4,..

2PHOTOFACT PUBLICATION TTS-1

TRANSISTOR

TRANSMITTERS
for the
AMATEUR

Complete details, including cir-

cuit descriptions, construction
techniques, parts lists, and op-
erating instructions, for 12 use-
ful transistor devices. Valuable
projects for hams, CB'ers, stu-
dents, and experimenters.
k
 TRANSISTOR

TRANSMITTERS
for the

AMATEUR
by Donald L. Stoner, W6TNS

HOWARD W. SAMS & CO., INC.

THE BOBBS-MERRILL COMPANY, INC.
Indianapolis New York
 Preface
Amateur radio has served to introduce thousands of enthu-
siasts to the field of electronics. For many, amateur radio is the
FIRST EDITION start of a rewarding profession; for others, it is a helpful supple-
ment to their formal education in engineering.
FIRST PRINTING - FEBRUARY,
NOVEMBER,
1964
1966 Constructing your own equipment is a satisfying and educa-
SECOND PRINTING -
tional segment of the hobby. Also, you can keep abreast of the
Co., Inc., Indian- latest developments in circuits and components at little
Copyright © 1964 by Howard W. Sams & States of America. expense.
apolis, Indiana. Printed in the United Construction projects for the beginner, as well as for the
Reproduction or use, without express experienced ham, are described in this book. Included are
All rights reserved.
pictorial content, in with
permission, of editorial orliability
any manner,
respect
projects for novice, technician, general -class, and Citizens -band
is assumed licensees.
is prohibited. No patent herein.
to the use of the information contained For example, the "Novice Powerhouse" in Chapter 5 is an
Number : 63-23372 ideal rig for the beginner or the QRP general -class operator.
Library of Congress Catalog Card The modulator described in Chapter 6 can be added to the
"Novice Powerhouse" for phone operation. The 10 -meter rig
in Chapter 8 can be used to communicate over distances of
thousands of miles under proper conditions.
The latest design and construction techniques are incor-
porated throughout, including transistor circuitry and circuit -
board layouts.
Every piece of equipment has been built and thoroughly
tested on the air. I hope that you will enjoy building and
operating these projects as much as I did preparing them for
this book.

DONALD L. STONER, W6TNS

 Contents
CHAPTER 1
TRANSISTOR TRANSMITTERS ARE DIFFERENT 7
History of transistor development, power and heat dissipation,
transistor packaging, high -frequency operation, transistor pa-
rameters, circuit configurations, power supplies, bias, and test-
ing and alignment.

CHAPTER 2
OSCILLATORS 29
Tickler -coil oscillators, Hartley oscillator, Colpitts oscillator,
Clapp oscillator, oscillator stability, crystal -controlled oscil-
lators, and overtone oscillators.

CHAPTER 3
BUILDING OSCILLATORS 49
Building and testing a crystal checker and calibrator, an 80/40 -
meter "Peanut Whistle," the CB Cyclone, and recommended
transistor types.

CHAPTER 4
POWER AMPLIFIERS 61
Flywheel action, frequency multiplication, interstage coupling,
collector matching, the "Peanut Whistle II," and the "CB
Cyclone II."

CHAPTER 5
THE NOVICE POWERHOUSE 77
How it works, construction, adjustment, and operation.
 CHAPTER 6
89
MODULATION modulation, collector modu-
Frequency modulation, amplitude build a half -watt modula-
lation, build a low -power modulator,
construction, and testing.
tor, build a 5 -watt modulator,

CHAPTER 7
1
111
TUNNEL -DIODE TRANSMITTERS build a 60 -second transmitter,
How it works, the tunnel diode,
construction, and adjustment. Transistor
CHAPTER 8 Transmitters
HIGH POWER FOR CB OR 10 METERS and obtaining components
How it works, construction, testing,
121
Are Different
127
INDEX
There is always an element of glamor and just a little extra
thrill when operating a short-wave transmitter that employs
no vacuum tubes. This is particularly true when you are using
it to communicate with someone several.hundred or even a
thousand miles away.
Before plunging into the construction and operation of these
solid-state "peanut whistles," let's review a bit of theory to
obtain a clearer understanding of how they work. Even if you
are an "old pro" in building rock -crushing tube rigs, transistor
transmitters are different. Let's see why.

THE HISTORY OF COMMUNICATIONS TRANSISTORS

Although it may not be immediately obvious, much credit
should be given to the computer industry for the great progress
which has been made in the field of communications transistors.
In the mid -1950's there were few manufacturers who had the
foresight to see the vast potential of the transistor for elimi-
nating vacuum tubes in communications equipment. They
seemed content to fabricate noisy, delicate devices suitable for
use in raspy -sounding portable radios.
7
 be a poor second compared The designers could not tolerate such sluggish performance
Early transistor devices provedcommunications
to transmitters. in equipment which had to make hundreds of thousands of
to the vacuum tubes used in high -power, high -frequency "decisions" every second. Transistor manufacturers were asked
For several years it appeared that to fabricate. Fortu- to make faster and faster switching devices in order to raise
transistors were theoretically impossible this has not proved the IQ of the next model computer. Devices were also needed
nately for the communications industry
to be the case. was enthusiastic about this ON

The computer industry, however,

immediately visualized the possi-
TD -= DELAY TIME

new electronics infant. Theythousands of heat-generating and

bility of eliminating the
OFF
TR = RISE TIME

computer. Some of the

inefficient tubes used in a typical table -top computers
"dreamers" went so far as to predict
= FALL TIME
TF
secretary or TR TD

( Fig. 1-1) which could be operated by an average TD

ON
bookkeeper.

OFF

Fig. 1-2. Rise, fall, and delay time characteristics.

which would deliver watts of power, rather than milliwatts or

microwattg, in order to drive memory cores in the complex
electronic cerebral cortex of the computer.
This constant pressure on transistor manufacturers for
faster and more powerful devices brought about a revolution
in the semiconductor industry. Vast sums were invested in
research program.13to perfect new ways of making better tran-
sistors for less money. The infant industry progressed from the
humble junction transistor, through the drift field, the diffused,
and the mesa, to the present-day planar transistor, a miracle
of semiconductor technology.
Modern transistors are able to operate at the speed of light.
Fig. 1-1. Transistor IBM 1620 computer. A planar computer transistor is capable of switching from the
of the tubes in the off state to the on state in less time than it takes the light
The transistor was able to replace many but the designers reflected from this page to reach your eyes (remember that
slower and less "intelligent" computers,shortcomings. When light travels at the speed of 186,000 miles per second). Today
had many
soon found that the transistor was accomplished rather it is possible to purchase devices which deliver 20 watts or more
instructed to "turn on," the process
instructions was overcome, of radio -frequency energy while switching on and off 50 million
slowly. Once the delay in following and continued times (50 megacycles) each second. Although these transistors
the "turn-off" signal
it was reluctant to obeymicroseconds. time characteristics
The are very expensive, they are available. Further pressures on
to plod along for several the semiconductor industry, and the ever increasing com-
are illustrated in Fig. 1-2.
9
8
 reduced prices of com-
petition within it, will result in greatly rate of progress con-
transistors. If the present
munications
tinues, it will soon be possible to buy a device which will
tube for less than $10.
perform the same as a 6146

POWER AND HEAT DISSIPATION

wishes to make a more powerful vacuum
If a manufacturer and encloses them in a larger
tube, he employs larger elements capacitance may increase,
bottle. Although the interelectrode will be the same as that
generally speaking, the performance that the larger tube will
of a physically smaller tube, except
without damage. Unfortunately
be able to handle more power
this is not the case with transistors.lossless device, it will always
Since the transistor is not a
is passed through it. Its design,
produce heat when a current improved manufacturing
which minimizes these losses through rid of the heat once generated,
techniques, and its ability to get of the transistor.
determine the power-handling abilitymaterial and the junction larger
Making the amount of active but it also impairs its high -
decreases losses in the transistor,techniques must be employed
frequency performance. Special junction to
low while maintaining a largerather weird
to keep capacitances results in some
dissipate heat. Often this conflict and the
such as the "comb," the "star,"
internal structures,
Fig. 1-3. Construction of a typical transistor.
"snowflake." power transistor (Fig.
If you have ever torn apart a large piece of semi-
1-3) you may have been surprised to see the tiny square inches of copper or aluminum for every watt of power
inside. The large metal slug which com- input. It is preferable to use commercial heat sinks (Fig. 1-4)
conductor material is designed to draw heat from
prises the case (usually copper) which the which compress 100 square inches into a space of approxi-
through the heat sink on tend to mately 2" x 3" x 4". Never run the transmitter at its maxi-
the junction and radiate it
transistor is mounted. Theheat temperature will always out of mum power level in an effort to get the last possible watt into
will tend to be drawn the antenna. Remember, if the power output is reduced by
equalize and, in effect, the metal to which it is attached. half, it only causes a loss of one "S" unit on a distant receiver.
the junction into the cooler are concerned, the junction Experimental evidence indicates that for every 10°C. that the
Where operating life and reliability is the single most important
of the transistor junction temperature is reduced, the expected life span of the
temperature transistor transmitters,
consideration. When building or using tem- device will be doubled.
should be placed on the operating As a "rule How hot can a transistor get? That depends entirely on the
particular importance it as low as possible.
perature of the device to keep manufacturer's rating, which is determined by the package the
should be mounted on at least 10
of thumb," the transistor 11

