Performance 5 horsepower @ 3,000 r.P.m.
lubricant relative to shaft speed provecl entirely satisfactory. 8 horscpower @ 4,700 r.P.m.
'I'he first test engines used special spark plugs, the electrodes of which (tolerancc range * 5%)
were connected to the conrbustion chamber by a channel. These spark Ignition Bosch flyrvhecl magneto-lighting coil on request
plugs were very expensive ancl not readily available; therefore a way was Spark plug Charnpion L-85 or Bosch W 190 MllS
sought to utilize standarcl spark plugs without a connecting channel. Ignition timing l0 clegrces before TDC
The solution made use of the fact that the spark plug was located Breaker gap 0.014-0.0 1 8"
exactly at that point of the trochoid at rvhich equal pressures existed Carburetor Bing 8/14/l
Air filter \\/et t],pe
on both sides of the apex seal. One chamber has compression pressure,
Starting method Rccoil-hancl start
the other has a similar pressLlre from the burning gases. By moving the Top specd limit (governor)
Control
spark plug, somewhat better consumption values could have been ob- Weight )7 lbs., engine including starter, governor, car-
tained, but the advantage of being able to use an inexpensive standard buretor and nluffier
spark plug, available ever),where, was regarded as being more important. Lubrication oil Shell Rotclla SAII l0
The Fichtel & Sachs Wankel engine now uses standard Bosch W 150 lissolub I{D 30
MllS spark plugs. They are quite satisfactory for 150 to 200 hours BP Outboard-Ir4otor Oil
service, which is comparable to the life of spark plugs used on the I\'Iobiloil Outboard
smaller F & S two-stroke engines. I\zlobiloil TT
Nlobilmix TT
KM37 SPECIFICATIONS Pre mix ratio 40: I fuel to oil
ITuel Cnscllitrc, rcgttlar gracle
Direction of rotation Countcrclockwise (when looking at drive end )
Displacenrent 6.6 cu. in. (108 cc.) KN{9I+ SPECIITICATIONS
Cooling Air-cooled (blower)
Cornpression ratio 8.5: I Direction of rotation Countcrclockrvisc (lvhcn lookirrg at clrive end)
Performance 6.5, horsepower @ 5,500 r.p.m. (tolerance range Displacemcnt I 8. 5 cu. in. ( 30 3 cc.)

+r%) Cooling Air-coolcd (irrtcgral blower)

Ignition Bosch flywheel rnagneto I\,IZIURB l/116 R 2 or Compression ratio 8:l
Bosch l5 W Pcrf ormance 20.0 horsepower @ 5,000 r.p.m. (tolerance range
Spark plug Champion L-85 or Bosch W 190 Ml lS +5%)
Ignition tirning l0 degrces before TDC Ignition Bosch fl1,u,heel nragneto (rvith lighting coil ) 40-
IJrcaker gap 0.011-0.01 g,, \t'lltt, l2-r'olt
Ca rburetor Bing 8/l+/l Spark plug chaml>ion L-86 or L-90 or Bosch w t 50 Ml ls
Air filter Wet type Ignition timing l0 dcgrces bcfore TDC
Starting method I land start Brcaker gap 0.014-0.018"
Control 'l'op speed linrit (governor) Carburetor Tillotson IlL242^ with fixed main iet or HL252A
\Veight 3+ lbs., cnginc including starter, governor, ca:- adi. nrain ict
burctor and muffier Air filter Inlct nrcsh filter
Lubrication oil Shcll Rotclla SAE 30 or 40 Starting method I Irrncl rccoil starter
Prc mix ratio 50: I fuel to oil \\reight 56 lbs., crginc includirrg starter, carburetor and
Iruel Gasoline, rcgular grade rnufficr
330
DEVELOPMENT
I\zluffier Flat
I-trbrication oil Shcll Ilotella SAE 30
Ilssolub HD30
BP Outboard-N4otor Oil
15
Mobiloil Outboard
Mobiloil TT
Mobilmix TT
Prc rrr ix ra tio 40: I fucl to oil
Iruel Gasoline, regular grade
I?trel pump Diaphragnr-type p.mp fittecl to carburetor, oper-
atecl by impulse connection with engine

Exhaust
Emissions

IN TrIE pnrivrorJs cllAp'r'rir(s, I havc

clescribed in consiclcral;lc dctail hol thc vurious conU)iurics cnqrgccl iu
Wankcl enginc rcsearch artcl «lcvclopntcut havc attackccl arrcl o\/crconrc
all problcnrs affccting tlrc succcssful urr«l cflicicrtt o1>critti«»r «lf thc rotrtry
conrbustion enginc. lJut thc uclricr,cnrcut of lriglr [lrcnrutl lttttl tttccluttti-
cal efficiency is not enough to guarantce the ru:.lss procluctiorr «lf any
power unit in toclay's environnlcnt-consciotrs rvorlcl. 'l'lrc crrginc, no rrurt-
ter how aclvancccl technologically, Irorv irtcxpcusit,c to 1lr«ldui:c, ltou,
attractive for a variety of installations, still nrust fill one all-irnportaut
I requirement-it must not pollute thc atmosphere.
Thc matter of the Wankcl engine's emissions has not becn fully
mappccl as yet. It is An areil that was ignorccl for the first .six or scvcr)
I

years of development u,ork, but nruch has beeu clonc recerrtly to iclcntify
the problems and to find solutions. Solutions had to be sought, not in
absolute ternts, but within the franrcrvork of the laws cnactccl by the
government of the state of California ancl the U.S. fccleral govcnrnrent.
The automobile inclustry is confident that the reciprocating piston
engine can be "cleaned up" u,ell enough to nreet the 1975 stanclarcls,
and the Ieading Wankel engine nrAnufacturers u'ill not concecle that the
rotating cornbustion engine has \4,orse problcnrs. ße1'oncl tlret, uncoll-
I
ventional powerplants nray bccclnrc ncccssary. I?or thc tirnc being, two
Wankel-powercd cars have mct the current 1970 stitnclarcls.
I
l
 337, DEVELOPMENT EXHAUST EMISSIONS )33

EXIIAUST IIMISSION STANDARDS seconclary 50,000 nrilc test. Thc prinrary test was concluctccl cluring thc
perio«l *ia-1une to rni«l-|uly, 1969 ancl thc sccouclitry tcst was conrplctecl
Exhaust on Octobei 13, 1969. The prescribed lirlit of exhaust gas depcnds on
Crank- Evapora-
partic_ cctse vehicle weight. That is, allowable carbon rnonoxicle is 1.7% and hydro-
tion
carbon .or,-t.r,t is 305 parts per million for a car in the category of the
Year HC CO NO, ulates HC HC
R-I00.
196) ,.7 g7 .2 5.9 0.3 ).2 2.9 NSU went to work on a similar program. In the middle of lune, 1969,
(uncontrolled
veh icles)
an Ro-80 was launched on a 4,000 mile circuit through the Swabian
lowlands. After running for two weeks, it completed its initial trials in
1966 Fed. N.S. N.S. N.S. N.S. N.S. N.S. West Germany. Th.r. *.re followecl by final measurements cln the NSU
1966 Cal. 3.4 3+.0 N.S. N.S. 0.0 N.s. test beds and äir fr.ight to Ypsilanti, Michigan, for testing by the Divi-
1968 Ired. 3.3 34.0 N.S. N.S. 0.0 N.s. sion of Motor yehiile Polluiion Control. In the middle o! Iglv, 1969,
1968 Cal. ),4 )1.0 N.S. N.S. 0.0 N.S. another Ro-80 began a non-stop 50,000 mile test. Here, 50,000 miles
are travelled on a course which has been tested and approved by the
1969 Fed. ).7 34,0 N.S. N.S. 0.0 N.S.
Anrerican authorities. Mountains and valleys, a certain atnount of city
1969 Cal. 2.2 2).0 N.S. N.S. 0.0 N.S.
driving, haltin g at traffic lights and crossroads for a minimum of 5 sec-
1970 Fed. 2.2 2).0 N.S. N.S. 0.0 N.s. onds, togeth.r *ith stretches at full acceleration-all have to be included.
1970 Cal. 2.2 2).0 N.S. N.S. 0.0 6.0 G For this purpose, the NSU development department has its own special
197 | I?ed. 2.2 23,0 N.S. N.S. 0,0 0.5
l97l Cal. z,z 2).0 4.0 N.S. 0.0 6.0 G lvlazda R-100 passirrg the federal etttissions tests.
1972 ITed. 2.2 2).0 N.S. N.S. 0.0 0.5
1972 Cal. 1.5 23.0 3.0 N.S. 0.0 6.0 G
197) Fed. 2.2 23.0 3.0* N.S. 0.0 0.5
197) Cal. 1.5 8.0 ),0 N.S. 0.0 6.0 G
1974 Iied. 2.2 2).0 3.0* N.S. 0.0 0.5
197+ Cal. I,5 23.0 13 N.s. 0.0 6.0 G
197 5 Fed. 0.5# ll.0r o.gs 0.1* 0.0 0.5
197 5 Cal. 0.5* 12.0f 1.0* N.s. 0.0 6.0 G
1980 Fed. 0.25* 4.7* 0.4* 0.03* 0.0

I IC - Hydrocarbons (grams pcr mile)

CO = Carbon rnorroxidc (granrs per mile )
NO, - Oxidcs of nitrogcn (gr.rns per nrilc)
N.S. = No standard
t - Proposed
G - Crams 1>cr test (California spccifications).

In 1969, Toyo Kogyo successfully subjected the R-100 productiop car

to the American test conditions. This removed the obstacle to importing
cars fitted with Wankel engines into the U.S.A. under l97O rules. The
federal government's test consists of a primary 4,000 mile test and a
r00
00
t0 I
-
o-
t0.t0
t.50 ae
r !.00
r2.40 EXTIAUST EMISSIONS 335

I
,.t 0 E
J t 0.00
,0 5.05
r
C'
t.05
track which is about 25 nriles in length. On this circuit, rally Proccdrlrc
80 {.85 ry 6.40
50
I\-
3.60 5.05
is followed using a stopwatch arrcl rigid schcclulc. It is lreccssary to acl-
40 2.60

\
3.E5
30
\ / t.85 2.10
here to an average circuit specd of around 30 rn.p.h. in sumlltcr ancl
\
20 r.t 5
t0
/ \ .50
t.70
.00
0
llta tl
0 0
winter and, after all, it takes more than an occasional thunderstorm or
t{ 31 80 r0r
I lllra.l

60
ilIr.stc0xDs
il14 13, t3, 0 3f 60 t0{ r3' stretch of sheet ice to discourage the marathon drivers (rraturally, in
ilxt.§EG01103 IlrE-stc0lt0s
alternating shifts) from completing their total of I,900 circuits. l)uring
The Federal Test Cycle for Exhaust Emissions specifies starting with a cold engine this phase, the car returns at 4,000 rnile intervals to the test becl wlterc
and letting it idle for 20 seconcls. Measurement of emissions begins at the moment
of startup. After 20 scconcls, thc car spccds up to 30 mph in 14 seconds, and runs the "California test," a prescribed sinrulittccl sequence of specifiecl driv-
at that spccd for l5 seconds before slowing down to 20 mph. The slowdown is per- ing conditions, is concluctecl (irr tcn stagcs frotn idling to partial loacl.
formed with a closcd, or nearly closed throttle, in a span of I I seconds. After running with pushing operation and acceleration) over a total period of 137
at 20 nrph for I 5 seconds, the car accelerates to 50 mph in 29 seconds, decelerates
to 25 mph in 16 seconds, and finally comes to a full stop in a l7-second span. seconds. This procedure is carried out seven times in zuccession, äc om-
T'he sccond graph shows carbon monoxide (CO) emissions, as measured on a panied by , simultaneous check of exhaust gas collstitucnts. According
typical piston engine cluring this cycle. CO enrissions are high while the engine is to somc expert conclusions, which are not crt tircly tlnoPP«lsccl, this
idling, and reach their pcak durirrg decelcration. They are lowest during acceleration
and at stcacly spccds, with snrall peaks occurring as a result of sudden changes in represents, on the averitge, an accurate ancl virtuall)' true-to-life oPerat-
throttle opening cluring accelcration. Complete closing of the throttle produced the ing situation. Runs one to four ilre rcgarclccl as colcl stttrt ct'clcs. '['llc
sharp rises in ttre curve after a1>proximately 50 and 105 seconds running.
Irr the third graph the levcl of carbon dioxide (CO, emissions, which are not
fifth run is disregarclecl, because it is usecl only to rvurn'r up thc crrgittc to
considcrcd polluting, is shown. Ilorvcver, nrapping CO, volume during the test cycle ideal temperature. lluus nrrmbcr six arrcl scven theu arc used as the
gives researchcrs valuable data.'l'he proportion of carbon dioxide remains fairly con- "hot" phases for conclusive checking of tlre actual cxhaust collstituettts.
starrt, rcgardless of vehicle - speed and engine r.p.il., although a drop in CO,
ernissious occurs whenever the accelerator is suddenly depressed or released. Closing
Most of the noxious agents in exhaust gas are odorless and nrust be
thc throttle at high spced recluces CO, output, as more carbon monoxide is then analyzed with the aid of cornplicated and expensive apParatus.
formed.
On December 10, 1969, the final decisive tests were carried out at the
Iaboratory of thc National Air I'ollution Control Aclutinistnrtiou at
Ypsilanti. On December 18, 1969, thc official ccrtificate of succcssful
-
t200
- completion of all of the various tests was signecl. This certificate issued
525

a
00 3500 465
e.
by the United States Department of Health, Education and Welfare
C'
E(l
d-
70 L, t..
2800
2430
O-
1
407
355
CJ
60 -sA
50 F
I 085
I 570
300
24A finally clearecl the NSU Wankel enginc in the lto-80 series for travel in
t_
the rarified atmosphcre of North America, and thus gave NSU every
I

{0 I tg0 t0s
=
30
20
t0
\ \r
{ T
t40
530
240 t-
145
85
46
reason to expect success in the conring year. 'l'he l1o-80 vehicle gave t 50
0
ll
0
- tl
0
parts per million hydrocarbon emission and only l% carbon monoxide.
!r 00
lt.aa lll.ll
il00 tf G0
a

Ilrt.stc0ilot
t0t t3,
Tlrt-ttt0ll0t
toa t3,
ilrE.stG0[Ds
tol 13,
The dramatic recluction in cmission levels \r':ls accomplished by irr-
stalling a thermal reactor as part of the exhaust manifold.
The first graph nreasures intake air flow in cubic feet per minute. This is an indica- In order to understand the emissions problenr, onc ntust hat'e alt
tion of the anrount of ox),qcn available for combustion. Air flow is approximately unclerstanding of gasoline as a motor fuel. Gasolinc is a hyclrocarbon,
1>roportional to vehicle speed and throttle opening. Lack of oxygen during decelera- nrade up of about 15% hydrogen and 8r% carbon. Air is a mixture of
tion is a prirlrary factor in the formation of high volumes of carbon monoxide output,
with a loss of carbon dioxide content in the exhaust gases. 2I% oxygen , 787o nitrogen, and l% other gascs. But onl1, the oxvgell
In the secorrd graph, the emission level of oxides of nitrogen (NO,) are plotted combines u,ith the gasoline. Gasoline rvill not burn in a liquid state; it
(in 1>arts per nrillion) against tinre. NO, occur as nitric oxide and nitrogen dioxide.
They are toxic, and contribute to the formation of photochemical smog. NO, nrust be changed to vapor ancl supplied u'ith an adcquttc volunle of air
crttissions rise sharply under acceleration, but have little relation to vehicle speed or to allou, combustion. The air/fuel mixture nrust be acljusted to speed,
r.p.m. Lowest NO, emissions are recorded during idling and during deceleration. load conditions, and temperature, and the total volume of the nrixttrre
Strangely, a rich mixture helps reduce NO, content in the exhaust gases.
'l'lte third graph shows unburned hydrocarbon (HC) emissions (in parts per also must vary accorcling to these conclitions. F-or cold starting, the
nrillion) plotted against time. IIC emissions are the direct result of incomplete air/fuel ratio must be as rich as 3:l or 4:l; a ratio of l0:l is best for a
contbustion, and the pattern of I IC emissions during the test cycle is closely related warm engine at idle. For part throttle, Iight load operation, a Iean ratio
to engine r.p.nl., load antl road spcecl. A peak is reached during deceleration from
high speccl (up to 2,000 parts per million). In general, HC emissions are high when
air consumption is low, and vice versa.
336 DEVELOPMENT EXI{AUST EMISSIONS 337

of I 5 or 16.5 : I will suffice,but for f ull throttle operation, the ratio bccomcs iso-octane l)lus r ccrtairr arnourrt of tctra-ctlryl lcacl. Tlte knoc:k
slrotrld rise progressively to about l2:l
. vuluc of such fucls is «lcfiuc«l u.s nrillilitc:rs of tctrrr-cthyl lcrrrl pcr gnllorr
In general, a rich mixture rvill produce Inore toxic emissions than a of iso-octauc.
lean mixture. f'he exact emission quantity and composition clepends Anti-krrock cornpourtcls usually are lcacl-basccl. Thc active arrti-knock
mainly on throttle opening and engine r.p.m. With a rich and,/or Poor agcnts are alkyl-lcacl conrpouncls consisting oI cthyl ancl ,/ot rtrcthyl
mixture there is the risk of incornplete conrbustion in the Wankel groups attachccl to an utour of lcacl-tctra-ctlrvl lcacl is thc nlost cottr-
engine. T'he problem is most serious under Iight load conditions, when mon of these. Recause lcacl is a know,n poison, arrcl pctrolcuur inclustrl,
cxhaust dilution is sigrrificant. The Wankel cngine seems to have an spokesnlen gcncrllly aclrnit tlrnt about 50",'r, of tlrc lcrrcl contcttt in gflso-
advant age over the piston engine in this area bccause tests havc shown lirre escapcs into the atnrospllcrc via tlrc crlr's taill>i1lc, tlrc ittttotttol>ilc
it to operate well on a lean mixture. In addition, tlte Wankel has intense industry, Ied by General Motors, has prcssurccl the oil cornparrics into
conrbustion chambcr turbulence, which tends to pronrote mixtrtre honto- providing non-lcaclcd or l«lrv-lcad contcnt grrs«rlirrcs of 9l (x'trlllc. r\ll
geneity and complete combustion. American piston engines now irr procluctiorr will rurr on this 9l octltttc
GasoJine must be in vapor form at low temperature to ensure easy fuel. FIow will this affect the Wankel cnqirrc? As part of its rescälrch
starting. It must vaporize at an increasing rate as carburetor and mani- on the use of non-Jeadcd gasoline ancl lou'-gnrclc gasoline in rotary cn-
fold temperatures rise to allow fast warnr-up, smooth acceleration and gines, Toyo Kogyo carriecl out tc.sts on thc fucl octaue value requirccl.
even fuel distribution among the cylinders. The vaporizing cltaracter- They wanted to see to what extent the octane value of the gasoline
istics must be in keeping with the clirnate ancl altitude to prevent vapor could be lorvcrcd rvithout nclvcrscly affcctirrg;rcrfornt:rncc.'l'ltc tcsts
lock and fuel boiling inside carburetors, fuel pumps and lines. Gasoline were carried out at the Miyoshi Proving Grouncl beginning on Mry 29,
should contain few extremely high-boiling hl,clrocarbons to ensure goocl 1970, in collaboration with An oil conrpanv. 'f'wo l\{azcla It-100 cotU)es
fuel distribution and freedom from crankcase cleposits and dilution. A were usecl for the test, ancl various ulca.surcnrcnts of llotlr fucls ancl
lrigh anti-knock quality (octane number) throughout its boiling rangc running 1>erformance wcrc nraclc. The rcsults wcrc nrorc favoral>lc thtrr
is needed to give freedonr from knock at all engine speeds ancl loads. predicted:
Gum content rnust be lorv to prevent valve stickirrg, carburetor difficul-
ties and deposits inside the engine and intake manifolds. Gasoline also
Octane value required 67 (full throttle
at 2,000 r.p.m.)
must have good stability against oxidation to prevent deterioration and
66 (part throttle
gum formation in storage. at 2,000 r.p.m.)
The anti-krrock property of a gasoline is indicated by its octane num- speed
I\{aximurn I 13 m.p.h.
ber. The octane scale was created by giving the number 0 to heptane Acccleration 16.6 sec. stirndirrg-start
(C?flr0) and the number 100 to iso-octane (C8FIr8). Numbers between % mile
0 ancl 100 indicate the proportion of each if the two are mixed. The
octane number of a gasoline is determined by a test comparing it with Piston engine cars of sirnilar perfonn:lnce recluire a high-octane gaso-
a mixture of heptane and iso-octane. Iior example, if the gasoline shows Iine of 98 to 100, but the rotary cars are an cxception. NIazda rotary
the sanre tendency to knock as a mixture containing 6% heptane and vehicles normally run on regular gasoline with octarre values of 87 to 91.
94% iso-octane, its octane number is 94. There are two methods of The test proves that the Wankel-porvered car fully retains its per-
establishing a fuel's octane number. The research ntethod requires that formance, without any clrop in maxinrunr speecl ancl acceleration, even
the test engine be run under closely controlled conditions of speed, air with far lower octane gasoline totally free of lead aclditives. This ability
intake temperature, and ignition timing. The motor method requires of the rotary engine to operate on lorv-octane gasoline attractecl Toyo
that tlre test engine be run rvith variations in speed, air intake tempera- Kogyo's attention at earlie r stages of dcvcloprnent, and research has
ture and ignition timing. The clifference between motor and research since continued on the relationship between non-leaded gasoline and
rnethocl ratings for the same fuel is called sensitivity.It is impossible to combustion. Toyo Kogyo engineers explain that the rotary engine does
obtain more than 100% iso-octane in a reference fuel blend, but some not require leaded gasoline because of certain basic characteristics in-
fuels have ratings above 100 octane. This means that the reference fuel trinsic to its combustion process.
338 DEVELOPMENT
Generally, knocking (abnormal combustion) occurs due to slow flame
propagation or high tcml>erature at some corner in the combustion
cltanrber. In the combustion charnber of the rotary engine, the mixture
moves at high speed in the clirection of rotor rotation, thus causing
rapid flame propagation in the same direction. Also, the temperature at
the trailing cnd in the combustion chamber is low due to the cooling
effect of the large surface area at this portion of the chamber. Dr.
Iiroede, of NSU, rentarkecl at or)e tinre that displacing the rotor face
recess in thc leading direction increased the tangential gas velocity con-
siderably ancl was an effective way of speeding up flame propagation.
NSU also points out that, in standard Ro-80 form, their KKM-612 en-
gine produces less than one-half the unburned hydrocarbons emitted
by the Mazcla engine clcspite the apparent clisadvantage of considerable Elevation arrcl cross-section of the aftcrburner used for the NSU Ro'80.
port overlap, NSU also confirnrs that the KKM-612 has low-octane, non-
lcaclccl fucl capability.
The various pollutants from automobile engines were Iisted earlier,
but perhaps a little more should be saicl about the nature of these 30" vor OT. 30'vor O.L

pollutaltts. More than 957L of the exhaust is composed of innocuous

substaltces : rvater vapor, carbon dioxicle, and nitrogcn. 'I'he main air
pollutants crnittecl bv the internal combustion engine are hydrocarbons
(I IC), ccrbon monoxiclc (CO), and oxicles of nitrogen (NO.). Carbon
ntonoxide ancl oxicle.s of nitrogcn are formed primarily in the bulk gas;
hydrocarbons are fornred in the quench area. Automobile exhaust gas 2 x 250 cmJ
I x 500 cm3 P^o, = 25,5o; K=7,15;B=1,8'e
is a nrinor source of particulatcs arrcl sulfur dioxiclc (SOr) content is in- Pro, = 25,5o ; K=f,, 15 ; 8= 1'8' e F/V = 1,5 cm-l
significant. F/V= J53 cm-l
Enrissions parallel power output. Complete utilization of the fuel 2 Zytinder je 500 crn3
Bohrung/Hub = t : F/ Y= 157 c m-t
gives 100% power ancl no undesirable exhaust emissions; it's the un-
burned portion of the fucl that produces the problem. Carbon monoxicle Comparisorr ratios in a two'cylinder _piston
emissiort is cau.secl nrainly by poor cornbustion, in combination with engine with ment of- 500 cc. to that of
a fingle-roto and a twin'rotor Wankel
rich air/fuel ratios and partial combustion in the quench areas. Most
of the conditions that result in reduced CO emissions also result in re- ratio. Y*"- - Lealing angle. K : K-factor, or R/e ratio. B = Bore
duced HC emissions. |ust as is the case with the reciprocating piston (in piston engirre). Rotor width (in Wankel engine). e
B- =
Eccentricity.
engine, carbon rnonoxide emissions in Wankel engines depend on the
ür /fuel ratio of the mixture. The leanest mixture gives the lowest emis-
sion levels, but the ratio must be kept high enough to maintain stable are slightly lower than in piston engines, NO. emissions are less of a
combustion. Atomization of the fuel (pre-heating of the mixture and its problem. One way that has been suggested to prevent the formation of
flow speed in the intake manifol,l), distribution of the mixture and the NO, is the addition of an inert gas to the mixture to reduce the peak
turbulence of the mixture also are factors that influence exhaust emis- cycle temperature without increasing the oxygen concentration. Experi-
sions. With relation to these three factors, the Wankel engine has ments along these lines have been run with convcntiorral piston engines,
clcfinite advarrtage over the reciprocating piston engine. using exhaust gas recirculation to provicle the inert Bäs, and one source
NO, entissions probably are created in the lrigh-temperature flame gathered convincing data shorving that there was a reduction in the
front. Because peak combustion temperatures in the Wankel engine amount of NO. formecl as the anrount of exltaust gases recirculated
340 DEVELOPMENT EXFIAUST EMISSIONS 341

was increased, There also was good correlation between the predicted tion and the frequency of combustion in one specific area of the
reduction and the observed reduction. housing.
When the exhaust emissions of the Warrkel engine are compared to Thc ts of recent tcsts pro\/c that the shape of tlrc contbttstion
rcsul
those of the reciprocating piston engine, the hydrocarbon emission level chamber cavity forrnecl on the rotor face, the position of the spark
of the Wankel engine is higher and the oxides of nitrogen Ievel is lower. plugs, and the ignition advance are closely relatecl to hydrocarbon con-
Iror this reason, most research and development of rotary engine emis- centration. Going to a lorver compression ratio aids the surface-to-
sion control systems is being directed to the reduction of hydrocarbons. volume ratio by increasing the volume of the combustion chamber
The factors that deternrine the hydrocarbon concentration in the ex- without adding much surface. Thc exact surface-to-r'olume ratio is de-
haust gas are not easily isolatecl, but the main cause of FIC emission is termined by the engine's basic geometry. For engines that are otherwise
considered to be the flanre quenching action which takes place near the iclentical, it is fixecl by the angle of obliquity.
wall of the conrbustion chamber, plus the misfiring and leakage of The angle of obliquity is half of the largest possible angle contained
fuel into the exhaust port. This flame quenching phenomenon is by the rotor's radial centerline (passing through the apex seal), and a line
thought to be due mairrly to the very poor surface-to-volume ratio at normal to the tangent of the epitrochoidal surface where the apex seal
the trailing end of the combustion chamber. A large quench zone is contacts the surface. With raclially disposccl apex seals it corresponds to
formed in this section, and it becomes difficult for the flame front to the leaning angle. The higher the leaning angle, the more favorable tlre
rcach the end gas. The following example may help illustrate the se- surface-to-volurne ratio. F or a given total engine displacement, tlte
verity of the problem. If a handkerchief is dipped in gasoline and held single-rotor Wankel offers an advantage in surface-to-voltlllte ratio. A
in the open air, sctting a match to it will cause it to burn very quickly. twin-rotor unit, with 250 cc. chamber volume, has a less favorable sur-
If another handkerchief, soakecl in gasoline, is wrapped tightly around face-to-volume ratio than a single-rotor 500 cc. Wankel engine.
a flat iron and a match held to the gasoline-wet fabric on the iron, it In acldition to the quench zone problcm, the Wankel cngine also
will not burn. The conclusion is that a thin layer of explosive mixture suffers from a "crevice effect," the inability of gases to bttrn wltcn the
on a relatively cold surface will not ignite. In Wankel engines, it means surfaces enclosing them mo\/e very close together. This crevice effect
tlrat fresh air/fuel mixture in the far portions of the combustion cham- results in unburned gases being trapped along the edges of the rotor
ber nlay stay unburned throughout the combustion process and then sides, in a small band limited by the side seals. The corresponcling phe-
escape as unburned hydrocarbons, partially burned or cracked fuel, car- nomenon in reciprocating engines is a thin layer of unburnecl gas belorv
bon monoxide, and otller less harmful products. the piston crown arourrd the piston top land, limited by the toP com-
The possibility of wall quenching irr any one engine is a function of pression ring.
that engine's combustion space surface-to-volume ratio. A piston engine A Wankel engine's thernral efficienc)' is raised wlten the comPressi«rn
rvith a big bore arrd a pancake-shaped combustion chamber has a cham- ratio is increased. This rneans more power ancl recluced emissions.
ber with small volunre but Iarge surface area. This combustion chamber Similar improvements can be procluced by allorving the engine to run
is more prone to wall quenching than the compact space formed by a at higher operating ternperatures, simply because an engine that is
srnall bore and a cylinder head cavity shaped like half a pear. The small running hot will provide the bcst rnixiug of the charge and the ntost
bore engine tends to have a low surface-to-volume ratio. In the Wankel complete combustiorr. In piston engincs, thcre are a number of factors
engine, the combustion chamber is inherently flat (except for the cavity that limit operatirrg temperaturc, tlre ttrost critical of which is the cx-
in the rotor) and elongated. Many experts have maintained that the haust valves. The \\/ankel engine prornises rcliability at far higher
Wankel engine, with its long, flat combustion chamber, is inherently at temperatures than are now possible with pistort engines.
a disadvan tage in respect to surface-to-volume ratio. While the cornbus- Fuel supply at high temperatures may be a problcm, ancl this is the
tion chamber passes through the minor axis, quenching occurs in the area where direct fuel injection inside thc rvorking cltantber ltas mucll
space on the trailing side, and, therefore, high concentrations of hydro- to offer over carburetors. The validity of this claim is provccl by tlte
carbons occur in this trailing area. Quench may be partly compensated performance of the I\{ercedes-Benz C-l l l. In 1967, Curtiss-\\/right be-
for, however, by higher wall temperatures due to the Iocalized combus- gan tests with the Conelec electronic f uel inicction system (uncler
I
 I
I
I

34?. DEVELOPMENT EXHAUST EMISSIONS )47

I
I

in one rvorking chanrbcr corresl>oncls to bottom clcacl ccutcr aftcr

I

rln
Rendix patents). This system's characteristics are low-pressure, Iong- I
i

cluration nretering, ancl Curtiss-Wright was looking for improved fuel intake stroke. Remember that operational events iclentifiecl by time in I

the piston engine are clefinecl as points on the workirrg surface in the
I

vapori zation I-Iowever, because the Conelec system works on low i

pressure with port-mountecl injectors, it did not give the predicted re- Wankel engine. The area in question was cooled by the intake mixture,
sults. and it was at one time thought that the extremell, lou, tcrnpcrature of
Pre-heating of the intake air is part of the emission control systems this portion was detrimental to the atomization of the fuel mixture. This
on many cars powercd lry reciprocating piston engines. Toyo Kogyo so Ied to the design of a systenr for prc-heating thc intakc air by exhatust
far stancls alonc in using such a device on Wankel engines. It performs gas. The housing was reclesigned to function as an exhaust heat ex-
a dual function on the Toyo Kogyo Wankel. The low-temperature area changer. Toyo Kogyo's heating system used only a part of the exhaust
gas for circulation in the lorv-temperature areas of both the rotor hous-
ing and the sidewalls. The svstem also servecl to eclualize housing tem-
I\{az.cla engines rcsponded well to secondary air iniection. The exact peratures and thereby minimize the dual 1>roblcms of thermal distortion
position of the air hole or nozzle had strong influence on emissions.
Position A (iniection holes in the wall of the working chamber, near
in the housing and deterioration of the lubricant on the sliding surfaces.
thc exhaust ports) , gave the lowest emissions. Position B (air nozzle It proved impossible to completell, climinate the temperature differences
inside the exhaust port) rcsults were acceptable, and there was no and thermal clistortion, and An amount of clistortion ultimately was
risk of clogging the nozzles with combustion products. Position C
(air nozzle downstream from the port) showed only a small improve- iudged tolcrablc insofar as it did not noticeably hinder the operation of
nrcnt over runniug rvithout secondary air iniection. the moving parts in relation to the housirrg or cause rnalfunctions in
overall engine operation irr service.
700
Blowby gases in a piston engine are gases that force their 'rvay past
the piston rings into the crankcase. Blowby is a significant source of
4
600 pollution in these engines. In the case of the Wankel engine, blowby
l Ieaks past the apex seals in both directions. Leakage at the leadirrg end
500 \o aJ I of the combustion chanrber means that raw mixture is vented into
the exhaust port, u,hile gas leakagc at the trailing cncl is includecl in the
next charge and is not lost. Thc gas sealing performance of the sicle
400
E
seal also has an influence on blowby. Gas leakage hcre goes into the
&
(J- housing oil sump ancl reduces engine performance.
r 300 If both the oil seal and the sicle seal function eclually u,ell, the harm
C B
(€ brought about on the oil seal by the blowby gas is far smaller in the
200 )po.E P case of the side port design than in the case of the peripheral port. This
is because the side intake port is located betrveen the side seal and the
\ oil seal so that if there is a gas leak through the side seal, escaped
too
V A gascs will be drawn back into the intake port. This nrcAns that the gas
pressure acting on the oil seal's periphery rvill bc kcpt lorv.
o
r.6 ?,o 2.4 The utilization of thermal reactors, rnade by f)tr Pont, allorved botlr
?.8
FUEL FLOW lblhr tlre NSU Ro-80 ancl the N{a zcla Wankcl to pass thc current crnission
standards. Thermal reactors similar of these nuly play an important
EMISSIONS, ppm
part in thc acccptAncc of thc Wnnkcl cngirrc irr conriug ycurs. A tlrcnrur]
SIhJGLE.CELL TYPE DUAL.CELL TYPE reactor is simplv an "afterburner" tlrirt exists solcly for the purpose of
WARMUP 8to 376 burnirg leftover combustion products to prevent thenr fronr escaping
HOT r87 42 into the atmosphere. Afterburners do not contribute to the efficiencl'
COMPOSITE 406 t59 of the engine; in fact, they detract fronr engine performance. Te.sts
 3M DEVELOPMENT EXHAUST EI\4rSSrONS 3$
with thermal reactor systems have indicated power Iosses up to ZO% stallations yielded Ioltmission Ievels. Because it is fearecl that the
ancl increases in fueJ consr.rrnption of about 107".
iniection holes can become clogged by combustion products during
The thermal reactor provicles a high-temperature zone in whicS hy-
operation, tests are at present being performecl with the ortftngon.I
drocarbon ancl carbon monoxide emissions are burned almost
com- nozzle installation. However, if the iniection lrole configuratiol can be
pletely before being passcd on to the atmosphere. The reactors,
con- further improved, there is a great possibilitv that the test position hav-
sisting of a casing containing a tubular core, replace tlre conventional
ing the greatest potential, the onc nearest thc port, can be restorecl.
exhaust manifolds. Hot exhaust gase.s are mixed in a tubular
core rvith
air from a separate air iniection System operatecl by an engine-driven
pumP. Oxiclation takes place in the core, and the oxicl ized
Irr., Ieave
through a series of hoJes alorrg tlre lcngth of the core, pass aroupcl I Curtiss-Wright RC2-60 U5 automotive engine fitted with an afterburner.
tle
core's outsicle to help keep it at a higli tenrperature, tlren
are expellecl
into the exhau.st pipe.
Research and development conductecl bv Toy,o Kogyo on
exhaust
emission control of the Mazda 0813 engine was carriecl out
with t6is
tyPe of exhaust reactor. At first, tests were macle to
fincl which seconcl-
ary air iniection location was the nrost efficient, Injection
holes were
provided inside the exhaust chamber close to the exhau.st
port, by utiliz-
ing the unique strttcture of thc rotary engine. In anothär
test, an air
nozzle was installed insicle the exhaust port orthogonally
to t5e exSaust
gas flow, and, in a thircl test, an air iozzle was
installccl ancl aimecl
counter to the exhaust gas flow. At idline speecl, both the
first two in-

Gas flow in the thermal reactor used on the Curtiss-Wright engine.

