£ 4
$2.70
FUNDAMENTALS
OF
TRANSISTORS
by
LEONARD M. KRUGMAN B.&, M.S., P.E.,
Signal Corps Engineering Laboratories
JOHN F. RIDER PUBLISHER, INC.
480 Canal Street • New York 13, N.Y.
FIRST EDITION
Copyright 1954 by
JOHN F. RIDER PUBLISHER, INC.
All rights reserved. This book or parts thereof may
not be reproduced in any form or in any language
without permission of the publisher.
Library of Congress Catalog Card No. 549064
Printed in the United States of America
PREFACE
Ever since the point contact transistor was announced by the Bell
Telephone Laboratories in 1948, a considerable effort has been directed
toward the improvement of transistor manufacturing and circuit design
techniques. As a result, the transistor has now evolved to a point where it
is suitable for many applications, both as a direct replacement for and
as a supplement to electron tubes.
That the transistor caused an immediate and intense interest in
the electronic field is not difficult to understand. Here is a device which
acts like a triode, yet together with its protective housing is smaller than
a jellybean. It requires little power, no warmup time, is extremely
rugged, promises indefinite life, and in addition, has certain unique
characteristics which make it suitable for many novel applications.
The superiority of the transistor over the electron tube in applica
tions where miniaturization of space and power requirements are primary
factors has already been established. Transistor hearing aids and car
radios that operate directly from a battery supply are typical applica
tions. The RCA Princeton Laboratories built and demonstrated a com
pletely transistorized television set some time ago. There is no doubt
that everyone connected with the electronic arts will have to meet this
latest addition to the family — the transistor.
While a massive quantity of literature is available for the physicist,
the mathematician, and the research and development engineer, little
has been consolidated in practical form for the technician and the ama
teur. "Fundamentals of Transistors" is intended for this group. It is also
intended that this book will serve the initial needs of engineering stu
dents and engineers who are confronted with transistors for the first time.
For this reason, advanced physical and mathematical concepts have been
purposely avoided. However, all the fundamentals necessary to assure a
complete understanding of basic transistor operation, performance, and
characteristics have been included.
Space limitations, and the large amount of duplication in existing
material, preclude listings of exact credits. However, the author would
be lax indeed if he did not offer his gratitude to Mr. Seymour D. Uslan
and Mr. Sidney Piatt for their patience and assistance in editing the
manuscript; to Mr. C. L. Hunter of the Signal Corps Engineering Labor
atories for his assistance in circuit fabrication; and in particular to Mr.
C. E. Bessey of the Signal Corps Engineering Laboratories for his assist
ance and guidance.
May, 1954 L.M.K.
New York, N. Y.
CONTENTS
Chapter Page
1. Basic SemiConductor Physics 1
2. Transistors and Their Operation 8
3. The Grounded Base Transistor 19
4. Grounded Emitter and Grounded Collector
Transistors 44
5. Transistor Amplifiers 71
6. Transistor Oscillators 96
7. Transistor HighFrequency and
Other Applications 119
Appendix 135
Index 137
Chapter I
BASIC SEMICONDUCTOR PHYSICS
Structure of Matter
For many years, the atom was considered to be the smallest particle
of matter. It is now known that the atom is composed of still smaller en
tities called electrons, protons, and neutrons. Each atom of any one ele
ment contains specific quantities of these electrical entities.
Physically, the electrons rotate around the core or nucleus of the
atom, which contains the protons and neutrons. Figure 11 illustrates
the layout of a carbon atom. A carbon atom contains six each of elec
trons, protons, and neutrons. Note that the six orbital electrons do not
rotate at equal distances from the nucleus, but rather are restricted to
two separate rings. With respect to the size of these electrons, tremendous
distances exist between the electrons and the nucleus. If it were possible
to magnify the atom by a factor of 10 14 , that is, one hundred thousand
billion times, the electrons would be the size of basketballs, with an orbit
spacing of approximately 12 miles.
The negative electrical charge of the electron is exactly equal and
opposite to the charge of the proton. The neutron has no charge. The
electron is three times larger than the proton, but its mass is only .0005
that of the proton. In an electrically balanced atom, as illustrated in
Fig. 11, there is an equal number of electrons and protons.
Gravitational, electric, magnetic, and nuclear forces all act within
the atom. These forces tend to keep the electrons revolving in their orbits
around the nucleus at tremendous speeds. As might be expected, the
electrons located in rings close to the nucleus are tightly bound to their
orbit and are extremely difficult to dislodge. The outer or socalled val
encering electrons are, comparatively speaking, loosely bound to their
orbit. The ease or difficulty with which electrons can be dislodged from
the outer orbit determines whether a particular element is a conductor,
insulator, or semiconductor.
Conductors, Insulators, Semiconductors
Conductors are materials that have a large number of loosely bound
valencering electrons; these electrons are easily knocked out of their orbit
and are then referred to as free electrons. Insulators are materials in
which the valencering electrons are tightly bound to the nucleus. In be
tween the limits of these two major categories is a third general class of
materials called semiconductors. For example, transistor germanium, a
semiconductor, has approximately one trillion times (1 x 10 12 ) the con
ductivity of glass, an insulator, but has only about one thirtymillionth
(3x 10 s ) part of the conductivity of copper, a conductor.
The heart of the transistor is a semiconductor, generally the ger
manium crystal. Other semiconductors such as selenium and silicon have
1
FUNDAMENTALS OF TRANSISTORS
/ TIGHTLY  BOUND
/ INNER RING
NUCLEUS
/ /
/_N J / 6 PROTONS \ 1
Vv* ( \6 NEUTRONS/
^ /
/
/
LOOSELY BOUND
OUTER RING
(VALENCE RING)
Fig. 12. The carbon atom in
short form for trantistor phytic!.
Fig. 11. The carbon atom.
been used in transistors, but germanium has proved to be the most widely
applicable material. The general semiconductor principles discussed in
this book apply to all elements used as transistor semiconductors.
Insofar as transistor operation is concerned, only the loosely bound
orbital electrons and their associated protons are of importance. For the
purposes of future discussion it is therefore convenient to picture the
carbon atom in the short form illustrated in Fig. 12. Note that in this
figure only the valencering electrons and their associated protons are
indicated; the tightly bound inner orbit electrons and their respective
protons are not shown. Thus, the carbon atom in the short form contains
a nucleus with a +4 charge around which the four valencering electrons
rotate. The short form simplifies the graphical representation of semi
conductor operation, as will be seen later.
Crystal Structure
Covalent Bonds. Carbon is occasionally found in nature in a stable
crystalline form, the diamond. In this form, each valencering electron,
moving around the nucleus of a carbon atom, coordinates its motion
with that of a corresponding valencering electron of a neighboring
atom. Under these conditions, the electron pair forms a covalent bond.
Equilibrium between the repulsion and attraction forces of the atoms is
reached at this time, the previously loosely bound valencering electrons
now are tightly bound to their nucleus, and cannot easily be dislodged.
This effectively reduces the number of available free electrons in the
crystal, and hence reduces its conductivity. Thus, carbon, generally a
semiconductor, becomes an insulator in the diamond form.
The Germanium Crystal. Like the carbon atom, the germanium
atom has four valencering electrons. Thus a shortform illustration of
BASIC SEMICONDUCTOR PHYSICS 3
the germanium atom would be similar to that shown for the carbon atom
in Fig. 12. In addition, when germanium is in crystalline form, the four
valence electrons of each atom form covalent bonds, and are tightly
bound to the nucleus. Figure 13 (A) is a shortform illustration of the
structure of the germanium crystal in this perfect state. (For simplifica
tion, the atoms in this figure are shown in a twodimensional plane
rather than in the three dimensions found in nature.) Note that all co
valent bonds are complete and that no atoms or electrons are missing
or misplaced. The pure germanium crystal is an insulator and is of no
use in transistor work. However, pure germanium can be changed into
a semiconductor by adding minute quantities of certain impurities, or
by adding heat energy (phonons) , or by adding light energy (photons) .
Any of these actions increases the number of free electrons in germanium.
If an excess free electron could be added to a pure germanium crys
tal without changing the structure of the crystal, the electron would
move through the crystal as freely as an electron moves through a vacuum
tube. However, when pure germanium is treated so as to become a semi
electronpair
■COVALENT BONDv
FREE ELECTRON FROM
DONOR ATOM
I 1 I' I' l
© "= © ■= © "= ©
I 1 l l l»
62) =.. (g*\ = (el) = (S?)
<&<& ~=®
l> I 1 I 1
© ^©^=~©
. I INDIUM I
1 1 II IMPURITY) I
1 ' ATOM '
(£) = (oe) = Q)
I'
(A)
■=@
I 1
i» i 1 i 1 L i 1 i 1
© ^ © ^ ©^ ©^©
■X
I I ,1 ARSENIC .1 .1
I 1 I 1  'IMPURITY  ' I'
® ~= ® ~= ® ~= ® = (&
,1 ,1 ,1 ,1 ,1
©^©^©^©^®
(B)
Fig. 13. (A) Pure germanium in crystalline
form. (B). Ntype germanium. (C) Ntype ger
manium.
HOLE CAUSE0 BY
ELECTRON TAKEN ( C J
BY ACCEPTOR ATOM
4 FUNDAMENTALS OF TRANSISTORS
conductor, the symmetry of the crystal is destroyed. Consequently, any
one excess electron moves a short distance, bounces off an imperfection,
and then moves on again. This collision is similar to the collision be
tween an electron and a gas molecule in a gas tube.
Donors — NType Germanium. When impurities having five elec
trons in the valence ring are added to germanium, each impurity atom
replaces a germanium atom. Four of the impurity atom's valence elec
trons form covalent bonds with the valence electrons of neighboring ger
manium atoms. The fifth electron is free and is available as a current
carrier. Pen tavalent type impurities are called donors because they donate
electrons to the crystal transistor germanium thus formed. Such transistor
crystals are referred to as N type because conduction is carried on by
means of the Negatively charged electrons, contributed by the donor
atoms. This action is illustrated by Fig. 13 (B) , with arsenic acting as
the pen tavalent impurity.
The application of a dc potential across the Ntype crystal forces
the free electrons toward the positive voltage terminal. Every time an
electron flows from the crystal to the positive terminal, an electron en
ters the crystal through the negative voltage terminal. In this manner a
continuous stream of electrons flows through the crystal as long as the
battery potential remains.
Acceptors — PType Germanium; Holes. Figure 13 (C) illustrates
a second method of forming transistor germanium. In this case an im
purity having three valence electrons (indium) is added to the pure
germanium crystal. Each such trivalent impurity atom replaces a ger
manium atom, and in order to complete its covalent bond with neigh
boring germanium atoms, the impurity atom borrows a fourth electron
from any one of the other germanium groups. This destruction of a ger
manium covalent bond group forms a hole. A hole is an incomplete group
of covalent electrons which simulates the properties of an electron with
a positive charge. These trivalenttype impurities are called acceptors be
cause they take electrons from the germanium crystal. Germanium con
taining acceptor impurities is called P type because conduction is effected
by Positive charges.
Connection of a battery across a Ptype crystal causes the holes to
move toward the negative terminal. When a hole reaches the negative
terminal, an electron is emitted from this battery terminal and cancels
the hole. At the same time, an electron from one of the covalent bonds
enters the positive terminal, thus forming another hole in the vicinity
of the positive terminal. The new hole again moves towards the nega
tive terminal. Thus the battery causes a continuous stream of holes to
flow through the crystal. Insofar as the flow of current is concerned,
hole flow from the positive to the negative terminal of the crystal has
the same effect as electron flow from the negative to the positive terminal.
BASIC SEMICONDUCTOR PHYSICS 5
Transistor Germanium Properties
Impurity Concentration. It is interesting to note the important role
that donor and acceptor atoms play in determining the conductivity of
germanium. If one impurity atom is added for every 100,000,000 germa
nium atoms, the conductivity increases 16 times. This concentration
forms germanium suitable for transistor work. If one impurity atom for
each 10,000,000 germanium atoms is added, the conductivity increases
160 times, and is too high for transistor applications.
Other types of impurities which are neither trivalent nor penta
valent may be present in the crystal. These impurities are not desirable.
Although they do not affect the conductivity, they introduce imperfec
tions in the structure, and cause degradations in the transistor character
istics. Conductivity is affected, however, by the presence of Ntype im
purities in Ptype germanium and by Ptype impurities in Ntype ger
manium, since in either case the holes furnished by the P type will can
cel the electrons furnished by the N type. If both N and P types were
present in equal amounts, the germanium would act as if no impurities
were present. To avoid these possibilities, the germanium is purified so
that the impurity ratio is considerably less than 1 part in 100,000,000 be
fore the desired impurity atoms are added.
Intrinsic Germanium. In those cases where the germanium is ex
tremely pure, or where there are equal numbers of donor and acceptor
atoms, the germanium is called intrinsic. Conduction can take place if
electrons are forced out of their valence bonds by the addition of external
energy to the crystal in the form of heat or light. Although the disruption
of the covalent bonds by these processes creates equal numbers of elec
trons and holes, intrinsic conduction is invariably of the N type, because
the mobility of the electrons is approximately twice as great as that of
the holes.
In the case of thermal excitation, the higher the temperature, the
greater the number of electrons liberated and the higher the germanium
conductivity becomes. This explains why germanium has a negative
temperature coefficient of resistance, i.e., the higher the temperature,
the lower the resistance. Intrinsic conductivity can adversely affect im
puritytype conductivity. As the temperature is increased to 80° C, the
electrons produced by thermal excitation cause the conductivity of the
germanium to become too high for satisfactory transistor operation.
The disruption of covalent bonds by the addition of light energy
is discussed under PN junction photocells in Chapter 2.
PN Junctions
Potential Hills. Alone, either P or Ntype germanium is capable
of bidirectional current flow. This means that reversing the battery will
reverse the direction of the current flow, but will not affect the magni
tude of the current. When P and Ntype germanium are joined as shown
FUNDAMENTALS OF RESISTORS
r+i
1 «
1
P 0*0 t *©
+
+ t ©
+ + +
© t ♦©
© " © © .©"
© ©©©
© ©_©_"©
+ HOLES
 ELECTRONS
ACCEPTORS
DONORS
t
h +
0* © t ©
'e/e © ©
© © © ©
© © ©_©H
© © © ©
+ HOLES
 ELECTRONS
ACCEPTORS
© DONORS
Fig. 14. PN junction at equilibrium.
Fig. 15. PN junction with reverse bias.
in Fig. 14, an effective rectifying device is formed. The junction, desig
nated ab, is called a PN junction. In the illustration the f and — signs
represent holes and electrons, respectively; and the f and — signs with
circles around them represent the donor and acceptor atoms, respectively.
It might appear that the holes of the Pregion would diffuse into the
Nregion and the electrons of the Nregion would diffuse in the Pregion,
eventually destroying the PN junction. Instead, the holes and electrons
concentrate away from the junction. This phenomenon is caused by the
fixed position of the donor and acceptor atoms in the crystal lattice struc
ture, as compared to the mobility of the electrons and holes. The donor
atoms repel the holes to the left in the diagram, while the acceptor atoms
repel the electrons to the right. This barrier to the flow of holes and
electrons is called a potential hill, and it produces the same effect as a
small battery (shown dotted in Fig. 14) with its negative terminal con
nected to the Pregion and its positive terminal connected to the Nregion.
To use the PN junction as a rectifying device requires connection of
an external battery to either aid or oppose the equivalent potential hill
battery.
Reverse and Forward Bias. The connection of an external battery,
illustrated in Fig. 15, is an example of reverse bias. The negative termi
nal attracts holes and concentrates them further to the left, while the
positive terminal concentrates the electrons further to the right. There
is no flow across the junction, since the effect of this connection is to
increase the potential hill barrier.
Consider now the connection illustrated in Fig. 16 (A) . This is an
example of a forward bias connection. The positive terminal pushes the
holes towards the Narea, while the negative terminal forces the electrons
toward the Parea. In the region around the ab junction, holes and elec
trons combine. For each combination, a covalent bond near the positive
terminal breaks down, and the liberated electron enters the positive ter
minal. This action creates a new hole which moves toward the Nregion.
Simultaneously, an electron enters the crystal through the negative bat
tery terminal and moves toward the Pregion. The total current (I ) flow
ing through the crystal is composed of electron flow (I N ) in the Narea,
BASIC SEMICONDUCTOR PHYSICS 7
hole flow (I P ) in the Parea, and a combination of the two (I N and I P )
in the region near the junction. The forward bias connection, then, re
duces the potential hill by a sufficient amount to allow current to flow
by a combination of hole and electron carriers, as illustrated in Fig.
16 (B).
One may well ask, "How much battery voltage is necessary?" Offhand,
since the equivalent battery potential is in the neighborhood of a few
tenths of a volt, an external battery of equal value should normally be
considered sufficient. Unfortunately, a large part of the battery potential
© © © t ©
p + © + © + © ©
© *© + © + V
© ©"© ©
+
©_ © © © .
© © © ©"
G
<5
b
1 1M=
HOLES
ELECTRONS
ACCEPTORS
DONORS
(A)
Ip+I„
(B)
Fig. 16. (A) PN junction with forward bias. (B) Carrier conduction in PN junction.
is dropped across the resistance of the P and Nregions before the poten
tial hill is reached. The voltage drop in these regions is proportional to
the current flow through them; as the current increases due to the re
duction of the potential hill, the drop across the P and Nregions also
increases, leaving even less of the external voltage available to reduce
the junction barrier potential. An external battery of approximately
one to two volts is required because of these factors.
Chapter 2
TRANSISTORS AND THEIR OPERATION
In this chapter, the basic concepts concerning P and Ntype ger
manium are applied to an analysis of the pointcontact, the junction, and
certain other modified transistors. The construction, operation, gain, and
impedance characteristics of typical transistors are considered.
PointContact Transistors
Construction and Electrode Designations. The elements and basic
construction details of the pointcontact transistor are shown in Fig. 21.
This transistor consists of two electrodes (emitter and collector) which
which make contact with a germanium pellet, and a third electrode (the
base) which is soldered to that pellet. (It is common practice to desig
nate the electrodes by e, c, and b— emitter, collector, and base. The prac
tice will be followed in this book.) The entire assembly is encased in a
plastic housing to avoid the contaminating effects of the atmosphere.
The pellet is usually Ntype germanium, roughly .05 inch in length
and .02 inch thick. The emitter and collector contacts are metallic wires,
approximately .005 inch in diameter and spaced about .002 inch apart.
These contacts are frequently referred to as "cat whiskers." The bend
in the cat whiskers, illustrated in Fig. 21, is required to maintain pres
sure against the germanium pellet surface. The practical man will cer
tainly ask, "Why use cat whiskers which are obviously difficult to manu
facture and which produce a mechanically weak contact? Let us eliminate
the cat whiskers (he goes on) and use a lowresistance soldered contact
similar to that used on the base electrode." An answer to this question
necessitates an analysis of the pointcontact transistor. Transistor opera
tion requires an intense electric field. If the external battery potential
is made high enough to produce the required field intensity, this poten
tial has adverse effects on the transistor. The high voltage, in the input
or emitter circuit, produces a high current which burns out the tran
sistor. In the output or collector circuit, the high voltage causes a break
down. Thus, since the battery voltage is limited, as shown by the con
siderations, the use of the pointcontact cat whiskers is a convenient
method of obtaining the required highintensity field. The electrical ac
tion of the points in concentrating the battery potential to produce a
concentrated electric field is analagous to the increased water pressure
which is obtained by decreasing the nozzle area of a garden hose.
SurfaceBound Electrons. The fundamental concepts of current flow
in the pointcontact transistor are illustrated in Fig. 22. Physicists have
found that those electrons which diffuse to the surface of the germanium
pellet not only lose their ability to return to the interior of the germa
nium but also form a skinlike covering over the surface. Because of this
8
TRANSISTORS AND THEIR OPERATION
CAT
WHISKERS
EQUIVALENT
POTENTIAL
HILL BATTERY
DONOR ATOMS
ELECTRONS
ITTER
COLLECTOR
BASE
Fig. 21 (left). Construction of pointcon
tact transistor. Courtesy CBSHyfron.
Fig. 22 (right). Basic pointcontact tran
sistor operation.
phenomenon, they are called surfacebound electrons. For the Ntype
transistor illustrated, the surfacebound electrons combine with the layer
of donor atoms just below to form a potential hill.
The proper battery connections for a transistor can be determined
as follows: The emitter is always biased in the forward or low resistance
direction. Since this is accomplished by reducing the potential hill, the
positive battery terminal is connected to the emitter. Conversely, the col
lector is always biased in the reverse, or highresistance, direction. There
fore, the negative battery terminal is connected to the collector in order
to increase the potential hill.
Hole Injection. To understand hole and electron flow in the point
contact transistor, observe that in Fig. 23 the surfacebound electrons
near the emitter contact are immediately removed by the positive emitter
electrode. This is due to the intense emitter field which breaks down co
valent bonds of atoms in the vicinity of the emitter electrode. The lib
erated electrons are immediately attracted to and enter the emitter ter
minal. These electrons are the emitter current carriers. For every electron
which leaves the pellet, a hole is left behind. This creation of holes is
called hole injection, since the effect is the same as if holes were injected
into the transistor through the emitter. The holes immediately diffuse
toward the collector because of the negative potential at that terminal.
The need for the extremely close spacing between the emitter and
collector is now apparent. Many of the holes may meet with and be
cancelled by the free electrons in the Ntype material. Therefore, the
flow path between the emitter and collector must be small to keep the
hole and electron recombinations to a minimum.
At the collector electrode, the potential hill produced by the sur
facebound electrons limits the current flow. However, holes that reach
10
FUNDAMENTALS OF TRANSISTORS
Fig. 23 (top). Magnified view of hole and
electron flow into pointcontact transistor.
Fig. 24 (bottom). Ntype and Ptype
pointcontact transistor connections.
SURFACE BOUND
ELECTRONS
REMOVED BY
POSITIVE EMITTER}
I
1
©©©©©©
J_i
© 00N0R ATOMS
StSNAL
the collector area combine with the surfacebound electrons and reduce
the potential hill. This permits the collector to inject more electrons into
the germanium, thus increasing the collector current.
Holes travel through the transistor from emitter to collector in many
indirect paths. The holes set up a net positive space charge in the areas
of their flow paths, due to the combined effects of their positive charges.
The resultant positive space charge attracts electrons from the more
remote areas of the Ntype transistor into the hole flow path between
the collector and base, thus effectively increasing the electron flow. While
some of the electrons emitted by the collector neutralize holes, the ma
jority flow toward and enter the base terminal. The electrons which
flow between the collector and the base are the collector current carriers.
Current, Resistance, Voltage, and Power Gains. In the average point
contact transistor, an increase in emitter current of one milliampere will
cause an increase in collector current of 2.5 milliamperes. In physical
terms, this indicates that one million holes injected by the emitter causes
2.5 million electrons to be injected by the collector. One million of the
collector electrons neutralize the holes. The remaining million and a half
electrons flow to the base.
The ratio of change in collector current to change in emitter cur
rent is called the current gain a (Alpha) . Thus
where a — current amplification, i e — change in emitter current, and
i c = resulting change in collector current.
In the typical case described above, a = — i = 2.5.
;r 1 ma
At first glance, the current gain factor of a transistor is disappoint
ingly low when compared with the amplification factor of a vacuum
TRANSISTORS AND THEIR OPERATION
11
tube. However, another consideration enters the picture: The input
resistance between the emitter and base is relatively low (300 ohms is a
typical value) , while the output resistance between collector and base
is relatively high (20,000 ohms is typical) . Thus, in addition to the cur
rent gain, the transistor has another gain characteristic, namely the
ratio of output resistance to input resistance. For the typical pointcon
. . 20,000
tact transistor, the resistance gam is — sku — = 67.
300
Since the input voltage is the product of the emitter current and
the input resistance, and the output voltage is the product of the col
lector current and the output resistance, the transistor voltage gain equals
the current gain times the resistance gain.
e.i ii»r ft r ft
Voltage gain =
x c x o
e, 1.J, r 4
where: ei = input voltage, e = output voltage,
i e = emitter current, i c = collector current,
a = current gain, r = output resistance, and
r, = input resistance.
For the typical case under consideration, the voltage gain equals
2.5 x 67 = 167.5. Furthermore, since the input power is the product of
the input voltage and the emitter current, and the output power is the
product of the output voltage and collector current, the transistor power
gain equals the current gain squared times the resistance gain.
Power gain =5£_ = a r °(? c> \ = tt 2 5l_
e,i e M,W ii
For the typical transistor, the power gain equals (2.5) 2 (67) = 419.
PType Transistor. The Ptype pointcontact transistor operates sim
ilarly to the Ntype unit, except that the emitter and collector battery
polarities are reversed. Fig. 24 illustrates the essential difference between
the battery connections for the two types of pointcontact transistors.
EUNTTOI
(SOUXREO
MTM. C*SE VACUUM
S£ALS omm unit
•MMIST MOISTURE
fa r*h
n
t
© ©. ©
©
© © ©
© ©©
©
© © ©
© ©"©"
N
©
P
© © ©
  _ . N
© DONOR ATOMS
ACCEPTOR ATOMS
 ELECTRONS
+■ HOLES
5
n
j
Fig. 25 (left). Construction of basic NPN
junction transistor. Courtesy CBSHyfron.
Fig. 26 (right). Basic NPN transistor.
12 FUNDAMENTALS OF TRANSISTORS
Junction Transistors
Construction and Operation. In Chapter 1, it was observed that a
combination of P and Ntype germanium form a PN junction. In ef
fect, this combination produces a germanium diode. The germanium
diode has been incorporated in television circuits for several years to
serve as second detectors, and has been used in other circuits where its
excellent rectifying characteristic is useful.
Consider now the effect of combining two germanium diodes into
one unit and, for further simplification, make the Ptype section com
mon to both. This new device, illustrated by Fig. 25, is the basic NPN
junction transistor. In actual construction, the P section is very narrow
as compared to the strips of Ntype germanium. As will be seen later, the
narrow middle section is required for proper transistor operation.
While the pointcontact transistor required the relatively highre
sistance point contacts used for the emitter and collector electrodes, all
junction transistor electrodes . are soldered to their respective sections,
and make low resistance contact. The designations of the electrodes are
the same as for the pointcontact transistor, namely: the emitter (biased
for forward or high conductivity) , the collector (biased for reverse or
low conductivity) , and the base (which connects to the common P
juncti on area) .
Although the junction transistor is a physical combination of two
germanium diodes, conduction in the transistor is decidedly different
from that in the diode or pointcontact transistor. Dbserve that in Fig.
26 the negative potential at the emitter electrode pushes the free elec
trons towards the PN junction. At the junction, as discussed in Chapter
1, a potential hill is set up by the action of the fixed donor and acceptor
atoms. Since the emitter battery acts to flatten this emitterbase potential
hill, a number of electrons pass this barrier and enter the P base region.
The number of electrons crossing the barrier is proportional to the value
of emitter battery potential. Some of these electrons combine with holes
in the P base region, but most pass through and enter the N collector
region. The loss of electrons in the P base region remains low (approxi
mately five percent) because: (1) the base section is thin, and (2) the
potential hill at the collectorbase junction acts to accelerate the electrons
into the N collector region. In the N region, the electrons are attracted
to the positive collector.
PNP Transistor Operation. A PNP junction transistor, shown in
Fig. 27, is formed by sandwiching a thin layer between two relatively
thick P areas. As in the case of the NPN junction transistor, the elec
trode on the left is designated the emitter, the electrode on the right is
designated the collector, and the common electrode is designated the base.
However, the polarities of the potential hills formed are opposite those
formed in the NPN junction transistor. In order to adhere to the general
TRANSISTORS AND THEIR OPERATION
13
rules of biasing the emitter in the low resistance direction, and biasing
the collector in the high resistance direction, the polarities of the ex
ternal bias batteries are also reversed.
Conduction in the PNP junction transistor is similar to that in
the NPN type. The holes in the P emitter region are repelled by the
positive battery electrode toward the PN junction. Since this potential
hill is reduced by the emitter bias, a number of holes enters the base N
area. A small number of holes (approximately five percent) is lost by
combination with electrons within this area, and the rest move toward
the collector, aided by the action of the collectorbase potential hill. As
each hole reaches the collector electrode, the collector emits an electron
to neutralize the hole. For each hole that is lost by combination within
the base or collector areas, an electron from one of the covalent bonds
near the emitter electrode enters that terminal, thus forming a new hole
in the vicinity of the emitter. The new holes immediately move toward
the junction area. Thus a continuous flow of holes from the emitter to
collector is maintained. It is evident that in both types of junction tran
sistors discussed, the collector current is less than the emitter current by
a factor proportional to the number of holeelectron recombinations that
take place in the base junction area.
This analysis of the junction transistor leads to the following gen
eral observations:
1. The major current carriers in the NPN junction transistors are
electrons.
2. The major current carriers in the PNP junction transistors are
holes.
3. The collector current in either type of junction transistor is less
than the emitter current because of the recombinations of holes and elec
trons in the base junction area. As an example, a typical rate of recom
bination is five percent. If an emitter current of one milliampere is as
sumed, the collector current is 1 ma. — (1 ma. x .05) = 0.95 ma.
Gain Factors. Since the current gain (a =Aof the junction tran
sistor is always less than one, it might be expected that its voltage gain
will be less than that of the typical pointcontact transistor. In an actual
case, however, the voltage gain of the junction transistor is considerably
II
I
H
i
i r
l
®* ®1® +
© + © © +
+ + +
© © ©
©
©"
©~
N
©*© + ©
© + ©*©
+ + + +
© © ©
+ + p
© DONOR ATOMS
ACCEPTOR ATOMS
 ELECTRONS
+■ HOLES
Fig. 27. Basic PNP transistor.
fl
I
14
FUNDAMENTALS OF TRANSISTORS
larger than that of the pointcontact type. Since the voltage gain is the
product of the current gain and the resistance gain, it must be expected
that the resistance gain is large. Typical values of emitter input and col
lector output resistances are 500 ohms and 1 megohm, respectively. Thus,
the voltage gain
^=m( 1 ^°)= 1,900
r, \ 500 /
VG =
PG
•>&) wn^r ■•
The high voltage and power gains of the junction transistor as com
pared to the pointcontact transistor are due primarily to the high col
lector resistance.
Transistor Comparisons. To understand the factors which cause the
relatively high collector resistance of the junction transistor as compared
to the relatively low collector resistance of the pointcontact transistor
requires the aid of typical collector current voltage characteristics. Figure
28 (A) illustrates the "VVI,, characteristic of a typical pointcontact tran
sistor. As the collector voltage is raised above 5 volts, the current con
tinues to increase, although at a diminishing rate due to the lack, of
available electrons in the transistor. As discussed previously, the holes
in the pointcontact transistor set up a positive space charge in the vici
nity of their flow paths, attracting electrons from the more remote areas
of the pellet. Thus, the electrons available for collector current decrease
gradually.
Figure 28 (B) illustrates the V c I c characteristic of a typical junction
transistor. Here the V c I c characteristic again follows an Ohm's law re
lationship at small values of collector voltage. The point of electron
exhaustion is reached very abruptly, since there is no hole spacecharge
effect in the junction transistor to increase the available supply of elec
trons. After the critical voltage is attained, a large increase in collector
2 4 6
COLLECTOR CURRENTI,. (MILLIAMPERES)
(A)
1.0 2.0
COLLECTOR CURRENT I c (MILLIAMPERES)
(B)
Fig. 28. (A) Typical pointcontact V c I c characteristic. (B) Typical junction
V c I c characteristic.
TRANSISTORS AND THEIR OPERATION 15
voltage causes only a very small increase in collector current. The col
lector resistance is equal to the change in collector voltage divided by
the resulting change in collector current: r c = — A. For the typical
junction transistor characteristic illustrated, the collector resistance from
05
point A to B is ' ■ n , — 250 ohms, and from point B to C the col
2x 10 3
50
lector resistance is ^ — , n „ = 1,000,000 ohms.
.05 x 10~ 3
Because of the large collector resistance, and the resultant high re
sistance gain, the junction transistor is capable of far greater voltage and
power gains than the pointcontact types. Commercially available tran
sistors with collector resistances in the neighborhood of 3 megohms are
common; silicontype junction transistors are inherently capable of far
greater values.
The basic transistors have an upper frequency limit due to the
small but finite time it takes the current carriers to move from one elec
trode to another. This limit, called the "alpha cutoff frequency," de
fines the point at which the gain is 3 db down from its low frequency
value. The frequency response characteristics of transistors are considered
more fully in Chapter 7.
