Aviation Throttle Body Injection – an update

One goal of this article is to provide some information on both Rotec and Ellison TBI systems
including some of the commonalities and differences in these units. Another is to speculate on
where we currently stand in industry in terms of purchasing and maintaining these systems. It
would also be beneficial to develop or strengthen the user community for those who are
interested in self maintaining their TBI systems within the boundaries of safe practices and the
limitations of our capacity due to availability of equipment and accurate understanding of how
these systems need to be set up.

Ellison TBI’s were invented, manufactured and marketed by Ben Ellison in Seattle area who
retired and eventually sold the business to Steve Glover, in Southern California. Glover is currently
overhauling units and working towards manufacturing new units. Rotec is in Australia and along
with TBI’s they manufacture a series of small radial aircraft engines.

The best description for these units is they are pressure regulated variable venturi throttle body
multiport spray injection. They somewhat mix technologies from several different fuel-metering
systems. Like Bendix pressure carburetors they use a diaphragm with air at inlet impact pressure
on one side and fuel at venturi pressure on the other to regulate fuel pressure to a semi-constant
that is not affected by changes in inlet pressure, without the use of a float. Like a Mikuni or Bing
slide carburetor they use a slide throttle valve (versus butterfly) to emulate a variable diameter
venturi that approximates a constant pressure delta (impact to venturi) across the full range of
throttle opening. Because engine load will also affect MAP and RPM, fuel pressure and air pressure
will vary but as a ratio they are held fairly constant as the throttle moves. Where the Mikuni or
Bing change fuel flow by moving a tapered needle in a jet, the TBI’s maintain controlled air/fuel
ratios during throttle changes by opening, or closing a series of tiny fuel ports in the metering tube
running through the center of the throttle slide.

Having mentioned a constant ratio of fuel to air it must be pointed out that this is not actually
desirable for air cooled engines because they often produce more heat than they can reject during
certain phases of engine operation. In general we use the evaporation from extra fuel to lower
both inlet and subsequently lower combustion temperatures when the engine produces too much
heat for the cooling available. So we run it rich, typically during high power settings and in some
cases at idle settings (because the cooling is so poor). Meanwhile, during cruise periods where
heating is low, cooling is high and economy is desired we will want to lean the ratio of air to fuel.
So any fuel metering system must be able to somewhat autonomously provide these modified
ratios depending on what power setting is being used. The Ellison and Rotec TBI’s do this by
uncovering a series of asymmetrically spaced fuel ports as the throttle is opened. While this valve
mechanism is very different it is the same principle used on a Bing or Mikuni carb, when the fuel-
metering pin valve is moved by the slide throttle. The Mikuni/Bing pin is cut at differing tapers
throughout its length. Unlike the Bing or Mikuni the other thing the Ellison/Rotec units can do to
alter the ratio, in this case manually, by changing the orientation of the fuel ports within the
surrounding airstream. By facing them into the impact air the increased fuel rail pressure will
close the fuel regulator some and produce less flow. Alternately, facing them at 90 degrees to the
airflow, at the apex of the venturi siphoning action will open the fuel regulator some increasing
flow.

Because of this design, changes in atmospheric pressure will have a similar affect on both fuel and
air pressure such that the need to adjust mixture for changes in altitude is quite a bit less than
many other fuel metering systems.
This image from the Rotec manual shows the basic
arrangement of the regulator, slide throttle and fuel
spray/mixture tube with fuel ports arranged to increase flow
appropriately as the throttle opens.

The conversion process of an RV-6 with a Lycoming O-360

engine from a Marvel/Schebler MA4-5 float carburetor to the
Rotec MKII 48.4/5 TBI was fairly straight forward, very much
on a par with the installation of an Ellison TBI. Both require
some rigging modifications because the action of the throttle
and mixture control are not on the same axis and one is linear
while the other rotates along an arc. Ben Ellison was very
adamant that the throttle action must be oriented so that its
slide motion was at 90 degrees from the crankshaft axis. The RV-6 exhaust does not leave much
room in this area so the throttle slides forward to
open. At this early point in operational testing the only
difference I’ve noted is when leaning to extreme lean
of peak the hottest cylinders tend to be one and three,
which are on the side away from the direction the
spray bar emits fuel in this installation. This may be
why Ellison was emphatic about installation
orientation. The Rotec also provides a button on the
fuel regulator, which can be pushed in flight with the
addition of a lever and cable. This opens the regulator
thereby priming the engine. They claim it can be used
to run the engine at full throttle but when I tried it
there was way too much fuel, even at full power. It
does prime very nicely and I should add the engine will not start unless primed., cold or hot
Starter cranking speed does not produce enough venturi flow to get the fuel regulator to open.
While I haven’t tested it, the engine easily starts with a little prime so I’m reasonably sure it will
also hand prop easily. But you will need enough battery power to run a boost pump for a bit.