,. I
DIRTY"
EXHAUST GAS
ADDITIONAL AIR
,,DIRTY"
ExHAUST GAs 'trli
ADDITIONAL AIR i-r

,.
CLEAN"
EXHAUST GAS
---t-t'/'
346 DEVELOPMENT EXHAUST EMISSIONS 317

Certain design criteria were established by the Toyo Kogyo engineers. The exhaust emissions of the RC2-60 U 5 were measured, with ancl
The reactor had to allow sufficient time for the efficient mixing of ex- u,ithout an exhaust reactor, both under stcady-state conditions at the
haust gas ancl injected air, hacl to have efficient heat insulation. rapicl University of Michigan and in a vehicle operated on the simulated
warm-up tirne, durability, arrd had to be made of lorv-cost material. California cycle at an independent facility under contract with Curtiss-
In order to hold the ternperature of the reactor core below 1,800oF., a Wright. Tlrc big qucstiorr was whcthcr tlrc Wankcl crrgirte had any
clurability determining factor, the supply of secondary air had to be inherent features that tended to produce high exhaust emissions. Next
automatically cut off by a control device when engine revolutions ex- came the identification of the design parameters that affected exhaust
ceed a certain limit. emissions, and the determination of how they could be modified to
The base line of hydrocarbon emission level on the 0813 engine was reduce the emission levels. The study began with an investigation of
recorded as being 2,000-2,100 parts per million under the California how changes in aft /fuel ratios and spark advance affected hydrocarbon
cycle. With various engine modifications and the adoption of a thermal emissions. The study was restrictecl to hydrocarbons because only limited
reactor with air in jection, this was reduced to 120-200 parts per mil- equipment was available. In addition, it was believed that carbon
Iion, and the CO and NO* emission levels were 0.6-l .0To and 400-600 rnonoxide emissions rvould follow a siurilar pattern, and consequently
parts per million, respectively. The hydrocarbon emission Ievel in the no separate study on CO was undertaken.
cold cycle could not be considered as being satisfactory, but the level First, the engine was tested in its basic form to establish a base line
in the hot cycle was down to a very Jow figure. for results obtained with various emission control systems. The test
The installation used by NSU for the lto80 is essentially the same, engine was run at 1,000,2,000, and 3,000 r.p.m. under varying load con-
and Daimler-Benz has done a considerable amount of work with the ditions, connected to the dynamometer through an automatic trans-
Du Pont reactor. In 1969, Uhlenhaut felt that the C-tll would meet the nrission. Instead of merely adopting the federal test, which seems ar-
California exhaust emission standards when hot, but would fail the bitrary although it includes a variety of common city and suburban
cold-starting test. The problem was that very high temperatures are nec- driving conditions, Cole was more interested in studying the overall
essary for proper reaction, and the materials in use up to that time did emission characteristics of the Wankel engirre than in making nrodifica-
not have enough durability. Exotic metals had to be avoided because of tions that would suit the present federal test method. He discovered that
their high cost, which would prohibit mass production, then and in the the emission problem was most severe at low engine speeds. Conse-
future. However, in April , 1970, a Mercedes-Benz spokesman said, quently, only Iow-speed dynamometer tests were made. At high speeds,
"Mrior progress in the area of exhaust emission control has been made the emission levels were surprisingly low, but the problem with high
with the use of afterburners." I{e added that a new thermal reactor r.p.m. was that as breathing and emission levels improvecl, the rubbing
sy.stem, developecl specifically for the C-l I I engine, is currently under- speed of the seal tips became very high. At low r.p.ffi., the engine proved
going intensive tests at the company's research department in Stuttgart. to have considerably higher emission levels than modern automobile
Curtiss-Wright ancl the University of Michigan have jointly developed piston engines. That was because the Curtiss-Wright RCZ-60 U5 had
an insulatecl cylindrical reactor manifold. A series of baffies within the been designed without any regarcl for emissions. Detroit's engine de-
manifold slow down the exhaust gases and keep the particles suspended. signers have fought with the emission problem for many years and
At the same time, air is pumped into the exhaust manifold by a vane- have had time to incorporate their knowledge into all-new engine de-
type Saginaw compressor. This air mixes with the exhaust gases and signs as well as to apply it to existing engines in the form of detail
helps complete the combustion process at 2,000oF. modifications.
Professor David E. Cole of the University of Michigan, son of GM The engine first was tested with a thermal reactor witltout secondary
President Edward N. Cole, and Charles |ones of Curtiss-Wright de- air injection, thcn with air injection. To ensure complete burning, tenr-
Iivcrecl a ioint paper to the annual SAII Congress irr Detroit in |anuary, peratures in the area of l,l00oF. were consiclercd to be the allowable
1970, concerning the emission control studies conducted at the Uni- minimum. Oxygerr shortage is a const'ant inrpcclinrertt to proper bunt-
versity. The paper explains a full-scale investigation of the emission ing, but this can be overcome by iniecting air into the reactor. AIso, the
characteristics of the RC2-60 U 5 engine and presents many interesting gas flow inside the reactor has to be turbulent to ensure conrplete burn-
conclusions. ing, and the reactor nrust be dcsignecl to givc a«lcquate residcnce tinrc
348 DEVELOPNIENT EXIIAUST EMISSIONS 349

for thc gases to complete proper reaction. This meant a long and sinuous
off at about 1,700oF., I{C enrissions also levelecl off-to 150 parts per
gas flow pattern was necesstty, maintained by internal baffies in thc nrillion. Within tlo rninutes, thc reactor hacl reachcd 90% of its full
reactor core. Such a baffie sy'stenr was developed, ancl the baffies also
effectiveness.'Iest results with the reactor showed temperatures of
1,900"F. in the core ancl 700'F. on the outcr shcll surface.
ensured mixing of the secondary air rvith the exhaust gascs. FIot "clean"
gas was recirculatecl arouncl the core to minimize heat loss from the core. It is characteristic of the Wankel engine to run with high exhaust gas
The hot "clearr" grs a]so circulatecl around the reactor neck to maxinrize temperatures and thereby contribute to the efficiency of the thermal
cxhaust gas tenrpcrature as it entcred the reactor. Air iniection was ar' reactor. Each exhaust port receives exhaust gas during one-half of the
ranged by n-lounting a clirect port nozzle upstreanr from the reactor, four-stroke cycle. T'his conrpares to bctween one-sixth and one-eighth
and an insulation blanket was attached to the inside of the casing to in a piston engine. T'he highcr frequency of exhaust pulscs, directed
minimize heat loss by racliation. The high-tcmperature parts were nrade against the sanle sector of thc reactor, rxiscs the tcrnperature in that
of a high-grade turbine alloy called l{astelloy X. area. The Wankel engine was founcl to have a very high tolerance to
Dynamometer tests indicated considerable aclditional improvements retarded ignition and lean mixtures. In combination, these conditions
in emission levels were obtainable through adjustnrents in spark timing, produce late burning and lowcr flarne propagation, rvhich in turn means
carburetor setting, choke control, heat riser flow, and air pump pressure. hotter exhaust gas. The Wankel engine, then, is well suited to rvith-
The total reactor volume \t,as equal to the displacement of one chamber stand the high exhaust gas tenrperatures and backpressures created by
of each rotor; the volunre of the inner core was equal to one chamber. the Llse of an afterburner. Because there is no exhaust valve, high
Cole reportecl that he found inclications that larger reactor volume exhaust gas temperature is less critical than in a piston engirre to begin
could have furtlrcr recluced emission levels by allowing longer resiclence with, and high backpressure is no threat to valve life and operation. The
timc for the exhaust gases irrsicle the reactor. The air iniection nozzle Iack of a valve guide aud its support, and the sirnple shape of the port,
was positionecl at the earlicst point behincl the exhaust port consiclerecl makes design for aclecluate cclolant florv arouncl the port a sirnple task.
feasible for maxirnunl nrixing of the second ary air and exhaust gas before Hydrocarbon emission levels provecl cxtrernely sensitive to air/fuel
it entered the reactor. No tests \l,ere made using alternate positions. ratios. With 17 .5:l air /fuel ratio, the RC2-60 without the reactor
Thc enginc responcled very rvell to the acldition of the thcrmal re- real iz,ed a Zr%, rccluction irr lryclrocarbon cmissions. With the same
actor. I-Iydrocarllon crnissions at 1,000 r.p.m. were recluced from o\Ier 17.5:l air/fvel ratio, plus the thernral reactor without air iniection, a
900 to less tharr 100 parts per rnillion. Running at 2,000 r.p.ffi., witlr reduction of 75 to 90oÄ in hyclrocarbon ernissions resultecl.
a constant rnanifolcl vAcur.un of 20 inclres of mercury without the re- In a parallel prograru, anothcr ItC2-60 U 5 cngine was installed in a
actor, gave a hyclrocarbon en'rission Ievel of about 500-510 parts per 1964 Ford Galaxie 500 seclan. This experimental vehicle was equipped
milliorr with 50 degrees spark aclvance. Retarcling the spark to 30 cle- rn,ith an air iniection reactor .s),stem incorporating a Saginaw air pump
grees (before the minor axis) brought the lo,el clorvn to 340-3 50 parts n,ith a capacity of I 9.3 cul>ic inches, a l0 inches of mercury relief
pcr million. Parallel irnprovenrents urere founcl possible with similar valve, a Delco back florv check vAlve, artd Rochester diverter valves.
spark rctarclation under Iower loacl conditions. The air pump was driven at 1.3 tirnes rnainshaft speed and vehicle
Airlfuel ratios wcre studiecl at 2,000 r.p.m. rvithout the exhaust re- weight was 4,366 pounds. T'he vehicle was tested on a Clayton chassis
actor, running on regular gasoline. Manifold vacuum was varied in the dy'namonreter by the Ilsso Rcsearch ancl Ilngineering Conrpany, Lindcn,
tests (15, l0 ancl 20 inchcs of mercury). Air,zfuel ratios around 17.5:l New ferse1,. Exhaust enrissions nreasured and recorded in this test in-
ga!'c the lorvcst hydrocarbon concentration in the exhaust; except at the cluded hyclrocarbon, carbon mor)oxide, ancl oxicles of nitrogen. In gcn-
highest n)anifold vacur.lnr, wlrich gavc tlre clcartcst cxhaust at I5.5:l cnrl, the vclriclc tcsts confinrrcd thc finclirrgs of tlrc cnginc bench tc.st
air/fucl ratio. results. Flydrocarbon lcvcls were arouncl 100 parts pcr urillion, carborr
Next, the engine was fitted with a reactor on the test bench and put nronoxicle variecl betw'een 0.77 ancl 1.47 ,,qL ancl oxicles of nitrogen varicd
through a colcl-starting test to sce rvhat effect temperature had on start- from 378 to 607 1>arts pcr rrrillion. Thc tcst vclriclc, sai«l Colc, illustratccl
ing al>ilitv ancl emission levels. T'he engine \r/as started and run at 1,000 that current lcgal limits could be met for h1'clrocarbon ancl carbon
r.p.m. Ilydrocarbon cnrissiorrs clropped clrasticallv after 30 seconds, when monoxide emissions. Thc NO- emissions were above the anticipated re-
corc tcrnperaturc reachcd I,l00olr. When core tcmperature Ieveled strictive level, but wcrc bclievcd controllable by relativcly minor acl-
750 DEVELOPMENT
iustment. The reactor principle was provecl to be fully compatible with
the Wankel engine, and the development of low-cost versions was
recommended.
The last paragraph in David Cole's paper suggested exploring far
higher speeds because the engine appears to have reduced emission
levels in the upper r.p.m . rangc§. "The RC engine's rapid decline of
hydrocarbon emission levels as a function of increasirg speed conr-
bined with the engine's propensity for higher r.p.ffi., by virtue of com-
plete balance and no valves, may be exploitable. Raising the entire
speed spectrum 50 to 100 percent is not an unreasonable development
goal today."
Secti on lll

APPLICATIt)NS
 16

lntroduction

Tnu DUTY oF PRovIDINc IuoTIvE

power for a passenger car is only one of a multitude of applications-
actual and potential-of thc Wankel cngine. It can be regarclecl as a
prime nrover with thc capability of replacing the piston engine through-
out its whole range of applications. It also has the san-re nrulti-fuel
capabilities as the conventional pistorr engine.
The Wankel engine does not, in principle, suffer from scale effects. It
can l>c built to any scalc, in ilny sizc, ancl opcratc succcssfully. At tlrc
small encl of the scale, intenral friction coulcl beconte too great in
proportion to the po\r,cr output, but Wankcl engincs havc pro\/c(l prac-
tical cloln to a displacernent of 18.5 cubic inchcs (305 cc.) per u'orkittg
clrarnber. At the otlrer encl of thc scalc, llroblcnrs «lf flame front travel
may ultimatcly restrict rotor clinrensions. I-lowevcr, Curtiss-Wright built
ancl testccl an experimcntal \Vankcl cngine lravirrg 1,920 cubic inches
chanrber displaccmcnt. It relr) rvcll, arrcl proviclccl anrl>lc proof of the
feasibilitv of up-scalirrg thc enginc to clinrensi«rns usually rssociatcd with
nrarinc navigutiou rather tlran u,ith lancl tntttsportttiort.
Iivcn if thcre shoulcl l>c :ln optinrunr sizc for thc Wankel engine
rotor, thcre is not, nccessaril\', a linrit to the powcr rangc of Wankcl
engincs.'['llcrc is no lirnit to t]rc rrurnllcr of rotors thtt cfln bc conl-
bined in one power unit, so Warrkel engincs can be corrU)arccl with
multi-cylinder reciprocating piston cngines. |ust as a V8 delivers torque
nlore sttroothly'tlran it four-cvlinclcr piston crrgiuc, thc ruttlti-rotor
Warrkel enginc offers supcrior instrrntaucous torque ancl elinrinatcs the
negative torque periocls of the sirrgle-rotor unit. The most obvious
rncthocl of crcating a nrultiplc-rotor Wankcl cngirre is to acld rotors t«l
 TNTRODUCTTON 355

it is conceivable that a Z4-rotor crrginc rnight yield 6,000 horsepotver,

rvith extraordinarily lolv vibration levels, ancl only nominal torque fluc-
tuations, while allou,ing full sen'ice accessibility to all modules.

q,I?UI SXT'I The full spectrum of possible Wankel engirre applications must in-
clude the following:

Branch Application Ilorsepower Range

Industrial Portable wclding gcnerators r5-2,

engines Stand-by clcctric gcncriltors 5-50
Portable batterY chargers 5-25
Portable power tools l-10
Portable lifting equiPrnent 10-20
This is R. Ir. Ansdale's proposal for a modular multi-rotor Wankel engine. Three 0-40
shafts, each with four rotors, geared to a common output shaft, produce a compact
Industrial utility vehicles I

twelve-rotor unit.
Engines for Portable building equiPment l0-20
building & Concrete mixers 5-20
construction Portable conveyors I0-20
the original nrainshaft. The engine is so constructed that the designer equipment Portable conrpressors 540
has freedom to lengthen and add eccentrics to the mainsh aft, fit rotors Construction machinerY 10-50
on thern and build up a scries of outer housings separatecl by partitions Lawn tractors Gr2
Engines for
until all rotors are located inside working chambers. agricultural Lawn mowers 3-5
The most pronrisirrg layout of a multi-rotor Wankel engine was pro- equipment Soil & harvesting equiPment l0-50
posecl in I968 by thc inclel>endent consulting errgineer, R. Ii. Ansclale, in I-Iorticultural equiPntcut 5-50
an article pulllished in the magazine Combustion Engine Progress, en- Irrigation boostcr PumPs I 0-40
titled "l?easibility of I-Iigh-Output Wankel RC Engines" (RC stands Automobiles 40-400
Aircraft,
for Rotating Combustion). Ansdale proposed coupling a number of automotive & Trucks 100-600
twin-rotor Wankel engine modules, suitably geared, to a common out- marine engines Motorcycles I0-80
put slraft. Twin-rotor engines need only two main bearings and are Scooters 3-5
easy to asseurble as a unit. Using four or more rotors on a common Outboard motorboats 20-200
shaft wouJcl require internrecliate main bearings between the rotors, Auxiliary yacht engincs 5-r,
which would acld conrplications to both production and maintenance. Life boats I5-50
Placing tu,o twin-rotor engines errd-to-encl, with a spur gear on the shaft Light aircraft I 50-500
Auxiliary glider engines 5-I
that links them together, would result in a four-rotor, inline unit. If the
5

spur gear was nreshcd with a larger gear "below" it, whic]r was attached Engines for Small lawn mowers 2-'
to a new output shaft, the rnainshafts of the two combined engines household & Snowmobiles l0-60
would add torclue to tlre new output shaft. This new shaft coulcl be recrea tional Golfcarts 5-40
nreshecl with two adclitional spur gears, each of which would be con- equipment Snowblorvers z_5
nectccl to two nrodules. If the gears were properly arranged around e for Railroad locomotives 3,000-6,000
Engines
common output shaft, spacecl 120 degrees around for example, a highly heavy trans- ShiPs 3,000-6,000
cornpact IZ-rotor Wankel u,ould result. Two such engines could be portation and Pumping sets for atomic reactors 3,000-6,000
arranged in scries to make a }4-rotor installation. According to Ansdale, stationary uses Generator sets 3,000-6,000
756 APPLICATIONS
Branch

Miscellaneous
Altplication

Irire fighting pumps

Air-conditioning units (buses)
Rcfrigeration units (trucks )
Horsepower Range

I 0-40
3_5
ll
5-10
Special military purposes t 00-500
(generator sets, gäs turbine starter
unib, etc.)

The scope of this book is largely restricted to the application of Wankel

engines to automobiles. Some of the results of applications in this area
are found in the following chapters.
Th
Wanl«el

UNprn rHE TERMs of the standard

Wankel engine license contract, all licensees are given the benefit of
NSU's know-how, and they in turn pledge to share their experience and
developments with NSU and each other. The "partners" hold regular
technical conferences, and the open excltange of information is com-
pulsory for all parties in the sense expressed in the agreement. This
pooling of data and information relates to all aspects of design and
manufacture of Wankel engines ancl their parts. This includes un-
Iimited numbers of blueprints, samples and models, nranufacturing
drawings, specifications, parts lists, engineering data, test reports, design
information, performance charts and similar engineering, manufacturing
and technical data and information. In short, anything that is deemed
necess ary to the design, redesig., adaptation, operation, construction,
manufacture, production, assembly, maintenance, service and repair of
Wankel engines by each licensee, including equipment and tooling used
in the manufacture of such engines.
As part of this program, NSU development engineer Hans Paschke
was sent to |apan to observe Toyo Kogyo's work in 1962 ancl 196).
After his first ride in the Cosmo sports car, he sent a most enthusiastic
report back to Neckarsulm. It was not received with great cheer at the
NSU headquarters. NSU was not itself ready to start production of the
Spider, and such a glowing report frorn ]apan could only serve to stir
up more troublesome questioning from minority stockholders. Then von
Heydekampf, ready to make his epoch-making announcement of pro-
duction plans for the Spider, received news on |une 7, 1963 that Ma zda
(Toyo Kogyo) wanted to show a Wankel-powered car at the Irrankfurt
show that September!
358 APPLICATIONS BALANCE WEIGHT
AND PULLEY
ENO COVT,R ROTOR HOUS NG

STARTER MOTOR

The Toyo Kogyo announcement emanated from one of its vice presi-
dents during a press conference, and the remark was picked up by the
wire services and broadcast all over the world. The news hit Neckarsulm
Iike an earthquake. Von Heydekampf discussed possible countermeas-
ures with Dr. Henn, head of the legal department, and Dr. Henn sent
Toyo Kogyo a brief Ietter from which the usual polite phrases were
omitted, pointing out that Mazda's exploitation rights for the Wankel
engine were confined to East Asia and that consequently the company
had no right to exhibit its machines in Europe. This letter had the
effect of a cold shower in l{iroshima. The }apanese coolly replied to the
IFFThENTIAL

Ietter that the whole thing must have been a misunderstanding. Toyo
Kogyo had no intention of exhibiting a Wankel-engined car in Irrank- O'TIVESHAFT

furt in Septernber, but were arranging to show such a prototype at the

OHV
PLATE CLUTCH

Tokyo Autonrobile Show in October. Thus, the NSU Wankel Spider WATEß PUMP

did not have to share the limelight at the Frankfurt auto show with tcnrrtbn coNTAcr
a Mazda, and is recognized as being the world's first production car BREAKER

powered by a Wankel engine. OIL SUMP

rccgutntc

Cutawity clrnrvirrg of thc NSLI Wirrrkcl Spitlcr cttgitrc.

EXIT FROM ROTOR

lot -
Endtollc

FßESH OIL ENTRY

FROM OIL COOLER

MAIN BEARING
OIL FEEO

, MAIN RETURN LINE

ROTOR Orlo ,oroff
BEARING OIL FEEO

To noroR cAvlTY

The NSU Wankel Spider engine and drive train, ready for installation. Lubrication and rotor cooling system in the KI(M'502 engine.
 360 APPLICATIONS
TTre Spider body was a convertible version of the Sport Prinz, witS
some special features made possible by the use of a Wankel engine,
such as a trunk of useful shape above the engine. The car was strictly a
two-seater but also hacl generous Iuggage space behind the seats. It was
well designed and the bodywork was cleverly thought out.
Quality seemed
uP to or above Prinz standards. The convertible top was easy to put
uP or down, being a one-man operation taking about half a minute. It
was stored under a metal panel rather than under one of those stretch-
tyPe vinyl covers usually found on American convertibles, and was
completely out of the way. The chassis components for the Spider were
taken directly from the Sport Prinz. The architectural Iayout remained
unchan$ed, with the single-rotor Wankel engine positioned behind t6e
rear wheel centerline, iust as with the stock piston engine, Steering,
front and rear suspension, and transmission remained essentially the
same.
My first experience driving the NSU Spider came in 1965 when I
borrowed one from the importers and distributors in New York, Trans-
continental Motors, Inc., and took it to Bridgehampton race circuit on
Long Island for a test. There were n'rany conventional aspects to driv-
ing the NSU Spidcr, just as was the case with the lto-80 ancl the Mazcla
R'100. You could actually perform all the normal driving functions NSU Wankel Spider cooling system. I - Cooling fan. 7 Radiator. ) = Branch
without knowing the car had a Wankel engine, and find t[at it would to heater. 4 - Heater element. 5 _ I{eater fan. 6 : Hot line.7=Returnline
from tfiermostat. 8 _ Therl-rostat. 9 - Water punrp. t0 :
respond by doing more or less what you woulcl expect. It started up i1 lleat exchanger (oil/water).

ulrt I0 txsHt

Cross-section of the KKM- 502. Cutaway drawing of the NSU Wankel Spider.
 SPIDER
I

THE NSU WANKEL 363

I

the normal way, by turning a k.y, but as soon as it started running

I

you knew it was not a conventional piston engine. Its silence, lack of
I

:I
I

vibratiotr, and the 7,000 r.p.m. redJine on the tachometer immediately I

il
told you that this was somethirg different. More than the redline itself, rl
it was the striking ease and rapidity with which the needle would climb I

there that impressed me.

The driving position was unusually correct for such a small car, in that
the front wheel housing did not rob the driver or front seat passenger
of too much foot room, and the stecring colun'rn hacl a very slight tilt.
I expccted the steering to be both light ancl quick, and I rvas not dis-
appointed. It was also uncommonly precise in its action. The car could
be placecl within an inch of its intendecl path at any speecl.
'\- The brakes were powerful and stable. But the most remarkable thing
. ".
\
.r\

-"'.\--
-- \-
-\. \ ..: about driving this car on an extremely fast road circuit was the handling.
I have driven a lot of rear-engined cars with swing-axle rear suspension
)..

The NSU wankel

spider showed such reliability, that German tourists showed no
hesitation about taking them deep behind the Iron Curtain, under the same conditions, and expected a certain amount of over-
far from parts and serv-
ice facilities. steer with the NSU Spider. But it was perfectly balancedl There was
no breaking loose, front end or rear. The whole car stuck to the chosen

It was in international road rallies that the NSU Wankel Spider showed its mettle.
The NSU Wankel Spider. Its reliability rccord is even nlorc inrprcssive than the number of victories, with nearly
all starters finishing.
I
 TI{E NSU WANKEL SPIDEI1 36'
gone on to 9,000 r.p.nl. or n)ore if the throttlc hacl bccn kcpt oPen
a little too late beyond the reclline. Rut over-rel'ving was not etrcotlragcd,
sipce tests hacl shown that engine wear ancl oil collsulrtptiort increasc«l
at an alarming rate when shaft r.p.m. excceded 6,000. There was no
torque below 3,000 r.p.m., so that the shift hacl to bc uscd frequently.
For any unexpected opening irr trafEc, a clorvttshift was trecessary t«l
give proper acceleration. J'h is was not oll jectiorrablc to the clicntele
that was attracted to the NSU Spider, since they were performance-
oriented enthusiasts and were used to small piston engines with sirnilar
torque characteristics.
The long linkage between the floor-mounted shift lever and the
actual transmission made some lack of precision in the gate inevitable,
but shifts were fast because the syrrchromesh was unbeatable, and the
shift movements were Very, very light. In turnpike drivirg, the car was
more than able to hold its own. Acceleration was more than adequate as
Iong as the engine could be kept above 4,000 r.p.m. Engine noise levels
were extremely low at all steacly speeds, and cruisirrg at 70 m.p.h. in the
The_ racing career of the NSU wankel spider began in 1966, when private owners
combined with NSU engineers acting independently, to get more power and speed Spider was as pleasant as in some far more powerful sports cars.
from the single-rotor engine.

Iine around each curve, in an attitude that can best be described as SPECIFICATIONS
neutral with an understeering tendency. With a front roll center Io- MAKE NSU
cated l.l8 inches above ground level and the very high rear roll center MODEL SPIDER
on the Spider (11.37 inches above ground level), body roll was prac- Year introduced 196+
tically non-existent. Front wheel bump travel was more than aclequate, Year discontinued t967
at 5 3l inches, and rear wheel bump travel was an even more generous Price $2,979
Type of body Open trvo-seat roadster
5.56 inches. As a result, ride comfort was excellent for a sports car, but
Type of construction Unit body
still would be deemed harsh by family car standards. Because the car
Driving wheels Rear
was so light, it bounced a bit on big bumps.
Power unit position Rear
f)irectional stability was strong for such a light car, mainly due to the Curb weight 1,543 pounds
very high front wheel caster angles (12 degrees) combined with a swivel Weight distribution front/rear 4+/56%
axis inclination of 7o)0'. Still, steering effort, with its 16.25:l overall Power/weight ratio Z+.2 pounds per horsepower
ratio, was quite Iight. Static front wheel camber was ? degrees positive. Fuel tank capacity 9.2 gallons
The rear suspension was designed for zero static wheel camber, but Fuel tank position Front
tended to go positive under transient conditions. The tires were small, Power unit Wankel
but they were radial-ply Michelin X, and I could not have felt more Number of rotors I
confident. Chamber displacement 30.51 cu. in. (498 cc.)
Equivalent total displacement 6l .0 2 cu. in. (996 cc. )
Acceleration, too, was a surprise. The first thing that had to be learned
Compression ratio 8.6: I
about this car was to watch the tachometer, simply because the engine 6{ horsepower
Power output
gave no hint of audible protest when it reached its rev limit. On the
at r.p.m. 6,000
contrary, it seemed happier the faster it spun, and no doubt would have Torque 52 foot pounds
366 APPLICATIONS TI-IE NSU WANKIiL SPIDER 367

SPECIIIICATIONS SPECIFICATIONS
MAKE NSU MAKB NSU
MODEL SPIDER MODEL SPIDER
at r.p.m. 2,500 Test Results
Carburetion system One Solex l8/)Z HHD Acceleration times
Ignitionsystem Coil and battery 0-30 m.p.h. 3.8 sec.
Cooling system Water and oil 0-10 m.p.ll. 6.3 sec.
Clutch Single dry plate 0-50 m.p.h. 9.8 sec.
Transmission Irour-speed all-synchromesh 0-60 m.p.h. 1 J.8 sec.
Cear ratios I 3.08:l 0-70 m.p.h. 20. t scc.
2 1.77 :l 0-80 m.p.h. 28.2 sec.
3 I .17: I 0-90 m.p.h. 44.0 sec.
4 0.8 5: I 0-100 m.p.h.
5
R ).43:l
Top speed I n m.p.h.
2 47 m.p.ll
Final drive ratio 1.13 l ) 7r nr .p.h.
F ront suspension Upper and lower A-frame conhol arms 9' m.P.h.
4
with stabilizer bar 5
Front springs Coil springs Avcrage fuel consumption 2u nl.P.g.
Rcar suspension Swing axles and semi-trailing arms
Rear springs Coil springs
Steering system Rack and 1>inion
Turning diameter 2 8.8 feet
Overall steering ratio 16.25:l
Turns, Iock to lock 2.8
Brake system ATE-Dunlop discs F; drums R
Disc dianreter F inches
Disc dianreter R
Drum diameter F :-
Drum diameter R
Lining area
Swept area
Parking brake _
Tires Michelin X
Tire size t), x l2
Wheelbase 79.53 inches
Iiront track 49 inches
Rcar track 48.3 inches
Overall length l4l inches
Overall width 59.84 inches
Overall height 49.61 inches
 THE NSU Ro-80 369

passengü car such as NSU was proposing to build. The Wankel Spider
18 was a small sports car without the luxury and refinement of the Ro-80
and, while it had the same problem, it was considered tolerable in this
type of car.
Nobody at NSU made any secret of the fact that a completely new
car was being designed around this engine, but no details were available
on the proiect that was to become the Ro-80 car. It was officially intro-
duced a little over a year later, at the Frankfurt auto show in Septem-
ber, 1967 . Production began in October of that year. I saw a few of them
on European highways from time to time, but it was not until 1969 that
The NSU I actually got to drive an Ro-80. By then, NSU had produced over IU

14,000. T.

Ro-80 The Ro-80 is not a sports car, but an extremely well-equipped and
ü
rl
trj
l:i
H

well-finished four-door, four-passenger sedan. The first thing you notice ii

when sitting down behind the wheel is the nearly flat, tunnel-free floor,

ECKARST'LM IS A SMALL INDUSTRIAL

town in the valley of the river Neckar, about an hour's drive north of
Stuttgart in southern West Germany. I went there in f un€, 1966, to look
at the twin-rotor Wankel engine that NSU was experimenting with. A
young liaison engineer named von Manteuffel let me see the new engine
undergoing dynamometer tests, showed me the dismantled test engines,
and explained the design features. Then he took me into the garage and
pointed to a |eep-like vehicle. It was a DI(\M Mung ?, a four-wheel-drive,
off-road car built by Auto Union from 1954 to 1965, This was the test
car for the twin-rotor NSU Wankel engine.
At von Manteuffel's invitation, I clambered in behind the wheel. The
engine started easily but seemed noisier than the single-rotor unit that
powered the NSU Spider. That was mainly because the engine was in-
stalled in the front of the chassis, without any kind of sound deadening
materials around it. It had excellent power and gave the Munga a higher
top speed than its chassis was ever intended for. It pulled smoothly from
quite low revs, but had little torque below 4,000 r.p.m. It took a lot of
shifting to drive it on the hilly country roads in the Neckar valley.
In town traffic, we experienced what von Manteuffel described as the
biggest drawback to this engine. It was a snatch in the drive train that
occurred at low speeds on a closed throttle, such as when easing towards
a street corner. The result was a stop-and-go bucking motion in the
whole car, which disappeared when the clutch was diseng aged. This had Ro-80 instrumentation is clear and sirnple, the entire interior package having been
designed with full regard for safety. Spacious and comfortable, the car is also very t,
to be cured before the engine could be iudged suitable for a high-grade well appointed and finished. t'ir
si
i.iri
..
r!: r
,t
,.!
rl
i; .i

"!
 TI IE NSU l{o-80 371

eration from a standstill is good, but not startling in vicw of the en-
gine's 136 horsepower rating. Iior brisk driving at speecls under 50
nl.p.ll., you Iravc to n,ltch tlrc tlclr«lrtrctcr ltrtd usc thc gcllrs.
T'lrc Saxonrat transnrission usccl on tlrc I(o 80 is a "sclcctivc auto-
matic" made by l,'ichtel & Sachs to NSU specificntions. It hus it hyclrau-
lic torque converter witJr an elcctro-pneumatic clutch and a nlanual-shift,
three-speed gearbox. A rocking switch in the gearshift knob disengages
the clutch instantly when you grab it to shift. There's no need for a clutch
pcclal. This scnri-automatic transnrissiou wa.s the solution to the "slliltch"
problem that occurred in thc stanclarcl-transnrission Spiclcr ancl the
experimcntal Nzlunga. By interposirrg a hyclraulic torque convertcr be-
tween the road u,hecl.s and the cuginc, thc drive train ollttinecl tltc
neccssary cushioning. Naturally, the cngine clocs not give nruch braking
effect on a closed throttle, but that is not consiclered significant. 'f'hc
All drive trai, components are located aheacl of the passenger compartment, Ieaving absence of a clutch pcdal does not mean that the Ro-80 can be driven
nll spilcc frce for occupants and luggage rear of the cowl structure. as a car with autonratic traursrnission, thc drivcr still has to shift for hirn-
sclf. Thc car clocs not, howcvcr, posscss tlrc sp«lrting clurracter tlrat ir
sunk six inches below the sills. Wide, inclivi«lual seats arc bocly-con- 1>urcly rnccltarricll four- or fir,c-.spcccl flo«lrshift u,«lukl givc.
toured to cradle you on corners, and backrest angle is fully acljustable. The car will start off in second, or evcn third, gear, but only at a
In addition to more-than-gcncrous forc-and-aft rrror,.,ncnt, tfic c]river's crarvl. It is possible to stay in the original gcar:rs Iong as desired without
seat also has height acliustrnent. J'his is done by ingenious clouble run- ever shifting. The lruy mcthocl is O.K. for lcisurely clriving, but for bcst
ners, with the ]orver pair sliding up ramps to raise the seat. T'5e broacl results all the gears have to be used; the fine synchromesh gives quick
seat back ancl the firm cusltion ntake for tirelcss clriving over long clis- and smooth shifts. But, you have to remember not to rest your hand
tances. TIte steering wheel is large and has a paclded ;safety" centcr. on the geu knob. The moment you touch it, the clutch disengages.
The instrument panel is well laid out, with big ,llrh grouped in front of The long, floor-mounted stick shift works the three-speed gearbox with
the driver. A small padded hood above thern prevents reflections on the secure and precise mo\/ements. Iiour speeds instead of three would have
winclshield. improved perforrnance, although to sonre extent the torque converter
The vital switches are carried on stalk.s branching off from the stecring makes up for the Iack of low-range torque. The Saxomat is cheaper to
colunrn for finger-tip control. A four-way switch «ln the rigfit o1>cratcs make than a fully automatic transnrission ancl it absorbs less po\\,cr.
tlle tlrree-tone horn, two-speed wipers ancl electric was|ers. l}e left-hancl Because a full automatic requircs a large-capacity hydraulic punrl> and arr
stalk dims the headlights and handles the turn signals. Small'controls expensive controJ unit to actuate the bancls ancl clutches, whiclr can
on tJre instrument panel inclucle a pull-out rnastei Iight switch and a take as much as I 5 horsepo\4'cr to operate even at idling speecl, the Saxo-
srvitch for four way emergency flashers. An easy-to-read I30 m.p.h. mat is a good cornpromise for snrall-clis1>lacenrent engines. NSU sees
speedometer is matched with an 8,000 ,.p.m. tachometer. this low-cost arrangement as thc transmission for futurc cars with srnall
The engine sounds smooth, even when idling, but has little torque power reserves. It is already available on the Saab, Fiat 850, Simca 1000,
below 3,500 r.p'm. Acceleration at lower .ngin. speecls is definitlly and Volkswagen 1500.
sltrggish. On hills, there is enough power to maintain high speecls, bJt In highway driving, the Ro-80 is incredibly quiet ancl deceptively fast.
little Power reserve for acceleration.Top gear acceleration Ir poor in There is no wind noise at all. The body is so beautifully wincl-cheating
relation to the claimed horsepower. To get performance, the g.ri, must that it just skips through the air without causing any auclible disturb-
be used frequently. Top speed is remarkabiy high, and the Ro-80 rvill ance. The engine and clrive train are quiet, too. There seerl-rs to be a
cruise in uncanny silence at 100 mlp.h. At top speed, the engine is still complete absence of stress ancl vibration, even rvhen clriving at very
quiet, and there are no signs of stress or vibration. Full-thrättle accel- high speeds. f)irectional stability at highu,ay speecls is excellent, prob-
 3
Im
I
o)
.s
.=
-c,

§5m
o
o
.=
o
TU

F
am

rverter
ement
roint

knr,h

This performance graph relates tlre actual perfornrance potential of the Ro-80 to the
ideal tractive effort h1'pcrbola at wide open throttle (WOT). Under 45 k.p.h.(28
nrph) its potential is unconrfort:rbly distant frorn the ideal curye, but at all normal
driving speeds, its potential is only a ferv points fronr the ideal curye.
I
ably mostly as a result of its front u,heel drive. All earlier NSU cars \\'ere
of the rear-engine ty'pe. \\ras there a risk of unacceptablc rveight dis-
tribution with the twin-rotor engine urounted in the back? Probably not.
NSU chose to position the Wankel engine ahead of the front rvheel
Full-scale wind tunnel tests were made with the NSU Ro-80 axis, thus freeing a maximum amount of interior space for the passcngers
body, to achieve low
aerodynamic drag and high stability in crosswinds. The
wool tufts-several hundreds and driver. The front mounting also enabled the NSU errgineers to
of thcnr -per car-show tl,e «lirection of air now a multitude of points
,t simplify the coolirg systern ancl radiator installation. NSU also fclt that
on the
surface of the body. The Ro-80 lras both l;; ä;;g ,nd low lift. thc driving wlrccl.s slroulcl carry nrost of tlrc wciglrt; tlrcrcforc oncc tlrc
 Compact on the outside, yet full-size.inside, is how some critics have
summed up
the Ro-80. Tlrc bllv ltas e. tremely rvicle cloärs, yel the rear ääor does
the fender well. Glass arca is g.närous, and ,iriuiiity 1 ä..uent in not cut into
all directions.

forward location of the engine had been cleciclecl upon, front wheel drive
followecl as a nratter of course. NSU never consiclered rear wheel I
dri'e
for their front engined car. Doing away with the long driveshaft meant NSU Ro-80 shows high cornering ability, with moderate roll, on this narrow country
a Iower fo9r, and a possible reduction in frontal ,r., and aeroclynamic lane. Superior roadholding means extra safety in everyday conditions, and can be an
important difference in an emergency situation.
drag. It also meant a lower center of gravity, with the consequent
promise of improved ride and handling. A Iow center of gravity gives
a
good ride because carrying the load of occupants and lugIag.
ttär., to
the ground means snral]er sic]e movements in the vehicle-*f,.n crossing Iack of it. The reason is that putting tlre clrivc through the front wheels
asvmmetrically applied bumps. It also permits a lorver roll center without produces an inherent sclf-ccntcring tcnclcncy that is effcctive at all
excessive risk of body roll on curves. Elirninating the driveshaft
also cut speeds. It also means that front u,hecl clrive cars unclcrsteer.
procluction costs. I
The kind of understeer that comes rvith front rvheel drive makes for
With a rear engine, the engine and transaxle fonn a unit. This unit is excellent safety. It nroderates the curve path if the clriver nrakes a sudclcn
mounted in tlre tail of the clrassis, and connectecl to the driving wheels steering wheel movement, or accclerates or brakes too hard for thc curvc
via a transaxle and short driveshafts. The same situation exists with indicated by the steering angle. Such features may not givc the greatest
front wheel drive, with the extra complication that the wheels also are maneuverability, but clo give rnost drivers a fceling of bcing at case whcrr
steered. Irront wlleel drive offers better traction uncler most conclitions.
driving such a car. Irront rvhccl brcakawily, at thc Iimit of aclhcsion, is a
The need for directional stability in NS[.J's new high-speed car also stable motion with a self-diminishing tenclency. Getting off the throttle
favored the choice of front wheel clrive. Iiront wheel drive usually
brings will bring the car back on a sltortcr tunt raclius, thc car u,ill slnrv clon'rr
improved directional stability, while many rear-engine cars suffer
from a and the driver will retain control.
 l5 tt.8 ln.

