I»l»
Radio
1
Pitman
F^Derbacks
J
H REYNER
P J REYNER
Communication
/
/
/
^
h
zz mm
mm <•<
J3J3
$
PFI
Radio
Communication
J. H.
B.SC, A.C.G.I.,
D.I.O.,
Reyner
MEMBER
I.E.E.E.
C.ENO., F.I.E.E., LIFE
and
P. J.
Reyner
M.A. (CANTAB.), C.ENG., M.I.E.E.
Third Edition
SI Units
Pitman Publishing
Third edition 1972
The paperback edition of this book may not be lent, resold, hired out or otherwise disposed of by way of trade in any form of binding or cover other than that in which it is published without the prior consent of the publishers.
SIB ISAAC PITMAN AND SONS LTD. Pitman House, Parker Street, Kingsway, London, WC2B 5PB
P.O.
Box
46038, Portal Street, Nairobi,
Kenya
SIB ISAAC PITMAN (AUST.) PTY. LTD. Pitman House, 158 Bouverie Street, Carlton, Victoria 3053, Australia
PITMAN PUBLISHING COMPANY
P.O.
Box
S.A. LTD. 11231, Johannesburg, South Africa
6 East 43rd Street,
PITMAN PUBLISHING COBPOBATION New York, N.Y. 10017, U.S.A.
SIB ISAAC
PITMAN (CANADA) LTD.
495 Wellington Street West, Toronto, 135, Canada
THE COPP CLARK PUBLISHING COMPANY
517 Wellington Street West, Toronto, 135, Canada
J.
H. Reyner and P.
1972
J.
Reyner
273 36164 3 Cased edition: ISBN Paperback edition: ISBN 273 36165
1
Text set in 10/H pt. Monotype Modern printed by letterpress, and bound in Great Britain at the Pitman Press, Bath
G2— (T.5426/1352:74
Preface to Third Edition
The response
the scope
to the earlier editions has encouraged us to enlarge
by the inclusion of a treatment of logic circuitry and similar
developments of interest to the practising electronics engineer. The basic coverage remains that required for the City and Guilds Telecommunications Technicians' Course, with a slight change of emphasis. Despite the almost universal use of solidstate techniques in many fields, thermionic devices still play an important role in certain applications, and will continue to do so for many years. The treatment of valves in Part I has therefore been maintained, though certain obsolete matter has been omitted and the space devoted to an extended discussion of semiconductor usage, including
fieldeffect transistors.
The chapters in Part II, dealing with practical applications, treat valve and transistor circuitry separately, as before, but the discussions of transistor usage have been appropriately extended, while Chapter 16 has been enlarged to cover colour television. Two new chapters have been added. The first of these deals with solidstate switching devices, including logic circuitry and the use of integratedcircuit techniques, while the second discusses briefly some of the specialized developments in space communications. We have retained the wide selection of examples from examination papers which we believe to be of value. A study of recent papers shows that they do not, in the main, include any material not already covered in previous years. The fundamental basis of the art is in fact stable; hence although some 1970 examples are included, extensive updating appears neither necessary nor desirable. Finally, Appendix 2 has been modified to cover the recently agreed SI units, which have been used throughout the book in place of the former MKS units, while a new Appendix 4 has been added summarizing the principal techniques of transistor construction.
Little Missenden March, 1972
J.
H. R.
P. J. R.
.
.
Contents
PAGE
Preface.
v
1.
Fundamentals 1.1. Basic Concepts 1.2. Ohmic Conduction; Resistance
.
....
.
Part
I.
Basic Principles
3 3 7 13
1.3. Electrostatics; 1.4. 1.5. 1.6.
Capacitance Blectromagnetism; Inductance Alternating Currents Conduction in Oases and Liquids
Vibrations and Waves Sources of E.M.F.
1.7.
1.8.
24 38 53 58
68
71
2. Circuit
Theory
.
2.1. 2.2. 2.3. 2.4.
Vectors
Network Theory Tuned Circuits
.
Transformers; Coupled Circuits
of Radio Communication Production of Radio Waves Modulation Reception
.
71 80 00 104 117 117
133 148
3. Principles
3.1.
3.2.
3.3.
4. Practical
Components
.
4.1. Resistors 4.2. 4.3.
Capacitors . Aircored Inductors 4.4. Ironcored Inductors and Transformers 4.5. Generators, Motors and Meters 4.6. Batteries .
4.7. 4.8.
5.
152 152 160 171 178
Microphones Loudspeakers
204 218 222 224
230 230 232 243 251 255
Thermionic Valves 5.1. Electron Emission
5.2. Practical
5.3.
Forms of Valve Basic Valve Usage
Tubes
. .
5.4. Gasfilled
5.5.
Photocells
.
.
W/7
6.
.
CONTENTS
Semiconductors 6.1. Semiconductor Physics 6.2. Semiconductor Diodes
6.3. 6.4. 6.5. 6.6. 6.7. 6.8.
The Transistor
.
Transistor Performance Transistor Parameters Practical Transistor Circuits Fieldeffect Transistors Photosensitive Semiconductors
259 259 262 267 277 286 294 307 312
Part
7.1. 7.2.
7.3.
II.
Practical Applications
319 319 329 337 348 351 355 360 360 369 377 392 405 406 410 423 434 450 464
472 472 475 493 509 516 539
The Radio Transmitter
Valve Transmitter Circuits Frequency Stability Modulation Terminal Arrangements F.M. Transmitters
Transistor Transmitters
.
7.4.
7.5. 7.6.
Aerials and Feeders
8.1.
8.2. 8.3.
8.4.
The Transmitting Aerial The Receiving Aerial Feeders and Transmission Lines
.
Directional Aerials
.
The Radio Receiver
9.1.
Radiofrequency Amplification 9.2. R.F. Valve Amplifiers 9.3. R.F. Transistor Amplifiers 9.4. The Superheterodyne Receiver 9.5 The Detector Stage 9.6 F.M. Receivers
. .
10.
Audiofrequency Technique.
10.1. 10.2. 10.3. 10.4.
Feedback. A.F. Valve Amplifiers
A.F. Transistor Amplifiers Sound Recording
Circuits
11.
Power Supply
12. Oscillator Circuits 13.
Shortwave Technique 13.1. Propagation 13.2. Shortwave Aerials
13.4.
.
13.3. "Ultrashortwave Aerials
Waveguides
Shortwave Receivers Shortwave Transmitting Valves Semiconductor Techniques
13.5.
13.6. 13.7.
552 552 555 570 573 586 593 605
CONTENTS
14. Filters 14.2. 14.3.
Ix
and Attenuators
14.1. Passive Filters
Active Filters Attenuators and Equalizers
15.
Measurements
15.1. 15.3. 15.4. 15.5. 15.6.
Voltage and Current 15.2. Resistance, Inductance and Capacitance
.
.....
Oscilloscope
.
.... ....
610 610 627 634 640 640 651 656 658 662 668 679 680 680 685 692 695 698 698 701 710 721 726 727 736 748 754 761 763 763 766 770 771 773
Frequency and Wavelength B.F. Measurements . Performance Tests
The Cathoderay
15.7. Ballistic
Technique
16. Picture Transmission
16.1. Principle of 16.2. Television 16.3. Television
and Television Scanning
. .
Cameras
Receivers 16.4. Colour Television
17.
Wave
17.1. 17.2.
Shaping and Pulse Techniques Shaping Pulse Generation; the Multivibrator
Wave
....
.
17.3. Operational Circuitry 17.3.
Sawtooth Generation.
and Control Circuitry
18. Switching
18.1.
Switching Devices 18.2. Logic Circuitry 18.3. Monolithic Circuits
18.4. Bistable Circuits 18.5.
Core Stores
19. Specialized
19.1. 19.2. 19.3.
Communication Techniques. Radar
Weaksignal Amplifiers Telemetry
.... .... ....
. .
.
19.4. Optoelectronics
Appendix
1.
Atomic Structure
Appendix Appendix
2. 3.
Units and Dimensions
778 784
785
791
The Binaby Scale
Appendix
Examples
Index
4.
...... ......
Teansistob Constructions
857
Plates
pig.
4.6.
facing page
Typical Fixed Capacitors
162
4.6(a).
4.7.
4.8.
A Variable Capacitor
Typical Receiving Coils
.....
.
162 163 163
Typical Constructions Using Ferrite Material
13.23.
Steerable Directional Downs Radio Station
......
Aerial
at
.
Goonhilly
572 573
13.24.
Aerial Array at Baldock Radio Station
.
Part
I
Basic Principles
Fundamentals
1.1.
Basic Concepts
Telecommunications engineering is concerned with the transmission of information by utilizing the effects produced when an electric current is varied. The variations may be transmitted by means of wires, as is done with telegraph and telephone systems, or by means of electromagnetic waves as used in radio communication. It is this second category with which this book is primarily concerned, though some discussion of the simpler aspects of telegraph and
telephone techniques will necessarily be required. The fundamental requirement for a proper study of the subject is a clear understanding of the nature of electricity, which is no longer regarded as a mysterious invisible fluid pumped through wires by a battery, but rather as the movement or displacement of fundamental particles called electrons in the atoms of the materials themselves. Hence the study and utilization of electrical phenomena constitute the twin sciences known as electronics and electronic
engineering.
Atomic Structure
All matter is believed to be composed of atoms which are like miniature solar systems containing a central positively charged nucleus surrounded by negatively charged electrons revolving in
definite orbits.
These atoms do not normally exist alone but associate themselves with other atoms, either of their own kind or of some other element, to form molecules in which a number of atoms group together to provide an equilibrium between the various internal forces, under conditions of minimum energy. In gases these molecules remain as independent entities moving at random through the space in which they are confined. If two molecules happen to collide they bounce apart like elastic balls. In
3
4
solid materials,
RADIO COMMUNICATION
however, the molecules have
much
less
freedom, and
generally associate themselves with other molecules to form symmetrical assemblies of either crystalline or chainlike formation.
be realized that all matter is very largely open space. of an atomic nucleus is of the order of 10 11 mm, while the electrons, which are of similar size, revolve in orbits of the order
It should
The diameter
diameter, i.e. some 10,000 times the size of the nucleus. of 10 7 It is the powerful internal forces resulting from the relatively compact grouping of the molecules in a solid which cause it to appear impervious and reflect light waves so that it looks solid. Liquids occupy a position between solids and gases. The internal forces are weaker so that the molecules have less restraint and the substance will behave in some ways like a solid and in others more
like
mm
a gas.