Performance of both units seems to be very good, both allow one to control mixture to lean of
peak quite well during lower power settings. Both were installed with identical air box systems,
modified from the Vans FAB vertical air box kit. Both Rotec and Ellison call for a large plenum with
stable high-pressure air available to the inlet with little turbulence. I did greatly increase the
volume of the air box relative to the Vans original design, which is strongly recommended by both
Ellison and Rotec.

The one idiom of the Rotec that I have not yet

worked out is higher EGT’s in the last 3/8” of
throttle opening. This is the one place where fuel
enrichment is critical, however… “high” is a very
relative term with respect to EGT’s. What is too
high? This is not well spelled out and can change
due to many variables. I found that this engine
tends to peak EGT’s at about 1450f during cruise
with the float carb and therefore I concluded that
going above 1480f was undesirable and much
above 1500f was probably bad. I openly confess there is no science to this but I know that the
Marvel/Schebler never went much above 1350f at full throttle and the Rotec will easily go higher
than 1550f at the same power setting. While there is no science defining what is too high, clearly
by comparison to the MA4-5 Carb (and inherently Lycoming in its choice of this model carb) it’s
running leaner than was intended, but only in that last bit of throttle opening. Rotec’s manual also
claims the unit can open too much depending on the engine so one needs to be sure MAP and RPM
keep going up, and a throttle stop is put in place where they stop. Well, on this engine MAP and
RPM increase throughout the full range of throttle travel.

Since mixture adjustability at all other power settings seems to be just fine I’m generally
convinced the fuel ports on the extreme open throttle end need to be opened up a small amount,
for this installation. For those who are wondering if the air box design might be a factor due to
high ram air, when I close off the ram air thereby opening heated air the EGT’s do drop a little as
expected, but the “too high” behavior is essentially the same. I also opened up the idle fuel screw
some to enrichen things. The manual says it comes from the factory at 1 3/4 turns from closed
however I found it at about 1/8 turns from closed. Using the mixture cut off/RPM rise technique to
determine idle fuel ratio doesn’t work that well on this kind of TBI because the mixture doesn’t
exactly shut the fuel off. It is running fairly rich at idle though with it set at 1.75 turns as evidenced
by loping and a significant rise when I go full lean at 1000 RPM. As I sort out the high power issue I
intend to reset the idle fuel back to leaner.

After I spend more time tweaking things I’ll update this section.

Physically the two units are similar. The Rotec is about one inch shorter from top to bottom flange.
Both flanges are dimensionally the same, although the Rotec is dimensionally smaller in other
ways. The Ellison uses a ball end clip that screws onto the end of a 10/32 threaded throttle cable.
The Rotec uses two provided spacers that also accept a 10/32 threaded shaft. But the travel
distance here may be offset by an inch or more because the Ellison ball end adaptor is no longer
necessary. The mixture rotates similarly and can be indexed to similar positions. Like the Ellison,
this is limited by interference with the throttle shaft. However, it should be noted that the shaft
rotates in the opposite direction in terms of going lean to rich, or rich to lean. (To update: it turns
out that Rotec will modify direction of rotation during the ordering process, but this will require a
phone call.)

The Rotec manual seems to be more comprehensive than that provided by Ellison. They also sell
rebuild kits… wherein a couple of things need to be said about other internal differences. The
Ellison uses a Grose-Jet fuel inlet valve that instead of being a needle has two balls (one glass and
one steel) stacked within a brass seat and a spring-loaded lever that is pushed on by the impact air
diaphragm. The Rotec just uses a spring-loaded valve that directly rests against the diaphragm. I’m
not sure either is better. Rotec recently made a change to their system to improve the valve by
bonding the seal ring to the valve stem.