In this chapter and those that follow, the germanium pointcontact
and junction transistors are considered at great length. This is not in
tended to create an impression that the entire field of semiconductors
is limited to these two fundamental types. However, since their charac
teristics are basic to other semiconductor devices, a thorough understand
ing of the prototypes is essential. At this time, it appears that an un
limited number of variations of the original transistors is possible. Sev
eral of the more significant devices will now be considered.
The PN Junction Photocell
Figure 29 (A) illustrates the essential construction and connections
for the PN junction photocell. The photocell is connected in series with
a battery and a load resistor. The cell is biased by the battery in the
reverse direction. Under these conditions, and with no light striking the
PN junction, approximately ten microamperes of current flow. The cur
rent value is low at this time because of the high resistance of the junc
tion. However, when light strikes the PN junction, the load current in
creases at a rate proportional to the light intensity.
These characteristics are illustrated by the typical operating curves
shown in Fig. 29 (B) . Notice that increasing the voltage from 20 to 100
volts, while holding the light constant, increases the current by less than
10 microamperes. However, increasing the light intensity from 3 to 6
millilumens increases the current approximately 100 microamperes.
16
FUNDAMENTALS OF TRANSISTORS
PN
— JUNCTION
ELEMENT
HI
LOAD<
RESISTOR <
(A)
12 MILLILUMENS
9 MILLILUMENS
6 MILLILUMENS
3 MILLILUMENS
DARK
=!=
zo 40 eo eo ioo
BATTERY VOLTS
(B)
Fig. 29. (A) PN junction photocell construction. (B) Typical junction photocell
operating curves.
Basically, light from any source is composed of tiny particles of
energy called photons. Thus, when light strikes the PN junction element,
in effect photons are bombarding the surface of the element and their
energy is being absorbed by the germanium. The total energy absorbed
is sufficient to disrupt some of the covalent bonds in the element, thereby
creating free electrons and holes in the germanium, and increasing the
number of available current carriers. When the light is removed, the
current decreases rapidly because of the recombination of holes and
electrons.
WideSpaced Transistors
It was noted previously that the emitter and collector contacts of
the pointcontact transistor must be closely spaced for normal transistor
action, since the functioning of this transistor requires an intense elec
tric field. In addition, the frequency response of this transistor decreases
rapidly with increased contact spacing. Theoretically, the frequency op
erating band varies inversely as the cube of the contact spacing. In spite
of this, it has been found that widespaced transistors have some novel
and useful characteristics. When germanium having lower conductivity
(fewer impurity atoms) is used, an increase in the normal contact spac
ing from .002 inch to as much as .015 inch has no effect on the transistor
current and power gains. At the same time, the effect of the emitter volt
age on the collector current is decreased, due to increased spacing.
The ratio of change in emitter voltage to the resulting change in
collector current defines the backward transfer or feedback resistance.
The feedback resistance in a transistor acts similarly to the positive feed
back parameter in a vacuum tube circuit. (The feedback resistance and
other related transistor characteristics are considered in greater detail in
Chapter 3.) The feedback resistance of a transistor with a normal .002
inch contact spacing is about 200 ohms. This resistance is reduced to
approximately 50 ohms when the contact spacing is increased to .015
TRANSISTORS AND THEIR OPERATION
17
inch. This low value insures circuit stability at relatively high values
of power gain.
Germanium with higher than normal resistivity is used to compen
sate for the narrowing of the usable frequency limits by the wide contact
spacing. Despite this, the usable frequency range is reduced to about 1/50
of its normal value. The increased contact spacing has little effect upon
other transistor characteristics.
The PNPN Transistor
Figure 210 illustrates the construction of the PNPN junction tran
sistor. This transistor, unlike the PNP or NPN junction transistor, is
capable of a current gain. For satisfactory operation, both of the central
P and N regions must be narrow.
In operation, the holes move in the direction from emitter to col
lector, but are trapped by the third potential hill in the collector area.
The holes pile up at this barrier, and their cumulative positive space
charge reduces the effect of the potential hill. As a result, electrons from
the collector area encounter a decreased resistance at the junction and
are able to flow into the central P region. Some electrons are lost through
combinations with holes, but most of them, aided by the action of po
tential hill number 2, move into the middle N region and enter the base.
The PNPN construction, because of the space charge effect of the
holes, allows the current gain to reach values in the vicinity of 20. In
comparison it must be remembered that the current gain of the proto
type junction transistor is inherently limited to values less than one.
Transistor Tetrode
The frequency response of the conventional junction transistor is
limited by several factors. First, the frequency cutoff (the frequency at
which the current gain drops sharply) is inversely proportional to both
the base resistance and to the square of the thickness of the junction
layer. In addition, the frequency cutoff is also inversely proportional to
the collector junction capacitance, considered only at high frequencies.
Figures 211 (A) and 211 (B) illustrate the structural and symbolic repre
sentations of the junction transistor tetrode.
In this transistor, a fourth electrode, designated as b 2 , is included.
The fourth electrode is connected to the P junction layer in the same
manner as the conventional base electrode, but the connection is made
on the opposite side of the layer. The base resistance is reduced sub
HILL HILL
#1 #2
Fig. 210.
transistor
Basic PNPN
construction.
18
FUNDAMENTALS OF TRANSISTORS
N P N
b
Fig. 211. Tetrode junction
transistor: (A) structural
representation, (B) symbolic
representation.
(A)
(B)
stantially when a negative bias is applied to the second base electrode.
The bias prevents that part of the emitter junction which is near b 2 from
emitting electrons into the P layer. Thus all of the transistor action takes
place near the base. This effectively reduces the base resistance; as a re
sult, the frequency response increases.
For proper operation, the second base electrode is biased to about
—6 volts with respect to the base. The resulting bias current is approxi
mately one milliampere. In a typical case this bias reduces the base re
sistance from 1,000 to 40 ohms; the change in emitter resistance is neg
ligible. The current gain is reduced from .95 to .75, and the collector
resistance is reduced from 3.0 to 1.5 megohms. The frequency response
cutoff is increased from 0.5 to 5 megacycles. Thus, an increased band
width is obtained at the expense of lower available gain.
The effect of the junction area thickness is decreased by using very
thin P layers (roughly .0005 inch) . The collector junction capacitance
is reduced by decreasing the collector junction area.
Chapter 3
THE GROUNDED BASE TRANSISTOR
This chapter deals with basic fourterminal analysis in general, and
the specific application of fourterminal network analysis to the tran
sistor. Hence, the important characteristics of the transistor, including the
opencircuit parameters, the current gain, the voltage gain, the power
gain, and the conditions for image input and output resistance match
are derived. The basic principles and connections for measuring transis
tor characteristics are discussed, and a comparison between the transistor
and the electron tube is considered.
While the mathematics involved in the analysis of the transistor has
been held to a bare minimum, some readers may be dismayed at what
appears to be an excessive number of derivations. It cannot be overem
phasized, however, that a thorough understanding of the transistor re
quires a general knowledge of the mathematical analysis leading to the
major design formulas. These important design equations are noted by
an asterisk (*) .
FourTerminal Networks
In all types of engineering circuit design, it is frequently convenient
to represent a device by an electrical equivalent. This invariably eases
the task of optimizing the design, since the device is, in effect, reduced
to a simpler equivalent form. One of the most useful methods of equiva
lent representations is by means of the fourterminal network.
The fourterminal network (also called a coupling network, or two
terminal pair network) is shown in Fig. 31. Terminals a and b represent
the input to the network and terminals c and d the output. The network
itself, which represents the equivalent of a device or any combination
of devices, is located between the input and output terminals, and is
considered sealed, so that electrical measurements can be made only at
the input and output terminals.
The sealed network may be, and often is, very complex. As an
example, consider the case of relating the acoustical input to a micro
phone in a multilink transmission circuit to the acoustical output of a
receiver. This system involves transmission lines, electronic circuts, acous
tical, electrical, and mechanical power and transducers. In the fourtermi
nal method of analysis, however, the complete intermediate system be
tween the microphone input and the receiver output is represented by
the sealed box.
The advantage of this type of representation is that only one basic
analysis of a particular device or system is required. Once accomplished,
problems involving the same system or device are a matter of routine
and become simple substitutions of numbers. For electronic devices, other
19
t
\
E
fz
\
1
20 FUNDAMENTALS OF TRANSISTORS
Fifl. 3). Fourterminal network, conven
tional designation.
advantages are that the basic equivalent circuit can be modified to in
clude the effects of highfrequency operation, and that the equivalent
circuit invariably contains a minimum number of parameters which
can be directly related to external measurements.
Fourterminal networks are divided into two general classifications:
active and passive. Passive networks are those that contain no source of
energy within the sealed box; currents and voltages within the box are
a result of the application of energy to the external terminals. Examples of
passive networks include filters, attenuators, and transmission lines. Ac
tive networks, on the other hand, do contain internal sources of energy.
Examples of these, therefore, include all types of amplifying devices, in
cluding the transistor. Although the conventional transistor has but three
external connections, fourterminal network analysis is applicable because
one of the electrodes is common to both the input and output circuits.
The performance of the transistor can be completely defined by the
voltage and current measured at the input and output terminals. Actually
only two of the four values are independent, because if any two are
specified, the other two values are automatically determined. This situa
tion is exactly the same as that in the conventional triode electron tube,
where the four values are the grid current, grid voltage, plate current,
and plate voltage. The grid and plate voltages of a tube are usually con
sidered the independent variables, and their respective currents then
become the dependent variables.
General FourTerminal Network Analysis
The general fourterminal active network is fully described by the
relationship between the input and output currents and voltages. Refer
ring to Fig. 31, the general voltage (loop) equations are:
Ei = ZjjIj + Z 12 I 2
and E 2 = Z 21 Ij + Z M I 2
where Z n is the input impedance with the output open.
Zn = Ej/Il when I 2 = 0.
Z 12 is the feedback or reverse transfer impedance with the input
open.
Z12 = Ei/I 2 > when Ij — 0.
Z 21 is the forward transfer impedance with the output open.
Z 21 = E 2 /I 1( when I 2 = 0.
Z 22 is the output impedance with the input open.
Z 22 = E 2 /I 2 , when I x = 0.
THE GROUNDED BASE TRANSISTOR 21
The equivalent current (nodal) equations are
Ij = Y n Ei \ Y 12 E 2
and I 2 = Y 21 Ei + Y 22 E 2
where Yji is the input admittance with the output shorted.
Y n = Ij/El where E 2 = 0.
Y 12 is the feedback or reverse transfer admittance with the
input shorted.
Y 12 = Ii/E 2 , when Ei = O.
Y 21 is the forward transfer admittance with the output shorted.
Y 2 i = I 2 /E 1( when E 2 = O.
Y 22 is the output admittance with the input shorted.
Y 22 = I2/E2. when E t = 0.
Amplification factors are the best general index of an active net
work. Since the general case may have amplification in both directions,
definitions are included for forward and reverse directions.
The forward current amplification factor, a 2 i, is equal to the nega
tive ratio of the current at the shorted output terminals to the current at
the input terminals.
~t&
a 2 i = — t— = — Iwhen E 2 =
Then = E 2 = Z^ + Z 22 I 2 .
Solving these equations 021 = — I . 2 ) = „ 21 and in terms of admittance
\ lj / Z 22
The reverse current amplification factor, oi 2 , is equal to the negative
ratio of the current at the shorted input terminals to the current at the
output terrminals:
ai2 = — ( j 1 ) » when E a =
Then = Ei = Z^ + Z 12 I 2 .
Solving as before, an —— I t 1 ) = 12 , and in terms of admittances
The forward voltage amplification factory ^ 21 , is equal to the ratio of
open circuit output voltage to the input voltaj
I 2 = 0. On this basis, Ei = Z u Ij and E 2 = Z 21 Ii.
the open circuit output voltage to the input voltage. ^ 2 i = p 2 , when
22
FUNDAMENTALS OF TRANSISTORS
Thus ,12!
and on an admittance basis ^ 2 i
VW'
Ej Z n
The reverse voltage amplification factor ^ 12 is equal to the ratio of
£
the open circuit input voltage to the output voltage. ^, 12 = ^= — when
Ei
Z] 2
22
I t = 0. Then E x = Z 12 I 2 and E 2 = Z 22 I 2 . Thus ,n 2
In terms of admittance ^ 12 = — [ 12 1
VacuumTube Analysis on a FourTerminal Basis
Figure 32 (A) illustrates the familiar case of a conventional ground
edcathode triode operated at low frequencies with its control grid biased
sufficiently negative so that no grid current flows. (It should be noted
at this point that the current arrows in this diagram and those that follow
indicate the direction of electron flow.) The applied grid signal causes a
voltage ^e g to appear in series with the plate resistance r p . Since the grid
current i g is zero, the network is completely described by a single
equation:
P e g
+ e D
The fourpole equivalent network for this same circuit when the
grid draws current is illustrated by Fig. 2 (B) . In this case, the grid
voltage acts across a series circuit consisting of the voltage ^,e p and
the grid resistance r g . The term ^ equals the reverse voltage amplifica
tion factor: ^ p =
As in the previous case, the grid signal voltage
causes a voltage ^e g to appear in series with the plate resistance. Since
there are two voltage loops in the case when the grid is driven positive,
two equations are required to describe the network completely; these are
!»£* = Vp + e p
e g = i^g + rt> e p
This analysis of triode vacuum tubes on a fourpole basis is not
limited to the groundedcathode operation of these tubes. The choice of
this type of operation is dictated on the basis of reader familiarity with
OGRIO
O PLATE
(B)
Fig. 32. Equivalent circuit of a triode: (A) with negative grid bias, (B) with
poiitive grid bras.
THE GROUNDED BASE TRANSISTOR
23
#
(A)
Fig. 33. Fourterminal network ground connection: (A) cathode, (B) grid,
(C) plate.
the circuit. The groundedgrid and groundedplate connections (which
have useful counterparts in transistor circuitry) may be analyzed in simi
lar fashion. The basic fourterminal currentvoltage relationships for all
three cases are illustrated in Fig. 33.
Vacuum Tubes Compared with Transistors
Representation of a vacuumtube circuit by an equivalent circuit
which includes its transconductance, amplification factors, plate resist
ance, and grid resistance is particularly useful in design applications.
This treatment greatly simplifies analysis in those applications of the
tube's operating characteristics where a linear approximation is valid. A
similar type of linear analysis is applicable to the operation of the tran
sistor. As will be seen shortly, transistor parameters correspond closely
to tube parameters. The main factor contributing to differences between
tube and transistor characteristics is that the transistor is primarily a
current operated device, while a vacuum tube is a voltage operated
device.
When the grid of a vacuum tube is held negative with respect to
the cathode, only three tube variables exist, since the grid current is
zero. The transistor, however, always has four variables. As a result, four
independent parameters are necessary to specify its characteristics com
pletely. The analysis that follows is based on small signal inputs which
satisfy the requirements of linearity. For this reason, the resulting para
meters are called the smallsignal parameters.
In the transistor, both the input and output currents and voltages
are significant. In addition, it is possible to have two or more sets of
currents for one set of voltages. This situation is somewhat similar to
that existing in a vacuum tube that draws grid current, in which there
may be two possible grid voltages for a given set of grid and plate cur
rents. In the transistor, there can only be one set of voltages for a speci
fied pair of input and output currents. This reason governs the choice
of current as the independent variable in transistor work as opposed to
the choise of voltage in the representation of vacuumtube character
istics.
In vacuum tubes the input grid voltage is plotted against a plate
characteristic, because the output voltage is approximately a linear func
24
FUNDAMENTALS OF TRANSISTORS
tion of the grid voltage. In transistor circuitry, a similar curve can be
formed by plotting the collector voltage as a function of collector current
for a fixed value of input current. Note again that the transistor input
current is selected as the independent variable rather than input voltage.
The groundedcathode vacuum tube is a voltage amplifying device hav
ing a high input impedance and a relatively low output impedance. Its
equivalent transistor circuit, the grounded emitter transistor, is a current
amplifying device with a low input impedance and a relatively high
output impedance.
Several types of equivalent circuits can be used to represent the
transistor under small signal conditions. Figure 34 represents only three
of the many possibilities. The indicated circuits are equivalent in that
they all give the same performance for any given set of input and output
characteristics. Examples (B) and (C) are particularly well suited to
transistor application because the resulting parameters are of significance
in transistor physics. In addition, the parameters are readily measured,
are usually positive, and are not extremely dependent on the exact
operating point chosen. The significance of the impedance parameters
is covered later in the chapter.
The derivations of these equivalent circuits are based on the relation
ship between the input and output currents and voltages. For example,
assume that for the sealed network of Fig. 31 the input and output re
sistances remain constant with frequency and are each equal to 200 ohms.
Then the network may be a shunt resistor equal to 200 ohms (Fig. 35A) ,
a "T" pad of three equal 100ohm resistors (Fig. 35B) , a "pi" pad of
three equal 300ohm resistors (Fig. 35C) , or any other combination meet
Zc Zml,
o— AAA/ » — \AA/ — (~) — o
>z ll z 22<
Mzi 2 i2 ("M^aiii
o o
(A)
(B)
Fig. 34. Typei of fourterminal
equivalent circuit!.
(C)
THE GROUNDED BASE TRANSISTOR 25
o o— AAA/ — t — VW— o
100 100
OHMS I OHMS
INPUT £ 200 OUTPUT INPUT S OUTPUT
RESISTANCE > OHMS RESISTANCE RESISTANCE >i°° RESISTANCE
■ 200 OHMS > • 200 OHMS ■ 200 OHMS > 0HMS ■ 200 OHMS
(A) (B)
AAAr
300
Fig. 35. Examples of four
Ohms I terminal networks.
INPUT > > OUTPUT
RESISTANCE >300 300% RESISTANCE
« 200 OHMS >OHMS OHMS S. • 200 OHMS
O
(C)
ing the required input and output characteristics. The derivations of ac
tive networks are admittedly more complicated than this simple example,
but the basic principles are exactly the same.
FourTerminal Analysis of Transistors
Like the vacuum tube triode, the transistor has useful properties in
any of the three possible connections: grounded base, grounded emitter,
and grounded collector. Most of the present literature starts with the
grounded base connection because this configuration is the most con
venient for describing transistor physics. In circuit work, however, the
grounded emitter connection is most popular because it provides maxi
mum obtainable power gain for a specified transistor and is well suited
to cascading without impedancematching devices. The vacuum tube
counterpart of this circuit, the grounded cathode connection, also pro
duces maximum power gain and is adaptable to cascading without im
pedancematching devices.
Typical characteristics for a junction transistor in grounded base
connection are shown in Fig. 36 (A) . Since the collector current is the
independent variable, it is plotted along the abscissa, in apposition to
the method used in plotting vacuum tube characteristics. Notice the simi
larity between the junction transistor characteristics in Fig. 36 (A) and
those of the typical triode vacuum tube illustrated in Fig. 36 (B) . Based
on this similarity, it is reasonable to assume that the transistor collector
voltage, collector current, and emitter current can be compared with
the plate current, plate voltage, and grid voltage of a triode vacuum tube.
Examining the tube characteristics, it is seen that a signal applied
to the grid shifts the plate voltage along the load line. The numerical
plate voltage shift caused by a change of one volt in the grid voltage is
defined as the amplification factor ^ of the tube. In a like manner, apply
ing a signal to the transistor emitter shifts the collector current along
the load line. The numerical shift in collector current caused by a
26
FUNDAMENTALS Ot TRANSISTORS
i 2 3 4 5
COLLECTOR CURRENTMILLIAMPERES
(A)
50 100
PLATE VOLTS
(B)
Fig. 36. (A) Typical junctiontransistor characteristics. (B) Typical vacuum
tube characteristics.
change in emitter current of one milliampere is defined as the current
amplification factor a of the transistor. The current amplificaton factor
of a transistor, then, corresponds to the voltage amplification factor of
a vacuum tube.
Insofar as input characteristics are concerned, the vacuum tube nor
mally operates with its grid biased in the reverse or high resistance direc
tion, while the transistor operates with the emitter biased in the forward
or low resistance direction. In the output circuits, a similar relationship
exists. The plate of a vacuum tube is biased in the forward direction,
while the transistor collector is biased in the reverse direction. These
biasing conditions produce the high input and low output impedances
in the vacuum tube circuit, and the low input and high output impe
dances in the transistor circuit. This reemphasizes the basic difference
between the vacuum tube and the transistor: the vacuum tube is a volt
age controlled device, while the transistor is a current controlled device.
Equixtalent Passive "T" Network. In the analysis of the transistor
on a fourterminal basis, the entire device is treated as a sealed box with
three external terminals e, b, c, designating the emitter, base, and collector,
respectively. This basic fourterminal network is illustrated in Fig. 37 (A),
in which the input signal is applied between emitter and base. The input
signal E g is taken from a signal generator that has an internal resistance
R g . The output circuit is between the collector and the common base
and consists of a load resistance R L . In the smallsignal analysis which
follows, it is assumed that the transistor is biased in the linear region
of its operating characteristics. It is also assumed that the operating
frequency is low enough so that the transistor parameters may be con
sidered pure resistances, and the capacitive junction effects may be con
sidered negligible.
THE GROUNDED BASE TRANSISTOR
27
The simplest method of approaching the analysis of the equivalent
transistor circuit is by an equivalent "T" network with no internal gen
erating sources (passive basis) . This circuit is illustrated in Fig. 37 (B) .
Under these conditions, the transistor parameters can be completely
specified by the following terminal measurements:
A. Input resistance with output terminals open,
when i c = 0, r u = r e + r b .
B. The forward transfer resistance with the output terminals open
r  e °
when i c = 0, r 21 = r b .
C. The output resistance with the input terminals open,
r  e "
r 22 ■
when i e = 0, r 22 = r c + r b .
D. The backward transfer resistance with the input terminals open
.. _ e,
when i e = 0, r 12 = r b .
Notice that the forward transfer resistance is equal to the backward
transfer resistance. This is typical of a fourterminal passive network. In
the practical case, then, it is only necessary to measure r 12 or r 21 .
Equivalent Active "T" Network. While the passive network serves
as an interesting introduction to transistor analysis, it does not describe
this device completely, because the transistor is known to be an active
o — WV
i
«i
■AAA/ — o
'ii
"e + 'i.
r 2l " r b
' 22 =' c +'b
f l2 ■ r b
(B)
o — VW t VsA/— vM— °
r 2l' r m +r b
'I2 
'22
'»!>* 'r + 'b
(C)
Fig. 37. (A) The basic cir
cuit for transistor fourter
minal network analysis. (B)
Transistor equivalent "T" on
a passive basis. (C) Tran
sistor equivalent "T" on an
active basis.
28 FUNDAMENTALS OF TRANSISTORS
network. The equivalent circuit of the transistor can be represented in
a number of ways; the most widely used configuration is illustrated in
Fig. 37 (C) . The basic difference between the equivalent circuits repre
senting the active network and the passive network is the voltage source
e = r m i e inserted in the collector arm. In general, the passive network
determines three of the characteristics of the active network. In the
case under consideration, the input resistance r u , the output resistance
r 22 , and the backward transfer resistance r 12 are the same for the passive
and active networks shown in Figs. 37 (B) and 37 (C) . The only differ
ence is the value of the forward transfer resistance r 2 i, which in the case
of the passive network equals r b , and in the active network equals
r m + r b .
Thus, the equations for an active network under ideal conditions
(that is, when the resistance of the signal source is zero, and the resist
ance of the load is infinite) become:
ru = r e + r b
i"i2 = r b
r 2i = r b + r m
r 22 = r c + r b
The four parameters in this active network are the emitter resistance
r e , the base resistance r b , the collector resistance r c , and the voltage source
r m i e . The parameter r m is represented as a resistance since it acts as the
proportionality constant between the input emitter current and the re
sulting voltage source in the collector arm. The mathematical logic of
the resistance r m is easily derived as follows: In preceding chapters, the
current gain a was defined as the ratio of the resulting change in collector
current to a change in emitter current, o = — A — . The equivalent volt
age introduced into the collector circuit is e = i c r c . Since i c = oi e » e =
oi e r c  Since a is a dimensionless parameter, it can be related to the col
lector resistance by a resistance parameter r m . Thus, o = — — • Substi
—ldiy. Imltt
tuting this latter equality, e = oi e r c =
There is no phase inversion in the grounded base connection; a posi
tive signal applied to the emitter produces an amplified positive signal
of the same phase at the collector.
Measuring Circuits. Figure 38 illustrates the four basic circuits for
measuring fourterminal parameters. The double subscript designations
on the general resistance parameters of the fourterminal network (those
designated r u , r 12 , r 2 i, and r 22 ) refer to the input terminal 1 and the
output terminal 2. In addition, the first subscript refers to the voltage,
and the second subscript refers to the current. For example, r 12 is the
ratio of the input voltage to the output current, while r 21 is the ratio of
THE GROUNDED BASE TRANSISTOR
29
la
'2
'*'% (C) '"
Fig. 38. Basic circuits for measuring fourterminal parameters.
(D)
the output voltage to the input current. These designations also indicate
whether a test signal is applied to the input or output terminals, since
the current will always be measured at the terminals where the test sig
nal is applied.
There are several sources of error inherent in this use of smallsignal
inputs to evaluate the parameters of a transistor. The first error is due
to the nonlinearity of the characteristic curves. The larger the signal
input, the greater the error. Thus, to make this error negligible, the signal
must be held sufficiently small. In the practical case, the minimum useful
signal is limited by the transistor thermal noise.
A second error results from the internal resistance of the signal
source. In measuring any of the parameters, the amplitude of the input
signal is assumed to be independent of the transistor resistance. However,
this is true only if the source impedance is very much greater or very
much smaller than the transistor input resistance. If a high resistance
source is used, the magnitude of the resulting error is proportional to
the ratio of the transistor input resistance and the resistance of the signal
source. If a low resistance source is used, the error is proportional to the
ratio of the internal resistance of the signal source and the input resist
ance of the transistor.
The third source of error, when small signal inputs are assumed,
results from the shunting action of the input voltmeter and the input dc
bias supply on the input signal. The first effect may be made negligible
by using a very high resistance voltmeter. The magnitude of the error
caused by the shunting action of the dc bias supply is proportional to
the ratio of the transistor input resistance and the dc supply resistance.
Errors are also introduced by the capacities between emitter and base,
and between collector and base. These capacitance effects are comparable
30
FUNDAMENTALS OF TRANSISTORS
to the plate to grid and cathode to grid capacitances in a vacuum tube.
In the audio frequency range, these errors are generally neglected.
The Grounded Base Connection
Equivalent Operating Circuit. At this point, the transistor equivalent
circuit must be considered using a practical circuit, such as illustrated
in Fig. 39. The signal generator E g , having an internal resistance R g , is
connected between the emitter and the base. A load resistance R L is
connected between the collector and the common base. The input cur
rent is designated i u and for the common base connection is equal to
the emitter current i e . The collector output current is designated i 2 . A
cursory look at Fig. 39 makes it fairly evident that the input resistance
r n as seen by the signal generator depends to some extent on the value
of the load resistance R L , and the output resistance r 22 as seen by the load
resistance is determined to some extent by the value of the generator's
internal resistance Rg. On a basis of Kirchoff 's Law, the loop equations
for the circuit of Fig. 39 are:
Input loop 1: E g = i, (R„ f r e f r b ) f i 2 r b Eq. (31)
Output loop 2:  r m i e = i,r b + i 2 (r e f r b rf R L )
Since i e = i 1( then
O = ij (r b f r m ) f i 2 (r c + r b + R L ) Eq. (32)
Since these two loop equations are independent, they may be solved si
multaneously for the two unknown currents i x and i 2 . Then
, Egfrb + rc + RQ
12
(R g + r e f r b ) (r b f r c f R L )  r b (r b + r m )
Eg (r b + r m )
Eq. (33)
Eq. (3J)
(Rg + r e f r b ) (r b f r c + R L )  r b (r b + r m )
Under ideal conditions, namely, when Rg equals zero, and R L is infinite,
it was previously found that r u = r e + r b , r 12 = r b , r 21 = r b + r m , and
r 22 = r e  r b .
If these values are substituted in equations 33 and 34, i t and i 2 can be
evaluated in terms of the ideal or opencircuit parameters.
fm'«
Fig. 39.
Equivalent circuit for grounded
base connection.
Fig. 310. Simplified transistor equivalent
circuit for analysis of input resistance r t .
THE GROUNDED BASE TRANSISTOR 31
. E g (R L + r 22 )
11 ~ (R, + r n ) (R L + r 22 )  r 12 r 21 ^ V*>
• Egr 2 i p ,, xx
12  (R, + r u ) (R L + r 22 )  r 12 r 21 ^ ( ^>
Current Gain. The current gain, a = — * — , when the circuit is work
h
ing into the load R L , becomes the ratio of equation 34 to equation 33
and in terms of the openrcircuit parameters, the current gain is the ratio
of equation 36 to equation 35
The current gain as derived in equations 37 and 38 indicates the effect
of the load resistance R L on a , but does not take into account the effect
of mismatch between the signal generator resistance and the input re
sistance of the transistor. It is evident that the maximum current gain
is obtained when the load resistance R L = O. Thus, the maximum
 = a =^= rb + rm Eq. (38A) •
r 22 r b T r c
Since r b is very small in comparison with either r m or r c , it may be neg
lected in equation 38 A, and a rather accurate estimate of the maximum
current gain is
a = ao=£= Eq. (38B)*
A frequently used form for the current gain, which incorporates
the maximum current gain Oo » is
»= n , Rl Eq. (38C)*
r 22
Input Resistance, r,. The input resistance of the grounded base tran
sistor shown in Fig. 39 can now be computed in terms of the transistor
parameters and the transistor fourterminal opencircuit parameters.
Since the input resistance as seen by the signal generator is r^ Fig 89 may
be simplified as shown in Fig. 310. This series circuit is expressed
E, = i,(R, + r,)
or R^ + r.^J^ Eq. {39)
Substituting equation 33 for i x
R +r  E * £(*» + r ° + fb) (r b + r c + R L )  r b (r b + r m )]
g ' E g (r b + r c + R L )
32
FUNDAMENTALS OF TRANSISTORS
*l* 250 OHMS
WHEN R L °°
r; = 50 OHMS
WHEN R L »0
^^
r M .250 OHMS
r l2  100 OHMS
r a > 24,000 OHMS
r 22 = 12,000 OHMS
Fig. 311.
LOAD RESISTANCE R L (OHMS)
Input resistance vt load resistance for typical pointcontact tran
sistor (grounded base).
then
r, = r„ + r b
and in terms of the opencircuit parameters
r 12 r 21
/ r b (r b + r m ) \
V r b + r c + R L ;
r, = r x
Eq. (312)*
Eq. (313)'
r 22 + Rl
The effect of varying the load resistance on the input resistance can be
best appreciated by examining Figs. 311 and 312, which illustrate the
r, vs R L characteristics for typical pointcontact, and junction transistors,
respectively. For the typical pointcontact transistor, r u = 250 ohms,
r 12 = 100 ohms, r 21 = 24,000 ohms, and r 22 = 12,000 ohms. For the typi
cal junction transistor, r n = 550 ohms, r 12 = 500 ohms, r 21 = 1,900,000
ohms, and r 22 = 2,000,000 ohms. Notice that in the case of the point
contact transistor, the transistor input resistance varies from 50 to 250
ohms as the load resistance changes from zero to infinity. The junction
transistor input resistance varies from 75 to 550 ohms as the load resist
ance is varied from shortcircuit to opencircuit conditions.
Output Resistance, r . The output resistance can be found in a
similar manner. Consider Fig. 313 (A) , which illustrates the equivalent
circuit for analyzing the output resistance. The equations for the two
loops on the basis of Kirchoff's law are:
Loop 1: O = \ (R g + r c + r b ) + i 2 r b Eq. (314)
Loop 2: E 2  r m i e = i x r b + i 2 (r c + r„ + R L ) Eq. (315)
Since i e = i lt then
E 2 = ii (r b + O + i 2 (r c + r b + Rl) Eq. (315 A)
THE GROUNDED BASE TRANSISTOR
33
Solving the two independent equations 314 and 315 A for the unknown
load current,
: (R 8 + r e + r b ) E 2 „
2 (R g + r e +r b ) (r c + r b + R L ) r b (r m + r b ) * ' v '
Looking back into the transistor, the generator E 2 with its internal re
sistance R L sees the output resistance r . Again the circuit may be simpli
fied as shown in Fig. 313 (B) .
Then E 2 = (R L + r ) i 2 or R L + r = — 4 Eq.
(317)
Substituting equation 316 in 317,
» , r _ E * [(R g + r e + r b ) (r c + r b + R L )
L+ ° , M'b + r. + R.)
In terms of the opencircuit parameters
r b (r m + r b ) ]
Eq
Eq,
Eq.