Free Space
Since the structure of the
forces, there
atom involves the interplay of internal must be a medium through which they operate. It is
clearly nonphysical since the forces are found to exist in a complete vacuum, and 19thcentury physics postulated the existence of what was called the ether to fulfil this function. This has been replaced today by the concept of free space, which is said to possess
certain specific electrical properties, as
is
discussed later.
One can then regard the subatomic
particles as local distortions
or condensations of free space, which create a sphere of influence in their immediate environment. They are said to carry an electric particle carrying a negative charge, which may be of two kinds. charge is called an electron, while a positively charged particle is
A
called a proton*. The nature of the force is such that similarlycharged particles will repel one another, while dissimilar charges will be mutually attracted. The nucleus of an atom thus contains an appropriate number of protons surrounded (normally) by an equal number of electrons rotating in orbits at distances and speeds which produce a centrifugal force sufficient to counteract the electrical attraction between the two opposite charges, and hence maintain a state of equilibrium. (The nucleus may also contain some additional particles called neutrons, which carry no charge but merely add to the mass, as
discussed in Appendix
1.)
* The electron is assigned a negative charge because in the classic early experiments with minute oildrops carrying an accumulation of electrons it was found that the force exhibited was in opposition to that of gravity.
FUNDAMENTALS
Electric
5
Charge
it is
Basically,
disturbances of this state of equilibrium which
give rise to the whole range of electrical (and magnetic) phenomena. The electrons in the outermost orbits of many atoms can be detached by the application of a suitable force and thereby become free. One of the simplest examples of this is the behaviour of certain insulating materials under the influence of friction. If a glass rod is rubbed smartly with a silken cloth some of the electrons
in the material are transferred to the cloth leaving the rod with a
deficiency.
The rod is then said to be positively charged and it will actually exercise appreciable mechanical forces. If, for example, it is brought near some small pieces of tissue paper it will attract them like a
magnet.
derived from the Greek meaning amber, because it had been observed that amber had this property of becoming "electrified." The mechanism of this attraction is itself based on disturbance of the equilibrium. The atoms on the surface of the rod, being deficient in electrons, exhibit an invisible force seeking to attract electrons from any neighbouring material to restore the equilibrium. This force acts on the atoms in the paper attracting electrons to the surface and if the pieces of paper are fight enough the mechanical attraction between these electrons and the positively charged (deficient) atoms on the surface of the rod is sufficient to lift the
The word
electricity, in fact, is
word
rjkeKTpov (electron),
paper to the rod.
Electromotive Force
This displacement of electrons constituting an electrio charge can be produced by other means than surface friction. In fact, except for certain specialized types of equipment, factional methods are
rarely used.
There are, in fact, several ways of producing a force which will cause an electron displacement. One is the familiar electric battery hi which chemical interaction between two different materials produces an electromotive force, or cm./. Alternatively an e.m.f. may be produced by mechanical means, as in a dynamo, which converts mechanical forces into electrical energy. Then, finally, there are arrangements which will convert heat or light into electrical
energy.
The mechanism of these various processes is discussed later. All that need be noted at the moment is that, by suitable techniques, it is possible to produce an e.m.f. which will cause the displacement of electrons which constitutes an electric charge or current.
6
RADIO COMMUNICATION
Electric Current; Conductors and Insulators
The movement of
electric current.
electrons produced
by an
e.m.f. is called
an
In the
effects so far described the
movement has
only been momentary because the freedom of the electrons is Materials of this type are known as insulators or dielectrics. There are, however, materials, mostly metallic in nature, in which the internal forces are appreciably less strong so that electrons are continually escaping from their parent atom and drift at random. They are soon captured by another atom which is deficient of an electron but there is always a quantity of stray electrons in
limited.
transit.
If an e.m.f. can be introduced into the system, however, these random electrons will all be subjected to a force tending to move them in the same direction, producing a continuous flow of electrons
through the material. In this case the current is sustained, and materials of this type are called conductors. Hence it will be seen that there are two possible kinds of current: 1. Conduction current which is a continuous movement of electrons through the material. 2. Displacement current which is a momentary movement of electrons within the material. This might appear to be of limited importance but, as will be seen later, if the direction of the e.m.f. is continually reversed the electrons can be caused to surge to and fro in the material so that a sustained oscillating current is
obtained.
In either case the intensity or strength of the current depends on the rate of flow of electrons; the more electrons passing a given point in a given time, the greater the current; and since an accumulation of electrons at rest constitutes an electric charge we can define the intensity of an electric current, whether momentary or sustained, as the rate of change of charge. Gases normally behave as insulators but can under certain conditions behave as conductors. Liquids are usually (but not necessarily) insulators except for aqueous solutions (salts dissolved in water) which conduct in a special way, to be discussed later. There are also certain materials known as semiconductors which occupy an intermediate position, behaving partly like insulators and partly like conductors. An understanding of their behaviour requires a more detailed discussion of the mechanism of conduction than is necessary at this juncture. It will suffice to note that developments in the use of these semiconductors have produced an entirely new range of techniques which are discussed in detail in Chapter 6.
FUNDAMENTALS
Units
7
may conveniently conclude this section with a brief reference to the units adopted for the measurement of the various quantities so far discussed. It might appear that the unit of charge could be the charge on a single electron but apart from the fact that this is far too small for practical purposes it is not commensurate with the units normally adopted. All engineering quantities can be expressed
in terms of the three fundamental measurements of length, mass, and time and for many years the centimetre, the gramme, and the second were used as primary units (the CGS system). From these a series of derived units can be obtained but unfortunately there were two schools of thought, one based on electrostatic phenomena and the other on electromagnetic effects. In 1950 it was agreed to bring into use a common system of units known as the MKS system. This was based on the metre, kilogramme, and second as the fundamental units, which did not, in itself, make a significant change; but by assigning certain specific values to the characteristics of free space, as explained in Appendix 2, a simplified and uniform system was developed. This is, in the main, still in use but its scope has been widened to include other than electrical quantities to form what is known as the International System of Units, and these SI units have been adopted in the present text. In the International System the fundamental electrical quantity is not the charge but the current, which can be more conveniently measured in terms of the physical forces developed. The unit of current is the ampere, which is defined as the value of a constant current which, when maintained in two straight, parallel conductors of infinite length and negligible circular crosssection, separated by a distance of one metre in vacuo, would produce between the conductors a force of 2 X 10~' newton per metre length. The unit of charge is the coulomb which is the quantity of electricity which results when a current of one ampere flows for one second. This is actually 624 x 10 18 times the charge on an electron, but this figure is not of importance in engineering, though it is of significance
in electron physics.
We
The unit of e.m.f. is the volt which is derived from the relationship between current and resistance as explained in the next section.
1.2. Ohmic Conduction; Resistance have seen that a movement of electrons constitutes an electric current, but this can take many forms. The simplest form is that which occurs in those materials in which there is a proportion of
We
8
RADIO COMMUNICATION
relatively free electrons
which can be caused to move through the material by the application of a suitable e.m.f. Such materials are known as conductors and are, in the main, metallic in nature. The atomic structure of metals (as explained in Appendix 1) is such that the outermost orbit only contains one or two electrons which in consequence are relatively loosely held. From time to time an electron may leave its parent atom, but in the absence of any directing force any such free electrons will circulate at random through the material.
In the presence of an e.m.f., however, this movement will be coordinated into a steady drift towards any point or points of positive potential, which may be the result of a battery or other source external to the material, or may be the result of potentials actually produced + within the material itself. Battery Consider the effect of a battery connected
±
to a length of wire, as in Pig. 1.1.
The
_ electrons in the wire next to the positive , „ , Fio. 1.1. Simple Electric <• . , .,, , , ,i terminal will be drawn out of the wire, CracinT
_
leaving behind them a deficiency which will be filled by some of the random free electrons and so in turn along the wire, the deficiency at the negative end being made up by electrons supplied (by chemical action) by the battery itself. There will thus be a continuous flow of current which will persist as long as the battery maintains its e.m.f. (Note that this current is an electron flow towards the point of positive potential. The accepted convention of current as flowing from positive to negative originated before the mechanism of current flow was understood.)
Resistance
The magnitude of the current depends on the material of the If an atom loses one of its outer electrons it becomes positively charged or ionized, and any free electrons are soon recaptured by an ionized atom. The pressure of an e.m.f., however,
conductor.
increases the velocity (and hence the energy) of the free electrons,
some of which may thereby overcome the attraction of the ionized atoms and continue on their journey. The number of electrons
which successfully evade recapture (and hence the current) determined by
is
The applied e.m.f., and The strength of the internal forces. The second factor is clearly dependent on the
(a)
(b)
internal structure
of the material.
It
is,
in fact, a definite
and measurable physical
FUNDAMENTALS
9
property of the material, which is called the resistivity, and a particular length of conductor or circuit is said to possess a certain
resistance.
The
length, for obviously the greater the distance
less
actual resistance of a piece of material depends upon its which the electrons
have to travel the more the chance of their recapture and hence the the current produced by a given e.m.f. It is also inversely
proportional to the crosssectional area since the larger the area the greater the number of free electrons available. This is expressed where I and are the plj mathematically by saying that length and crosssectional area and p is the resistivity, which is the resistance of a cube of the material of unit dimensions. Typical
R=
A
A
values of p are given in Table 4.1, p. 153. The reciprocal of resistivity is called the conductivity, for clearly the less the resistance the more easily will the material conduct. The resistivity is dependent upon the temperature. The atoms of the material are in a state of vibration which increases with temperature so that the chances of recapture are increased. Hence the
normally increases slightly as the temperature is raised, though there are some materials, notably carbon, in which the resistivity decreases with temperature. * Conversely, if the temperature is reduced, the resistance normally decreases and in fact with certain metals a point is reached, near to absolute zero ( 273°C) at which the resistance almost disappears and a current, once started,
resistivity
—
will continue after the removal of the e.m.f. for several days.
is
This
known
as superconductivity,
and
is
utilized in certain special
applications.