10
 ADIA

re --.1 in DIA
.185 MAX

009
125 +.002
-.001 .500- M- IN

T cil DIA
.017
DIA
Hu {fr-
I 5 MIN
200
.209
.230
-6
EMITTER
100 .242-
BASE
.200 .278

.041

Fig. 1-4. Commercially built

heat sink. 9
045
.100
COLLECTOR

the type of material it

transistor chip is mounted in, and on should not feel heat (A) TO -5 outline. (B) TO -18 outline.
speaking, one
is composed of. Generally -18 case (Fig.
device in either the TO -5 or TO
in a germanium devices in the diamond or round case (TO -3
1-5). Germanium should never be painful to touch,
and TO -36, respectively) safely around 130°F., which
although they are often operated of pain. Silicon transistors are
is just below the threshold than are
operation at much higher temperaturesTO -5 and 100_.01
capable of transistors in the TYPE No
transistors. Silicon
germanium at 130°F. or higher.
TO -18 package can safely be operated
DATE CODE

in packages which can be mounted on a MAX DIA

Silicon transistors beyond the temperature point 250

heat sink often are operated ,.-± 005

where they are painful to touch. 30

M,
.36 tMAX

their maximum temper-

Whenever you use remembernear
transistors
that it is excessive junction M N MU 77
12
-4-312
MAX .500
375 .710
.C". DIA.
ature rating, always the transistor, and the junction
052 .610

temperature that destroys

010 EMITTER
us, -±-
than the case. .140 MAX DIA
temperature
itself will always be at a higher
.010 'INSULATED
4,0
temperature can rise suddenly LOCATOR PIN .072 MIN

More important, the junction the increased temperature 410.32 --I. .140

and destructively long before

MAX
NF -2A THREAD

reaches the outside of the package.

BASE
Y .005 t- .010
.156 215
RASE

HIGH -FREQUENCY OPERATION

of high -frequency
Vacuum tubes are excellent amplifiers is capable of very
for example,
radio energy. The Nuvistor,low cost. The primary limiting 302, 07
r
15 -
0.2

impressive
performance at in vacuum tubes are .3409
EMITTER

factors of high -frequency performance

20 1.1_.°1

and transit time; that is, the (C) TO -3 outline. (D) TO -36 outline.
the interelectrode capacitances
reach the anode after leaving the
time it takes the electron to Fig. 1-5. Typical transistor cases.
cathode. 13
12
 a much lower frequency of, say, 1,000 cycles. The term ex-
transistors, although it is com-
A similar condition exists in the junctions are in inti- presses the loss of gain in the common -base configuration and
is not valid for the more useful common -emitter circuit. Fur-
pounded many times. For example, attached, and they generally
mate contact, being physically ther, unless input and output impedances are known, it is not
(Ce on the data sheets) . To make possible to determine power gain. The alpha -cutoff frequency
exhibit a high capacitance into the transistor must
matters worse, the signal injected (called Rb) to reach the curve for a typical transistor is shown in Fig. 1-7. It should be
electrode resistance
pass through the C forms a tiny low-pass
junction. The presence of both R andof high -frequency signals
filter which inhibits the amplification 6 db PER

The fact that the electrons must negotiate through

3 db OCTAVE

( Fig. 1-6) . z
TRANSISTOR

SIGNAL FREQUENCY--- lab

SIGNAL IN
Fig. 1-7. Alpha -cutoff frequency curve.

pointed out that frequency is shown on a log scale which makes

the high -frequency gain appear to fall off quite rapidly.
Another term which expresses the high -frequency operation
IN
Rb
SIGNAL
of a transistor is ft, the gain -bandwidth product frequency. It
>SIGNAL is a convenient figure, for, as a "rule of thumb," it reveals the
gain of the transistor at any particular frequency. For example,
a transistor with an ft of 60 would have a power gain of 2 at 30
mc, 4 at 15 mc, 8 at 7.5 mc, etc. Thus, by knowing the frequency
Fig. 1-6. Rb and Ce form low-pass
filter. and the desired power output it is possible to approximate the
confines of a vacuum driving power required.
relatively effortless A third term, Lt.) will probably become the industry stand-
a solid rather than the all the carriers reach the collec-
tends to slow them down. Not ard; this term is defined as the theoretical or computed maxi-
travel different paths and tend to mum frequency at which the transistor is capable of oscillation.
tor at the same time; they accounts for the inability of the
Therefore at fn,a, the power gain is equal to unity and the device
disperse. This storage effect instantaneously. As the fre-
transistor to follow instructions reached where the transistor is not capable of amplification.
a point is it The high -frequency characteristics of a typical transistor,
quency is increased,
whether to turn on or off, and as a result such as the Philco Micro Alloy Diffused (MADT), is shown in
does not know
Fig. 1-8. Note that it has a flat portion where gain is relatively
does nothing. generally becomes quite con- constant. However, as the frequency is increased, a point is
The transistor experimenter have arisen in an effort to define
fused with all the terms that At one time it reached where the internal workings of the transistor combine
performance of a transistor. to oppose the amplification of high frequencies. Above this
the high -frequency the alpha -cutoff frequency (fab) point the power falls off at a rate equal to 6 db/octave. Each
was necessary to know only possible to predict roughly how
time the applied frequency is doubled, the power gain drops
By knowing this figure, it was frequency, and one transistor by 6 db. The point where the rolloff curve intersects the unity -
a transistor would work at a given simply defined
Alpha cutoff isdown
could be compared with another.current 3 db from gain line is known as Lit,.
gain is
as the frequency at which the 15

14
 leads can be heated, as in dip -soldering of printed -circuit
- 60
70
boards, for as long as 10 seconds without internal damage. All
"fi 50 6 db PER OCTAVE silicon transistors can be soldered into a circuit without
20 db PER DECADE
z 40 FREQUENCY damage.
c.6 30