(318)
(319)
(320) •
The latter equation indicates that the output resistance depends to
some extent on the value of the signal generator input resistance. The
variation of r vs R g is illustrated in Figs. 314 and 315 (for the same
pointcontact and junction transistors considered in the preceding sec
tion) . In the case of the typical pointcontact transistor, the transistor
l£» 550 OHMS
^•75
OHMS
AT R L »0
*— ^
r 2 > 500 OHMS
r 22  2 MEGOHMS
L_
1
1
I MEG 10 MEG
LOAD RESISTANCE R L 10HMS)
Fig. 312. Input resistance vs load resistance for typical junction transistor
(grounded base).
34
FUNDAMENTALS OF TRANSISTORS
>r o it
Fig. 313. Analysis of output resistance r : (A) equivalent
circuit, (B) simplified circuit.
output resistance varies from 2,400 to 12,000 ohms as the signal genera
tor internal resistance increases from zero to infinity. The junction
transistor output resistance varies from 270,000 to 2,000,000 ohms as the
signal generator internal resistance increases from zero to infinity.
Voltage Gain VG. Looking again at Fig. 39, it is seen that the
voltage gain VG — 2 . Since E 2 = i 2 R L and E g = i t (R g + r,) ,
Since 
12
a, if equation 38 is substituted for
in equation
ii * ii
322 and if equation 313 is substituted for r, in equation 322 the volt
12 pOO
iopoo
r = 12,000 OHMS
A
1 Rg
E O*.
i
tn
Z
o epoo
Hi
o
5>
OUTPUT RE
*
"I
r„ » 2400 OHH
s
—
t Z 'IOO OHMS
r 2  ■ 24,000 OHMS
T22= 12,000 OHMS
IOO IK
GENERATOR RESISTANCE Rg (OHMS)
Fig. 314. Output resistance vs generator resistance for typical pointcontact
transistor (grounded base).
THE GROUNDED BASE TRANSISTOR
35
2.4
2.2
2.0
1.6
1.4
1.2
1.0
.8
.6
.4
■ Z MEGOHMS 
U Rg"«=
r 27
~ AT Rg
0,000
•0
OHM!
T2> 900 OHMS
<
rj" 1.9 MEGOHMS
.2
T a  2 MEGOHMS
L_
1
1
1
I
10 100 IK
GENERATOR RESISTANCE Rg (OHMS)
Fig. 315. Output rcsittanc* r* gamrator rnbtanc* for typical junction
tramiitor (grounded pom).
age gain becomes:
VG =
r 21^I.
Eq. (323)
Notice that the voltage gain is maximum when R L is infinite and R,
is zero. Under these conditions the maximum
I*L_ Eq. (325)*
VG =
Til
For the typical pointcontact transistor, the maximum VG =
24,000
250
— 96. Assuming typical values of R L = 25,000 ohms, and R, = 200
ohms
(24,000) (25,000)
VG
= 42.1
1,900,000
550
(25,000 + 12,000) (200 + 250) 100(24,000)
For the typical junction transistor, the maximum VG =
3,450. Assuming typical values R L = 1 megohm, and R g = 200 ohms
vr (2,000,000) (1,000,000)
(1,000,000 + 2,000,000) (200 + 550)  550 (1,900,000) '
A comparison of the maximum and operating gains of the typical
pointcontact and junction transistors indicates that the junction is cap
36 FUNDAMENTALS OF TRANSISTORS
able of furnishing much larger voltage gains. This explains why the
junction transistor is invariably used in audio amplifier circuits.
The power gain (PG) of the transistor can be calculated from the
product of the current gain and the voltage gain or found directly
from the ratio of output power to input power.
PG = a(VG)
The theoretical maximum power gain is the maximum current
gain and the maximum voltage gain. However, the condition for maxi
mum current gain is R L = 0, and the condition for maximum voltage
gain is R L = infinity. Since these conditions are in opposition, the
problem of finding the maximum power gain involves matching the
input and output resistances of the transistor. The maximum power
gain is obtained when the internal resistance of the signal generator
is equal to the input resistance of the transistor, and the load resistance
is equal to the output resistance of the transistor, that is R g = i t and
R L = r . When these conditions are simultaneously satisfied, the tran
sistor is image impedance matched.
Input and Output Impedance Matching. Equations 313 and 321
indicate that the input resistance is affected by the load resistance and,
conversely, the output resistance depends on the generator internal re
sistance. Thus, starting with a given load resistance, if the generator
resistance is changed to match the input resistance, the output resistance
of the transistor changes, thus requiring a change in load resistance, and
so on. In the following analysis, the proper values of generator and
load resistance which satisfy both the input and output matching con
ditions at the same time are determined. Let r x equal the proper value
of input resistance and generator resistance. Let r 2 equal the image
matched value for the transistor output resistance and the load re
sistance. Then: r 2 = R s = r t and r 2 = R L = r .
Substituting for R L and r 4 in equation 313
ri = r, = R g = r n  7 r " r " ) Eq. (326)
Solving in terms of r 12 r 21
(ri  r n ) (r 2 + r 22 ) =  r 12 r 21 Eq. (327)
Substituting for R g and r in equation 321
r, = r = Ri = r M  /"*" Eq. (328)
r i "I" r n
Again solving in terms of r 12 r 21
(r 2  r 22 ) (r, + r n ) =  r 12 r 21 Eq. (329)
Equating equations 327 and 329
(ri  r u ) (r 2 + r 22 ) = (r 2  r 22 ) (r x + r n ) Eq. (330)
Cross multiplying and cancelling equal terms,
r l r 2 — r 2 r ll + r l r 22 — r ll r 22 = r l r 2 ~~ r l r 22 + r 2 r ll — ril r 22
2x l r„ = 2tj[ ll Eq. (331)
THE GROUNDED BASE TRANSISTOR 37
Tl r ll
This latter equation indicates that matching the input and output re
sistances for maximum power gain requires their values to be in the same
ratio as the opencircuit characteristics of the transistor.
The absolute value of the generator internal resistance and its
matched input resistance in terms of transistor opencircuit parameters
can now be determined. Substituting the equality r 2 = 1 22 into
r u
equation 326,
_ / r 12 r 21 \ _ / r 12 r 2l r U \
1 " f rir 22 + r 22 J " ^r 22 ( ri + r n ) / Eq. (333)
(ri  r u ) (r x + r n ) =  (^f") = r,»  r n » Eq. (334)
r i a = r n' ( r ^ rU ) E 1' ('■»)
* \ r 22 / ¥ r 22
In terms of the stability factor, 8 = — 12 21 , which will be defined later
r ll r 22
in the chapter, the input image resistance
For the typical pointcontact transistor previously considered, when
r n = 250 ohms, r 12 = 100 ohms, r 21 = 24,000 ohms, and r 22 = 12,000
ohms, the numerical value of r! is
11 "\H2lSoO~t 250 < 12 ' 00 °) ~ 10 ° (24000)] =112 ohms
For the typical junction transistor, when r u = 550 ohms, r 12 = 500
ohms, r 2 j  1,900,000 ohms, and r 22 = 2,000,000 ohm s,
ri =v / 2 ooo° oo E 550 ( 2 ' 000 ' 00 °) ~ 500 (1,900,000)] = 203 ohms
The output image resistance of a transistor can be determined in a
similar fashion from the ratio
'22 . _ _ r 2 r n
Tx — 
r l r ll r 22
Substituting this equality into equation 328
r 12 r 22
r 2 — r 22 ~j
'"*+ r„.
)='»~km) «■*«>
38 FUNDAMENTALS OF TRANSISTORS
(r. ~ r 22 ) (r 2 + r 22 ) = _ f T ^ T ^A = ^ _ r ^ Eq . (3 . 39A)
r 2 * = T 22 * ( ri * ilT *A Eq. (339B)
t2 = J r^^ 2 ^ =^i (r u r 22  r 12 r 21 ) Eq. {340)*
In terms of the stability factor 8 = — 12F21 ; the output image resistance
r 2 = / r 22 * pi^2 li*LJ = r22 Vl _ S Eq. (341)
V/ \ r n r 22 r n r 22 /
For the typic al pointcontact transistor.
r 2 = J 12 '^ [250 (12,000)  1 00 (24,000)] = 5,370 ohms
For the typical junction transistor
" 740,000 ohms
or trie typical junction transistor
r 2 /^^^L^o (2,000,000)  550 (1,900,000)]
These values may be checked on the R L vs r, and R r vs r„ characteristics
plotted for these typical transistors in Figs. 311, 312, 314, and 315.
Negative Resistance and Transistor Stability. Consider the general
expression for input resistance
* = (ot) *■ <™>
It is evident that the input resistance can have a negative value. The
input resistance r, is positive as long as r n is greater than — = — .
r 22 + K L
This condition is most difficult to attain when the output is shorted,
namely when R L = 0. For the transistor to be stable under this condi
tion, r n r 22 must be greater than r 12 r 21 . The stability factor is the ratio
of r 12 r 21 to r u r 22 . The stability factor 8 = 12 21 must be less than
r ll r 22
unity for shortcircuit stability. Substituting the equivalent transistor
parameters for the grounded base connection into the stability equation,
the following relationship is obtained:
riir 22 > r 12 r 21 becomes (r c + r b ) (r e + r b ) > r b (r m + r b ) Eq. (342)
Expanding equation 342,
r^ + r c r b f r b r e + r b 2 > r b r m + r b 2
Dividing through by r b
r c + r e +^>r m Eq. (343)
r b
This equation emphasizes the importance of the backward transfer
resistance r b , since when r b = 0, the transistor must have a positive input
resistance.
On the other hand, if the value of r b is increased by adding external
resistance, it is possible to reach a condition where a normally positive
THE GROUNDED BASE TRANSISTOR 39
input resistance becomes negative. Notice, however, that increasing the
total base resistance eventually causes the input resistance to become
negative only if r e + r c is less than r m . In the case of the junction tran
sistor, r c is always greater than r m , and increasing the base resistance
cannot produce a negative input resistance.
The conditions for negative output resistance are obtained simi
larly. In the general output resistance equation,
r = r 22  (V^ ) Eq. (321)
the output resistance r is positive provided that r 22 is greater than
r 12 r 21
Rg + r u
This condition for stability is most difficult to meet when the
generator resistance is equal to zero. For the transistor to be stable
under this condition, r u r 22 again must be greater than r 12 r 21 . The same
stability factor and equations then exist for both the input and output
resistances. It is evident, then, that one method of fabricating a tran
sistor oscillator is by adding sufficient resistance to the base arm. Typi
cal circuits incorporating this principle will be considered in Chapter 6.
Power Gain. Before determining the power gain included in tran
sistor circuits, some definitions must be considered. Figure 316 illus
strates a signal generator E g with an internal resistance R g feeding into
E 2
a load R L . The total power delivered by the generator P = ? — ;
£ 2 E R
the power dissipated in the load P L = L . Since E L = K T
then P  Eg2Rl '
Rg + Rl
(Rg + RO 2
By using conventional calculus methods for determining conditions for
maximum power, it is found that the load power is maximum when
Rg = Rl Under this condition the power available from the generator
F 2 R F 2
p ———gJ—B— — _zi_
a "(2R g ) 2 4R,
K
The operating gain, G, of a network is defined as the ratio of the
power dissipated in the load to the power available from the generator.
Fig. 316. Simplified rrantistor equivalent circuit for
analysit of power gain.
M
40 FUNDAMENTALS OF TRANSISTORS
For the general transistor circuit of Fig. 39
The power dissipated in the load
E2 2 _r r 2i
i ' RL (R, + r 11 )(R L Vr»)r M r„ ^ { ™ 4)
Pi =
Rr.
(R g + rn) (R L + r 22 )  r 12 r 21
The operating gain
r 21
G = jr = 4R,R,
(Rjr + rn) (R L + r 22 )  r 12 r 2 J
led as flu
E g *R L Eq. (345)
" 2 Eq. (346)*
The available gain, AG, of a network is defined as the ratio of the
power dissipated in the load to the power available from the generator
when the load is matched to the output resistance. When R L = r =
H«r?y d  AO £
Substituting in equation 346, the available gain
4R g (r 22 MpW
AG= ^ ^^ Eq. (347)
[(R, + r„) (r 22  Tl2 *\ + r 22 )  r 12 r 21 ] *
4R * ('■"^R^» a
AG \ r n + Rg/ Eq. (348)
4(r. + , u >. [*.(5Jf%y
The maximum available gain, MAG, of a network is defined as
the ratio of the power dissipated in the load to the power available from
the generator when the generator internal resistance is matched to the
input transistor resistance, and when the load resistance is matched
to the transistor output resistance. In order to solve for the maximum
power gain in terms of the opencircuit parameters, the imagematched
input and output resistances, previously determined, are substituted
in the operating gain equation 346. Then, the maximum available gain,
MAG = tV j_ w^f 2 ^ ir Eq. (350)
[(ii + ru) (r, +r 22 )  r 12 r 21 ] 2 * v '
where rj = rn Vl— 8 Eq. (337)
and r 2 = r 22 y/l8~ Eq. (341)
Substituting equations 33 7 and 341 i n equ ation 350, for r t and r 2 ,
MAG " L(r„ + V 1 ~ * + rn) (r 22 V 1  » + r«) r 12 r 21 J » ^ <*">
THE GROUNDED BASE TRANSISTOR 41
which is equal to
4r n r 22 r 21 2 (l8) 4r u r 22 r 21 2 (1  8)
[rur M (1 + VT=F) 2 r 12 r 21 ] 2 ^^ T (1 + ^7^)2 _g£?J
Eq 2 \352)
MAr = 4r 2 21 (18) 4r 2 21 (l8)
~r n r 22 T (1 + VM 2  «J a r u r M I 1 + 2yiS + 18Sj*
L J L Eq. (353)
from which derives
4r 2 21 (18) r 2 21 (yi8) 2
4r u r 22 t\ + Vl8 S] 2 ~ r u r M (VI*)* + V**) *
L J £g. (554)
r
2„
For the typical pointcontact transistor, when r n = 250 ohms, r 12 =100
ohms, r 21 = 24,000 ohms, r 22 = 12,000 ohms, and when assuming R g =
50 ohms and R L = 8,000 ohms, the operating gain G, becomes
G= 4R g R L r* 21
[(R g +r u ) (R L + r 22 )r 12 r 21 ] 2
4(50) (8,000) (24,000) 2 _
[(50 + 250) (8,000 + 12,000)  100 (24,000)] 2
The available gain, AG
V 2 21
[HOT7)] (r " + R « ):
Ar (50) (24,000) 2
r 2
The maximum available gain MAG = ■ ., , 21 /Y ■■■ ! „
5 r n r 22 (1 + yi8) 2
MAG = , < 24 'T ,= 92
Notice that the stability factor, 8 = ^^~ , is *!£ {?*S =
r n r 22 250 (12,000)
0.8. If the stab ility factor is greater than one, the numerical value of
the quantity \/l— 8 must be negative, which indicates an unstable con
dition. For the typical junction transistor in which r n = 550 ohms,
r 12 = 500 ohms, r 21 = 1,900,000 ohms, and r 22 = 2,000,000 ohms; when
assuming R g = 100 ohms and R L = 1,000,000 ohms, the operating gain
42 FUNDAMENTALS OF TRANSISTORS
4R,Rtf* M
G =
[(R, + r„) (R L + r 22 )r 12 r 21 ]s
4(100) (1,000,000) (1,900,000) «
G _ [(100 + 550) (1,000,000 + 2,000,000) 500 (1,900,000)] 2 '
The available gain AG =
V81
Ar ^ 100(1,900,000)'
'[wwTyn^ 1 "'
From equation (i55)
MAG = 0,900.000) 2>400
Junction Capacitance
Before the actual transistor circuit is considered, some additional
and important characteristics must be defined. In Chapter 2, the col
lector junction capacitance was mentioned in connection with tran
sistor highfrequency characteristics. In the equivalent network, this
parameter acts in parallel with the collector resistance. The value of
collector junction capacitance C c varies in units of the same type, but
for a typical junction transistor is approximately 10 /*/tf. The value
of capacitance is primarily a function of the junction area, although
it also depends on the width of the junction layer and the resistivity
of the base layer.
Zener Voltage
If the reverse voltage applied across a PN junction is gradually
increased, a point is reached where the potential is high enough to
break down covalent bonds and cause current flow. This voltage is
called Zener voltage. In transistor application the Zener voltage has
the same design importance as the inverse voltage rating of a vacuum
tube, since it defines the maximum reverse voltage which can be ap
plied to a junction without excessive current flow. The Zener potential
for a transistor junction can be increased by widening the space charge
layer, or by forming the junction so that the transition from N region
to P region is a gradual process. In germanium, the Zener voltage field
is around 2 x 10 5 volts/centimeter. A transistor junction with a Zener
THE GROUNDED BASE TRANSISTOR 43
potential of 70 volts would therefore have a space charge layer of
70
^ t; = 35 x 10 5 centimeters.
2 x 10 6
Saturation Current, l ce
Another important transistor characteristic is the saturation cur
rent I co . This is the collector current that flows when the emitter cur
rent is zero. In properly functioning transistors, I^ is in the vicinity
of 10 microamperes; the value is considerably higher in defective tran
sistors. The saturation current is composed of two components. The
first is formed by thermally generated carriers which diffuse into the
junction region. The second component is an ohmic characteristic
which is caused by surface leakage across the space charge region, from
local defects in the germanium, or from a combination of these two
factors. The ohmic component may be separated from the true or
thermally caused value by measuring the collector current at different
values of collector voltage.
Chapter 4
GROUNDED EMITTER AND GROUNDED COLLECTOR TRANSISTORS
The design and servicing of the transistor circuit is more compli
cated than that of the vacuum tube, because transistor input and output
circuits are never inherently independent of each other. This makes
it difficult for a newcomer to get the "feel" of the transistor. In the
long run, however, these same complex characteristics provide for a
more flexible device, one capable of many circuit applications beyond
the range of the vacuum tube.
This chapter deals with the extension of the fourterminal charac
teristics developed for the grounded base to encompass the two remain
ing connections, the grounded emitter, and the grounded collector; a
comparison of the major features of the three basic connections; limita
tions of the transistor; and transistor testing methods.
Introduction
In the following analysis of transistor performance in the grounded
emitter and grounded collector connections, the same typical point
contact and junction transistors discussed in Chapter 3 will be used
for numerical examples. For the pointcontact transistor in the ground
ed base connection, the parameters are:
r 12 =100 ohms = r b
r n = 250 ohms = r e \ r b ; then r e = 150 ohms
r 21  24,000 ohms = r m f r b ; then r m = 23,900 ohms
r 22 = 12,000 ohms = r c { r b ; then r c — 11,900 ohms
For the junction transistor in the grounded base connection:
r 12 = 500 ohms = r b
r u = 550 ohms = r e f r„; then r e = 50 ohms
r 21 = 1,900,000 ohms = r m f r„; then r m = 1,899,500 ohms
r 22 = 2,000,000 ohms = r c + r„; then r c = 1,999,500 ohms
Notice that since r m and r c are so much greater in value than r b , par
ticularly in the case of the junction transistor, for all practical purposes
r 2 i = I'm and r 22 = r c .
m «
o WV • — V\A — y^r — °
•i '•
o
(A) (B)
Fig. 41. (A) The grounded emitter connection. (B) Equivalent active "1" for grounded
emitter connection.
GROUNDED EMITTER AND COLLECTOR TRANSISTORS 45
Fig. 42. Operating circuit, ground
ed emitter connection.
The Grounded Emitter Connection
Equivalent Operating Circuit. The grounded emitter connection
is illustrated in Fig. 41 (A) . In this case the input connection is made
between the base and emitter electrodes (conventionally the emitter is
shown schematically as an arrowhead resting on the base) , and the
output is taken between the collector and the emitter. Thus, in this
case, the emitter is the common electrode. Figure 41 (B) illustrates the
equivalent active "T" circuit for the grounded emitter connection.
Figure 42 is the complete operating circuit of the grounded emit
ter connection. Notice that although the negative side of the signal
generator is grounded, the polarity of the signal in this connection is
reversed with respect to the emitter and base terminals shown in the
grounded base connection of Fig. 39. Since this effective reversal of
input leads is the only physical difference between the two connections,
the grounded emitter, unlike the grounded base, produces a phase in
version of the input signal.
Circuit Parameters. The general opencircuit characteristics derived
for the grounded base connection apply equally well to the grounded
emitter and grounded base connections, since the characteristics were
determined on the basis of a sealed box. However, since the internal
parameters of the transistor have been rearranged, the values of the
general characteristics are different. It is necessary then, to evaluate
the opencircuit characteristics r u , r 12 , r 2 i, and r 22 in terms of the tran
sistor internal parameters r e , r b , r c , and r m . The same basic measuring
circuits, illustrated in Figs. 38, may be used to determine the four
terminal parameter for the grounded emitter connection:
A: r u = e t /ii wheni 2 = 0, r n = r e + r b Eq. (41)
B: r 21 = e 2 /i t when i 2 = 0, r 21 =r e — r m Eq. (42)
C: r 12 = e x /i 2 when i t = 0, r i2 = r e Eq. (43)
D: r 22 = e 2 /i 2 when i t = 0, r 22 = r e + r,.  r m Eq. (44)
These groundedemitter relationships are derived as follows:
A. Using Fig. 41 (B) , the input loop equation on the basis of Kirch
off's law is:
ei = ii (r e + r„) + i 2 r e
46 FUNDAMENTALS OF TRANSISTORS
when i 2 = 0, e! = i t (r e + r„)
then ril= iL = ^i^^r e + r b
B. For the same input loop equation, when ij =
ei = i 2 r e ;
_ e i »2r c _ ,
then r 12 = ± =
C. The output loop equation for Fig. 41 (B) on the basis of Kirch
off's law is:
e 2  r m i e = iir c + i 2 (r c f r c )
Also i c =  (i, + i 2 )
Substituting for i e ,
e 2 + r m (ii f i 2 ) = iir c + i 2 (r c + r c )
e 2 = ii (r e  r m ) f i 2 (r c f r c  r m )
when i 2 = 0, e 2 = i x (r e  r m )
then r 21 = r = ; r c — r m
D. Using the same equations as in C above, when
ii = 0, e 2 = i 2 (r e + r c  r m )
then r22= Jl == i2(re+rc r ) , = re+rc _ rn
i 2 i 2
The opencircuit characteristics can now be numerically evalu
ated for the typical pointcontact and junction transistors previously
considered in Chapter 3. For the pointcontact transistor in the ground
ed emitter connection:
in = i"e + ib = 15° + 100 = 250 ohms
r 12 = r c = 150 ohms
r 2 i = r c  r m = 150  23,900 = 23,750 ohms
raa = r e + r c  r m = 150 + 1 1,900  23,900 = 1 1,850 ohms
For the junction transistor in the grounded emitter connection:
r u = r e + r b = 50 + 500 = 550 ohms
r i2 — r e = 50 ohms
r 2 i = r e  r m = 50  1,899,500 = 1,899,450 ohms
r 22 = r c + r c  r m = 50 + 1,999,500  1,899,500 = 100,050
Because of the large values of r m and r c with respect to r e , r 21 in
the practical case can be approximated by — r m , and r 22 by (r c — r m ) .
The emitter resistance, r c = r 12 , is the feedback resistance and is equiva
lent to r b = r 12 in the grounded base connection. Notice, however, that
since there is phase inversion in the grounded emitter connection, r e
produces degenerative (negative) feedback, rather than regenerative
(positive) feedback. The degenerative effect of the output current
GROUNDED EMITTER AND COLLECTOR TRANSISTORS 47
through r e is similar to the degenerative action of an unbypassed cath
ode resistor in a grounded cathode vacuum tube.
Current Gain in the Grounded Emitter Connection. The current
gain in terms of the general fourterminal parameters was defined by
equation 38 as:
In terms of the transistor parameters in the grounded emitter connec
tion now being considered, the current gain is
Eq. (46)*
RL f r e + r c — T m
In the case of the groundedemitter pointcontact transistor, r 2 i and
r 22 are both negative. The value of the load resistor, R L , determines
whether the current gain is positive or negative. If Rl is less than the
absolute value of — r 22 , a is positive; if R L is greater than the absolute
value of — r 22 , a is negative. A negative value of current gain indicates
simply that the input current is inverted in phase. This is normal in
the grounded emitter connection. Theoretically, an infinite current
gain is attained when R L = — r 22 . The current gain of a typical point
contact transistor with a load R L = 15,000 ohms is
 23,750 _ = _ 1M
15,00011,850
(Equation 38 A for maximum current gain, oo = — ^— , does not apply
r 22
in this connection, since it is found that the pointcontact transistor
is unstable when R L is less than — r 22 .) Notice that the current gain
becomes very large for values of R L slightly larger than — r 22 . For ex
ample, if
R L = 12,500 ohms, « = nWUm = " 366
The current gain in the junction transistor is always negative in
the grounded emitter connection, since r 22 is always positive. The cur
rent gain for the typical groundedemitter junction transistor with a
load
— 1 899 450
R L = 100,000 ohms, is . = m ^ + ^ m =  9.5
The maximum current gain, as in the case of the grounded base con
nection, is:
_r 21 _ 1,899,450 _ ian
a °~~r^"~ 100,050 ~
48
FUNDAMENTALS OF TRANSISTORS
3,000
_ zpoo
ipoo
r;
=50 OHMS
A
= 250 OHMS
T R L =~>
AT R L =0
r n = 250 OHMS
r, 2  150 OHMS
r 2 , = 23,750 OHMS
r 22  11,850 OHMS
1
2£00
IK IOK IOOK I MEG
LOAD RESISTANCE R L (OHMS)
Fig. 43. Inpuf resistance vs load resistance for typical pointcontact transistor
(grounded emitter).
Input Resistance r ( for the Grounded Emitter Connection. The in
put resistance was defined in equation 313 in terms of the general
/ Tl^l \
opencircuit parameters as: r, = r u —  — — J
V r 22 + Kl /
The input resistance in terms of the transistor parameters in the
grounded emitter connection becomes:
»' + H.r"'£~rI + Eq{4 ' 7) '
_ IJ600
1^00
fl "1.
■iOO
^v.
AT R L =0
'll
'\2
 50 OHMS
= 1,899,450 OHMS
= 100,050 OHMS
r»
t\  550 OHMS
AT R L = ~
IOK IOOK I MEG 10 MEG
LOAD RESISTANCE R L (OHMS)
Fig. 44. input resistance vs load resistance for typical junction transistor (grounded emitter).
GROUNDED EMITTER AND COLLECTOR TRANSISTORS 49
The effect of the value of the load resistance on the input resistance
of typical transistors is illustrated in Figs. 43 and 44. The input re
sistance for the pointcontact transistor starts at a value of —40 ohms
for R L = 0, and becomes more negative as the load resistance increases.
When R L = — r 22 , the input resistance is infinite. As the load resist
ance increases beyond this point, the input resistance becomes positive,
decreasing in value to the limiting condition r, = r n = 250 ohms when
the output is opencircuited. Negative values of input resistance indicate
circuit instability; consequently, the pointcontact transistor can be
used as an oscillator in the region where R L is less than — r 22 . Circuits of
this type are called "collectorcontrolled oscillators."
The input resistance of the junction transistor is always positive.
In the typical transistor considered, the input resistance decreases from
a value of 1,500 ohms at R L = 0, to 550 ohms for an infinite load.
Output Resistance r for the Grounded Emitter Connection. The
output resistance was defined by equation 321 in terms of the general
fourterminal parameters as:
'■H«^
The output resistance in terms of the internal transistor parameters in
the grounded emitter connection becomes:
r = r. + r t r m (j' ( I' " ^.1 Eq. (48)*
The effect of the value of the signal generator resistance is illustrated
for the pointcontact and junction transistors in Figs. 45 and 46, re
spectively. Notice that the output resistance of the pointcontact type
is positive at R g = 0, and decreases rapidly to zero when R g is slightly
less than 50 ohms. As R g is increased further, r„ becomes negative and
gradually approaches the limiting condition, when R g is infinite, r, =
r 22 = —11,850 ohms. Thus, the pointcontact transistor can have a nega
tive output resistance over a large range of generator resistance values,
and this characteristic can be used in transistor oscillator design. Cir
cuits of this type are called "basecontrolled oscillators."
The output resistance of the junction transistor is always positive,
and for the typical type considered, r„ gradually decreases from approxi
mately 273,000 ohms to 100,000 ohms as R g is increased from zero to
infinity.
The range in which both the output and input resistances of the
pointcontact transistor are positive can be increased by adding external
resistance in the emitter arm. This increases the effective value of
r e = r i2 Notice that if enough external resistance is added so that the
effective emitter resistance r e + R L is equal to or greater than
50
FUNDAMENTALS OF TRANSISTORS
4,000
Ul — fipOO
8JDO0
AT
R
,2
OC
OHMS
'n
= 290 OHMS
= 150 OHMS
' 23,750 OHMS
j= 11,850 OHMS

r
=11,850 OHMS
fl"
10 100 IK
GENERATOR RESISTANCE Rg(OHMS)
Fig. 45. Output resistance v« generator re«ittance for typical pointcontact transistor
(grounded emitter).
^=H
r = 2.73 MEGOHMS
AT Rg >
r '.1 ME60HMS
AT Rg ■ oo
f l2
• 550 OHMS
■ 50 OHMS
■1,899,450 OHMS
■ 100,050 OHMS
r 2 i
LJ
100 IK
GENERATOR RESISTANCE Rg (OHMS)
Fig. 46. Output resistance vi generator resistance for typical junction transistor
(grounded emitter).
GROUNDED EMITTER AND COLLECTOR TRANSISTORS 51
— (r c — r m ) , the input and output resistance is positive. This stabilizing
effect of adding resistance to the emitter load is frequently used in
transistor circuit applications.
Voltage Gain VG in the Grounded Emitter Connection. The volt
age gain was defined by equation 324 in terms of the general four
terminal parameters as:
VG= r " RL
(R L + r 22 ) (R, + r„)  r 12 r 21
The voltage gain in terms of the transistor parameters for the grounded
emitter connection becomes:
VG = ( r e ~ r m) R L £• /jm •
(R L + r e r c r m ) (Rg + r e + r b )  r e (r e  r m ) H ' v ;
The maximum voltage gain defined by equation 325 is:
Max. VG =^i,
j j
which in this case becomes Max. VG = e m Eq. (49 A)
r « + r b
For the pointcontact transistor, assuming R L = 30,000 ohms and R„ =
10 ohms,
vr= 23,750(30,000) = _
(30,000  1 1,850) (10 + 250)  (150) (  23,750)
— 23 750
and the maximum voltage gain is ^r = — 95.0
For the junction transistor, with R L = 100,000 ohms and R g = 10
ohms, the voltage gain is
vr 1,899,450(100,000) =
(100,000 + 100,050) (10 + 550)  (50) ( 1,899,450)
• • 1,899,450 . .„
and the maximum voltage gain is ^^ = — 3,460
Notice that the voltage gain in this connection, like the current
gain, is negative. Again this merely indicates that the input voltage is
inverted in phase.
Impedance Matching in the Grounded Emitter Connection. As in
the analysis of the grounded base connection in Chapter 3, the stability
factor 8 = — 12 21 must be less than unity for shortcircuit stability.
r ll r 22
The numerical value of 8 for the typical pointcontact transistor is
150 (23,750) ,_ _.. . . . . . .
■ ; — ' = 1.2. This reemphasizes the fact that the pomtcon
250 ( — 11,850)
52 FUNDAMENTALS OF TRANSISTORS
«
tact transistor in the grounded emitter connection is unstable when the
output is shortcircuited. The stability factor for the junction transistor
. 50 ( 1,899,450) , _ a , . , ,. , , ...
is — ' ' = — 1.73, which confirms the fact that the junction
550 (100,050)
transistor is shortcircuit stable in the grounded emitter connection.
The in put image resistance is defined in equation 337 as: ri =
r u V 1 —8 The input im age resist ance of the p ointcontact transistor is
numerically equal to 250\/l — 1.19 = 250>/— 19. Since the quantity un
der the square root sign is negative, r t is imaginary and cannot be built
into co nventional si gnal sources. For the typical junction transistor, r x
is 5500  (  173) = 908 ohms.
The output image resistance defined by equation 341 as r 2 =
r 22 \/l — 8 also is imaginary for the pointcon tact transistor. For the
junction transistor, r 2 is 100,050 V 1 — (— 173) = 165,000 ohms.
Power Gain in the Grounded Emitter Connection. The numerical
values of the voltage and current gains are always negative in the stable
range of operation of the grounded emitter connection. The negative sign
is merely a mathematical indication of the phase inversion of the ampli
ifed signal. Since the power gain is a function of the product of the volt
age and current gains, its numerical value must be positive.