Ohm's Law
In normal circumstances the current, as we have seen, is directly proportional to the e.m.f. and inversely proportional to the resistance. This relationship is known as Ohm's Law, which states that kE\R, where I is the current, E is the e.m.f., R is the resistance, I and k is a constant. In practice the units of e.m.f. and resistance are E[R. so chosen as to make k 1, so that / The unit of resistance is the ohm, named after the scientist who first formulated the relationship. It is a practical unit, being that resistance in which a steady current of one ampere generates heat
=
=
=
energy at the rate of one joule per second
(1 watt).
(See page 12.)
* Negative temperature coefficients are frequently found in materials of the semiconductor class, wherein the mechanism of conduction is somewhat different, as explained in Chapter 6; and although carbon is not normally regarded as a semiconductor, its atomic structure is of the same form.
10
RADIO COMMUNICATION
The unit of
e.m.f., the volt, is thus the e.m.f. which will produce a current of one ampere in a resistance of one ohm. The metals used for conductors all obey this strict proportionality and are therefore said to be ohmic conductors. There are certain materials and alloys which do not behave in this way and these are known as nonlinear or nonohmic conductors.
Kirchhoff's Laws;
In the circuit form E — IB =
offset
Concept of Back E.M.F. of Fig. 1.1, / = E/B. This may be rewritten in the
by the
0, indicating that the applied e.m.f. is exactly voltage drop produced by the passage of the current I
The relationship, however, may be expressed slightly differently by suggesting that the passage of the current through the resistance has developed a back e.m.f., —IB, so that the total e.m.f. in the circuit is zero. The advantage of expressing the relationship in this way is that, as will be seen shortly, there are other forms of circuit element in which the passage of current develops a voltage drop, particularly when the current is varying; moreover there are certain forms of "active" circuit element which contain or generate e.m.f.s of their own. If any voltage drop is regarded as producing a back e.m.f., we have a consistent form of relationship which applies to any form of circuit. The relationship is known as Kirchhoff's Second Law, which states that the total e.m.f. in a circuit at any instant is zero. Kirchhoff's First Law relates to the current distribution in a circuit. Circuits are rarely as simple as that of Fig. 1.1. There is often more than one path for the current, as for example in Fig. 1.3. In such cases Kirchhoff's First Law states that at any point in a
through the resistance B.
sum of the currents flowing at any instant is zero. Thus at the point in Fig. 1.3 there will be a current I flowing towards A and two currents Ix and Ja flowing away from it. Kirchhoff's First Law says that
circuit the
A
7
so that I
I I =
x 2
7a In this simple example this may appear selfevident but with more complex circuits the answer may not be so readily apparent. It should be noted that in circuits where the current is varying it is the instantaneous values of current or e.m.f. which must be used.
X f.
=7
Series and Parallel Connection
Practical circuits nearly always contain
and
it is
necessary to
know how
more than one element, the effects of the various parts may
m. so that Ix = E/R 1 and I2 so that if B is the Ix J2 EJRZ But we have seen that I = . hence for any number of resistors in parallel the total conductance is .3. \\r. Fig.2 the resistances often required in practice to provide units having a given value of resistance.2. = effective / = EIB = E\B + E/B and B 2 together. be expended in overcoming the total circuit in proportion to the resistance of the various portions. the back e. 1. i —<ww» Fig. MW *wv> Resistances in Sebies 1 \\r.2 is a circuit containing two resistances connected end to end.f. Hence E = I(BX + B which means that B and B in series are equivalent to a X 2 ). Such an arrangement is called a resistor and the In Fig. the connecting wires possess some normally have to be taken into account.FUNDAMENTALS 11 be added together. Bt . for example. The e.1. Bx might be a resistor while B a might represent the resistance of the wiring. resistance which will for convenience.f. It  construction of typical forms of resistor is discussed in Section 4. ? Fig.f. Here it is the e. in fact. Thus if 1 ohm then an e. 1. If this IB1 and current is I. across the circuit which is the same. of 10 volts will 9 ohms and z x produce a current of 1 ampere and there will be a voltage drop of circuit resistance and will be divided across the B = B = 9 volts across Fig. is called the conductance. reciprocal of the resistance.f. 1.e. The resistance + . will. In Fig. h Resistances in Parallel however.m. on the resistances will be — — IBs respectively so that E — IB — IB 2 = 0. it must be the same throughout the circuit.m. Since there is only one path for the current.3 Bx and 1 volt across B%. This is sometimes expressed by referring to the voltage drop across the circuit. have been shown as separate units. Hence l/B = l/B1 + 1/JB2 2 X This again may be applied to any number of resistors. in aeries. 1/JB. shows two resistors in parallel. t 2 single resistance of value equal to the this same reasoning applies to is 1. sum of Bx and B 2 Moreover any number of resistances. i.2. of the two resistors . assumed to be concentrated In any circuit. 1. 1. '.m.
any Hence Power 2 of which forms may be used according to convenience. Hence the energy W oc EQ. Energy and Power We have seen that the electron drift which constitutes an electric current is not unrestricted. x Quantity per second = E.m. and the unit of power. x Current In symbols P (watts) = E (volts) X J (amps).F. It is more exhausting to run uphill than to amble up at a leisurely pace. = E. These expressions only apply when the current is steady. but have to is now used * Mechanical power was formerly expressed in different units. If the current is varying they are still fundamentally correct.M. Collisions occur between the electrons and the atoms and these absorb energy.F. which is called the watt. With only two resistors it is sometimes convenient to rewrite the above expression as follows: 1 — 1. The rate of doing work (expending energy) is called the power. usually as heat.f. and the unit of energy.12 RADIO COMMUNICATION equal to the sum of the conductances of the individual parts. (volts) x Quantity (coulombs) The effect of this energy depends upon the rate at which it is expended. See Appendix 2. it JL J?2 8 Mi _ (fl + Ri) Heating Effect of a Current. which is called the joule. In due course the atom reverts to its former condition and the surplus energy is released in the form of electromagnetic vibrations.F. But by Ohm's Law / = EIE. though it may take other forms such as light or Xrays. is so chosen as to make this an exact relationship. . is the expenditure of one joule in one second*. which is determined by the e. so that Energy (joules) = E. The energy developed by these collisions is proportional to the quantity of electrons and their velocity.M. but the watt for all power.M. a collision between an electron and an atom displaces one or more of the constituent electrons from their normal orbits. so that we can write P = EI = E /R = I*R. In simple terms.
. called the load resistance. in order to bring the . 1. The charged body is said to produce an electric field which X Y will influence any other charge in the vicinity. 13 1. EN Two Plates Eq where given by the expression F E is the intensity of the field and q is the charge. »» Consider a uniform field such as is shown in Fig.5 Concept of Load The resistance of a practical circuit is not simply that of the conductors. . The force will. be BI:IW]. . or light. . It may be a device designed to convert electrical energy into mechanical energy. such as an electric motor. Electrostatics. It constitutes —>— — — . The intensity E is sometimes called the field strength. Such a field is usually repre»sented diagrammatically as a number of lines of force. work will have to be done charge into in overcoming this force. two parts. LinesotFobceEep.° RESENTING AN ElLECTRIC FIELD m. This useful part of the circuit is called the load and since energy is expended therein it can be and usually regarded as a resistance. in fact.4. It may be a device deliberately designed to produce heat as in an electric fire or furnace. as in a radio set. The object is to pass the current through some device from which useful work can be obtained. A current / flowing in the resistance r and in the load and a (smaller) power circuit will develop a power I*r which is dissipated (mainly as heat) in the conductors. while yet again it may be an arrangement to convert one form of electrical energy into another. If a charged body charge) is (or simply an electric brought into this field there will be a force acting upon it depending on the strength of the charge and the intensity Fig.. as in an electric lamp. . 7° c .3. as explained in Section (page 48). In fact. Capacitance has been shown that the disturbance of equilibrium which an electric charge results in a force. This latter power is not useful and is called the conductor loss.j r of the field. = Potential To move the charge against the field.4. is Any practical circuit thus consists of PR 1. 1.— FUNDAMENTALS be modified to express the average power. The form of the load varies.. the conductor the load resistance R.
4 there will be a potential difference or p. can regard a charged body as being surrounded by an electroIn the case of an isolated charge this flux radiates uniformly in all directions as in Fig.e. this is called the potential at the particular point. which is the flux per unit area. Potential gradients vary widely in practice from a few microvolts/metre in an electric wave to many thousands of volts/metre in the insulation of highvoltage transformers. Permittivity What is the relationship between this flux density and the permittivity intensity? The answer depends on what is called the of the medium. Now the influence of this electric field is determined not by the amount of the flux but We by its relative concentration.5.4) is confined mainly to the space between them. The work done in moving a charge against the . work has to be done and the weight acquires some potential energy which is given back as kinetic energy if the weight is allowed to fall to the ground. i. If the distances of the points and from the source of the field are x and x dx respectively and the potentials are V and V 6V. then by definition the rise in potential 6V will be equal to the work done in moving a unit charge over the distance dx against the field which will be Edx. 1. Thus the intensity of the field at any point is equal to the potential gradient and this applies whether the field is uniform or X Y — X Y + — — = varying. If a weight is lifted against gravity. the rate of change of potential or potential gradient.d. between them.5. 1. while the flux between two bodies (such as the pair of parallel plates in Fig. The flux is The Electric regarded as being numerically equal to the Field prom a Charge Radiates Uniformly charge. the flux density (sometimes called the displacement). so that the flux density D = QjA in all Directions where A is the area over which the flux is Fig. 1. Flux Density static flux. 1. Similarly the work done in bringing a unit electric charge from infinity to any point in an electric field is called the electric potential at that point and if we consider two points and in Fig. distributed. Hence E —dV/dx which as dx —> becomes dV/dx.14 RADIO COMMUNICATION all the field at some work will have had to be done .