2°
The electrical ruggedness of the transistor is not quite so
e 10 impressive, however. For years the author used an 807 tube
0
max max which had a hole about the size of a lead pencil in the plate.
100
At one time or another the tube had flashed over internally
characteristics of
Fig. 1-8. High -frequency because of excessive voltage and vaporized some of the metal.
typical transistor.
The tube, even with its "lung" punctured, is probably working
or f,,, it would seem
If one works back from unity gain at some low frequency.
today.
In contrast to this, the author currently has a collection of
possible to achieve infinitely high gain transistors which were subjected to excessive voltage. Unlike
a modern high -frequency
As Fig. 1-8 shows, this is not the case;than the durable 807, these transistors are no longer of any electrical
capable of more 50 db of gain. Maxi-
transistor is easily value.
mum available gain (MAG
in Fig. 1-8) is a function of the
impedances. However, Excessive voltage across either of the junctions in a tran-
device current gain and the operating in so small a space sistor can result in instantaneous destruction due to localized
50 db of radio -frequency gain concentrated impossible to tame heating. The carriers, propelled at enormous speeds and in
tends to be highly unstable andneutralization
virtually
and shielding great quantities, tend to "gang up" in one area while taking
without extremely complex the avenue of least resistance. This causes a sudden tempera-
usually held down to 40
schemes. For this reason gains are ture rise which virtually melts the transistor and almost in-
mismatching in the input or output
db/stage by impedance is usually considered stantaneously shorts the junction. The destructive avalanche
circuit, or in both. Thus the 40-db figure Although some power is much faster than the finest fuse or circuit breaker can cope
to be the maximum useful gain (MUG) . reduced circuit com- with.
gain is lost, the increased stability and If a resistance were inserted in series with the device to limit
plexity make it a small price to pay.assumed and La. (which is current flow below the destructive heating point, no damage
Once a 40-db, flat -gain line is sheet) is known, it is possible would occur, even if the junction voltage (VrE ) were increased
given on the manufacturer's data for the transistor at any far beyond the maximum rating. For example, if a 1-meg re-
to predict the maximum power gain sistor were inserted in series with a general-purpose PNP audio
frequency.
The rolloff is more or less constant at 6 db/ octave transistor and connected to a 500 -volt supply, no damage would
40-db maximum usable gain, or
or 20 db/decade. Assuming the rolloff knee can be calculated occur to the junction because the current would be limited to
two decades, the frequency of 100. If the device Li is a few microamperes.
simply by dividing the fn. figure by frequency above this Unfortunately, we cannot do this in a transistor transmitter;
100 mc, the knee occurs at 1 mc. At any the same resistor which limits destructive currents also limits
point the power gain will diminish. the power output below the useful level. Thus the maximum
voltage ratings must be closely observed. It should also be
RUGGEDNESS stressed that the maximum voltage supplied by the DC source
temperatures. must always be much less than the maximum transistor rating.
Newer transistors can withstand high externalsoldered into a For example, in a class -C RF power amplifier, the back emf and
Most modern germanium transistors may be
the case itself. The flywheel action of the coil can generate peak voltages twice
circuit without heat -sinking the leads or
17
16
 high-level can cause this junction to break down. Drive -limiting resistors
1-9A and B). If the stage is four
the supply value (Figs. peaks will reach at least times or peak -voltage limiting devices are often used to prevent this
modulated, the RF or form of destruction. These techniques are shown in Figs.
collector
voltage (Fig. 1-9C). If the antenna loading
the supply reach even higher levels, 1-10A and B.
matching is improper, the voltage can
CIRCUIT CONFIGURATIONS
With the exception of grounded -grid amplifiers, most
vacuum tubes are employed in a circuit where the cathode is
48V common to the grid -input and plate -output circuit. Likewise,
12VDC
24V
this is the customary arrangement found in transistor circuits.
However, there are several other interesting and unusual tran-
oV sistor configurations employed to achieve impedance matching,
improved high -frequency performance, and better heat dissi-
pation.
(C) Increase due to The three basic methods of connecting a transistor into the
(A) Power -supply (B) Increase due to high-level collector circuit are common -emitter, common -base and common -
voltage of 12 volts flywheel effect and modulation.
on collector.
back emf. collector configurations ( Fig. 1-11). The common -emitter con-
back emf, flywheel figuration (Fig. 1-10A) is generally used in radio-frequency
Fig. 1-9. Voltage increase due tomodulation. circuits, since it provides the highest power gain of the three
effect and high-level
to
configurations. For example, in a transmitter the common -
of these peaks can have sufficient duration emitter configuration would require approximately one -fifth
and any one destruction of the transistor. as much driving power as the same transistor when connected
cause instantaneous transmitters, always select in a common -base configuration. The common -emitter con-
Thus, as a rule of thumb for AMat least four times the supply figuration almost approximates the comparable tube circuit,
a transistor with a Vcc rating until it is one-fourth the maxi- which appears to make the operation of common -emitter cir-
voltage, or reduce the supply
with a six -times rating would
be more cuitry more easily understood.
mum rating. A transistor safety margin in the pres- A third advantage is not quite so obvious. Of the three con-
conservative and result in a greater In CW, FM, and SSB figurations, the common emitter is the only circuit capable of
ence of transients or voltage surges.
two-times supply rating should
be
phase inversion. Thus, in radio -frequency amplifier circuits
equipment a device with a
times the rating is preferable. the output is degenerative (opposite phase) with respect to
selected, and one with threenotably the mesa types, the base - the input. Inherently the stage becomes more stable, particu-
In certain transistors,
voltage is also important.
Excessive drive larly on the flat portion of the transistor response curve, when
emitter breakdown connected common emitter. Neutralization will result in an
increase in power gain of as much as 3 db. Because of increased
stability in the common -emitter mode, circuits tend to become
DRIVE I more reproduceable: that is, the gain and bandwidth variations
DRIVE I from transistor to transistor are greatly reduced.
(B) Peak -voltage limiter. There is considerable confusion regarding the relative merits
(A) Drive-limiting resistor. of common emitter versus common base in
protective circuits. transmitter circuits.
Fig. 1-10. Base -emitter breakdown
19
18
 common -base configuration (Fig. 1-11B) would be used on the
6 db/octave slope frequencies. This is because the common -
base configuration becomes somewhat regenerative due to the
in -phase feedback energy. This regeneration increases the
power gain of the stage noticeably and tends to produce more
output power for a given amount of driving power. This be-
comes particularly true and much more pronounced when the
transistor is operated near and only a few db of gain are
available. It is quite usual to find common -base circuitry em-
ployed in equipment operating on VHF and UHF bands. Like
the grounded -grid, vacuum -tube amplifier, some of the driving
power appears in the output tank. This makes the stage appear
to have a higher power gain than is actually the case.
The common -collector configuration (Fig. 1-11C) is an
impedance -matching scheme and is seldom found in trans-
(A) Common emitter. mitter circuits. Its vacuum -tube equivalent is the cathode
follower, and like this circuit, the emitter follower is capable
of matching a high impedance to a low impedance. It is not
capable of voltage amplification but does have current gain.
The emitter follower may occasionally be used for impedance
.matching into high -power stages when link or capacitive
coupling is undesirable. The input impedance of this configura-
tion is determined primarily by the input impedance of the
driven stage multiplied by the current gain of the transistor
in the driving stage.
In transistor transmitters (particularly those where the
transistors operate at elevated temperatures) a variation of
OUTPUT
INPUT

INPUT TANK INPUT TANK

DRIVE DRIVE
(C) Common collector. OUTPUT
OUTPUT
(B) Common base.
and vacuum tube equivalents.
Fig. 1-11. Basic configurations
and the OUTPUT TANK

is the same for all configurations OUTPUT TANK

Actually the fn, of choice. How-

used would seem to be a matter
actual circuit which are not immediately
are subtle differences that the tran--
B+
ever, there of operation is such
1

(A) Common emitter, grounded (B) Common base, grounded

obvious. If the frequency of the frequency collector.
flat -gain portion collector.
sistor is operating on the -emitter configuration should be
common the Fig. 1-12. Variations of basic circuits.
response curve, the just described. However,
employed for the many reasons
21
20
 shown in Fig. 1-12. This is
these circuits may be employed, asof the collector to the heat
done to permit direct grounding
insulating mica washer. Elim
sink without the use of an
transfer greatly reduces the
transistorinating

this resistance to heat that common

temperature. It should be stressed if the tran-
junction still common emitter even
emitter, for example, is This is why the term grounded
sistor collector is grounded. emitter can be quite misleading.
emitter instead of common consider a batterY, a pilot
lamp, and a
To illustrate the point, could be
connected in series. Any part in the circuit
chassis without
meter grounded to a
and could be
called common is true in transistor circuits,
affecting the circuit. The same complex than the simple
although they are considerably more arbi-
described. Ground or common is strictly an rela-
circuit just long as the
be placed at any
trary term and cancomponents
point so
this common point is f aith-
tionship of other to Fig. 1-13. Inexpensive bench voltage source.
fully observed. = = . 12V AUTO BATTERY

POWER SUPPLIES
transmitter circuit may require any-
A particular transistor
volts for proper
operation as con- _°T T COPPER WIRE
SHUNTS

where between 1.5 and 24 involved in vacuum -

potentials
trasted to the dangerously high two sides, however, because for
This coin has flow
tube equipment. considerably more current must heavier
a given amount of power circuit. This requires
in the transistor transmitter and greatly increased bypass FUSE - RATING

wiring, larger meter ranges,

AS REQUIRED

capacitor values. the transistor transmitteravoltage (A) Schematic diagram.

It is recommended that
in this 0
Although this may seem rather unglamorous
source. solution to the power - SW ITCH 8 0
day and age, it is the most practical obtained for less than _*= FUSE 0
A used battery can be drug store
from the supermarket or
than,
supply problem. PANEL MOUNTED

$5 and even a new one than $10 to $12. A

VOLTAGE
WITH SHEET
SELECTOR
METAL SCREWS
seldom runs much more an electronic
times as much. If the lead tie
five to ten for connec-
power supply costing be drilled and tappedshown in the (B) Construction details.
lugs are exposed, they may
tions to make an adjustable voltage source, as The case Fig. 1-14. Details of bench voltage source.
(Figs. 1-13 or 1-14A and B) .
accompanying photos
23
22
is usually thick enough to permit drilling and tapping for the cease. In effect the transistor is self -protecting, much the same
installation of a control panel containing a current meter or as the 6V6-807 transmitter would be if a screen -clamp tube
meter jack, a voltage switch, a fuse holder, etc. ( 1-13B). (controlled by grid bias) were employed.
The adjustable -voltage feature is a useful addition since it Class -A transistor operation is almost identical to tube oper-
permits testing and tuning of transmitters at reduced voltages ation; that is, the device is biased near the center of the linear
before applying full power. curve, and the output is considered to be a true replica of the
In an effort to standardize voltages to aid constructors, all input waveform.
circuits in this book are optimized for use at 12 volts DC. Class -B transistor operation is similar to zero -bias, class -B
Further, all circuits are designed with the common bus con- tube circuitry. A small amount of forward bias may be applied
nected to negative so that they may be used in an automobile, to the stage. Half of the signal cycle pulses the stage into oper-
if desired. ation, while the other half cycle drives the stage well into
cutoff.
Class -C transistor circuits seldom use the large cutoff bias
BIAS AND CLASSES OF OPERATION found in vacuum -tube circuits. The transistor will have neither
Although the theory of operation for tubes and transistors is a forward nor a reverse bias applied. To achieve class C, a re-
quite different, they are both amplifiers and the classes of oper- sistor is usually inserted in series with the drive signal (Fig.
ation are usually considered to be approximately the same. 1-15B). Base current, caused by conduction on one half cycle,
The bias schemes for transistors and tubes are often quite causes current to flow in this "base -leak" resistance, which gen-
different, however. For example, in a simple 6V6-807 trans- erates reverse bias for the stage.
mitter, the RF driving signal may bias the final tube. The Class -A operation is seldom found in a transmitter circuit
positive half cycles cause the 807 to draw grid current, and the other than for oscillators. This mode is characterized by a pure
subsequent current flow through the grid -leak resistor causes sine -wave output independent of the flywheel action of a tank
the grid to become negative with respect to the cathode. Should circuit. Class -A operation results in a freedom from harmonic
the drive signal fail, the bias voltage also ceases and the tube generation and improved frequency stability. This condition
will very likely draw excessive current ( Fig. 1-15A). is difficult to achieve in oscillator circuits because the transistor
usually swings violently between saturation and cutoff and the
output approaches the appearance of a square wave. The fly-
wheel action of a tank circuit helps restore the sine waveshape,
but it is seldom perfect if the original waveshape is badly dis-
torted. To obtain true class -A operation, careful attention must
be paid to such details as biasing, impedance matching, and,
(A) Tube signal bias. (B) Signal -induced bias. most important, the drive level of the feedback energy.
Fig. 1-15. Signal -bias systems for tube and transistor.
Actually, class -A operation is only desirable in variable -
frequency oscillators where stability is of paramount impor-
tance. It is much less important in crystal oscillators, since the
In most transistor transmitter circuits the signal also sup- driven stage contributes far more distortion and subsequent
plies some or all of the bias ( Fig. 1-15B). However, it will be harmonic radiation.
remembered that a transistor, unlike a tube, does not conduct
Class -B operation is usually found in transmitters employ-
until a forward bias is applied. In a transistor transmitter the
forward bias is also supplied by the drive signal. However, in ing transistors well down the 6 db/octave slope where adequate
drive is lacking or more stage power gain is required. Many
the absence of signal the forward bias and the collector current