F___ 5ft.01n. _-+

passenger compartment'
The Ro-g6 was designed arouncl a comfo*abJv dimensioned-
-5ft.01n. I The next consiclerations *;;;;ginä bulk, drive system, cooling lequirements,. basis of
wgigh!
fixed on the
distribution, and target price. T[re overali dimeniions äte then
.-i
sF all other design Parameters.
C
t

4 tr. l0 lt2ln._-+ 4 ft- E ll2ln.---)

i
o)
.d
)o

Dimensional sketches of the Ro-80. a = 38.8 inches. b 38.6 inches. c

- 47.3 :
inches. d -- 16.0 inches. c ll.8 inches. t : 19.3 inches. g ll.4 inches.
=
-
h = l8.l inches. i - 17.? inches. i = 20.5 inches.

The need for power steering with the front wheel drive Ro-80 stenrs
from the fact that a Iarge percentage of the car's weigltt is resting on t|e
front u'heels, the same feature that gives superior traction. The drive
wheels on a front wheel drive car always point in the same cJirection
the car is being steered. fnertia in the vehicle itself thus causes under-
steer, ancl a clifference betw'een steering angle and clirection of travel, but
traction is always aPplied in the direction the front wheels are pointecl.
Because the car is never in any doubt about the clirection it is headecl, which
its structure can be ligltter, and the need for suspension reinforcernents The hydraulic power cylinder for the steering is driven by a vane'tYpe PumP,
is drivän by tlle engine.'l'5e power c1,li,der is not integral with the steering gear,

but deliveri hydraulic pressure to t5e stcering valve at the head of the
is smaller than on rear wheel drive cars. steering gear'
)78 APPLICAT'IONS 2-DruREE luconez
Tlre Ro-80 is a well-balancecl car with mild understeer. When pressed
to the lirnit, it rvill not "plow" Iike a Toronado but drift on all four like
a Lancia. It is not vcry particular about taking a corner with power-on I

or power-off. Most front wheel clrivc cars unclersteer with power-on, and L

oversteer with power-off. With its low center of gravity and wide track,
there is little roll when taking curves at highway speeds, but on sharp,
slow turns, the car lcans over more than expected. With Michelin XAs
tires, rvlrich give an exccllent grip ancl unfailing sicle bite in all kincls of
u,eather, the lto-80 can bc taken around curves at arnazing speeds, in
unusual stability, conrfort ancl safcty. The steering is power-assistecl and
is gcarccl to give 3.7 5 turns from extreme left to extreme right. It is I
possibly lightcr than somc drivers like, but is extremely accurate. It rvorks
rvit]r unusual smoothness, without loss of precision in its response, or any
wH.' a RAI r sERyO -ASs t sT E o
Ioss of feedback. BRAKTNO 5y6rEn',t

The steering wheel is cornpletely free of drive train vibrations and masses in the Ro-80, its general
road shock. Return is firnr ancl pleasant, making for relaxed cornerirg This sketch shorvs the distribution of the principal zones for occupant protection
construction, and the prouirion of front lrr,ä ,.rr-crush
even under high centrif ugal force conditions. The steering gear is trurrk absorb the impact energy,
in front and rear impact ,itu^tionr. The hood and
intact.
compartme,t rernains relatively
mounted on the cowl structurc at the rear of the engine compartment, *,hile the passeuger

far back frorn the front of the car. A rack-and-pinion gear, pivoted at its

SHOCK ABSOREER

{
I
! STEERING ARM
t
t
tI
I
I
T UB CARRIER
$

i
I

LOWER CONTROL
ARM

The rack and pinion steering gear of the Ro-80 is self-adiusting and offers resilient
and phase«l application of stccrirrg force. IIigh stresses in the steering geü occurs lr{acPherson type, in which a tall
only when thcre is no hydraulic servo assist (i.e., when the engine is not running) . Iiront suspension of the NSU Ro-80 is of the
spring leg gives the wheel hub its upPer locating Point, and no upper control arrn
In normal clriving, it is the driver's steering effort which controls the valves that
adnrit hyclraulic pressure to the servo cylinders, which are pin-iointed to the drop arm. (A-frime) is needed.
 380
APPLICATIONS COIL SPRING
outer end, is coupled to-a lo-ng drop-arm. The
hycrrauric power cyli,der
is placecl on the other side of the arm,
which in turn is linked to trre two
ms on each wheel.
endow the Ro-90 with such exccllent
without belonging in the ,.break-
of the Macpherson type,
ension is

spring reg. rhe conkor il,""n::#'r"j I't,ält,,X1r'jl':i::T:[?] SPRING LEG

and supports the wheer 1,T

I

hub. The spring Ieg is mount.a in-r-ro.r..t on

the spindle support arm and its a spring tower WRIGHT.SENSITIVE
BRAKE FORCE
PROPORTIONING UNIT
Between them, I
eel. The spring

2/2
ber. rhe roil center
above grounrr lever. For added roil stiffness,
f[ää':ilär? .FRAME CONTROL ARM

-incbe.s
used' This stabirizer con-sists of a s1 eciar
,-rt.uiii., u* i,
torsion bar, which is attached
to the left.and right A-frames and whose center piece BODY SUB FRAME MEMBER
is reversibre. It
turns elastically
-and
resists the twisting motion wrren one
wheer is €
raised and the other
tä tt" maÄr'rr*ry, i,"rp.r,
lowcred in relation
when taking a curve. when.the springs of
botrr *t."t, or" i,i'oätion, it I .t

revolves as a whore around its center "piece

without .äi;; ,"irr,irg ," I 'l'he Ro-80's rear suspcnsion rclics on a lower A-frame and a spring Ieg incorporating
the spring rates. a hydraulic shock absorbcr. T"he lower A-franre is linked to a u,eight transfer sensor,
. ltle rear iuspension is designed on the same principres, but, because which regulates brake force distribution to the rear wheels. This reduces the risk
this is a front wrreer drive car,lhe rea, *helrs of premature rear wheel lockup during hard braking. This suspension system keeps
have ,"ärii", than the rear wheels parallel to the car at all times, which means that single-wheel bumps
to carry the load-trrey-are
iust arong for the ride.
dead I'beam axle and Ieaf springs rlke a trairer
Rrth.;;;;
"irrlr
rsing n
I
I
ncvcr produce any unclcsirablc cambcr changcs. On the other hand, the camber angle
changes with body roll, which adcls a cert:rin roll steer effect.
lrna rrroreä'bv ,o*e i

constructors of front rvheer drive cars), the NSu'engine..,

independent rear suspension. The *freer, are
a..iä.a * I
t
weight to allow bctter adhcsion and traction. The rear «liscs are canied
i, fir""-if
{

,.-i_
I
herd I on the wheel hubs. Iior parking, there are cable-operated clrums insiclc
trailing arms and almost verticat spring legs. The 1

arms are Iike A-frame i the rear discs. Thc brakes arc powcr-assistccl and arc ils scnsitive as
control arms turned around so thät tr-re
fivot axis runs across the car those on an American car. They sometinres tencl to conre on too strong
r_

instead of lengthwise. Tire spring legs are {I

front ones.
ärmost exact dupricates of the
D

for the pedal pressure, but arc easy cnough to nroclulate so tlut the
I car decelerates smoothly at an even rate and comes to a complete stop
sulgelsion systems on the Ro-g, ensure precise
,Tl1e guidance of the I
at the intended point. Brake action is always prompt, almost sudden,
wheels-all four of them-when hitting bumps
or pothores as weil as
J
t
in-cornering situations,with minute chaiges I
and effective. Brake force distribution betrveen front and rear is flaw-
in track, camber less. There is little no.se dive, and r)o threat of premature rear rvheel
All four wheers are linked to the car soärat tr,.y pr*ia, and caster. .t
t

resistance to centrifugal force.

r,igi tterrt locking on hard stops due to a load-sensing valve that limits pressure
t

If handling is an important erement of vehicre safety, to the rear brakcs accorcling to u,eiglrt. The car stops frour 60 nl.p.h. in ,I

more important. The Ro-g0-has big disc brakes

brakes are even under 150 feet, without any thrert of side pull or locking. Iirorn 90
on ,il'iour rurr..ir. rr," m.p.h. it stops in just o\rer 300 fcet. Strch brake perfornrance is con-
front discs, which do two-trrirds oi the braking,
flanking the transaxlq wrrere they're .*por.d
,.. *orntlä inboora, sidered well above average.
tä a flow or ,ir. In a land like German1,, u,ith its high-speed Autobahnen, normally
Because they are away from t'e wheels,
th.r l;;; ä.-;;;rrr*
"ooiirg
unrestricted by speed limits, the ability to cruise near maximum speed
is den'randed of even high-po\r,crecl cars. To make high-speed cruising
 THE NSU Ro80 38)
The general design goals for this vehicle were the production of a
car u,ith the same seating ca1>acity, Iuxury aucl perforrlrAncc as the nrost
popular Mercedes-Benz and BN,{W seclans, while taking tlrc fullest
possible advantage of the lightness and conrpactness of the Wankel
engine. NSU saw the possibility of making a lightcr ancl lowcr vehicle,
with a better power-to-weight ratio, tharr was available frorn the conr-
petition. NSU had no expcrience with cars of this caliber, and they
were strangers to the high-priced markct, too. But there was one thing
more-NSU hacl no existirrg parts that they rrright bc tcnrptecl to in-
clude in the new clesign for reasons of procluction econonry and, äs a
result, they were able to start with a truly clcan sheet. The entire Ro-80
package was designecl around the \\/ankel engine, in contrast with the
earlier NSU Wankel Spicler.
-fhe Spider was basically a convertible-

ü
t
Irour'wheel powcr-assistecl disc brakes are stan«lard *
on t
but the systcm cliflers from -the
.t

t
$

harr rncrcly sellarating froni ,na ,ü

'il

;f i:'i,,?"r31,tffiY
t
i
:l
pnrnary systcnr docs not affcct the secondary i
of nraximum braking cffcctiveness. j'
1

t
I
I
i
comfortable, qtt:et, ancl economical, NSU ,\
decicled when the design I

goals for the Ro-80 were set that a "streamlined"

;l

body
T

of extreme importallce. The basic shape of the car was going to be

196l by a gracluate of tlre Schule Für i-or*gestaltung
*rr- ä.rrg".a in
now enrployed by NSU. Tlren began four
in UIm, w5o is
years of rvind tunnel testing,
in collaboration with the Stuttgart Polytechnic j

Institute. Wind tunnel t'

rvork began with a one-fiftlr t.rl. nroclel,

and full-scale tests startecl in
Atrgust of 1963. As was the case with Daimler-Benz i

and the C-lll, I

these tests werc nlacle not only for clrag,
but also for lift and stability i
in crosswittcls. After a number of minor nrodifications, p

the final car had t

a drag coefficient of 0.355, rvlrich compares I

favorably with many popu-

Iar shapes and aPproaches the bc.st ever
real ized, in ihe way of a four-
door sedan:

19 30 Irord Nlodcl A .g3Z .

1960 Volksu,agcn . 50
1960 Chcvrolct Con,air .q3
q

1960 Pontiac .Sj These three designs were considered rvhen the final prototypes for the Ro-80 were
1966 Citroön DS Zt to be built. The first and second designs wcre rejected for lack of practicality.
.326 Number three was the choice.
384 APPLICATIONS THE NSU Ro-80 385

bodied Sport Prinz chassis, rvith a single-rotor Wankel engine merely SPECIIIICATIONS
taking the place of the stock Prinz,piston engine.
NIAKE NSU
The engine used in the Ro-80 is the KKM-612. It weighs 280 pounds I\4ODEL Ro-80
and develops 136 horsepower at 5,500 r.p.m. The unit is only l8 inches Ycar introduced 1967
long, I6 inches high and l7 inches wide. It was small and light enough Ycar discorrtinucd
to be installed in any conveniently available space, and for reasons we Price $3,7I2 in thc honlc rllarket
have already discussed, NSU decided to place the power unit in the T1'pc of body 4-door +/ 5 seater sedan
nose of the car. Because of the small bulk of the KKM-612, the hood T),pe of construction Unit bodv
is short and has a low profile. Because of its low rveight, positioning Driving rvhcels liront
the engine overhanging the front wheel centerline posed no problems in Pou,er unit position ITron t
terms of weight distribution. It raises the polar moment of inertia, Curb weight 3,035 pourrds
Wcight distribution front/rear $/37%
which in turn strengthens the inherent understeering characteristic of Porvcr/wcight ratio 223 pounds pcr horsepower
the vehicle. Understeer remains, however, well within acceptable limits. liucl tank capacity 2L9 gallons
Now that the Ro-80 has been thoroughly tested, NSU has clecided I,-ucl tank position Il.car
to increase daily production in l97l to some 75 units per day. Poler unit Wankel
Nurnbcr of rotors 2
Chaml>cr clisplaccnrent 30.i cu. in. (497.5 cc.)
Iiquivrtlcnt total displacenrent l2l .? cu. in. ( I .9 861 liters)
Conrpression ratio 9.0: I
Pou,cr output 136 horscpower
at r.p.m. 5,500
Torque 117.2 foot pounds
at r.p.m. 4,500
Carburetion system One Z-stage Solcx l8ß2 HHD
Ignition systcnr
Cooling system Water and oil
Clutch Singlc dty plate
'f ra ns rn ission Torquc cout,crter with manrlal three-
spccd gcarbox
Gear ratios I 2.0 56: I
z 1.208: I
3 0J88: I
4
5
R 2.105:l
Final drive ratio 4.857 :l
Iiront suspension NlacPherson spring legs and A-frame
lorvcr coutrol arms with stabil izet bar
I?ront sprirrgs Coil spring and tclcscopic shock absorber
Rcrrr suspensiou iling A-f ranre control arms and
Sc:nri-tra
spring legs
Rear springs Coil spring and telescopic shock absorber

With the Ro-80 in the background, the K-70 shows a clear family relationship. Stccring sl,stem Porver-assisted rack and Pinion
Both have front wheel drive. The prototype I(-70 had a four-cylinder water-cooled Turnirrg dianreter 38 fcct \Yr inches
cngine, but the engine conrpartment had room for the KKM.6l2.
 386
APPLICATIONS
SPECIFICATIONS
MAKE
I\{ODIiL
NSU
Ro-80
r9
Ovcrall stccring ratio 18.3:l
Turns, Iock to lock 3.9
Brakc systcm ATli-Dunlop 4-wheel disc brakes
Disc cliarncter F Il.l8 inches
Disc diamcter R l0 .71 inches
Drum dianreter F
Druln cliarneter R -6.5 inches (parking brake only)
I-,ining arca 13.07 squarc inches
Srvcpt fl rcA
Parkirrg brake -Mcchanical cluo-servo on rear The
wheers
Tires
Tirc size
Michelin XAs
17, HR t4 M azda Cosmo
\Vhcclba.sc 112.6 inchcs
Iiron t track 58.27 inches
Ilca r track 56.16 inches
Ovcrrrll lcrrgth I 88.19 inchcs
Overall rvidth 69.29 inches
Ovcrall hcigh t 5r.51 inches Ry rrrE riND on 1962, the clesign ancl
basic hyorrt of thc Cosnro Sllort Il0 S \\,crc clctc:rnrinccl, rtncl crtrly iu
'lest Resu/fs
tlre following year, dctailccl clesign work conrnrenced full scale. In Oc-
Accclcra tion times tober of 1963, Toyo Kogyo clisplayed two types of rotary engines-a 24
0-30 nr.p.]r. 4.3 sec.
0-40 nr.p.h. cubic inch (400 cc.), sirrgle-rotor cngine an«l a 18 cubic inch (800 cc.)
6.0 sec.
0-50 ur.p.h. twin-rotor engirre. In 1964, a trvin-rotor cngine with a single chamber
8.8 sec.
0-60 rn.p.h. 12.8 displacenrent of 400 cc. ancl a four-rotor crrginc of the same charnbcr
sec.
0-70 nl.p.h. I6.0 scc.
size were displayed at the I I th Tokyo l\{otor Show.
0-80 m.p.h. 22.4 sec. The production Cosmo engine is a water-cooled tandern twin-rotor
0-90 m.p.ll. 32.3 sec. unit, not unlike the NSU KKN{- 612 in basic configuration, having
0-100 m.p.h. 52.0 sec. 491 cc. chamber displacement. Here are its vital statistics:
J'ol, spced I 46.6 m.p.h.
80.2 m.p.h.
Radius l0 5 mm.
Eccentricity (e)
2
15 mm.
) I I1.8 nr.p.h.
4
Rotor width 60 mm.
R/e ratio 7 zl
+e
5
Avcragc f ueJ consumption R 120 mm.
2l m.p.g.
The Cosmo engine puts out Il0 horsepower at 7,000 r.p.rn. and
weighs only ZZ5 pounds. It
is also small enough to fit comfortably in
existingMazda cars-length is 70.2 inches, width Z3.l inches and height
2l .4 inches. The engine housing is built up as a sandwich structure
composed of two rotor housings and three sicle lousings, which are
clamped together in one body by 19 tension bolts. The rotor housings
are made of rnolded aluminurn alloy castings for higher radiation effi-
ciency. Their strength at high temperature has been improved by heat
 THE MAZDA COSMO 389

prns also serve as buffers against oscillating loads acting on the rotor
gear.
By Iate 1963, the first layouts for the Cosmo sports car had been
completed, though they differed in some cletails from the final design.
The fuel tank was shown uncler the trunk floor instead of above the
rear axles, where it latcr was located, and drunr brakcs originally were
specified for the front wheels. Using that first twin-rotor engine, the
prototype Cosmo was complete and running by Iuly, 1964. It then
weighed 1,980 pounds; the engine weighed 198 pounds without the
transmission. Field tests on the N{azda ll0-S vehicle were performed in
parallel with engine bench tests. Early in L965, 60,000 test miles had
been accumulated.
During the next 12 months, Toyo Kogyo switched from peripheral
ports to side intake ports to improve the engine's torque and low-spee«l
performance. The revised engine idled at 500 r.p.m. and pulled smoothly
in any gear from 800 r.p.m. Peak torque occurred between 3,000 and
4,000 r.p.m.; however, peak power and overall performance were sacri-
ficed.
As a countermeasure, a new engine was designed in 1965. This engine
had 120 cubic inches displacement instead of the original 98, and was

The lvlazda Cosmo I l0'S was first shown to the public at the Tokyo
in 1966, Auto Show

treatment. I{ard chromium plating, which has a high resistance to wear

and possesses outstanding friction characteristics, is applied on the
trochoidal surface against which the apex seals slide, and the surface is
finished by profile grinding.
Both front and rear side housings and the intermediate housing are
made of a special high rigidity cast-iron. The castings are macle Uy ttre
shell-molding Process, which has high procluctivity and produces upi-
forrn thin-wall castings. The sidewalls rrä incluction-lrrrdened in a radial
stripe pattern. By this method, a superb resistance to wear is obtained
and at the same time the clistortion causecl lry the hardening process
can be held to a minimum.
The two rotors are placed in line on the eccentric shaft with a dis-
placenlent phase 180 clcgrees apart. The rotors are made of cluctile
cast-
iron for durability and their interiors are partitioned ipto several cSam-
berl by ribs so that they can be efficiently cooled by Iubricating oil. T5e
carbon steel rotor gear that controls the trochoidal movement of
the
rotor is compactly fittecl on the rotor with six double spring.pins. These Powerplant for the Cosmo, complete with gearbox and all accessories.
 390 PPLTCATTONS THE MAZDA COSMO 39I
equipped with dual ignition. The larger engine also was given side 80 Cosmos. Of these, 60 were loaned to the firm's key suppliers and to
intake ports, two per "cylind et," the inner ports being fed by the pri- their 79 factory-owned main dealers for field evaluation, not unlike that
mary throat of a progressive carburetor and the outer ports being sup- undertaken by Chrysler with its turbine cars. It was a six-month test
plied by the two larger secondaries. program, initiated in April, 1966. At the end of its self-imposed field
With this Power unit, Toyo Kogyo produced a small series of some test program, Toyo Kogyo knew it had a reliable engine with seals
good for 80,000 miles. Its users reported good acceleration and fuel
mileage, returning 24 m.p.g. at a steady 70 m.p.h., but the car's high-
speed performance, with a maximum of only 98 m.p.h., was below Par
for a twoJiter vehicle. At this point, the decision was made to keep the
inner side intake ports for low-speed operation, and to add peripheral
intake ports to fill out the high-speed part of the curve. The change, ef-
0000p fected by late 1966, extended the power range front 6,000 to more than
-q00oo 7,000 r.p.m. Early in 1967, final clcsigrr cltangcs were utacle to inrprove
distributor drives and general accessibility.
With the production engine, the leirding plug fires 2 degrees after top
dead center and the trailing one at 7 degrees after top dead center.
Due to higher temperature in the combustion chamber (cornbustion is
always occurrirrg at one side of the rotor casing), high temperature
plugs are specified (NGK B-8EPD for normal use and B-9EPD for high-
speed runs).
With correctly metered breathing and efficient flame propagation,

;Cä
the twin-rotor engine has attained a remarkable degree of flexibility
through a wide r.p.m. range. The engine is slightly quieter than piston
engines with similar power and high r.p.m. Mostly it offers a different
sound, particularly at high speed. Instead of the clattering roar of a
Honda, for example, the Cosmo gives off a humming growl at 8,000
r.p.m. At idle, and up to 2,000 r.p.m., this Warrkel engine is distinctly
quieter than a piston engine. Acceleration, particularly in the lowcr

SPRING STRUT
WITH SHOCK ABSORBER
4r 40

Dimensional sketches for the I l0-S Cosmo. Cutaway drawing of the Mazda Cosmo.
392 APPLICATIONS THE MAZDA COSMO 797

speeds, is sensational. A four-speed manual transmission takes it

from damper units and stabilizer bar. F ront spring rate (i..., wheel rate in
0-60 rn.p.h. in 8.9 seconcls and from 25-70 rn.p.h. in 9 seconcls flat. Up- bounce) is Il2 pounds pcr inch; bunrp travcl 2J4 irrches artcl stirtic
shifting is recommended at 5,000 r.p.m. and the engine seems to grab wlreel deflection is 3.55 inches. The front end of the car has a ratio
hold of each gear as if it were waiting for it. A torque curve that is re- of sprung-to-unsprung weight of 5.3:1, which is excellent. The stabilizer
markably flat on top provides quick response to all demands. From 65 bar is strong-168 pounds per inch. Total roll stiffness is 494 foot
foot pounds at 1,000 r.p.ffi., torque climbs to 80 at 1,500 and 85 at pounds per degree; enough to keep roll angles down even at extreme
2,000. At 3,000 r.p.nl., 92 foot pounds is procluccd, 98 at 4,000 and Iateral accelerations.
99 at 5,000. At 6,000 r.p.ril., torque falls off slightly to 96 foot pounds.
The Cosmo is capablc of loafing along at I,000 r.p.m. in top geat,
without lugging, at a steacly 12 m.p.h. Wherr the accelerator is floorccl,
it movcs out quickly ancl without cornplaint. The rotors also rccluce
vibration to the point where it is only barely perceptible. This is as true
at idle, just over 800 r.p.m., as it is at 8,000 r.p.m.
The cooling systern is sealed, and has an expansion tank which col-
Iects coolant that would otherwise be lost through the overflow pipe.
This system also keeps the radiator permanently filled with liquid in-
stead of air, so that coolant foaming is prevented and cooling efficiency
is kept high at all times.
The car handles with assurance. The steering ratio is 17.3:1, giving
three quick turns of the wheel from full right turn to full left turn (lock- FitrNrtttttut\
to-lock). Very firm suspension and a low profile give the car good ^uttp12211u*l
I trrlltrrrrlerrfl

cornering characteristics. It is stable, even in sharp turns on rough

.rtr,rrrrr&,u»r}.$

;i,';ü!1;iäi
roads. Front suspension is a double wishbone arrangement, with un-
equal length fabricated upper and pressed lower A-arms, coil spring/

I,
600
'/
li

7
500 EI

{a
I,
{00
1
70m
300 6000

200 4000

r00 2000

Tractive effort and hill-climbing

{0 80 t20 t60 200 ability in all gears, for the Cosmo
+iß (km/h) I l0-s. The Cosmo cockpit.
3e4 APPLICATIONS THE MAZDA COSMO 795

The Cosmo front encl has some built-in roll understeer, amounting SPECIFICATIONS
to about 0.557o. Static canrber on the front wheels is I degree positive I\,IAZDA
I\,TAKE
ancl the roll center is located 2.5 inches above ground level. Swivel axis COSN,IO SPORT I IO-S
MODEL
inclination is 8 degrees and the caster angle 3"30'. On the rear, the Year introduced 1967
Cosmo has a de Dion axle, with semi-elliptic leaf springs and trailing Year discontinued I 969
radius arms. The de f)ion tube which connects the two rear wheel hubs Price $ 5,81 2 in the honre market
is solid, and the halfshafts incorporate ball sliding splines. The drive- Type of body Tu,o-scatcr coupe
shafts are open, and each has two universal ioints. The shafts, in other T)'pe of construction Unit-construction body
words, are relieved of all locating duties ancl have only to transmit the Driving wheels Rear
Power unit position Iiro n t
drive to the wheels. The cle Dion tube is free to move up and down
Curb weight 2,223 pounds
with the wheels, but is anchored to the chassis by radius rods which
\Veight distribution front/rear 47.9/52.r%
keep it in proper fore-ancl-aft location under all conditions of load and pcr horscpower
Porver/rvcight ratio 20.7. pouncls
thrust. 'l'he final drive unit is boltecl to the unit-construction bocly. The
Iiucl tank capacity I 5 gallorrs
de Dion axle is not clarnped directly to the larninated springs, but is Ii'ucl tank position Ilcar end of chassis
attachccl to them via pivots. The de l)ion tube runs between the frame Powcr unit \\/rrnkcl
and the springs. Raclius rods on each side help take up the driving Nunrber of rotors 2
thrust and brake torque reaction forces; the leading portions of the Chamber displacement 29.96 cubic inches (491 cc.)
Ieaf springs take the rest. Iiquivalent total displacenrent I19.7 cubic inches (1,994 cc.)
Rear roll center is about 14 inches above ground level and total roll Compression ratio 9.4: I
stiffness is 142 foot pouncls per degree. Static rvheel camber is zero, Power output r l0 (DIN)
at r.p.m. 7,000
wheel rate in iounce is I40 pouncls pcr inch, static wheel deflection is 102 f«ltlt pouucls
Torque
3.5, inchcs, and rvheel travel in iounce is 2.16 inches. Rear sprung-to- 3, 500
at r.p.m.
unsprung weight ratio is 6.9:1, even with the de Dion tube. Braking, Orrc Flitrrchi KCA -306'l triple-throat
Carburetion system
with discs up front and drums in the rear, is straight and sure. The citrb.
original drum brakes were replacecl by discs on the front wheels in Ignition system Coil ancl lxtttcry
1966, prior to the beginning of production. The rear drums have one Cooling system \\/atcr ancl oil
Ieading and one trailing shoe each. There is no power brake option. Clutch Singlc clrv plate
.'orrr-spccd all-s1'ltchrorttcsh
On May 30, 1967, the Mazda Cosmo Il0-S, with the world's first Tritnsmission I

production rnoclcl twin-rotor Wankcl engine, was placed on sale. This Gcar ratios I ) 379:l
2.077:l
car was built rnainly for the purpose of determining the market ac- 2
I .390: I
ceptance of the rotary engine; therefore it was produced only on a
3
+ 1.0: I
Iimited scale. It was followecl on Iuly 13, 1968, by the introduction of
the It-I00 Coupe on the domestic market. That date nrarks the arrival R
5
-3.389:l
of the first volume-production, Wankel-powered car. I?inal drive ratio 4.lll:l
I?ront suspension Coil springs and tclescopic shock ab-
sorbcrs w'itlt stabiliz,cr bar
Rcar suspension Dc Diort tubc arrd rrrdius arms
Rcar springs Scnri-clliptic lcaf springs and telescopic
shock absorbcrs
Stccring systcnr l(irck itncl pinion
'l'urning diameter 32.2 fect
Ovcrall steering ratio 17 .3:l
I
I

I
I
 396 APPLICATIONS
SPECIITICATIONS
IvtAKI,l
N,IODEL
MAZDA 20
COSI\IO SPORT IIO-S
Tnrns, Iock to lock 3
Br:rke system brakes front-drums rear
Disc diametcr F
Disc diarneter R
Dnrnr diameter F :r.
Drum diameter R
Lining area -52.6 square inches
Srvcpt area
Parkirrg brake
The M azda
-
'l'ires
Tire size 165 HR 14
I
R-l00
Wheelbase 86.6 inches
Irront track 49.2 inches
Rcar track 48.8 inches
Ovcrall Iength I 6 3 inches A unr,roNSrRATIoN of the high-speed
Ovcrall rvidth 62.8 inches reliability of the Wankel engine was given in Iuly of 1969, when the
Ovcrall height
|apanese firm of Toyo Kogyo entered three Mazcla R-100 coupes in a
+5.85 inches
Test Results 24-hour endurance race at the I?rancorchamps road course in Belgium.
Acceleration times The Wankel-powered cars were rnatched against the fastest produc-
0-30 m.p.h. 3.1 sec.
tion touring cars in the world and finished an impressive fifth and sixth
0-10 m.p.h. 4.) sec. overall after one car crashed during practice. The success of the Mazda
0-50 m.p.h. ,.9 sec. R-100 in this race was but a small indication of the basic soundness
0-60 m.p.h. 8.9 sec. of the standard production model.
0-70 m.p.h. I 1.0 sec. Unlike the C-lll and the Ro-80, the I\,[azda R-100 is not a car
0-80 nl.p.h. I4.0 sec. created around the \\rankel engine, but rather the installirtion of the
0-90 rn.p.h. 17.8 sec. Wankel in an existing car. The R-100 body and chassis are borrou,ed
0-100 rn.p.h. 22.0 sec.
Top speed I from the 1200 Coupe, normally powered by a 71.3) cubic inch four-
37 m.p.h.
2 6l m.p.h. cylinder sohc piston engine developing 75 horsepower at 6,000 r.p.rn.
3 9l m.p.h. and having a top speed of 93.?. m.p.h. The twin-rotor Wankel engine
4 m.p.h. fits easily inside the same package and gives far superior performance.
The Wankel is rated at ll0 horsepower at 7,000 r.p.m. and gives a top
,
Average fuel consumption It speed of I l0 m.p.h. The chassis Iayout is conventional, with a front
engine and transnrission and a one-piece open clriveshaft to the I-beam
rear axle. The only modification nrade to the 1200 to prepare it for the
additional power of the Wankel engine \r,as to install a conrpletcly ne\\'
brake systenr, including front cliscs. TIrc body structurc, suspension
rnembers, ancl drivc train conrlx»rcnts wcrc corrsiclUrcd to bc aclequatc
and wcrc not clrangccl.
The Wankel engine is an outstandingly snrooth and quiet engirrc in
tJre Mazda, having a flexibility tlrat is founcl irr very few drive train com-
 turn, there
In racing trim, the Mazda handle<l outstanclingly well. On this hardground.
is hardly any bodY roll, and the inside front tire is not lifted from the

JFt ,
l.-*i
The lr{azcla R- l 00 stays irr front of two Porsche coupes on a curve during the
endurance racc.

binations. The engine will pull smoothly from I0 rn.p.h. in fourth

gear (about 600 r.p.m.) all the way to the 7,000 r.p.m. liruit. At 7,000
r.p.m., an electrical device shuts the secondary chokes of the twin
tandem-choke carburetor, ancl immediately reduces the engine speed
to the factory-prescribed limits. The Wankel also produces a surprising
braking effcct when the throttle is closed. Closing the throttle closes
both main throats in the carburetor and the engine immediately tends
to revert to idling speed. This is not a new innovation-the same thing
is accomplishcd by the same proceclure with a piston-type engine-
but the Wankel will slow the car to such an extent that often the
brakes need not be usecl except for a complete stop. A "long" final drive
ratio gives restful gearing, which allows the car to reach 64 m.p.h. in
seconcl gear and 85 m.p.h. in third. At ,,250 r.p.m. in top gear, the
speedometer reads 100 rn.p.h. and the engine is still smooth and totally
unstrained.
In |anuary, 1970 ToYo Kogyo entered a team of Mazda R-I00 coupes in the Monte
The steering lacks precision, but is unusually light. The linkage is Carlo Rally. Here one is shown on an AlPine road.
4OO APPLICATIONS
completely free of kickback caused by ruts, holes and bumps in the roacl.
The steering gear is a low-friction recirculating ball design, having a
variable ratio that causes a certain vagueness near the straight-ahead
position. On very slight curves, it has a strangely "dead" feeling.
The chassis may be more than adequate for the 1200 piston engine
but, no matter what the manufacturer says, it can hardly cope with the
additional speed capacity and torque of the Wankel engine. Its short-
comings exist mainly in tlre rear suspension system. The handling
characteristics are not exactly poor, but they detract from the dis-
tinctive exccllence of the power unit. On smooth surfaces the car handles
very well, with slight understeer, ancl it can be taken around curves at
surprisingly high speeds. Bocly roll is slight, and the ride is pleasing
without being soft. On rough roads, however, it soon becomes clear
that the shock absorbers allow the car to float badly on wavy surfaces.
They are too weak for their task. But the really rveak point is the rear
suspension. The axle bounces at the slightest provocation, resulting in
lost traction, and clirectional corrections beconre necess ary , On old
country roacls, driver fatigue sets in earlier than in other light cars of
comparable power and speed. lhe R- 100 cockpit has a sportl, apperlrance. Instrurnentation is ample, and the
controls are conveniertt and functional.
In certairr cornering situations, a change in vehicle attitude takes
place. This probably can be tracecl to roll oversteer. The basic char-
acteristic is understeer, q,hic| can be coupterccl b1, power incluced over-
steer due to the ease with which the insiclc rear wheel will lift and spin.
The Bridgestone "supcrspeecl Raclial" tircs nracle under Goodyear li-
cense provide less adhesion thart goocl Iiuropcan raclials. Tltev squeal
even on relatively easy turns ancl brcak away carly.
The axle is carried by trvo almost straight leaf springs placcd vcry
close to the wheel hubs .so äls to give maxinlunl stability. There are no
other links or arms locatirrg the axle in relation to the body. This axle
suspension systern is called IIotchkiss clrivc, artcl is the least ex1>ensive
way known to hang a back axle on a car. Thc h1'draulic, telcscopic shock
absorbers are positioned "sea-l eg" fashion, rvith axle mountings about
halfway betrveen the clifferential and thc hubs. Tlte body mountings are