One thing I did not like about the Rotec is the threads cut for the bolt holes on the air box side
flange were kind of torn up. They were cut too fast and tore some in production. Their manual also
claimed lock washers were on all the screws, which apparently no longer get used. However, one
aspect of the Ellison that I’ve never liked is the heavy reliance on retaining parts by staking dents
into the body. This might work marginally, but one can only stake so many dents before the body
is no longer serviceable. Since these major parts are currently not available this limits the
serviceability of these units.
Ellison was never willing to sell parts for these units, in part for good reason. While they are
inherently simple they can be easily misassembled and set up so badly as to be unsafe. So the
question at this point is, is there anything Ellison owners can do to ensure continued safe and
reliable operation and do this within our capabilities and resources?

I do think that there is some work owners can do on a periodic interval that is within the capacity
of many homebuilders and the equipment available to them that could increase service life
between overhauls. However, like Bendix pressure carburetors, full overhauls of these units must
be done by the current manufacturer, Ellison Inc. under Steve Glover.

I’d like to suggest a users group might be able to compile a basic parts list of what I call class A
parts that one can use to clean up and inspect their units. Having said this let’s talk about what it
means to overhaul or rebuild something. Ultimately several things occur during an overhaul.
Everything gets cleaned and inspected. Class A parts are discarded and replaced, no questions
asked. All other parts are reused, replaced, or restored to serviceable condition and reused. The
system is then reassembled and adjusted or setup to some potentially exacting specifications. It is
this latter part that I would argue most of us are not well equipped to accomplish. However, with
respect to cleaning things up and replacing all O-rings, top plate seals and the diaphragm there are
many of us who could accomplish this without too much difficulty.

Anytime the fuel regulator valves, springs and levers (all with staked screws) need to be removed
the unit must be readjusted and set up with bench testing. But one could easily remove the
diaphragm plate, clean/rinse out any debris, and replace the diaphragm and gasket without
affecting the regulator valves. The same can be said for removing the top plate where one can
easily clean out the gunk around the throttle slide and replace the few O-rings therein. However,
removal of any other staked in, or riveted parts, removing the throttle shaft from the slide (lots of
locktite) or replacing any of the three Bal-Seal rings on the mixture and throttle shafts are all
things that need appropriate equipment, materials and post-assembly adjustments.

Not only are we not equipped to replace these parts and make these adjustments, sourcing some
of these parts seems to be problematic. I have found most of the other O-rings, roll pins, screws
and retention internal push rings. I also think a users group could compile a list of known wear
issues to look for whenever anyone choses to clean and inspect their units. Anybody doing such an
inspection should have clear criteria to determine when it should instead be returned to Ellison
for a full overhaul.

In my recent visit to Steve’s shop it is clear he is gearing up both to produce more new units and to
increase their overhaul operations. Now, having typed all this, existing Ellison users as well as
those new to the TBI idea might also consider the option of converting to the Rotec, which is
clearly a viable option. The conversion should be fairly easy. For some they’ll need to build an
adaptor to fill in the missing one inch, and for others this is not a factor. While Rotec is in Oz, they
have a knowledgeable rep in Florida who has been helpful. They do claim their units are
warrantied and their forum seems to reflect an interest in serving their clients well.
Here is a link to a number of images reflecting the installation I did. Also there are some images of
an Ellison EFS-4-5 disassembled.
www.miravim.org/4RE

Ellison is now in Southern California owned by Steve Glover and he can be reached at info@ellison-
fluid-systems.com http://ellison-fluid-systems.com/index.htm

Rotec just recently completely upgraded their website with the introduction of their new MKII
product line. The earlier units had the fuel regulator separated from the TBI. The new ones have
them integrated just like the Ellison units.
https://www.rotecaerosport.com/tbi

Brian Kelly is their Florida rep… [email protected]

Additional Notes:
Grose-Jets were originally made by D&G Valve Manufacturing and is no longer in business.
However, LLP Manufacturing in NY seems to still be making them. Several of the Ellison units have
been having these valves jam open (while running), and in one case jam closed (while shut off).
This does not appear to be caused by wear because in one case they were brand new. So, at this
point it is not clear if wear, substandard manufacturing or contamination is causing this. The
pictures below show this behavior being replicated for both valves. A general guess would be
contamination as the likely culprit. Filtration recommendations by both Ellison and Rotec must be
followed.