is the charge which must be acquired by one body to achieve a potential difference of one volt relative to the other. Inverse Square Law how The the intensity varies with distance. where E Qj^ns. 1. the effective area over which this flux is distributed will be the surface of a sphere of radius d. This may be written E E /d 2 .FUNDAMENTALS 15 field is not the same for different materials. Mica. and in its simplest form consists of two parallel plates as shown in Fig. is five times more permissive than air so that only onefifth of the energy is required to move a charge against the field. the field strength at unit distance. . any body possesses a certain capacitance but in practice we are more concerned with the capacitance between two bodies. d.6. In fact. this capacitance is a hindrance which has to be circumvented while in other applications it is useful. If the charges are Q x and Q 2 the intensity Let us consider an isolated charge. and hence the charge. D/e and if we call the permittivity e. for example. for which purpose special arrangements are made to obtain the maximum capacitance in the space available. If a source of . Capacitance From what has been between electric charge said it is clear that there is a connection and potential. This is known as the inverse square law. by definition. are pro = D= portional to the permittivity. In some circumstances. of the field produced = = = . which. E Since the intensity E is equal to the voltage gradient this means that for a given voltage the flux. Consider total flux will be distributed uniformly through the surrounding space and at a distance. any variation in the amount of charge is accompanied by a change of potential. which is 4iTd 2 Hence the flux density will be QI4nd* and the intensity or field strength Ql^ned*. Such a device is called a capacitor or in older parlance. The produced by this field on the charge Q 2 is EQ 2 force Q1Q2 l4med i Hence the force is dependent on the product of the two charges and is inversely proportional to the square of the distance between them. Hence for a given flux density the intensity is dependent on the permittivity so that Ee. We can now consider the force developed between two charges separated by a distance d. which is the Theoretically charge (in coulombs) necessary to raise the potential by one volt. This relationship is called the capacitance of the system. some of which permit the flux to be established more readily. a condenser. — — D/e by Qx will be E QJ4jrd 2 e.
It should be noted.. In practice the capacitor would be discharged in time by leakage currents which would flow through the dielectric and even the surrounding air (particularly if the air were moist) since no material is a perfect insulator. and V the applied (or acquired) p.f. 1.m. Conversely if we have two plates as in Fig. producing charges of Q and Q on the opposite faces.16 e. by the application of a suitable e. which will become similarly charged.m.6 and we introduce between them a thin sheet of metal. there will be a momentary flow of electrons resulting in an accumulation at the (negative) plate and a deficiency at the other. If. is connected to the plates. however. as in a practical case.. RADIO COMMUNICATION such as a battery. that the action of a capacitor does not require the presence of a dielectric. The presence of a dielectric merely modifies the amount of charge produced. = . 1. Parallel Plate Capacitor established between the plates which will allow the accumulated electrons to return to their normal condition. so that the arrangement becomes electrically charged and there will be an electric field between the plates. there is some insulation or dielectric between the plates the electric field will cause a corresponding displacement of the electrons in the dielectric.d. of one volt we can write G Q/V where is the capacitance. = D + — Capacitance Relationships The capacitance of a parallelplate capacitor is readily calculable. an electric field is set up between them. and the capacitor will remain "charged" until some conducting path is This condition Metal plaits O/e/ectr/c Fig. the electrons in this sheet will be displaced by the influence of the electric field.6. Q is the charge. where Q Da. Provided the plates become charged.d. will persist even if the battery is removed. being the flux density of the field and a the area of the (inserted) plate. Since by definition the capacitance is the charge acquired for a p.f.
1. that V = Hence C = QfV = AEe\Ed = Aejd The undetermined factor here is the permittivity e. sometimes called micromicrofarads 1 1 (fiftF) microfarad (/iF) picofarad (pF) = = 10~ farad = 10~* microfarad 10 _e farad 12 Practical Forms of Capacitor is insufficient. across the plates is d times the voltage gradient. For most requirements the capacitance of a single pair of plates For example. So we can = AD = AEe Again assuming a uniform field. sometimes called the electric space. Hence e e er where e is the permittivity of free space. As already stated.10. the p.d. to use assemblies of plates connected . two plates 10 cm x 10 cm would have apart in air the capacitance they were 1 an area of 102 m2 would be . where d is the distance between them. so Ed. The presence of any material between the plates will simply modify the amount of charge to an extent depending upon the relative permittivity of the material. constant. while for many purposes even this is too large and the capacitance is expressed in picofarads so we can parallel plates in the = (pF). and er is the relative permittivity of the dielectric used. which is itself made up of two factors. In SI units the value of e is = 1/(36tt X 10 9 ) = 8864 X 10~ 12 write the expression for the capacitance between two form C (8854 Aer /d) x 10 12 farads.FUNDAMENTALS density write 17 Now assuming a uniform electric field between the plates the flux D will be Q\A where A Q is the area of the plates. In practice this unit is too large and capacitances are specified in microfarads for which the symbol /iF is used. H mm O = 5^^. the capacitor of Fig.) It is customary.6 does not require any material (even air) between the plates.= 885 pF only (In fact it would be slightly less because there is some dispersion of the electric flux at the edges of the plates so that the flux density is not quite uniform. therefore.
f.f. by Kirchhoff 's Law. which thus becomes and less as the charge builds up.m. The relative permittivity was formerly known by various other names. When the e. will produce a charge which will generate a back A e. is determined experimentally and is known for a wide range of materials. For many purposes variable capacitors are required. that E — iR— q\Q = where E is the applied e. But current = we can write rate of change of charge. Q/G is equal and opposite to the applied e. is first applied the capacitor is not charged and therefore offers no back e. limited only by the resistance of the connecting leads (or any other resistance in the circuit). of Charge have seen that if a capacitor is connected across a source of there will be a displacement of the electrons in the dielectric will and the capacitor will adjust itself so that become charged. V. is not an instantaneous process. is discussed in Section 4. Time is always required for any circuit to readjust itself to changed conditions. — Relative Permittivity The value of the relative permittivity e.m.18 RADIO COMMUNICATION together alternately. The amount of the charge the back e. however. current will therefore flow. until the condition of equilibrium reached such that Q = CV.f.m. The relative permittivity of some other materials in common use is given in Chapter 4. This current.2. notably specific inductive capacity or dielectric constant.f. is the number of plates. It is found that the capacitance with air dielectric is not appreciably different from the capacitance in a vacuum so that er for air is taken as unity.m. Hence i = dq/dt so that E— Rdqjdt — qjC = . Such an arrangement is equivalent to (n 1) capacitors in parallel.f.m. however. These may be constructed by assemblies of plates so arranged that either the area or the separation can be varied.m. both fixed and variable. i and q are the instantaneous current and charge. where n.m.f. Let us consider what happens more closely. The construction of practical forms of capacitor. Growth We e. q\C. At any instant we can say..f. and R and C are the resistance and capacitance. less is This will begin to limit the current. This.
The voltage across the capacitor obeys a similar law since v q\C. — = — — = = .7. In other words its rate of change is exactly equal to its value at any instant. . The term x is called the exponent. . There are many physical processes which behave in just this way so that mathematical calculations frequently involve this fundamental value which is called the exponential constant. to the capacitor t 0. t/CR 1. but there is one particular value of a for which the derivative xa x l is equal to a". is the exponential* constant (27183) and t is the time Let us see what this means in physical terms. As 1) (1 the time increases the term &~ llCR which is l/e' /CB becomes smaller and smaller until finally it becomes negligible and q simply CE. At the instant of connecting the e. v £ Fia.. It is not an exact number. . When t CR.f. In other words the 1/27183) capacitor has acquired 6321 per cent of its full charge. .e"Cii = ) Time Constant How long does this operation take ? The answer depends on the value of the product CR. When t 2CR the charge is 865 per cent of the final value and if we plot — — = = = = 100 t. * If a quantity a is to be multiplied by itself the process is expressed in the form a 2 and the expression a" means a multiplied by itself x times. The capacitor is then fully charged. = = Hence v = E(l . being actually 27183 and is written as e.7. ^ JC/t /^ Discharge CR 2CR 4CR SCR Exponential Growth and Decay charge against time we obtain the curve shown in Fig.m. Hence q 0. so that the term in the brackets becomes (1 e°) 0. and q CE(\ 1/e) =CE(l 06321 CE. This is known as an exponential curve and this type of variation is frequently encountered in electrical engineering. . Sometimes (but not in this book) e is used.FUNDAMENTALS This is 19 is a differential equation of which the solution q = GE{1 — e«CJJ ) where e elapsing.. . 1. Now if we differentiate ax the result is xa"' 1 This may be more or less than ~ a". BO \ Zharge b 60 f\ i k s 20 O 1.
In this case v of its charge in time t = = . E/R. Charging Current The current which flows during the charging (or discharging) process is determined by the rate of change of charge. For example with and 1 MD. losing 63 per cent E e>tlcR CR. in raising the potential of a capacitor is Hence the work done not . But as we have just seen. The energy input into the capacitor may therefore be taken as the total charge Q multiplied by the average p. the charge in a capacitor due to the application of an e. where V is the final CV. \CV* since Q voltage. = = Energy Stored in a Capacitor Since work is done in raising the potential of a capacitor it follows that. during the period. the capacitor would take nearly 0=1 R— GR = 5 seconds to acquire 99 per cent full charge. Similar conditions obviously obtain when a capacitor is discharged. and then falls off exponentially as shown by the dotted curve in Fig.e* = 1 i *^ The constitutes distortion of the atomic structure in a dielectric which an electric charge is not a perfectly elastic process. as long as the capacitor remains charged. this means that initially (when t 0). If we differentiate the expression for q just t deduced we find that t' = {EIR)e CR i In physical terms again. The work dw done in bringing an elementary charge 6q up to a potential v is vdq. The charge does not immediately disappear but decays exponentially in inverse manner to the charging curve of Fig. across the capacitor gradually builds up to the full value.20 RADIO COMMUNICATION The product GR is called the time constant of the circuit.7.f. Hence if Q is the final charge on the capacitor the total work done is A W—G " Dielectric Loss 4^ 1 ro f° _ = ^. energy is stored in it. which as we have just seen is the time required for the capacitor to acquire 6321 per cent of its full charge. 1.m. If the time constant is small the capacitor will charge up virtually immediately. 1. is not immediately established but requires a certain (small) time during which the p.7.d. Hence the energy stored is %QV = = — more rigorous derivation of this expression is as follows. but with a larger (iF value appreciable time is involved. = dqjdt. making 1.d. \V. But v = q\G so that dw = qdq/G.