The R-100 utilizes MacPherson-

type f ront suspension, rvith an
alnrost vertical spring lcg and a
transvcrse lott'er control Arnl. The
stabilizcr bar is not shorvn in this
The Mazda R- l 00 coupe is a two-door four-seat sedan of subcompact dimensions, sectional drawing.
but surprising 1>erfornrance. Based on a production model using a four-cylinder
piston engine, the R-100 design has not taken advantage of the small size of the
Wankel engine to make gains in useable space.
 THE MAZDA R.IOO 40)
Why does the axle rvant to make all these undesirable movements?
They are mainly caused by torque reactious, which are the result of
putting drive torque into the axle. This goes for axles just as for all
ERAKE ORUM
other bodies when any force exerted in any direction is met by an equal
and opposite force. As the driveshaft turns the pinion in the axle, the
whole axle housing tends to revolve irr the direction opposite to the
AIR CLEANER pinion's rotation. The ring gear in turn drives the ring, which sets up
DISTRIBUTOR another reaction which tends to force the axle housing to revolve in
RAOIATOR the direction of the ring gear's rotation. This torque reaction to the
ring gear's rotation causes a lift in the forward encl of the pinion shaft
OVERFLOW TANK

on acceleration and bends the springs into An S-shape. This is called

spring wind-up, and usually affects the left spring more than the right.
Uneven spring wind-up twists the axle diagonally in the chassis ancl
MULTI.LEAF SPRINGS causes rear wheel steering. The torque reaction to the pinion's rotation
ORIVESHAFT
tends to lift the right wheel on acceleration rvhile pressing the left one
4.SPEED
TRANSMISSION harder to the roacl surface. This causes axle tramp, which usually nrani-
SPRING LEG
fests itself as a periodic jumping up and down of the right wheel and
leads to loss of traction and wheelspin. To regain traction, the driver
BRAKE OISC

has to let up on the accelerator. Because of this same torque reaction,

The \\/ankel cngine is practically invisible berreath the air cleaner, and partly hidden
bv thc battcry. .l'l,g igrrition clistributor sticks up on the left of lhe irgi";: Chassis the left wheel has extra traction until the moment the right rvhecl
cngincering iL sirnple and straightforward, the overall concept being siäilar to that starts to spin, and this is what rnakes the left spring more susceptible to
of thc Irord Escort.
wind-up than the right one. Wincl-up ends, of course, when wheelspin
sets in.
Inuch closer together. This has no effect whatsoever in reducing body
The front suspension is MacPherson type, rvith near-vertical spring
Iegs and a lower A-frame control arnr. This is identical in principle to
roll, Ilut helps somewhat in restricting lateral movement of tlie axle
the front suspension design of the NSU Ro-80, urinus the conrplication
in relation to the body on bad roads.
J'he problem with Hotchkiss drive is that in adclition to the normal of the front wheel drive components. The brake pedal is very hard, but
springing and load-carrying functions, the springs are callccl upon to
it is easy to nrodulate brake force orrce thc high pedal effort is over-
come. The front discs are strongly fadc-rcsistant and perfectly stable in
carry out the axle locating duties. They have to take the full äriving
operation. The front to rear balance in the brake system is remarkably
thrusf arld whatever braking Ioads the rear w]reels car) produce. Usinl
springs as susPension links also has the drawback that they tend to flc*,
well chosen, and the brake force seems to be distributecl iust right
not only in the intencled plane but also in other planes. Above all, t5e whether the car is driven with one occupaut or four, on a wet or dry
road.
springs give the axle the freedom it needs to move up and down; one
wheel uP and the other down, or both up and down together. They
The R-100 obviously was not put into production for a long-ternr
permit some fore-and-aft ntotion in the axle-which causes a rear wheel run. It must be regarcled as an interirr-r rnoclel, rcprcscnting a hybrid
steering tendency when one wheel is twisted forward and the other
solution relying partly on obsoletc chassis enginccring ancl partly on
innovative and thoroughly refirred porver unit clesign. The i\,[azda range
backward. The leaf springs cannot positively prevent sideways move-
now includes two newcr, and far rnorc aclvarrccd, Wankcl-powcrcd cars
nlelrt of the axle whetr subiected to side force, ancl they clo not elirtripate
that ltave not ),et beconrc a\/ililablc for roacl testipg. IJotS :lre clcscrihccl
axle rottttion arouncl the driveshaft axis. The shortcomings of springs
when used as control arnrs are not so obvious in low-powerecl .rri, Uut
in detail irr the ncxt chapter.
in high-powered cars the springs must be stiffenecl or supplementecl by
traction bars.
404 APPLICATIONS THE MAZDA R-IOO 405

SPECIFICATIONS SPECIFICATIONS
MAKE MAZDA MAKE MAZDA
MODEL R-100
MODEL R-100
Ycar introduced l 969 Turns, lock to lock ).7
Year discontinued Brake system brlkes front; drums rear
Price $3,045 in the home market I Disc diameter F
Type of body Trvo-door 4-passenger sedan Disc diarneter R lt.
Type of construction Unit-construction body f)runt diamcter F
Driving wheels Rear Drum diamcter R
Power urrit position Front Lining area ,:0, square inches
Curb weight 1,77 5 pounds Srvept area
Weight distribution front/rear 56.5/1).SVo Parking brake
Porver/weight ratio l6.l5 pounds per horsepower Tires Bridgestone SuPersPeed Radial
ITuel tank capacity l6.Z gallons Tire size l+5 SR 14
Iiuel tank position Rear end of chassis Wheelbase 88.98 inches
Power unit \Vankel F ront track 47.24 inches
Number of rotors z Rear track 46.85 inches
Chambcr displacement 29.96 cubic inches (491 cc.) Overall length 150.79 inches
Equivalent total displacement I19.7 cubic inches (1,994 cc.) Overall rvidth 58.27 inches
Conrpression ratio 9.4:l Overall height 5?..9, inches
Pou,er output I 10 horsepower
Test Results
at r.p.m. 7,000
Torque 100 foot pounds Acceleration times
at r.p.tn. 4,000 0-30 m.p.h. 3.6
Carburetion system One I{itachi-Stromberg 4-banel KCB 306 0-40 m.p.h. 5.7
Igrrition system Coil and battery (dual) 0-50 m.p.h. 7.8
Cooling system Water and oil 0-60 m.p.h. 10.7
CIutch Single dry plate 0-70 m.p.h. r 5.l
'l'ransnrission Iiour-speed all-synchromesh 0-80 m.P.h. 19.6
Gcar ratios I ).7)7 :l 0-90 m.p.h. 28.2
2 2.202:l 0-100 m.P.h. 44.0
3 | .435:l Top speed I 34
4 1.0:l z 6+
3 85
5
R 4.024:l I r r0
Final drive ratio ).70:l 5
Iiront suspension MacPherson spring leg and lower A- Average fuel consumPtion 20.2 rn.P.g.
frame control arm with stabilizer bar
Irront springs Coil springs and telescopic shock ab-
sorbers
Rear suspension I-bcarn axle and Hotchkiss drive
Rcar springs Serni-clliptic leaf springs and telescopic
shock absorbers
Stccring system Ball and nut; recirculating ball
Turning diameter 3 0.0 feet

Overall steering ratio (r'ariable)

 NEW MAZDA ROTARY CARS 107

ance of the RX-Z as comparecl with the R-100.

2l In announcing the Capella series, a Toyo Kogyo official saicl: "Tlte
RX-Z is an ideal car for the Seventies, combining high performance,
luxury, safety and economy. We introduce it on the market in the be-
lief that the rotary engined cars will take over the rnain current of
vehicles in 1970."
Demonstrating the performance of the Warrkel engine through par-
ticipation in racing continues, but Toyo Kog1,o has not shown any in-
clination towards building racing prototypes. They have founcl a simpler
method, by which they nrerely develop a special 200 horsepower racing
The New M azda version of the 0823 engine for installation in a ready-made racing proto-
type. The car chosen by Toyo Kogyo is British-built, carrying the little-
Rotary Cars known name of Chevron. The Chevron car is the creation of Derek
Bennett, who built himself a clirt track nriclget car at the age of 18. FIe
raced it for two years all over Englancl and won l6 events, oftcn setting
new lap records.
The first Chevron was a Ford-powered 1,172 cc. sports car. Bennett
stopped racing the Chevron in 1962 arrcl bought a Lotus Elite; then, in
Prnneps rr wAS THE TNSPTRÄTroN of
1964, he switched to a Iiormula 3 Brabham. The following year he built
the Ro-80, or perhaps it was a desire to obtain some first-hand experi-
another car in his Bolton, Lancashire workshop, a club-formula sports
cnce wit]r front whcel clrive. Whatever the reasons, Toyo Kogyo's en-
car with the Chevron name. Then orders begarr to come in, and he
gineering staff went to work combining an enlarged version of the 0820
started production. The next step was a Group 6 GT car. For 1967,
cngine with a front rvheel drive sports coupe. The result, the RX 87,
Chevron concentrated on the tu,o-liter class and used BMW engines.
was shorvn to thc public at the 1968 Tokyo Auto Show. In October of
Later that year, Bennett decicled to make a scries of Iiomtula 3 car.s,
1969, a procluction version was announced under the name R-I30. Pro-
followed by a Formula 2 nrodel. Late in 1968, he began to builcl
duction is linritccl, however, and the R-l 30 is sold only on the ]apanese
home market.
On Mny 13, 1970, the Capella Itotary series and the Capella 1600
series were introduced on the dornestic market. The Capella series is a
,

Itew line of passenger cars, available with either a sohc piston engine or
a Wankel engine developecl from the 0820. The piston-powered Capella
replaccs the Mazcla Luce, and will be marketed as the Mazda 616; the
rotary engine version is clesignated RX-z.
The RX-z fills the gap betrveen the Mazda R-100 and R-l30, and is
pou'ered by , high-performance twin-rotor Wankel engine (57) cc. x 2)
clcvclopecl cspccially for this vehicle. The R6l2 engine is basically the
same as the It-100 engine (491 cc. x 2), but its rotor width has been
incrcasecl to raise engine volume and its exhaust port shape has been
changecl to minimize exhaust noise. By modifying the exhaust port, the
inherent quietness of the rotary engine has been further improved. Like
the R-100 engine, this engine is highly flexible over its u,ide speed range
and runs practically rvithout vibration. Interrnediate-range and high-
speed pcrfonllance have been further irnproved, adding to the perform- The l\{azda RX-2 coupe.
 F

The Mazda RX-z cockpit. The lüazda-Chevron at Le N'lans, 1970.

 !i
l

il

ri
...1

ilr1

t)
it,l
li

)l
1,,

..!
,i

rl
.l
rl
;i
., l
j
ii
ii
1i
:i
:..J

i':l
tti
l-r
'i'i
i,l
i',
,.1
!":
iri

The Mazda R-l30 coupe. :t

.-{
r;!
r.i
;,,

ii:ti
i;
.,1

(
;,1
I

Mazda engine installation in the Chevron car. ti,

..
l

.,i

During 1969, the Familia sedan was made available with the 0820 engine and
marketed as the Fanrilia Rotary SS.

The Mazda-Chevron at speed (Spa, 1,000 k*.).

412 APPLICATIONS NEW MAZDA ROTARY CARS 4t3
Iiormula B cars with 1,600 cc. twin-cam engines for export to America. SPECIITICATIONS
During 1969 and 1970, the 88 Chevron proved almost invincible in
the two-liter class of international sports car racing. For the Ma zda MAKE MAZDA
engine, Bennett designed a new moclel, the B-16. The B-16 is a sports- N,IODEL RX-Z
Ycar introduced 1970
racing prototype with the engine placed midships in a tubular steel
Year discontinued
chassis. The whole car weighs only 1,298 pounds. The Chevron-Mazda
Price -Tu'o-door couPe
campaigned in Europe during the 1970 season. It finished second in its 'Iype of body
class in the l,000-kilometer endurance race at the Nürburgring in Ger- 'l'i,il. of construction Unit bocl1, u'ith front sub-frame
many in Mry, but retired in the opening stages of the Z4-hour race at Driving wheels Rcar
Le Mans due to unspecified engine trouble. Porver unit position Iiront
Strangely, the 200 horsepower Wankel engine is one of the noisiest Curb weight 2,1 16 Pounds
power units on the scene. It has proved to have an impressive speed \Veig[t distribution front/rear 5?..+/+7.60/o
capability, but it may be overstressed, ä point which is supported by its ratio
porver/rveight 17 .6 pouncls per horsepower
fragility. As an engineering research exercise, the racing program may Ii,el tank capacity 18 gallons
bring valuable lessons, but results in terms of publicity or sales pro-
t
f, Fucl tank position Rcar
I Porver unit \\/ankel
motion are elusive. I
Nurnber of rotors z
If Toyo Kogyo decides to get serious about racing, the company will II
t Charnbcr displacement 34.9 cubic inchcs (57) cc')
have to field a team. They are equipped to run every race on computers Ilcluivalcnt total displaccment 139.6 cubic inchcs (2,292 cc.)
and make every conceivable test in the laboratory before their future Compression ratio 9.4:l
racing cars even have to come face to face with competition. But at this Porver output 120 horsepower
writing, Do firm decision has been made. at r.p.m. 6,500
Torque 108 foot pounds
at r.p.m. 3, 500

Carburetion qystem Strombcrg 4-barrel

Ignition s1'stem Coil and battery
Cooling system Watcr and oil
Clutch Single dry plate, diaphragrn-tYPe
Transurission Iiour-specd
Gear ratios I 3.683: I
2 7.76):l
3 l.)97 :l
+ I .00: I
5
R
-).692:l
Ilinal drive ratio 3.70:l
Iiront suspension I\'IacPhcrson spring leg and lower control
arm
ITront springs Coil springs lvith concentric hydraulic
The RX-85 prototype was the final pre-production version of the R-100. shock absorbers
Rcar suspcnsion I-bcunt etxlc and four-link location system
Rertr springs Coil s1>rings and ht'draulic shock ab-
sorbcrs
Stecring system Ball ancl nut
Turning cliaute ter 3 I fcct

{i
I
I
 1r4 APPLICATIONS
SPECIIIICATIONS
MAKE
MODEL
MAZDA 22
RX-2
Ovcrall stecring ratio l7-19 : I variable
T'urns, Iock to lock _
ßrakc systcnt Disc Ii, clrums R
l)isc dianrcter It
Disc diamcter R
f)runr cliamctcr F
Drum dianretcr R
Lining arca
Su,cpt area The Citroän
Parking brake
Tircs
'l'ire size *l to 13
Wheelbase 97 inches
I?ront track 5l inchcs
Rcar track 50 inches IN fanuanv oF 1970, Citroön an-
Ovcrall length 163 inches nounced a program to build 500 Wankel-powered cars, with the model
Ovcrall width 62 inches
designation M-35. The M-35 is not a high-spced car. It's not a fanrily
Ovcrall height 4, inches
car either, as it has only two doors and back seat room is too crarnpcd
Test Results for two passengers on a Iong trip. But it's a grcat little econorlry car for
Accclcration times cornurercial travelcrs or ),oul'rg nrarriccl pcoplc. I t's rrot il s1l«rrts car,
0-30 nr.p.h. although it is not cntirely dcvoid of sornc sportiucss in its chanlcter.
0-40 rn.p.h. Over the past several years, Citroön, acting within the framework of
0-50 m.p.h. Societe Comotor, has studied the rotating piston engine in collaboration
0-60 m.p.h. with NSU. When an advanced prototype was finally built, Citroön de-
0-70 nr.p.h.
cided to invite the public to cooperatc with its engineering office in a
0-80 nr.p.h.
0-90 nr.p.h. Iarge-scale field test. They offered a ne\trr vehicle to errthnsiasts of the
0- I 00 rn.p.h. new technology.Tlte M-35 bocly was clcsignccl strictly as rl functional
Top speed I structure for this research vchicle. It does not indicate the shape of a
2 new production model or the possiblc evolutionary direction of other
) models currently in production; it is only a prototype and production
4 will be limited to 500 cars. The M-35 is asscrnblecl in the La ]anais
, plant, near Rennes, at a rate of two cars a dry. The cars are to be in
Averagc fucl consumption continuous operation until a certain rnileage is achieved. The engine
is guaranteed for two years and the rest of the vehicle for one year
with no mileage limit.
I
No M-35 is to be sold outside Metropolitan France. But Citroön
made sure that M-35 cars would be delivered in all provinces of France,
so as to obtain field expcrience under a great varietl, of conditions suclr
as mountain roads, motorways, and urban traffic. It is expected tltat
the lessons of this field test will bring Citroön's engineering office a
i
I

I
116 APPLICATIONS
sum of knowledge unobtainable by any in-house test method. A full
evaluation of the 500 prototypes is scheduled for 1974. Not until that
time will Citroön reach a decision about mass producing a Wankel-
powered car.
When you get into the driver's seat of the M-3 5, your reactions de-
pend on whether or not you have had previous experience with Citroön
cars. If you have, things will be as expected. If you have not, you will
be surprised. In typical Citroön fashion, the M-35 has a one-spoke
steering wheel and a gearshift lever that is actually a bent rod sticking
out of the instrument panel, with a knob at the end, reminiscent of
some of the less practical umbrella handles I have seen. Some of the
instruments are built into a little cluster attached to the steering post.
The speedometer is a horizontal design, placecl in the panel itself, but
the tachometer is mounted in an "adcled-on" position to the left of the
other instruments.
By their appearance, the seats give no hint of their real superior
quality and comfort. Not excessively spongy or soft, they still give way

Citroön M'15 is
-lt with the Ami-8, thefuture
Sharing a number of bocly a!,1 chassis p_arts
;;iil; o"fv *itr, regard io its engirrc. is not proposcd as a production
*äd.t, but strictly .J , large-scale field test experiment.

but also give support exactly where neeclcd most-including lateral suP-
port. Tü seati irru. reclining backrests. They are not hingecl at the
bottom, äs is the case with most other typcs, llut higher uP, in the
small of the back above the hip ioirrt. The scat ctln be sct to
ahnost
perfect position for any clrivcr and the clriving positiorr is colllfortable,
,ltl-,orgi, t5e almost-nat steering u,hcel positiorr looks utrttsttitl at first.
T5e driver soon gets usecl to this, ancl in fact, it gives a better griP for
a hard pull. The unassistecl steering sometimes calls for
the use of force,
locatiorl
as the combination of front u,heel drive ancl the foru'ard engine
pinion
tend to keep the car traveling in a straight lirre. The rack arrcl
steering is faultless as far as precisiorr is cotlccrltccl, bttt
gcts ilrvfully
heavy at low sPcecls.
Tlie funny gcarshift lever in the dash is harcl to rcrtclt, but it nloves
witS reasonrUf. precisiorr.'Ihc single-r«ltor ertgirtc is pcrlraps a llit slow
in starting, but icspo,r,ls instantly to tlrc throttlc. [ingine rroise_i.s we]l
muffiea ana insulatecl. The cxhaust tonc is pleilsatrt, ancl tlte nrechanical
noise from the power unit ancl clrive train is ututed by * Iayer of
souncl-
deadening matärial on the uncle rsicle of the hood. Tlterc is only a
The I\,1- 3 5 is a four seat business coupe, powered by a single-rotor Wankel engine I small amount of gear noise that filters through frorn tlte tratrsnrission.
which has been built by NSU and partly designed by Citroön. As with all other
The engine pulls strongll, irr top gear from ZS nr.p.h. (2,000 r.p.rl')
I

currcnt Citroön models, it utilizes front rvheel drive.

I
I

I
t
 i
t
I
APPLICATIONS
a

418 I
t
I

without snatch or vibration. The ignition distributor is a new type of I

SEV unit with two-stage advance to overcome snatch on the overrun

without loss of performance during acceJeration.
As was noted with all other Wankel-powered cars, the engine seems
to thrive on high rotational speeds, ancl never gives any hint that it is
spinning at or near its permittecl peak. There is an audible warning to
prevent overrevving; when engine speecl exceeds 7,500 r.p.m., a loud
buzzer goes off. In my opinion, it would be preferable to just cut the
ignition, as Mercedes does on the C-l I l. The engine runs satisfactorily
on regular-grade French gasoline. Like other Wankel engines, it may
Iose a fraction of its power potential when running on low-octane fuel,
but abnormal conrbustion phenomena do not occur.
Citroön claims a top speed iust short of 90 ffi.p.h., but the M-15
actually will go faster on superhighways when given full throttle for
Ionger periods. However, the engine does not have the power to main-
tain speed on uphill gradients. Low-range torque is poor, acceleration is
mediocre, and top-gear torque is decidedly lacking. Fortunately, the gear
ratios are well spaced, and third gear can be used to accelerate to 80 I

m.p.h. The shift is easy to operate, once you get used to it, and the I
add-on tachometer and
synchromesh is flawless. The cockpit of the M-35 reveals its z-cv ancestry, -with its table, and the cdr has
dash mounted g.rrrriiri i.r.r. The seats are extremely comfo
The technical details of the M-35 cannot be profitably discussed until
been designed *itt full regard for safetY.
you know what the basic package is. First of all, it has front wheel
drive. As is the case with the NSU Ro-80, the engine is located right up
in the nose of the chassis. This is the arrangement that gives the best factured by NSU, it carries many signs of clevelopment work carried
space utilization within a short wheelbase. out by Citroön.
The M-35 however, was not created as a design study in the manner When the Wankel engine went into the Arni-8 chassis, the transmis-
of the Mercedes-Berrz C-l I I or the NSU Ro-80, but utilizes a large sion rausmissiorr is an entirely new design
proportion of off-the-shelf parts. It has a platform frame, borrowed di- I
that the gearboxes used on the f)yane or
Ami tyPe, usirrg only two gears to obtain
I
rectly from Citroön's two economy cars, the zCV Dyane and the Ami-8.
t

Both the Dyane and the Ami-8 have two-cylinder, horizontally opposecl,
i
the possible because of the front wheel
Other
air-cooled engines, positioned ahead of the front wheel centerline, in drive, and the rotation can be reversed without complications.
Citroöns with similar chassis and piston engines do not reverse
the ro-
addition to front wheel drive. Some of the body panels are the same as
tation but pass the power flow through two gearsets. The other
g€at-
those used on the Anri-8, but the Ami-8 is a family car and the M-35
the
had to have a coupe body. An adaptation was quickly concocted by the boxes (Ami-8 and Dyane) have three separate shafts. Unforturtately,
M-3 5 shift pattern is 'ibackwards" in rclation to collventional
cars.
Citroön styling department. The same hood, fenders and doors have
been adopted unchanged. The rear quarters resenrble those usecl on the w5en upshirting from first to second, it's easy to go straight into top
Ami-8 station wagon. The front sheet metal has been extended forward gear by mistake.
to accommodate the racliator needed for the water-coolecl NSU engine, Citroön's other light vehicles use suspcnsion s1'stenrs with horizontalll'
which takes up Zy2 inches more room than the air cooling fan used oil springs to provide eclualizing susPension, but the
on the Ami-8. The M-35 body is built by Carrosserie Heuliez and u.rrior, äf the oleo-pncumatic sllsPension system de-
shipped to Citroön for assembly. Technically speaking, however, the
I
D series, the long-wheelbase "stanclard-size" cars in
IU-35 is an Ami-8 with an engine swap. Although the engine is rnanu- line. Citroön has been a pioneer of self-equalizing sus-
420 APPLICATIONS THIi CITI{OTX M-35 lZI
camc tic-rocls). Volutc sprirrgs inscrtccl llctu,ccrr tlrc ct,lindcr urrcl tlrc
pension systems, ancl the M-35 c«lmbines ccrtain elenrents of the first
tie-rod guides ensurccl the rccluisite stability. 'l'he susperrsion linkagcs
with certain elements of the newest. providecl a roll axis at grouncl level, giving the car vcr), low roll stiffness,
Citroör's, and the worlcl's first, mass-producecl car with an equalizing and even the pitch resistance was disappointingly low (rvhich is uncler-
suspension system was the 2CV. Design work began in 1936, testing and
standable u,ith a u,heelbase of only 95 inches). Neither did it provide
development went on throughout World War II (the Citroön manage-
automatic leveling to compensate for loacl variations, and it could not
ment apparently convinced the German occupation force.s that they
be considercd suitable for higlr-perfornrancc crlrs.
were working on a revolutionary type of military vehicle somewhat less
Citroön's oleo-pneumatic suspension has been used on the series since
than Ieep size), and the car was presented to the public in 1948. Front 1955, Its irnmcdiatc forcrunncr \r/AS an olco-1>neunutic, sclf-levelirrg rcilr
and rear wheels were carriecl lry bellcrank levers, Ieacling at thc front
suspension usccl on six-cylinclcr Citroöns sittcc 1953.'l'ltc l)S-19 was in-
and trailing at the rear, pivotecl in such a fashion that they moved in troduced at the Paris Auto Show in Octobcr, 1955, and caused a gcnuine
vertical planes. A single spring assembly on each side of the car was
sensation, although its existence had been rumored for a Iong time. The
connected to both front and rear suspension levers. The spring unit
car had a curb weight of 2,775 pounds, and 66% of the weight was
consisted of two coil springs enclosed in a common cylirrder, one fac-
carried on the front wheels. The wheelbase was extremely long, lZ3
ing forward, the dther backward.
inches, while front track was 59.7.5 inches, considerably wider than the
For the initial production run, the springs were loaded in cornpres-
5l .5 inch rear track. The front wheels were also larger than the rear
sion, but in 1955 the Iayout was modifiecl so that the main springs
ones on the initial procluction run, llut the rcar whecls were latcr
worked in tension (the pushrods from the suspension arms simply be-
brought up to front whecl size. All the u,ltecls h:tcl independent sus-
pcnsion. 'l'hc fr«lnt rvhccls wcrc locatcd lry tq)pcr itttcl lolcr trtttst,crsc
The Citroön M-35.
control arms and the rear u,hecls by trailing arnrs. T'lte frottt arttts wcrc
not triangular but elbow-shaped, tapering from the chassis attachment
points to the spindle holes for the ball joints in the wheel hubs. A
splined extension of the upper arm was coupled to a rocker arm which
transmitted suspension rnovements to the air-and-oil spring unit. A
similar rocker arrn system was coupled to the rear trailing arms.
Each rocker arm was connected to a spherical container in which a
flexible rubber diaphragm separated the gas elenreut of the suspension
from the oil-basecl elenrent. The wciglit of the car and the suspension
movements were sustainecl by rritrogen gas uncler pressure in the cylinder
on which the spherical container was mounted. The effective length
of the suspension leg was dictated by the arnount of oil in the cylinder.
The oil volume also controlled the height of ttre car above the ground
and its automatic leveling, ancl the oil also worked as a shock absorber
for all suspension movements. The danrping action was obtained by
fitting a two-\ ,ay restrictor valve betwcen the cylinder ancl the spherical
container. Lamination of the fluid reduced the florv rate, rvith thc re-
sult that nlovements u'ere continuous ancl progressit'e, artcl always pro-
portiorral to thc s1>cccl of flui<l clisplAcenrcrtt. Whcn pilsscngcrs erttercd
the car, the gas in tlre cylinclcrs \1,4s courprcsscd ancl thc citr sank closer
to the ground. A central h1'clraulic brain then restored the car to its
ät
normal height by pumpirrg nrore oil to the cylinders in which the gas
 4ZZ APPLICATIONS THE CITROBN M-35 423
was comPressed. Conversely, Passengers getting out caused the
car to near the center of the car, filling the same space normally occupied by
rise somewhat before the Ieveling valves went into action and
releasecl the coil spring units. They are attached to the srlnre levers and control
oil (which then returned to the central reservoir).
The Citroön system is noted for its extremäly low ride-frequencies arms that are used on the zCV Dyane and Ami-8. In acldition to pro-
(i'e., the motion frequency of the sprung mass in free osciilation). viding automatic constant level control, this system gives the I\,I-35 ad-
Modern passenger cars, touring cars, ,na spärts cars have frequencies iustable ground clearance, just as on the D series. Iior superhighway
in driving, the car is Iet dorvn to the lowest suspension lcvel. This has the
the range of 70-80 cycles per minute. Tüe Citroön design ällo*,
fre- effect of reducing suspension travel, ancl any sudden burnps then will
quencies as low as 40'cyclei per minute, and, as adirect
result of such be felt. Normal clriving is done with the suspension system at medium
frequencies, it offers a unique degree of ride comfort.
The ricle rate is height. Iior rough roads, the high position is used, which, inciclentally,
autonratically variable according to load and deflection-a near-perfect
combination. However, the DS-I9 suspension geometry offers also gives the softest ride.
1.,o roll Aside from the engine, the greatest advantage the N{-35 has over the
stiffness at all, and heavy stabilizer bars are fitt;d both
front and rear. Dyane and Ami-8 is the suspension system. Thc irritating pitching nro-
This makes the car sensitive to changes in roacl camber as a penalty
for tions in the small cars has been practically eliminated in the I\{-35. Thc
its remarkrbly upright cornering slyle. The Citroön oleopneumatic
suspension has proved its practical value in international ride is luxurious, beautifully cu.shioncd, far in irclvancc of the avcragc
rallies over a for cars of its size. It is very stable, and rvhile roll is restricted by sta-
l5year period, and now has been adaptecl to the M-35.
On the M-35, the oleo-pneumatic spiing units are placed horizontally bilizer bars, it has not been eliminated. The car leans sidervays on
curves, iust'like the coil-spring moclels using the same basic chassis. The
M-35 has a firm understeering characteristic at nornlal road speeds. It
LTVEL CONTROT LEVCß
wants to go straight, and resists turning. Self-centering action is very
strong. The car is stable and highly predictable, with a generous margin
SUSPENSION UNIIS
of controllable front end slip before the wheels go into a slide. The in-
herent stability of the car is never upset by poor road snrfaces, potholes,
ripples or washboard pavement. It keeps right on its intended course, and
road shocks are fully absorbecl in the suspension system. BecAuse of the
Pn-€ssun€ AccuMullroß
shorter wheelbase, the ride is not cluite up to D series standards, but
it's close.
The M-35 is very safe on all kinds of roads, and takes L-berrcls, hair-
pins and S-bends with equal aplomb. When front end slip (understeer)
gets critical, merely backing off on the throttle brings the car vcry nicely
LEVEL COR߀CIOß
back to a normal control situation. The rear end never breaks loose, the
rear wheels merely track behincl the front oncs rvherever they go. The
'ROT{T

ßEAR IEVET CORRECIOF car is not only well balanced, but also unusually obedient. Tires are
ZX 135-15, using a new finer-mesh steel belt than the regular X tire, in
order to reduce any Iow-speed harshness. These tires have excellent side
bite, and run with moderate slip angles at very high lateral accelerations.
Steering is somewhat heavy for such a light car, but its tuming circle
,
ttf t ;ßoxT wxtEL
---- ./5 is tight, and parking and rnaneuvering are quick and easy. The steering
coxraof Iaiü HlGx pßEssunr uNCs
IEFT REAß WHEET
COI{IßOL ARM
also seems slow, but it is necessary to provide more leverage than on
thc Ami-8 in vicw of the hcal'icr front cncl. Thc Citroön cnginccrs clc-
Olco-pneunlatic suspension systern for the M- 3 5 has four suspension
located,_ two on each side, with hydraulic con nections
units, centrally cided that a slower steering ratio woulcl be m()rc acccl>tablc tlran thc
to the wheels' control arms.
Each wheel is located by a single cänirof arm (leading in front
and trailing in the
extra cost of porver steering. Contrary to general clairns for the \Vankel
rear). engine, there is no weight advantage in this case.
 125
TTIE CITI\ONN M.35
474 APPLICATIONS
SPECIFICATIONS
Power brakes are starrdarcl with clisc brakes front ancl drunr brakes
rear. The handbrake u,orks on the front wheels, and the front disc brakes IVIAKE CITROiiN
are mounted inboarcl for better cooling. The brakes are fed cool air via MODEL M-35
clucts, therefore they do not overheat and remain stable in operation Carburetion sYstem Onc two-barrel Solcx L8/32 I{I-ID
evcn after hard use. Nose divc under braking is less pronounced than on Ignition system Coil and batterY
tlre Anri-8. Wind noise at high speeds is bothersome, but the mechanical Cooling system Water
noisc leve] is very low, rvhich terrds to accentuate u,ind noise. Road noise, CIutch Single drY Plate
too, seenrs higher than irr the other small Citroöns, despite the hydraulic Transmission Iiour-spccd manual
suspension system. Probably the fault lies not with the M-35 but with Gear ratios I 3.81l:l
2.)17:l
the f)1'anc and Anri-8. With thcir noisy air-cooled piston engines, a lot 2
1.454:I
7
of the roacl noise and chassis rumble is drowned out. I .077: I
4
The I\,1-35 is experimental in character, even th«lugh the expcriment 5
is large-scale. Citroön has made it clear that the M-3 5 will not under any R 4.1 82: l
circumstanccs bcconre a production car. What, tlten, is its purpose? It Irinal drive ratio I
-1.12 5:
is a test for the economy car formula of the future, combining aclvancecl f,'rortt suspension Singlc lcaclirtg :.lrnl on each side with
conccpts such as a Wankel engine, hydraulic suspension, ancl front strrbilizer bar
wheel clrive irr onc vchicle. The fact that Citroörr is rnaking 500 of tltese I?rorrt springs Olco-pucttttrrrtic struts with automatic
leveling
cars is eloquent testirnony of the company's belief in this formula.
Rcar suspension Single trailing arms on each side with
stabilizer bar
SPECIFICATIONS Rear springs OIeo-pneumatic struts with autonratic
leveling
I\,IAKIi CITROiiN Ilack and pinion
Steering system
MODEL M-35 37 .) feet
Turning diameter
Year introduced t970
Ovcrall steering ratio l8:l
Ycar cliscontinued 7..25
Tunts, lock to lock
Price $2,740 in the home market Brake system Porver-assisted; discs front; drums rear
Typc of body Trvo-door four-passenger coupe 10.6 5 inches
Disc diameter F
Typc of construction Platfornr framc and steel body
Disc diameter R
Driving whcels ITron t
Druru cl iarneter IT
I'ower unit position Front .12 inchcs
Drum dianreter R 7
Curb wcight 1,793 pounds +6.0 5 square inches
Lining area
Wcight distribution front/rear
Swept area
Power/weight ratio )2.6 pounds per horsepower Ivlechanical, on rear wheels
Parking brake
Iiuel tank capacity ll32 gallons
Tires Michelin X radial PIY
Iiucl tank position Rear
Tire size t 3 5-15 ZX
Porver unit Wankel 94.5 inches
Wheelbase
Nurnber of rotors I +9.6 irrchcs
Iiront track
Chanrbcr displaccment 30.03 cubic inches (497.5 cc.) 48.0 3 inchcs
Rear track
Iiquivalcnt total displacerncnt 60.06 cubic inches (995 cc.) 159 irtclrt:s
Ovcntll lcrrgth
Conrprcssion ratio 9:l 61.25 ittc:ltcs
Ovcr:tll rviclth
Porvcr output 55 horscpowcr fi.2.5 inchcs
Ovcrall hcight
at r.p.trt. 5,500
Torclue 50.6 foot pounds
at r.p.m. 2,7 45
126 APPLICATIONS
SPECIFICATIONS
MAKE
MODEL
CITROiiN
M_35
23
Test Results
Accelcration times
0-30 m.p.h. 4.8 sec.
0-40 m.p.h. 8.2 sec.
0-50 nl.p.ll. 12.0 sec.
0-60 nl.p.h. 18.0 sec.
0-70 nl .p.ll. 25., sec. The
0-80 m.p.h. +0.0
0-90 nr.p.h.
scc.
Mercedes-Be nz
0-l 00 ur.p.h. I
Top spccd I 27.0 m.p.h.
2 44.0 nl.p.h.
) 68.5 m.p.h.
4 89.5 m.p.h.
5
Larr rN 1969, Mercedes-Benz an-
Average fucl consumption u m.P.g.
nounced the existence of a mid-engined prototype sports car powered
by the revolutionary Wankel rotary engine. This prototype, Iabeled the
C-lll, was designed to develop the Wankel engine, examine new sus-
pension systems, explore the use of plastic for body structures and to
pursue the study of high-speed aerodynamics. In accordance with these
goals, the C-l I I project is probably the most radical clesign that Mer-
cedes-B enz has ever undcrtakeu.
The engine in thc C-lll sits belrirrd the clrivcr but cloes rrot hang out
over the rear whcel ccrrtcrliue, AS iu a Volkswagcn.'IItc ertgine is car-
ried entircly r+,ithin thc rvlrecll)asc, rvhiclr has advantages ovcr a con-
ventional front cngine /rear clrive systcrn bccause of bettcr wcight
distribution, Iess weight, Ioler cost because of the fewer number of
drive train components neeclccl, Iess heat transfer to the passenger
compartrnent and a lower noise levcl in thc passcnger compartment.
The Wankel engine is extremcly rvell suited for nrid-engine installa-
tion in a sports car bccausc it is so light ancl conU)act. The C-l l l engine
clelivers 330 horsepowcr and wcighs orrly 308 pounds. In conrparison,
a 3)5 horsepowcr Iiorcl V8 u,cighs 630 pouncls ancl thc nc\\' I\'[crcccles-
Benz 230 horsepowcr V8 weighs 495 pouncls. T'he C-l l t threc-rotor
Wankel engine is only ?4.1 inches Iong, 35.8 inches high ancl 3).) inches
n,ide, while the2)0 lrorscpo\\,cr \r8 is 33.4 inches long, ?6., inches wicle,
art«l 27.15 irrchcs high.
The C-l l I has slight rernvurd bias in u,cight distributiorr ancl An
^
extremely low centcr of gravity. Static front/rclr u,eight clistribution is
I
45/55%. But, the Ieast praised of its design features is the low polar