The operating gain is defined in equation 346 as
_ 4R.R L r» M
[(Rg + Jii) (RL + r 22 )r 12 r 21 p
Note that since r 21 ana the bracketed quantity in the denominator are
squared, the numerical value of this equation is always positive. It is
certainly possible to obtain an apparently valid power gain in an un
stable portion of the transistor characteristic if numerical values are
haphazardly substituted. For example, evaluating the operating gain for
the typical pointcontact transistor when R L = 1,000 ohms and R g = 10
ohms,
r _ 4(10) (1000) (23 ,750)" = ^ 4
[(10 + 250) (1000  1 1,850)  150 ( 23,750)] ;
However, Fig. 43 indicates that at a load of R L = 1,000 ohms, the tran
sistor is unstable and will oscillate. This does not mean that the ground
ed emitter connected transistor can oscillate and supply a power gain
at the same time, but rather that the operating gain equation can only
be applied conditionally. Without going too deeply into the mathematic
al concepts involved, equation 346 can only be applied when
(R g + r u ) (R L f r 22 )  r 12 r 21
is greater than zero. When R g and R L equal zero, the worst possible
case, the condition equation becomes r u r 12 — r 12 r 21 > 0. This is just an
other way of expressing the requirement that the stability factor, 8 =
GROUNDED EMITTER AND COLLECTOR TRANSISTORS
53
r 12 r 21
rnro 2
,must be less than unity. As a result, the operating gain is con
ditional when 8 is greater than unity.
In the typical pointcontact transistor under discussion, 8 = 1.19,
the conditional equation is (R g + 250) (R L  11,850)  150 (23,7 50) >0.
A plot of this conditional characteristic is shown in Fig. 47. Any com
bination of generator resistance and load resistance in the stable region
can be used, but the selection of operating values close to the condition
al characteristic provides the greatest operating gain.
The following example illustrates the design of a grounded emitter
circuit for maximum power gain when the stability factor is greater than
one. Assume that the load R L is fixed at 10,000 ohms for the typical
pointcontact transistor. Figure 47 indicates than any value of R g less
than 1,530 ohms will provide stable operation. Thus for R g = 100 ohms
G=,
4(100) (10,000) ( 23,750) :
[(100 + 250) (10,000  1 1,850)  150 ( 23,750)] "
If, however, R g = 1,450 ohms were selected,
4(1450) (10,000) ( 23,750) 2
= 267
G =
[(1450 + 250) (10,000  1 1,850)  150 (  23,750)]
= 195,000
14,000
12,000
_ iopoo
SfiOO
6,000
4,000
1
R e '0
c
TAB
»ESI
LE
3N
"e
= 8!
DH
MS
t
\y
, oc
pe
UN!
R
TA
EGI
BL
ON
E
>Re
1
1
l"
GENERATOR RESISTANCE Rg (OHMS)
Fig. 47. Conditional stability characteristic (grounded emitter).
54 FUNDAMENTALS OF TRANSISTORS
These examples prove that extremely high values of power gain can
be attained by selection of R g R L values close to the stability character
istic. In the practical case, however, the selected values must be suffi
ciently removed from the instability limit to avoid the introduction of
circuit oscillation by normal parameter variations.
The grounded emitter connection can be stabilized by adding re
sistance in the emitter arm. As an example, assume that a resistor
Re = 850 ohms is added in series with the emitter. The fourterminal
parameters then become:
r n = r e + Re + r b = 150 + 850 + 100 = 1,100 ohms
r 12 = r e + Re = 150 + 850 = 1,000 ohms
r 21 = r e 4. R e  r m = 150 + 850  23,900 = 22,900 ohms
r 22 = r e + Re + r c  r m = 150 4 850 + 1 1,900  23,900 = 1 1,000 ohms
Substituting these new values, the conditional equation becomes
(Rg 4 1,100) (R L — 11,000) (1,000) (22,900) and must be greater
than zero. A plot of the modified conditional stability characteristic is
shown in Fig. 47. Notice the extent to which the stability area has been
increased. As before, the selection of values R L and Rg located near the
limiting line provide the greatest power gain.
The maximum available power gain defined by equation 355,
MAG= **" .
rur 12 (l + yply
can be applied to the grounded emitter connection provided that the
stability factor is less than one. The numerical value of the maximum
available gain for the typical junction transistor is:
MAG = C 1 ' 8 "' 45 , ) 2 9,340
550 (100,050) [1 + Vl(l73)] 2
The Grounded Collector Connection
Equivalent Operating Circuit. The grounded collector connection
is illustrated in Fig. 48 (A) . In this connection, the input signal is
connected between the base and collector electrodes, and the output is
taken between the emitter and the common collector. The equivalent
active "T" is illustrated in Fig. 48 (B) .
The general fourterminal parameters can be measured in terms
of the internal transistor parameters using the basic measuring circuits
of Fig. 38. The fourterminal parameter equations for the grounded
collector connection are:
A. r n = J— when i 2 = 0, r n = r b + r c Eq. (410)
B. r 21 = —£ when i 2 = 0, r 21 = r c Eq. (411)
GROUNDED EMITTER AND COLLECTOR TRANSISTORS 55
9°
(A)
Fig. 48. (A) The grounded collector connection. (B) Equivalent active "T" for grounded
collector connection.
C r 12 = ^ when i t = 0, r 12 = r c  r m Eq. (412)
D *22 = r 2  when i t = 0, r 22 = r e + r c r m Eq. (413)
These equalities are derived as follows:
A. Using Fig. 48 (B) , the input loop equation on the basis of Kirch
off's law is:
ei + r m i e = ii (r b + r c ) + i 2 r e
For this connection i 2 = i e
and e t = ij (r b + r c ) + i 2 (r c  r m )
when i 2 = 0, ei = i t (r b + r c ) ,
and r 11= f = ^ + r * = r b + r c
B. Using the same input loop equation, when i t = 0, ei = i 2 (r c — r m ) ,
and r 12 =4i = i2(rc .~ rm) =r (! r ni
C. The output voltage loop equation is:
e 2 +r m i e = iifc + k (r e + r c )
Since i 2 = i e , e 2 = i^,. + i 2 (r e + r c  r m )
when i 2 = 0, e 2 = ^r,.
_ e 2 ijr,. _
r 21 — =— j r c
D. Using the same output loop equation,
when ii = 0, e 2 = i 2 (r e + r c  r m )
r  e a _ *«(*« + *«*■) _ , L ,
r 22 . . r e f r c — r m
i 2 i 2
The numerical values of the fourterminal parameters can now be de
termined for the typical pointcontact transistor. The values are:
*n = r b + r c = 100 + 11,900 = 12,000 ohms
Ti 2 = r c r m = 11,90023,900 = 12,000 ohms
r 2i = r c = 1 1 ,900 ohms
r 22 = r e + r c r m = 150 + 11,900  23,900 = 11,850 ohms
56
FUNDAMENTALS OF TRANSISTORS
Fig. 49. Operating circuit, ground
ad collector connection.
The numerical values for the junction transistor are:
rii = r b + r c = 500 + 1,999,500 = 2,000,000 ohms
r 12 = r c  r m = 1,999,500  1,899,500 = 100,000 ohms
i"2i = r c = 1,999,500 ohms
r 22 = r e + r c x m = 50 + 1,999,500  1,899,500 = 100,050 ohms
Because of the very low values of r b and r e compared to the quan
tities r c and (r c — r m ) , r n is approximately equal to r 2 i, and r 22 is ap
proximately equal to r 12 .
Figure 49 illustrates the operating circuit for the grounded col
lector connection. As in the analysis of the grounded emitter circuit,
the performance characteristics for this connection can now be deter
mined by straightforward substitution in the general fourterminal cir
cuit equations.
Current Gain, a , of the Grounded Collector Connection. The cur
rent gain as defined in equation 38 is:
_ r 21
Rl + r 22
In terms of the internal transistor parameters in the grounded collector
connection, the current gain becomes:
 R , + ,, r ;r,r„ *• «">
The value of r 22 is always negative in the case of the pointcontact
transistor. Therefore, the load resistor R L must be larger than the ab
solute value of r 22 for stable operation, and the equation for maximum
current gain a<, = — — can only be applied to the junction transistor.
r 22
Numerical values for the typical junction transistor when R L = 100,000
ohms are
1,999,500 Q
a ~~ 100,000 + 100,050
and the maximum current gain
_ 1,999,500 _ on
° 0_ 100,050
For the pointcontact transistor when R L = 15,000 ohms,
GROUNDED EMITTER AND COLLECTOR TRANSISTORS 57
11,900
a 15,00011,850
It would appear that as the load approaches the absolute value of
r 22 , extremely high current gains are attainable. For example, for the
pointcontact transistor when R L = 11,950 ohms,
__n^oo__
a 11,95011,850
In operating circuits, however, the grounded collector current gain is
limited to the same order of magnitude as in the grounded emitter con
nection. This limitation is caused by the rapid increase in input re
sistance with an increase in current gain.
Input Resistance, r { , in the Grounded Collector Circuit. The gen
eral input resistance is defined by equation 313:
r, = r
In terms of grounded collector transistor parameters, the input resistance
becomes
r. = r b + r c
{<■;■£.£•; *• «">
The variation of input resistance with load is illustrated for the point
contact and junction transistors in Figs. 410 and 411, respectively. The
input resistance for the pointcontact transistor is negative from R L =
to R L = — r 22 = 11,850 ohms. Notice that when R L = — r 22 , the input
resistance is infinite or open circuited. As R L is increased further, the
input resistance becomes positive, and gradually decreases to a limiting
value of 12,000 ohms. The input resistance of the junction transistor
increases from a value of approximately 500 ohms to the limiting value
r i = r n — 2,000,000 ohms when the output is open circuited.
Output Resistance, r , for the Grounded Collector Connection. The
output resistance is defined by equation 321:
r o — r 22
/_£l2£21__\
In terms of the internal transistor parameters, the output resistance
becomes:
r„ = r e + r,
<~ \x:fh) *• <*«>
The variation of r with respect to the generator resistance R g is il
lustrated for both transistors in Figs. 412 and 413. In the pointcontact
characteristic, the output resistance is positive over the range of Rg from
to approximately 50 ohms. When the generator resistance is increased
beyond 50 ohms, r becomes negative, and gradually approaches a limit
ing value equal to r 22 (— 11,850 ohms) for large values of R K . The out
put resistance of the junction transistor starts at a value of approximate
58
FUNDAMENTALS OF TRANSISTORS
4Q0OO
jopoo
20POO
lopoo
4P00
I2POO
V
\
5=12,000 OHMS
AT R L = oo

rj =50 OHMS
AT R L =0
r n =12,000 OHMS
r 2 = l2,0O0 0HMS
r 2l = 11,900 OHMS
r 22 = ll, 850 OHMS
\
\
10 K IOOK
OUTPUT RESISTANCE R L (OHMS)
Fig. 410. Input resistance v< load resistance for typical pointcontact transistor
(grounded collector).
ly 75 ohms at Rg = and gradually approaches a value equal to r 22
(100,050 ohms) for large values of generator resistance.
As in the case of the grounded emitter, the grounded collector cir
cuit using the pointcontact transistor cannot be matched on an image
2.0
a
AT
!ME
V
GO
HM
S
1.5
.5
r
1 • 2 MEGOHMS
2 ' 0,1 MEGOHMS
rj • 500 OHMS
f
, > 1,999,500 OHMS
 —
1 —
1—
IK I0K IOOK I ME'
OUTPUT RESISTANCE R L (OHMS)
Fig. 411. Input resistance v< load resistance for typical junction transistor (grounded collector).
GROUNDED EMITTER AND COLLECTOR TRANSISTORS
59
zpoo
2poo —
6£00
izpoo
r . 50 OHMS
AT Rg =
r =11,850 OHMS
AT Rg = oo
•"II
' 12,000 OHMS
• 12,000 OHMS
• 11,900 OHMS
•11,850 OHMS
^
•21
1
GENERATOR RESISTANCE Rg (OHMS)
Fig. 412. Output resistance vi generator resistance for typical pointcontact transistor
(grounded collector).
basis without external modification, since the stability factor of this
circuit is greater than one. However, the grounded collector does ex
hibit a unique characteristic when external resistance is added in the
collector arm. For example, assume that a resistor R c is added to the
loopoo
r o
1 III
= 100,050 OHMS
80J000
r Rg
*
*•
<
60000
40D00
20P00
1 .2 MEGOHMS
2 '0.1 MEGOHMS
2I • 1,999,500 OHMS
22> 100,050 OHMS
r . 75 OHMS
_ AT Rg .0
„ 1
r
1—
LJ
IOOK I MEG
GENERATOR RESISTANCE Rg(OHMS)
10 MEG
Fig. 413. Output resistance v> generator resistance for typical junction transistor.
60 FUNDAMENTALS OF TRANSISTORS
collector arm so that R c + r c = r m . For this modification,
rii = r b + r c + R c
*i2 = *c + Rcr m =
r 21 = r c + R,. and
r 22 = r e + r c + Kc — r m = r e
Since r 12 = 0, the stability factor
8 =_£l£?i_=0
r ll r 22
Thus, the modified circuit is stable. The input image matched resistance
(equation 337) is then ^__
ri = r n VlS = r n = r b + r c + R^
and the output image matched re sistanc e (equation 341) becomes
r 2 —• r 22 v 1 — o —  r 22
Notice also that r, = r x = r u and r = r 2 = r 22 .
Thus, adding a suitable external resistor in the collector arm causes
the circuit to act as a perfect buffer stage in which both the input and
output resistances are independent of R L and R g .
Numerical values for the typical pointcontact transistor modified
to act as a buffer stage are:
r l = r 1 = r n = r b + r c + R c = 100 + 11,900 + 12,000 = 24,000 ohms;
r = r 2 = r 22 = r e = 150 ohms.
The image matched input and output equations can be applied
to the junction transistor since its stability factor is always slightly less
than one. A practical method to use in selecting values to be substituted
in these equations indicates that r t should be chosen to equal 2 percent
of r n , and r 2 equal to 2 percent of r 22 . The exact determination of the
image matched resistances in the grounded collector circuit is not im
portant, because the power gain is constant over a wide range of load
resistances when the signal generator is matched to the input resistance.
In the junction transistor, numerical values for image matched re
sistances are
„ . ,„^= ,000,00 y. iSagffgg  «"» —
If the approximate values are used
r! = .02r u = .02 (2,000,000) = 40,000 ohms
r 2 = .02r 22 = .02(100,050) = 2,001 ohms
Voltage Gain in the Grounded Collector Connection. The voltage
gain, as defined in equation 324 by the general fourterminal para
meters is:
vg=,^ y^
(R L + r 22 ) (Rg + r n ) r 12 r 21
GROUNDED EMITTER AND COLLECTOR TRANSISTORS 61
In terms of the internal transistor parameters for the grounded collector
connection, the voltage gain becomes:
VG = (Rl + r c + r e  r m ) (R g + r c + r b )  r c (r c  r n ) E + ( '" /7) *
Under the conditions for infinite input resistance and infinite current
gain (R L = — r 22 ) , the voltage gain becomes:
yp _ r 21 (~ r 22) £22 r c \ r e ~ r m
(~ r 22 + r 22) (Rg + r ll) ~~ r 12 r 21 r 12 r c ~ r m
For a perfect buffer stage Rc  r c — r m = 0. Thus, the voltage gain equa
tion becomes
YQ (Rc + r c) Rl
(R L + r e ) (R, + r b + r c + R.)
The maximum voltage gain as defined by equation 325 is VG = — — ;
r 21
this becomes p — . For the typical pointcontact transistor, when
r b + r c
R L = 15,000 ohms and R g = 10 ohms,
VG
I"2iRl
(Rl + r 22 ) (R g + r n ) r 12 r 21
11,900(15,000) = 98?
(15,000 11,850) (10 f 12,000) (12,000) (11,900)
Under the conditions R L = — r 22 = 11,850 ohms
, rn _ r 22 —11,850 oa _
G ~"^" = 12,000 ~ 985
Under the conditions R c f r c = r m (R c = 12,000 ohms)
YG (Rc + r c) Rl
(R L + r e )(R g + r b + r c + R c )
(12,000 f 11,900) 15,000
= .983
~ (15,000 + 150) (10 + 100 + 11,900 + 12,000)
The maximum voltage gain VG =igJ= " ,90 ? = .990
For the typical junction transistor, R L = 100,000 ohms, R g = 10 ohms
VG ^ f 2lRL
(Rl + r 42) (Rg + r u ) r 12 r 21
1,999,500 (100,000)
= .998
(100,000+ 100,050) (10 + 2,000,000) 100,000 (1,999,500)
The maximum voltage gain VG =^— = 1 ' 999 ' 500 =,999
^ 8 r u 2,000,000
Notice that in all of the above cases, the voltage gain is slightly
less than unity. This is typical of the grounded collector connection.
Power Gain in the Grounded Collector Connection. The operating
power gain as defined by equation 346 is:
G= 4R g R L r2 21
[(Rg + ru) (Rl + r 22 )  r 12 r 21 ] *
62
FUNDAMENTALS OF TRANSISTORS
12000
POpOO
2 WOO
3 spoo
g 4PO0
10 K 0.1 MEG
GENERATOR RESISTANCE Rj (OHMS)
Fig. 414. Conditional stability characteristic (grounded collector).
As in the case of the grounded emitter connection, this gain equation
is conditional for the pointcontact transistor when the stability factor
S is greater than one. The conditional equation is:
(Rg + r u ) (R L + r 22 )  r 12 r 21 >
Substituting the numerical values of the typical pointcontact transistor
into this equation:
(R g + 12,000) (R L  11,850) + 12,000(11,900) >
The conditional stability characteristic is plotted in Fig. 414.
The grounded collector circuit can be stabilized by adding an ex
ternal resistance R,. in the collector arm. For example, assume that a
resistor R c = 3,100 ohms is placed in series with r c . The opencircuit
parameters now become:
r u = r b + r c + R,. = 100 + 11,900 f 3,100 = 15,100 ohms
Tiz = r c + Rer,,, = 11,900 f 3,100  23,900 = 8,900 ohms
r 2 i = r c + Rc = 11,900 + 3,100 = 15,000 ohms
r 2 2 = r e + r c + R c r m = 150 + 11,900 + 3,100  23,900 ==8,750 ohms
The conditional equation now becomes:
(Rg + 15,100) (R L  8,750) + 8,900 (15,000) > 0.
This stabilized conditional stability line is also plotted on Fig. 414.
The maximum available gain, defined by equation 355:
MAG= 7T—7T
rnr 22 (1 + VT
W
GROUNDED EMITTER AND COLLECTOR TRANSISTORS
63
can be applied to junction transistors, since the stability factor is not
greater than one. In the grounded collector connection, since 8 is al
ways very near unity, this equation can be simplified as:
MAG=
r 2 i'
, _ Eq. (418)*
rur M (l + '/unj» r u r 22
Notice that this result is nothing more than the product of the maxi
mum voltage gain — — an d the maximum current gain — — . The
r ll r 22
maximum available gain in terms of the transistor internal parameters
is:
MAG= (r b + r,)(rf + r.U ** <*">
and since r and r e are negligible compared to the large values of r c and
(r c — r m ) , the maximum available gain is:
MAG= J^ — — = Eq. (420)
. r c (r c r m )
For the typical junction transistor
MAG =
**» *m
1,999,500
=20
r c  r m 1 ,999,500  1 ,899,500
Reverse Power Gain in the Grounded Collector Circuit. The
grounded collector connection also has the unique ability to furnish
power gain in the reverse direction. This characteristic might be antici
pated on the basis of the equivalent circuit, since the internal generator
r m i e is common to both the input and output circuits, and the values of
r b and r e are approximately equal. The equivalent circuit for the re
verse connection is illustrated in Fig. 415. The resulting fourterminal
parameters for this connection can be evaluated in terms of the internal
Fig. 413 (above). Equivalent "T"
for reverie operation of grounded
collector connection.
Fig. 416. (right). Tramittor col
lector l c Ec characteristic Illustrat
ing maximum limitations.
MAXIMUM COLLECTOR
DISSIPATION 100 MILLIWATTS
MAXIMUM COLLECTOR
VOLTAGE 30 VOLTS
4 6 8 10 12 14 16
COLLECTOR CURRENTMILLIAMPERES
64 FUNDAMENTALS OF TRANSISTORS
transistor parameters as before:
A. The input loop equation is:
ei + r m i e = ii (r e + r c ) + i 2 r c
Substituting i e = i lf
ei = ii(r e + r c r m ) + i 2 r c
when i 2 = 0, e x = i x (r e + r c  r m ) ; then r n = r e + r c  r m , which is
equal to r 22 in the forward direction.
B. Using the same input loop equation, when i t = 0, e x = i 2 r c , then
r 12 = 1 = r„ which is equal to r 21 in the forward direction.
i 2
C. The output loop equation is
e 2 + r m i e = iir c + i 2 (r b + r c ) ;
Since i e = i lt e 2 = i, (r c  r m ) + i 2 (r b + r c )
when i 2 = 0, e 2 = i t (r c  r m ) ; then r 21 = — p = r c  r m ; which is
equal to r 12 in the forward direction.
D. Using the same output loop equation, when ij = 0, e 2 =
i 2 (r b 4 r c ) ; then r 22 = e 2 /h = r b + r c , which is equal to r n in the for
ward direction.
Therefore, it can be seen that any of the equations derived for
operation in the forward direction can be revised for use in the reverse
direction by substituting r u for r 22 , r 12 for r 21 , r 21 for r 12 , and r 22 for r u .
For example, the maximum available power gain in the forward direc
j2 ]2
tion, MAG = — — , becomes MAG =  — in the reverse direction.
r 12 r 22 i"22 r n
Comparison of Transistor Connections
The analyses of the three basic connections and their operating
characteristics apply equally to both pointcontact and junction transis
tors. However, due to the difference in comparative values of the in
ternal transistor parameters, r„, r b , r c , and r m , the performance of the
two basic transistor types is considerably different. In practice, the
pointcontact transistor is unstable, and has negative input and output
resistances. On the other hand, the junction type is generally cheaper
to produce, has better reliability, better reproducibility, higher available
gain, and a lower noise figure than the pointcontact type. It is safe to
predict the gradual displacement of the pointcontact transistor by the
junction transistor in all but a few specialized applications, particularly
since the frequency range of the junction type is steadily being in
creased by new manufacturing techniques. In view of this, the remainder
of the book will deal primarily with the junction transistor, and unless
specified, typical junction characteristics will be assumed.
At this point in the book all the basic design formulas have been
derived for the three transistor connections. Thus, a comparison be
GROUNDED EMITTER AND COLLECTOR TRANSISTORS 65
tween the general characteristics of the three fundamental connections
is now in order.
The grounded base connection is similar to the grounded grid cir
cuit in electron tubes. This connection is characterized by low input
resistance, high output resistance, and no phase inversion. Although its
current gain is less than one, it provides respectable voltage and power
gains. It is well suited for dc coupling arrangements and for preampli
fiers that require a low input and high output impedance match.
The grounded emitter circuit is the transistor equivalent of the
grounded cathode connection in the vacuum tube circuit. This transis
tor connection is the most flexible and most efficient of the three basic
Connections. The grounded emitter connection reverses the phase of
the input signal. Its matched input resistance is somewhat higher than
that of the grounded base connection; its matched output resistance is
considerably lower. The grounded emitter usually provides maximum
voltage and power gain for a given transistor type.
The third connection, the grounded collector, is the transistor equi
valent of the groundedplate vacuum tube. It is characterized by a volt
age gain that is always slightly less than unity. Its current gain is in
the same order as that of the grounded emitter. It has a relatively low
output resistance, a high input resistance, and does not produce phase
reversal. It offers low power gain, but is capable of supplying reverse
power gain. The grounded collector circuit is used primarily as a
matching or buffer stage.
Transistor Limitations
Maximum Limits. To use the transistor in practical circuits, it is
necessary to be aware of its limitations. First, the transistor has limited
powerhandling capabilities. (The maximum power dissipation rating
of a transistor is always specified in the manufacturer's rating sheet.)
Because the dissipation rating is relatively low, the operating tempera
ture of the transistor is usually kept in the general temperature range
of 50°C to 60°C. Relatively low ambient temperatures are also desirable
because germanium is temperature sensitive, and behaves erratically at
higher temperatures. In addition, the operating range is limited by the
maximum allowable collector voltage (a function of the Zener voltage,
previously discussed) , and the maximum collector current. (The values
of these latter factors are also specified in the manufacturer's rating
sheets.)
Figure 416 illustrates the three maximum limitations of a typical
transistor having the following specified ratings: maximum collector
dissipation — 100 milliwatts; maximum collector voltage — 30 volts;
and maximum collector current — 15 milliamperes. The useful region
of the collector currentvoltage characteristics is necessarily limited to
the area contained within these boundaries. In circuit application, none
66 FUNDAMENTALS OF TRANSISTORS
of the limiting factors can be ignored; exceeding any of the limits may
damage the transistor. For example, assume that the transistor illustra
ted in Fig. 416 is operated as follows: collector current I c = 4 milliam
peres, collector voltage E c = 20 volts, load resistance R L — 10,000 ohms.
Assume also that the ac input signal causes a collector current variation
of ± 2 milliamperes. Thus, the output signal varies along the load line
between the collector current limits of 2 to 6 milliamperes. The col
lector current never exceeds the maximum limit of 15 milliamperes, and
at peak signal the collector dissipation is 40 volts times 2 milliamperes
(80 milliwatts) , which is well within the maximum power limits. How
ever, the collector voltage is now 10 volts greater than the allowable
limit of 30 volts. The transistor, therefore, would not be suitable for
the assumed operation.
Minimum Limits. The minimum limits of the transistor are gen
erally not critical in practical cases. The minimum collector voltage is
set by the nonlinear portion of the characteristic curve, which is not
reached until the collector voltage is reduced to a few tenths of a volt.
The minimum collector current must be greater than the saturation
current I co , which is considerably less than 100 microamperes in most
junction types. The error introduced by assuming the minimum limits
to be E c = and I c = is generally negligible.
Transistor Noise. The minimum signal that can be applied to a
transistor is limited by the internal noise generated by the transistor.
Since the transistor does not require cathode heating (one of the major
noise sources in the vacuum tubes) , it is inherently capable of operat
ing at lower noise levels than its vacuum tube brother. At present, the
junction transistor is equal to the vacuum tube, insofar as its noise
characteristics are concerned. The noise level of the pointcontact types
is between 15 and 30 db higher.
There is some confusion in the field as to what is meant by the
manufacturers' specifications on noise limits. This confusion is caused
by the various manners in which the noise level is specified. The noise
level, when specified "with reference to thermal noise," tells the most
about the transistor, because the reference value is reasonably fixed. The
noise factor on this thermal basis is the ratio of the noise power de
livered to a load compared to the power delivered if the only source
of noise were the thermal noise of the signal generator. A second method
of noise specification is the "signaltonoise ratio." The noise figure
on this basis does not tell as much about the transistor as the first
method, because the signal is not at a constant level. Another method
is specification of noise in db above one milliwatt (dbm) . This method
is least useful since it neither specifies amplifier gain nor bandwidth.
The noise figure of the junction transistor is about 10 db above
thermal noise at 1,000 cps; by selection, values as low as 5 db have
GROUNDED EMITTER AND COLLECTOR TRANSISTORS 67
1 r e r c '">'• E *
o — (Jw— W\ •— AA/V — (~) — (~)— °
► r b
Fig. 417. Effect of noiM on oqui
volant transistor circuit.
been found. These noise levels are comparable with those of the best
vacuum tubes available. In general, the noise energy in the transistor
is concentrated in the lower frequencies and, as might be expected, the
noise factor decreases as the operating frequency is increased. The
noise factor is affected by the operating point and the signal generator
resistance. It appears to be lowest both at low values of collector voltage
and when the generator resistance R g is equal to the input resistance r t .
In general, transistors with large collector resistance have a low noise
level. Figure 417 illustrates the equivalent circuit of the grounded base
connection, and includes the equivalent voltages E 1 and E 2 introduced
by transistor noise.
Testing Transistors
Basic Circuits. Although the manufacturer's data sheets for transis
tors are very useful in preliminary paper studies of circuits, it is often
necessary to make direct transistor measurements. The block diagram
of Fig. 38 illustrates the basic circuits for measuring the ac opencircuit
parameters r n , r 12 , r 21 , and r 22 . (Methods for measuring a and I TO are
indicated later in this section.) The following general rules aid the
experimenter in obtaining reasonably accurate results for all measure
ments.
1. Use an accurate meter calibrated for the appropriate operating
range. This is required since the transistor operates on comparatively
small values of current and voltage.
2. Measure the dc bias voltages with a very high resistance volt
meter, to avoid metershunting effects. Shunting errors are particularly
noticeable in the collector circuit which may have resistance of several
megohms.
3. Connect the test signal (usually 1,000 cps) through a stepdown
transformer that has an impedance ratio in the order of 500:1. This
keeps the measurements independent of R g and, at the same time, per
mits a low signal input without requiring a low oscillator gain control
setting.
4. Measure all calibrating resistors with an accurate bridge, or use
a calibrated resistor decade box for the resistors.
5. Check the waveform with an oscilloscope. The waveform quickly
indicates reversed bias connections and overloads.
68
FUNDAMENTALS OF TRANSISTORS
TRANSISTOR
SYSTEM
>°
(A)
rj < O SERIES CONNECTION
(B)
TRANSISTOR
SYSTEM
Fig. 418. Equivalent voltage
method of measuring system input
or output resistance.
H<0 PARALLEL CONNECTION
(C)
Equal Voltage Method. The equal voltage method is a quick way
of determining the input or output resistance of a system when the
equipment is limited. This connection is illustrated in Fig. 418 (A) .
Resistor R is a calibrated decade box or a helipot in series with
the effective input resistance of the system under test. Resistor R is
adjusted until its voltage drop V is equal to the input voltage V t . Since
the arrangement is a simple series circuit, the input resistance r 1 is then
equal to R.
Figure 418 (B) illustrates the equal voltage method for measuring
a negative resistance. In this case, a calibrated resistor R x having a
larger absolute value than that of the negative resistance is connected
in series with r^ Again resistor R is adjusted until V = V^ for which
R— K t = r t . For example, suppose a pointcontact transistor is operat
ing in its negative resistance region. When a resistor R x = 2,000 ohms
is placed in series with the input, it brings the circuit into its positive
input region (stable operation) . When the connection of Fig. 418 (B)
is set up, R = 1,225 causes V to equal Vj. Then r, = RRj = 1,225 
2,000 = 775 ohms.
Notice that this latter arrangement requires that R be greater than
the absolute value of r x . If the only calibrated resistors available are
low in value, the parallel method illustrated in Fig. 418 (C) can be
used. The procedure is the same as before except that when V = Vi, R
is equal to R x and rj in parallel,
R 7 Ri + r,
which in terms of the input resistance becomes:
RR t
ri R x R
GROUNDED EMITTER AND COLLECTOR TRANSISTORS
69
For the same transistor measured above, if R t = 500 ohms, R is ad
justed to 1,408 ohms, at which time V = V t . Then r,= g »» , , no =
500 — 140b
— 775 ohms.
Transistor Test Sets. Several elaborate transistor test sets are avail
able commercially. These testers are useful for largescale experimental
work, since they incorporate means for completely evaluating the char
acteristics of all types of pointcontact and junction transistors, and do
not require external test equipment and meters. The home experi
menter and the lab technician, however, can get satisfactory results on
breadboards, based on the techniques described on the previous pages.
In checking transistors during maintenance and repair, it is not
necessary to check all the transistor parameters. A check of two or three
of the performance characteristics will determine quickly whether a
transistor needs to be replaced.
Figure 419 illustrates a transistor check circuit which will measure
the current gain and saturation current with reasonable accuracy. The
operation procedure and general functional description of the circuits
follows:
1. With switch SW2 in the calibrate (CAL) position and switch
SW1 in the current gain (o) position, adjust the signal gain of the
R ■ 600 OHMS
R 2 ■ 0.1 MEGOHMS
Rj ' 60,000 OHMS
R 4 ' 10 MEGOHMS
R 5 = 100 OHMS
C,,C 2 ' 4pf, 23 VOLTS
L • 10 HENRYS
SWI.SW2 DOUBLE POLE
DOUBLETHROW
SWITCHES
6 VOLT
BATTERY ■^
Fig. 419. Tramiitor tetter for measuring a and l c
70 FUNDAMENTALS OF TRANSISTORS
audio oscillator for one volt across resistor Rj. Now throw SW2 to the
current gain (a) position. The signal is now connected to the base of
the transistor through resistor R 2 and the dc blocking capacitor C^.