1.FUNDAMENTALS difference 21 completely given out again when the capacitor is discharged.C#.d. one where the capacitors are in series as shown in Fig. we have i. Capacitances in Series and Parallel together. = Q\C + QIC I/O. and when a capacitor is being used under conditions where the charge is continually varying (i.8 (6). . Hence. The is called the dielectric loss. the charge Q) will clearly be the same whether we consider the whole system or the tors.8 (a) and the other where ^l^^^ tfUJ (a) Fig.e. = C. so that Vt=V + v QIC. while combinations of both are possible. = 1/Oj + l/Cjj 1 2 X 2 This reasoning may be applied to any number of capacitors in series so that the reciprocal of the total capacitance is the sum of the reciprocals of the individual capacitances. will be divided across the capaciin inverse proportion to their capacitance. but the p. For the series case the quantity of electrons displaced by the application of a p. as in Fig. p.d. the capacitor is passing a current) the effects of this dielectric loss may be appreciable (See Chapter 4. individual capacitors.e. 1. since C = QV.e. across the system (i. The above expression may be rearranged in the form a) O. Capacitors in Semes and Parallel they are all connected in parallel across the same p. It will be noted that the series capacitance is always less than either or any of the individual capacitances.d. 1.8. + C which is sometimes more convenient. 167). Occasions often arise where two or more capacitors are connected There are two possible arrangements.
9.22 RADIO COMMUNICATION In the parallel case the p. and = the electric space constant. . 1.e. For the solid portion we have: Cx where = Ae^\A x A= er the area of the plate. the total capacitance being the sum of the individual capacitances. = the relative permittivity of the material. across the capacitors is the same but the total charge is the sum of the charges on the individual capacitors. Hence i.d. sometimes necessary to allow for naturally existing capaciWe may therefore conclude this section with a discussion of some of the more common forms. ^ Fig. Nonuniform Dielectrics The capacitance of a system having a mixed dielectric (e. C2 dt e For the airspaced portion — AeqIu^ whence Special Configurations It is tances. C V t so that G this t = GJ + C = C\ + C2 2 F Again can be applied to any number of capacitors. Mixed Dielectric ^ Fig. the distances between the plates being the thickness of the respective dielectrics. = thickness of the material. ebonite and air) may be obtained by considering the system as made up of two capacitors in series. Capacitor with 1.9 illustrates a mixed dielectric.g.
Hence . Hence the flux T II density will be Q\2ttx and the intensity // of the field will be Q\2ttxb. Fla 1 10 PABALLEL (since we must measure our distance from "w ir°s the same point). Similarly the other wire carrying a charge of Q will produce an intensity —Q\2tte[x d). QI27re(d The total intensity at x is the sum of these two and this is the potential gradient at the point. large compared with r. \x d 1 _ dv — xf — dx x] \ . Under these conditions the charges on the wires may be assumed to be Fig. * 10 3 fiP per metre n is. D — — '  = — E— Q 2tts /l (\ . this simplifies to = (Q/««)loge(*/r The capacitance G Q/V. ..e. For air spacing s which is 1/367T X 109 and substituting this in the e expression for C we have = = = .10 represents uniformly distributed. an area of 2irx. This is a maximum when x <L(\ ftre \r = for which the expression becomes ' _±_\ = JL + d^r) = 2^ Veiy nearly . i. potential gradient at a distance x as stated above. This potential gradient is not constant so that the voltage between the two wires must be obtained by integrating dvjdx over the range x = r to x = d — r. C The = tt 36 loge d/r ^tt.e. = &I and since r is small + dhc) dx = (x F d £e l0g < —r compared with d. i. Let the total charge \\ f ^ per unit length be Q. ± 2ire ( \x L xj +— A d 1 r. At a distance x this * j6x will produce a flux operating through the surface of a cylinder of unit length and ()»—X *»»!»CX>< radius*. whence C 7re/log e (dlr). .FUNDAMENTALS 23 Two Paballbl Webbs by a two long parallel wires of radius r separated distance d. 1. x).
But from the expression V above. Inductance We have seen that the disturbance of equilibrium in a dielectric which results in an electric charge gives rise to an electric field.11 represents two coaxial cylinders. By similar reasoning to the above it can be shown that at a radius x the intensity is E = Q\2ttsx = whence dv/dx V = „— = Q 2lTE I .4. Electromagnetism . log {r /r) = QfC .24 since l/(d for RAD/O COMMUNICATION — r) is small compared with 1/r.11. . Concentric Cylinders V — 2r loge (d ry — r)lr volts/metre ' Fig. — j • 10. of radii rt and r2 respectively. per metre ^ The potential gradient at a radius x Substituting for Q this becomes * This is as stated above. such as would be found in a coaxial transmission line. r^ is a maximum when x =r where the potential gradient volts per metre loge rt lrt 1. . c hence the expression becomes C = £r . 1. Q 2ne V 2 log e (d — r)\r Hence the maximum potential gradient (which occurs at the surface of the wires) is —. Q\2nex. 1.3 18log e r2 l ri uF * is.dx Two ConCENTRIC CyLINDEBS Fig. Concentric conductors writing e =ss r may well have a solid and substituting l/36w X 10 9 for dielectric. loge x.
13. one 25 in the system while this condition persists. as in Fig. It does not radiate from the wires like an electric field but surrounds the wire as a sort of sheath of gradually decreasing intensity as the distance from the wire increases. If we consider a crosssection of this sheath in a plane normal to exists between Two wires L/'nes of force >. electrons in motion do produce a force similar in some respects to that which electric charges. as shown in Fig. 1. This is known as a magnetic field. the current in any segment of the turn producing a field at rightangles. In practice we are usually more concerned with a complete circuit rather than an isolated portion and the simplest form of circuit we can shown in Fig. Magnetic Field around a Wire the wire Magnetic Field Produced by a Single Turn of Wire Pig. 1. represent the field by a series of lines of force.13.12. may expect to find some similar manifestation and.• Fig. the direction of the field being at right angles to the direction of the current. would be a single turn of wire. as It will be seen that they form a series of concentric circles. If the wire takes the form of a number of turns in series the magnetic fields of the individual turns reinforce one another so that .FUNDAMENTALS and that energy is stored If the electrons are free to move. 1.12. as they are in a conductor. in fact. 1. carrying an electric current exercise a definite force on each other which is evidence of the existence of some effect due to the current. At various points round this turn we can draw the lines of force and it will be seen that they combine to give a force acting along the axis of the turn. Using the conventional direction of current flow as being from positive to negative (which we have seen is actually the opposite direction to the electron flow) the direction of the field is clockwise when the current flow is away from one. 1.12. It will also be seen that the clockwise rule still applies.
which will be defined shortly. 1.m. 1.m.m. indeed. the product of these two.f.14. the m. which magnetic field are said to constitute a magnetic flux produced by a magnetomotive force (m. illustrating distribution of magnetic flbld abound an Inductor The Magnetic There fields. to the current in the coil It is. in fact. defined as (/) Now. is an important difference between electric and magnetic namely that the magnetic field is always a complete loop. Circuit.14. is shown in Fig.) analogous to the e. which produces current in an electrical circuit. a magnetic circuit similar to a conducting circuit. =a quantity analogous to resistance. There is. . so that M — IN. as is clear from the various diagrams. called the reluctance.m..f. called ductance coil or inductor. It forms. lines of is The namely <& = M/S where O M = the m. in fact.f. and in many respects behaves similarly.F. and S = the flux. M. is clearly proportional and also to the number of turns (N).26 RADIO COMMUNICATION field is increased. the total an in Vooofofa^ /' / /6o6Q&p0boq£i^^ \ flo.M. a relationship similar to Ohm's Law in a conducting circuit. Such an arrangement.f.
f. In certain materials. 27 Magnetic Path For many purposes. Materials which have a greater permeability than air are called 'paramagnetic though this term is usually applied only to materials in which the increase is relatively small. gradient or field strength. the magnetic field can be concentrated within certain limits (as in the toroid of Fig. operating in a material of given permeability will produce a certain total flux. gradient at any point is the change of m. Units A given m. thus increasing the effective flux. It is found. It is not always easy to decide what the length of the magnetic path is. 1.m. an external m.m. Hence H = I\1nr. Strictly speaking the m.m. The m. 1.f. The concentration of .16) so that the path length can be defined. Flux Density. In any material the electrons revolving in their orbits produce heir own magnetic fields but these have a purely random orientation. the field strength at a distance r from the single wire of Fig. however.FUNDAMENTALS Field Strength .14 is provided with an iron core the flux produced by a given current is increased many hundreds of times.e. gradient is assumed to be uniform we can say simply that M\l IN \l where I is the total length of the magnetic path. There are also some materials which are called diamagnetic. but if the m.m. namely the relative 'permeability of the material. i.m. The substance which exhibits the effect most strongly is bismuth which has a relative permeability of 09998.f. is simply /. For example. Materials which exhibit considerably increased relative permeability are called ferromagnetic (though as explained in Chapter 4 some do not actually include iron in their composition).m.12 can be calculated without difficulty. therefore. however. we are more concerned with the m. for example. r Permeability There is. as will be seen shortly. having less permeability than air but the difference is only slight.f. can produce an alignment of some of these orbits. Otherwise it is necessary to state clearly what path is being considered. over a small length of path at that point..f.f. that if a coil such as that shown in Fig. since the wire is part of a single turn so that H H= H— — = N= 1. another factor involved in determining the total magnetic flux produced. dMjdx.f. Sometimes.m. for at a radius r the magnetic path is clearly the distance round the circle of this radius 2tjt. 1. for which the symbol is used.