I
 I

428 APPLICATIONS I

moment of inertia. The polar moment of inertia is an indication of the

car's resistance to a change in direction of travel, and is determined not
only by the front/rear weight distribution but also by the exact disposi-
tion of the various weight concentrations in the chassis (such as engine,
transmission, fuel tank, radiator, battery, spare wheel and occupants).
Flaving them all within the wheelbase gives a low polar moment of
inertia, which makes it easier to steer the car from one path to another.
I-Iaving the engirre in the front of the car with a bie fuel tank and a

;e in Stuttgart, the C-l I I shines in the

,i
parked in the middle of Mercedes-Stras
n
-a

*";*g sunlilhi. M.rcecles-strasse is the main approach road to the Daimler-Benz ,:l

headquarters. :l
ii-l
,,i

heavy spare wheel in the rear gives a high polar moment of inertia and
the car resists any change in directioll.
Because .orn.iing is iuch an important part of racing, it is under-
standable why the nrid-engine racing car has becontc universal irr the
past decade. i)ai,llcr-Rcnz has bccn awarc «l[ this fact for t vcry lorrg
time. Benz built a mid-engine racing car back in 192?., and there was a
Mercedes-Be nz 150 FI mid-engine roadster in 1934. But the C-l I t is the
compaly's first atternpt in moclern tirnes to irrvcstigate the concept of
t5c inicl-cpginc irrstailirtion as ap1>liccl t«l n lriglr-1:crftlrtllrltlcc Crlrtcl
Touring car.
Irr the early spring of 1970, Mercedes-Benz displayed a new version of the C-lll, The architectural pattern for the moclern rnicl-cngine sports car was
the Mark II. It is shown in thc foreground, in front of the rnodel driven by the author Iaicl clown by Porschc as carl), ärs 1954,Iiarlicr cilrs fr«lttr l'«rrschc carricd
in Scptember 1969, with the original prototype in the upper left-hand corner. The
Mark II has a four-rotor Wankel engine. the engine in the tail, giving a high polar momettt of inertirt and irrtro-
 THE MERCEDES-BENZ C-III 131
their iobs. The C-I11 has a unique suspension system. Whcels must be
free to travel up and dorvn when they meet une\/enness in the roadway.
Some horizontal compliance in such situations is also desirablc. Whilc
they travel up or clown, the whcels should have a nrinimunr of frecclorn
to move in other planes, such as in'camber and toe-in. The extent to
which such changes are allowed to occur is dictated by the suspension
linkages-the same control arms that also determine the roll center
height. In the C-lll, the front roll center is Iocated about 2 inches above
ground Ievel and the rear roll center about 5 inches above ground level.
Coupled with a center of gravity positioned about l5 irrches above
ground level, this results in a naturally Iow tenclency to roll. Roll, in this
context, means lean or sway.
The Palatal body was built at Wagonfabrik Rastatt and shipped to Mercedes-B enz,
where the complete vehicle was assembled . (Photo: 'Weitmann)

T'he redesigned rear end of the C-l I I Mark II offers considerably improved rear
visibility in addition to increased luggage capacity. This test car is fully equipped
for road use.

ducing an unpredictable elernent in the handling characteristics by the

extreme rearwarcl bias in weight distribution. For the Carrera model
(Typ.7l8), Porsche turned the entire power train around, placing the
engine in the micldle. This school of design quickly dominatecl the
smaller categories of sports and racing cars. It was popularized by Cooper
in ITormula I Grancl Prix racing in the late Itifties, and is now univcrsal
in that field, and has been regarded as a prerequisite for Le Mans proto-
types of all sizes since the early Sixties. Beginning with the I)e Tomaso
Vallelunga in 1963, a fcw urid-engine production cars have appeared.
The most widely publicized have been the 1966 Lamborghini Miura,
1967 Lotus Europa, 1967 Dino (by Ferrari), 1968 Matra 5)0, 1970 Dc
Tomaso Pantcra, 1970 lMonteverdi FIai and the 1970 Porsche 914.
Good weight distribution, a low center of gravity, and suitably low
polar nroment of incrtia do not guarantee good roadholding and great
handling precision-the suspension ancl steerirrg systcrns'still must do
 THE MERCEDES.BENZ C.II I 473

The front suspension has many novel features. The lower control arms
are attached to the wheel at hub level, not below as is common on pro-
duction cars. The lower control arms are not of an A-frarne construc-
tion, but each is an I-bar located in the fore-and-aft plane by a senri-
leading diagonal drag strut. The drag strut provides the necessary tri-
angulation required for handling the weight transfer loads that occur
during braking and hard cornering, while giving some measure of hori-
zontal compliance. The upper control arms are constructed on the same :

principle, although the I-bars are shorter and the drag struts are almost
I

parallel with the centerline of the car. The decision was made to give
the car true center-point steering. That means letting the steering axis
coincide with the center of the tire footprint. This has the effect of
reducing self-centering action (or "returnability," as it is often called),
but it does provide plenty of self-aligning torque, which is a measure of
cornering force. The lower control arms are attached to the wheels
at hub level, because of space considerations. T'he clesirccl gconretry j
could not be obtained with a mounting below the enormous brake disc,

o I
I

a i
,.1
r.i
I

i{
t.t
.1

!.i
'i ,1

'{,,i
,{
.il

il
,1

I
i{
i,l
,.-4

;i
't
;i,'.{
il

:l
:lri
,.
.,t
:l

il
il
II
li
ri
l1
IJ
l
,i
,I

I
i
li
t,
Thc rear suspension on the C- l I I has tal i:
Ir
corrcentric shock absorbers. Two steel bars
li
arc uscd as radius rods. T'he tubes in front
transnrission can be seen behind the brake [;
ti
t;
The front suspension on the C- I I I uses a tall coil spring leg with a built-in shock
t;
il
absorber. The disc brake is ventilated for maximum cooling. (Photo: Weitmann) lj
ll
l,
lr
tt
I

i:
I
 434 APPLICATIONS THE MERCEDES.BENZ C.III ß,
but it seemcd quite feasible to place the mounting in the center of the is pointing ancl the line it actually will follorv. 'l'his slip anglc is usc-
disc and acljust the Lrpper control arm accordingly. To understand this, ful, for it creates side bite or cornering force. The cornering force
you have to consicler the slip angle of the tires. opposes the side force that producecl the slip arrglc to begin with. It
J'he slip angle is thc angle formccl bctwecn tlre linc whcre thc tire cnablcs the crlr to follow :l curvccl 1>a tlr. As thc crrr crr tcrs a crrrvc, tlrc
tire is deformed ancl the footprint is distortccl. At thc front, thc rubber
has just been placed in contact with the roacl arrcl is taking littlc or no
Ioad. At the rear, there is maxinrum sicle strain and force, rvhile the
vertical loads are ccntered on tlrc f«lotprirrt's gconrctrical ccrrtcr. T'hc
distance between the ccnter of thc vertical loacls and the ccnter of thc
side forces is called the "pneumatic trail." The amount of cornering force

The four-rotor engine is slightly longer than the three-rotor unit, and remains the
same height and width. Engine accessibility for service and maintenance is very good.

il

,l
ij
il
:i
t,
t:
!i
t

,ü

)x

,ii
:
,,i

I
ir
'\
,:l

i
/t

:i
rt
t,t.

;i
i
,!'
,i:

ic
'(
r.i

The three'rotor C-l I I engine fits snugly into the space between the passenger
con)partnrent and the rear wheels. Behind it, there is ioom for a small trunk. T-he
protctty'pc enginc makes use of modi6ed, standard Mercedes-Benz components
§'hcro,er possiblc, in order to minirnize clevelopnrent time.
436 APPLICATIONS
generated in the tire, multiplied by the pneumatic trail, gives the
amount of self-aligning torque. This assists caster action ancl tends to
keep the wheels on a straight course, and it also gives the driver an indi-
cation of the force neecled to steer the car.
Thc simplest way for the Daimler-Benz engineers to build return-
ability and clirectional stability into this suspension system was to
incorporate arr cxtreme caster arrgle in the steering system. Caster is set
at 9 degrees, which makes the steering harder, but straightens the car
after a curve ancl keeps it pointed straight,
The rear suspension design looks as if it came directly from a Grand
Prix I?ormula I chassis. It has the same massive hub carriers, with hefty
anchorage points for control arms. Fore-and-aft location is determined by
four radius rocls, tlo frorn each hub carrier, upper and lower. Tlre lolver
ones arc slightly longer ancl point torvards the centerline between the
two front wheels. The upper radius rods run almost straight forward.
Lateral location of the hub carriers is taken care of by three control rods
on each sicle, the two longer ones being attached at hub level in front
II rear suspension show the details of its
and in back of the hub itself. The single short control rod is held by the ber cannot occur.
top of the hub carrier. Rear stabilizer bars add roll stiffness with very of the upper and
center of gravity,
Iittle effect on spring rates during single-wheel deflections. Adding roll how the absence
stiffness also adcls to Iateral weight transfer. For the road version, moder- ed system of rear
ate weight transfer in the rear end was desired, to keep the rear wheels whecl location.
trailing the front at all times. For the track version, high weight transfer SPRING
UPPER CONTROL ARM

was iudged clesirable, so as to enable the skilled clriver to provoke larger R CONTßOI
slip angles in the rear tires and thcreby assist the vehicle around the SIAEILIZER BAR ^RMS

curve. This should be thought of as adding oversteer rather than as COMPßESSION STRUT

rcclucing unclerstecr, because the basic understcer in the front errcl was SPßING LEG

Ieft intact and oversteer was added to the rear.

The present research program on plastic body construction at Daimler-
Benz is clirected towards the use of fiberglass-reinforced plastic as the
material for a non-stressed body structure to be mounted on a metal
frame. Other firms in Germany are experimenting with monocoque
stressecl-skin structures made of plastic materials, but the C-lll has a
sheet steel frame, partly riveted and partly welded. The fiberglass-rein-
forced plastic body structure is bonded and riveted to the chassis frame. HUB CAßRiER

Daimler-Benz found plastic an attractive enough material to include

it in the specification for the C-I I l. There are overwhelming reasons why
any automobile manufacturer should want practical experience with it. PPER CONTßOL AßM

Iriberglass-reinforced plastic has several important advantages over steel LOWER CONTROL ARM

for car boclies. TIre ma jor one is lower weight. F or a sports car, a weight
sreeninc KNUcxLE
saving of 40% coulcl be realized. Fiberglass-reinforced plastic will not
rust or corrode, which means longer body life in many parts of the world. C-1I Mark II chassis has minor susl>eusion mo«lifications from the earlier model.
'l.his sketch shows the overall configuration.
 THE MERCEDES.BENZ C.III 439
sion of fillcrglass arrcl rcsirr irrto rnolrlcrl p:trts irrvolvc.s ir lriglrcr lnlror
content than i.s rrccclccl for rrrctrrl startrllir)g.'l''llirdlt,, finishccl lltlclv ntolcl-
ings havc to bc storccl for «lryirrg, u,hich n)cans Iargc storagc spaccs, a Iot
of tinrc, ancl tlrc risk of distrlrtions cluc to irnprr)l)cr storlgc.
From the desire to learn more abclut fibcrglerss-reinforcccl plastic,
Daimler-Ben z gained the extra aclvantage of being able to nrake frequent
body changcs on short rrotice. But Dairnler-Bcnz did not actually make
the C-lll body in their own plunts, although it was desigrrecl b),them.
They farr:recl out t]rc actuul construction to a srnall finu tlrat lras bccrr
active in plastic body production for several !'ears-the Wagerrfabrik
Rastatt.

'Irunk in the C'lllN{ark II is small compared with a typical Detroit sedan, but
quite roomy for a sports car of its type. Enlargement was made possible by rerouting
thc exhaust pipes and muffier.

Dcnts or clnmagc fronr srnall blows tcncl to llc local izccl, and therefore
c:tsicr to repitir. Fibcrglass-reinforced plastic bodies do not respond to
vibration the way sheet stecl does, and the risk of stress failures is con-
.sequently reclttced. It has acoustica] propcrties that muffie road noise,
ancl plastic clairns to offer extra safety. There is some merit in this. Pound
for pouncl, it is strongcr tltan nretals used for car bodies. Parts are made
thrce tirnes as tltick as steel to clbtain comparable structural stiffness,
ancl this extra thickness gives the fiberglass-reinforced plastic twice the
tensile strength for about half the weight of similar steel parts. Uncler
severe impact, tlte plastic body absorbs a Iarger portion of the impact
force and gives addecl occupant protection. The body structure itself
is frce of squeaks and rattles. It offers better heat insulation and gives
improvecl noise insulation for an equal amount of sound-deadening
nraterial. Thcse advantagcs have been arnply demonstrated since Cfievro-
lct began production of the Corvette back in 195). But a largc-scale
rcplaccmen t of steel lry fiberglass-reinforced plastic has not occurred.
This is clttc mairrly to economic factors, and partly to a lack of public The Mark II has an entirely revised interior. Instruments are regrouped and air
clenrancl. vents relocated. The accelerator retains the auxiliarl,tab used on the original venion.
This tab is placed so that it is natural for the right heel to touch it under braking,
Irirst of all, textile glass fibers are more expensive than steel. This is to enable the driver to increase engine speed when downshifting. The tab works on
not becarlse of the nraterial itself, but because of the huge investment a scparate throttle linkage, with a cliffercnt, slower ratio than the nrain accelerator,
reqtlirecl for .special equipntent to procluce them. Secondly, the conver- in order to prevent over-revving of the Wankel engine under zero-load conditions,
such as when the clutch is disengaged.
440 APPLICATIONS
The C-l I I is not the first experimental Mercecles-Benz with a plastic UPPER CONTROL ARMS
body. The first experinrents with fiberglass-reinforced plastic bodies were
maclc in 1950 and 1951 on prototype chassis for the 300 SL sports car.
At this time, no conclusions can be drawn about possible applications
of this rnaterial to futurc production cars-it's just a research proiect.
The bocly styling for the C-l I I stems from a shape that Karl Wilfert,
chief of bocly developnrent at Daimler-Benz, designed in 1961. At that COMPßESSION STRUT
tirne he rvas toying with a proiect for a Le l\{ans prototype in the hope
that the managenrent would allow the engineering department to form
a new official factory racing team, as they had done so successfully in BRAKE DISC

1954 and 1955. But, in November 1968, the engineering department got
the orclers to go aheacl with another project-the C-I01. C-l0l was the
original proiect clesignation for the C-l I l. The project number was
changecl to C-l t I when it was announced to the public to avoicl any
con flict with existing or future Peugeot car model designations. AII COMPRESSION STRUT
Peugeot numbers have one thing in common-the middle zero. AIso,
tlre clircctors of Daimler-B enz felt that three ones in a row symbolized The front suspension s1'stem for the C-lll includes a thick stabilizer bar. The coil
springs- (n-ol shown) .are _rnountcd over the shock absorbers. Strong anti-dive effect
the thrce-rotor Wankel engine much better than two ones separated by under braking is designed into the upper control arms. The lower -control anns are
a zero. located at hub level.
Aeroclynamic drag is an obstacle to economical high-speed driving, and
one of the clesign objectives was a wind-cheating body shape. Wilfert RADIUS RODS
and his men rvent to work on a GT coupe bocly based on his 196l
clesigns for a Le Mans prototl,pe. (Jnfortunately, the cars with the lowest
air clrag are usually the ones with the least resistance to crosswinds. SHOCK ABSOREER

Iror the C-lll, Wilfert had to find a shape that would not disturb the air
very much rvhen traveling through it, yet would remain relatively insen- UPPER CONTROL ARM

sitive to sidewinds and changes in wind direction. These goals are

incompatible because sidewinds act through a car's center of pressure HUB CARRIER
in the same way that centrifugal force acts through a car's center of
gravity. It makes no difference whether the pressure angle is straight
from the sicle or just a little off the car's direction of travel-the center
of pressure clepends on shape alone. The lower the clrag coefficient of a
LOWER CONTROL ANMS
car, the farther forward its center of pressure. On most modern cars, the
center of pressure is located in the area betrveen the windshielcl corner
posts. With more clrastic streamlining, the center of pressure moves
further forrvard. On speed-record cars with very smooth airflow, it can
actually be ahead of the car itself. Sidewinds pushing on a center of
pressure that is well ahead of the car's center of gravity rvill produce
),rrvin g, and tlte car will veer off course in the direction of the sideu,ind. BRAKE DISC

I?or many )'ears, aerodynamicists and engineers have been exploring

The C'l I I rear suspension is based on modern Granct Prix racing car practice, with
various ways of pulling the center of aerodynamic pressure further back,
long radius rods and transverse corttrol arnls. This system permils minimal changes
tow'arcl the car's center of gravity. Iior this purpose they have addcd in camber angles, track ancl w'heclbase cltrring wheel clcflcCtiorrs. It is a very light-
weight construction.
412 APPLICATIONS THII MERCEDIIS-BENZ C-III 41)

tailfirrs urrcl "spoilcrs." spoilcrs arc airfoils of various sizcs ancl slta;>cs, oI corlr.sc, trrrtil tlrrlrorr64lr lcsts lrrrrl llct'u ('otttlttc'lc«I.'l'll('tt'sts w(lr(:
in various positions orr the cilr, sorlre lorv clorvtt in frortt, otlters
n-ror,rr)tccl rttltclc irr I)lrirnlc:r Ilcuz's ()\\'u tvirrtl trttrttt'ls, s()lttt: tt'itlt s(':tl('ttttltlcls lttttl
«rtlrcrs n,itlr ftrll 'l'lrc witrrl ttttttrcls l)l'ovitlctl tvilrtl ltrlttt irll
higlr up irr tlrc tail.'l'hc irritial purposc of s1>oilcrs wils to rc«lucc acro- si 7,c urock-nl)s.
clvuanric Iift forccs, but it was soon founcl that thcy also aclclccl drag, <ltrlrrtcrs, irrrcl tlrc courl)lllry cvcrr riggcrl ttl) rl sllct'irtl lnrtclrilrc l'tlr ct'clrtirrg
rvhich irr turn improvecl clircctional .stability. Tailfins present a greater sud«lcrr siclcn,irrds. It «lclivcrcd gusts u1> to 50 rrt.p.h. right rtt tlrc rtxttlsiclc.
w,inclagc arca bchirrcl the ccntcr of gravity, which tcncls to n)ovc the At thc cncl of fir,c rrrontlrs «rf irrtcrtsivc \t,ork, ll lltltll'slutllc tlf tlrc
ccrltcr of prcssure to t]rc rear of tllc car. DAirrtlcr-llcuz wrttttccl to achicvc dcsirccl strrbility, rvitlr l«»r, lift c'lulr':rctcristic.s rtttrl rt rlt'rtg c'oc[Tit'icrtt tlI ottlt'
0.33, Iracl bccn crcrltcd. 'l 'hc clr:rg cocffic:icnt irrrlicutcs tlrc ilct'(ltl1'ttittttic
its goals without recourse to fins ancl spoilcrs.
'l'hcy startecl off rvith a thorough examination of the car's airflow efEcicncy of a bocly. A lroclv u,itlr Iorv clnrg cocfficicrrt clisttrrbs tlrc air vcry
ncccls-air for engine brcathing, cooling air for racliator arrd brakes, ancl littlc; il bocly u,itlr u higlr clrilg cocfHcicrrt crcrttcs:t sullstutrtirtl ttrrlrtrlcrt('c.
vcntilation air for tlrc occr.lpants. By choosing thc right sizc, shapc ancl 'l'lrc rlrlg cocflficicrrt rrrrnrllcr \\'rls ittvcn tt:cl ottlv f tlr ('otl tlwring lltltlt'
location for tlre scoops or slots to be used as air intakes, corrsidcral>le shapcs-bocl1' sizc lras no llcarirrg on it. A clrag ctlcfl'icicrtt of 1.00 is
gains werc macle irr both clrag and stability. The choice was not made, assignccl to il partic:ulrtr sltrtpc-n ct'littclcr scvcll tirttcs:ls high tls it is
rviclc. l)rag cocfficicrrt.s arc clctcnnirrccl l>r,u,irtcl turtucl tcstirrg. A s(ltlrlrc

1Il

rL Y.

The lrastily cobbled-up C- I I I test car began road testing in May, 1969. In this
photo thc air intake was positioned betrveen the faired-in headlights, and the outlet
\\,as 1:lacccl at thc windshielcl base. Air flow for racliator cooling was considerably Both full-scale and a one-fifth scalc rnodel of the C- I I I were constructed for
rnodificcl after wincl tunncl tests, taking in air below the stagnation line, and letting
^
rvind tunnel test purposcs.
.l-his
u'as the stage of progrcss as late as February, 1969.
it out in a low-pressure area located in the middle of the short hood. (Photo: \Y eitmann)
444 APPLICATIONS THE MERCEDES.BENZ C.III +45

plyu,ood boarcl has a clrag coefficient of l.ll,while modern U.S. passen- and country driving. Therefore, the Mark II bocl1, has an enlarged and
gcr curs ha\/c clrag cocfEcicnts ranging fronr 0,42 to 0.50. rc.shApccl windshicld, urrcl biggcr sicle rrrrl rerlr u,indou,s. Rcnr visicln in
T'he main cause of lift is thc high velocity airflow ovcr thc roof of a particular has becn inrprovccl with a rcstylccl engirte covcr.'l'ltc rooflinc
car-it crcä'rtes a negative pressrlre area and aggravatcs thc effects of crclss- was completely changed. Iustcacl of thc rear quarter-panels, with louvres,
rvinds. A lrigh pressure arca under the car adds to the total lift force. A the roofline is simply marked by a rib on each sicle, rvhich improved rear
t1'pical Arnerican seclan has a lift force of some 300 pounds at 100 nr.p.h. visibility considerably. The panels were done away with altogether. In
I-ligh speecl racing cars, like Irord's GT-40, have experienced situations addition, the Mark II has 36% more glass area, and correspondingly
s,lrcrc thc lift was grcater than the entire loacl on the wheels, and the improved visibility forward and to the sicles. Both nose and tail were
car was actnally fl1,irrg. If the front wheels lift, tlre car cannot be stecrecl slightly lengthened. Although front spoilers ancl reverse airfoils at the
at all, or brakccl vcry u,cll; if the rear encl lifts, traction is lost. At I 50
nr.p.h., the C-lll has a 33 pourrd aerodynamic lift force acting on thc
tail of thc body, wlrile the nose is pressecl clorvn lr),, 66 pound negative
Iift force.
Dcsign work on the C-l I I began in November of 1968. By the first of
April, 1969, the car was completed. The first one was equipped rvith a
provisional body arrd a three-rotor, 335 horsepower engine. Tests at the
Daimler-Benz proving grounds in Stuttgart, on the open roacl and on
racing circuits were satisfactory, and by Iuly, the first of a small series
of six cars ccluipped with a streamlinccl body were seen in various parts
of Germany and in the neighboring AIps. Both the body and the chassis
underwent nrany changes that summer. After some redesign, three new
vehicJes wcre lluilt in August of 1969. The research division was author-
izecl to build more C-llls, a limited nunrber depencling on the require-
ments of tlre research program, to enable the engineers to carry out
research irr several areas on different cars at the same time.
The C-l I I was redesigned in ttre final months of L969 to accommo-
date the four-rotor version nf the Wankel engine. Adding one more rotor
to the engine did not charrge its package size and weight very much. The
four-rotor engine weighs 371 pounds complete-only 66 pounds more
than the three-rotor version. The length of the four-rotor unit is only
29.4 inches. This insignificant growth did not necessitate a stretching of
the wheelbase or redesign of the frame because the chassis of the first
C-l I I had, in fact, been clesigned with the ultimate installation of the
four-rotor powcrplant in nrind. There is some vacant space ahead of the
thrce-rotor unit, while the four-rotor engine fits snugly.
The body was complctely redesigned to eliminate the most severe
clrawbacks of the first prototypes: poor visibility, lack of luggage capacity,
ancl Jack of elegance. The C-lll Mark II \4/as first shown at the Geneva
Auto Show in Switzerland in March of 1970, then it was flown to New
York for its American debut. The original C-l I I body was designed to
have lorv air drag and to be functional only. Although it provided suffi- The C-l I I instrument panel has large dials for the speedometer and tachometer
(partly hidden by the steering wheel), plus small gauges for fuel tank, water tempera-
cient visibility for turrrpike clriving, improvement was necessary for city ture, oil pressure, and oil temperature. The gauge adjacent to the air vents is a clock.
446 APPLICATIONS THE IvTERCEDES.BENZ C.III 447

rear rvere tested in the wind tunnel, high speed driving demonstrated Thc Mark II has a new intcrior, rvith a nrorc civilizccl appcaralrcc. Thc
that these aerodynamic aids were not necessary. new instrument pancl alonc nurkes it sccrn ulnrost a cliffcrcnt cilr from
T|c Mark II also has a real trunk on top of the engine compartntent. thc one I hacl clrivcn cight nronths earlicr. '[']rc ncw irrtcrior has inr-
It IAS room for three suitcases, and evcn though it is located above thc 1>rttvccl Ircrttirrg nrrcl vcntilatiorr rtrrcl rtllows cnsy irrstnllittiorr of a ru«lio
errgine it is completcly heat insulated. Uhlenhaut pcrsorlally carried ottt ancl alt air conclitioning s),stcttt.'l'he clull wirrclshiclcl u'i1>crs \\'erc rc-
t[e firral "test" of its insu]ation. He put a pound of butter in the trunk, placccl by a single rr,,iper, rvhich clears a largcr area ancl is concealccl
clroyc several hundred miles at high speeds, then looked into the trunk when not in use. Iior better iclentification, the square tail lights v/ere
again. The butter hadn't melted. replaced by large round ones.
T[e redesign brought about an 8% reduction in the drag coefficient, What about the future of the C-lIl? Will l\4ercedes-Benz go racing
and the high speed lift characteristics are somewhat improved. The again, and try to rvin the Z4-hour race at Le Mans with the C-lll? Not
Mark II fias a slight aeroclynamic lift on the front end at speeds above very likely-the C-l t t is not a racing car. The C-l l l has many luxuries
125 m.p.ll. but the lift force never rises to the point where it is a problem, that are never found on racing cars, and much excess weight that would
even at the amazing top speed of 186 m.p.h. No changes were made in ruin its clrances. The engine is not clevelopecl or prepared for racing,
basic chassis clesign or weight distribution. The front suspension was and is certainly less powerful than the piston engirres in thc cars that
modifiecl in detail, and the steering gear was changed to give lighter now dominate the sport. Because of ncw clisplaccnrent linrits conring
steering. Springs and shock absorbers are better matched and spring into effect in 1972, the C-l I I engine wor.rld not be pcmrittecl in the races
travel was increased for a more comfortable ride at high speeds. that decide the \Vorlcl l\,lanufacturcrs' Chanrpionship.
Because of the higher output of the four-rotor engine, the cooling air The car is strictly an experinlent. Dainrler-Benz decidecl to let the
intake in the front grille and the air outlets on the hood have been public see it because they are proud of it, and because they fclt they
enlargecl and reclesigned for approximately 50% greater cooling capacity. neeclecl to charrgc thcir inrage sonlcwhat. Thc bulk of N'[crcccles-Benz
production cars are lou,-powcrcd family seclans. The C-l l l is the only
car that Daimler-Benz has ever shou,n in public that \\ras neither a
production-readv vehicle nor a racing car. A spokesmau for the companlr
said, "We decided that although we hacl a long traclition of keeping the
Iaboratory doors closecl, thcre \\,as no point in llcing bourrd b1'tradition."
There are no plans to put the C-l1 I irrto procluction at this tinre, but
it cannot bc rulecl out as a possible future rival to existing low-volumc,
Irigh-priced Grand Touring cars, such as Aston l\{artin , ßizzarrini, Iier-
rari, Iso Grifo, Lanrborghini, N{aserati, and N'{orrto,erdi. This is a glamor
nrarket in which Mcrcccles-Bcnz. has strclng traclitiorrs, ancl thc C-lll
would bc a fine reprcsentativc. It is a car that stancls out for its un-
cquallecl conrbination of technical rrovclty, high perfornlance, first-class
comfort, accurate stecring, powcrful brakes, arrcl balancccl roacl behavior.
SPECII-,-ICATIONS
MAKE I\l}lRCIiDlrlS-BIiNZ
NlODEL C.III IVIK I
Year inhoduced I 969
Year discontinued Ncvcr in production
Price Ncr cr listcd
'I)'pe of body 'I'\\'o-scir tcr tu'o-cloor coupe
'l.his cutaway drawing of the C- l I I
clearly shows the details of its construction.
'l'he entire drive train is positioned behind the seats, the front part of the car being Type of construction Shcct stecl platfornr franrc rvith fiberglass
takcn up with controls and instruments. (Drawing: T'heo Page) body
448 APPLICAT'IONS TI-IE MERCEDIiS-BENZ C.I I I 4{e
SPECIFICATIONS SPECIIi'ICA'I'IONS

MAKE MIiRCIiDIiS.BIiNZ MAKI] I\ I ll ltCI:DtrtS,B tiNZ

MODEL c-Iil Mk I MODEL c-lll Mk r
Disc diameter F 10.8 inches
Driving wheels Rear
Disc cliameter R I 1.0 inches
Porvcr unit position Miclships
Drum diameter F
Curb weight 2,125 pounds
Drum diameter R
Weiglrt distribution ftont/rear 45/5r%
Lining area )9.9 square inchcs
Porver/rveight ratio 7 .7 pounds per horsepower
Swept area
Iiucl tank capacity 3 1.6 gallons
Parking brake iluo-r.rvo drums on rear wheels
[.'uc] tarrk position Midships in door sills
Tires Michelin XVR
Polcr unit Wankel
Number of rotors
Tire size t95 VR 14
)
Wheelbase 103.2 inches
Charnber displacement )6.6 cubic inches (600 cc.)
Equivalcnt total displacement 220 cubic inches (3.6 liters)
Front track 54.3 inches
Rear track 5).9 inches
Conrpression ratio 9.3:l
Porver output 3 30 horsepower
Overall length 166. 5 inches
at r.p.m. Overall rvidth 70.9 inches
7,000
Torque 220 foot pounds Overall height 1+.3 inches
at r.p.m. 5,000-6,500 Test Results
Carburetion system Bosch direct injection with mechanical Acceleration times
pumP 0-30 m.p.h. l.g
Ignition system Transistorized coil and battery 0-40 m.p.h. 2.2
Cooling system Water and oil 0-50 m.p.h. +.0
Clutch Single dry plate 0-60 m.p.h. +.e
'Ira nsm ission Irivc-specd ZF (5DS-2 r/l) 0-70 m.p,h. 7.2
Gear ratios I 2. 58: I 0-80 nr.p.h. 9.3
z l.6l : I 0-90 m.p.h. I 1.9
l.2l:l 0-100 nl.p.h.
i I .00: I Top speed I
14.9
5'
5 0.84 : I 2 8t
R 3 tt2
Final drive ratio 3.77 :l 4 r40
Iiront suspension Triangulated control arms with non-paral- 5 r62
Iel pivot axes. Stabilizer bar. Average fuel consumption
Iiront springs Coil springs with concentric gas-pressur-
ized shock absorbers.
SPECI*A'IONS
Rear suspension Trailing and transverse links with stabi- MAKE MERCEDES-BENZ
Iizer bar. MODEL C.III MK II
Rear springs Coil springs with concentric gas-pressur- Year introduced r970
ized shock absorbers. Year discontinued Never in production
Steering system ball. Price Never listed
Turning diameter Type of body Two-seater two-door coupe
Ovcrall steering ratio :.t'.ulating Type of construction Sheet steel platform frame with fiberglass
Turns, lock to lock body
Brake system Ventilated disc F and R Driving wheels Rear
4r0 APPLICATIONS TI-IE MERCEDES-BENZ C.I I I 45r
SPECIFICATIONS SPIiCIFICA'f IONS
MAKE MERCEDES-BENZ MAKE I\,IERCEDES-BENZ
MODEL C-III MK II MODEL C.III MK II
Power unit position N{idships Disc diameter R I 1.0 inches
Curb weight 2,774 pounds Drum diameter F
Weight distribution front/rear 45/5r% Drurn cliameter R
Porver/wcight ratio 6.8 pounds per horsepower Lining area -)9.9 square inches
Iiuel tank capacity 31.6 gallons Swept area
Iiucl tank posftion Midships in door sills Parking brake ilro-r.rro drums on rear wheels
Porver unit Wankel Tires Dunlop
Number of rotors 4 Tire size F:4. 50/lI.60-1 5; R: 5. 50/13.60-l s
Chamber displacement 76.6 cubic inches (600 cc.) \Vheelbase 103.2 inches
Ilquivalcnt total displacement 29) cubic inches (4.8 liters) Front track 56.9 inches
Compression ratio 9.3: I Rear track 54.7 inches
Porver output 400 horsepower Overall Iength 17+.8 inches
at r.p.m. 7,000 Overall width 71.8 inches
Torque 289 foot pounds Overall height 44.1 inches
at r.p.m. 4,000-5,500
Carburction system Bosch direct-iniection with mechanical Test Results
Pump Acceleration times
Ignitionsystem Transistorized coil and battery 0-30 m.p.h. l.g
Cooling system Water and oil 0-40 m.p.h. 2.0
Clutch Dual dry plate 0-50 m.p.h. 3.8
Transmission Five-speed ZF (5DS-2 5/l) 0-60 m.p.h. 4.7
Gear ratios I 2.58:l 0-70 m.p.h. 6.6
z l.6l : I 0-80 m.p.h. 8.8
7 l.2l : I 0-90 m.p.h. rr.2
I 1.00: I 0-100 nl.p.h. r 4.0
5 0.84: I Top speed I 62
R 2 96
Final drive ratio ).77 zl ) r2+
Irront suspension Triangulated control arms with non- 4 r56
parallel pivot axes. Stabilizer bar. 5 ,tu
Front springs Coil springs with concenhic gas-pressur- Average fuel consumption
ized shock absorbers.
Rear suspension Trailing and transverse links with stabi-
lizer bar
Rear springs Coil springs with concentric gas-pressur-
ized shock absorbers.
Steering system ball.
'Iurning diameter
Overall steering ratio I.,'.ulating
Turns, lock to lock
Brake system Ventilated disc F and R.
Disc diameter F I 0.8 inches
 THE MUSTANG RC2-60 $3
r.p.m. For best performance, they should have been delayed until some-
24 where between 5,400 and 6,000 r.p.m.
The RC2-60 engine behaved normally during my entire time wittr it.
I drove it from New York to Indianapolis in 1966, then to Chicago,
Detroit and back to New York. It gave better fuel €conopy at high ,p..d
than the original Ford V8. Low speed economy was sacrificed f-or tract-
ability and smoothness. Oil temperature rose slowly to l50oF., where it
stabilized, and water temperature hovered around the l80oF. mark all
the time. Oil pressure was a rock steady 45 psi. Under load, throttle
response was immediate. When coasting, the engine acted Iike a high-
The Mustang comPression piston engine. Engine braking power was probably less than
with a piston engine of comparable power, but the automatic trans-
RC2-60 mission precluded any testing of this aspect of perforrqspss. Under full
throttle acceleration, the RC2-60 was noisy. The noise was partly a roar
from the air intake, partly exhaust noise. There was a disiinct lack of
vibration and no mechanical noise from the power unit. At steady speed
ON THE oursrDE, the Mustang in
front of me was indistinguishable from one straight off the River Rouge
assembly line. But instead of a 200 horsepower 289 cubic inch V8 under
the hood, I found a much smaller unit. This was Curtiss-Wright's RCz-
60 twin-rotor Wankel engine. When I turned the k.y, the engine came
.to life instantly. It had a steady idle at 800 r.p.m. and a dab at the
acceJerator sent the revs up to 2,000 instantly. A bigger dab sent the
needle on the tachometer up to 4,000 or 5,000 almost as fast. Without
Ioad, the acceleration of the RC2-60 was as rapid as would be expected
from a'Formula I Grand Prix racing engine.
The sound was totally unlike that of the NSU Spider, with its single-
rotor Wankel engine. It sounded almost like a six-cylinder piston engine
at idle, and almost like a gas turbine at speed. It shared with the NSU
engine the same lack of indication that there was any peak r.p.m.
Curtiss-Wright indicated the redline to be 6,000 r.p.rr., conservative for
a Wankel engine, but I went beyond it many times. The sound did not
change at higher engine speeds, and the engine gave the impression that
it was capable of going right on accelerating forever. It was extremely
smooth in terms of torque output, and was completely vibrationless.
Despite the fact that the engine mounts had not been tuned for the
RC2-60, the engine sat perfcctly still even at idle speed. The RC2-60 was
coupled to a standard Ford three-speed Cruise-O-Matic automatic
transmission. The transmission was the same one that had been bolted
to the original Irord V8, and the shift points remained as they had been : I ' " ', *. .
for the piston engine. As a result, the transmission was poorly mated to Irord Mustang experinrental vehicle equipped with the Curtiss-Wrigtrt RCZ-60 U5
the torque curve of the RC2-60. Upshifts occurred at a little over 4,000 en$ne.
4r4 APPLICATIONS TI-IE MUSTANG RCz-60 45,
the engine was quiet, no matter whether it was running at 3,000 r.p.m. The original V8 was rated at 200 horsepolver, the RC2-60 at 185. With
or 5,500 r.p.m. under Iight loacl. Wlrenever it was laboring, however, the a manual transmission, there is no doubt that the Wankel would have
noise lcvel immediately rose. Acceleration tirnes were as follows: beaten the reciprocating piston engine on acceleration.
The engine in the Mustang carried the designation RCZ-60 U5. This
Ford V8 289 RC2-60
version had been created by Curtiss-Wright specifically for passenger car
0-30 m.p.h, 2.8 sec. 2.9 sec.
propulsion. The RC2-60 U5 differed from its preclecessors in several small
0-40 rn.p.lr. 5.0 sec. 5.2 sec.
ways, though the geometry was identical to the RCI-60.
0-50 m.p.h. 6.8 sec. 7 .2 sec.
0-60 m.p.ll. I 0.0 sec. 10.6 sec. Rotor radius was 5.75 inches and rotor width 3.0 inches. Rotor eccen-
0-70 m.p.h. I 3. 5 sec. l+.7 sec. tricity was .87 inch, giving a K factor of 6.9 and a potential compression
0-80 m.p.h. 18.4 scc. 20.0 sec. ratio in excess of 10.0:1. The rotor sealing configuration and basic rotor
and housing cooling systems remained unchangccl from tSe RCI-60, but
the U5 modification included minor improvements in the rotor and rotor
housing, the end housirrg, bearings, arrd mainshaft. Complctely new
were the intermediate rvall, the ccntcr bearing support, gas ancl water
manifolds, accessory drives and housing, oil pumps, power take-off
aclapters, starter mounting, and oil transfer bearings.