Since resistor R 2 is 100,000 ohms, the base and emitter resistances of
the transistor are negligible; ac base current i b , 10 microamperes.
2. The dc base current bias is controlled by resistors R 3 and R 4 ,
which permit a variation of from about 1 to 100 microamperes. R 4 is
adjusted until the collector dc bias current, measured by meter M,
is one milliampere.
3. Practically all of the ac output appears across the 100 ohm re
sistor R 5 , because of the high impedance of choke coil L (over 60,000
ohms at 1,000 cps) , and the high output resistance of the transistor
(usually more than a megohm) . The output voltage across R 5 is a i b R 5 ,
and since i b = 10 microamperes, R 5 = 100 ohms, this voltage equals
.001a. The value of a may vary from 10 to 100. The ac voltage may,
therefore, range from .01 to .1 volt. Thus, the current amplification
can be taken directly on a low scale of a good voltmeter.
Due to the comparatively low value of R 5 , the measured reading
closely approximates the maximum current gain a = r 12 /r 21 . This value
of current gain for the grounded emitter connection can be converted
into approximately equivalent values for the grounded base and ground
ed collector circuits by means of the following conversion formulas:
0015 Eq. {421)
OQB
1 +
OGE
and ^=22* Eq. (422)
where o GE = maximum current gain for grounded emitter connection;
oqb = maximum current gain for the grounded base connection; and
oqc = maximum current gain for the grounded collector connection.
These relationships are derived by neglecting r e and r b in comparison
with r m , r c and (r c — r m ) . Error in this approximation is negligible.
For example, assume that a transistor is tested in the circuit of
Fig. 419 and produces a reading of .022 volt on the ac output volt
meter connected across R 5 . The current gain
.022 22 _ 22
<*ge = oQi ' = 22; oob — — j . oo — = 96; oo C = gg = 22.9
The saturation current is read directly on the milliammeter M if
switch SW1 is now placed in the I„, position. This switch opens the
base lead, removing the bias, and also shorts out the inductor L so that
the sixvolt battery is across the emitter and collector electrodes.
The circuit as shown is only suitable for NPN junction transistors,
but can be modified easily for the PNP type by incorporating a switch
to reverse the battery, the meter connections, and the dc blocking elec
trolytic capacitors.
Chapter 5
TRANSISTOR AMPLIFIERS
This chapter deals with the design and operation of the transistor
as a lowfrequency amplifying device based on the transistor character
istics and limitations discussed in the preceding chapters. Since it is
impracticable to cover every useful type of connection, the emphasis in
this section is on fundamental illustrations, such as choosing the tran
sistor dc operating point, stabilizing methods, matching, direct coupl
ing, and cascading class A and B singleended and pushpull transistor
amplifiers. Some of the unique properties of transistors that are at
tained by the symmetrical operation of the NPN and PNP types in
the same circuit are also considered.
Grounding the Transistor System
Some confusion exists about which electrode should be connected
to ground in a transistor system. The basic reason for the difficulty lies
in the terminology: grounded base, grounded emitter, and grounded col
lector. Actually, these designations do not refer to the circuit ground,
but only specify which of the three electrodes is common to both the
input and output circuits. A better way to specify the three basic con
nections would be: common base, common emitter, and common col
lector. These latter designations are used by many authorities. In gen
eral, the system ground can be made at any convenient point in the
circuit, without consideration to the type of connection.
The DC Operating Point
Limitations, Supply Voltage and Load. As in the case of the vac
uum tube, the problem of designing a transistor amplifier is somewhat
Inpn I
\ I.670X
INPUT S
800 MO 1
PEAKTO PEAK
2 4 6 6 10 12 14 16 18 20
COLLECTOR CURRENT MILLIAMPERES I
Fig. 51
dc
MAX DISSIPATION
 100 MW
operating point.
I C MAX
Fig. 52 (above). Fixedbias
operation.
71
72 FUNDAMENTALS OF TRANSISTORS
simplified if the ac signal is treated independently of the dc operating
point. The first step in the design could logically be the selection of
the dc operating point. (Actually, three separate conditions must be
fixed; the operating point, the load line, and the supply voltage. In
general, the selection of any two automatically limits the determination
of the third.) The dc operating point may be placed anywhere in the
transistor characteristics, limited however by the collector maximums
of voltage, current, and power dissipation. The final selection of the
operating point is based primarily on the magnitude of the signals to
be handled.
Suppose, for example, a transistor, whose characteristics are illustra
ted in Fig. 51, is to be used with its operating point set at E c = 10 volts,
I c = 6 ma. Assume, also, that the maximum limits of the transistor are
I c = 18 ma, E c = 30 volts, and collector dissipation = 100 milliwatts,
as shown enclosed by the dotted lines. The supply voltage required is
the value at the intersection of the load line and the collector voltage
axis. Thus, for a fixed load of 1,670 ohms, the necessary supply voltage
is 20 volts. If, however, the supply voltage is fixed, then the load re
sistance is determined by the line joining both the supply voltage (E c
at I c = 0) and the operating point. As an illustration, assume the sup
ply voltage is to be E bb fixed at 30 volts. The resulting load resistance
E bb E c = 30 10 Qhms
I c 6 x 10 3
For any selected operating point there are many combinations of load
resistance and supply voltage that will permit the load line to pass
through the dc operating point.
The usual problem is one in which both the load and supply
voltages are fixed. The problem then resolves itself into a choice of
the operating point. In Fig. 51, for the conditions R L .= 1,670 ohms,
and E bb = 20 volts, the dc operating point may be placed anywhere
along the load line. It is usually desirable to design the amplifier for
maximum signal handling capacity. In this case, then, the dc operating
point should be midway between the extreme limits of the base current,
namely and 800 microamperes. The choice of I b = 400 microamperes
sets the operating point for maximum signal capacity at I t . = 6ma, and
E c = 10 volts.
Fixed Bias. The collector bias conditions, then, fix the dc bias cur
rent I b of the input base electrode; conversely, the base bias current
fixes the collector bias for a given load and supply voltage. The desired
base bias current can be obtained by connecting a resistor between the
base and the collector terminal of the supply voltage as shown in Fig.
52. For E bb = 20 volts, and I b = 400 microamperes, the total series re
F 20
sistance is^ = „ *" , = 50,000 ohms.
I b 400 x 10 6
TRANSISTOR AMPLIFIERS
73
4 6 S 10 12 14 16 18 20
COLLECTOR CURRENT I c
NORMAL CASE
(A)
4 6 8 10 12 14 16 18 20
COLLECTOR CURRENT I e
LOW I c0
IB)
30"
o
1
1
20
10
Fig. 53. Variation of operating
point.
B io
2 4 6 8 10 12 14 16 18
COLLECTOR CURBENTI e
HISH I C
(C)
This value includes the emitter to base resistance, but since r c  r b is
generally only a few hundred ohms, they can be neglected. The result
ing circuit, with the calculated values, is illustrated in Fig. 52 for a
NPN transistor. If the same characteristics were applied to a PNP
type, the only circuit change would be a reversal of the supply battery
potentials. The transistor bias indicated in this figure is called fixed
bias.
SelfBias. Unfortunately, transistors are temperature sensitive de
vices; in addition some variation usually exists in the characteristics of
transistors of a given type. These factors may cause a displacement of
the constant base current lines along the collector current axis. Figure
53 illustrates the effect of this variation; the abnormal cases are pur
posely exaggerated. Notice the effect of this shift on the relative posi
tions of the dc operating point. In the low I ro unit (Fig. 53B) the
collector voltage is too high; in the high I^unit (Fig. 53C) the collector
voltage is too low. To overcome this, the circuit needs degeneration,
similar to that produced by an unbypassed cathode bias resistor in a
vacuumtube circuit. In transistor circuitry, this method of degeneration
is a form of automatic control of the base bias, known as self bias.
A simple method for establishing automatic control of the base
bias requires the base bias resistor to be tied directly to the collector,
as in Fig. 54. Thus, if the collector voltage is high (Fig. 53B) , the
base current is increased, moving the dc operating point downward
74
FUNDAMENTALS OF TRANSISTORS
rzT} —
I 25K Y
• 1 NPN
INPUT
eoo pa
PEAKTOPEAK
IS70<
OHMS<
E bb±
20 VOLTS T
*bl 
■ 700 pa I
Fig. 54 (above). Selfbias operation.
Fig. 55 (right). HunterGoodrich bias
method.
along the load line; conversely, if the collector voltage is low (Fig.
53C) , the base bias current is decreased, moving the dc operating point
upward along the load line. The value of the selected base bias resistor
is different in the selfbias case from that computed in the fixedbias
connection. For self bias, the resistor is tied to the collector voltage,
E 10
which in this case is 10 volts. Then R B = — ^— = . — . 6 = 25,000
ohms. The base bias resistor performs the double duty of determining
the value of I b and preventing those excessive shifts in the collector dc
operating point due to temperature change and transistor interchange.
The principal limitation of self bias is that it still allows some variation
of the dc operating point, since the base bias resistor is fixed by the
required operating point, and the stabilization produced by it is only
a secondary effect. In addition, self bias also introduces ac negative
feedback which reduces the effective gain of the amplifier. Despite its
limitations, however, self bias is very useful and works well in many
applications.
The importance of temperature stability with respect to the dc
operating point cannot be taken lightly. ,One of the effects of a tem
perature rise is to increase the saturation current I co , which, in turn,
increases the collector dissipation. The increased collector dissipation
increases the temperature, which increases I co , and so on. Thus, poor
temperature stability almost certainly will cause transistor burnouts,
particularly if the transistor is operated near its maximum dissipation
limit.
HunterGoodrich Bias Method. A method of establishing tighter
control on the base bias current illustrated in Fig. 55 is the Hunter
Goodrich method. This involves the addition of a fixed base bias op
erating in the reverse direction of the normal self bias. The fixed bias
is introduced by resistor R F and separate voltage supply E P . To over
come this reversed fixed bias, the self bias resistor R B must be de
creased to maintain the same base bias current. The reduced value of
R B increases the available negative dc feedback from* the collector cir
cuit, thus providing greater transistor stability.
TRANSISTOR AMPLIFIERS
75
As in the preceding cases, the effect of the base and emitter circuit
resistances (r e f r b ) can be neglected in the calculations. The values
of R F and E F depend upon the value of fixed bias desired. For example,
assume that a fixed bias value I b2 of 300 /ia will provide the additional
stability needed, and a battery E F = 10 volts is available. Then R P —
F 10
— — ~ 37^ — TTcz = 33,300 ohms. The current through the self bias
300 x 10 8 °
I
b2
resistor R B is I M = I„ + I b2 = 400 f 300 = 700 ^a; then R B =
I,
10
bl
= 14,300 ohms.
— 700xl0«
In comparison, R B = 25,000 ohms in the simple self bias case. Since
the input resistance of the transistor is small compared to R F , prac
tically all of the stabilizing current flows into the baseemitter circuit.
The HunterGoodrich bias method is extremely useful when a high
degree of circuit stability is needed. Its particular disadvantage is that
it requires two separate battery supplies.
Self Bias Plus Fixed Bias. One method of obtaining additional sta
bilization with only one battery is shown in Fig. 56 (A) , the basic fea
tures of which are often used in transistor power stages. The fundamen
tal differences between this circuit and the preceding fixed plus self
bias method are the interchange of R L and E bb , and the connection of the
reverse bias resistor R F into the collector circuit. Interchanging the
supply battery and the load resistor provides two points at which varia
tions in collector voltage will appear. However, this interchange does
not affect the dc operation of the circuit. Connecting R F , as illustrated,
produces essentially the same result as the HunterGoodrich arrange
ment, except that the reverse bias is no longer fixed. If the previous
E c _ 10
circuit constants are desired: R P =
I,
All the other values remain the same.
b2
300xl0«
33,300 ohms.
l bi
 700pa I
o
14.3 K
 — VW
lb
c 400(1
I — vw
[npn __
N ibz f
E bb
Ibz "T 20
300110 V °>TS
(A)
Fig. 56. (A) Stabilization of dc operating point with one battery. (B) Typical power
output stage.
76
FUNDAMENTALS OF TRANSISTORS
In power amplifier circuits, the load usually consists of a trans
former plus an additional stabilizing resistor. Figure 56 (B) illustrates
one possible form of this arrangement for use as a transistor power
amplifier stage.
A disadvantage of this bias method is that the dc degeneration
feedback is reduced, due to the shunting effect of resistor R F , thus re
ducing the stabilization. On the other hand, this method provides for
greater stability than does the simple selfbias method. It provides less
stability than the HunterGoodrich method, but requires only one bat
tery supply.
Current Sources. Notice that all the bias requirements are supplied
by conventional batteries, which act as constant voltage sources. At this
point the conscientious reader may wonder if this does not conflict with
the statements in earlier chapters that transistors are currentoperated
devices. Actually, the term "current source" is more than just a math
ematical concept. The practical aspect can be shown as follows: Assume
that a sixvolt battery with negligible internal resistance is connected to
a variable load resistance. Except for very low values of load, the bat
tery terminal voltage remains constant as long as the battery remains
fully charged. Now assume that a one megohm resistor is connected in
series with the battery and the load resistor. In this case the current re
mains reasonably constant while the load resistance is varied from zero
to about 0.1 megohm. Thus, the addition of a series resistor has con
verted the constant voltage supply into a constant current source over
a fairly wide range of load resistance values. The range depends upon
the value of the series resistor. Figure 57 illustrates the basic equivalent
interchanges of supply sources. Mathematically, all that is involved is
sfsxc
■v )e
E'IX
VOLTAGE
=FXc
Fig. 57. Equivalent volt
ag«curr»nt tourcei.
x c
CURRENT
E
X L
TRANSISTOR AMPLIFIERS
77
2lf
jCT
835 5
ohms;
AAA/ — "
25K
835 <
OHMS"
OUTPUT
O
Fig. 58. Clan A amplifier.
50
=J
the movement of the impedance proportionality constant from one side
of the equation' to the other.
That these circuits are equivalent can be shown by a simple ex
ample. Take the case of a sixvolt battery in series with a resistor R = 1
megohm and a load R L =: 1 megohm. Then the load current equals
E 6
d — "p = "71 r\ — wm = ^ microamperes. The equivalent circuit on a
K. \ Kl (1 f 1) x 10°
current basis is a current generator I =
R ~ lxlO« ~
microamperes, which is shunted by both a resistor R = 1 megohm, and
Fig. 59. (A) Distortion due to
crowding of collector characteristic.
(B) Effect of input distortion on
output wave.
78 FUNDAMENTALS OF TRANSISTORS
a load R L = 1 megohm. Now the load current equals the source cur
rent less the amount shunted by resistor R. Since R and R L are in paral
lel, the voltage drop across each resistor must be the same, and the
load current equals
IR 6xlO«(IxlO«) Q .
R+R L = (1 + I)xl0« = 3 rmcroamperes.
This checks with the previous result. The same procedure can be
used to convert an ac voltage source into an ac current source.
Class A Amplifiers
Basic Circuitry — Efficiency Stabilization. Figure 58 represents a
typical Class A transistor amplifier using dc operating biases as de
scribed in the preceding paragraphs. The stabilizing resistor in the
emitter circuit is made equal to the load impedance in this case. This
condition provides maximum protection against variations in 1,^, since
the power available from the battery is effectively limited to the maxi
mum collector dissipation. While the arrangement shown in Fig. 58
prevents transistor damage due to excessive collector variations, half
of the dc power is dissipated across the stabilizing resistor. The maxi
mum efficiency of a class A transistor is 50%; using a stabilizing re
sistor whose value is equal to the load resistor reduces the efficiency
to a maximum of 25%.
In general, this amount of stability control is needed only in mass
production applications if transistors having a wide tolerance range are
to be used. Actually, the reproducibility of transistor characteristics has
improved rapidly during the past few years. There is no reason why
this trend should not continue, and eventually permit the attainment
of amplifier efficiency values very close to the theoretical maximum.
In most circuits, between 5 and 10% of the collector dc power (E c I c )is
satisfactory for normal stabilization. In the circuit shown in Fig. 58,
for example, a resistance of 100 ohms between the emitter and ground
would be sufficient.
Bypass and Coupling Capacitors. If the stabilizing resistor in the
emitter lead is unbypassed, the amplifier gain is decreased. This is simi
lar to the action of an unbypassed cathode resistor in a vacuumtube
amplifier. A value of 50 /J works out well for the bypass capacitor in
most audio frequency applications. The self bias resistances from col
lector to base may also be suitably bypassed to avoid ac degeneration.
In cascaded stages, the load is unusually low, and the ac collector volt
age is also low. In this case, the bypass capacitor can be omitted with
only a slight loss in the stage gain.
The value of the coupling capacitor C c must be large enough to
pass the lowest frequency to be amplified. Usually a maximum drop of
3 db in gain is permitted. At this value, the reactance of the coupling
capacitor is equal to the input resistance of the stage. Since the input
TRANSISTOR AMPLIFIERS 79
resistance is low for the grounded emitter and grounded base connec
tions, relatively high capacity coupling condensers are required. For ex
ample: What is the minimum value of C c necessary for coupling into
a stage whose input resistance r, = 500 ohms, if a frequency response
down to 100 cps is required? At 100 cps,
x < = air ri = 50 ° ohms ' and Cc = 2ir = 2.(ioo) (5oo) = 32 f
In typical circuits, the required value of the coupling capacitor varies
from 1 to 10 pi.
Distortion. Another characteristic of a transistor amplifier that
should be mentioned is the harmonic distortion. If the circuit is wired,
the distortion can be measured directly, using a suitable wave analyzer
or distortion meter. In addition, distortion can be calculated under
given operating conditions from the collector characteristics, using the
same methods as in vacuumtube amplifiers. These methods are de
scribed in detail in most radio engineering handbooks. The total har
monic distortion is about 5% in the typical transistor amplifier. It is
caused mainly by the decreased spacing between the collector current
collector voltage curves for equal changes in base current. This crowd
ing effect occurs at the higher values of collector current. Figure 59 (A)
is an exaggerated illustration of this type of distortion in transistor
circuits.
If the input resistance of the amplifier is high compared to the
source impedance, another type of distortion, due to variations in the
input circuit, is introduced. In the region of low collector current, the
input resistance increases, thus reducing the amplitude of the input
signal. In the high region of the collector ac current cycle, the input
resistance decreases, thus increasing the amplitude of the input signal.
This type of nonlinear distortion is illustrated in Fig. 59 (B) .
Since the two major types of distortion described above have oppo
site effects, it will be possible to counteract one with the other by ad
justing the value of the signal generator impedance. In troubleshooting
multistage transistor amplifiers, an output waveform similar to that
indicated in Fig. 59 (B) would probably indicate a defect in one of the
preceding stages.
When computing the harmonic distortion of a transistor amplified
by conventional vacuumtube graphical methods, the computed value
is generally in the order of 1% less than the measured value. This is
caused by the assumption that the signal generator resistance is neg
ligible, a condition seldom realized in low input resistance transistor
circuits. The source resistance in vacuumtube amplifiers does not af
fect the determination of the harmonic distortion, since the grid current
IS zero. However, the equivalent parameter in transistor amplifiers, the
baseemitter voltage, is not zero. The effect of the source may be taken
80
FUNDAMENTALS OF TRANSISTORS
&: OPTIMUM LOAD FOR
LOW SIGNAL
LIMITING CHARACTERISTICS
B: DPTIMUM LOAD FOR
LARGE SIGNAL
r EC MAX
(BL'UOOOHMS
(A)
4Uf
\" 'o'SSJcLH
INPUT 1 ^/V/^.
66.7K
J
OUTPUT
o
(C)
Fig. 510. (A) Load fines for maxi
mum power. (B) Determination of
optimum toad. (C) Power amplifier
circuit.
into account by considering it as part of the base resistance. However,
in most applications, it is more than satisfactory to simply add 1% to
the calculated value of percentage distortion.
Maximum Output Conditions, Since the power handling capacity
of the transistor is small compared to that of the vacuum tube, it is
usually necessary to drive the transistor to its maximum limits. When
a transistor amplifier is designed for maximum power, as in Fig. 5
10 (C) , it is properly termed a power amplifier, although the actual
power involved may only amount to a hundred milliwatts. To obtain
maximum power, the load line is selected to include the maximum
possible area concurrent with the fixed limitations of maximum collec
tor dissipation, current, and voltage. The ideal load for maximum
power would be one which followed exactly the transistor limit bounda
ries illustrated in Fig. 51.
In the practical case, it is usually necessary to settle for a load
line that is tangent to the limiting characteristic line. Since this curve
is nonlinear, there are several possible choices of load. The final choice
depends primarily on the signal requirements of the circuit. Figure
510 (A) shows the two extreme cases: line A h the optimum load for
a small signal input; line B is the optimum load for a large signal input.
Since, however, the supply voltage is usually specified, the load line
chosen is the one that is tangent to the limiting curve and that passes
through the specified supply voltage (E,. at I c = 0) .
The calculations for determining the conditions for maximum out
put in a power amplifier stage, assuming a transistor with the charac
teristics illustrated in Fig. 510 (B), are as follows: Assume a battery
TRANSISTOR AMPLIFIERS
81
supply E bb = 20 volts. This determines a point on the load line for
I c — 0. Now hold one end of a straight edge on this point and swing
the edge until it just touches a point on the maximum collector dissi
pation line. Draw a line through the two points. This is the optimum
load line for the given conditions. The maximum input signal, the dc
bias resistors, and the distortion can all be computed directly from
the figure by means of the methods discussed in the preceding para
graphs:
E e (atl e = 0) _ 20
_ I c (atE c =0)
E bb _ 20
R T
1,100 ohms.
R„ =
.0182
= 66,670 ohms.
I b 300x10°
Using the standard equation derived for vacuumtube circuits, the tran
sistor ac output P ac is oneeighth of the product of the peaktopeak
collector voltage and collector current:
P =
EpAp .20(18.2x103) _
45.5 milliwatts.
8 8.
Since the maximum dc power is E C I C = 11 x9x 10~ 3 = 99 milliwatts,
P 45 5
the efficiency becomes p flC x 100 = ' x 100 = 46%.
EX
99
The maximum ac signal current can be taken directly from the char
acteristic curve. In this case, i„ = 600 ^a peaktopeak. (The stabilizing
resistors have been omitted to simplify the illustration.)
PushPull Operation. Whenever possible, transistor power ampli
fiers should be operated as pushpull stages. Pushpull operation has
several desirable features, including the elimination of the evenorder
Fig. 511. Class A pushpull
amplifier.
82
FUNDAMENTALS OF TRANSISTORS
OHMS 21 12 OUTPUT INPUT
20 VOLTS I £)
Fig. 512. (A) Clan B circuit (constant voltage). (B) Class B pushpull operation.
harmonics and the dc component in the load. The first factor is par
ticularly fortunate, insofar as transistor applications are concerned. It
was noted previously that operation at high values of collector current
introduces a distortion due to crowding of the collector currentvoltage
lines. Thus, for a given value of allowable distortion, pushpull opera
tion will allow the transistors to be driven into the higher I,, regions.
In turn, each transistor delivers more power to the load than when it
is connected for singleended operation.
The operating point, load, and biasing resistors for the Class A
pushpull stage are determined for each transistor exactly as if it were
a singleended type. A typical pushpull transistor amplifier is illustra
ted in Fig. 511, based on the same transistor characteristics used pre
viously. The separate biasing arrangement indicated in this illustration
permits a more exact match of the transistor characteristics. Notice that
the load is twice the value computed for the singleended stage.
Class B Transistor Amplifiers
Basic Operation — Quiescent Point. While the efficiency of a Class
A amplifier is good under operating conditions, the collector dissipation
is approximately the same whether or not a signal is applied. Its effi
ciency for intermittent or standby operation is poor. For standby opera
tion, as in the case of the vacuum tube, Class B operation is preferred,
and the operating point of a Class B transistor amplifier should be on
the E c = line. This bias condition, however, would require an ex
tremely high resistance in series with the battery. Thus most of the
available supply power would be lost in the series resistor, the only
function of which was to convert the voltage source into a current
source. As an alternate method, a constant voltage battery is used. This
sets the dc operating point at the collector voltage E c = E bb on the
I c = axis. Figure 512 (A) shows a typical Class B transistor amplifier
with a constant voltage source, using the same transistor as in previous
calculations.
TRANSISTOR AMPLIFIERS
83
PushPull Circuitry. Two Class B amplifiers connected as a push
pull stage, using two of the circuits illustrated in Fig. 512 (A) , will not
operate. One transistor will always be biased in the reverse direction
by the input signal, thereby causing its input resistance to become very
high. This condition can be eliminated by using a centertapped input
transformer and connecting the center tap to the common emitter elec
trodes. This circuit is characterized by a distorted output wave. The
distortion is particularly evident when the signal generator resistance
is low. However, the distortion can be reduced within limits by intro
ducing base bias into the circuit.
Figure 512 (B) illustrates one possible form of this latter arrange
ment. The value of the base bias resistor R F for minimum crossover
distortion can be determined by the conventional graphic methods of
vacuumtube Class B pushpull amplifiers when using the composite
transistor characteristics. The proper bias setting may be determined
experimentally by direct measurement with an oscilloscope or a distor
tion meter. If the experimental method is used, care must be taken to
avoid setting the base bias too high. This would cause a relatively high
quiescent dc collector current to flow, and the circuit would perform
in a manner similar to that of a Class AB amplifier in vacuumtube
circuits. Resistor R c may be a thermistor or some other temperature
sensitive device. R c is usually required in stages, subject to large changes
in temperature to prevent excessive variation in the collector dc op
erating point.
Another arrangement for a transistor pushpull Class B stage is
illustrated in Fig. 513 (A) . This circuit permits the elimination of the
input transformers. The diodes Di and D 2 prevent each transistor from
DIODES OMITTEO
BIAS RESISTORS
OF INCORRECT
VALUE
ALL ELEMENTS
CONNECTED AND
CORRECTLY CHOSEN
(B)
Fig. 513. (A) Class B pushpull operation without input transformer.
(B) Output waveforms.
84
FUNDAMENTALS OF TRANSISTORS
cutting off when it is biased in the negative (reverse bias) direction
by the input signal, since the diodes effectively short out the signal
induced bias. The point at which this bypass action occurs is deter
mined by the bias due to resistors R F and R c . These resistors also fur
nish base bias to the transistors to minimize crossover distortion. Figure
513 (B) illustrates the effect of diodes and bias resistors on distortion
of the output signal.
The detailed operating characteristics of a Class B transistor push
pull amplifier are determined by the same methods used in similar
vacuum tube circuits. The approximate values of the major character
istics can be calculated as illustrated in the following example: Assume
that the transistors to be used in the Class B pushpull circuit have a
maximum collector dissipation rating of 100 milliwatts, and assume
that a battery E bb = 10 volts is specified. The collector dissipation P c in
each transistor is approximately
EbtJpc
where I pc is the peak collector
current. Then I pc =
8P„
8(100)
10
•= 80 milliamperes. The required
load for maximum power output is: R L =
4E bb _4(10)
and the power output is approximately
Ebblpo
.08
10(80)
= 500 ohms;
400 milli
watts, or four times the maximum collector dissipation of each transistor.
Phase Inverters
Function. Transistor pushpull amplifiers, like their vacuumtube
counterparts, require the use of a phase inverter to supply the required
balanced signal input. Transistor inverters are more complicated than
conventional vacuumtube types in that they must provide a balanced
current, rather than a balanced voltage, input signal. However, the prin
ciples of operation are essentially the same.
44f
Fig. 514. Transistor phase
inverter.
25K
TRANSISTOR AMPLIFIERS
85
»£Ht
SATISFACTORY OUTPUT GAIN CONTROL (A)
UNSATISFACTORY OUTPUT GAIN CONTROL (B)
SATISFACTORY INPUT GAIN CONTROL (C) UNSATISFACTORY INPUT SAIN CONTROL (D)
Fig. 515. Gain controls.
Typical Circuit. Figure 514 illustrates the basic circuit of a tran
sistor phase inverter, which provides a reasonably wellbalanced output.
The basic operation is as follows: the upper transistor operates as a
conventional grounded emitter amplifier except that the emitter is
grounded through the parallel circuit, consisting of the lower transistor
emitterbase path and resistor R E . The emitterbase path has a low re
sistance, less than 50 ohms, so that practically all of the ac emitter cur
rent of the top transistor flows through this path. Since the emitter
current value for each transistor is the same, the collector currents are
also equal if the current gains from emitter to collector are equal. For
proper operation, the load resistances should be small compared to
the output resistances of the transistors, and the emittertocollector cur
rent gains should be well matched. For the circuit illustrated, the out
put resistance of each transistor is the collector resistance shunted by
R B . Since r c is much greater than R B , the output resistance is equal to
R B . Thus R B should be about ten times R L . It is not necessary for the
current gains to be exactly matched. Values which fall in the range of
.92 to .97 are usually satisfactory. R L and R B are selected to provide the
operating biases, which in this case are E c = 10 volts, I c = 4 ma, and
I b = 400 ^a. The value of R E is particularly important. It must be large
compared to the emittertobase resistance path of the lower transistor;
if it is not, an appreciable portion of the ac signal will be shunted
through R E and the currents in the emitters will not be equal. In
general, a value of R E that is ten times the emittertobase circuit resist
ance is satisfactory.
Transistor Gain Controls
Despite the relatively low gain of transistor amplifiers, a gain con
trol is frequently necessary to compensate for changes in the input sig
86
FUNDAMENTALS OF TRANSISTORS
a* ■ ie.8 ct2 *ie.i a,3«i9.a«
13
— 
— *•
X i
a:,.
*z
<*3'
k
^3.
'2
'3
*
rKxi
Fig. 516. Block schematic of caicade
operation
SYSTEM OPERATING POWER CAIN ■ 4,650,000
Fig. 517. Calculated threestage cascade.
nal, the ambient noise level, and other variations. The design of volume
controls for transistor circuits is not a difficult problem if the fact that
transistors are current operated devices is kept in mind. Figure 515 (A)
illustrates one possible form of output gain control in a RC coupled
stage. In this circuit, the output potentiometer sets both the collector
dc operating point and the level of the output signal. The coupling
capacitor blocks dc current from flowing into the load. The value of
this capacitor must be large enough to pass the lowest frequency to be
amplified. If the output load is a transformer, this same form of gain
control is not satisfactory, since, as illustrated in Fig. 515 (B), the load
impedance varies with the potentiometer setting. If the coupling capa
citor is omitted, circuit operation is poorer because the volume control
setting changes the dc operating point.
Figure 515 (C) illustrates a satisfactory form of input volume con
trol in a transformer coupled stage. The resistance of the potentiometer
should be at least ten times the value of the secondarywinding imped
ance to make its loading effect negligible. The arrangement illustrated
in Fig. 515 (D) , however, is not satisfactory, because the base bias varies
with changes in the volume control setting.
In multistage operation, the gain control may be located in the
input or output circuit of any stage. It is usually desirable to place the
control in the first stage if the signal amplitude is likely to vary ap
preciably. This arrangement helps to prevent the system from overload
ing on large signals.
Cascade Operation
Design Considerations — Overall Power Gain. In any given prob
lem requiring more than one stage of amplification, several cascade
arrangements are possible. This flexibility is a desirable design feature;
however, it complicates the problem of selecting the best combination
of the three general forms of transistor connections with respect to the
input and output resistances, and to the required gain of the system
Every design is fixed to some extent by the function of the circuit, but
the requirement for maximum gain is invariably included.
Figure 516 is the block schematic of a three stage circuit. It is
evident from inspection that the overall current gain of the system is
TRANSISTOR AMPLIFIERS 87
the product of the individual stage gains, thus a = aia 2 a3 The operat
ing gain as defined in equation 345 is
G = 4RgR L
which can be modified to
r 2 i
_(R g + r n ) (R L + r 22 )  r 12 r 21 _
G = 4R « R fcw)
R + r »bSM
Since the current gain as defined in equation 38 is: a = p .
KL + r 22
and the input resistance as defined by equation 313 is r ( = r u —
[ ^24^ — \ , these values may be substituted in the operating gain
\ *22 j K L /
equation, which then becomes
g =tof *•■<">*
This is a useful form of the equation. For the cascade stages, illustrated
in Fig. 516, the overall power gain based on equation 51 can now be
written as
g =(t^f] * R '
1st Stage "2nd Stage 3rd Stage
On this basis, a cascade system has maximum gain when each of the
stages is separately designed for a maximum value of its associated gain
factors.
Selection of Stage Connection. The first stage requires that its
R. 2
gain factor — /r /^ 1 be as large as possible. The following general
(K g f r,) *
rules for this stage are based on an analysis of the gain factor vs R g
characteristic:
1. When R g has a low value (0 to 500 ohms) , use either the
grounded base or the grounded emitter connection.