24 25
power amplifiers which are supposed to be operating class C
may actually be in self -bias class B because of insufficient drive
to create a large reverse bias. In such cases a small amount of
forward bias ( just enough to place the operating point at the
bottom of the linear portion of the curve) will improve output
by overcoming the base barrier potential. This form of class -B
biasing also seems to improve the output of a frequency
multiplier.
The characteristic of class -C transistor operation is
with vacuum -tube operation, although bias may be obtained
a slightly different manner. In a typical PNP common -emitter
amplifier, for example, the negative peaks drive the stage into
conduction while the positive peaks drive the stage beyond
cutoff.

ALIGNMENT
Fig. 1-16. Practical dummy -load system for transmitters.
A transistor transmitter is considerably easier to align than
a vacuum -tube rig. It is only necessary to connect a suitable The impedance of the bulbs may be easily calculated by
dummy load to the output and tune everything for maximum dividing the voltage rating of the bulb by the current rating.
power output. Unless the final transistor is severely overloaded This will equal the bulb resistance (and therefore impedance)
because of excessive drive (which is seldom the case) , no dam- at full brilliance. For example, a No. 47 pilot lamp (0.15 am-
age will occur due to excessive loading, out -of -resonance tuning, pere at 6.3 volts) has a full -brilliance resistance of 42 ohms,
insufficient drive, etc. The only precaution, which cannot be which is a satisfactory match for 50 ohms. A No. 44 pilot, rated
stressed too often, is always have the output properly loaded. at 0.25 ampere at 6.3 volts, has a resistance of 25 ohms, and two
If this precaution is not observed faithfully, excessive voltage in series make a perfect 50 -ohm match.
of sufficient amplitude to damage or destroy the transistor can The amount of power required to light the lamp to full bril-
be generated. liance can be determined by multiplying the voltage and cur-
An excellent dummy load can be constructed by employing rent rating. For example, slightly less than 1 watt will fully
pilot lamps (Fig. 1-16). Although the brilliance determines illuminate a No. 47 lamp, while a No. 44 requires slightly more
the actual impedance of the bulbs (the impedance will be less than 1.5 watts. Combinations of these bulbs require propor-
at reduced brilliance) and tuning changes the brilliance, this tionately more power. By knowing these two facts, dummy -load
technique is very successful for adjusting transmitters. For strings can be constructed for any power or impedance. This
equipment in the 100-milliwatt (0.1 -watt) class, a No. 48 or 49 technique also makes an excellent way of estimating trans-
pilot bulb makes an excellent output indicator. For 1 -watt mitter power output simply by comparing it to a fully lit bulb
transmitters, a No. 47 pilot is excellent. Four series -parallel as illustrated in Fig. 1-17. Power can be measured quite ac-
No. 47 bulbs can be used for transmitters in the 5 -watt class. curately by varying the voltage to a comparison bulb until it is
Above this power No. 44 bulbs can be series -parallel connected the same brilliance as the dummy -load bulb or bulbs. By know-
to make up the correct impedance. ing the power consumed by the comparison bulb, the power

26 27
 PILOT LIGHTS
MILLIAMPERES

TO TRANSMITTER VOLTAGE
OUTPUT SOURCE

Fig. 1-17. System for estimating power output.

required to light the dummy -load bulb to the same brilliance

2
can be accurately calculated.
It should be pointed out that the comparison -method be-
comes increasingly inaccurate as the transmitting frequency
is raised. This is due to the inductance of the spiral-wound
lamp filaments. The error introduced by this inductance is
Oscillators
not particularly severe below 30 mc, however. There is one
other pitfall which almost always traps transistor transmitter
experimenters when measuring power output. One common Oscillator circuits employing vacuum tubes generally are
method of determining the power output of a transmitter is confined to two or three well-known types. The transistor,
to measure the voltage developed across a known value of load however, can be used in several configurations, as well as varia-
resistance. The voltage is then squared and divided by the tions of these configurations. A myriad of circuits result when
value of load resistance and the resultant is the power in watts. these configurations are multiplied by the various types of
However, few RF voltmeter probes are accurate, and small oscillators.
errors in voltage produce large errors in power. Even more Figs. 2-1 and 2-2 illustrate a few of these circuits. None of
important, the probes are calibrated to read rms volts. As is these oscillators are crystal controlled; the addition of a
often the case, transmitter output may not be a pure sine wave
and the voltmeter may indicate something other than the true
rms value. The voltage error increases as the transmitter
output continues to depart from a true sine wave. For this
reason it is strongly recommended that pilot lamps or other
heating devices be used as power -output indicators.
Diode voltmeters frequently create other problems. It is
often noted, when observing the output of a transistor trans-
mitter on an SWR bridge or field -strength indicator, that
modulating the transmitter makes the meter kick downward
rather than showing the increased reading customarily asso-
ciated with upward modulation. However, this is an effect
common to germanium diode detectors and can generally be
ignored. The most accurate test of upward modulation, other
than an oscilloscope, is to observe the dummy -load pilot lamps.
They should increase in brilliance a little more than 20% for
true 100% upward modulation. Fig. 2-1. Variations of tickler -coil oscillators.

28 29
 ALL ABOUT OSCILLATORS
Fig. 2-3 represents the "black box" equivalent of an oscil-
lator circuit. By definition an oscillator is simply an amplifying
stage connected in such a manner that some of the output

TRANSISTOR
OUT LOAD
AMPLIFIER

(A) Basic Hartley circuits.

FEEDBACK

Fig. 2-3. Black box equivalent of an oscillator.

power (PO is coupled back into the input circuit. The feed-
back energy (13f,,,i) must be in -phase or regenerative to sustain
oscillations. An important point to remember is that the energy
used for feedback is no longer available as useful power output.
The circuit must also contain a frequency -selective network
to determine the frequency of oscillation. This circuit can take
(B) Basic Colpitts circuits. the form of an LC network (a tuned circuit) , a quartz crystal,
Fig. 2-2. Hartley- and Colpitts-type oscillators. or in the case of audio -frequency oscillators, an RC network.
Also necessary, though not shown in Fig. 2-3, are the bias
frequency -stabilizing crystal results in another group of similar components required to bias the transistor and stabilize the
circuits. The circuits of Fig. 2-1 are variations of the tickler - operating point.
coil oscillator. Fig. 2-2A represents the basic Hartley circuit, The input and output impedances of a vacuum -tube oscil-
while Fig. 2-2B is the Colpitts-type oscillator. These circuits lator are both high; therefore, losses in the feedback network
and their variations are discussed in this chapter. seldom cause a serious loss of power output. On the other hand,
In the circuits to follow, exact resistor values are replaced the common -base oscillator is characterized by a relatively high
by symbols to simplify the schematic and the discussion of their output impedance but an extremely low input impedance. If
operation. However, if you wish to experiment with a particular some form of impedance matching is not employed, it is neces-
circuit, the following values will get the circuit going even sary to use "brute force" to supply sufficient feedback to sus-
though the values may not be optimum: tain oscillations. For all practical purposes, this power is wasted
and is not available for driving succeeding stages. The same
base resistor RH = 10,000 ohms thing is true for common -emitter oscillators, although not to
forward -bias resistor RE = 100,000 ohms the same extent.
emitter resistor RE = 1,000 ohms Although a transistor oscillator will operate well above the
collector resistor Rc = 4,700 ohms alpha -cutoff frequency (see Chapter 1) , it can never operate
The particular coil and capacitor values in each circuit depend at fn... This is a computed frequency, based partly on RR and
Cf
on the frequency of operation. ', where the gain of the transistor is zero. There is always

30 31
loss in the feedback network and the transistor must supply winding polarity, dots are usually included near the coil or
the power to overcome this loss. As a rule of thumb you can transformer to indicate the start of the winding ( Fig. 2-1).
successfully use a transistor to approximately one-half of f-max,
although the power output at this frequency will be very small. Tickler -Coil Oscillator
Let's see how a typical radio -frequency amplifier can be used If the voltage at point C is compared to the signal at point
as an oscillator. Fig. 2-4 is the schematic for a PNP RF ampli- A, the two will be in phase, since the transistor has introduced
180° of phase shift and the coil (connected as just described)
contributes another 180', making a total of 360°. If points C
and A are connected, the stage will oscillate immediately. The
instant that power is applied to the circuit, forward bias on the
transistor causes a surge of current to flow in the tank circuit.
This causes a damped wave ( Fig. 2-5) of current to flow in
the inductance which is coupled to the link winding. If there
is insufficient feedback, the waveform decays as in Fig. 2-5A
Fig. 2-4. PNP RF-amplifler circuit.