'l-he automotivc version of the RC2'60 U5 engine. Closeup of the RC2-60 U 5 installation in the Irord lr,lustang.
456 APPLICATIONS THE MUSTANG RC2-60 +57

The dual side intake ports were combined with peripheral exhaust Opens Closes Opens Closes Durdtion
ports to give the following timing:
Porting limits available
RC2.6O U5 PORT EVENTS
in present castings
Ponr Truwc Intake-max. 583" 855' 29"ATC 30"ABC I8l'
EQUTVALENT FOUR-STROT(E Exhaust-max . ?02" 6l I " 15'BBC 47"ATC 272'
RECIPROCAI'INC ENGINE Overlap-max. 28" 18'
RC CRANK ANGLE CRANK ANGLES

Opens Closes Opens Closes Durdtion The spark plugs were mounted 25", or 2.13 inches, before the minor
Present porting axis and the distributor was set to give 40-4r" spark advance at 5,000
Intake 583' 840' 29"ATC 20"ABC 174" r.p.m. All other accessories \4,ere standard automotive units. The two-
Exhaust 202', 598' 45"BBC 39"ATC 266. barrel carburetor came from a Buick, 'and the alternator and ignition
Overlap 15" 10" system was adapted from the system that came with the original 289
Ford V8. The RC2-60 U5 had dty sump lubrication ancl carried a
oil tank and oil cooler, neatly installed under the hood of the
separate
?6t
Mustang.
Irdl.o[
o.I
o zc[
il .'f RCz - OO 8E8T ECOITOTY
.,1
H o
IlI
H 2.r
-l
,,1
9
i .,f
STD. 6 CYL;
E.,f
3 aol- § "'l
ül
E tgt
EI
I;f
RC2 - 60
BEST [email protected].
E
r! '01' g rcl-
r rr[ Z,
t.ol rrF
I

ä ,.1.
EI !II ,rl-,

5
ol '0r -l' ,.1-
ol
g ''f E
II
,ol-

*ol ,.f
ctTY TRAFFTC AvO. 20-GO AVG.50-rc
H ''r HFO MFC TPO
RCa-60oESTECOI.OIY t6.5 ?2.e t7.0
-Go t 3.2 2l . 3 J rsF
3l
ncz t7.6 CITY TRAFFIC
IPG
sTo v-t t6.2 zr. t 17. I
L' rel ice-6() tEsr EcoxorY ta.3
STO 5 CYL BEST ECOTÜOTY t5.6

il STO 6 CYL PROOUCTIOil ra.7

30 +o oo 60 *
VEHICLE SPEEO - MILES PER HOUR L, 50 .o 00 60 70 00
Tests showed that the RC2-60 had better high-speed fuel economy YEHICLE SPEEO.TILES PEß HOUR
than a production V8 of similar performance, but worse low-speed
economy. If tuned for economy, the RC2-60 could surpass the Comparison of fuel consumption at constant speed, with similar
econorny of the V8 throughout the speed range, but at the cost vehicles using three different power units-one RC2-60 U5 and two
of performance. different production six-cylinder piston engines.
458 APPLICATIONS THE MUSTANG RC2-60 459

Colcl-starting tests with the Iatest version of the U5 demonstrated SPECIIIICATIONS

consistent automatic starts at - 20"F'. with 5W-20 oil ancl a 90 A/H
MAKE, I.-ORD
battery. The most dranratic inrprovement resulted from a drain in the MODEL MUSTANG RCz-60
exhaust manifold to prevent flooding. As a result of space limitations in
the Mustang installation, the exhaust manifold was designed with an 5

2.20:l
initial su,€ep upwards rvhich would trap fuel against the rotor flank, mak- Final drive lrr,o 3 .20:l
ing a start clifficult unless it occurred on the first few turns. Other Iiront suspension Upper A-f ranre and lorvcr I-rod control
moclifications were also made, inc]uding an automotive type starter. The arm rvith trailing clrag strut and stabi-
automotive ignition systcm also helped, particularly when the spark plug Iizer bar.
electrodcs were brought close to the trochoid surface. Front springs Coil springs
Rear suspension I-beam axle
SPECIIIICATIONS Rear springs IVIulti-lcaf scmi-clliptic leaf springs
Recirculating ball
MAKE IIORD
Steering system
Turning diameter 39.5 fcct
N4ODEL MUSTANG RC2.6O
Overall steering ratio
Ycar introduced 1965
Turns, Iock to lock +.6+
Ycar discontinucd
Brake system Drums all around
Price Iixperimental only
Disc diameter F
'l'ypc of body 2-door coupe
'l'ypc of construction Disc diameter R
Platform frrrnre rvith steel body
Drum diamcter F
I)riving rvhccls Rcar
Drum diarneter R
l'orver unit position liront
Lining area 212 square inches
Curb wcight 2,57 5 pounds
Srvept arca
\Vcight clistribution fron t/rear 53/47 %
Parking brake
Pou,cr,/rvcight ratio 13.+ pounds pcr horscpower 'I'ircs Co«ldycar powcr cushion
I,'ucl tank capacity l6 gallons 'Iire size 6.95 x l+
I,'ucl tank position Rcar (undcr trunk floor).
\Vheclbase I08 inchcs
Pou'cr unit \\/ankcl
Front track 56 inchcs
Nur"nbcr of rotors 2
Rear track 56 inches
Cha nrl>cr clisl>lacenrent 60 cubic inches (989 cc.)
Overall length I 8l .6 inches
Iicluivalcnt total displaccment 2+0 cubic inches Q,992 cc.)
Overall rvidth 68 inches
Cornprcssion ratio 8.5: I
Porvcr otrtput I 8 5 horscpower
Overall height 5l .l inches
fl t r.1>.rtt . 5,000
'l'orqttc l9Z foot pounds
at r.p.m. 3,7 50
Carl>urction systcnr Onc trvo-barrel Rochester
lgn ition
systenr Coil arrd battery
Cooling sl,stcrn Watcr and oil
Clutclr I-I1'clraulic torque converter
'I'ril n srn ission Cru isc-O-Ivlatic torcluc converter and
threc-spced planetary transnrissiou
Gcrr ra tios I 2.46:l
2 1.46:l
.l
)
4 1o''
 CONCLUSIONS 46r

25 engine except as it applies to aircraft propulsion. The agreement is sub-

iect to termination by General Motors at the end of each contractual
year on one day's notice. The agreement provides that General Motors
will pay $5,000,000 by December 31,1970 upon approval by the parties
concerned. Subject to GM's right to terminate the agreement General
Motors will pay $10,000,000 at the end of the fifth year. At the comple-
tion of these payments, General Motors will have the right to manu-
facture and sell the Wankel rotary combustion engine on a worldwide
basis without further payments."
The contract was signed formally on November 10, 1970. The $50
Conclusions million will not be paid directly to Audi-NSU Auto-Union AG and
Wankel G.m.b.H. It is estimated that Curtiss-Wright will collect $23
million over the five-year period spelled out in the contract. Of the $27
million to be paid to the German firms, NSU will gain $16,200,000 and
Wankel G.m.b.H. $10,800,000. Because Audi-NSU Auto-Union AG is a
subsidiary of Volkswagenwerk, it is important to note that only 30% of
Dors rHE WeNxru ENGTNE HAvE A the $16,200,000 collected will be kept in the company's coffers. That
future? We have seen that it has potential to match the reciprocating amounts to $4.86 million for NSU. The other $11,340,000 will be clis-
piston engine in terms of life between overhauls as well as emissions. tributed to former NSU stockholders in proportion to their holdings of
What is more uncertain is whether both will be replaced by an uncon- Wankel vouchers.
ventional powerplant, forced upon the world by strict regulations insti- Why does GM want the Wankel engine so much that it is rvilling to
tuted to protect the environment and avoid imbalance in the ecology. pay $50 million for the rights to it? Obviously, even a company the size
Arguments have been raised by our automobile manufacturers to the of General Motors does not propose to invest the sum of $50 nrillion in
effect that it will be easier to meet future clean-air laws by developing the a routine investigation of a possible future source of motive power. It is
reciprocating piston engine than by investigating and experimenting a definite commitment, and one that comnlancls other automobile
with other power sources on the assumption (which may turn out to be manufacturers throughout the world to take similar steps. It is clear that
unfounded) of their being pollution-free. Ten years ago it was impossible GM is serious about building Wankel-powered cars at some future date.
to make any realistic estimate of the status of the Wankel engine in The question has been asked, "Where Chevy goes, can Ford be far
world industry as of 1971, and today, it is iust as senseless to try to set a behind?" Ford is planning to acquire a 35% interest in the Toyo Kogyo
timetable for the future adoption of the Wankel engine by the leading Company-a step which would put Ford in on the ground floor as far
autornakers. As is the case in all responsible reporting, it is best to try as Wankel engine research and production technology is concerned, but
to look at the future of the Wankel engine by examining what is affect- which would bring the Dearborn giant no closer to the manufacturing
ing its status today. and marketing of Wankel-powered cars in the United States. That
On November 2, 1970 GM issued the following press release: "General could follow only after a negotiated contract-necessarily costly-with
Motors has agreed to enter into a worldwide, nonexclusive paid-up Curtiss-Wright Coqporation, Audi-NSU Auto-Union AG, and Wankel
Iicense agreement with Audi-NSU, Wankel G.m.b.H., and Curtiss- G.m.b.H. When this step will be taken is mainly dependent on hou,
Wright to facilitate its further intensive research and developmerrt much and how fast GM can progress with the Wankel engine.
studies of the Wankel rotary combustion engine to determine whether it Because a great deal is known about the operations of General Motors,
is suitable for GM automotive applications. The tentative agreement the timetable for the Wankel engine can be set rvithin certain linrits. To
among the companies, subiect to the approval of the managements of the understand this, we must look at how teclinical innovations are developed
otlrer companies on or before December 31, 1970, covers the manu- at GM, brought to maturity, and placed in production. Consider GM's
facture and sale by General Motors of the Wankel rotary combustion record of progress in applying gas turbines to autourobiles. Last year,
 462 APPLICATIONS CONCLUSTONS 463

General Motors' Detroit Diesel Engine Division announced its intention search organization of its type in the rvorlcl. Dr. Laurence R. I{afstad
to be the first on the market with a commercial turbine engine for heavy was vice president of GM in chargc of the research staff when the GN,l
cluty vehicles. Detroit Diesel's engine is scheduled for initial production Technical Center at Warren, IVfichigan, was dedicated in 1916. In his
in mid-1971. Developing the heavy duty gas turbine engine from an opening address, he said: "It is in the tradition of scicnce first to ob-
interesting but impractical concept into a commercial reality was a serve, then to understand, ancl finally to utilize the forces of nature. N,Ian
process sparrrring more than 20 years. The GM Research Laboratories has been doing this since the clawn of Jristory, but at one time discoveries
brought the engine fronr the concept stage to the point of technical were made in a 'hit or miss' rnanner by lone investigators poorly sup-
feasibility, and in 1961 turnecl it over to Detroit Diesel. In the ensuing pliecl with information and equipment. We have now learned to bring
years, Detroit Diesel has met the challenge of making the turbine engine trained scientists and engineers together in such rvell-equippecl labora-
cconomically practicable. tories as are at the Technical Centcr, to make cliscoveries aucl clo,cloll
The experimcntal vehicle series began with the Firebird I in 1953, the new ideas. This is an inrportant new conception in our modern econonry
Unitecl States' first gas turbine powered automobile. A non-regenerative which has resulted in greatly accelerated technological progress."
370 horsepower engirre propelled the sleek mobile laboratory-which "General l\4otors has supported a rescarch organization for over 40
today is a museum piece. Also in 195), they built the Turbo-Cruiser I, years. Research discoveries and developments have contributed to all
a turbine-powered bus. 'Ihe GM family of experimental turbine vehicles the products of GN{'s n)anufacturirrg Divisions. Our automobiles, diesel
burgeoned with the arrival in 1955 of the ITirebird II, a family car with engines, household appliances, Ioconrotives, ancl jct cngincs have all «lc-
a regenerative engine; the Turbo-Titan I, a heavy duty truck, in 1956; pended upon a continuous research program for their constant intprove-
the Irircbird III, with a more powerful GT-305 engine in 1958; and the ment. The research staff is the one organization in General l\{otors that
Turbo-Titan II the following year. deals solely with funclamental, long-range research. Its scierrtists ancl
The fifth generation GT-309 engine, boasting a rugged and simple engineers are concerned with proiects that continuously cxplore the
design, was introduced at the New York World's Fair in 1964. It was future. The program is divided into fundamental scientific rcsearch,
fielcl testccl in several heavy duty vehicles, including Chevrolet's Turbo- Iong-range engineering research, and advancecl engirreering devclopnrcnt.
Titan III truck and GMC's Turbo-Cruiser II bus. In the new Research Staff facilities at the Technical Ccnter emphasis
Genera] I\{otors has enornrous investnrcnts in machine tools that pro- is placecl on basic proiccts u,hich, whcn succcs.sful, will rc.sult irr tcclr-
duce parts for piston engines, and there is little more than normal inertia nological improvement."
invoh'ccl in getting a totally new kind of powerplant through its re- "The Research staff has a trvo-folcl responsibility to Gcneral N,Iotors
search and dcvelopnrent phase ancl up to the point where it can be management ancl the manufacturing divisions. First ancl forernost, we
proclucecl ancl nrarketed economically. Will GN,l put the \Vankel engine discover and develop fundametrtal irrformation which will becopre t|e
through a similar program, and would a Wankel-powered Chevrolet be basis for the new products of tomorrow. Second, our spccialized pcr-
20 years away? Convincing feasibility stuclies have been undertaken sonnel and facilities are available for usc by the engineering clepart-
elservhere. The work of NSU, Toyo Kogyo, l)aimler-Benz and Citroön ments of the variotts GM manufacturing units shoulcl they choosc to use
covcrs all the fundamentals, and furnishes eloquent testimony as to the them."
advisal:ility of (l) dcveloping Wankel engines for passenger car applica- The matter under stud1, b1, thc GIVI research staff is not siurply to fincl
tions artcl (Z) clesigning uew passenger cars around such power units so a "Yes" or a "No" for the Wankel cngine, but to spccify tcchnical
as to take full aclvantage of their small bulk and low weight. clirectiolls for all proclttcts that fall rvithin thc corporatiou's splrcrc of
If GIvl top nranagenrent were to give the go-ahead order on Wankel interest. Any tinrc thc subicct of a ncw ty1:c of powcrl>lrrnt courcs gp,
enginc.s toclay, it woulcl nevertheless be years before you would be able tllc GM rescarclt engirtccrs gct clowrt to tlre lla.sics, which nuly rrot llc
to btry a Wankel-powered Chevrolet, Pontiac, Oldsnrobile,' Buick or involved specifically with rotary or reciprocating nrotion, but g,ith tfic
Cadillac. It is naive to think that basic research has been completed. basic energy ancl materials.
Research, testing, redesign, ntore testing, design alterations, and material Being the world's Iargc.st ntallufacturer oI grouncl transl>ort cquil>nrept,
clevelopment coulcl take ycar.s. GM has unquestioned ability to under- GN4 is extrernely interestecl in future sources of encrg),. Licluicl pctrolcunr
take proiects of this kind. The corporation has created the Iargest re- fuels are the mainstay of its proclucts toclu1,. It is pos.sil>lc that in thc
464 APPLICATIONS CONCLUSIONS 465

future uranium and thorium may be just as commonlv understood words tended for production at the GM scale has to be redesigned for mass
as gasolinc and fuel oil are today. Gasoline engines, diesel engines, gas production, with due regard to nraterials handling, parts handling and
turbines, steanr crrgines, free-piston engines, ancl the fuels which they assembly. These programs can be conducted side by side only to a certain
burn all provide pronrising projects for better powerplants in the vears extent, for there are distinct limits. Production engineers cannot get very
ahead. How to make the best use of atomic energy in industry and far beyond preliminary studies until the design has been finalized; other-
:'

transportation of the future has become GM's newest long-range prob- rvise, much production engineering work could be suddenly rendered
lem. A small example will illustrate the scope of nuclear power: one worthless by a basic design change originating in the research depart-
ri

pouncl of uranium, U-Z)S, has as much energy as 1,300 tons of coal, ment. Only when the final production design is ready can tooling begin.
F

.1

which is a ratio of I to 2,600,000-this is the incentive. It takes time to make tools, dies, iigs, and special machine adaptations. :,

Another big problem is materials. At present it takes 18 tons of For an all-new engine line, it takes about I8 rnorrths fronr placirrg the
1l

material per vear to keep an American citizen at his present standard of first orders until the last deliveries are made. Then the purchasing de-
livirrg. T'his aclds up to the astronomical figure of ovet 21,/z billion tons of partment comes in. Parts not manufacture.l by the conlpany itself have
nratcrial per year for the Urritccl States alone. Last year General Motors to bc designecl, clcvclopccl, ilpprovccl, arrcl orclcrccl frclul outsidc suppliers.
procluccd over ten nrillion tons of automobiles. T'lte GM research labora- Stocks have to be built up. All this takes time.
tories support a never-ending search for new materials, new processes,
:

Even when production can begin, all is not cut and dried. No matter
and new methocls of fabrication. A part of this program is a group of en- how thorough the methods, the engine line cannot run at full speed
gineering projects aimecl at increasing the fatigue life of parts and utiliz- right from the start. Engines from the pilot line have to be tested to :

irrg materials more efficiently. make sure they perform according to specifications and match the proto-
The research staff's projects are divided between two main groups of type. Minute design or nraterial specification changes may be needed. I

departments-applied science on one hand, and engineering research And above all, tooling changes may be called for. A running-in period
on the other. The applied science departments are concerned primarily for any production line turning out an all-new engine is taken for
rvith fundamental investigations into such fielcls as engine combustion, grantecl. It is all the more irnportarrt irr the cnse of the Wankel engine.
elcctronics, ultrasonics, paints and finishes, electroplating, instrumenta- On this basis, assuming a full-speed-ahead order, a blank-check budget,
tion, and radioisotopes. The engineering research departments conduct and no unforeseen setbacks, the Wankel-powered GN{ car is a mininrum
basic studies in spark-ignition engines, gas turbines, vehicle suspension of five years away, more likely seven to ten years. That may seem a
components, various types of bearings, and the fatigue life of automotive long time, but we are not talking about iust another license agreement.
components. While these are only a few of the many areas in which We are talking about mass production by the largest auto company
research staff engineers and scientists work, they are typical and serve in the world. There can be no doubt that if GM goes to Wankel en-
mainly to illustrate the general nature of the investigations continually gines, Ford and Chrysler will have to follow. So will Volkswagenwerk
in progress. (who may well beat GM's timetable, now that VW owns NSU), Fiat,
TIre Autontotive Engines Department centers its activities around the British Leyland, and fapan's auto giants, Toyota and Nissan (Datsun).
development of rnore efficient piston-type automotive engines, engaging It is no mystery rvhy GM is so interested in the Wankel engine. GM
in research on basic engine design and on specific engine components. takes an active interest in all types of automotive power units. The
Studies of fuel economy, octane requirements, carburetion, and other particular advantages of the Wankel engine have been set forth in an
pcrformance factors are made on dynamometers and by road testing. earlier chapter, but there is one area of unknown potential that has ncrt
\\/ankel engines have not yet been incorporated into this group. They been mentioned-automation.
are the resporrsibility of another section, the Mechanical Development Labor is a big part of the production cost, and labor costs are rising,
Department, rvhich is concerned with a variety of proiects including un- while material costs have not increased proportionately. The industry,
conventional powerplants, fatigue testing, diesel engines, free-piston Ied by GIvI, is trying to cut costs by reducing the labor content in every
engines, friction, and bearings. product, along every step of the way. This is where the Wankel engine
When the research and development engineers have a design they feel could revolutionize powerplant production. Because of the inherent
is satisfactory, it is still a long way from mass production. Any item in- sirnplicity of the Wankel engine and its low number of parts, it holds
466 APPLICATIONS CONCLUSIONS 167

the promise of automated rnanufacture and assembly. Highly complex The licensing agreement with Audi-NSU Auto-Union AG, Wankel
automatic transmissions are being assembled with a high degree of G.m.b.H., and Curtiss-Wright should allow GM to begin arming itself
automation right now. The assembly processes of a Wankel engine fully for a major assault on whatever problems still remain, and at the
seem considerably simpler. There are no basic tooling problems involved same time should advance the progress of the Wankel engine not only
in manufacturing parts for Wankel engines, for even epitrochoidal with strict reference to technological ancl industrial considerations but
chambers can be machined. All Wankel engine parts probably can be also in terms of public acceptance and support.
made on automatic transfer lines, the number of operations varying
depending on clesign, finish and tolerances required.
When GM reaches the point where the research staff is ready to
release the Wankel engine for production, the management has more
decisions to make. GM could create a special Wankel engine division,
similar to its Detroit Diesel Engine Division. The car and truck divisions
then would be supplied with power units from the Wankel engine di-
vision to whatever extent the divisions choose, within the framework
of their own autonomy. Detroit Diesel, by the wäy, has no monopoly on
diesel engine production within General Motors. There are other pos-
sibilities. GM could assign Wankel errgine production to an existing
division, such as Detroit Diesel, AIIisoD, or Electro-Motive. Or the
corporation could give the Wankel engine to one of the car divisions,
such as Oldsmobile which has a record of pioneering new concepts (auto-
matic transmission, high-compression V8s, front wheel drive). The
choice will be made depending on what scale the corporate management I

wants to build the Wankel engine. We have already seen that produc- t

tion would have to start gradually, which means that the Wankel I

engine could not suddenly become the standard power unit in a large- I

volume car line such as the Chevelle. Again, GM has a multiple choice.
I
I

The Wankel engine could be rnade optional for one or more low-volume
car lines, from one or more divisions. Or it could be made standard in
one low-volunre car line, from one or two divisions. The latter course
is more likely; GM must be anxious to get the Wankel-powered cars
out in the field, and that is best controlled by using it as standard I
I

equipment. Sales of options are less predictable. I

It is conceivable that GM could assign Wankel engine production

I
I
I

to one of its overseas divisions, such as Opel in Gernrany, Vauxlrall in

England, or Holden in Australia. However, if the corporation intends
to install Wankel engines in a large number of American cars, it
would make better economic sense to manufacture the engine within I

the U.S. When the decision is made, the license agreement allowing
t

GM to exploit the Wankel engine will work in two ways. GM will get ,l

full access to test reports and research material compiled by other Ii-
censees, and the corporation will be required to share its own findings
I

with the other companies engaged on Wankel engine development.

{

\
I
 OTHER ROTARY ENGINES 169
drum able to rotate about an cccentric irxis relative to that of the casing.
The rotor was cylindrical and revolved eccentrically within a cylinder of
greater diarneter. Rotor rnovernent was eccentric in relation to the outer
cylinder, so that the rotor touched the casing at only one point on its orbit.
The rotor had four radial groovcs at right unglcs, rvith seal strips to separate
thc four working chanrl>crs.'fhc vilncs wcrc loosc in thcir slots, free to move
in and out. Thc vanc cdgcs werc maintairrcd in continuous contact with
the inner surfacc of the casing. However, the rotational speed of Ramelli's
purnp can hardly have been sufficient for centrifugal force to make the vanes

Appendix
OTHER ROTARY ENGINES

Sor.,rr oF THE MECHANICAL move-

ments necessary to make rotary engines possible were invented long before
the stcam cngine or the internal combustion engine had been thought of.
These mechanisms were used mainly for water pumps. Since pumps share
certain characteristics of variable displaccment with heat energy machines, all
types of pumps can be converted into heat engines of some kind. The re-
ciprocating engine, for instance, corrcsponds to the kind of pump used to
inflate bicycle tires. The bicycle pump has a piston and a cylinder. What the
engine has in addition is fucl, and a spark plug. Naturally the valving differs
in exccution if not in principle. A centrifugal pump could be converted into
a gas turbine. A vane type pump could become a rotary engine. Rotary
engines in the strictest sense have unidirectional movement with uniform or
variable velocity. They are not affected by alternating inertia forces due to
changes in the direction of movement.

RAMELLI

The first evidence of a practical rotary punrp appeared in I 588 in Genova,

Italy, in a book published by a leading military engineer of the time, Captain
Agostino Ramelli (1530-l 590), €Dtitled "Le Diverse e Artificiose Macchine
del Capitane Agostino Ranrelli." In it were illustrations of 195 inventions,
ranging from pumps to gear transnrissions. Ramelli's designs included over
100 different pumps, plus a variety of windmills, sawmills, screw iacks and
derricks. They were designs-not realities. Some of the machines he illus-
trated and described could not have been made with the tools and materials
therr in cxistcnce. Ramelli's rotary pump, in an installation proposed by the inventor,
The Ranrelli rotary pump consisted of a cylindrical casing containing a driven by a water wheel.
470 APPENDTX OTI-IER ROT'AITY ENGINIIS 47r
providc adequate sealing. As no spring-loading was supplied, he seems to watcr fountairrs in tnnny liurollcnn citics: R«lntc, Pntgue, Regcnsburg arrd
have relied upon gravity, placirrg thc ccntcr of drunr rotation in the upper Salzburg. It wns clrivcrr not lly' ttuul-powcr llttt frotn lln ovcrshot rvrtcr whccl
part of thc casing. It is possiblc that this rcliance upon gravity accounts for sct in a stream.
the usc of four vanes instcacl of two, which ought to have been sufficient for
water-pump applications. Ramelli spoke of the vanes as "paddles" and
described the punrp as a "machine designed for the drawing of water from
docks or foundations by the power of two nlcn who turn the two cranks of
an eccentric wheel within the covering cylinder. The cylinder is made of
metal or other suitable material closed and well fastened by screws. It has but
one aperture for the entrance of water, and is firm and immovable. When the
water entcrs the cylinder it is forced around by the movement of the wheel
and its four sliding pieces (which move easily back and forth as required)
continually through a tube." The principle of Ramelli's pump is used to this
day in certain types of compressors and pumps, and has inspired a host of
invcntors through the ages to try to improve upon the basic idea.

Schematic of the Ramelli pump, H

showing a version equippeä wi[h

e single floating radial vane.

PAPPENHEIM

Thc gear-§pe oil pump most commonly used in modern automobile

engincs is based upon a rotary pump dating back to 1636. It is known as the
Pappenheim pump, but the invcntion is variously ascribed to Grollier de
Sen'i6re or to Pappenheim. A grandson of the French inventor published a ExOrxGlt" twerl3c.

collcction of his mechanical and other devices in 1719, which would indicate
The Pappenheim pump.
that it originated around 1640. But the punrp was described by a number of
Gcrman writers on record as early as 1636. The gear-type pump is in fact
so old an invention that it remains uncertain whether Pappenheim was the
nAmc of its invcntor or the village in which he livedl
The Pappenhcim water pump used six-toothed gears of equal size, WATT
mounted side by side and mcshing. The two shafts were parallel, but rotated
in opposite directions. The tceth mcshed in the center so as to assure positive The man most frequently credited with the invention of the piston-type
displacement of the fluid. No provision for sealing was made beyond working steam engine, |ames Watt (1736-1819), also ,Cid much work on steam
with minimum clearances. Still, the Pappenheim pump is considered more powered rotary engines. The son of a carpenter and shipping merchant, he
advanced than Ramelli's because it did not have sliding, reciprocating vanes spent much time in his father's shops and became a skilled artisan. At the
and no off-balance eccentric rotor. The Pappenheim pump was used for age of 19, Watt went to London for training as an instrument maker. When
 172 APPENDIX OTHER ROTARY ENGINES +73
he returned to Scotland several years later, he was appointed instrument cngines of diffcrcnt typcs.'l'hc first itncl rathcr primitivc Watt rotary nrachinc
maker at the University of Glasgow. The idea of a steam engine was sug- was ntitdc of a rving-shupccl rotary blade insicle a cylindcr. The blade was
gested to him onc day in l7r9 by Professor |ohr Itobison of the University not pcrrnittcd to conrplctc onc full revolution. It opcncd arr intake port
of Glasgow. and approached an cxhuust port, thcn retunred to its original position by
ro,ersing its dircction of r«ltation. Thc int:rkc and cxhaust ports were sepa-
rated by a curved radial rvall anchored at the rotarl, blade axis and providing
a seal against the housing.
\Vatt's enthusiasrn for the rotary engine secms to have run hot and cold.
Apparently he was rcady to abandon all work ou rotary machines as early
as 1777., beforc the first one had even been tcsted. The first test report dates
from 1776, rvhcn Boulton wrote that their last wheel had a power equal to
three horses. Thc dcfects of thc machine had then become apparent. There
was too nruch friction and too nruch leakage.
In 1782 Watt patented another form of rotary engine which consisted
of a rotor platc rnountcd on an axlc aud adapted to turn inside a cylindrical
casing. An "abuturcnt" was hinged to thc casirrg and bore against the axle
at the outer end. Stcam was adnrittcd oll one side of this abutnrent and
Ih. rglqy engine designed by
- exhausted fronr thc other. It is doubtful that this engine was ever built.
fames Watt (1782). Holo,er, thcrc arc clistinct srrggcstions that the machine \\,as patented rvith
thc ob jcct of discourrging onc of thc finn's rvorkuren (Crnreron) fronr
spending furthcr tirne on dcvcloping the idca.
The problenrs of inventor (cnrplol,ee) and inclustrialist (employer) relation-
ships so conrnlon no\\'adu1,s irrc rrot a phcrronrcllon unicltrc to the 20th
century. There is reason to bclieve that \\/att's 1782 rotary nrachine patent
did in fact cover not onc of his own idcas, but Cameron's. Cameron had
ilpproached Watt rvith an idca for nraking a rotarl' or perhaps a semi-rotary
virtual monopoly for the production and marketing of steam engines for stcam engine, but \\/att said that hc had himsclf thought of the same type
many years. of eugine sonre ycars l>cforc and hacl evcn nrirdc n model of it. Cameron
rcfusccl to bc clisc«runrgccl urrrl got pcnnissi«lrr to rvork on his o\\,ll idea, brrt
all his cfforts rvcrc in vain.
Shortlv after \Vatt's patent was issucd, Canlcron lcft Boulton & Watt. In
a patcnt application datcd 178+, Carncron dcscribcd a rnachine much like
Watt's but having tu,o rcciprocating abutments instcad of the single hinged
flap, plus another intcrcsting nrachinc, incorporating a rotor which moved
on the idea, in it deep hclix, right-hancl for half a turn and lcft-hand for the rest, so that
l'lte first mention of Watt's rotary steam engine occurred in a letter to his it completcly cncirclcd thc shaft. This rotor was contained in a cylindrical
casing of the san're dianreter as that covering the top of the helix blade. A
sliding franrc projcctcd into the sidc of thc rvorking chamber and 6tted over
the helix, it was capable of moving up and down as the helix turncd, acting
as an abutnrcnt. T'hc frarnc adnrittcd the stcarn. This frame acted as a slide
valve, and could bc givcn a clual-dircctional action so as to give a continuous
irnpulsc torvarcls turnirrg thc hclix. Sincc thc rnachinc was dependent upon
schcme could be for a turbine, but sounds more like two intermeshing right thc usc of a separatc conclcnscr, it infringcd Watt's patent and was refused.
and left hand screws inside a cornmon casing. The machine seems to have
been built about 1771, but it was not till t77+ that it was actually tested.
Between 1772 and 1782 Watt designed a series of rotary-piston steam
474 APPENDIX OTFIER ROTARY ENGINES 175

HORNBLOWER MURDOCH
Watt had many rivals, insidc and outside his own firm. One of the outside William Murdoch, another enrployee of Boulton & Watt, used the Pappen-
rivals who gAve hinr thc nrost troublc over patent clainrs for rotary machines heim gear-type pump to contrive a rotary steam cngine in 1799. He fitted
was fonatharr I-Iorrrblorver, the inventor of thc cornpound engine. His rotary, rvoodcn apex seals in each gear tooth to inrprove gas sealing.Steam was
llatented in 1781, is renrarkable for its ingcnuity, in terms of its conception aclnritted through a mushroonr valve and was cxhausted to a separate con-
as wcll as its construction. It consisted of two hollow vanes, mounted to- denser mounted below the rnachinc. The date of its appearance and the
gcther likc thc hingcs of a door insidc a cylindrical casing concentric with usc of a separate condenser suggest that Murdoch may have delayed
their axles. The axis of one vane projected through the hollow center of application for a patent until Watt's patcnt covering the condenser expired.
thc axis of the otller vane, so that the vancs could move independently. The The air pump for tl-re condenser was driven by a crank from the main shaft.
inside of the vane which was nrounted on the longer shaft was supplied Each tooth of each gear carried iu its head and around its flank a layer of
rvith steanr through a passage in the shaft-the other vane was connected packing in a slot.
rvith thc cxhaust pipc in n sirnilar rnanner. One vane had mushroom valves Murdoch's engine had no external gears to kecp the two rotors in correct
ärrrangecl in cach of its faccs and adapted to lift outwards.'Ihe other had angular rclationship and to pro'cnt wcar of thc contacting faces. Despite
sinrilar valvcs but lifting inwards. The two valves on each vane were inter- the rotor seals, the cnginc was still quitc incfficient, as other sealing problerns
connccted so that if one was open the other had to be closed. Despite its remaincd unsolvcd. It also suffcrcd from cxcessive friction. An engine of this
nrcrits, thc l-Iorublowcr engine coulcl not be nrade to work on account of its kind dcveloping about % lr.p.w:rs sct up at thc Soho Foundry in 1802 to drive
frictiorr and lcakagc problcnrs. the machines in Murdoch's workshop. It pro','ed not to be capable of meeting
any useful pllrpose.