2. When R g has an intermediate value (500 to 1 ,500 ohms) , use
the grounded emitter connection.
3. When R g has a high value (over 1,500 ohms) , use the grounded
emitter or the grounded collector connection.
In the intermediate stage a2 2 is made as large as possible. This re
quirement generally can be met by either the transistor grounded emit
88 FUNDAMENTALS OF TRANSISTORS
ter or grounded collector connection. The intermediate stage equivalent
load should be less than (r c — r m ) . If r c is nearly equal to r m , the ground
ed collector should not be used. This equality would cause the input
and output resistances of the stage to become independent of the values
of the connecting circuits. (An analysis of this buffer effect was covered
in the discussion of input and output resistance of the grounded col
lector stage in Chapter 4.)
It must be noted that the intermediate stage represented by the
current gain 02 in this discussion may actually consist of several inter
mediate stages having a total current gain equal to 02 This analysis
of the threestage circuit of Fig. 514, therefore, is applicable to any
number of cascaded stages.
In the final stage, the gain factor R L a 8 2 is made as large as possible.
The following general rules for this stage are based on the analysis of
the gain factor vs R L characteristic for the three basic transistor con
nections.
1. When R L has a small value (0 to 10,000 ohms) , use a grounded
collector or a grounded emitter connection.
2. When R L has an intermediate value (10,000 to 500,000 ohms) ,
use a grounded emitter connection.
3. When R L has a high value (over 500,000 ohms) , use a grounded
emitter or grounded base connection.
(The numerical values listed above apply to those junction tran
sistors with characteristics similar to the Western Electric Type 1752
transistor; however, the general values can be extended on a relative
basis to cover all types.)
Based on the foregoing rules, it might appear that the choice of
the grounded emitter connection is the best under all conditions. How
ever, specific design problems often dictate the use of grounded base
and grounded collector circuits when the coupling network, biases, feed
back, and other factors are taken into consideration.
Cascade Design. As an illustration of these principles, consider the
design of a threestage cascade system using the typical junction tran
sistor with r e = 50 ohms, r b = 500 ohms, r c = 1,999,500 ohms, and
r m = 1,899,500 ohms. Assume that R g is adjustable but limited to low
values. R L = 150 ohms requires the use of the grounded emitter or
grounded collector connections. Assume that other design factors limit
the choice to the latter case. Then for the last stage:
r n = r c + r b = 1,999,500 + 500 = 2,000,000 ohms;
r 12 = r c r m = 1,999,5001,899,500 = 100,000 ohms;
r 21 = r c = 1,999,500 ohms;
r 22 = r c + r e  r m = 1,999,500 + 50  1,899,500 = 100,050 ohms.
The input resistance of the last stage (equation 313) is expressed as:
TRANSISTOR AMPLIFIERS
89
I r 12 r 21 \ _
~ Tll ~ lr 22 + Rj
2 x 10 6 
0.1 xlO 6 (1,999,500)
100,050+150
5,000 ohms,
and the current gain (equation 38) is:
r 21 1,999,500
az — 
==== 19.99
Rl + t 2 2 ~" 150 + 100,050
Since r m is close to the value of r c , the intermediate stage is re
stricted to the grounded emitter connection. For this stage:
r n = r e + r b = 50 + 500 = 550 ohms;
r 12 = r e = 50 ohms;
r 21 = r e r m = 50  1,899,500 = 1,899,450 ohms; and
r 22 = r e + r c  r m = 50 + 1,999,500  1,899,500 = 100,050 ohms.
Since the input resistance of the last stage is the output resistance of
the intermediate stage, R L = 5,000 ohms. The input resistance of the
intermediate stage is
[50 (1,899,450)"
rn
_/__£i£ 2 j_\
U 2 + RJ"
l r 22 +
and the current gain is:
550
a 2 — 
r 2 i
100,050 + 5,000_
1,899,450
= 1,455 ohms
18.1
Rl + r 22 5,000 + 100,050
Since a low value of R g is specified, the first stage must use either
the grounded emitter or the grounded base connection. The load of
the first stage equals the input resistance of the intermediate stage and
is a low value. Therefore, the best choice for the first stage is the
grounded emitter connection. Since R g was specified as being adjustable,
its value will be made equal to the input resistance,
"50 (1,899,450)
= ril ir 22 + RJ~
^22 +
The current gain is: ai =
550
r 2 i
100,050+1,455.
1,899,450
1,487 ohms.
= 18.75
r 22 + R L ~ 100,050 + 1,455
The overall current gain of the cascaded system
a= ai a 2 a s = ( 18.75) (18.1) (19.99) = 6,780
The operating gain (equation 51) is
_ 4R,R L q* _ 4(1487) (150) (6780) » _ lfi5om
(Rg + r,) 2 (1487 + 1487) 2 W '
The resulting cascade circuit is shown in Fig. 517. This circuit does
not include biasing arrangements, coupling networks and feedback
loops. The values of the elements necessary for introducing these re
quirements may be computed by the methods in preceding paragraphs.
90
FUNDAMENTALS OF TRANSISTORS
The cascade system may be changed considerably by the addition
of external resistance arms to the circuits. These have the effect of in
creasing the effective values of the transistor parameters. For example,
consider the effect of adding a stabilizing resistor R E = 50 ohms in
series with the emitter arm of the input stage. The effective resistance
of the emitter is now r e } R E = 50 + 50 = 100 ohms, and the general
fourterminal parameters are now:
r n = r e f R E f r b = 50 + 50 + 500 = 600 ohms;
r 12 = r e f R E = 50 + 50 = 100 ohms;
r 21 = r e + R E  r m + 50 + 50  1,899,500 = 1,899,400;
r 22 = r e + R B + r c  r m = 50 + 50 + 1,999,500  1,899,500 =
100,100 ohms.
The input resistance
" 100(1,899,400)\ _ .
100,100 + l,455/ _ '
r, = r„ —
r 22 + RL
and the current gain ai =
= 600
1,472 ohms
1,899,400
= 18.72
6,770
r 2 2 + R L •100,100 + 1,455
The overall current gain a = aia 2 a 3 =  18.72(18.1) (19.99)
4(2472) (150) (6770) 2
(2472 + 2472) 2
and the operating gain G :
4R,R t
(Rg + r.) 2
= 2,790,000.
Thus, a simple change reduces the overall system gain by a factor of
onehalf. It is evident that even after the basic stage connections are
fixed, a considerable variation in the cascade performance and resist
ance terminal characteristics can be attained by changes in the effective
value of the transistor parameters.
Coupling and Decoupling Circuits. To obtain the absolute maxi
mum gain from a cascaded system, image resistance matching between
stages is required. The analysis and conditions for matching the three
basic transistor connections are covered in Chapters 3 and 4. The stage
<
<
lb
soo ua
<?
Cc
Fig. 518. (above). RC interstage coup
ling; X c less than r i at lowest frequency
to be amplified; R at least 10 times r,.
Fig. 519 (right). Typical decoupling
network.
PRECEDING
STAGE
Rl
fAAAr
NEXT
STAGE
TO COMMON
B+
TRANSISTOR AMPLIFIERS 91
can be matched by interstage transformers. In if strips, transformer
coupling is convenient and invariably used, because the transformers
are also required for selectivity. In audio circuits, however, the increased
gain due to the transformer is seldom worth its expense. In audio cas
cades, therefore, resistancecapacitance coupling is the most practical
and economical choice. Figure 518 represents a typical RC coupled
stage. The capacitance must be large enough to pass the lowest fre
quency to be amplified. Its value can be computed as indicated in the
preceding paragraphs dealing with single stage amplifiers. Resistor R
must be large compared to the input resistance T. The interstage loss
in gain is less than one db if R is chosen to be ten times as great as r t .
When cascaded stages are connected to produce an overall gain of
60 db or more, consideration must be given to the addition of a de
coupling circuit, as indicated by the combination RiC^, as shown in
Fig. 519. Decoupling is required to prevent positive feedback through
the battery resistance which is common to all the stages. Highgain
transistor cascades almost always require a decoupling network, since
even low values of battery resistance are significant when compared
to the low input resistance of transistor stages. The product of R t and
Ci (time constant) should be equal to or greater than the inverse of
the lowest frequency to be amplified by the stage. While this specified
frequency sets the time constant, there are any number of combinations
of C t and R t which can be used. In general, R t is made small enough
so that it does not affect the supply voltage greatly, and at the same
time is not made so low that a very high value of C t is required. The
following example illustrates the calculation of the decoupling network:
Suppose that for the circuit illustrated in Fig. 519, the dc base bias
I b = 500 fj.a, and a drop of onequarter of a volt in the battery supply
through R t can be tolerated. The maximum value of R t equals the al
.25
lowable voltage drop divided by the base current, R, = ■,..' , n =
° r ' 500 x 10 6
500 ohms. If 100 cps is the lowest frequency to be passed, then— = R^
and Ci = =r — = in/wen^ =20 fd. (In this equation, f is expressed
in cycles per second, R t in ohms, and C t in farads.) The value of Ci
depends on the allowable voltage drop through R t . If a larger drop is
allowable the value of C t will decrease proportionately. In this ex
ample, assume that only a 10 /if capacitor is available, and that the
maximum drop through R t can be increased. Then R lt for the same
cutoff frequency, equals — ■ —y^ — ^ — ,_ . =1000 ohms, and the
^ ^ fC x lOOxlOxlO 8
92
FUNDAMENTALS OF TRANSISTORS
voltage drop through R x equals Rjl b = 1000 (500 x lO" 8 ) =0.5 volt.
The base bias resistor now must be adjusted to compensate for the re
duced value of the effective supply voltage. Thus
E bb K 1 l b _ 120.5
R«=
I b 500xl0 6
as compared to the value (without decoupling) ,
E bb 12
= 23,000 ohms,
R«=
24,000 ohms.
I b _ 500xl0«
In general then, when the value of the decoupling resistor is significant
in comparison to the value of the bias resistor, R B must be decreased
by an amount equal to that of Ri to maintain the specified dc base
current. In the form of an equation, this condition can be specified as:
'bb _
lb
Rb+Ri
Figure 520 illustrates an experimental twostage amplifier using
grounded emitter circuits designed specifically to amplify the output
of a 50 ohm dynamic microphone. The output terminates in a 600 ohm
line. The overall gain of the system is 46 db.
ComplementarySymmetry Circuits
Basic Theory. The circuits discussed to this point can be used with
either NPN or PNP transistors. It is necessary only that the battery
supply is connected with the proper polarity. For other applications, it
is possible and often very profitable to combine the two types of junc
tion transistors into one circuit. This technique permits the design of
many novel configurations that have no direct equivalent in vacuum
tube circuits, since no one has yet invented a vacuum tube that emits
positive particles from its cathode. Some of the characteristics of this
unique property of transistors can be illustrated with the help of Fig.
52 1 (A) , which is the composite curve of NPN and PNP units having
identical characteristics except for polarity. (Practical circuits are never
designed for an exact match, because of the expense of selection.) For
33 K 6.8K
rA/WtAA/V
4 if
K
Fig. 520.
Experimental
omplifier.
twostage
4lf
K
\ PN
K RAT
■\CK7
RATHEYON
CK720
1.2 K
fWVVo— 1( — o
4Mt
\ PN
K RAT
" NCK7
= 6 VOLTS
I
RATHEYON
CK720
600
OHMS
TRANSISTOR AMPLIFIERS
93
+ EC VOLTS
S o NPN
10 13 20 +IcMa
IclPNP)
(B)
Fig. 521. (A) Composite characteristics for NPN and PNP transistors. (B) Waveforms of
composite characteristics.
each operating point E^ in the NPN unit, there is an equivalent
operating point (— Ej) (—It) for the PNP unit. These symmetrical
properties offer innumerable possibilities in circuit applications. For
example, if a peak signal current of 20 ^.a is applied to the base of each
transistor simultaneously, the operating point of the NPN transistor
has shifted to E 2 = 8 volts, I 2 = 15 ma at the instant that the input
signal reaches a value of + 10 y&. But at the same instant the operating
point of the PNP unit is at E 2 = —22 volts, and I 2 = —5 ma. An in
crease in the base current of the NPN unit causes the collector current
to increase; the same variation causes the collector current of the PNP
unit to decrease. When the signal is reversed, the opposite effect occurs.
The complete waveforms for this operation are shown in Fig. 52 1 (B) .
Since the output of the transistors are 180° out of phase, it appears that
the NPN and PNP types will operate, with their input circuits in
<'
<
t bbH
Fig. 522 (left). A symmetrical pushpull
stage.
Fig. 523 (above). A directcoupled sym
metrical cascade.
94
FUNDAMENTALS OF TRANSISTORS
1 NPN 1 T PNP _J_
Fig. 524. Twostage symmetrical
pushpull amplifier.
1
dt&SLrf^_
VOICE
COIL
1 — kHU — K NPN T
parallel, as a pushpull stage. Furthermore, due to the complementary
action of the NPN and PNP types, the circuit does not require an
input transformer or a phase inverter.
Symmetrical PushPull Operation. Figure 522 illustrates the basic
symmetrical pushpull circuit with numerical values based on the same
typical transistor characteristics used in previous examples. The opera
tion of this circuit is the same as that of the transistor pushpull Class
A amplifier that uses only one type of transistor. The circuit is capable
of supplying a high voltage gain when operating into a high impedance
load. The voltage gain of the circuit shown in Fig. 522 is in the order
of 250 (48db) . If the transistors are exactly symmetrical, the dc collector
currents supplied by each transistor cancel each other, and no dc com
ponent flows in the load. The circuit is easily adaptable for direct con
nection to the voice coil of a speaker. Notice also that the same circuit
can be modified by proper adjustment of the base bias for Class B
pushpull operation.
Cascade Operation. One type of symmetrical circuit that proves very
practical is the cascaded arrangement illustrated in Fig. 523. This tan
dem circuit represents the simplest possible cascade, since the only com
ponents of the system are the transistors and the battery supply. The
gain per stage is low compared to the maximum available gain because
of the mismatch existing between the stages. However, the reduced
number of components and the simplicity of the design often outweighs
this disadvantage.
A circuit which incorporates the major features of both pushpull
and cascaded symmetrical configurations is shown in Fig. 524. This
arrangement can serve as a singleended power amplifier to feed a low
impedance speaker from a relatively high resistance source. The two
transistors in the output circuit are operated in the grounded emitter
connection. Therefore, the phase of the input signal is reversed in going
from base to collector. The base of the last stage is connected directly to
the collector output circuit of the input stage. Since the signal also
undergoes phase reversal in the first stage, the output of the transistors
on each side of the load are in phase. The stability of this circuit is very
high because it incorporates 100 percent degenerative feedback. The
large amount of feedback keeps the distortion very low, and also allows
TRANSISTOR AMPLIFIERS 95
the load to be very small. Since the circuit is in effect a twostage Class
B pushpull amplifier, the standby collector dissipation is negligible.
The amplifier is capable of delivering a constant ac output of about
400 milliwatts using transistors rated at 100 milliwatts. In intermittent
short term operation, the same amplifier can deliver about a watt with
out damage to the transistors.
It is apparent that complementarysymmetry circuits offer consider
able promise for further investigation. Their use in the field of high
quality, lowcost portable audio systems is particularly attractive be
cause the output can be fed directly into a voice coil, thus eliminating
the expensive and often troublesome output transformer.
Chapter 6
TRANSISTOR OSCILLATORS
This chapter deals with the operation and circuitry of transistor
oscillators. In general, these fall into two categories: the feedback (or
vacuum tube equivalent) types, and the negativeresistance (or current
multiplying) type. Transistor oscillators are capable of sinewave genera
tion by every mode of operation now feasible in vacuumtube circuits,
plus some additional novel modes. This chapter covers the capabilities
of the transistor as an oscillator in basic rather than specific designs. A
number of numerical examples and specific values are included to il
lustrate the fundamental concepts involved. An analysis of relaxation,
frequency multiplication, frequency division, and triggering in the tran
sistor is also included.
Feedback Oscillators
Transistor Hartley Oscillator. In the earlier chapters, it was shown
that transistor properties, in every important respect, are equivalent to
those of the vacuum tube. It is reasonable then to assume that any vac
uumtube oscillator configuration has an equivalent transistor circuit.
For example, consider the vacuumtube oscillator, illustrated in Fig.
61 (A), which represents one form of Hartley oscillator. Positive feed
back is accomplished by arranging the resonant tank E to be common
to both the input grid and output plate circuits. The equivalent tran
sistor circuit using a grounded emitter connection is illustrated in Fig.
61 (B) . Again, positive feedback is provided by placing the resonant
tank so that it is common to both the input base and output collector
circuits. If ground is removed from the emitter lead, and placed at the
bottom of the tank circuit, the electrical operation of the oscillator is
i I
(A) VACUUM TUBE (B) TRANSISTOR HARTLEY (C) TRANSISTOR HARTLEY
HARTLEY OSCILLATOR OSCILLATOR (GROUNDED OSCILLATOR (GROUNDED
EMITTER CONNECTION) BASE CONNECTION)
Fig. 61. Vacuum lube and transistor Hartley oscillator circuits.
96
TRANSISTOR OSCILLATORS 97
unchanged. Notice that when this circuit is rearranged as illustrated
in Fig. 61 (C) , it is now in the groundedbase connection. While the
grid bias of the vacuumtube oscillator in Fig. 61 (A) is regulated by
the grid leak resistor R G , the equivalent transistor base in Fig. 61 (C)
is selfbiased through resistor R B . In all three circuits, the battery supply
is decoupled by an RF choke.
The major difference between the operation of the vacuumtube
Hartley oscillator and that employing a transistor lies in the loading
effect of the emitter resistance on the tank coil. This resistance is re
flected into the tank circuit and acts as an equivalent shunting resist
ance. The tank is also shunted by the collector resistance, and the equi
valent shunt resistance of the resonant circuit becomes R = ^r~ Oscil
lation starts when the equivalent shunt resistance of the tank is coun
terbalanced by the reflected negativeresistance of the emitter. The op
timum tap point of the coil (as determined both mathematically and
experimentally) is T = £> where T is the ratio of the feedback turns
included in the emitter circuit to the total number of tank coil turns,
and a is the emittertocollector current gain. Notice that when a ap
proaches unity, the transistor oscillates at highest efficiency with a cen
tertapped tank coil. Under this condition the minimum allowable par
4r
allel resistance of the tank circuit is R = — ~ , which sets the Q of the
a
R 4r
circuit at Q = —  — = — > e j , where <o = 2irf , r e is the transistor
resistance, f is the resonant frequency, and L is the inductance of the tank
coil. The operating resonant frequency is klways lower than the isolated
resonant tank frequency, because of the change in effective value of in
ductance caused by the coil tap.
The disadvantages of tapping the coil can be avoided by using a
direct feedback path from the resonant circuit to the input terminal. Fig
ures 62 (A) and 62 (B) illustrate two such possible arrangements. In both
examples, the feedback resistor R F (a choke may be used) and the effec
tive impedance of the resonant circuit form an ac voltage divider. The
value of R F can be adjusted to obtain the required amount of feedback
for sustained oscillation.
Transistor Clapp Oscillator. The transistor equivalent of the Clapp
oscillator is illustrated in Fig. 63 (A) . The operating frequency is set by
the series resonant circuit in the collector circuit. Feedback is taken from
the voltage divider consisting of capacitors C^' and C 2 . The upper fre
quency limit depends largely on the transistor in use, and can be increased
98
FUNDAMENTALS OF TRANSISTORS
Fig. 62. Direct feedback connec
tions: (A) collector ta base, (B) col
lector to emitter.
L 3~C
(A)
considerably by careful selection of the unit. Upper frequencies as high
as 3 mc can be attained using typical junction types in this basic circuit.
The numerical values shown are based on the average of a group of Ray
theon CK720 transistors. If crystal control is used, the frequency stability
is improved, and there is also a considerable increase in the upper fre
quency limit of the oscillator. One method of achieving crystal control
in this circuit is to replace the collector resonant circuit by a crystal.
Transistor Colpitts Oscillator. The transistor Colpitts oscillator is
similar to the Clapp type except that the resonant load is a parallel ar
rangement in the collector circuit. Thus the circuit becomes voltage,
rather than current, controlled. The feedback is again taken from a
point between the two series capacitors connecting the collector to
ground. The upper frequency limit for this oscillator is in the same
range as that of the currentcontrolled Clapp arrangement. The typical
values illustrated in Figure 63 (B) are again based on the average of
a small group of Raytheon CK720 transistors.
In both cases, the parallel combination C b Rb provides the necessary
emitter bias. This arrangement provides some degree of amplitude
stability similar to the control provided by bypassed cathode or grid
leak resistors in vacuumtube oscillator circuits.
In servicing transistor oscillators, the emitter bias measured at the
base end of the C B R B combination is a useful indication of the signal
200 11 f
, R E J_C,
— ^T~'
'ZOOlif
T*
Fig. 63. (A) Transistor Clapp oscillator. (B) Transistor Calpitts oscillator.
TRANSISTOR OSCILLATORS
99
Fig. 64 (above). Transistor multivibrator.
Fig. 65 (right). Basic resistance controlled
negative resistance circuit.
amplitude. In addition, the variation of the emitter bias over the fre
quency range indicates the relative uniformity of the signal output.
Special care is necessary during these measurements to avoid affecting
circuit operation. A vacuumtube voltmeter may be used without caus
ing additional loading. The use of a high resistance meter also mini
mizes that oscillator loading due to the stray reactance of the measuring
probe. While direct current measurement is better, it requires disturb
ing the circuit wiring.
Transistor Multivibrator. Figure 64 illustrates the transistor equi
valent of a basic multivibrator circuit. This configuration is generally
useful in the frequency range of 5 to 15 kc. The parameter values shown
are for an 8.33 kc oscillator which uses two Raytheon CK720 transistors.
The frequency is determined by the RbC b time constant. The value of
R B is limited to a maximum of about 200k ohms. C B is limited to a
minimum of .002 /*f. The frequency stability is poor compared to the
types previously discussed. The output collector waveform is almost a
perfect square wave. The advantages of the transistor multivibrator are
its simplicity and the small number of components required.
NegativeResistance Oscillators
Conditions for Oscillation. The preceding paragraphs indicate that
transistor oscillators can be designed as equivalents for all the known
types of vacuumtube oscillators that use an external feedback path. In
addition, the unique property of a transistor that furnishes current gain
can also be used to design many other novel types of oscillators. In the
earlier chapters it was found that the pointcontact transistor, by virtue
of its ability to multiply the input current (r m greater than r c ) , is char
acterized by negative input and output resistances over part of its op
erating range. It is feasible, therefore, to use the pointcontact transistor
in this region to design oscillator circuits that do not require external
feedback paths. As one engineer put it, "An oscillator is a poorly de
signed amplifier." This observation is particularly applicable in the
100 FUNDAMENTALS OF TRANSISTORS
case of the negativeresistance oscillator. The conditional stability equa
tion for a point contact transistor was specified in Chapter 4 as:
(r u \ R g ) (R L \ r 22 ) — r 12 r 21 must be greater than zero. Thus for the
transistor to be unstable, that is for it to exhibit negative resistance
characteristics, requires:
(in + Rg) (Rl + r 22 )  r 12 r 21 < Eq. (61)
In general, external resistance can be added to any of the three electrode
leads, as illustrated in Fig. 65. Substituting the transistor parameter
values into equation 61 results in:
(r e + r„ + R B + R g ) (Rl + r b + R B + r c )  (r„ + R B )(r b + R B + r m ), < 0'
Neglecting r e and r b as compared to R B , r c , and r ra , this becomes:
(R B + R g ) (R L + R B + r c )  R B (R B + r m )< and multiplying out
R B R L + R B 2 + R B r c + R g R L + R g R B + R g r c  R B 2  R B r m <
which becomes:
R g (R L + r c ) + R B (R L + R g )  R B (r m  r c )<
Notice that when r m is less than r c (as in the case of the junction tran
sistor) , the condition for oscillation cannot be satisfied. This reempha
sizes the fact that negativeresisttnoe oscillators can only be designed
using the pointcontact transistor. Notice also in this equation that if
both R L and r g are small compared to the value of (r m — r c ) , the con
ditional equation is primarily controlled by the value of R B . The
higher the value of R B , the more definite the instability. Furthermore,
as the external collector and emitter resistances are increased in value,
a higher resistance of R B is required to assure circuit oscillation. The
control of oscillation in negativeresistance transistor oscillators, then,
is determined by the following three factors, either separately or in
combination: the external resistance of the emitter lead (a low value
favors oscillation) , the external resistance of the base lead (a high value
favors oscillation) , and the external resistance of the collector lead (a
low value favors oscillation) .
Basic Operation. If the control of an oscillator can be maintained
by simple high or low resistance values in the three transistor electrode
arms, the substitution of series and parallel LC resonant circuits in
their place is a natural step. The insertion of a parallel resonant circuit
in the base lead will cause the circuit to oscillate at the resonant fre
quency because of the tank's high impedance at resonance. On the
other hand, placing a series LC circuit in the emitter or collector arms
will cause oscillation at the resonance frequency due to the tank's
characteristic low impedance at that point. Fig. 66 illustrates the ac
equivalent circuit of a negativeresistance oscillator that includes all
three methods of controlling oscillation. Since LC resonant circuits pro
duce sine waveforms, the oscillators using LC resonant tanks are gen
erally referred to as sinewave oscillators.
TRANSISTOR OSCILLATORS
101
POINT
CONTACT
x
Fig. 66. Basic impedance controlled
negative resistance oscillator.
The use of only the pointcontact transistor for the negativeresist
ance oscillator is readily explained on an electronic basis. Assume that
for the conventional grounded base connection, a disturbance or electri
cal charge of some sort causes an ac emitter current to flow. This re
sults in an amplified collector current i c = a i e in the collector circuit.
Since there is no phase inversion, the current flows through the base
in phase with the emitter current. If the base resistance is large, the
regenerative signal will be larger than the original signal. This in
creased current is again amplified, causing a greater collector current
to flow, which again is fed back to the emitter, and so forth. In a short
time, the current passes out of the linear dynamic operating range, and
the circuit breaks into oscillation. The frequency of this oscillation is
determined by the time constant of the circuit. In brief then, the point
contact transistor is capable of basic oscillation, without external feed
back path, because of its ability to provide current gain and internal
feedback path without phase reversal through the base lead.
General Types. Negativeresistance oscillators may be divided into
two general classes: voltage controlled; and current controlled. The volt
agecontrolled oscillator is characterized by a high resistance load, and
a low resistance power supply (constant voltage) . The fundamental
schematic of a typical oscillator of this type is illustrated in Fig. 67 (A) .
jF*
llll
(A)
Fig. 67. (A) Voltage controlled negative resistance equivalent circuit. (B) Idealized
currentvoltage characteristic.
102
FUNDAMENTALS OF TRANSISTORS
/<
"bb
,,E
I
I OPERATING
s^ POINT
t J
uu /
kJ?2 /
3 = 2 /
rui /
AC /
* /
'resonant
CIRCUIT
LOAD LINE
(B)
U — CONSTANT
CURRENT
BIAS
(A)
Fig. 68. (A) Currentcontrolled negative resistance equivalent circuit. (B) Idealized
currentvoltage characteristic.
Ec
Ic
y*
1
y\ i /
s \[ /
OPERATING  \ /
POINT \y
LOAD
— BIAS
(C)
Fig. 69. (A) Basecontrolled negative resistance oscillator and
idealized characteristic. (B) Emittercontrolled negative resistance
oscillator and idealized characteristic. (C) Collectorcontrolled
negative resistance oscillator and idealized characteristic.
TRANSISTOR OSCILLATORS 103
This oscillator is composed of three major parts: the resonant LC cir
cuit, the negative resistance of the oscillator, and the dc supply volt
age E bb .
Figure 67 (B) represents the idealized current voltage character
istics of this oscillator. It is typical of the negativeresistance oscillator
that the resistance remains negative only over a limited portion of its
operating range. The bias is established somewhere in the middle of
this useful section to guarantee oscillation. It is evident that a constant
voltage bias is required. A remaining condition for sustained oscillation
is that the resonant load have a higher absolute value than the negative
resistance presented by the oscillator at the operating point. The par
allel LC circuit that approaches an infinite impedance at resonance,
then, is ideal for this purpose.
The currentcontrolled type is shown in Fig. 68 (A) . This oscillator
is characterized by a low ac load and a high dc power source (constant
current) .' Figure 68 (B) represents the idealized currentvoltage char
acteristics for this negativeresistance oscillator. As in the voltagecon
trolled type, the negativeresistance region is limited to a section of the
operating range, and the bias is established somewhere in the middle
of this negativeresistance region using a constant current source. The
last condition to be satisfied for sustained oscillation is that the ac
load of the resonant circuit must be less than the absolute value of
negative resistance of the oscillator at the operating point. The series
LC circuit, the resonant impedance of which is close to zero, is the
ideal load for this application.
Sinewave Oscillators. These principles can now be applied to the
three basic methods of controlling oscillation in the pointcontact tran
sistor: the insertion of low impedance loads in the emitter or collector
circuits (current control) , or the insertion of a high impedance load
in the base lead (voltage control) . Figure 69 (A) illustrates the basic
basecontrolled oscillator and its idealized current voltage characteris
tics. This circuit is the most often used because it offers the best pos
sibilities of the three types. Its main advantages are that it employs a
constant voltage source (the easiest type to design) , and that the regen
erative feedback is through the resonant tank in the base lead. This
latter feature assures frequency stability, because maximum feedback
occurs at the resonant frequency of the tank circuit. The effect of the
internal base resistance is negligible due to the extremely high value of
the parallel circuit at resonance in comparison to r b .
Figure 69 (B) represents the basic emitter controlled negative re
sistance oscillator and its idealized currentvoltage characteristics. Fig.
69 (C) is the basic collector controlled type. The fundamental opera
tion of both is essentially the same. Oscillation occurs at the series
resonant frequency of the LC combination because at this point the
104
FUNDAMENTALS OF TRANSISTORS
©VARIABLE/^
CURRENT ( V )
SOURCE V^
(A) (B)
Fig. 610. (A) Basic measuring circuit for obtaining negative resistance
characteristics. (B) Typical negative resistance characteristic
effective resistance in either the emitter or collector arm is at minimum.
The base resistance must be large enough to furnish positive feed
back in order to sustain oscillation. The base resistance r b is generally
large enough to cause instability when either the emitter or collector
is shorted to ground, on the basis of equation 61. In practical circuits,
however, r b alone is rarely enough for dependable operation. An ex
ternal resistor R B equal to at least 2,000 ohms is generally added.
Negative Characteristic Measurements. The characteristics of the
three basic negativeresistance connections are not generally supplied
by the manufacturer. These, however, may be obtained by a point plot.
This is not too arduous a task since the curves are reasonably linear
and the changeover points are well defined. For most purposes it is
sufficiently accurate to insert a sweep signal into the controlled electrode
and observe the response on an oscilloscope. Figure 610 (A) illustrates
the basic measuring circuit for this application when the transistor is
in the emitter controlled connection. A typical resulting E e — I e char
acteristic is shown in Fig. 610 (B).
The measuring circuit is easily modified for application to the
base or collector controlled type. The plotted curve is similar to those
illustrated in Figs. 69 (A) and 69 (C) .
Bias Selection. It can be shown mathematically that the condition
for locating the operating point in the center of the negativeresistance
region is: E e (2 a Rc + R E ) = E C R E . This relationship indicates that the
extent of the negativeresistance range depends upon the bias batteries
and the values of R E and R c . The emittertocollector current gain a is,
of course, fixed for a given transistor. For the characteristic in Fig.
610 (B) , then, all the parameters are specified with the exception of
E e and R E . Notice, however, that these quantities are related to the
value of dc emitter current bias I e that is required to establish a dc
operating point in the center of the negative resistance region. This
condition is: E e = R E I e .
TRANSISTOR OSCILLATORS
105
The two conditional equations can be combined to evaluate R E in
terms of known quantities:
R H =4 £ — 2 a R c Eq. (62)
Since R E also equals
, equation 62 limits the value of the emitter
bias battery to less than that of E c . The limiting value of E e = E c is
reached when R c = 0.
As a numerical example, if the values associated with Fig. 610 (B)
are used so that the bias current at the center of the negative resistance
region is I e — 0.75 ma, then
R E = A /2aR,V= 75x ^ _3 [2 (95) 15 x 10 3 ] = 31,500 ohms,
and E e = I e R E = .75 x 10~ 3 x 31.5 x 10 3 = 23.6 volts.
The negative resistance of the oscillator is equal to the slope of the
characteristics in that region. Then
r =
e (max)
■e (min)
I.