fier. The resistive divider network (RF and RB) provides for-
ward bias, while RE is connected in series with the emitter
circuit to provide DC stabilization with varying temperatures
and supply voltages. Resistor RE is bypassed so that the signal (A) Decaying due to insufficient feedback.
current which flows through this resistor does not oppose the
input signal and lower the stage gain. An output tank, resonant
at the signal frequency, completes the circuit. A signal applied
to the input of the stage will appear amplified many times
across the tank circuit. As the input signal becomes more posi-
tive, it decreases the forward bias, and at the same time the (B) Sine -wave output with proper feedback.
voltage across the tank decreases and is said to be negative - Fig. 2-5. Damped -wave oscillator output.
going. Thus, in this circuit the output signal is always out of
phase with the input signal. If the transistor is operated on and the stage refuses to oscillate. Assuming, however, that all
the flat portion of its frequency -response curve and the tuned conditions are present to sustain oscillations, the positive -
circuit is resonant at the signal frequency, this relationship going half cycle is link -coupled to the base or input of the
will remain constant at 180° of phase shift. Signal coupling stage. The action of the signal induced in the link winding
back through the transistor capacitance lowers the stage gain inverts the phase so that the feedback at this instant is
since it opposes the input signal. negative -going. This, of course, causes the transistor to draw
The signal voltage induced in the link winding may be either even more current and drives the stage further into conduction
in phase or out of phase, depending on the winding direction until the transistor saturates and can no longer amplify. Since
and the polarity of connections. If, for example, the start of no further change in collector current
can occur, the feedback
both windings is common and they are wound in the same signal ceases. This represents the peak of the sine wave. At
direction, the voltage at point C will be 180° out of phase with this point the static bias
and attempts to decrease theapplied to the oscillator takes over
respect to point B. Where confusion can arise as a result of collector current. This decreases

32 33
the current through the coil and causes the direction or polarity one -tenth the number of turns on the link as are on the primary.
of the feedback to reverse so that is is now positive -going at The circuit shown is used as a beat -frequency oscillator and
the emitter, thus driving the stage into cutoff. When cutoff is the tuned circuit consists of a J. W. Miller No. 2045 IF trans-
reached, which represents the negative peak in the cycle, the former (T1). If the circuit does not oscillate immediately,
stage can once again no longer amplify and the feedback ceases. reverse the connections to the link winding. Any general-
As before, the forward bias tries to bring the collector current purpose RF PNP transistor can be used with the typical values
up to the static -bias point by overcoming the signal -induced shown so long as it is capable of a reasonable amount of ampli-
reverse bias. This causes the feedback polarity to reverse and fication at the frequency of oscillation.
the cycle repeats. This second cycle reinforces the damped Fig. 2-7 illustrates another form of tickler -coil oscillator that
wave so that the output signal consists of a series of sine waves employs the tuned circuit in the base or input section. A con -
as in Fig. 2-5B, rather than the decaying waveform in Fig.
2-5A. This oscillation will continue until the feedback path is
broken or until the supply voltage is disconnected. The fre-
quency of oscillation is determined almost entirely by the
resonant frequency of the tank circuit. This type of circuit is
called the tuned -collector, tickler -coil oscillator.
A practical working circuit of the tickler -coil oscillator is
shown in Fig. 2-6. The circuit can be easily constructed to

Fig. 2-7. Alternate tickler -coil oscillator circuit.

siderable mismatch occurs in this circuit due to the high tuned -

circuit impedance to be coupled through C11 to the low -
J. W. MILLER 1 2045 I F
impedance base of the transistor. The mismatch causes severe
TRANSFORMER OR EQUIVALENT instability; for this reason the circuit is seldom used.
.01

Hartley Oscillator
For reasons of circuit complexity, the use of a link winding
100K

10K
may be undesirable. Also the link may be required for coupling
Ti driving power to a succeeding stage. In such cases the tran-
120

Fig. 2-6. Practical circuit of a tickler -coil oscillator.

investigate the principles just discussed. The coil and capacitor

may be resonated at any desired frequency within the operat-
ing range of the transistor. There should be approximately Fig. 2-8. Hartley -type oscillator.

34 35
sistor may be used in a Hartley -type oscillator (Fig. 2-8). In moving the ground symbol has changed the Hartley from a
this configuration the tuned circuit is common to both the common -emitter to a common -base configuration. However,
input and output circuits. The coil is tapped, and the tap point the performance of either circuit is identical.
is at ground potential. Thus the RF at each end of the coil is
180° out of phase with respect to the other end. The tran- Colpitts Oscillator
sistor provides the remaining 180° of shift required to sustain Another oscillator circuit which generally operates in com-
oscillations. If the coil is center -tapped, capacitor CB is made mon base is the Colpitts. The Colpitts circuit may take on
small in value to match the high -impedance tuned circuit to many similar forms with only minor differences tailored to suit
the low -base impedance. As mentioned earlier, this results in the frequency of operation. The common -base Colpitts circuit
poor frequency stability. The reason for this is that Cil forms seldom fails to oscillate. Because the impedances may be
a capacitive voltage divider (and therefore an impedance di- closely matched, the Colpitts has the best frequency stability
vider) in conjunction with the base -emitter capacitance of the of any oscillator circuit. The Colpitts has an additional advan-
transistor. However, this junction capacitance is subject to tage in that it does not require a tap on the coil.
wide changes in value as the transistor parameters are changed. Consider the circuit shown in Fig. 2-10 and assume that it is
For this reason a preferable impedance -matching method is to a common -base, radio -frequency amplifier. Signals applied to
place the coil tap near the base end of the coil (where the im- the emitter will appear greatly amplified but not phase -inverted
pedances are more nearly matched) and use a large value of
CB. The impedance transformation would then be determined P FEEDBACK

by the turns ratio between the winding above and below the E

tap.
The Hartley oscillator illustrates the point mentioned ear-
lier, that ground or common is an arbitrary term. Observe the
circuit in Fig. 2-9. Although this appears to be an entirely dif-

Fig. 2-10. Colpitts oscillator circuit.

in the collector circuit. If a capacitor is now connected between

the collector and emitter, the in -phase energy is returned to
the input of the amplifier and a feedback path is created. This
configuration is particularly well suited for VHF applications
where only a small amount of feedback capacitance is required.
Many transistors have sufficient junction capacitance to oscil-
Fig. 2-9. Common -base Hartley oscillator. late without the aid of the external capacitor. This is
particu-
larly true if an RF choke is inserted in series with the emitter
ferent form of oscillator, careful examination reveals that th resistance.
only actual changes are moving the ground symbol from th Fig. 2-11 illustrates another form of Colpitts oscillator which
center tap to the bottom of the coil and a slight physical re is more suitable for the high -frequency
portion of the radio
arrangement of the position of the components. Note also tha spectrum. Here a capacitive voltage divider is connected di -

36 37
 tion of operation is somewhat different. Since the dividers place
an imaginary tap on the coil and since this point is at RF
ground, the coil is capable of phase inversion as in the Hartley.
Thus the signal fed back to the base through Clt is 180° out of
phase with respect to the collector signal. With the additional
180' of phase shift supplied by the transistor, the stage is
capable of sustaining oscillations. The impedance -matching
functions of the capacitors in this circuit are the same as for
Fig. 2-11. Colpitts with capacitive -divider feedback.
Fig. 2-11.
One disadvantage of the Colpitts which may have been un-
rectly across the coil with the emitter connected to the junc- covered by the reader is the fact that the circuits shown so far
tion of the two capacitors. As mentioned earlier, this capacitive require a variable inductance to change resonant frequency.
divider makes an ideal system for matching the high impedance If either of the two divider capacitors is varied, it will change
of the collector to the low impedance of the emitter. Because the resonant frequency but will also upset the impedance
the emitter has a lower impedance than the collector, capacitor match. Introducing this mismatch will degrade the oscillator
CE is always much larger in value than C, to provide a correct performance. The problem can be avoided by using a dual -
the
match. Capacitor CE should be approximately 10 times section capacitor; however, a simpler scheme is shown in Fig.
value of Cc. By changing the ratio of capacitance, the feedback
2-13. The divider capacitors are made 5 or 10% smaller than
can be varied to increase or decrease the vigor of the oscillator.
A tap placed at an equivalent point on the coil results in exactly
the same operation, but the circuit then becomes a Hartley
TUNING/
oscillator. CAPACITOR
Another form of Colpitts is shown in Fig. 2-12. This circuit
is similar to that in Fig. 2-11. The ground has been moved

Fig. 2-13. Colpitts with tuning capacitor.