This shows the inner valve jammed open.

And this shows outer valve jammed,

thereby jamming both open.

Most of the class A parts are reasonably priced, but the square PTFE .125” bar is not cheap. PTEF
is generally not cheap anyways. The Bal Seal seals are also pricy. The Grose-Jet valves have yet to
be identified. #07-415 or sometimes sold as #415 seems to be close but it has an unthreaded
segment that needs to be gone and the inlet port is smaller. So it is not clear if these could be
modified because the lower steel ball may be sized to the diameter of the inlet hole.
Ellison EFS-4-5 Parts

O-rings
2 - AS568-008 Top plate small seals, AS568-009 might also work
2 - N2.00X12.5 metric Inlet fittings, or AS568-112
1 - AS568-156 Top plate main (square shaped groove, AS568-157 might also work)

Screws – (Aircraft Spruce may sell these if you call)

6 - AN500 A8-8 8/32 x 1/2" Drilled head fillister head screws
6 - MS24677-24 Drilled head cap screw 10/24 x 5/8” (Two may
be MS24677-25 10/24 x 3/4" for throttle bracket)

Roll Pins
2 - 1/16” x 7/8” Mix tube
1 - 1/16” x 1/2" Slide/throttle shaft
1 – .100” x 3/4" (possibly longer) Mix lever stop

Diaphragm and gasket

Winderosa LP07402 – diaphragm kit for Tillotson DG1HD, the full repair kit also includes a 155A-
23 lever and 32-79 pin and a bunch of stuff you won’t need. Winderosa 451466

The following parts require special tooling and post installation adjustments.
Lock ring and seals for mixture and throttle shaft seals
2 - Daemar P/N TI-0044-PA, TI - Toothed Internal Stainless Steel "Push on" Ring 7/16” x 1/4"
(www.rotorclip.com seems to offer them under the P/N TI-44SS, not sure about retail sales,
https://www.huyett.com/Products/Fasteners/Retaining-Rings/Push-On-Rings/TI-044-
SS?searchText=TI-44 )
3 – For mixture and throttle shaft seals, they are a plastic V lip seal with spring 3/8” x 1/4" OD/ID,
made by Bal Seals, Inc. https://www.mcmaster.com/#13125k75/=1b66ry8 McMaster Carr has
them under the PN 13125K75 spring loaded seal, $22 ea
An AS568-010 O-ring might also work but probably won’t last as long.

Fuel inlet valve (2 in the EFS 4/5, 1 in the other Ellison units)
2 – Grose-Jet, similar to #415 common to some Tillotson carburetors, but the #415 is slightly
longer and its inlet port is smaller.
Dimensions .500” open length, .451” closed length, 5/16 hex, threads 5/16-32, inlet port .125”, six
outlet ports centered on each the hex flat .075”.
2 – Seat gasket 16B-199, often comes packaged with seat/Grose-Jet.
2 – Tillotson levers 155A-23
2 – lever pins, Tillotson 013406, or 32-79
4 – retention screws, brass, pan head, slotted, 3-48 x 3/16”

Teflon/PTFE Parts (McMaster-Carr sells these PTFE parts by the foot.)

4 - .125 x .125 x 2.375”apprx. square bar
1 – 3” x 6” x .040” sheet, trimmed to fit
4 – AN426-3 rivets, length to fit (apprx. ½”)

If anybody ever figures out what model number of Grose-Jet is used and a source, or has any other
questions feel free to drop a note to: [email protected] or [email protected]
Additional measurements not shown on the installation diagrams for Ellison and Rotec.

These are for general comparison only. Although the top and bottom flanges are respectively the
same the units are quite different dimensionally. The Ellison is 4.95” tall and the Rotec is 3.93”
tall. I have not weighed them but I suspect the Ellison is somewhat heavier.
CHT/EGT/MAP/RPM data from the Rotec MKII 48.4/5 Lyc. O-360, Jan 2018