I.
_ (25) (5)
(1.6 0.1) x 10 3
 13,300 ohms
A e (max) "e (mln)
This value defines the maximum limit of the impedance of the LC
series emitter circuit at resonance.
Oscillator Stabilization. The generated signal of the sinewave os
cillator becomes badly distorted when the dynamic operating range of
the circuit exceeds the negativeresistance region; excessive and uncon
trolled distortion causes frequency instability. Obviously, the reduction
of the harmonic content to a minimum is particularly important in those
applications that require a stable and pure sine wave. But even in those
cases where a high harmonic content is desirable, steps are necessary
(A)
j OUTPUT
VOLTAGE
SIGNAL
Fig. 611. Effect of ac load on harmonic content: (A) idealized ac resonant load;
(B) Ac load slightly less than negative resistance of characteristic.
106 FUNDAMENTALS OF TRANSISTORS
to keep the harmonic content of the signal constant to insure frequency
stability of the oscillator.
The value of the ac load impedance has a large effect on the
amount of harmonic distortion in the signal. This effect is illustrated
in Fig. 611 for the emitter controlled type. Figure 611 (A) illustrates
the distortion in the voltage waveform when a sine wave of current is
generated in an ideal series LC load having zero impedance at reson
ance. Figure 611 (B) illustrates how the distortion is reduced to a sat
isfactory level by increasing the resonant impedance of the load. The
increased ac load effectively limits the dynamic range of the oscillator
to the negativeresistance region. Thus, when a currentcontrolled os
cillator is required to operate with a low harmonic content, the ac
load impedance should be chosen to be slightly less than the absolute
value of the resistance determined by the slope of the negativeresistance
characteristic. This same condition applies when the oscillator is col
lector controlled. A similar situation exists in the basecontrolled nega
tiveresistance oscillator except that, since this is a voltagecontrolled os
cillator, the distortion occurs in the current waveform. In this latter
circuit, low distortion operation is attained by reducing the value of
the resonant impedance so that it is slightly greater than the negative
resistance slope of the characteristic curve.
The operating point must be stabilized in the center of the nega
tiveresistance region in order to avoid distortion from unequal positive
and negative signal amplitudes. When the external resistances are fixed,
the main causes of operating point shifts are changes in the bias sup
plies. An effective method of stabilization is the use of one supply
battery for both the emitter and collector bias. This assures that the
ratio —A will remain constant in spite of variation in the battery
potential.
Increasing the resonant impedance of a series resonant arm is ac
complished by selecting a higher resistance inductor, or by increasing
the value of the series resistor in the emitter or collector lead. Decreas
ing the resonant impedance of the base controlled tank circuit is not
as simple. A reduction of the tank Q will, of course, decrease the reson
ant impedance, but a low Q tank tends to promote frequency instability.
A more satisfactory method of decreasing the impedance is to tap the
base lead at some point in the tank coil. This permits the retention of
a high Q tank, and, at the same time, reduces the effective impedance
connected in the base lead. In addition, this connection helps to re
duce the effects of internal transistor reactances on the operating fre
quency.
These internal reactances, primarily caused by junction capaci
tances, are particularly troublesome because their values do not remain
TRANSISTOR OSCILLATORS
107
POINT
CONTACT
TRANSISTOR
TO
TO
TO
EMITTER
BASE
COLLECTOR
CIRCUIT
CIRCUIT
CIRCUIT
— AAAr
AAA — '■
■m
(B)
Fig. 612. (A) Stabilized basecontrolled highfrequency oscillator. (B) Alternate
method of providing common bias supply.
constant with changes in temperature and changes in operating currents
and voltages. However, loose coupling between the tank and the base
circuit minimizes the effect of internal transistor reactance. While this
reduces the available power of the oscillator, the sacrifice of power for
stable operation is generally justified. The level of the signal can always
be increased by a stage or two of amplification.
Amplitude stability in negativeresistance oscillators is generally
accomplished by incorporating some form of automatic bias control in
the circuit. Sometimes the required amount of degenerative feedback
is obtained through a nonlinear resistor, placed in either the collector
or emitter circuit. In this case, the main problem involves finding a
nonlinear element that is sensitive to the small current changes in
volved. Amplitude stability may also be obtained by a loosely coupled
tank in the basecontrolled oscillator, since it automatically decreases
positive feedback at frequencies off resonance.
Stabilizing Circuitry. Figure 612 (A) illustrates one arrangement
of a highfrequency basecontrolled oscillator that incorporates the vari
ous stabilizing features discussed in the preceding paragraphs. C t and C 2
are phase compensating condensers. The base lead is connected to a
tapped tank coil as a means of reducing the resonant impedance while
maintaining a high Q tank. Bias stability is accomplished by using one
common battery source. Notice also that positive emitter bias is supplied
by the bypassed resistor R B . Figure 612 (B) illustrates an alternate meth
od of providing a constant collectortoemitter bias ratio by means of
a common battery supply. The advantage of this circuit is its design
simplicity, since it is basically a voltage divider network. The values of
C x and C 2 are not critical; they complete the ac circuit between the
collector, base, and emitter leads, and bypass the battery and bias di
vider network.
Except for the inductance and capacitance elements of the resonant
network, the values of the external components in negativeresistance
108
FUNDAMENTALS OF TRANSISTORS
oscillators are not critical. The values of R E and R c should be large
enough to limit their respective currents to safe values, but not so large
that they cause excessive degeneration. The value of the base resistance
R B must be large enough to provide sufficient regeneration for sustained
oscillation. Typical values for these parameters are: R E = 50 to 2,000
ohms; R c = 2,000 to 10,000 ohms; R B = 10,000 to 20,000 ohms.
Transistor Phase Shift
Contributing Factors. In general, transistor oscillators make use of
their nonlinear characteristics. While there has been considerable pro
gress made in the mathematical analysis of nonlinear circuits, particu
larly in the past few years, oscillator design is invariably based on the
static characteristic curves. This is true since even the simplest math
ematical approximations of nonlinear operation are too involved for
the average experimenter or engineer to handle.
When the operating frequency becomes more than 100 kc, the in
ternal transistor parameters can no longer be considered as simple re
sistances. At this frequency, the values of the transistor reactive com
ponents become appreciable. In addition to the fixedresonant circuit
parameters, there are also stray reactances due to lead inductance, and
others that have a considerable effect on the transistor characteristics.
Static curves, then, are extremely useful to set bias points, and to ap
proximate the negativeresistance range, optimum load, and wave
shapes. However, circuit values based on the low frequency transistor
characteristics are not exact. The experimenter finds that every high
frequency transistor oscillator requires some readjustment for optimum
operation.
Phase Shift and Feedback. One effect of the reactive components is
to cause a phase shift between the input and output terminals. Phase
shift reduces the inphase component of the positive feedback signal.
This is illustrated in Fig. 613 (A) where E F is the feedback signal and <j>
E F  FEEDBACK SIGNAL
 PHASE ANSLE ^ C l
Fig. 613. (A) Effect of phase shift on feedback signal. (B) Basecontrolled phase
compensated oscillator.
TRANSISTOR OSCILLATORS 109
BAND
ELIMINATION
FltTER
NETWORK
Fig. 614. Phaseshift oscillator.
► R 8
i IiMh
is the phase angle between the input and output signals. E F1 represents
the feedback amplitude at low frequencies when the reactive effects are
negligible. As the operating frequency is increased, E F2 and the input
signal are no longer in phase. Thus, only the inphase component of E F
is useful for maintaining circuit oscillation. When the phase angle be
comes so large that the inphase component is less than the critical
minimum required value, oscillation stops.
Phase Shift Compensation. The reduction of value of the inphase
feedback signal requires either an increase in feedback E F or a form of
phase compensation to decrease the angle <f>. In the basecontrolled os
cillator, some phase shift compensation is provided by shunting either
or both the emitter and collector electrodes to ground through a small
capacitor (3 or 4 p.(J) . This simple modification usually doubles the
upper frequency limit of a transistor.
One method of increasing the available feedback is to connect a
resistor from the emitter to a tap point on the base tank coil. This pro
vides regenerative voltage feedback to supplement the inherent current
feedback of the circuit. The value of the resistor R r is critical. The
upper limit of the oscillator frequency drops as R F is either increased
or decreased from its critical value. For this reason, the feedback re
sistance is best determined on an experimental basis. Figure 613 (B)
illustrates a basic oscillator incorporating these two methods of phase
shift control and compensation.
Phase Shift Oscillator. One very stable negativeresistance oscillator
is the phaseshift type illustrated in Fig. 614. This circuit is particularly
useful in the audio range when a low distortion sinewave signal is re
quired. The resistances R c , R B , and R E are determined by the condi
tion for instability specified by equation 61. The phase shift network
used is a bandelimination filter at the desired operating frequency. At
this frequency, the filter offers maximum attenuation (theoretically an
open circuit) . At any other frequency, the network attenuation de
creases, thereby providing a degenerative feedback path into the base
lead. This degeneration counteracts the positive feedback through the
base resistor R B . Thus, oscillation is favored only at the operating fre
quency, namely, the frequency eliminated by the phase shift network.
If the network is designed for both phase reversal and minimum attenu
no
FUNDAMENTALS OF TRANSISTORS
Fig. 615. Crystal oscillators:
(A) base controlled; (B) emitter
controlled; (C) collector con
trolled.
ation at the operating frequency, it will also be a useful oscillator.
Under these conditions the network provides positive feedback into the
base, which supplements the normal regenerative signal through the
base resistor. The bandelimination filter oscillator is limited to the
lower frequencies since proper operation depends on a zero phase shift
through the network at the operating frequency.
NegativeResistance Crystal Oscillators
Basic Types. The negativeresistance oscillator is easily adapted to
crystal control, since crystals can operate as either series or parallel
tuned circuits. Figure 615 (A) illustrates the basic circuit of the base
controlled crystal oscillator. The RF choke which bypasses the crystal
provides a dc path to the base. A choke coil is used rather than a re
sistor for two important reasons: first, a resistor lowers the Q of the
crystal; second, a resistor provides a positive feedback path for frequen
cies off resonance, thereby eliminating the major advantage of the base
controlled circuit, namely, maximum regeneration at resonance, mini
mum regeneration off resonance. Since, in this case, the crystal is op
erated as a parallel resonant circuit, this oscillator is electrically equi
valent to the basecontrolled circuit illustrated in Fig. 612 (A).
Figures 615 (B) and 615 (C) represent the basic circuits of the
emitter and collectorcontrolled crystal oscillators. The circuit shown
in Fig. 615 (B) will operate satisfactorily if the base tank is replaced
by a resistor. The inclusion of the tuned circuit, however, provides in
creased frequency stability and decreased harmonic distortion in the
output signal. The series resonant circuit in the emitter arm of the
collectorcontrolled oscillator illustrated in Fig. 615 (C) is added as
a means of increasing the frequency stability. It can be replaced by a
resistor.
TRANSISTOR OSCILLATORS 111
Frequency Multiplication. Since the power handling capacity of
the transistor is small, it can seldom provide enough energy to excite
a crystal into oscillation at the higher frequencies. For this reason, high
frequency crystalcontrolled oscillators usually incorporate some form of
frequency multiplication. Figure 616 illustrates one basic circuit for a
crystalcontrolled frequencymultiplier oscillator. The emitter and base
circuits in this basecontrolled oscillator are conventional. The collector
lead, however, contains a parallel resonant circuit tuned to the desired
harmonic of the crystal fundamental frequency. At first glance it may
appear that the inclusion of this network in the collector arm violates
one of the fundamental requirements of negativeresistance oscillators,
that is, the need for a low resistance collector circuit (equation 61) .
However, the collector tank is tuned to a harmonic of at least twice
the fundamental frequency. Insofar as the fundamental crystal frequency
is concerned, then, the collector tank is a low impedance. The tank
offers a high impedance to the required harmonic, and consequently
establishes a good feed point for this frequency into the output circuit.
Proper operation of the frequencymultiplier oscillator requires
that the fundamental frequency be rich in harmonics, since low distor
tion contains little harmonic energy. The inherent nonlinearity of
negativeresistance oscillators [Figs. 69 (A) , (B) , and (C) ], makes it
easy to generate a distorted waveshape. This necessitates the use of a
high impedance resonant circuit in the basecontrolled oscillator, and
the use of a low impedance circuit in the emitter or collectorcontrolled
types. Tight coupling of the base tank also promotes increased harmonic
generation, but this feature is generally unsatisfactory because of its
adverse effect on frequency stability.
Relaxation Oscillators
Basic Characteristics and Operation. One of the most inviting ap
plications of the negativeresistance oscillator is as a relaxation type, par
ticularly since its power requirements are low. Transistor relaxation os
cillators have almost limitless use where a complex waveform, pulse
generation, triggered output or frequency division is required. Like the
equivalent vacuumtube types, the periodic operation of the transistor
relaxation oscillator usually depends on a RC or RL combination for
output Fig. 616. Crystalcontrolled frequency
^ multiplier.
TANK IS TUNED
TO A HARMONIC
OF FUNDAMENTAL
CRYSTAL
FREQUENCY
112
FUNDAMENTALS OF TRANSISTORS
► T*E (SECS1
Fig. 617. (A) Basic emittercontrolled relaxation oscillator
with (B) idealized characteristic, and (C) waveforms.
the storage and release of signal energy. For this reason, they are char
acterized by abrupt changes from one operating point to another. This
makes relaxation oscillators particularly useful for generating sawtooth
waveforms.
Figure 617 represents the basic emittercontrolled relaxation os
cillator and its idealized currentvoltage characteristic. The location of
the frequencydetermining network in the emitter circuit provides the
largest measure of control. This basic type; therefore, is the most usefuL
The fundamental operation is involved, but not difficult to understand.
For simplicity, assume the operation starts at point A (Figure 617B) .
At this point the transistor is cut off, since the emitter is biased in the
reverse direction (— E A ) . Because of this reverse bias, the input circuit
offers a high resistance path. The charge on capacitor C E (equals — E A )
has to leak off through R E , and the rate of discharge is determined by
the time constant R B C E .
When the voltage across the capacitor is reduced to — E E , operation
is at point B, which represents the point of transition from the cutoff
to the negativeresistance region. The values of the emitter and collector
resistances drop quickly to near zero, and the battery current is then
limited only by the value of R c . If the small effect of the saturation cur
rent I co is neglected, both the emitter and collector current increase
E
from zero to ^ almost instantaneously. In this instant, the operat
ic
ing point moves rapidly from point B through point C to point D. At
TRANSISTOR OSCILLATORS 113
the same time, the voltage across the capacitor starts to increase to its
original value of — E A . The rate is fixed by the time constant of C E and
the parallel equivalent of R B and R c . In the meantime, the emitter cur
rent decreases at the same rate, thereby moving the operating point back
toward point C.
When the current reaches point C, operation passes from the satu
ration region to the negativeresistance region. Instability in this area
causes the current to drop instantaneously to its value at point B. Be
cause of this rapid drop, the condenser voltage does not change. The
operating point returns to point A, and the condenser discharge action
starts the cycle again.
Note that there are two time constants during a complete cycle.
The first one T r — R E C E controls the discharge rate of the condenser
when operation moves from point A to point D. The second time con
R R
stant T 2 = •= — B ° (C E ) controls the charging rate when operation
R B \ R c
moves from point D to point A. The sawtooth voltage generated by this
circuit is illustrated in Fig. 617 (C). The frequency of operation is
approximately
1 1
F =
Ti + T 2
Ce ( Re +r*+r c )
The frequency of the current wave is the same, but the waveform ap
proximates a pulse, since the current only flows during the period when
the condenser is charging (T 2 ) . This simple oscillator, then, is useful
as a voltage sawtooth or a current pulse generator.
The following problem will be used as a numerical example of
basic relaxation oscillator design. Assume that a sawtooth voltage wave
is required for use in a sweep circuit, and that the following character
istics are specified: frequency is 5 kc; the charging rate interval T 2 is
limited to 10% of the total cycle; R B is 2,000 ohms, required for sus
tained oscillation; E c is fixed at 12 volts. The numerical values of the
major operating points shown on Fig. 617 (B) are: for point A, I E —
0.1 ma, E E = 10 volts; for point B, I E = 0.01 ma, E E = 2 volts; for
point C, I E = 3 ma, E E = 10 volts; and for point D, I E = 5 ma, E E = 2
volts. From the preceding analysis, I E = ^ .
R c
E 12
Thus at point D, R c =j£ = _ — ^5= 2,400 ohms
The overall time constant T = T t f T 2 = L= , * —200 u sec
r 5 x 10 3 r
Onds, T 2 = 10% (T ) = .10 (200) = 20 ^ seconds, and T, = T  T 2 =
200  20 = 180 M seconds.
114 FUNDAMENTALS OF TRANSISTORS
Since T 2 = R ^^(C E ),
_ (R B + R c ) T 2 _ (2,000 + 2,400) 20 x 1Q~« _
Ce R^ (2,000) (2,400) • ° >*■
Since T, = R E C E , R E =^= ^™*^_ e = 10,000 ohms
Base and CollectorControlled Oscillators. Basecontrolled and col
lectorcontrolled relaxation oscillators are illustrated in Figs. 618 (A)
and 618 (B). Both operate very much like the emittercontrolled type,
and are analyzed on the basis of their respective operating characteristics,
illustrated in Fig. 69 (A) and (C) . The main difference is that the
basecontrolled type uses an inductance for the storage and release of
circuit energy.
The fundamental difference between the sine wave oscillator and
the relaxation oscillator is determined by which of the circuit para
meters control the repetition rate. This, in turn, is determined by
which has the lowest period of oscillation. For example, if in Fig.
612 (A) the time constant of the emitter network CjR E or the collector
network C 2 R C is greater than that of the base LC tank, the circuit be
comes a relaxation oscillator. If a properly designed basecontrolled high
frequency sinusoidal oscillator suddenly switches to a different fre
quency and produces a distorted waveform, the trouble is most likely
in the base resonant circuit.
While the RC time constant of the collector and emittercontrolled
relaxation oscillator is fixed by the required operating frequency, the
C to R ratio should be as high as possible. This causes minimum de
generation in the circuit, and, at the same time, increases the surge
current handling capacity of the condenser. As before, the value of the
base resistor R B is determined by the amount of positive feedback re
quired for sustained operation.
SelfQuenching Oscillator. The relaxation oscillator in combination
with the regular basecontrolled type can be used to form the self
quenching oscillator. Figure 612 (A) illustrates a selfquenching type
if the value of either Cj or C 2 is increased sufficiently to make the
Fig. 618. (A) Batecontrol
led relaxation oscillator. (B)
Collectorcontrolled relaxa
tion oscillator.
TRANSISTOR OSCILLATORS 1 1 5
Hf
>«c
, T~ E Sl ic 1 F'9 619. Basic selfquenching oscillator.
E.
■#"
T, GREATER THAN Tj WHERE T, ■ R E C E
AND Tj ' 2 IT 4LC
emitter or collector time constant appreciably greater than that of the
LC tank circuit. Figure 619 represents the basic selfquenching oscilla
tor. Due to its time constant, the RC emitter network has primary con
trol of the circuit and produces the sawtooth voltage and pulsed current
waveforms illustrated in Fig. 617 (C). The operation of the relaxation
section of the circuit is independent of the base tank. The base network,
however, depends entirely on the relaxation operation. Assume the cycle
is moving in the charging direction (B of Fig. 617) , operation from
point C to point A. When the operation reaches the negativeresistance
region where sufficient regenerative energy is supplied, the base tank
oscillates at its resonant frequency. The amplitude of the resulting
wave is small initially, but rises to a peak at the point when C E starts
its discharge cycle (B of Fig. 617), operation from point A to point
D. The duration of the oscillation in the base tank is a function of the
Q of the network, the amount of stored energy and the loading effect
on the tank by the rest of the circuit. The relaxation or quench fre
quency in this case is f Q = — r= — — , while the resonant fre
quency of the tank is f T = — — ,.  . Notice that f Q must be less than
f T for proper operation. The basic circuit becomes collector controlled
if capacitor C E is moved into the collector circuit. The circuit operation
is exactly the same.
Synchronized Relaxation Oscillator. The operation of a synchro
nized relaxation oscillator is easily understood in view of the funda
mentals of operation covered in the preceding paragraphs. The basic
circuit is the same, but the relaxation frequency is made slightly less
than the synchronizing frequency. Referring to Fig. 617 (B), assume
that operation is moving from point A toward point B, and that a posi
tive pulse, large enough to instantly move operation to point B, is ap
plied to the emitter. The effect, as illustrated in Fig. 620 (A) , is the
same as decreasing the time constant R E C B , and the relaxation fre
quency becomes the same as that of the applied synchronizing pulse.
The actual point at which the synchronizing signal arrives is not criti
cal as long as the pulse amplitude is large enough to carry the opera
116
FUNDAMENTALS OF TRANSISTORS
Fig. 620. Synchronized controlled (A) waveform, (B) frequency multiplication wave
forms, (C) frequency divider waveformi.
tion into the negativeresistance region. Notice, however, that the magni
tude of the sawtooth voltage is reduced by an amount equal to that of
the pulse. The dotted line represents the voltage waveform without
synchronization.
The synchronized oscillator can be used as a frequency multiplier.
Figure 620 (B) illustrates one application in which the input frequency
is approximately half that of the relaxation frequency. Any submultiple
of the normal rate will work. The chief disadvantage of this type of
operation is the lack of control over the frequency in the interval dur
ing synchronizing pulses.
Figure 620 (C) illustrates the application of the synchronized re
laxation oscillator as a frequency divider. In this example, the input
frequency is three times that of the relaxation rate. As long as the
synchronizing rate is an integral multiple of the basic frequency, the
oscillator remains under control. Theoretically, any division ratio is
possible, but in practical circuits the ratio is limited by the nonline
arity of the sawtooth wave near the critical voltage E B . Consistent opera
tion for division ratios up to approximately 10 to 1 can be easily at
tained. Ratios higher than these require critical design for reliable op
eration.
Negative synchronizing pulses can be used to operate the base or
collectorcontrolled oscillator types. The many ramifications of the basic
relaxation oscillator are too numerous to cover, but the experimenter
may find many useful applications for this circuit. If, for example, time
constants are inserted in both the emitter and collector circuits, the re
laxation oscillator can be synchronized by a pulse applied to either
TRANSISTOR OSCILLATORS
117
electrode. The circuit may also be biased in either the saturation or
cutoff region, so that it remains nonoscillatory until pushed into the
regenerative region by an external pulse. The last type falls under the
general category of trigger circuits.
Trigger Circuits
The transistor oscillators considered to this point have one feature
in common: the controlling electrode is biased in the negative resistance
region. These types, whether synchronized, sinusoidal, or nonsinusoidal,
come under the general classification of astable operated.
Triggered circuits, on the other hand, are biased in one of the
stable regions and are nonoscillatory until the trigger pulse is applied.
These types are classified as either monostable operated or bistable op
erated oscillators.
Monostable Operation. The basic monostable circuit is illustrated
in Fig. 621 (A) . The only difference between this circuit and the emit
tercontrolled relaxation oscillator illustrated in Fig. 617 (A) is the
elimination of the emitter resistor R E . Since this action removes the dc
emitter current bias (I E = 0) , the operating point shifts from the neg
ativeresistance region (P 2 ) to the point intersection of the voltage axes
at I E = 0(P!) . This change is illustrated in Fig. 621 (B) . The circuit
is no longer capable of selfsustained oscillation since it is biased in the
stable cutoff region. Now, if a pulse of sufficient magnitude (at least
equal to I P ) is applied to the emitter, operation is forced into the re
generative region. The current jumps to its value at point D, and the
negative charge on the emitter condenser starts to build up. When the
charging current is reduced to its value at point C, operation again en
ters the regenerative region, and the current is quickly reduced to its
value at A. The charge on the condenser gradually leaks off through
the emitter base circuit (r e \ r b f R B ) until the stable operating point
P x is reached. The circuit is now ready for another trigger pulse.
i
i
(A)
Ee
,
1
1 I.
Eb
1 0.
P M
1 /
Ea
MINIMUM
TRIGGER
CURRENT
PULSE
ST 2 /
l\ /
1 c
BIAS LINE WITH
R, IS CONNECTED
R, IS NOT CONNECTED
(B)
Fig. 621. (A) Basic monostable trigger circuit. (B) Idealized characteristic.
118
FUNDAMENTALS OF TRANSISTORS
(A)
t ■*
i
B
1
Ia
Ea
Et
1
T "
V
^\~
j%/
'LOAD
/
LINE
>
f
(B)
3
'V
E B
E B' E A
Fig. 622. (A) Basic bistable trigger circuit. (B) Idealized characteristics.
The emitter resistance is very high in the cutoff region due to the
reverse bias. As a result, the same constant C E (r e ) r b ) R B ) is large
compared to that of the relaxation type illustrated in Fig. 617 (A).
This is the major factor limiting the repetition rate of the trigger pulse
if sensitive operation is required.
Bistable Operation. Figure 622 (A) illustrates the basic bistable
circuit. The fundamental requirement for this type of operation is that
the load line intersects the characteristic curve once in each of the three
operating regions. This automatically establishes three operation points:
one in the unstable negativeresistance region; one in the saturation re
gion; and one in the cutoff region. The last two points are stable, hence,
circuit operation is properly defined as bistable. The operation shown
in Figure (622 (B) is as follows: When operation is at point P lf the
circuit is stable, since the current is low; this is referred to as the off
state. If a positive pulse is now applied to the emitter, operation enters
the regenerative region at point A. The operation swings rapidly to
the saturation region where, at point P 3 , the circuit is again stabilized.
Since the current at this point has considerable magnitude, this is re
ferred to as onstate. To move operation back into the offstate requires
a negative trigger pulse whose magnitude is at least equal to E 3 . This
pulse moves operation back into the unstable negativeresistance region
at point B, where it rapidly swings back to the stable offstate point P v
The value of R E is selected to provide the three necessary operating
points. It is not critical and may vary considerably but, in general, it
should be fairly low. Notice that the potential of the emitter battery E e
fixes the location of V lr which in turn determines the required value
of the trigger pulse E v A low battery voltage, then, causes sensitive op
eration, since the triggering can be accomplished with a small pulse. A
large value of E e results in less sensitive but more reliable operation,
since the circuit is less likely to be triggered by noise or other unwanted
circuit disturbances. The final choice of both E e and R E should be
based on the most sensitive combination providing reliability.
Chapter 7
TRANSISTOR HIGH FREQUENCY AND OTHER APPLICATIONS
The preceding chapters discussed the basic operation, circuitry,
applications and limitations of the transistor. This chapter contains im
portant miscellaneous considerations, including transistor operation at
high frequencies, if and rf amplifiers, limiters, mixers, handling tech
niques, hybrid parameters, and printed circuits.
The Transistor at High Frequencies
Transit Time, Dispersion Effect. In the earlier chapters it was
noted that the lowfrequency, smallsignal parameters change as the op
erating frequency is increased appreciably above the audio range. Figure
71 illustrates the lowfrequency equivalent circuit of the transistor in
cluding the collector junction capacitance C,.. At higher frequencies this
equivalent circuit must be modified to include the effects of the current
carriers' transit time on the transistor parameters. The transit time of
the carriers (holes or electrons) is one of the major factors limiting the
high frequency response of the transistor.
The movement of holes or electrons from the emitter through the
base layer to the collector requires a short but finite time. In the tran
sistor, as noted earlier, the electron does not have a clear and unim
peded path from emitter to collector. As a result, the transit time is
not the same for all electrons injected into the emitter at any one
instant. The effect of an identical transit time for all electrons would
be a simple delay in the output compared to the input signal. Because
the injected carriers do not all take the same path through the transistor
body, those produced by a finite signal pulse at the emitter do not all
arrive at the collector at the same time. The resulting difference is very
small and is of no consequence in the audio frequency range. At the
higher frequencies, however, this difference becomes a measurable part
of the operating cycle, and causes a smearing or partial cancellation be
tween the carriers. Figure 72 illustrates the dispersion effect in a tran
o — \A/V~
_A/y\ r _ frJ\Lm O
INPUT
SIGNAL
PULSE
n
LOW FREQUENCY *.r~
HIGH FREQUENCE
Fig. 71. Lowfrequency equivalent circuit
of the transistor (including collector junc
tion capacitance).
Fig. 72. Transistor highfrequency
dispersion effect.
119
120
FUNDAMENTALS OF TRANSISTORS
0.1 fif
o VW
OUTPUT
Fig. 73. Transistor high
frequency equivalent circuit.
Fig. 74. Typical transistor if amplifier.
sistor at high frequencies. Notice that, in addition to the increased
period, the signal has also suffered a reduction in amplitude (the time
delay results in a phase shift) . The decrease in the output signal means
a decrease in the current gain o =
The degradation in frequency
response becomes steadily worse as the operating frequency is increased,
until eventually there is no relationship between the input and output
waveforms (and no gain) .
Another factor that limits the high frequency response of the tran
sistor is the capacitive reactance of the emitter input circuit, which
behaves as if r e is shunted by a capacitor. This reactive parameter is
reduced if the source impedance is made as low as possible. Since r b is
also effectively in series with the source, a good high frequency tran
sistor must have a low base resistance. If the source impedance and base
resistance are low, the upper frequency response limit is determined
primarily by the collector junction capacitance and the variation in
the current gain.
Alpha (o) Current Frequency. In view of these limitations, the
basic circuit illustrated in Fig. 71 is not a useful approximation of
transistor performance at high frequencies. To modify this circuit for
accurate representation of high frequency equivalence requires that all
of the internal parameters be specified in a complex form (magnitude
and phase angle) as functions of the frequency. In most cases, however,
it is sufficiently accurate to modify Fig. 71 to include only the varia
tion of a with frequency, since few design problems justify the details
required for exact equivalence. The variation in current gain can be
satisfactorily approximated by the relationship:
Ol
where o is the current gain of the operating frequency f; ai is the low
frequency current gain; and f c is the frequency at which the current
gain is 0.707 of its low frequency value (3 db down) .
TRANSISTOR HIGH FREQUENCY 121
As a numerical example of the above, compute the current gain
tor a junction transistor having a low frequency current gain of ai —
0.95, an a cutoff frequency of f c = 10 mc, and an operating frequency
of 7.5 mc. Then
01 .95
V^STiHW
0.76
Including only the junction capacitance and variation in a in the
low frequency circuit makes all the computed values far from exact. In
addition to the capacitive reactance of the emitter, there is also con
siderable variation with frequency in the collector resistance and col
lector junction capacitance. The collector resistance r c decreases rapidly
£
for a ratio of 7 — greater than 0.15, falling to about 10% of its low
£
frequency value at —? — = 1, and then remains at that value. The col
lector junction capacitance C c also decreases as the operating frequency
£
increases above an — ; — greater than 0.15, but does not decrease as
rapidly as r c . In a typical characteristic, C c drops to approximately 75%
f f
of its low frequency value at 7 — = 1 and to about 50% at 7 — =10,
after which the curve levels out. Due to the coupling between the in
put and output circuits, r t = r u — i2 2 * — , the input impedance
r 22 + KL
contains a reactive component beyond the emitter shunt capacitance.
At the a cutoff frequency f c , the reactive component is approximately
equal to the resistive input component. This causes the input imped
ance to be inductive for the grounded base connection, and capacitive
for the grounded emitter connection (due to phase reversal) .
High Frequency Equivalent Circuit. Because of these factors, rep
resentation of the transistor highfrequency operation by any linear four
terminal equivalent network is at best a rough approximation over any
substantial frequency range. This is especially true if the circuit is to be
reasonably representative of the physics of the transistor, and if the
number of crcuit parameters are to be kept within reasonable limits.
One form of equivalent circuit, suggested by Dr. W. F. Chow of the
General Electric Company, has worked out well. This involves the in
sertion of a low pass RC filter network in the low frequency circuit,
derived for an equivalent current generator in the collector arm ( a i e ) •
The modified equivalent circuit illustrated in Fig. 73 takes into ac
count the variations of t c and C c with frequency. This circuit provides
a fair representation of transistor performance through the range below
122 FUNDAMENTALS OF TRANSISTORS
the a cutoff frequency. If the operating frequency is greater than f c ,
the low pass filter must be replaced with an RC transmission line.
Frequency Comparison of PointContact and Junction Transistors.
At this point, a brief explanation of why the pointcontact transistor is
capable of a higher operating frequency than the junction type is in
order. The high frequency effects on the equivalent circuit parameters
are essentially the same for both types. Actually, the major difference
is in the mechanics of conduction.