which is 1 weber/metre 2 The permeability.m. of course. The flux density is measured in terms of the number of lines of force per unit area. The value of fi in SI units is 4tt/10' so that B= B= . and as has been stated this is not always easy to define.f. and the unit of flux density is the tesla (T).m.F.m. is itself the product of the permeability of free space. To evaluate the reluctance in any particular instance it is necessary to know the length and area of the magnetic circuit. however. In the old CGS system the unit chosen was 1 line/cm 2 which was called the gauss. So far it has been assumed that the current in the wire or coil.f. p . Induced E. It is discussed further in Chapter 4. But HI and reluctance (equivalent to resistance in a conducting circuit) is directly proportional to the length of the magnetic path and inversely = M= O— $= = = proportional to the area and the permeability. B = )j. m./reluctance i. This is an important principle. or the magnetic space constant. . is called an induced e. and the relative permeability (ir which for air is 1. It was stated that the flux Jf/<E>.f. ft. . in his classic research into electromagnetism. which in magnitude and direction is such as to try to preserve the original conditions. as in the electrostatic case. in fact.H = iirfirH Reluctance . defined by the relationship [iH. so that the flux density 0/^4. The reciprocal of the reluctance (Apll) is called the permeance. is proportional to the field strength and is.e. is of constant strength. The flux density. the only back e. If the current changes.f. to which reference was made earlier.28 RADIO COMMUNICATION this flux will depend on the area over which it is distributed. so that S Hence the HA/x so that S l/Aft. an imaginary concept but a greater field strength will obviously produce a greater concentration of lines of force. known as Lenz's Law. Faraday. MfS. in the circuit being that due to the resistance.f. These lines of force are. where p. is the permeability.m. this equilibrium is disturbed and forces are brought into play which resist any change. 10~ 8 teslas We are now able to define the reluctance of a magnetic circuit.M. analogous to the conductance in an electric circuit. and hence the magnetic field. The SI unit of flux is the weber (Wb) which is 10 8 CGS lines of force. discovered . be ignored for the present. The force takes the form of an additional e. This is an equilibrium condition. and the e.m. 10' = 1257 ^(IN/l) . however. This may.
1. In Fig.m. i. but it is clear that this is not necessarily the case.f. = length of flux path. If all the flux linked with every turn the linkage would be simply NQ>. Hence e = kNdQfdt = kN(pANfl) difdt = L difdt where i is the instantaneous current. though with suitable arrangements it may be made very close to unity. the manner in which the flux interacts with the turns of the coil. may be written as rate of change of current. the flux linkage per unit current.f.14.f. or In a coil electrically as in an oscillating circuit. The expression JfcJVO/7. which is denned as the inductance in which an e. so that if I and L= 1257 kfrAN*/! X 10« henrys . of selfinduction which is dependent upon the fhtx linkage. where k is a constant which depends on the configuration of the circuit. and hence it is called the e.m.FUNDAMENTALS flux 29 that the induced e. We where have already seen that = [iANI\l. N Inductance The linkage kN<& may be written (kN<PII)I and hence the induced e.m. so that in a practical case the linkage is kNQ>.e.m. of one volt is induced when the current changes at a rate of one ampere per second. 1.) A Since pt we can substitute 47r/10 7 for firfi are in metres and square metres.e.f. I /i The e. such as that of Fig. A = area of flux path.f. e which is = (kN<t>II) which the rate of change of linkage. = permeability.AN*ll <b N = number of turns. It is always less than unity. where is the number of turns. = /*„. will clearly depend on the influence of the flux changes on the turns of the coil itself. was proportional to the rate at which the was varied. The unit of inductance is the henry.m. e oc d<&/dt and this is is a fundamental relationship which applies whether the flux varied mechanically as in a dynamo. and k = linkage factor.m.14 the e. for example. so that the inductance L = kN<&/I = kfjt.f. i. it is only the turns in the middle portion of the winding which are linked with the total flux. of selfinduction is the rate of change of linkage.. (Note that when dealing with varying quantities it is customary to use small letters to represent the instantaneous values whereas capital letters are used for steady or average values. is called the inductance is x of the coil.
14 is called a solenoid and it will be seen that its inductance is considerably influenced by its shape. I The dimensions are in inches few per cent within the limits IjD and the formula is valid within a = 2 and djD — 03. = axial length of winding. D = outside diameter of coil. The variation of this factor with relative dimensions is shown in Fig 1. and a further correction factor has to be introduced dependent on the ratio of radial thickness to mean diameter. Valot: of Linkage Factor fob Single layer Coils If the coil is wound with more than one layer so that it has appreciable radial thickness the flux linkage is again reduced. 1.30 Cylindrical Coils RADIO COMMUNICATION A coil of the form of Fig. Length/ Diameter Fig. With a coil of short axial length the linkage factor k may approach unity. are readily available today. Curves or abacs for calculating inductance. and d = radial thickness of winding. Rectangular Coils Inductances are sometimes wound on rectangular formers.15. In such cases the inductance may be calculated to a sufficient accuracy .15. 1. which allow for these various factors. D microhenrys . but as the length increases k becomes progressively less. For a quick approximation the following empirical formula be used may n2 02 D 2 35D where + 81 X D225d .
and is normally quite small (a few pF only). Its influence is only of importance at frequencies where its reactance (see page 44) becomes comparable with that of the inductance. Toroidal Con. 1. and between the terminal connections with the material of the coil former as dielectric. being nd. there is also some capacitance between the various turns. so that there is a "stray" field. reduces to 04 firAN^fd microhenrys (where d Q in metres and A in square metres). but is obviously considerably larger with a multilayer coil. entirely confined within the winding. no coil is a pure inductance. minimum of often important. This is known as the distributed or selfcapacitance. The ring may be of insulating The inductance of a toroid is = = . which.16. Such a coil is called a foroid (Fig.FUNDAMENTALS for practical purposes drical coil 31 by regarding the coil as equivalent to a cylinhaving a diameter equal to the mean of the inscribed and escribed circles. since (jL ixr 4tniJLT IW. where d is the mean diameter is fi is thus simply ixAN 2 fird. Toroidal Coils have seen that the magnetic flux has to form a complete If the turns of a coil could be arranged to enclose the flux throughout. It is not readily calculable. as discussed later. while the escribed circle diameter is y/{a? b 2 ). Thus with a former having sides a and b (a being the larger) the inscribed circle diameter is b. which is magnetic path of the ring. 1. It forms part of the unavoidable stray capacitances which are present in any practical circuit. Apart from the fact that the winding necessarily possesses resistance. The equivalent circular diameter is then £IV(a 2 b2) &]• + + + SelfCapacitance In practice. We Fig.16) and has the merit that the flux is almost circuit. the linkage would be virtually perfect. while the length of the definite.
Mutual Inductance If the flux produced by a coil is allowed to link with the turns of another coil as in Fig.17.m. and also a flux due to the mutual inductance from the other coil. The total linkage is M therefore ^series = (Lt + M) + (£a + M = L + L + 2M ) x 2 . In symbols e2 = M dijdt This possibility of inducing e.17 any variation in the flux will induce an e.m.f. If the coils are connected in series. in the second coil as well as the first. This is called an e. each coil will have associated with it a flux due to its selfinductance. this is an important part of practical communication engineering and is discussed more fully in Chapter 4.f. The alternative method of confining the flux is to provide a path for the flux made of iron or other paramagnetic material.m.32 RADIO COMMUNICATION material or of some magnetic material. _ fir>r>rir* " JVJ~T ** — _ Fig. of one volt is induced in the secondary if the current in the primary changes at a rate of one ampere per second. which is the mutual inductance for which an e. according to requirements. their combined inductance will be the total linkage produced by unit current flowing through the combination. Illustrating Mutual Inductance the coils.m. 1. in an entirely separate circuit of great value in practical engineering.f. but toroids are difficult to wind and are only used for special purposes. is Inductances in Series and Parallel If two inductances L t and L% are connected together. Two coils so arranged that the flux from one links with the other are called a transformer. of mutual induction and there is said to be mutual inductance between if\r>ft . The unit of mutual inductance is also the henry.f. 1. which is discussed more fully in Chapters 2 and 4.
Similarly if the e.f.f. As in the case of a capacitor. If the coils are in opposition — must be substituted for M.m. When the current begins to flow a back e. while a third method is to use two coils in series or parallel and to vary the mutual inductance between them. the term e~ miL becomes smaller. unity and i 0. if the which = . time is required for the circuit to adjust itself to the new conditions. 1.f. M M Variable Inductors The inductance of a coil (or circuit) may be varied by varying the number of turns or the permeability. is removed the current does not immediately cease but decays exponentially. It may be noted here that.e. are discussed in Section 4.m.m.L + L 2M t 2 1 i The above expressions assume that the coils are wound in the same direction so that is positive. At the instant of switching on. is applied across established an Inductive Circuit an inductor the magnetic field is not immediately.f.3. At any instant the current is given by the relationship E — iR — L dijdt — This is a differential equation of which the solution i is = (E/R)(l .m L l ) is an exponential relationship similar to that of Fig. until when the current finally reaches its steady value there is no inductive back e. Current If in an e. is almost entirely occupied in overcoming the back e. Both methods are used. of selfinduction. in fact.m. the rate of change of current is very large and the applied e.m. each coil will only carry a portion of the total current.JH . Practical forms of inductor.FUNDAMENTALS 33 If the coils are connected in parallel.m. and the current is simply limited by the resistance. As the current begins to flow the rate of change decreases. The appropriate proportion can be calculated and hence the combined linkage can be deduced as JjwlMel L L . The term ea" i is. will be produced proportional to the rate of change of current.f.f. both fixed and variablo. This will limit the current which will thus build up to its final steady value in an exponential fashion similar to the growth of charge in a capacitor.7.