their normal value. A variable capacitor is then connected in

parallel with the divider and used to adjust the resonant fre-
quency of the circuit. This does not in any way upset the
impedance -matching characteristics of the divider, and the
theory of operation remains the same so long as the fixed
capacitance (Cc) is at least 10 times the value of the variable
Fig. 2-12. Alternate Colpitts circuit.
capacitor. It should also be pointed out that for variable -
frequency oscillators this scheme will provide straight-line
frequency tuning of the oscillator. For example, 1 mmf of
from the bottom end of the tuned circuit to the junction of the change at minimum capacitance will shift the frequency of
capacitive divider, and a collector DC return resistor is em- oscillation the same amount as for a 1-mmf change at maximum
ployed. Although the action and performance of the oscillato capacitance.
are similar to those of the one shown in Fig. 2-11, the explana
39
38
 Still another form of Colpitts is the Clapp oscillator illus- OSCILLATOR STABILITY
trated in Fig. 2-14. The Clapp varies from the Colpitts in that The oscillators which have been discussed so far are free
a series -tuned resonant circuit (consisting of L and C) is running as contrasted to an oscillator which is stabilized by a
substituted for the parallel tank. Capacitor C, shunting the
quartz crystal. When a crystal is not employed, the circuit is
generally referred to as a variable -frequency oscillator. To ob-
tain the greatest frequency stability in any oscillator, three
factors must be taken into consideration: (1) bias, (2) junc-
tion capacitance, and (3) temperature.
For best stability the oscillator should be biased near the
center of the transistor Lt L curve so that operation occurs
TUNING
over the linear portion of the curve. If excessive drive or im-
CAPACITOR
proper bias carries the transistor violently into the nonlinear
region, frequency instability will occur. For the same reason,
the oscillator voltage source should be stabilized to prevent
Fig. 2-14. Clapp oscillator configuration. frequency instability.
The experimenter may note an interesting effect when work-
ing with transistor variable -frequency oscillators. The use of a
collector junction and CI, in parallel with the base junction are common bias source for the collector and base -emitter junction
made quite large, being in the order of 1,000 mmf at 5 mega- will maintain a relatively constant ratio of the two voltages
cycles. This large capacitance tends to mask the variations in with varying supply potentials. The effect of these variations
the junction capacitance and its detrimental effect on the on frequency is partially canceled out as the change in one
oscillator frequency stability. The frequency of oscillation is voltage is somewhat counteracted by the change in the other.
determined almost entirely by the tuned circuit consisting of The reason for this is that an increase in collector voltage
L and C. causes an increase in frequency while an increase in emitter
Another version of the Clapp oscillator, shown in Fig. 2-15, voltage causes a decrease in frequency. The experimenter will
permits grounding the tuning capacitor. Since the value of find that a particular value of emitter resistance and emitter
current will cause variations in supply voltage to have a mini-
mum effect on the oscillator frequency. The exact value can
be found empirically by starting with an RE value of 1,000
ohms. Experiment with the value of RI: while noting the exact
amount of frequency shift for a given supply -voltage change.
Increase RE to 1,500 ohms and repeat the experiment. A
Micro Alloy Diffused Transistor (MADT-Philco), for example,
will show the best frequency stability with an RE value of
3,300 ohms and approximately 1 to 2 ma of emitter current
Fig. 2-15. Alternate Clapp oscillator circuit.
when operating from a 12 -volt source.
The supply voltage changes the frequency of oscillation be-
CB is large with respect to CE, the base is more nearly at R cause of its effect on the transistor junction capacitance. It is
ground and the transistor is considered to be operating in th logical to assume that by using high values of C across the
common -base mode. inductance in the tuned circuit, small changes in junction
40 41
capacitance will have less effect on the frequency of oscillation. Crystal -Controlled Oscillators
Obviously a change of 1 mmf out of 1,000 mmf will have less In this type of frequency -stabilized oscillator a quartz crys-
effect on the frequency of the tuned circuit than will 1 mmf tal either replaces the tuned circuit or takes over its frequency -
change out of 100 mmf. For this reason always use as much determining function. Quartz crystals are used because of their
capacitance as possible in parallel with the tank coil, particu- extremely high Q and excellent frequency stability over a given
larly in the Hartley circuit. Excessive capacitance will lower temperature range.
the tank impedance to the point where oscillations may cease. Fig. 2-16 shows the electrical equivalent circuit of the crys-
The exact amount of capacitance in the tank is a matter for tal. Capacitor Cs, inductance L, and resistance R in series
experimentation. To get in the "ballpark," a good approxima-
tion is to divide the frequency of oscillation into the constant -
number 5,000. Thus at 5 mc, one uses 1,000 mmf of capacitance
in the tank. The inductance value then must be juggled to
Fig. 2-16. Equivalent LCR circuit
of a crystal.
F.
make the circuit oscillate at 5 mc. If it does not oscillate with
this much capacitance, the value can be reduced and the
inductance increased. Conversely, if the circuit still oscillates
strongly, try additional capacitance for even more stability. represent the mechanical vibration characteristics of the crys-
It might be pointed out that such high -C circuits tend to make tal. These three components will exhibit a low impedance at
the inductance more critical. This is quite true, but the overall one frequency, which is determined primarily by the physical
stability with high -C circuits is better due to the masking of dimensions of the quartz and which, in turn, determines its
junction -capacitance variations caused by parameter changes. vibrating characteristics. This point is called the series -
For the same reasons, make the capacitive -divider values as resonant frequency of the crystal.
large .as possible (while maintaining a 10 -to -1 ratio) in the The equivalent circuit also shows a parallel capacitance
Colpitts end Clapp circuits. (Cr) , representing the electrostatic capacitance between the
Temperature change causes some unique problems in tran- crystal electrodes. These electrodes may be either pressure
plates or plated silver discs in intimate contact with the quartz.
sistor oscillator circuits. Changes in temperature have the same
effect on components as in tube -type, variable -frequency oscil-
The shunt or parallel capacitance CI., in conjunction with the
inductance of the crystal, forms what might be thought of as
lators. In tube equipment, however, the temperature of compo-
nents is usually elevated above ambient because of the heat a second tuned circuit in the crystal. This is called the parallel -
generated by the tubes. Thus, changes in ambient temperature resonant frequency and is always slightly higher in frequency
have less effect on the tube oscillator. In transistor equipment
than series resonance. The difference between the two fre-
quencies may be anywhere between a few hundred cycles and
the components usually operate at ambient temperature, and a few kilocycles, depending on the type of crystal. The differ-
changes in ambient have an adverse effect on the frequency of
ence is always very small compared to the actual frequency of
oscillation. For this reason temperature compensation of tuned
either point. The impedance of the crystal at parallel resonance
circuits may be required. For best stability always use good
quality components, such as silver mica or dipped mylar ca- is extremely high.
pacitors and temperature -stable coil forms. Employ transistors Fig. 2-17 shows a curve of impedance versus frequency. It
which have a low value of Ce, such as mesa and MADT types. should be pointed out that the impedance of the crystal is also
Silicon transistors are less affected by temperature than ger- high at frequencies other than the resonant frequency. How-
manium and therefore are recommended for variable-frequenc ever, the highest impedance occurs exactly at parallel reso-
oscillator applications. nance while the lowest impedance occurs exactly at series

42 43
 inf.

Fig. 2-17. Impedance -versus -

frequency curve of a crystal.
CE

0
F1 F2 FREQUENCY

Fig. 2-19. Crystal -controlled equivalent of a

resonance. Note the steep slope between the two points; this Hartley oscillator.
indicates that it is a very high -Q tuned circuit.
Fig. 2-18 shows the crystal -controlled version of the tickler -
coil oscillator discussed previously. The crystal (Y) is con-
nected in series with the feedback path and operates at series

Fig. 2-20. Crystal -controlled version of the

Colpitts oscillator.

Fig. 2-18. Crystal -controlled version of a

tickler -coil oscillator.

resonance. Since the crystal impedance is lowest at series reso-

nance, it may be thought of as a switch. Above or below series Fig. 2-21. Alternative version of Colpitts,
resonance the impedance is so high that it attenuates the crystal controlled.
feedback energy below the point where the circuit can sustain
oscillations. Fig. 2-22 is somewhat different from the previous circuits
Fig. 2-19 is the crystal -controlled equivalent of the Hartley. in that the crystal replaces the tank circuit and operates at
Capacitor Ci, in Fig. 2-8 is replaced by the crystal in Fig. 2-19. parallel resonance. Although this is another variation of the
Once again the crystal operates at series resonance as in the Colpitts, it is sometimes referred to as a Pierce oscillator. If the
previous circuit. crystal were replaced by a parallel -tuned circuit, the configura-
Figs. 2-20 and 2-21 are the crystal -controlled equivalents of tion would be identical to the Colpitts discussed earlier.
the Colpitts. As in the Hartley, the crystal replaces CI, and Capacitors C(' and CE form an impedance -matching capaci-
operates at series resonance. tive divider in shunt with the equivalent parallel -tuned circuit

44 45
 II with respect to the collector signal. When the phase is again
cc inverted by the transistor action, the stage oscillates.
Fig. 2-24 illustrates a rather unusual oscillator. At first
RC glance it appears to lack most of the necessary requirements
CEI

LB

Fig. 2-22. Circuit of a Pierce oscillator.