BRAMAH

)oseph Bramah, versatile eugineer, large-scale inventor, and operator of a

big factory in Binninghanl, was also attractecl to the idca of rotary nrachines.
Branrah, of coursc, is fanrous for othcr invcntions such as the hydraulic press
ancl the flush toilct.
He patentcd his hydraulic prcss in 1796. I{c patented a screw propeller in
1785, the modern flush toilct in 1778. FIe also manufactured unpickable

Murdoch's rotary steam engine

of 1799 borrowed its geometry
from the Pappenheim engine,
but had apex seals in the form
of wooden scrapers in the gear
tooth tips. Various viels of tlrc 1890 Brarnah rotary steanr engine.
176 APPENDIX
door locks and made many improvements to machine tools. In 1802 he
built a rotary wood-planer for Woolwich Arsenal. But his first patent for a
rotary cngine datcs back to 178 5 ancl included a machine very much like that
dcscribed in \Vatt's patcnt spccification of 1782. The hinged abutment was
to be closcd by weights. Another had a sliding block on the cylinder actuated I
by a rack and sector rvhccl. Brarnah thought of using rotary engines for
marine installations. His patcnt spccification of 1790 includes two different
versions of a rotary engine.
The first type consisted of a piston mounted on a drum, and able to
revolve inside a cylindrical casing. Two reciprocating abutments projected
from the walls of thc casing towards the drunr, operated from an external
guide ring mountcd orr thc shaft. One had to be closed while the other was
open to lct the rotor pass. The steam and exhaust ports were positioned on
opposite sides of the rotor, fed by passages in the shaft. The second design
consisted of a hollow drunr placed eccentrically inside a casing, having two
blades proiecting across it on dianreters at right angles, and fitting against
the walls of the containing cylinder. This was essentially nothing more than
an elaboration on the Ramelli pump principle. Nothing more was heard of
IJranrah's rotary nrachincs, but it is intcrcsting that hc was to play a sig-
nificant part in the dcvelopment of nrotorized highway travel in Great
Britain. In l82l |oseph Bramah had under construction in his Birmingham
factorl,a stcanr stage-coach designcd by fulius Griffiths of Brompton. In 1828
Brarnah built thc oft-dcpicted Chruch vchicle-a fancy-bodied steam-driven
road vchicle built to ply between London and Birmingham on a regular
schedule.

FLINT Two cross-sections of the Flint

rotary engine of I 80 5.
The abutnrents used to control steam flow in the engines we have studied
so far were nrounted on sirnple hinges. The idca of using rotary abutments
for rotary engines was not far off, however. And, of course, the rotary abut-
rnent is the anccstor of thc rotary valve. The honor of having been the 6rst
to suggest a rotary engirre rvith rotary abutments must go to Andrew Flint,
who rcceived a patent in 1805. In his machinc there was to be a 6xed outer
cylinder carrying two nearly semicircular abutments operated by an external
strikcr mechanisnr. Thesc abutnrents did not rotate continuously but were to
bc srvung asidc to allorv the rotor to pass. The single rotor was to have intake
and exhaust ports arranged on each side. The engine would therefore, had it
rvorked, havc had no clead ccntcrs.

POOLE

Moscs Poole patcntcd a rotary errgine in 1817, It had a cylindrical rotor

irrside a cylindrical casing. The rotor shaft providcd no eccentricity, but gas Cross-section of the Poole engine
sealing wils su1>1>osccl to bc pcrformed by hinged flaps carried on the rotor. of 1817.
 478 APPENDIX OTHER ROTARY ENGINES 179
These two flaps were positioned diametrically opposed. A control piston was
intcgral rvith the stationary casing. A small wheel mounted on the casing
MARRIOTT
forced down the abutmcnts as they approached, so that they would p6i
under the piston, and an cxternerl "striker" gear re-erected them after they
In 1828 Henry N{arriott and August Siebe, one an ironmonger and the
other a machinist, patented a rotary pump which might have proved reason-
had passed.
ably practical if it could hAvc bccn made lvith sufficient accuracy. Sealing
The Poole patent dcscription contained a great deal of thought on sealing
blrrclcs wcrc to sliclc in thc slots of it rlrunr-t),pc rotor, forccd outwards by a
1>roblcms. Poolc did not approve of the sealing systems then common in
stationarl,cfln) insiclc tlrc rotor to nrairrtain corrtnct with thc cylinder. To
steam systcms. Ife showed a clear understanding of the problems involved in
cnsurc proper scaling, a part of thc carn wils t«r bc mnde as a curved arm
prcvcnting wnstc nrtd cscal>c of stcartt bctwccn the sidcs of the rotor and the
flat cncl co\/crs of the hollow casing as wcll as betrveen the rotor edges and llivotccl at «»rc crrd rtnd frlrccrl orrlu'ltrcls lr)'n sprirtg. Thc stop wrls to carry
a sliding brass piece forccd outrverrcls irrto contact rvith the drum by springs
the circular chamber. Othcr steam engines used a packing of hemp or cotton,
and the ends of thc bladcs were to be packed in a similar manner. The
Iubricated with tallow. This forrn of sealing produced great friction losses
pump barrels were to be lined rvith copper or brass "by casting the iron over
particularly when comprcssion was high enough to make the fittings per-
the copper or brass."
fectly steam tight.
The stuffing Poolc specified was to consist of gaskets of plaited yarns in-
sertecl "in a groovc cttt in a serpentine or waving line, so near the outer edge
of thc rvlrccl as to adntit thc packing at certain intervals to lie close to the TROTTER
stcam stop and prct,cnt the rvastc of steam." This statement is proof that
even at this early clate, the main difficulty in the construction of a rotary Trotter's engine was patentcd in 1805 ancl hacl a cylindrical housing with
machine was recogn ized by capable engineers. It was an insurmountable a concentricallv mountcd drum. It was thc first rotary engine to use true
ccccntric rotation mther than canr-shapcd rotors to obtain volunre variations.
problem at the time, dtrcto the lack of precision in machining methods
and equipment. But it rvas the outcr rotor, not thc irruer one, that rvrrs cccentrically nrounted.
The inner rotor carriccl a bladc which projcctcd through an eccentric drum
revolving about its own axis. On admitting steam, the blade revolved,
causing the drum to turn about its own axis. In this engine, either shaft
WRIGHT
could be made to revolve at a steady unvarying rate, rvith the other turning
at a varying speed.
Poole's patent secms to have led to nothing as far as Poole's own activity
Steam could be applied simultaneously on both sides of the inner rotor,
is conccrnecl, but was probably thc irrspiration behind Wright's invention.
Lenrttel \Vcllman Wright patentcd a rotary engine in 1825 which appears
forcing the blade around rvith \/ery rrcarly even torque, the steam acting
always on the whole of the blade face, except that part actually passing
to be bascd on Poole's. The rotor carried flaps for sealing. Instead of using
a rvhcel to dcprcss thc flaDs, thc steanr stop itself was formed like a ramp through the slot in the eccentric drum. As late as 1882 an example of an
engine constructed to this desigrr was shown at an exhibition at the Agricul-
so as to force them gratlually clown. They were raised again by tail pieces
mountecl cxtcrnallv on their shafts striking against a wheel. But again-as tural Hall in London.
in Poole's dcscription-the specification is principally concerned with sealing
problcms.
Wright proposed to adapt for use in his rotary engine some of the same GALLOWAY
types of metallic packing rvhich were then coming into use in reciprocating
piston engincs. \Vright mcntioned packing the ends of the rotor with plaited
Trotter's ideas were taken up by Galloway 35 years later. In 1846 Eliiah
yarn in a scrpentinc slot. He also called attention to the packing of the
Galloway patented the first rotary engine rvith an epicycloidal rotor and an
flaps. The outer edge of each flap was to be made of a separate piece carried
outer envelope. The engine could be described as "paracyclic." The five-lobe
by nreans of slots on pins fixed to the main body of the flrp. This loose piece
rotor ran in a five-lobe housing, giving a I : I ratio of rotational movement.
was to be presscd outwards by steel strip spring, as the spring working
under the apex scal in a Wankcl ^cngine. Effective packing of the steam stop
It had a circular rotary housing u,ith a five-lobcd inner surface. The inner
rotor was a spider with five arnrs, each arrlr bcirrg captive within one lobe of
\1'as to be obtained by a picce of brass, spring-loaded to keep it continuously
the housing throughout the opcrational c),clc. The inncr rotor was eccen-
pressed against the pcriphery. Packing at the ends of the rotor was effected
trically mounted on a crank. As thc crank rcvolvcd, the center of the inner
by rings forccd against thc stationary part by helical springs.
rotor described a circle arouncl thc ccnter of thc outcr rotor. Correspondingly,
480 APPENDIX OTHER ROTARY ENGINES 48I
the end of erch arm travcled in a circle. It traveled along the curvature of into two paracyclic curves, one inside the other, with a common center. The
the lobe from about 120 degrces before top dead center to 120 degrees after outer one corresponds to the casing and the inner one to the rotor. All that
top dead center, then making a return, in circular motion, across the open is missing is the eccentricity. Move the rotor center away from the housing
gap from 60 degrees after bottom dcad center, rvithout being in contact with center in any direction until contact has been obtained. That fixes eccen-
the housing surface at all. No other guidance than that provided by the tricity at twice the short radius uscd for the initial sets of partial circles.
crank position existed. During this process the spaces in the lobes were caused Galloway made the rotor a loose fit on the crank and spring-loaded the
to expand or contract in a preciscly timed fashion. crankpin to assure permanent contact between rotor and casing at four points
The geometry of the Galloway engine is simple in principle. It is based on simultaneously. The spring is a circlip-type inserted in a slot on the crankpin.
a pentagon. If circles with a very short radius are drawn from each point of Four points maintain contact, because as soon as one arm leaves the working
the pcntagon, the ends of the five arms have been determined. Delete the surface, another has just made renewed contact. This fact ensured some
inner segment of about 120 degrecs. Next, determine the midpoints of the degree of sealing between the lobes, except the pair which communicated
Iines betwecn the points of the polygon and make them centers of five circles when the arm diametrically opposed to their dividing proiection was at top
of the same radius. On these circles, delete the outer 180 degrees, Now we dead center.
go back to the points of the pentagon, this time rvith a longer radius. This
Iong radius is equal to the half-distance betrveen the points less the radius Galloway's 1846 engine showed considerable advancement.
of the partial circles drawn before. Use this new radius to draw partial circles
around the earlier ones, using all the centers used earlier, points and mid-
points betwecn thcm. This has the effect of linking all the segments together
,
'^a /
/
16,;', I
a

,
I
tf
fv
(

/
I

I
\:'\

§*"{/Y
I

\r
§

.§
\r

This engine was Galloway's first

(
approach and dates from 1840.
182 APPENDTX OTI]ER I1OTARY ENGINIiS 183
The Galloway engine ran on steam admitted through circular side ports in step on the way towards the Wankel engine. It had cylinders and pistons, but
a cover plate which revolved with the rotor, into the several lobes. Each it was not a rotary in the samc sense as the radial engines once popular in
Iobe had its own intake and cxltaust ports. The exhaust ports branched out
aircraft. The four cylinders were positioned on two diameters at 90 degrees,
into spur-shaped recesses, curved to suit the motion of the arms. Due to its irr "+" fornration. The cylinders revolved on trunnions, and their rotation
geomctry, the Galloway engine produced torque over about 200 degrees of
produced a relative movcment bctwcen each cylindcr and its piston. Several
mainshaft rotation (40 degrccs per rotor lobc). Although some packing was units were built and tcstcd by Kitsou's of Lcccls. Thc Parsons engine was re-
provided around the ports, Galloway did not apply rotor seals to his engine. ported to have excellent balance, but it used up steam at the rate of 40
Steam consumption was therefore rather high and efficiency quite low. It pounds per horsepower-hour.
was used as a marine engine and developed about l6 h.p. at 400-480 r.p.m.

ROOTS
PARSONS The Roots pump would not make an efficient engine, chiefly because of
low compression potential and cooling problems, but its mechanism is pure
Sir Charles Parsons (1854-1931), father of the steam turbine, invented a rotary, and has been extensively uscd for air compressors. The Roots-type
rotary engine in 1882. It was not a true rotary engine, and was in no way a compressor is a sirnplified version of the Pappenheim pump, with dual two-
lobe rotors engaging to assure positive displacement. It was invented in 1860
by two Americans, Philander IJ. Roots and Francis M. Roots. Their father
va operated a .wool rnill, and for years their thoughts had mainly been occupied
with problems of belt gearing and other machinery.
The two interlocking rotors revolve in an oval-section casing. A gear train
keeps the rotors in phase with each other arrd nraintains the inter-rotor
clearance. Air is scooped in at the inlet opening on one side of the casing
and discharged into the outlet duct on the opposite side. The only known
rotary engine based on the Roots compressor was built in New York in 1867
-.?aa Cvlarra-
by an engineer named Behrcns. The rotors overlapped in the same way, but
nk,!.r5+- the lobes had a completely different shape.

sEcTloNAL ELEVATION cRoss sEcTtoN oN Xy

Elevation and cross-section of the Parsons rotary steam engine.

Cenhc of Cylrnilu
.\
.t'\'
ta t
Cloch frtnntsnl
Cttclc " l
z\
\\
)
t
a

a
a

Palh o/ Cranhpin
htcte ' 0' IMPE

(t
\
Operating principle of the Par- TOOTS ILOUEP
sons rotary engine. Roots-type blower, used for mine ventilation at Chilton Colliery.
 181 AppENDrx OTHER ROTARY ENGINES 18'
The most important theoretician on the subiect of rotary piston engines workable rotary engine: fohn Francis Cooley, Dimitri Sensaud de Lavaud,
was Franz Reuleaux, Professor at the Technical Institute of Berlin, and and Bernard Maillard. Each was on the track of the geonretry of the Wankel
author of many textbooks on machines that remain classics to this day. One engine. But they lacked Felix Wankel's persistence and singleness of purpose.
book was "The Constructor" published in 1893. Earlier he had written: They gave up when their engines were still in the concept stage or in the
"Kinematics of Machinery" in 1876. But he was never able to de6ne the early stages of development.
workable rotary piston engine,
For the rotary engine inventors of the 20th century, there was a wealth of
empirical knowledge to draw on, but no proper science, Do clear directions. COOLEY
Still, three inventors are to be singled out for their contributions towards a
)ohn F. Coolev from Allston, Suffolk County, Massachusetts, was a pupil
of Reuleaux, and had been engaged to work on rotary piston engines in
Berlin under his direction. After his return to the U.S.A. at the turn of the
century, Cooley continued his research and experimental work on rotary
engines, and formed a corporation to undertake this project: Cooley Epi-
cycloidal Engine Development Company, f ersey City, New ]ersey, and
Boston, Massachusetts.
His invention was covered by U.S. patent number 748,348 dated Decem-
ber 29, 1903. The Cooley rotary engine was a twoJobe inner epitrochoid
with a three-lobe outer envelope. Cooley was first to apply the inner
epitrochoid shape to the rotor, with an outer hypotrochoid to form the
working chamber. This is exactly the opposite relationship to the Wankel
engine's geometry. Cooley has one more lobe in the chamber than on the
rotor; Wankel has one more lobe on the rotor than in the chamber. He
Simplified cross section of the Roots-type pump, as applied to e patented his engine for use both as a pump and as a steam engine. He
modern supercharger. described the operational cycle as allowing steam to enter through a port in
a "spacer" into the pressure chamber, where it would force the rotor to turn.
The spacer was actually the outer rotor, corresponding to the outer rotor in
Wankel's original DKM-54.

Swedish engineer Fredrik Liungström invented this rotary steam engine in 1900.
It has four radial cylinders cast into one block, mounted ön , revolviäg shaft. The
engine is cran-kless, and each piston has a roller pin instead of a wrist pi;. The roller
Prns c?rry rollers which bear- against an elliptical guide in e statiönary housing. Simplified cross-section of the
Centrifgsal force kept the rollers in p_ermanent contäct with the guide. The engin:e
is stated to have run at speeds up to 1,840 r.p.m. Cooley engine.
 186 APPENDTX OTTIER ROTARY ENGINES 187
The inner rotor was eccentrically mounted, and this was the feature that center. Bearings for the eccentric shaft were provided by the end covers of
assured rotation. The spacer surface was in continuous contact with the the casirg, outside of thc phasing gears. Rotor movement and phasing were
periphery of the rotor. It was this permanent contact which assured the
dirccted by internirl gcaring. Both rotors rcvolvcd in thc sanrc dircction, but
volumetric variations and enabled the Cooley machine to operate as a steam at diffcrent specds. Thc rotor gcur was internreshed with the spaccr gear
engine. The itrtlcr rotor (piston) and outcr rotor (spacer) were mountecl on ancl asstrrccl thc clircctiorr of rotutiun.
axes parallel to each other inside a cylinder (casing). Cooley described the
The spacer had entry and exit ports for steam bctrveerl its points of contact
cylinder as having radial partitions. These partitions were bearings for the with the rotor. The combinecl rnotion of rotor ancl spacer opened and closed
cer.
sPa
the ports at the propcr tinre. Thc sealing systenr was primitive, and steam
The rotor was mounted on a shaft eccentric in relation to the cylinder consurnption high. Apex seals \\,crc carried in the spacer, not in the inner
rotor, as preferred by Wankel. Coolcy was forced to discontinue development
Simplified cross-section of the rvork on his engine. He ntade no attcurpt himself to convert it to operation
Umpleby conversion of the Coo-
ley steam engine, to operation on the four-stroke cycle as an intemirl combustion engine. But an English
on the four-stroke internal com- engineer named Umplcby did just that in 1908, only 6r,e years after Cooley
bustion cycle. had obtained his patcnt. Florvo,er, Urnplcby failcd to develop the engine.

CHARGE
't' OUI

The Larsen engine, from Den-

mark, is based on Ramelli's prin-
ciples, as applied to the four-
Swedish inventors Wallinder and stroke internal combustion cycle.
Skoog weie issued a patent in
192) for this true rotary engine,
incorporating a hypocycloidal
inner envelope, a five-lobe rotor,
and a 5:6 rotation ratio, assured
by central phasing gears. The
patent covered both two-stroke SENSAUD DE LAVAUD
and four-stroke operation.
Dinritri Sensaud clc Lavaud \r,as a Brazilian of Irrcnch origin, who first
gaincd pronrincncc as thc iuvcntor of a novcl nrcthod of casting pipe cen-
IR II{TAIG
trifugally, now used extensively throughout the world. He was a pioneer
aviator, having built his own airplane ancl floln it in Brazil in 1909. He
SIALING
ROIOR
arrived in Paris in 1920 u,ith thc idca of going into retirement in the most
civilized surroundings he could imagine. Irrstcad, his inventive mind turned
WORK to ideas for perfecting thc autornotive transmission system and he invented
ROIOR
a stcpless drive systcnr that rvoulg| autonratically sclect the right gear ratio for
thc car according to speed and load.
The Baylin engine, frorn Canada, He took out patents for independent flont rvheel suspension and limited-
is similar to Wankel's concentric SIip differentials back in the trventies. In thc thirtics he began to study power
BYPASS
COMBUSIION.
CHAMSIß ROI()R
engine of 193+, with its center plants and was athacted to the principles of the rotary engine. His work in
rotor and dual rotary valves. this area resulted in the issuc of Iircnch Patent 853 807 on December 16,
188
APPENDIX OTHER ROTARY ENGINES 489
1938. The engine consisted of a fivelobe outer
rotor with the shape of an Since the inner and outer rotors were kinematically interlinked, they re-
inner hypocycloid, and an inner rotor with six Iobes,
reduction ratio.
running oneSto6 volved at speeds inversely proportional to the number of their lobes. The
inner rotor had concave ferces, nlatching an outer rotor with a working surface
The outer rotor was cylindrical in shape and was
held inside e stationary made up of six convex segments. The relationship between the inner and
casing. The inner rotor was eccentrically mounted
inside the outer rotor. outer rotors provided a closed chambcr between cach two apices on the inner
.F=-i G.1 . rotor. The volume of each chanrber changcd during rotation. Minimum
volunre was obtained rvhcn two inner rotor apiccs coincided most closely
.l:i G.11 . with two lobcs of thc working surfucc. N{axinruur vohrnlc wAS obtained when
an inner rotor apex reachcd thc pcak of a convex scgnrent-or cxactly half-
way between two lobes of thc outcr rotor. It had two coaxial shafts. The
6E

outer rotor completely enveloped the inner rotor and was fixcd to the output
I

i
I

shaft on one side and running in a ball bearing on a support shaft on the
I
I
I
I

other side. This support shaft carricd thc inner rotor on an eccentric, ending
ä
i

before meeting the output shaft. Its ccccntric crrd was carried in an eccentric I
B
I

bearing. The outer rotor had a cylindrical shape on the outside. The inner I
§

rotor was split down the middle-as in two discs. They were fixed to each I
other by springs loading them laterally to provide side sealing. The gap
fi

between them had strip-type gas seals.

If you follow the volume change during rotor rotation, it will soon be seen
l6A that one set of rotors only constitute a compressor-not an engine. The four-
.FlG.1Z.\ stroke cycle cannot be completcd in one full revolution of the two rotors

W:\
.fi G.19. relative to each other. Starting from the point rvhen the intake port opens,
maximum compression will be reached when both rotors return to the same
'-f- relative positions. For the Sensaud de Lavaud compressor to be converted
into a four-stroke internal conrbustion engine, two sets of rotors are needed,.
nl
»l
r'
with one chamber pcrforming the duties of intake and compression, the
other going through the expansion and exhaust phases. For this to be pos-
1_ sible, they need a comrnunicating port rvhich admits cornpressed mixture
a\ into the combustion chanrber at the right timc and in the right volume to
maintain its state of compression up to the moment of ignition. r

Fresh mixture is admittcd inside the casing through an axial pipe. An

.FI G.1B annular chamber around the ontput shaft feeds gas inside the outer rotor of
the compression chambcr. The gas passages have seals to avoid gas leaks
betrveen the housirrg and outer rotor surfaccs close to output shaft bearing
give low rubbing speecl.s. The sealing problenr here is simple-because
-to
it is a depression area during the intake phase, pressure never goes higher
than atmosphcric. A side port in the outer rotor admits the conrpressed gas
to the rvorking chamber.
On the exhaust side, the outer rotor carries vanes to speed up scavenging
from the working chambcr into thc annular exhaust outlet. AIso, fresh air is
injected into the exhaust chanrber rvhich lowers the exhaust gas temperature
and adds to its volume. The outer rotor, according to the patent claims, is ;l
Illustrations from the D' Sensaud de Lavaud
patent of 193g. The four- therefore a partial gas turbine! As in thc Coolcy engirtc, thcre is permanent ttr

stroke cvcle is completed in ü; l.ä"tioni s

oa iir;'inner rotor, sliding contact betwccn thc tip of each lobc on thc intrer rotor and the rvork- FJ

1rY5 and
ing surface. Each lobc apcx wus proviclcd with a radial groovc for a seal strip,
,T

rotations of the outer rotor,

-[- in the example shown.
spring-loaded and under centrifugal load, against thc working surface. The
n
lii
ri
':.)
lla

it
t{
"t;
ti
Elevation of the Sensaud de Lavau«l engine. Below, details of the phasing gears
and gas seals are shown. OTHER ROTARY ENGINES 49r
seals are counterrveighted. The countcru,eights are subject to an opposite
ccntrifugal forcc to that acting on thc scals, thcrcbl, thel' Iinrit pressure on
.FiG.2. the working surface.
Gearin g of the Scnsaud de Lavaucl cngine is simple. One side of the outer
rotor carrics a ring gcar u,ith intcrnal tecth, coaxial rvith the rotor support
,13 83 shaft. It mcshes u,ith a smaller planet gear located inside it, mounted on the
1Z
83 94 eccentric part of the support shaft. It is therefore coaxial with the inner
rotor. The number of teeth on each must be full multiples of the number
of lobes on both rotors.
Ignition is performecl by trvin spark plugs rnounted in the end plates of the
outer rotor, directlv opposing cach other. Plugs are carried on the flanks of
the outer rotor. Both plugs havc contact rings for high tension current. The
ports are so positioned tha t rotor rnovement controls their opening and
closing. If the plugs are at 6 o'clock, thc ports are at l2 o'clock. Ports oppose
eachother-intake on one siclc, exhaust on the other.
De Lavaud speciEed thc possibilitv of diescl operation. Its operation on
An),cycle rvould have renrainccl hvpothctical if it hacl only bccn D. Sensaud
de Lavaud's private affair. But one test engine \r,as built bv the Ateliers de
Batignolles at the instigation of the Iirench Air I\4inistry, with the support
of Citroön and Renault. Ii[ou,ever, power output nevcr came up to the ex-
pccted figures, ancl thc project was abandonecl in 1941.
iot
MAILLARD

One American tutored in Germany; one Brazilian u,orking in Paris: what

next? One citizen of Switzerland, living peaccably in Gencva. Bernard
I\zfaillard had no exotic background in an!'wry, but he seems to have come
closer to anticipating the \Vankcl engine than anyone else. He was a
8? mechanical engineer, ernplol,ed bv Adolph Saurcr AG of Arbon, builders of
.§
high-grade trucks and clicsel cngincs of all sizcs. I-Ic inventcd a rotary
macliine and receivcd a British patent in 1943.
His design shows compressor activity in two phases per rotor revolution,
with intake ports at I and 7 o'clock, and exhaust ports at 5 and t I o'clock.
58 ls 62 ,t6 This did not allow much time or volume for compression, as the exhaust
port was opened only about five degrees after the intake port closed. As a
.FI G.17 result, the Ivlaillard design is not adaptable to torque-producing engincs.
Maillard's basic inventiorr consists of a static trvo-lobc housing containing
a three-lobe rotor, with chanrber volumes varying to provide the necessary
compression and expansion to operate on a four-shoke cycle. It was intended
{ot as an air compressor. Both sides of the minor axis had an intake port and an
exhaust port. As an air puurp, it has thc «lrau,brck thirt the intake period is
short and ntaxinrunr chanrbcr r'«rlunrc is rcachcd after the intake port has
been closed. \Vhatever provision for sealing Irlaillard envisaged is.not
dcscribcd in the patcnt. If N'Iaillard's invcntion looks like a poor compressor
 !

192 APPENDIX oTHER ROTARY ENGINES 19)

and an unworkable enginc, rvhat is thc value of his patent? Quite simply All the rotary cngincs we havc revicrvccl thus far have had an essential
this: I Ic showed the 6rst combination of an inner hypotrochoid rotor, dcfcct when erialuatccl as an intcrnal contbustion cnginc bccause prirnary
ccccntrically, nrounted in an outcr hypocycloid working chamber. Not being attention was paid only to modification of a mechanisnr, or there was no
acquainte«l rvith Irelix Wankel and his post-war work, Maillard let his patent advantage ou.i tJre reciprocating engine. Another reason why all research
lapse in 1948. work madc in the distant past on various types of rotary engines ended in ..

failure was the lorv level oi machining technology and quality of materials
' rvhich form the background for thc development of the engine.
'1,

Ceomet ry ofthe Maillard patent .nr',n..'

:

Cross-section ofthe Ma its adaptability

to both two'lobe and th
494 APPENDTX OTHER ROTARY ENGINES 195
The Wankcl engine was the first rotary piston engine on
which systematic At the concept level, the Wankel engine has many rivals. Strange engines
studics were carricd out in full scale with iespect to
the cooling system, the
lubricating nlethods, the combustion process, rvith rotating pistons, curved pistons, virne-type pistons, orbiting pistons, and
lr_ort timing, position of spark
plugs, effect of ntachining accuracy and gr, ,.rling. scissor-action pistons. Rotary engines conre in three main groups:
It wäs unfortunate that l. Eccentric-rotor types, such as the Wankel, f ernaes, and Renault.
no widc-basc rescarch such as this w
c,ginc. Thc arri'al, clcvelopnrentand :ldJil,ot|il 2. Scissor-action typcs rvith pistons or vanes, such as the Tschudi, Kauertz
and Virmel.
dctcrrcd other invcntors fronr llursui ;r;;;;
Every year new ones come up, ,nd' ). Revolving-block types, such as the Mercer, Selwood and Porsche.
commoll witlr invcntions of the distant past, none of ;j|,ffl ;:?i:
thcse have had the
llcncfit of a concLrtecl rcscarch, tcst ancl development program
such as
brought the Wankel engine to the forefront in record time. IERNAES

c Irinn f ernacs, a Norrvcginn invcntor livirrg in Kristiansand near the

southern tip of Norway, has takcn the thrce-lobe rotor and trochoidal en-
velope from Wankel's clesign but has conre up with a novel type of gearing
for his "Planet Motor" patentccl in the U.S.A. in 1965. The fernaes engine
is built up around a central ourtput shaft fitted rvith an integral circular plate
which works as a rotor hub. Unlike thc rotor in the Wankel engine, the
|ernaes enginb rotor has no internal gearing.
The hub supports three planet gears running on eccentric shafts. These
planet gears mcsh with a statiorrary reaction gear mounted concentrically
with the output shaft and thc hub. Each planet gear has the same eccen-
tricity and they are spaced 120 degrees apart. As the rotor turns, it carries the
planet gcars along. They are phased to let the output shaft turn at rotor
speed (instead of thrcc times rotor spccd as in the Wankel). Finn fernaes

REACTtOil CEAR ROTOR PIANET GEAR

SPARK PIUG

OUTPUT
EHAFT €NO COV€RS

Eievation ancl cross-section of a proposed prototype Maillard One possible version of the |ernaes engine with twin rotors. (Drdw-
compressor.
ing: Roy Pioch)
 OTHER ROTARY ENGINES 497

claims that his gear mechanism makes for a big increase in torque
at rela-
tively low ,.p.*l In other words, a four-stroke single-rotor |ernaes engine
produccs the same nunrbcr of powcr impulses pcr output-shaft revolution as
är six-cylinder piston cngine. 'itr. rotoi side-sealirrg is ingenious: the end
co'ers revol,u,e rvith thc ätors ancl are scalccl with trvo sinlple rings
-on
each

side. Intake and exhaust ports in the casing are opened and closed by the
overs.
ariations on thc engine described' A
D1'namics, Lycoming Divi'
rr of North American Rock'
Cross-section of one proposed
version of the fernaes engine, al progrcss has becn tnade.
using a three-lobe chamber with
an inner epitrochoid. I = Sta-
tionarycasing.2 = Working
chamber. 3 = Working cham-
ber. I - Working chamber. 5 = The |ernaes engine- - discs are
Rotor. 6-Stub shafts. 7- provid-ed with suitablY shaPed
Planet wheel. 8 = Disc. 9 = openings, which uncover gas Pas'
Sun wheel. ll=Seal strips. sages leading to the wolkin-g

N
-feeds
I 3 = Drive shaft center. chambers. Intake slot 2l '
fresh mixture through the 8äs,
passage 2), as long as the oP-en'
ings öoincide. The same aPPlies
to the exhaust slot 22, with
respcct to gas evacuation, as the
rotor and disc turn together.

Side section of one proposed

form of fernaes engine, showing
one rotor and its chamber. Code
numbers as in Fig. 28 with the
following additions: 17. = Shafts
ca rried on discs I 4. I 3 = Drive
I [],
;='J-
I

shaft. l1 = Cover discs. 15 = t 'r|rv'

Sealing rings. 16 : End cover
@ shields , 17 = Mainshaft bear- \
ings. l8 = Coolant passages.
19 = Intake manifold. 20 =
@ Exhaust r,tanifold. 2l = Port.
23 - Gas passage.

@ o
498 APPENDIX OTHER ROTARY ENGINES 499
lobe'
RENAULT chrrmber, with a separratc intakc and exhaust vah'c for each chamber
piston seal tips are nrounted
Another notable fcature lies in the fact that the
The rescarch dcpartmcnt of the Regie Nationale cles Usines Renault, in thc housing
"'b !'---
ancl not in the rotating piston, .i in the cooley engine' The
sngin. with Iow compression arrd a com'
France's govcrnmcnt-owned automobile company and the largest manu- patcnt s5o
facturer of passcnger cars in the country, began to study the Wankel engine ch
bustion e'
Renault Inoney on its nd
in 1958, but ccntcrcd its orvn dcsigns ancl expcrirnents on the Cooley patents.
intcrestecl in its lclrtptrrtion of
Rcnault startccl off with a rotary errginc with poppct valves instead of ports
Y bY RoY Abern
of
as in the Wankcl enginc. A four-lobe rotor ran in a fivelobe trochoidal American
Anrcricatt Motors Corporation, ancl Pir it f)rcyftts'
Presi ., ., . .r gie

des usines Re'ault, on Scpte,rbcr 19, 1962, snicl that thc trvo
Nationale
nö technical resources and costs in the
lex ancl more efficient powerplants'
of lrtttottrotivc rcscarclt and crrgi-
thc rotary engine then under con'
,'C'i. Further variations on the fernaes
ssential features and also drew a com-

,/ '\o/' parison with the Wankel engine'

t, '\r /'
en$ne.
of gas
Whcn thc rotor in thc llcnltrlt-Ranrbler cngirrc turns as a result the
pressure ofl'one or morc of its fatccs, it is .r,rii.,l
around its orbit bY

\ öl '{
\-\
- -.t'

The gear mechanism of the

f an internally
ernaes engine has
toothed stationary sun wheel,
with a set of planet gears mesh-
ine with it in a IzZ ratio. 5 -
Rotors. 6 - Eccentric pins. 7 -
Planet wheels. 8-Mainshaft Elevation of the geat system shown on Fig' 72-' ^14 = End
-Mrinshaft
piece'
flange. 9 - Sun wheel. l0 : 16 - End t o,rs,nä . L7 - bearing' . 18 - Water iacket'
Casing. 12 : Stub shafts. 17 : 3 Flywheel flange with starter motor geäring' 9 =
Sun wheel'
Mainshaf t. -
5 - Rotor,
5oo APPENDTX OTHER ROTARY ENGINES 5OI
eccentric shaft. A pair of gcars assurcs that the rotor rvill turn o1 the eccentric
in sttch a manncr that clcarancc is always nraintained betrveen the rotor and
has been proviclcd u,ith rcmclval>lc heacls to pcrnrit a thorough, long-range
research program with a r,,ariety of combustion chambcrs. While we do not
the stationary housirrg. The sanle sct of internal and external gears coordinates
cxpcct to offcr thi.s cnqirrc in a passcugcr car in thc ncar future, we feel that
thc rotor rotation ancl crartk rotittion to kccp thcm in phasc. Five radial scals
it has thcorcticll 1:ossibilitics, lncl \\'c thotrqht 1'ou rvoulcl fincl it of intcrest.
contact thc rotor's outcr colttour rnd scparatc thc individual combustion
In tltc corttinuing cffort to incrcrrsc thc totll cfficicncy of the passenger car,
chambcrs. Set'eral types of seal can be used to contact thc sides of the rotor
wc fccl that it is csscntial to cx1>lorc in clcpth clcvclopnrents such as this,
and conrplctc thc job «rf scaling off the chambers.
w'lticlt offcr thc pronrisc of basic gains in sinrplicity, rvc:ight, cost, reliability
In the \Vankcl cnginc, tltc gases arc carried around inside the outer hous- !i:

and conrfort."
ing, rvithirr the ntoving chanrl>er forrned by tlre rotor as it turns. Fuel-air
rttixture is takcn in on one sicle and conrpressed, fired at the top, and expanded
and cxhausted oll the other sidc. In contrast, the Renault-Rambler rotary
cngittc has statiorrart' conlbustion charnbers. In the Wankel design, the apex TSCHUDI
scals nlove u'ith the rotor ancl are oil cooled, while in the Renault-Rambler
ttttit, thc apcx scills are stationary,in the housirrg and therefore water-cooled. The same year that F-elix \Ä/ankel was born, another boy was born a few
Thc \Vankcl ttscs ports, rvhilc thc Rcnault-Rambler type uses conventional miles to the south, on the Swiss sidc of Lakc Constance, not far from Basel.
pt-tp1>ct valvEs, giving ntorc prccise tinring and closer control of overlrp.But Hc was the son of a blacksmith, and his name was Traugott Tschudi. Ile
thc cngirrc is soon to be al>andoned by both Renault and American Motors, went to America in thc Twcnties, settling in New York, and began design
arrd ncithcr colllparry is currcntll'engaged in rescarch or dcyelopment work on work on a toroidal engine about 1925.
anv forrtl of rotary cnginc. Ralph Isl>rarrdt's closing remarks seern particularly Thc'fschudi cnginc has four curvccl 1>istons nrr)ning a toroidal track. Each .t
I

rclevant: "This research engine was dcsigncd for convenience of experimental pair of pistons is dianrctrically opposcd and carriccl on a scparatc rotor. Iiach
rvork, with no provision for installation in a car, and no particular concessions rotor also carrics trvo rollers. 'l'hcsc rollcrs bcar irgainst trvo idcrrtical canrs
nradc to colllp:tctncss or cost. Since this type of rotary engine allorvs con- fixed to the outl>ut shaft. Thc rollcrs rotatc only bccausc of friction against I

siclerablc frccclont of contbustiorr charnber shape, the experimental engine the carns. Power flow is achieved by having onc roller push into a cam groove
rvhile the other assures positivc canl location. The caur disengages from the
rollcrs to allow onc rotor to stop. When two pistons stop and restart, the
engine obtains a chunge in the gas volunrc bctwccn the pistons. Piston travel,
controlled by cam an«l rollcr action, thercforc directs irrtake, compression,
combustion, and exhaust phasing. \Vhen one sct of pistons stops, the other
set continues to turn the output shaft. The shaft is ecccntric in relation to
the toroids and the rotors, and it makes 1.2 ro'olutions for each 360 degrees
of piston travel.
The type of motion given to the pistons is a hesitating progression. The
compression is followed by the trvo pistons containiug compressed air be-
tween them and they both nlove togcthcr to the spark plug region where
combustion occurs. At this point movcrncnt of the lagging piston momen-
tarily ceases whilc the leading piston procccds at full speed. Eventually, the
lcading piston rcachcs thc cxhaust lrort rvhcrc it conles to a halt while the I

lagging piston follorvs it, rapidly conrplcting thc scavenge. Induction of a

new chafge is carried out sinrilarly.
The piston nlovenrent in the Tschudi cngine is an inversion development
of the "Geneva Whcel" nrotion uscd in nrotion picture cilmeras, but it has
a reduction ratio of only 2:l instcad of the rnorc usual 4:l or 6:1. The drive
shaft is mounted cccentrically rclative to the center of movement to the
rollers rvhich then rnerely ictle round about tlieir own axis. For the rest of the
time the cam causes the pistons to rotate at a varying speed, depending üpon
cam profile. For every revolution of the shaft, one-half a revolution is im-
cross-section of the Renault-Rambler rotary engine. parted to the pistons.
 . - cYLINOEi CRANX Er{GrNE, FOUß
cYcLE, 4'EORE, ?'STROXE

ROTAF' ENGINE
F,OUR CYCLE
4',CORE. 7. STROI(E

FIG. 9
COMPARATIVE SIZE OF CTANX TYPE 3ROTARY
tOTH ENGINES HAVE Eotrrl'T'IDN OIS'I,(EI€'T
EACH ?rcOUCJS T\r!O EX'LOS|O|6 PCI tarourTtolt
GE^R FOR
M^cl€ro f
OL PtJT'P
'VALVES I VALVE.
OPERAIIONS
s|,|AFT .eßili{./

ROrOR EEArlxc WALL FRICTION OF PISTONS CAUsEs VEAI.

zowci Loss ANo LlMlTS IHE CI{OICE Ot
CONSTruCTI(»I MATERIALS

ONE.TOROIDAL CYLINDER ENGINE . FOUR CTII.E

PTOOI,CES TWO EIPLOSIONS PER SHAFT ßEVOLUTION
FIG. C CUBIC DISPLACEMENT Drt?.44
OPERATING DIAGRAMS
STATOR S POWER OI^6RAM OF
(auMrNrr{ euov) FlG. lO CRANX TYPE ENONE
SI{OYING SOME INHERENT
FAULYS Of CßANI( ENGINES FO|/ER g116,lAr of
THAT ARE EL]MINAIEO SITX ROTAßY ENGINE
rHC RST^rl TIPE
ROYOR BAII
BEARIM6
PACXING TII{GS XAYE
PnoTRUS|O|.5 AT Pr STOllS
TO INSUßC COMPLETE
p^^rslox
-- SEAL
sttr - rr
PACXIN6 RING
FlG.ll
(eturrnur euov )
SHOVING ADVA^'T^GE OF HIGH
PISTON SPEEO
P^CXIN6 iING WITH
@RR SPIDSC BAO(li,a'O

INVE NTOR : TRAUGOTT TSCHUDI

56- 16 I99TH ST , FLUSH ING 6' N Y
FIG. 3 STEEL RING il I x.!. I c..rr t' T TSClruOl
ROTOR ASSEMBLY coNN€cTeo{lTH U.SPATENT I lffi-
q .'..
DETAIL OF ROTOR ROTARY ENGINE-ASSEMBLY
SHOWI}JG SI<IION AT ROl I FR

i:äe-s81""'*-" 2 O O
Sections and details of the Tschudi engine' Sections and details of the Tschudi engine.