the attempt to maintain the current flow during this decay period will result in a very high momentary induced e. causing breakdown of the insulation. This is because of the interaction of the lines of force. Such permanent magnets are used to a considerable extent in electrical engineering practice and small magnetized rods are frequently employed to demonstrate some of the effects of magnetic fields.f. while if they are in opposite directions they will repel. . Now over a small element of time dt the energy is eq eidt and e L di/dt. they will try to locate themselves so that each lies along the direction of the magnetic field at that point.m. Hence if the final steady current is I.m. the current i == (E/R)(l 1/e) = — = Energy Stored in an Inductor it will be clear that work has to be done in establishing a magnetic field. exhibit different polarities. the total energy involved in establishing the field (and so stored in the inductor while the current continues to flow) is = = W= C1 di Li Jo at T dt = tLP to the increased permeability of ferromagnetic materials due to the internal realignment of the molecules under the influence of an m. in fact. I L\R. such as steel.m. when 06321 E\R.f. 1. has been removed and the material is then said to be permanently magnetized.f. as explained in the next section. With certain materials. and hence. Polarity Reference has already been made N poles attract. and. If the magnets are not directly in line.18. as with electric charges. which may reach a dangerous value. if one magnet is brought near to another magnet it will be found that one end will be attracted and the other repelled. this realignment persists after the m. During From what has just been said the process of establishing the field the current is varying exponentially and the total energy will thus be the sum of the instantaneous energies during the period. usually referred to as and S respectively. For example. energy will be stored in the magnetic field. as shown in Fig.34 circuit is RADIO COMMUNICATION broken so that R becomes infinite. like poles repel and unlike Permanent Magnets. If these are in the same direction they will combine. The term LjB is called the time constant and is the time taken for the current to reach 63 per cent of its final value since. in an inductor carrying current. The two ends of a magnet.
as it is called. however. are not isolated like electric charges though sometimes the arbitrary concept of a unit magnetic pole is employed. The field is not parallel to the surface of the earth except near the equator. The actual axis is displaced some 12° from the spin axis but varies slightly from year to year. As the latitude increases the field becomes inclined at an increasingly steep angle. Hence the poles of a magnet are termed and S respectively. which is true '<' ""\N ^ — ^ — _^ \ \ . of course. Bab Magnets so long as the field is not distorted. by reasoning similar to that used in the electrostatic case.FUNDAMENTALS 35 The poles. Actually the earth's field is relatively small (~ 14 A/m at European latitudes). Compass needles have therefore to be provided with a small weight on one pole to counteract the magnetic dip.18. The effect is thought to be due to circulating currents in the liquid core of the earth but there are many subsidiary influences which also contribute. are mainly of academic interest and are not employed in practical engineering. 1. Many hundreds of years ago mariners were assisted in their navigation by a property possessed by pieces of naturally magnetized iron ore known as lodestones which. Prom such a concept it can be shown. ' V_ ^v ^'' / J Fig. if freely suspended. Earth's Magnetic Field It has long been known that the earth itself is a huge permanent magnet. These concepts. It is also possible to introduce the idea of magnetic potential as the work done in bringing a unit magnetic pole from infinity to the particular point in the field (analogous to the definition of electric potential). It behaves as if produced by a magnet located approximately on the earth's axis. would take up a position with one end pointing North. that the force exerted by a magnet falls off inversely as the square of the distance away. which in European latitudes is some 67°. N .
The rule. This is a perspective Force Magnetic Current field \ k^2£ Fig. first memorized by using the lefthand and second fingers of the left hand are Motion FMd Currant Motor Fig. direction of motion can be If the thumb.20.20.19. Consider a wire carrying a current situated in a uniform magnetic field as shown in Fig. Force on a Conductor view representing a horizontal magnetic field with a wire. where B is the flux density. where 6 is the angle between the conductor and the magnetic field.) = . its effective length becomes I sin d. 1. Generofor Left. and at right angles to. 1. the field. then if the first and second fingers represent the direction of the field and the current respectively the thumb will indicate the direction of motion. (If the conductor is not at right angles.36 RADIO COMMUNICATION Force on a Conductor An important aspect of the magnetic field is the force exerted on a conductor.19. 1. at right angles to the directions of the field and the current. The wire will experience a mechanical force tending to move it out of the field. The force produced is given by F BIl. also horizontal but at right angles to the flux. as shown in Fig. 1.and Righthand Rules arranged at right angles. / is the current and I is the effective length of conductor within. carrying a current /. With the field and current in the directions shown the force will be vertically upwards.
f.5.12). The wire only possesses inductance as part of a circuit.m. The total flux will be the sum of all these elemental fluxes over the appropriate range of values for x. in which case the force produced is number of turns in the coil. (The inductance of an isolated wire. Consider the field existing at a distance x from the centre of a wire carrying a current I (Fig. the rate of change of flux which.f. The lower limit is clearly the radius of the wire itself. as we have seen. in fact. generated.f. is indicated by the second finger (Fig.) = = . being e = Circuit Inductance Although inductance is usually deliberately contrived in a concentrated form by winding turns of wire into a coil. if the wire is moved mechanically across (i.m. an e. the basic principle of the electric generator. for unit length. Where the current is varying very rapidly the effect of the inductance of the circuit wiring itself may be appreciable. In this case. 1.e. i. The field strength B nl\1itx per unit length so the flux [xldxj2mc. has no meaning. when the direction of the induced e. This expression may be deduced as follows. will be dx. will be induced tending to oppose the movement. r. In one second the conductor will move v metres across the field. will therefore be in the opposite direction to the current which would produce the same motion in a motor.m. This is the change of flux in one second. with the thumb and first finger representing the direction of motion and field. practical forms of which are discussed later. while the upper limit must be the distance d from the nearest return part of the circuit. is This e.f.e. is the e. it must be remembered that any closed circuit possesses inductance. From increased times.m.m. the crosssectional area of which. 1.20). hence the area swept out will be Iv and the flux enclosed during the period will be Blv. at right angles to) the field.FUNDAMENTALS Motors and Generators The existence of this force 37 is the basis of all electric motors and indicating meters. This field will surround the wire in a concentric sheath of elemental thickness dx.f. This e. where is the this the torque and power can N N be determined as explained in Chapter 4. therefore. Conversely. It is usual to construct the moving element in the form of a coil. the righthand rule should be used. the induced Blv volts where v is the velocity of motion in metres/sec and B and I are the flux density and length of the conductor as before.
Hence when a lead is required to be of low inductance it is best made of large diameter (or by using several wires symmetrically disposed round the circumference of a circle). 1. = 02 loge djr microhenrys/metre This is an expression which is easily memorized.f. X dx = (fil2n) loge d\r = U = 47r/10 so that L — 2 loge djr 10~' henrys/metre 7 . is the flux per unit current. CC of a singleturn which is . Note that the larger the radius of the wire the less the inductance. a slight correcting term which takes account of the flux inside the wire itself but this can usually be neglected. and hence the inductance which.s are set up according Fig. so that L = 02 loge Rjr microhenrys/metre 1.m. radio communication is largely concerned with currents which are not steady but flow first in one direction and then in the other. The inductance of the pair is therefore twice as great.5. hy definition. Two Parallel Wires For transmission lines consisting of two parallel wires the above expression gives the inductance per conductor. Such "alternating" currents. There is.11. 1. In this case d is the radius of the outer conductor. can be produced in various ways. . so that L In air /* = rd JT i (^/2tt)  . as they are called. in fact. As the coil is rotated. the same expression applies. so that L= Coaxial Cable 04 loge djr microhenrys/metre For a coaxial cable. Alternating Currents As has been stated. magnetic is the crosssection rotated at a steady speed in a uniform magnetic field. Due to this changing linkage.21 illustrates such coil an arrangement.38 RADIO COMMUNICATION With this understanding we can evaluate the total flux. one of which is to rotate a coil in a field. the flux linked with the coil changes. e. as in Fig.
the number of lines cut in a given time grows less and less until it reaches the position again and the e. 1.f. the rate at which it cuts the magnetic lines of force. Vectors will The manner in which the e.m. As the coil continues to rotate. is zero.f.22. it begins to cut more and more lines of force until at the points it is moving across the field at the maximum rate and a maximum e.m.f. Consider the e. When 6 90°. proportional to the rate of change of the linkage. At the points the coil is not moving across the field at all. E.m. the instantaneous e. Cross section at the point a maximum e. is induced.m.f. Illustrating Production of Alternating Current field. and Fig. induced moves across the = AN = =E . i. The rate of change of linkage depends on the rate at which the coil moves across the field.m.m.22 OA represents the position of the conductor in the p g s s r_ o. this time.e.m.m. so that if we say that OA represents the maximum e.f. From simple geometry this can be shown to be proportional to the ordinate AN.21.FUNDAMENTALS 39 to the laws of magnetic induction. 1. however. e.f.f. OA. is again or Simple Alternator induced in the coil. i. the component of the velocity at right angles to the field. If OA is rotating anticlockwise with a uniform velocity the is proportional to the rate at which the inductor flux. where the line vary is illustrated in Fig. 1. and the e.f. As the coil is rotated. when DD AA DD AA Sine Waves. in the opposite direction. however. after which it once more dies away to zero.e. induced in the coil it is uniformly rotated. e sin 6.f. induced will be zero. It then proceeds to move across the field in the opposite direction.m. ZZT\ Fig.
in a given direction.24 illustrates this point. as will be seen. OA Iucustrating Phase Difference between being in phase and 180° out of phase (i. The line OA is called the vector of the wave and." however. depending on the conditions. The wave marked "current lagging. is due to the vector OIv which is only lagging 45° behind vector being in question. 1. .23. . with time will be a rhythmic change from positive to negative which is called a sine. it is often convenient to consider a wave from the point of view of the vector producing it rather than in its actual sine wave form.23 by the dotted line. and in fact any intermediate position Wove generated by OA Wove generated by OB. the length of the made proportional to the actual strength of the wave The wave marked voltage (full line) is due to the vector OE. before the second wave. Fig. 1. and for this reason is said to lead by 90° on the wave OA. The following points should be noted with respect to phase differences. 1. For convenience and uniformity vectors are usually assumed to rotate in an anticlockwise direction. A leading wave reaches its maximum. as shown in Fig.e.40 RADIO COMMUNICATION Hence the variation of e. and vice versa. it also serves to indicate how two waves of different strengths may be represented. The wave marked "current leading" is due to 0Ia which is 90° ahead of OE. If two waves are in phase they rise and fall exactly together. however. whereas a lagging wave reaches its maximum after the other wave.m. If a wave leads or lags by 90° on a given wave it passes through its maximum when the original wave is passing through zero. exact opposition) may obtain. wave. 90° ahead of Fig. This wave will pass through its maximum before the original wave.f. Phase Difference Consider the wave produced by a second vector OB. Conversely the wave OA is said to lag 90° behind OB. A wave may lead or lag by less or more than 90°. at an angle of 90° with OA.