inside the crystal. The in -phase energy at the collector is

coupled back into the emitter circuit through C, to create
a feedback path. The frequency of oscillation is determined Fig. 2-24. Unusual transistor oscillator circuit.
primarily by the parallel -resonant frequency of the crystal
and to a lesser extent by the values of C,, and Cn. These capaci- for an oscillator circuit (tuned circuit, impedance -matching
tors should be made as large as possible to swamp the effect capacitors, etc.). If you ignore the ground connections for a
of junction capacitance on the frequency of oscillation. As the moment, the theory of operation becomes somewhat easier to
values are increased, however, the frequency of oscillation will understand. The crystal operates at series resonance. Capacitor
be pulled away from parallel resonance lower in frequency. A CE is relatively small and somewhat critical in value (500 mmf
point will be reached, as the frequency of oscillation approaches at 5 mc). The base -emitter capacitance of the transistor, in
series resonance, where the phase shift in the crystal induc- conjunction with CE, forms an impedance divider. Neglecting
tance is too great to support oscillations. the grounds, CE is in series with the crystal, and feedback oc-
Fig. 2-23 is the common -emitter version of the oscillator just curs between emitter and base through the crystal. Although
there is no phase inversion or voltage amplification, the tran-
sistor does have current gain. A transistor which has an f, of 100
mc or more will oscillate quite vigorously. The effect of junction
capacitance on frequency can be minimized by connecting a
capacitor equal in value to CE between the base and emitter
connections of the transistor. Output RF may be obtained
across CE. A low -impedance output, quite independent of the
oscillator section, may be obtained by lifting the collector from
ground and connecting a 50- to 500 -ohm resistor in series with
this point. The output voltage will vary between a fraction of
a volt and several volts, depending on the value selected.
Fig. 2-23. Common -emitter Pierce oscillator.
Overtone Oscillators
described. In this circuit the crystal assumes a phase -inverting Crystals have another characteristic which has not been
function the same as for a parallel -resonant L 'C circuit. Radio - mentioned. If the output tank circuit of Fig. 2-20 and 2-21,
frequency energy coupled to the base is 180` out of phase for example, is tuned to a harmonic of the crystal frequency,

46 47
the circuit will oscillate at this frequency. If a 10-mc crystal
were used in either circuit and the output tank tuned to 30
mc, the crystal would oscillate at the third harmonic. The crys-
tal is excited to vibrate three times as fast. Most crystals are
capable of establishing vibration modes at 3, 5, and 7 times
their natural resonant frequency. These higher -order modes are
called overtones. Because of the difficulty of fabricating crystals
for frequencies above 15 mc, oscillators above this frequency
usually employ overtone crystals. It should be mentioned that
3
crystals operating on an overtone always oscillate at series
resonance. Thus a fundamental crystal will have an output
slightly less than the actual marked frequency multiplied by
the order of overtone: -- Building Oscillators
Now that the theory of transistor transmitters has been
covered, let's take a look at the practical aspects and construct
several circuits which illustrate these principles. Since the be-
ginning of any transmitter is the oscillator, let's start there.
Several of the circuits which follow, such as the oscillators
in this chapter, are built on electronic pegboard, manufactured
by Vector and by Lafayette Radio. This method of construc-
tion allows simplified step-by-step assembly and results in the
most reproduceable performance short of using an etched cir-
cuit board. Several of the projects in the following pages in-
clude circuit -board layouts, in case the experimenter prefers
to use this method of construction.

1
A 0BCDE
O
FGH11
0 0 0 0 0 0 0 0
2 O 0 0 0 0 0 0 0 0 0
3 O 0 0 0 0 0 0 0 0 0
4 O 0 0 0 0 0 0 0 0 0
Fig. 3-1. Layout guide for 5 O 0 0 0 0 0 0 0 0 0
construction projects. 6 O 0 0 0 0 0 0 0 0 0
7 O 0 0 0 0 0 0 0 0 0
8 O 0 0 0 0 0 0 0 0 0
9 O 0 0 0 0 0 0 0 0 0
10 O 0 0 0 0 0 0 0 0 0

48 49
 The use of a pegboard chassis also provides an "electronic fundamental frequency between 100 kc and 15 mc will oscillate
roadmap" which is standardized at 10 holes square. Exact posi- in the crystal checker. This includes those hard -to -get -going
tions on the "map" can be located by key letters and numbers, surplus crystals between 370 and 500 kilocycles used in World
as shown in Fig. 3-1. For example, A-1, J-1, A-10, and J-10 War II tank transmitters. The crystal checker can also be used
are the four corner holes. as a 100-kc frequency standard for setting dial calibrations on
communications receivers. It is only necessary to insert a suit-
CHART 3-1. Recommended Transistor Types able 100-kc crystal and replace Cl (47-mmf mica) with a 7- to
Manufacturer Type Numbers 45-mmf rotary trimmer capacitor. This trimmer is used to
zero -beat the crystal with WWV at 5 or 10 mc.
Amperex OC-170, 2N2672
Motorola 2N741 How It Works
Philco 2N1727, 2N1745 The circuit is similar to the one described in Chapter 2, ex-
cept for the addition of Cl. This capacitor is added so that the
RCA 2N370, 2N371, 2N372
crystal "sees" the standard calibration capacitance of 32 mmf.
2N384 Thus the crystal oscillates at parallel resonance and the fre-
Texas Instruments 2N711 quency can be determined by measurement with an accurate
The circuits are designed so that the type of transistor used frequency meter or communications receiver.
is not critical. Chart 3-1 shows several popular transistors that Construction
are used in many of the projects. Each circuit was optimized Insert pegs at A-1 and J-1 and connect together with a piece
with the low-cost, high -gain Amperex 2N2672, then it was of solid tinned wire ( Fig. 3-3). This line becomes the positive
tested with the other types to verify operation. bus. Insert pegs at A-10 and J-10 ( Fig. 3-3) and connect to-
BUILD A CRYSTAL CHECKER
1
ABCDEFGH11
Fig. 3-2 is the diagram of a crystal tester which not only can 2

measure crystal activity but also can be used to check the fre- 3

quency of oscillation. Any reasonably active crystal with a 4

6
OUTPUT
7

8
C5 R4 2.a:
--- 270 9

C3 mmf
1K .05 10
X MFD
Fig. 3-3. Layout of the crystal checker.

gether as before. This is the negative bus. Install the pegs for
the transistor at E-5 ( base), F-6 (collector), and F-4 (emitter).
12VDC
° The t resistor can be soldered to these pegs or a socket can
Fig. 3-2. Schematic of the crystal checker. be use s shown in the accompanying photograph (Fig. 3-4).

50 51
 to 12 volts, positive DC, is available at the cathode of the
audio -output tube. The negative line should be connected to
chassis ground. A lead from the collector of X1 should be
wrapped around the receiver antenna terminal and coupled
for proper signal level.

BUILD AN 80/40 -METER PEANUT WHISTLE

It's truly amazing how far a fraction of a watt can be trans-
mitted when it is radiated by an efficient antenna. The "pea-
nut whistle" shown in Figs. 3-5 and 3-6 has been used to con-
tact stations more than 100 miles away on the 40 -meter band.
Signal reports indicated that more distant contacts could have
been made. The circuit can also be used for testing crystals on
their fundamentals at frequencies between 2 and 15 mc.
How It Works
Fig. 3-4. Photo showing crystal socket. The peanut whistle is a practical use for the circuit shown
in Fig. 2-24. However, in Fig. 3-5 a 330 -ohm resistor has been
The two remaining pegs are installed at D-5 and D-7 for the
crystal socket. If the unit is mounted in a metal box, these
points can be connected to a crystal socket on the front panel.
A socket suitable for hermetically sealed crystals (HC -6 and
CR-18/U style) can be connected in parallel with the large -pin
socket for testing both types of crystals.

Testing
The circuit is energized by connecting a source of voltage
Fig. 3-5. The 80-40 meter "Peanut Whistle."
(6 to 12 volts) in series with an SPST toggle switch and a 0 to
5 -ma meter to terminals A-1 or J-1 (positive) and A-10 or J-10
(negative). With 12 volts applied, the meter will read approxi- connected in series with the collector of X1. The signal when
mately 2 ma with no crystal in the circuit. When the crystal taken from this point is isolated from the oscillator section of
is inserted, the current will increase. The activity of the crystal the transistor.
is roughly proportional to the increase in the meter reading.
The meter is not required, of course, if the circuit is used as a Construction
signal generator or frequency standard. The anut whistle is built on a 21/2 " x 21/2 " board. The
If the oscillator is installed in a communications receiver, transist is positioned the same as in Fig. 3-3, and pegs are
there is no need to use a battery power supply. A source of 6 placed i the four corner holes. The crystal is positioned at

52 53
C-6 and C-8. A peg for the output is inserted at J-6. The exact To use the peanut whistle as a transmitter, the output ter-
layout is shown in Fig. 3-6 and the accompanying photograph. minals are connected to the folded -dipole antenna (Fig. 3-7).
The 330 -ohm resistor (R4) should be disconnected to avoid
absorbing transmitter power. The folded -dipole type of an-
tenna, such as the one shown in Fig. 3-7, must be used to com-
468

300 OHM LINE TO

TRANSMITTER (ANY
LENGTH )

Fig. 3-7. Suggested antenna for the "Peanut Whistle."

plete the collector DC circuit. This antenna is constructed

from 300 -ohm transmission line of the type used for television
reception. The length of the antenna section (in feet) can be
accurately determined by dividing 468 by the frequency in
megacycles. Thus, for the 40 -meter novice band (7.2 mc) the
AB CDEF GHI1 antenna should be cut to 65 feet. The two ends of this
1
piece are twisted together and soldered. One conductor is cut
2
exactly at the center and attached to the transmission -line part
3 of the antenna, as shown in Fig. 3-7. The other end of the trans-
4 mission line is connected to the transmitter output terminals.
5
A telegraph key in series with a 12 -volt battery completes the
6
circuit. Any of the transistors shown in Chart 3-1 can be used
in this circuit.
8
BUILD THE CB CYCLONE
9

10 The one -transistor transmitter shown in Fig. 3-8 is capable

Fig. 3-6. Layout for the "Peanut Whistle."
of impressive performance. The power output is less than one -
tenth watt, yet the transmitter is capable of transmitting
signals a thousand miles or more when the skip is in on ten
Testing meter and the Citizens band. The unit may be used on the
The circuit can be used as a crystal oscillator, exactly as Citize s band without a license, provided the circuit is adjusted
shown in Fig. 3-5. The output is quite strong, and there is no for no ore than 100 milliwatts input (8 ma at 12 volts DC)
need to couple the signal into the receiver antenna terminal. and is onnected to an antenna not more than 5' long.