Pointcontact transit time is determined primarily by the field
27rS 3
set up by the collector current. In equation form, T =—3 — = — where
cm /sec
S is the point spacing in centimeters, p. is the hole mobility in — pr — '
p is the germanium resistivity in ohmcm, and I is the collector current
cm /sec
in amperes. Typical transistive values are S = .003cm, u, = 100 T ^— l
r ' r r volts/cm
p = 12 ohmcm, and I c = 3 ma, for which T = 1,570 ^xsecs. Ignoring
all other factors, this limits the upper frequency response of the point
contact transistor to about 600 megacycles.
In the junction transistor, movement of the current carriers is
primarily by diffusion, and is not appreciably affected by the electrode
W 2
potential fields. In equation form T = — =r — , where T is the diffusion
time through the base layer, W is thickness of the base layer in centi
meters, and D is the diffusion constant in cm 2 /sec. Typical values for
a PNP transistor are W = 2x 10 3 cm and D = 33 cm 2 /sec (for an
NPN type, D is about 69 cm 2 /sec) , for which T — 0.121 /xsecs. Ignor
ing all factors but the diffusion time, the upper frequency for this typi
cal PNP type is approximately 8 mc, and for the NPN type about
16 mc.
High Frequency Circuits
IF Amplifiers. In general, the upper frequency limit of the junc
tion transistor is considerably lower than the limits of the pointcontact
type. On the other hand, the junction type has a lower noise factor, and
better stability in some applications. These factors frequently make it
advantageous to use the junction transistor in some high frequency
applications even if an additional stage or two may be required.
Figure 74 illustrates one stable form of if amplifier stage using
a WE 1752 NPN transistor. The operating frequency is 455 kc, and
the gain is 18 db.
Due to the natural regenerative feedback path through the collector
junction capacitance and the base resistance, and the close coupling
between the input and output circuits, the circuit, when connected in
tandem, is likely to oscillate unless the stage is carefully tuned. The
TRANSISTOR HIGH FREQUENCY
c
123
SERIES RESONANT COUPLING (A)
DIRECT PARALLEL COUPLING
(B)
PARALLEL CAPACITIVE COUPLING (C) DOUBLETUNED COUPLING (D)
Fig. 75. Transistor if coupling networks.
alignment procedure is easiest if the last stage is tuned first. For an in
put resistance R g = 500 ohms, the output resistance r averages 12,500
ohms, and C c is about 15 ^f.
The cascading of transistor if is more complicated than that of
vacuum tubes. The main contributing factors are the effect of the out
put load on the input impedance, and the effect of the generator im
pedance on the output impedance. These factors show up largely in
the design of interstage coupling networks.
IF Coupling Circuits. For interstage coupling, an if transistor
amplifier may use a series resonant circuit such as that illustrated in
Fig. 75 (A) . The main requirement for this type of coupling is that
the shortcircuit current gain is greater than unity. Thus, the series con
nection in the case of the junction type may only be used in the
grounded emitter connection.
Paralleltuned resonantcoupling circuits are applicable in if strips,
particularly when junction transistors are used. If pointcontact tran
sistors are used, special care is required to avoid oscillation due to the
inherent instability of these types when short circuited. Several types of
paralleltuned coupling circuits may be employed. Figure 75 (B) illus
trates one such possible circuit with the input of the coupled stage direct
ly connected into the resonant circuit of the first stage. This direct coupl
ing can also be used if the inductor and capacitor are interchanged.
Figure 75 (C) illustrates another coupling arrangement with the input
of the second stage connected to the junction of the two capacitors in
the resonant output tank of the first stage. In this case, the capacitors
can be used for matching the impedances between the stages. This
coupling arrangement can be made inductive by reversing the reactive
elements and connecting the input of the second stage into the tank
124
FUNDAMENTALS OF TRANSISTORS
inductance. This arrangement requires that a capacitor be inserted in
the input lead of th@ second stage. This capacitor blocks the dc bias
and also helps to ^yaid the excessive loading of the tank due to the
input circuit of the second stage. An alternate method is to couple the
second stage to the pank inductively. If the inductive coupling is also
tuned in the seconcl stage, the circuit becomes the doubletuned coupl
ing network illustrated in Fig. 75 (D) . The center tap in the induct
ance of the secondary circuit provides for the proper impedance match
between the stages.
Neutralization. The close coupling between the input and the out
put circuits of the transistor causes the resonant frequency of the coupl
ing circuit to be particularly sensitive to variations in the input and
output impedances. In general, the load impedance has a greater effect
on the input impedance than the generator impedance has on the out
put impedance. For this reason, the best procedure to follow in align
ing an if strip is to start with the last stage and work toward the first.
To avoid the critical tuning problem, the stage may be neutralized.
This allows each resonant coupling circuit to be independently ad
justed without introducing any detuning on or by the remaining stages.
One form of neutraljzation is illustrated in Fig. 76 (A) . For reasons of
clarity, only the ac circuit is shown. Neutralization is accomplished by
balancing the resistor R B and the equivalent impedance Z c of R and C
against this fransistor base resistor r b and the equivalent impedance of
the collector arm z c composed of r c and C c . The balancing conditions
are more clearly illustrated in the equivalent circuit of the neutralized
circuit, as shown in Fig. 76 (B) . The circuit is drawn in the form of a
Z,
conventional bridge. The bridge is balanced when ~— = j— . Un
Rp
der this condition there is no interaction between the input and output
o 1{
vn — \
Fig. 76. (A) Neutralized if amplifier. (B) Equivalent circuit of neutralized if
amplifier.
TRANSISTOR HIGH FREQUENCY
125
X^"
03
K
•tt x &
71
lu
<
<=2
TOW
AMPLIFIER
Fig. 77. Typical transistor rf amplifier.
Fig. 78. Junction transistor mixer circuit.
circuits. Then, when the stage is neutralized, the output impedance is
independent of R g and the input impedance is independent of R L .
In practical circuits the neutralization network design can be sim
plified by omitting the capacitor C if a pointcontact transistor is used,
or by eliminating R c if a junction transistor is used. This changes the
balance equations to ^— = ^ for the pointcontact types, and
Rn Rn
p^— for the junction types. These simplifications are pos
sible at the intermediate frequencies because feedback is governed pri
marily by r c in the pointcontact transistor, and by C c in the junction
transistor. The network components are not very critical. Values within
a 5% tolerance range are generally satisfactory.
Notice that the lower output terminal is connected to ground
through R B . This makes it important for the value of R B to be small
in order to avoid introducing too much noise through R u into the
output circuit. For satisfactory operation, the value of R B should not
be larger than the base resistance. This fixes the value of C in the range
of C c , and Rc in the range of r c . The loss in gain due to the neutralizing
network will be less than 10% of the total gain in a properly designed
circuit.
RF Amplifiers. Transistor rf amplifier circuits, like their counter
part vacuumtube circuit types, are most often used for improving the
gain, overall signaltonoise ratio, or selectivity characteristic of a multi
stage circuit. Figure 77 illustrates a typical transistor rf amplifier cir
cuit. The design is basically the same as that of an if amplifier. The
chief problem is the selection of a transistor having a sufficiently high
a cutoff. The power gain of a rf amplifier is inversely proportional to
the frequency. In a typical case a transistor having a maximum gain
of 40 db at 10 mc will have a maximum gain of 20 db at 40 mc. There
are two critical parameters, the emitter bias and the base resistance.
The base resistance is determined by the physical construction of the
transistor and, therefore, low baseresistance transistors, designed specif
126
FUNDAMENTALS OF TRANSISTORS
ically for high frequency applications, should be used. The importance
of emitter bias was considered in the analysis of oscillator circuits. The
bias should be selected to be far enough away from the unstable region
of the characteristics to avoid oscillation, and yet not so far away that
the gain is very low. Special care must be taken to avoid introducing
stray capacitance into the emitter input circuit. These reactances tend
to lower the input emitter impedance, and thereby decrease circuit
stability.
Limiters. Limiter circuits can be designed using transistors and
germanium diodes. These circuits operate much like vacuumtube limit
ers. In the grounded base connection, the input circuit acts like a diode
when the emitter electrode is biased slightly in the forward direction.
When the value of the input signal exceeds that of the emitter bias, the
signal is rectified by the diode action of the input circuit. The resulting
selfbias tends to keep the maximum emitter current constant. Since
the collector current is proportional to the emitter current, the output
signal is maintained at a constant level over a large range of input
signal values. The input rectification action is considerably improved
when the circuit is shunted by a junction diode. The diode performs
two important jobs. It clips large positive input pulses, and prevents
the coupling capacitor from charging on extraneous noise pulses. For
optimum operation, the output resistance is matched to the load, and
the generator impedance is kept as low as possible.
Mixers
The operation of the transistor in mixer circuits depends upon
the rectification and nonlinearity of the emitter circuit when it is
biased slightly in the forward direction. Figure 78 illustrates one basic
arrangement of a transistor mixer circuit employing a junction transis
tor. This circuit takes advantage of the relatively high gain of the
grounded emitter connection by injecting the BFO signal into the
common emitter lead. The junction transistor works well in mixer stages
despite its relatively low a cutoff. This is possible because only the if
frequency must lie within the useful frequency range of the transistor.
The pointcontact type also works satisfactorily in transistor mixer cir
cuit. Its utility is limited to some degree by its relatively high noise
figure and low gain.
Fig. 79. Transistor power supply.
TRANSISTOR HIGH FREQUENCY 127
Power Supplies. As the number of applications for transistors in
crease, many new power supply systems will be required to fit in effi
ciently with the particular design. The power requirements of an in
dividual circuit is very small, so small in fact, that quite often the life
expectancy of the bias battery is the same as its normal shelf life. Never
theless, in some applications it may be desirable to derive the power
supply from an existing ac source. Figure 79 illustrates an experimental
power supply, fabricated for a particular application, where a bias of
30 volts and a drain of 10 ma were required. The circuit is a basic full
wave rectifier terminated in an RC filter. With the values shown, the
ripple is less than one percent.
Miscellaneous Transistor Characteristics and Handling Techniques
Transistor Life Expectancy. One of the outstanding features of the
transistor is its practically indefinite life expectancy. Long life was ori
ginally predicted on the basis of the transistor construction and its con
duction mechanics, which indicate there is nothing to wear out. Al
though the transistor is still very young, enough experimental data is
now available to back the initial long life predictions.
The usual transistor failure occurs gradually over a long period
of time and after thousands of hours of operation. The performance
degradation generally shows up as an increasing saturation current.
(The effect of increasing saturation current was covered in the tran
sistor amplifier chapter.) While the various self and fixed biasing meth
ods may be used to minimize the effects of increasing I co , the system's
efficiency and gain suffer. In an amplifier circuit, this factor decreases
the available volume. Gradually, as the limit of the automatic biasing
arrangement is reached, there is also a noticeable increase in the dis
tortion content.
Another variation in the transistor performance characteristics is
a gradually decreasing output resistance. In systems designed for an
image impedance match (R g = r, and R L = r ) , this change introduces
a mismatch loss. In the usual amplifier design, however, the output re
sistance is in the order of 20 to 50 times the load resistance. The de
crease in r , therefore, is less serious than the increase in I co . The best
single maintenance check is a measurement of the current gain.
Sudden failures of transistors are not common in normal opera
tion, although open emitter and collector junctions were not too rare
in the early transistors. These defects were attributed to faulty assembly
during manufacture. Present manufacturing and quality control tech
niques have practically eliminated open junction defects. Transistor
shorts are more common since they are usually caused by overloading.
When the transistor power rating is exceeded, the junction temperature
rises quickly. The increased heating effect encourages diffusion of col
lector region impurities into the base layer, which, in time, will destroy
128 FUNDAMENTALS OF TRANSISTORS
the junction. In brief then, open circuited transistors generally result
from poor production; shortcircuited junctions generally mean im
proper circuit design.
Transistor Ruggedness. Insofar as ruggedness is concerned, the
superiority of the junction transistor compared to the pointcontact type
can be anticipated from a comparison of the basic construction details
(Chapter 2) . The emitter and collector electrodes of the pointcontact
type depend on a force contact with the germanium surface. These cat
whiskers, it will be remembered, are fastened to the main electrode con
ductors which are embedded in, and held by, the plastic stem. It is pos
sible, then, to vary the contact pressure of the catwhiskers by a twisting
force applied to the plastic stem. This distortion can be introduced by
direct mechanical force, humidity or temperature variations.
Most of the present transistors are hermetically sealed. Sealing is
important because of the ease with which an unprotected junction sur
face may be contaminated by water vapor. The contaminating effects
are particularly noticeable so far as the value of the saturation current
in an unsealed unit is concerned. In a typical case, the saturation cur
rent of a junction transistor will increase one hundred times its dry air
value when the relative humidity is increased by 50%.
The transistor can withstand shock, vibration, and drop tests far
beyond those of the vacuum tube. However, it is a good plan to treat
the transistor with reasonable care to avoid unnecessary damage. The ef
fect of distortion of the stem on electrode contact pressure was noted
in earlier paragraphs. Any damage to the hermetic seal is, of course,
serious. Transistor electrode leads are generally as flexible as those of
regular carbon resistors. These leads should not be subjected to con
tinual bending or flexing, or to pulls greater than a halfpound.
Soldering Techniques. Generally, junction transistors (Raytheon
types 720, 721, 722, Germanium Products Corporation types 2517, 2520,
2525, Western Electric 1752, etc.) have long pigtail leads. These types
can be soldered directly into a circuit. However, due to the temperature
sensitivity of the transistor, solder connections must be made quickly.
It is always a good idea to heat sink all solder connections by clamping
the lead with a pair of long nose pliers connected between the soldered
point and the transistor housing. This provides a shunt path for a large
part of the heat introduced at the solder joint. If it is at all possible,
transistors with short leads should not be soldered directly into the
circuit. Several types of sockets will accommodate these short lead types.
For example, the Cinch type 8749, type 8672, and regular 5pin sub
minature tube sockets will handle pointcontact transistors similar to
the Western Electric 1698, the General Electric Gil A, etc.
Temperature Effects. The physical location of the transistor is not
critical with respect to its mounting position, and since the heat gen
TRANSISTOR HIGH FREQUENCY 129
erated by an individual transistor is small, many may be packed to
gether. However, since the transistor is sensitive to the ambient tempera
ture, hot spot locations near tubes and power resistors should be avoided.
In this regard, a word of precaution on collector dissipation ratings is
in order. The maximum collector dissipation is specified at some defi
nite temperature (usually 25 °C) . This value must be derated if the
ambient temperature is greater than the specified rating temperature.
Usually this amounts to a 10% decrease in power dissipation for each
5°C increase in ambient temperature. As a numerical example, assume
that the maximum allowable collector dissipation for a transistor rated
at 250 mw at 25°C is required when the ambient temperature is 40°C.
The operating temperature represents an increase of 40° —25° = 15°C.
The power handling capacity should be derated 10% for each 5°C in
crease or 15/5 x (10) = 30%. 30% of 250 mw is 75 mw. Thus the maxi
mum collector dissipation at 40°C is 250 — 75 = 175 mw.
Whenever a transistor is operated near its maximum rating, it is
always good insurance to tie it to a metal panel or chassis. This connec
tion provides a large radiating surface which permits the collector dissi
pation to be maintained at higher levels. In typical cases, this procedure
increases the transistor power dissipation rating from 20 to 50%.
Transient Protection. In addition to its power handling limitations,
the transistor is susceptible to damage by excessive values of current
and voltage. It is particularly important to protect the transistor from
those transient surges which may be caused by switching or sudden
signal shifts. Transient effects are particularly predominant in oscilla
tor, if, and high frequency amplifiers due to the storage capacity of
the reactive components. Limiting devices are usually incorporated into
the circuit. The series resistors in the emitter and collector arms of the
basecontrolled negativeresistance oscillator are typical examples. In
more complicated circuits, transient limiting elements are usually selec
ted on the basis of tests made on experimental breadboard models. If
a scarce or expensive transistor is involved, the equivalent passive "T",
made up of standard carbon resistors, can be substituted for this meas
urement. When connecting a transistor into a live circuit, the base lead
must always be connected first. In disconnecting the transistor from a
live circuit, the base lead must be removed last.
It is an easy matter to mistakenly reverse the polarities of bias sup
plies, particularly in complementarysymmetrical circuits. Reverse polar
ities will not impair the transistor as long as the maximum ratings are
not exceeded. It is always a good plan to check for proper capacitor
polarity, since almost all of the circuits require polarized types.
Hybrid Parameters
Significance and Derivation. The opencircuit parameters, r u , r ]2 ,
r 21 , and r 22 are used exclusively throughout this book primarily because
130
FUNDAMENTALS OF TRANSISTORS
they are the most familiar fourpole equivalents. Some engineers prefer
the shortcircuit conductance parameters g n , g 12 , g2i. and g 22 . The con
ductance parameters serve well for the junction transistor, but do not
work out too well for the pointcontact type, which inherently exhibit
short circuit instability.
The disadvantages in both the r and g forms suggest a combination
or hybrid type of representation which will be applicable to all tran
sistor types without requiring elaborate measuring techniques. The so
called 'h' or hybrid parameters are becoming more and more popular.
Since many of the manufacturer rating sheets now specify the h para
meters, it is important to be able to convert the hybrid values into the
more familiar r form for use in the design and performance equations.
On a fourterminal basis, the hybrid parameters are equated as:
ei = hnii f h 12 e 2 Eq. (71)
i 2 = h^ii j h 22 e 2 Eq. (72)
The basic circuits for measuring the h parameters are illustrated
in Fig. 710, which define the values of the parameters in terms of the
input and output currents and voltages as follows:
1 when e 2 = (output shortcircuited)
h u =
«i2 =
h 2 i =
hoo =
ii
_£i_
ii
e 2
when i L = (input opencircuited)
when e 2 = (output shortcircuited)
when i t = (input opencircuited)
TEST
SIGNA
'1
L<~)*\
hn r When e 2 «o
11 (A)
«z / nj ) TEST
' ' ' SIGNAL
12
h 2  • — WHEN«2«0
(C)
Fig. 710. Basic circuitt for measuring fourterminal fi parameters.
TRANSISTOR HIGH FREQUENCY 131
Notice that two of the measurements are made with the output short
circuited, and the remaining two are made with the input opencir
cuited. Furthermore, none of the parameters are exact equivalents, since
r n is a resistance (ohms) , h 22 is a conductance (mhos) , h 12 is a numeric
(voltage ratio) , and h 21 is also a numeric (current ratio) .
Resistance Parameters in Terms of Hybrid Parameters. The rela
tionship between the r and h values can be determined by straightfor
ward substitution and the simultaneous solution of equations 71 and
72, as follows:
A. r n = A when i, = 0. Under this condition
ii
ei = hnii + h 12 e 2 Eq. (71 A)
= h 21 ii f h 22 e 2 Eq. (72A)
If these are solved simultaneously,
eih
n ll n 22 — n 12 n 21
^ — = , and therefore
h u h 22 — h i2 h 21 „ ,
r u p Eq. (73)
n 22
B. r 21 = A when i 2 = 0. Under this condition equation 72A still
applies and is solved
e = = (fe)' dr '={fe) *»™
C. r 12 = — i when ij = 0. Under this condition
ei = h 12 e 2 Eq. (71B)
i 2 = h 22 e 2 Eq. (72B)
If these are solved simultaneously
_£3_ = ilandr 12 =^ Eq. (75)
1 2 " 2 2 "22
e 2
D. r 22 = — r— when i t = 0. For this relationship equation 72B
*2
still
applies and is solved
e 2 =r^— and r 22 =r — Eq. (76)
i 22
E. The current gain a = — — = ^" aa ' = — h 21
_ v5r/
1 2 2
132 FUNDAMENTALS OF TRANSISTORS
As a numerical example of these conversions, the manufacturer's
rating sheet for the G.E. type 2N45 specifies the following typical values
for the hybrid parameters:
h u = 40 ohms, h 12 = 2.5 x 10*, h 21 =  .92, h 22 = 1.0 x 10— mhos.
Then r u = h nh 22  h 12 h 21= 40 (1.0 x 10) 2.5 x 10 (92)_ m ^
h 22 1.0x10—
r  = ^ = 0^r = 920 ' 000ohms
h 12 2.5x10— OKft ,
rM= S^ = i.Oxio =250ohl]M
r22 = "hlT == i.ox 1 io = l megohm
a = _h 21 =  (.92) = 0.92
Printed Circuit Techniques
One of the most promising features of the transistor is its ability
to fit into the new prefabricated wiring techniques, by which the maze
of handsoldered wires normally associated with electronic equipment
has been eliminated. Basically, a printed circuit starts with a metal foil
bonded to one or both sides of an insulating plastic material. The metal
foil may be copper (most popular) aluminum, silver, or brass. Most
types of laminated plastics are suitable as the base insulator. The circuit
is drawn on the foil clad laminate with an acid resistant ink. The com
plete assembly is then dipped into an etching solution which removes
the metal not protected by ink. Holes are then drilled or punched into
the assembly at appropriate points, and into these holes the various
circuit components are inserted and soldered to the metal foil. If the
circuit is at all complex, hand soldering is extremely tedious and diffi
cult, and the dip soldering technique is used. In this method, compo
nents with preformed leads are inserted into the holes, either manually
or by an automatic process. After fluxing, all the connections between
the component leads and the circuit pattern are accomplished by a
"oneshot" dip in a molten solder bath. Those portions of the circuit
which must be left free of the solder are coated with a protective lacquer
or masked before the solder bath. Dip soldering assures very reliable
solder joints in one simple operation, and also permits a greater reduc
tion in size by means of stacking techniques, which were previously lim
ited by the space requirements for hand soldering operations.
Complex circuits are normally laid out on both sides of the lami
nated base. Connections crossovers may be made by several methods.
The most common is by means of a tined eyelet. This is of particular
importance in those cases where connection is made to a component
TRANSISTOR HIGH FREQUENCY 133
Fig. 711. Experimental traiufetor if amplifier.
which may be soldered and unsoldered several times during the life of
the equipment. Repeated soldering at the foil will eventually cause
it to lift from the plastic base.
In spite of the small crosssectional area of the foil conductors, the
current carrying capacity of the printed circuit is good, due to the rela
tively large surface area and the heat conduction by the base material.
A 1/32inch copper foil conductor, for example, can safely handle about
five amperes. Increased temperatures caused by current overloads causes
the metallic conductor to buckle and separate from the base.
One of the major advantages of the printed circuit is its uniformity
from unit to unit. For example, the distributed capacitance between
foil conductors is in the same order of magnitude as that of a carefully
hand wired assembly. In the prefabricated type, however, the value re
mains constant from unit to unit because they are all produced from
the same master design.
Figure 711 illustrates the front and back of an experimental
printed circuit type of transistor if amplifier. The component arrange
ment can be seen at the left of the illustration and the printed wiring
can be seen at the right. Miniature components for use with transistors
are shown in Fig. 712. The top row of the figure shows a miniaturized
transformer and three resistors. The bottom row illustrates an inductor,
a capacitor, two junction transistor sockets, and two pointcontact tran
sistor sockets.
The marriage of standard and miniaturized components with the
basic printed circuit is, in essence, the "autosembly" technique devised
by the Signal Corps Engineering Laboratories. This method is best
suited to present production facilities, since it utilized components with
proven reliability. However, the recent progress in the development of
printed components indicates that most of the applications of prefabri
134
FUNDAMENTALS OF TRANSISTORS
Fig. 712. Miniature transistor components.
cated circuits are still to come. Printed resistors having values of 10
ohms to 10 megs and which are sprayed onto an area of 1/16 of a
square inch have been used successfully. Small inductance coils, having
values up to 20 ^Ji, can be etched into the printed circuit, and capacitors
ranging from 10 w f to .001 fd can be incorporated in the printed circuit
by etching opposite sides of foilclad glasscloth laminates.
The transistor, because of its mechanical ruggedness and long life
expectancy, is well adapted for direct assembly into printed circuit pat
terns. The minute heat generated by the transistor makes its future use
in compact packaged equipment particularly promising. The prefabrica
tion techniques will initially reduce the outofservice time considerably,
since complete circuits will be encapsulated in units no larger than
present vacuum tubes. On the other hand, assembly repairs will require
great skill and technical knowledge due to the complex arrangement
of the miniaturized components.
Appendix
COMMON TRANSISTOR SYMBOLS
AG
a
b
c
Co
C e
e
E 1
E
Ebb
E e
E b
E c
e b
U
G
gl2
gn
h tl
hit
h tl
h u
h
h
L.n
Available gain
Current gain
Base electrode
Collector electrode
Collector junction capacitance
Emitter junction capacitance
Emitter electrode
Input voltage • 4terminal network
Output voltage  4terminal network
Battery supply voltage
Emitter bias battery
Base bias battery
Collector bias battery
Ac base signal voltage
Ac collector signal voltage
Ac emitter signal voltage
Cutoff frequency
Operational gain
Small signal short circuit input
conductance
Small signal short circuit feedback
conductance
Small signal short circuit transfer
conductance
Small signal short circuit output
conductance
Small signal hybrid short circuit input
impedance
Small signal hybrid open circuit volt
age feedback ratio
Small signal hybrid short circuit for
ward transfer current ratio
Small signal hybrid open circuit out
put admittance
Input current  4terminal network
Output current  4terminal network
Saturation current  collector current
at zero emitter current
Dc base current
Dc collector current
h
h
h
Re
Re
To
r,
•1
Dc emitter current
Ac emitter signal current
Ac base signal current
Ac collector signal current
Base series resistor
Collector series resistor
Emitter series resistor
Input resistance 4terminal network
Output resistance 4terminal network
Image matched input resistance
Image matched output resistance
Small signal open circuit input
resistance
Small signal open circuit reverse trans
fer resistance
Small signal open circuit forward
transfer resistance
Smal signal open circuit output
resistance
Transistor equivalent base resistance
Transistor equivalent collector resistance
Transistor equivalent emitter resistance
Proportionality resistance constant be
tween emitter signal current and re
sulting voltage signal produced in col
lector arm
Internal resistance of signal generator
Load resistance
Base junction thickness
Voltage gain
MAG
Maximum available gain
NType
Transistor semiconductor with donor
type impurities (electron current car
riers)
PType
Transistor semiconductor with acceptor
type impurities (hole current carriers)
R
Rl
W
VG
135
INDEX
Acceptors, 4
Active networks, 20
Admittance, 21
Alpha, a, 2
cutoff, 15, 120127
definition, 10
measurement, 70
phase shift, 107109
variation, 120121
Amplification factors, 2122
Amplifiers, 7195, 122127
bias methods, 7275
cascade operation, 8690
class A, 7882
class B, 8284
complementarysymmetry, 9295
coupling and decoupling, 9092
current sources, 76
dc operating point, 7172
gain controls, 8586
if stages, 122123
limiter, 1 26
mixers, 126
neutralization, 124125
power supplies, 127
rf stages, 125
Analysis, four terminal,
active, 28
networks, 19
parameters, 20, 27, 28
passive, 26
power gain, 3942
transistor, 25
vacuum tubes, 22
Available gain, 40
B
Base,
control, 114
resistance, 28
grounded, (see Connections)
Bias,
battery, 7
fixed, 7273
for oscillation, 104105
forward, 6
HunterGoodrich, 74
reverse, 6
self, 7374
self plus fixed, 75
Bonds, covalent, 2
Buffer, grounded collector, 60
Capacitance, junction, 42
Capacitors,
bypass, 78
coupling, 78, 79
Collector,
capacitance, 42
characteristics, 14, 15
maximum limits, 65
minimum limits, 66
resistance, 28
Compensation, phase shift, 109
Complementarysymmetry,
advantages, 95
cascade, 9495
pushpull, 93
theory, 92
Conductors, 1
Connections,
comparison, 6465
current gain, 30
equivalent circuit, 30
groundedbase, 3043
impedance matching, 3637
input resistance, 31
137
138
FUNDAMENTALS OF TRANSISTORS
output resistance, 32
power gain, 3941
stability factor, 38
voltage gain, 34
Coupling circuits, 9091
Crystal,
control, 110111
structures, 2
Dc operating point,
fixed bias, 7273
fixed plus self, 75
HunterGoodrich method, 7475
selection, 72
self, 7374
Decoupling circuits, 9192
Diffusion constant, 122
Dispersion effects, 119
Distortion, 79
Donars, 4
Electrons, 1, 4, 6,
carriers, 13
surface bound, 8
Emitter,
capacitance, 120
control, 112114
grounded (see Connections)
resistance, 28
Equal voltage method, 68
Feedback oscillators, 9699
Fourpole networks, 1922
Frequency,
alpha cutoff, 15, 120127
dividers, 116
high, operation, 119126
multipliers. 111, 116
Gain,
controls, 8586
current, 10, 13, 30, 47, 56
overall, 87
power, 11, 14, 36, 3941, 5254,
6163
resistance, 11, 14
voltage, 11, 14, 34, 51, 6061
Germanium,
impurities, 5
intrinsic, 5
Ntype, 4, 6, 12
Ptype, 4, 6, 12
Ground, system, 71
Grounded base (see Connections)
Grounded collector, 5464
buffer stage, 60
conditional characteristics, 62
current gain, 56
equivalent circuit, 5455
impedance matching, 58
input resistance, 57
output resistance, 57
parameters, 5455
power gain, 6163
reverse power gain, 6364
stability factor, 5860
voltage gain, 6061
Groundedemitter, 4554
conditional stability, 5253
current gain, 47
equivalent circuit, 45
impedance matching, 51
input resistance, 48
output resistance, 49
parameters, 4547
power gain, 5254
stability factor, 5152
voltage gain, 51
tandem, selection, 8790
INDEX
139
H
Handling techniques, 127129
Hartley oscillator, 9697
Highfrequency operation,
alpha cutoff, 120121
circuits, 122126
effects, 119120
equivalent circuit, 121122
point contact vs junction, 122
Hills, potential, 56
Holes,
carriers, 13
injection, 9
theory, 4, 6
Hybrid parameters, 129132
I
lf alignment, 122
Impedance,
input, 31, 48, 57
matched, 3637, 51, 58
open circuit, 2022
output, 32, 49, 57
Impurities, 35
Insulators, 1
Intrinsic germanium, 5
Inverters, 8485
Junctions, PN, 5
Junction transistors,
compared to pointcontact, 1415
construction, 1 2
NPN, 12
PNP, 1213
Life expectancy, 127128
Limitations, 6567, 119121
Limiters, 126
Load lines, 72, 8081
M
Matched impedances,
grounded base, 3637
grounded collector, 58
grounded emitter, 51
Matter, structure of, 1
Measuring,
alpha, 70
circuits, 2829
negative characteristics, 104
saturation current, 70
Mixers, 126
Multivibrator, 99
N
Ntype germanium, 4
Networks,
active, 20
fourterminal, 1922
passive, 20
Neutralization, 124126
Noise, 66
Oscillation, conditions for, 99104
Oscillators, 96118
bias selection, 104105
Clapp, 9798
Colpitts, 9899
crystal, 110
feedback, 9699
frequency multiplication, 1 1 1
Hartley, 9697
multivibrator, 99
negative resistance, 99111
phase shift, 108
relaxation types, 111117
sinewave, 103104
stabilization, 105106
trigger circuits, 117118
140
FUNDAMENTALS OF TRANSISTORS
Ptype germanium, 4
Parameters,
admittance, 21
amplification, 21
hybrid, 129132
impedance, 20
measurements, 2829
smallsignal, 23
transistor, 28
variation, 119121
Phonons, 3
Photons, 3
Power factors,
available from generator, 39
available gain, 40
maximum available gain, 41 , 54, 63
operating gain, 39, 52, 6162
Power supplies, 127
Printed circuits, 132134
Pushpull operation, 8183
Q, oscillator tank, 97
Relaxation, oscillators,
base controlled, 114
basic operation, 111114
collector controlled, 114
emitter controlled, 112113
selfquenching, 114115
synchronized, 115116
Resistors, printed, 134
Ruggedness, 128
Saturation current, 43
Semiconductors, 12
Soldering techniques, 128
Stability, 5253, 62, 99100, 105107
Stability factor, 38, 5152, 5860
T
Tandem — stage connections, 8790
Temperature effects, 129
Test sets, 6970
Testing,
precautions, 67
transistors, 6769
Transient protection, 129
Transistor parameters, 28
Transistor types,
compared to vacuum tubes, 23
comparison of, 14
junction, 12
NPN, 12
Ptype, 11
PN junction photocell, 15
PNP, 1213
PNPN, 17
pointcontact, 8
tetrode, 17
widespaced, 16
Transit time, 119
Trigger circuits, 117118
astable, 117
bistable, 118
monostable, 117118
Variations,
alpha cutoff, 120121
high frequency, 121122
saturation current, 7374
W
Waveforms,
distortion, 79
sawtooth, 111112
Zener, voltage, 42