The Tschudi engine works on the four-stroke principle. Each toroid pro-
duces only two powcr impulses per outpnt-shaft revolution. A Tschudi must stacking a series of strrtors and rotor assenrblics around the conrmon output
have two toroids to become a practical torque-producing power plant. For- shaft. A double unit rvould be cquivalent to iln cight-cylinclcr rcciprocating
tunately, thc single-toroid unit lcrrds itself to power multiplication simply by piston engine. A 90 dcgrcc offsct of thc cams in contact with the respective
rotors woulcl cnsurc unintcrrulltccl trlrcluc ottty:ttt.
504 APPENDIX OTI]ER ROTARY IiNGINES 505
The Tschudi rotary enginc can be designed for either spark-ignition or cylincler is exposccl to thc cxplosion. 'l'hercforc, pistons arrd rotor f:rces will
compression-ignition (Dicscl) opcration. The principal advantage over the cool rvhile passing through thc othcr zonc. Pistons rvill be splash-coolcd with
rcciprocating piston arrd crank engine is its nrechanical simplicity. It scores oil contained insiclc thc rotors.
ovcr other rotary engines by thc fact that it is pressure sealed with conven- Onc of thc rvcntkcst fcaturcs of thc Tschudi cngine is the high stress level
tional piston rings. The re is also provision for piston cooling. These are at the point of contact of the cam and rollers, particularly at the tinre rvhen
important points often ncglectcd on rotary engines. the piston is forccd in a countcr-clockrvisc dircction by combustion pressure.
Tschudi chose the toroidal shape because it simplified pressure sealing. The rapid accelcration and stopping of the pistons and rotors relative to the
Thc piston rirrg.s are fixcd with a pin to prcvcnt their rotation on the pistons. canl is al.so botrnd to rcsult in shock loacling of thc cam surface. Experts who
Thc tangential rirrgs betwccn stator and rotors are the self-expanding type. Itave exanrincd it cxprcss d«rrrllts allout thc lullricltiorr, frictiou and wcar
Since their motion is radial, they are provided with a tongue and groove c\aractcristics of thc packing rings bctrvecn the rotors and stator, as well as
surface; a feature rvhich is uscd in the sealing of high pressure steam turbines. the use of alurninum for cylindcr nurtcrial. It is not considered likely that the
Tschudi's first rotary engine was designed in 1927. It was a scissor-action pistons rvill be frictionlcss against thc stator. Thc conrbustion pressures acting
type, using a ratchct device for power transmission from the pistons to the to deflect the rotors and seals may cilr.rse considerable friction.
output shaft. That proved to be a fragile arrangement and by l%5 it had Because the pistons are alternately moving and coming to rest, according to
been entirely redesigned, with a gear drive system. Then Tschudi becarhe the geometry of the carn, high inertia forccs rvill result. The large ball bear-
aware of modern cam roller drives and designed his own, which he patented ing which supports the piston carricr rotates at a mean speed that is half that
in 1953. The cam-roller drive is a positive contact mechanism; both rollers of the output-shaft. The axial load caused by high gas prcssures within the
of the rotors are in continuous contact with the cam faces. Aircraft engines toroid is eccentric and intermittent. The rotational speed is also intermit-
with cam-roller drives werc in use during the Irirst World War. However, tent. It is thcreforc, thought that the difficult operating conditions of the
thc cams of these cngincs were of the figure 8 shape, with the explosion bearing may impose a limitation on the engine.
taking place at the deacl point, a disadvantage fully avoided in the present Although rings are provided to stop gas lcaking past the pistons, the
invention. Oval shapcd cam-roller drives are now used on the latest European toroidarl passagc cAnnot bc pcrfcctly intact duc to its construction from three
diesel engines as a means to climinate the cranksltaft. separate parts. Three snrall fissures run dos,n the length of the toroid past the
The cam-roller drive system \r,as designed for smooth operation and a piston rings and they cannot be completely closed. This is a problem as-
minimum of vibration. This type of canr-rollcr drive was first used on a three- sociated with all engines of this t),pe and a full investigation into experience
cylinder two-stroke reciprocating piston diesel engine built in l9r4 by already accumulated u,ould show rvhether it is a serious matter or not.
Svanemölle Verft in Copenhagen, Denmark. It was completely crankless. A similar sealing problem exists on rotary heat exchangers and is, in fact,
Tschudi clairned that the pistons wcre friction-free in his engine and would the main obstacle to successful development of these urrits. In view of the
never wear. They arc rigidly nrounted on rotors and have clearance in their considerable developrnent u,ork carried out on this type of sealing problem it
toroidal chambers. This pcrnrits grcater choice in the selection of materials. is thought that this feature may bc the maior snag rvith this type of engine
Pistons can be die-cast and nccd no surface rnachining. The output shaft is combustion.
located eccentrically to the engine and contains two cams. Each cam contains
a segmcntal curve serving as the base circle, interrupted by one ovate curve
and by one concave curve. Both cams are mathematically identical but are
KAUERTZ
offset 180 degrees on thc driveshaf t, and axially spaced so as to engage in the
respective rotors.
f'hc Kauertz engine rcsenrblcs the Tschudi in that it is another scissor-
The Tschudi enginc has a continuously cool inlet region, ensuring adequate
action powcr unit. It is thc inveqtion of Eugen Kerucrtz, now Iiving in retire-
induction of air. It has low pressure Ioss through elimination of torturous
nrent in Hü6ngen, Germany, but still active as an inventor. The prototypc
valve passages. Prc-ignition, through hot spots that can occur in a conven-
engine has onl1, 22 parts. A tcst engine of 7 .5 cubic inch displacement has
tional cylinder hcad, can probably be climinated. The inventor claims more produced 5l horscl>o\\'cr at 4,000 r.p.nl. according to thc invcntor's claims. A
efficient operation due to a fastcr piston speed, but other experts consider larger 61.7 cubic inch unit has dcli','ercd 217 horsepower in tests. The
this doubtful. The Tschudi protot)'pe engine is designed with a iacket for
Kauertz crrginc has thc sarnc nnnrber of firiug impulses pcr shaft revolution
water-coolirrg but thc toroidal cylinder may be provided instead with fins for
as a V8 reciprocating piston cngirre, and runs with comparable smoothness.
air-cooling. Several factors indicate that the rotary engine will generate less
J'he Kauertz engine has vane-type pistons iil a circular-scction working
waste heat than a crank engine: rapid expanding pistons will transform the
chanrber. Two sets of vancs rotatc on thc selrne axis but they continuously
caloric cnergy of the fuel into additional power; only one-sixth of the toroidal
changc position relative to cach othcr.
506 APPENDIX OTHER ROTARY ENGINES 507

This specding up and slowing down of one set of vanes changes the volume VIRMEL
of gas bctwccn the tn,o scts. Changes in gas volume produce the pumping
action nccdccl for intake, compression, conrbustion, and exhaust. Four effec- The Virmel engine is sinrilar to the Kauertz because the vane-type move-
tive conrbustion chanrbcrs give four power phases per output-shaft revolution. ment is the samc, but thc powcr drive is cntirclv different. The Virmel is the
C«lrrcct phasing of thc vane motion is assured by a gear-and-crank system. invention of I\,{clvin Itolfsnrcycr, an crtgitrccr living in Lincoln, Nebraska,
T'hc output shaft carries a coaxial but stationary sun ger and a planet-gear rvho nanred it for his rvife Virginia arrd himself (Vir(ginia)Mel(vin)). The
carricr that rcvolves rvith the shaft. A crank fixed to the planet-ger axis prototype is undergoing testing and devclopment at the Lynx Corporation in
dcscrillc.s trvo conrplctc rcvolutions for one turn of the planet cArrier. The Lincoln, for use in cars ancl boats.
planct-gcar crank is linkcd to a countcrbalanced arrn that both rotates and Displacement is only 50 cubic inches, and claimed power output exceeds
oscillatcs around the output shaft axis. Both movements of this arm are 300 horsepower at 3,800 r.p.m. The Virmel engine has two sets of vane-type
produccd by the crank and its linkage. The arm is rigidly fixed to the second- pistons antl a gear-and-crank svstcm that controls piston phasing. It differs
ary set of vancs. Thcse vanes alternately catch up rvith and fall behind the from the Kauertz in several respccts, but works on a similar cyclc. It produces
prinrary vancs, rvhich arc 6xed to the output shaft and therefore drive the four power impulses for each output shaft revolution. The two vane sets are
fixed to two concentric shafts, and there is a stationary sun gear. But in the
Virmel engine, the sun gear has two satellites, each one with a crank linked
to a lever splined to one of the two main shafts. Rotation of the satellite
COlvtgUSfloAl
CONT/NUES
gears and cranks directs piston acceleratiorr and retardation. The satellites are
always diametrically opposed on either side of the sun gear. Satellite rota-
tional speed (and output-shaft speed) is steady, while one set of vanes is
monrentarily brought to a conrpletc stop twice during each cycle.
In the Kauertz engine, the primary vanes run at steady speed and second-
ary vane speed varies. In the Virmel engine both sets stop and restart.
Unlike the Kauertz, the Virmcl engine has no through shaft. The two con-
centric mainshafts do not continue beyond the sun gear and crank system.
Power flow is taken via the satellite gears and crankshafts revolving at
satellite speed to another planetary gear train. The satellites in both planetary
INTAIG EXHAUS T
gear trains simultaneously orbit and rotate. The final gear set reverses rotation
in a I : I ratio to turn the output shaft in the same direction as the twisting
Operational cycle of the Kauertz engine. All piston movement is unidirectional, but force applied to the engine mass. This practically eliminates torque reaction
it is interrupted in order to produce a scissor-action (with the scissors revolving
and simplifies installation.
around the pivot point).

OMEGA

planet carrier. Both faces of all four vanes rvork full-time, going through all Another scissor-action engine is the Omega, invented by Crant'tlle Brad-
phases of the four-stroke cycle. Vane sealing is obtained by free-riding blades sharv in l955..Granville Bradsharv designcd Pratt & Whitney's first radial air-
hcld against thc housing by ccntrifugal force. The same blades also offer craft engine during \\/orld \\/ar I, and dcsigncd the Belsize-Bradsharv car of
somc scaling against thc end covers. The Kauertz engine has no unbalanced 1921, using a light V-trvin oil-cooled engine. Latcr he joined BSA (Birming.
incrtia forccs, and very high speeds are possible because there is no valve ham Small Anns Conrpany, Ltd.) and designed several flat-trvin, air-cooled
gcar. Thc vancs thcmselvcs open and close the ports with accurate timing. motorcycle engines. In the early 19 50s hc was engaged in the design of
Thc nrain drawback in the Kauertz engine seems to be the high inertia motorscooter engines.
Ioacls gcncratcd on thc phasing vanes and transmitted to the frail-looking The Omega enginc sharcs the toroidal chamber and the four pistons with
linkagc. Docs it have a future? It may be developed under a contract with the Tschudi design. The pistons Are carried on rockcr arms and reciprocate
the South African governrncnt. And a Canadian firm, Hi-Pow'r Roto, which in pairs. The rocker arms are nlounted on scparate but concentric shafts, each
holds Kaucrtz rights for Canada, |apan and Australia, is reported to be build- fitted with a crankpirr that carries a connecting rod. The connecting rods turn
ing a factory in Victoria, B.C.
508 APPENDIX
a crankshaft locatcd inrmediately below. The toroidal engine block revolves
around thc oscillating pistons, and a 2:l gear set positioned behind the
crankcase assures one rcvolution of thc block for every two crankshaft revolu-
tions. The pistons are double-faced and float between gas pressure phases on
oppositc sidcs. Every swing of each piston, in either direction, is a power
stroke. Every power stroke is cushioned by a compression stroke at the other
cnd. This drive systcm elinrinates inertia losses caused by crankshaft reversals
and sharply rcduces bearing loads. Friction losses are claimed to be small,
since there is rro side thrust on the pistons. Operating on the two-stroke cycle,
the Omega engine has only one intake and one exhaust port serving the
combustion chanrbers. This is possible because the block itself rotates,
Sinrilarly, thcre is only one spark plug, mounted on the periphery of the
toroidal chamber.
The Onrega engine was handcd o\/er to Weslake & Co. of Rye, Sussex, for
development, in 1966. The Wcslake people rvere unable to start it and had to
overhaul the engine beforc it was put on the test bench. It never ran under
its own power, though occasional firings would occur. No actual development
was ever accomplished.

Cutaway drawing of Granville

Bradshaw's Omega engine.

MERCER

The Mcrccr engine begins the third category of rotary engines which have
reciprocating 1>istons and rotary blocks. The first engine of this type was built
in 1898 by I?. O. Iiarrvell. His design was used in the 1905 Adams-Farwell
autonrobile. Its principles reappeared in the French Gnone-Rhone aircraft
engine during \\rorld War I, and are now used, with variations, for the
experimental Mercer, Selwood, and Porsche rotary engines.
The lt{ercer engine is a rotary radial with four spoke-like cylinders spinning
inside a stationary rirn. It was the creation of Austin Mercer, an engineer and
inventor, rvho opcrates a ten-man machine shop in Bradford, Yorkshire. It
was 6rst tcstcd in 1963. Opcrating on thc trvo-strokc cycle, thc Mcrcer enginc
was designcd for high torque at low revs. Unlike other rotary engines, it has
a wide spread of power throughout thc speed range. Operational cycle of tlre Orncga cngine.
5IO APPENDIX OTHER ROTARY ENGINES 5II
The Mercer engine consists of two opposed pistons in a single cylinder. the piston crown. The large-diamctcr part of the pistons travels through an
Each piston pin works as the axis for two rollers. The rollers run on a hack intakc and partial conrprcssion chambcr connected to the cotnbustion
that looks trochoidal but is rnerely thc peripheral track of two intersecting chamber by a conventional transfcr port. In this rcspect it bears a striking
circles. Combustion in the center forces the pistons and rollers apart. The resemblance to the Pescirra typc of free-piston cnginc. As in the earlier Mercer
rollcrs can move outward only if the cylinder block turns so the output shaft engine each piston carries rollers that run along a stationary track consisting
revolves rvith the block. The only reciprocating parts are the pistons, and of two circular arcs. As the pistons are drivcn outrvard by the combustion
since they movc in and out in unison, the engine is inherently balanced. The pressure from thc centrally located spark plug, the rollers are pressed against
output shaft does not continue on the other side of the block. A hollow shaft the track, thereby forcing thc block to rotate. The outrvard motion of the
at the othcr cnd carries thc spark plug, feeds a gas mixture from the car- pistons also cornprcsscs a frcsh air-fuel charge that has entered behind the
burctor, and delivcrs lubricating oil. The exhaust ports are in the output pistons. \Vhcn the pistons are farthcst apart, the exltaust ports are uncovered
shaft. The pistons open and close the ports in the manner of a normal two- and vent the burned gases.
stroke piston engine. Since there is no crankcase, the fresh mixture cannot be This second-generation Mercer engine design is quite versatile because it
fed in lry crankcase comprcssion as in conventional two-stroke engines. In- can be applied to multi-cylindcr or bankcd arrangcments. This piston design
stead a blowcr is installed between the carburetor and the induction pipe. is intended to give inrproved charge sl,nchronization at all engine speeds.
This gives the advantage of forcing the gases into the cylinders instead of Detail dcsign for porting and cooling can be varied according to application.
relying on their haphazard passage through transfer ports uncovered and A sepirrate system of prcssure lubrication to all moving parts was considered
sealed by sliding pistons. The supercharging arrangement eliminates the necessary.
common two-stroke conrplaints: poor scavenging and disposal of exhaust However, thc latcst Nlcrcer engine shows no advantage for operation un-
gascs, resulting in Iow efficicncy and high fuel consumption. der part Joad (typical automobile cnginc conditions). Purging prcssure falls
Intake timirrg is variable to give peak performance at all speeds. The central off r'ery quickly, in the cerse of exhaust gas scavenging as rvell as feeding the
induction tubc can be trvisted axially to alter the point at which fuel is blown fresh charge under comprcssion to the combustion chamber. The high pres-
into the cylinders. The angular relatiönship between'the inlet valve and the sure drop is not conducive to high cfficicncy. Thc tinring of gas intake and
conventional cylinder-wall exhaust port is changed in accordance with speed exhaust is poorly controllcd by piston rnovement. There is a risk of 20 to
and load. Scavenging of the Mercer engine is facilitated by the transfer of a 50% loss of fresh charge to the exhaust system, rvhich would make the
fresh charge from the area behind the pistons where it has been compressed. emission levels unacceptable.
The pistons then move togcther to compress the new mixture prior to com-
bustion. J'he frcsh chargc is admitted to the combustion chamber via a
transfer channel similar to two-stroke piston engine practice. Some of the SELWOOD
frcsh charge is lost to the cxhaust during the transfer process. Stresses on
roller assenrbly and cylindcr walls are quite high. Each cylinder goes through William R. Selwood of Southampton, England, decided to sponsor de-
two conrplcte cycles for each revolution. Four pairs of explosions or eight velopment of a rotary engine invcntecl by Cecil Flughes in 1952. Progress
power strokcs accompany one turn of the shaft. At a given speed it makes the was slow, and now Selvi'ood has sold the engine to a new company, Orbital
cngine cqual to a l6-cylinder, four-stroke engine of four times the cubic Engineering, in Bristol. IJughcs is still rvorkirrg on its developtnent. The
caprcity. Sclwood engirrc is rotary in tlrc scnsc that the cl,linder block ro'olves, driving
The experimental model had a displacement of only 50 cc. (3.0 cubic an output shaft, u,hile the nrnrinshaft rvhich supports the spider and the
inches) with a lYa inch bore and Y* inch stroke, and ran with a 6:l com- pistons, is stationary. Thc spidcr is fixed at a I 5 clegree angle. The Selwood
prcssion ratio. Mercer claimecl a peak output of l0 horsepower at 5,000 orbital engine has toroidal char:rbers arrd turns reciprocating motion into
r.p.m. Thc little prototype was ten inches in diameter and weighed 56 pounds rotary nrotion by a kinenratic invcrsion of the swash-plate system (now
including blower and generator. commonll' used in h1'draulic puntps).
In 1965 Mercer built a7 50 cc. engine (46 cubic inches) that was tested in The cl,linder block is fixed to the output shaft and rotates as a unit. The
1966 and produced l6 horseporver at 1,500 r.p.m. In 1968 and 1969 Mercer rnotiorr of the pistons is the key to this engine: as the piston is forced down
revised his invention and built a second-generation rotary engine. The new into the cylinder, its pressurc on the u,obble plate causes the rotation. The
version is lighter, snraller, arrd simpler. The cruciform cylinder block has pistons run back and forth in thcir toroidal tracks. Piston travel is possible
been replaccd by one cylinder and the supercharger has been discarded. The only by Ietting the block tunr. But thc piston doesn't really reciprocate; it
cylindcr has dual bore and the two pistons have dual diameter. The small orbits around thc nrainshaft and gocs through 30 degrees of track travel in
bore part is ccrrtral, including thc combustion chamber and the full travel of half an orbit. Thc ncxt half-orbit, thc piston nroves 30 degrees back again.
,I? APPENDIX
Thcrc arc l2 cylinclcrs groupccl around the output shaft, and the engine is
statcd to be notably free from vibrartion. Like the Mercer engine, the Sel-
woocl rvorks on thc trvo-strokc principlc. When one side of a piston begins a
power stroke, thc othcr sidc starts a compression phase. Selwood's interest in
this cnginc was 6rst publicized in 1961, and little progress seems to have
been nradc sincc then. Its suitability for automative purposes remains
qucstionable.
C/I TTB UR F TC'R SPIO€R STON
rr----=

'T;-§.

trffi

of the
/
CYLtlYOe',1 I
cross'section proposed Porsche crankless rotary engine.

ELOCN /
ItUU§/N(t

The Selwood rotarv engine, invented by Cecil Hughes.

PORSCHE

Porsche (the wcll-known sports car makers in Stuttgart) patented a crank-

lcss enginc in 1963. Its four-cylinder crucifornr block revolves in a cage, but,
unlikc the N{crccr engine the Porschc gets its power on inward strokes. Each
piston has a roller connected by beltJike links to the other rollers. All rollers
are in permanent contact with a large two-lobe cam on the central shaft.
The housing has an intake port, an exhaust port and a spark plug. The
u,rist pins of thc four pistons are large rollers interconnected by beltJike
Iinks, and are always in contact with a large kidney-shaped cam on the out-
put shaft. The rotor brings each piston through the four strokes with each
one making onc strokc cach way twice pcr revolution. The patent shows a
very short-stroke rotor and sensibly-shaped combustion chambers. Experts
have cxpressed some conccrn o\rer internal friction problems, and the engine
has never been developed. In fact, Porsche has become a Wankel licensee. Cross-section of the Leath rotary engine.
511 APPENDIX
LEATH

Another rotary-block engine, invented by American engineer Harry Leath,

has four cylinders in a square block which revolves inside a cylindrical hous-
ing. Each piston axis runs at 90 degrees to that of its neighbor. The pistons
have an axle running diametrically across the bottom of the skirt and the
axle carrics a pair of rollers that run on an epitrochoidal track inside the
housing and control the motion of the block. The block is carried centrally
in the housing and the rollcr track is designed to correspond with the
rcciprocating tnotion of the pistons in their cylinders.
The inventors claim to have solved the basic problems of sealing
and cooling, but they still know relatively little about the effect of
changcs in porting, shape of combustion area, or spark plug location and
Bibliography
ignitiol timing and they have no clear idea of production costs, durability, ARTICLES, SAE, PAPERS, B()()KS
and ease of maintenance. Inventing a rotary engine is the easy part-the
«liffipulty lics in making it a good rotary engine. Continued experience with
the Warrkel engine and improvcmcnts on it may facilitate development
work on certain other types of rotary engines. As the Wankel engine con- ARTICLES
tinues. to progress, it will become increasingly accepted, and while we are
still thinking of it as an "unconvcntional" cnginc, it may bccorne a fairly ANoN.: Wunkclnrut. I)cr Spicgcl, Nr. 31, 1961.
typical autonrotive powerplant in thc future, whilc other rotary engines still ANoN.: |apancr irrr Nackcn. f)cr Spicgcl, Nr. )1, 1963.
rcrnain in their ernbryonic stages. ANoN.: A Ilistory of Rotary Iingincs and Punrps. The Enginccr, f anuary to
f une, 19 39.
ANso,rrr, R. Ii.: NSU-Wankel Enginc. Automobile Engineer, Mry, 1960.
ANs»ltn, R. I,'.: Rotary Conrbustion lingincs. Automobile Engineer, De-
cember, 1963, |anuary & Iiebruary, 1964.
ANsDALII, R. Ii.: Air-Coolcd Wankcl linginc. Aut«lnrobilc lingineer, August,
lg65.
ANsDAI.n, R. Ii.: 'l"hc lr'frrzdn 0813 \Vankcl lingiuc. Autourobilc Engineer,
April, 1968.
ANs»,*u, R. F'.: Ireasibility of Fligh-Output Wankel RC Engincs. Combus-
tion Errginc Progress, 1968.
BENsINGER,Dtpr. INc. W. D.: Dcr heutige Entrvicklungsstand des Wankel-
motors. Motortechnische Zeitung, fanuaty, 1970.
Bolmn, Mrcuerr.: \A/ankel Progress. Motor, August 24, I968.
BnocxHAUs, HrnnuRr AND Srnonrr, Rupou' M.: Ro 80-cin NSU Mittel-
klasscwagen nrit Krciskolbcnnrotor. Autornobiltechnische Zeitung, Sep-
tember, 1967 .
Burrmn, Cnanrrs: Serious Contender (NSU Ro 80). Motor, September 9,
1967.
voN Frnsrx, Ora,E: NSU-Wankcl Zrvischenbilanz. Autornobil Revue, Iuly
24, 1969.
Fn,tNcrs, Duvox: The Engine That's Giving Detroit "Wankel Fever."
Popular Science, April, 1966.
FREnn, Paur: Vierkammermusik. Das Auto, Motor und Sport, Heft 8, 1970.
FRornn, W.: Entrvicklungsarbeiten an Dreh- und Kreiskolben- Verbrenn-
5T6 BIBLIOGRAPHY ARTICLES, SAE PAPERS, BOOKS 517

ungsmotoren. VDl-Zeitschrift, March I l, 1960. WtrsrLMANN, H. U.: Wankelmut. Das Auto, Motor und Sport, Heft 20,
g69.
FnoE»r, DR. INc.'W^LIER: Kreiskolbenmotoren Bauart NSU/Wankel. I

Motortechnische Zeitschrif t, |anuary 10, 196l . YaprecucHr, f,r,cx: Wankel twin in production. Motor, ]uly 29, 1967.
FnorpE, DR. INc. WALTER: Auszüge aus neueren Entwicklungsarbeiten am YeprecucHr, ]lcx: Wankels for all. Motor, August 17, 1968.
KreiskoJbcrrmotor Bauart NSU/Wankcl. Motortechnische Zeitschrift, Y,ru.rrnroro, KnNtcru: Dic Krciskolbenmotoren clcr Bauart Wankel-NSU
April , 1963. von Toyo Kogyo. Automobiltcchnische Zeitschrift, |une, 1970.
Fnonon, Dn. INc. W'ALTER: NSU Wankel Engine. Automobile Engineer,
Iuly, 1963. SAE PAPERS
Fnornr, Dn. INc. WALTER AND |uNcnrurH, Dtpr.. ING. Gronc: Der Kreis-
kolbenmotor des NSU Spider. Automobiltechnische Zeitung, May and ANsoetu, R. F.: Rotary Engine Development and its Effect on Transport.
August, 1968. Paper #8, SAE of Australasia, presented in Auckland, New Zealand,
GnotsMANN, C,rnr.: fapanese Wankel-Powered Car. Popular Science, April, October l4-l 8, 1968.
1967. Corn, Devln E. AND |oNrs, Cnanrns: Reduction of Emissions from the
HrccrNs, L.R.: The Story of the Rotary Valve. Motor Cycling, fuly 30,1957. Curtiss-Wright Rotating Courbustiorr Engine with an Exhaust Reactor.
Hussn, EucrN WnurLM: Thermodynamische Untersuchungen an der SAE paper #70007+, presented in Detroit, Michigan, |anuary 12-16,
Kreiskolbenmaschine. VDl-Zeitschrift, March I l, 1960. 1970.
Hur, Drpr. ING. IinANz: Zur Geschichte der Rotationskolbenmaschinen. Fnopnn, WaLTER G. Dn.: The NSU Wankel Rotating Combustion Engine.
Automobil Revue, No. 49, 1961. SAE paper #288A, presented in Detroit, Michigan, |anuary 9-l),1961.
Konr, Dtnrrn: The Man Who Made Tonrorrow Come, Sports Car Illus- Fnon»r, Warrun G. Dn.: The Rotary Enginc of the NSU Spider. SAE
' trated, March, 1960. paPer #610722, presentcd irr Cleveland, Ohio, October 18, 1965.
Konr, Dtrrun: Der Wankel-Motor und die Aktien. Das Auto, Motor und FnoEpE, WILTER G. DR.: NSU's Double Bank Production Rotary Engine.
Sport, Heft 18, 1960. SAE paper #680461, presented in Detroit, Michigan, May 20-24, 1968.
Konn, DrBtEn: So entstand der neue Mercedes-Benz C-l I l. Das Auto, |ouEs, CuenrEs: The Curtiss-Wright Rotating Combustion Engines Today.
Motor und Sport, Heft 19, 1969. SAE paper #886D, presented in San Francisco, Calif., August 17-20,
LunvrcsEN, Kenr.: Technique of Tomorrow's Power. Car and Driver, Mry, t964.
l96l . foNus, Cuenrrs: New Rotating Cornbustion Powerplant Development. SAE
LuovrcsEN, K.rnr: NSU Ro 80. Car Life, Dccember, 1967. papcr #650723, prescntcd in Cleveland, Ohio, October l8-21, 1965.
LunvrcsEN, KAnL: Mazda Cosmo. Toyo Kogyo |oins the Rotary Club. Car foNus, CsanrEs: The Rotating Combustion Engine-Cornpact Lightweight
Life, Decenrber, 1967. Power for Aircraft. SAE paper # 670194, presented in Farmingdale,
LunvrcsEN, Klnr: Return of the Gullwing. Car, October, 1969. N.Y., November 7, 1966.
Morrun, GüurHER: NSU Ro 80. Road & Track, October, 1967. Kntlnn, Hnrvrur: Srnall Wankel Engines. SAE paper #680572, presented
MuNpv, H,rRnv: NSU-Wankel Rotary Expansion Engine. Autocar, February in Milwaukee, Wisconsin, Septernber 9-12, 1968.
l9 and 26, 1960. Ptcano, FTnNAND: The Future of Automobile Technique. SAE paper #980A,
Muxpy, H,rnnv: March 2, 1962,
Progress Report. Autocar, presented in Detroit, Michigan, fanuary I l-15, 1965.
NoRnvn, Iex P.: Test Drive of a U.S. Car with a Rotating Combustion Yeuauoro, KENtcsI AND Kunool, Texasxt: Toyo Kogyo's Research and
Engine. Popular Science, April, 1966. Development on Major Rotary Engine Problerns. SAE paper #700079,
RrxrvreNN, W.: Der Nsu-Wankelmotor im Fahrzeug. Automobiltechnische presented in Detroit, Michigan, |anuary 12-16, 1970.
ZeitunB, March, 1961.
ScHr'rtot, EnNsr: Die Drehkolben- und Kreiskolbenmaschine, Entstehung
einer neuen Bauart des Verbrennungsmotors mit überraschenden Eigen- B()()KS
schaften. VDl-Zeitschrift, March ll, 1960. ANsoatn, R. F.: The Wankel RC Engine. A. S. Barnes & Cornpany, South
WlNxnr, IirLIx ANo FnoEuu, DR. INc. WnrER: Bauart und gegenwärtiger Brunswick & Ncw York, 1968.
Entwicklungsstand einer Trochoiden-Rota tionskolbenmaschine. Motor- BuRsutt, Aunnnv Ii.: A History of Mechanical Engineering. The M.I.T.
technische Zeitschrift, February, 1960. Press, Massrrchusctts Institutc of Technology, Canrbridge, Mass., 1965.
WeNxm, Fuux: Rotary Piston Engine Pcrformance Criteria. Automobile Hürrnrv, I{.: Schncllc lr{otorcn scziert und frlsicrt. Richard Carl Schmidt &
Engineer, September, 1964, Co., Braunschu,eig and Berlin, l9rr.
5I8 BIBLIOGRAPI]Y
INlr,rNHuxtrn, Mlncus C.: Rotary Valvc Engincs. Hutchinson's Scientific
& Tcchnical Publications, Londorr, Ncrv York, Mclbourne, Sydney, &
Capc Town, l9 51.
Invruc, Cerr. f . S.: Investigation of the Developrnents in the German
Autonrobilc Inclustry during the War Period. British Intelligence Ob-
iectivcs Sub-Comnrittec, London, 1945,

The Author

I^* P. Nonnvn is the Automotive

Editor of Popular Scia nce Month ly and Contributing Editor of, Auto-
^
mobile Ouarterly, FIe is the author of Sports Car Suspension and The
New Fiat Guide. Norbye has been a full-time writer and editor sirtce
he came to the United States in l96l to take o\/er the position of Tech-
nical Editor of Car and Driyer, rvhere he worked for four years prior
to joining Popular Science. I{e had startecl rvriting about cars as a
spare-time occupation while still in high school in his native Oslo,
Norway. His all-absorbing interest in automobiles Ied him to ioin the
industry after graduating fronr Oslo Comrnercial College in 1951. FIe
worked for a year with I{RG Engineering Conrpany in Tolworth, Sur-
tey, England, and developecl his contacts within the British motoring
press. Returnirrg to Oslo, he ioined the aclrninistratiorr of the Royal
Automobile Club of Norway and handled many assignmcnts for the
Club's journal, Motorliv. In 1954 he became sen,ice manager for the
importers of MG, Morris and Studebaker in Norway but left in 1956
to take up a position with Iisso Touring Service in Paris, France. Trvo
years later he returned to Scandirravia as a service representative for AB
Volvo, working out of the Gothenburg, Sweclen, factory. He helped
establish Volvo's service organization in the United Kingdom and made
frequent visits to Iirance and Belgiunr. Ncrbyc has driven cars since
1945 and is an experienced, professional test driver. His journalistic
specialities are technical writing ancl automotive history. Norbye is
thirty-nine, an Affiliate Mernber of the Society of Automotive En-
gineers, and a Director of the International Motor Press Association.
He lives in |ackson Heights, New York City.