54 55
 h- 5/16"

GN

10 TURNS
Fig. 3-10. Winding details for Ll. NO. 28
ANTENNA CLOSE WOUND

COLLECTOR

L1 WINDING
+ 12 VDC DETAILS

Fig. 3-8. Schematic diagram of the "CB Cyclone."

Construction
Pegs for the power connections are inserted at the four
corner holes as before, but the transistor is positioned at E-4
(emitter), D-5 (base), and E-6 (collector). The crystal is
inserted at C-6 and C-8. A guide to parts layout is shown in
Fig. 3-9.
ABCDE F G H I 1

10
Fig. 3-11. Finished "CB Cyclone."
Fig. 3-9. Layout for the "CB Cyclone."
ordering a crystal for this unit, a third -overtone type for the
To insure duplicating the performance of the unit, the coil -
winding details shown in Fig. 3-10 should be followed closely. frequency of operation should be specified.
suitable antenna is shown in Fig. 3-12. The length depends
The exact type of slug -tuned coil form is not important so long he frequency of operation and can be determined by divid-
as the outside diameter is approximately 5/16" in diameter. A in the frequency in megacycles into 468. The coaxial shield
printed -circuit type of coil form can be seen in Fig. 3-11. The leo is connected to J-10, and the center conductor is con -
circuit will work equally well on the Citizens band or 10 meters ne d to J-4.
with the component values given and the coil shown. When
57
56
 468 CRYSTAL CHECKER
PARTS LIST
oMpMo
Quantity Item No. Description
1 Cl 47-mmf mica capacitor.
2 C2, C5 270-mmf mica capacitor.
RG-59/u COAX. TO TRANSMITTER
(ANY LENGTH) 2 C3, C4 .05-mfd disc capacitor.
1 R1 100K, 1/2 -watt, carbon resistor.
1 R2 10K, 1/2 -watt, carbon resistor.
1 R3 1K, 1/2 -watt, carbon resistor.
Fig. 3-12. Suggested antenna for the "CB Cyclone." 1 R4 2.2K, 1/2 -watt, carbon resistor.
1 X1 Transistor (see Chart 3-1).
Testing 1 Xtal See text.
As in earlier projects, the circuit can be used to test CB I Piece of "Vectorbord" (type A)
crystals for frequency and activity. To test the unit, connect 10 holes square.
a 12 -volt battery in series with a 0 to 10 -ma meter to the power 9 Push -in terminals
terminals on the board. The collector current will be approxi- (Vector T-30 or equiv.).
mately 3 ma with the crystal out. If the circuit is oscillating,
the current will rise when the crystal is inserted.
Tuning is similar to that of a tube transmitter. As the coil ap-
proaches resonance by tuning the slug, the collector current PEANUT WHISTLE
will dip and then rise on either side of the resonance point. PARTS LIST
Connecting an antenna, or dummy load (such as the No. 49
pilot lamp shown in the photo), will make the collector current Quantity Item No. Description
rise. With the antenna shown in Fig. 3-12, the collector current 1 Cl 510-mmf mica capacitor.
will be approximately 10 ma. The antenna loading can be de-
1 C2 .05-mfd disc capacitor.
creased by inserting a 50-mmf trimmer capacitor in series with
the output lead and coaxial cable center conductor. The pilot 1 R1 33K, 1/2 -watt, carbon resistor.
lamp will load the transmitter more heavily than the antenna, 1 R2 3.3K, 1/2 -watt, carbon resistor.
causing it to draw more than 10 ma. When the transmitter is 2 R3, R4 330 -ohm, 1/2 -watt, carbon resistor.
tuned properly, the bulb will glow brightly. l X1 Transistor (see text).
Although all the transistors shown in Chart 3-1 will oscillate
in this circuit, only the Amperex 2N2672 is capable of the Xtal See text.
power output described. The 2N711 and 2N741, for example, 11 Piece of "Vectorbord" (type A)
put out only enough power to barely light the bulb. For this 10 holes square.
reason the 2N2672 is highly recommended. 9 Push -in terminals
A suitable modulator to permit voice transmission is de- (Vector T-30 or equiv.).
scribed later in the book.

58 59
 CB CYCLONE
PARTS LIST
Quantity Item No. Description
1 C1 22-mmf mica capacitor.
1

1
C2
C3
Ll
68-mmf mica capacitor.
.05-mfd disc capacitor.
10 turns of No. 28 enameled wire
closewound on 5/16" -diameter
4
slug -tuned form. Link 3 turns
same wire closewound over cen-

1 R1
ter of primary coil.
22K, 1/2 -watt carbon resistor. Power Amplifiers
1 R2 2.2K, 1/2 -watt carbon resistor.
1 R3 330 -ohm, 1/2 -watt carbon resistor.
1 X tal See text. Although the low -power transmitters described in the
1 Piece "Vectorbord" (type A) previous chapter are capable of impressive range, it is con-
10 holes square. siderably easier to make contacts with more power going in-
15 Push -in terminals to the antenna. This can be accomplished by adding a power -
amplifier stage to the oscillator circuits described earlier.
(Vector T-30 or equiv.).

FLYWHEEL ACTION

To work efficiently, the power amplifier must be properly

matched in both the input and output circuits. The tuned -
circuit Q must be high to insure a minimum of harmonic
radiation. Fig. 4-lA is the circuit of a typical link -coupled,
power -amplifier stage. The drive from the oscillator consists
of a sine wave applied to the base -emitter junction of the
power amplifier. Negative half cycles forward -bias the stage
and cause a large pulse of collector current to flow. No current
will flow during the positive half cycle, and the output, there-
fore, consists of a series of pulses resembling the output of a
half -wave rectifier. This waveform is shown in Fig. 4-1B; it is
obtained by replacing the tank circuit in Fig. 4-lA with a
10 -ohm resistor. Alone, these pulses, although they are capable
of conveying intelligence, are rather useless. They are rich in
harmonics and should not be applied to an antenna where
60 61
 circuit to ring at the harmonic frequency. This process can be
repeated many times by cascading frequency multipliers to
permit a low -frequency oscillator to control a much higher
OUTPUTD carrier frequency.

INTERSTAGE COUPLING

To be useful, the power output of one stage must be coupled

(A) With inductance in output.
to the input of a succeeding stage. The coupling system should
introduce a nimum of loss since the power output is gen-
erally deter ned by the amount of driving power.
There are hree principal ways in which two stages may be
10 coupled to ther, through inductive or capacitive coupling,
OHMS and throu filter networks. Although inductive coupling is
undoubted the most popular system, particularly for high -
frequency equipment, the other two methods have advantages
which the experimenter should be aware of.
(B) With resistance in output. Inductive coupling generally involves the use of a primary
coil connected to the driving stage with a link coil wound
Fig. 4-1. Typical RF power amplifier showing input and output
RF waveforms. around the primary and connected to the driven stage. Occa-
sionally a secondary coil may be inductively coupled to the
the harmonics can be radiated or cause interference on primary to minimize harmonics. In this case, the link should
unauthorized frequencies. be wound over the secondary.
To minimize harmonic radiation, it is customary for the Link coupling is generally used for impedance matching.
collector current to flow through a tank circuit ( Fig. 4-1A ) If the loaded Q of the tuned circuit is not important, the col-
,

consisting of a high -Q coil and capacitor combination resonant lector is usually attached to the primary, as in Fig. 4-1A. The
at the signal frequency. Each time a pulse of current flows impedance step-down ratio is easily computed from the
through the coil, it causes a ringing similar to a damped wave formula:
and has a duration of several cycles. Each pulse reinforces this N=
damped wave, and the result is a continuous series of relatively les
pure sine waves that are free of harmonics. The action is similar where,
N is the turns ratio,
to that which occurs in oscillator circuits ( Chapter 2).
Z1. is the primary impedance,
Zs is the secondary impedance.
FREQUENCY MULTIPLICATION However, in transistor transmitters the impedances are some-
times difficult to determine, so the number of turns on the link
The tank circuit can also be resonated at a multiple of the winding and the degree of coupling are usually adjusted for
drive frequency to reinforce the production of harmonics. In maximum power output. A good starting point is 1 turn on
this type of frequency multiplier, every second or third pulse the link winding for every 8 to 10 turns on the primary. Maxi-
( for second- or third -harmonic production) will cause the tank mum coupling will occur when the link is near the center of

62 63