The number of such waves occurring in one second is termed the frequency. Referring again to Fig.FUNDAMENTALS 41 OE. to a time basis. Since the range of frequencies encountered in practice is very wide.M.g. at each of which the appropriate value of AN 1. while radio frequencies extend well into the gigahertz region.24. as in Fig. it is obvious that each complete revolution of the vector OA will produce one complete wave. though 25 Hz on H. The question of phase difference between sine waves is very important. Generators for aircraft or radar often use higher frequencies (e. suitable prefixes are used to denote appropriate multiples. 60 Hz is is (megahertz) (gigahertz) = 10 Hz = 10 Hz The frequency of the normal alternating current (a. kHz (kilohertz) = 10 8 Hz 6 9 MHz GHz 50 Hz. as detailed in Section 3. 400 Hz) as explained in Chapter 4.22. but this has now been replaced by the hertz (Hz) as the unit of frequency (= 1 c/s). 1.1 (page 134). Frequency a circle and marking off points on the circumference every 30°. Ships. At each position the projection (Fig. . Voltage and Cubbent may be Represented on the Same Diagram thoroughly understood. Here it will be observed it reaches its maxima after the fullline wave but before it has reached zero. These points represent. 1. If any difficulty is experienced he should draw sine waves for himself. This may be done by drawing it is AN is plotted. and the student will do well to study the subject until Vo/tage 'urrent leading 90° Current logging 45 Fig.) supply is occasionally found.22. A horizontal base is then taken and 12 points £ in.22) is measured. the successive points on the circle spaced 30° apart. apart are marked off. 1.c. Speech and music involve frequencies ranging broadly from 15 Hz to 20 kHz. and used. In the Americas. For many years the abbreviation c/s was used to denote the number of cycles per second.
of a sine wave is e = E sin 8. for the wave just considered. a e.f. and 8 is the angle through which the rotating vector OA in Fig. The time required for a complete cycle is called the period of the wave. so that This is. Hence the equation will have revolved through an angle of 8 to the wave generated by OB is A + = E sin (0 + 90°) = Since 360° = 2tr radians. where E is the maximum value.m. the period is clearly 1//. In Fig. which is 2w times the radius.42 RADIO COMMUNICATION Equation to a Sine Wave We have seen that the e. therefore. 1. 90° = tt/2 radians. and The term is usually represented 6 by co.m. e = E sin + t/2) e (cot cot. OB 90°. Hence and e = cat = E sin cot means of expressing the actual value of the particular instant. the frequency of the wave. in terms of the maximum value of the wave.) If the frequency of the wave is/.22 and considering successive values * A radian is the angle through which the vector must move for the arc traced out to be equal to the radius of the vector. 1. OA will revolve/times per second. Hence 360° = 2w radians. . then 8 2im radians.f. so that in time t the number of revolutions is n = Hence 8 2irf is called = 2nft radians the angular velocity of the vector. the sign being be expressed in the form e sign when it is lagging. at any the frequency / is sometimes called the periodicity. 8 depends on the time which has elapsed since the beginning of the cycle. so that In general if a wave is leading or lagging by an angle <f>. the wave E sin {cot <f>).23 if 8 is the angle through which OA has revolved. and the time. used when the wave is leading.* (Note that n need not be a whole n X 360° = = number. Referring once more to Fig. If n is the number of revolutions made in a given time t. lagging or leading wave may be expressed in the same form.22 has moved. Since OA is rotating uniformly. ft. 1. In one complete revolution the are traced out is the complete circumference of a circle. and the may = — ± + Rate of Change of a Sine Wave A sine wave is possessed of a very valuable property relating to the manner in which its value changes from instant to instant.
Qlt however. proportional to the rate of change of the current. Actually if e E sin cot the rate of change AN AN = of e. so that in a the voltage and current rise and together. applied across it. iR v\R. at that particular instant it is not changing at all and the rate of change is zero. It is found that if the rate at which increases or decreases is plotted against the time. i. an e. Hence if 0. This fall is circuit containing resistance only.FUNDAMENTALS of 43 very close to each other. a second sine wave is obtained leading by 90° on the original wave. neglecting the resistance time being. i — = = = .f. is de y.m.m. S. the value is changing rapidly and the rate of change is a maximum. a second sine wave in phase with the first. whose maximum value is eo times as great as that of the original wave. quite apart from any other considerations. . Current in a Circuit have discussed the simple laws relating current and voltage in a circuit carrying direct current. of selfinduction will be set up. = Eco cos cot at = Eco sin (* + I) This is a highly important property of sine waves which may be stated as follows The rate of change of a sine wave is a second sine wave leading by 90°. i. v v Consider first a circuit containing resistance only. Due to the varying current in the coil.e. Inductive Cieouit for the Consider the current in an inductance.m. Pt the curve is horizontal. for the higher the frequency.f. At each instant. t V = 5 sm cot K . We Resistive Circuit by Kirchhoff's Law. A little thought will show that the actual numerical value of this wave must be proportional to the frequency of the original wave. We will now extend this investigation and consider the current in a circuit having an alternating e.f. the more rapid must be the actual rate of change. it will be seen that at the points P. R.e. At the points Q. V sin cot.
: : Reactance In the inductive case we wrote V for I Leo so that / == V[Leo. are sine waves. the current in a capacitor when an alternating voltage is applied across the plates.e. V. . since C is constant and only v is varying. Here. C = = q = = Cv by definition (Section 1. The current in a capacitor Current is the rate at which the charge varying (from the definition of current) i. = . then capacitance. This is analogous to the resistive case where / VJR.44 RADIO COMMUNICATION Now. is = Rate of change of charge = dqjdt But q write i — Cv and. Capacitance Current leading 90° ahead of the voltage. and v charge. at any instant. cot. Hence v i — L di/di = 0.m. so we can write v = V cos cot — V sin (cot f W^) Both the current and voltage. the current lagging 90° behind the voltage. Capaoitive Circuit Consider. is in Section 1. so that an inductance behaves (numerically) like a resistance of value Leo. = I sin Hence that v — I Leo cos cot = is = Ieo cos cot and v = I Leo cos cot The term I Leo clearly the maximum value of the voltage. if q voltage. but are 90° out of phase. Inductance Current lagging 90° behind the voltage.3). of selfinduction L difdt. thirdly. As far as phase is concerned. therefore. the results marized as follows (a) (6) (c) : wave voltage may be sum Resistance Current and voltage in phase. we may — C dv/dt Let Then Therefore = dv/dt = v i V sin cot cot cot Vco cos = VCco cos — VCco sin (cot + w/2) Hence the current in a capacitor having a sine applied across it is also a sine wave leading by 90°. therefore.f.4 it was shown that the If dijdt e.
the term = /. X . The reciprocal of the reactance is called the susceptance. Hence / = VCco = V/il/Cco) The capacitance is thus equivalent numerically to a resistance of value 1/Cco. The quantity \/[R2 (Leo) 2 ] is called the impedance and is usually given the symbol Z. This is a fundamental relationship applying to any form of circuit. where is the reactance. Thus the impedance of a capacitance in series with where It <f> is is less than = + = +X X a resistance is Z = Vt^ 2 + (l/Cw) 2 ] The reciprocal of the impedance is called the admittance. c decreases.s due to both resistance and inductance.m. the phase angle is determined by the ratio of reactance = X/R. 90°. and it is sometimes more convenient to use admittance than impedance. then the voltage in the circuit must be current be i equal to the sum of the back e. X . VCco Similarly in the capacitive case where i is the maximum value of the current = VCco cos mt. 2 \/(R 2 It will be noted that the impedance Z ). so that in general terms tan <f> . L increases as the frequency is increased while capacitive reactance. It will be noted that the reactance is dependent upon frequency. as explained on page 86. which is called the capacitive reactance. for to resistance. Let the /sin tot. being such that tan <f> Lco/R. sin (cot + <j>) the phase difference between the current and voltage.FUNDAMENTALS This is 45 is called the inductive reactance and usually given the symbol X. However.f. Therefore = v = iR + L di/dt = IR sin + I Loo sin = I{R sin + Leo sin cot cot (cot (cot + ttJ2) + tt/2)} This expression may be simplified by trigonometry to the following: v = I^R + 2 (Leo) 2] . Inductive reactance. Phase Angle As just said. Impedance Consider a coil having resistance as well as inductance.
The current in an inductance lags 90° behind the voltage. in inverted form. (f> = Circuit Containing Resistance. 2. conversely. whose value is the numerical difference of the two component waves. to use the full expression for a sine wave. reaching their maxima at the same instant. we may say that the voltage leads by 90° on the we pass an current. Inductance.2).c). and Capacitance Consider the simple case of a coil in series with a capacitor. The reactance of the circuit is thus Leo ~ 1/Cto. — R. If the second term is the larger. in terms of the ratio of the resistance to the total impedance. behaves in an exactly opposite manner. we must insert the expression deduced above. as explained on page 48. the former term is greater than the latter the inductive effect preponderates. and the arrangement behaves as if it were a capacitance. and the whole circuit behaves like an inductance (of smaller value than that of the coil itself). The voltages are thus in opposition. the voltages across each will be out of phase with the current. of course. but in opposition. and they will go through every intermediate point together. the reverse is the case.S. This happens when Leo 1/Cco. but in opposite directions. Values not always necessary or convenient We can. and in fact can be replaced by one single wave. on the other hand. and the voltage lags 90° behind the current.c. simply use the maximum value and this is sometimes done. so that. The phase angle is then given by the relationship cos RJZ as will be seen from the vector diagram in Fig. a condition known as resonance. For the reactance X. They will thus tend to cancel each other out. however. .M. Where a circuit contains resistance. two the sign ~ If signifying the numerical difference between the quantities. while alternating current is abbreviated to a. but it is clearly more useful to adopt a value which would produce the same results as a steady or direct current (usually abbreviated to d. the impedance is obtained in an exactly similar manner to that for the simpler circuits already considered.46 RADIO COMMUNICATION many purposes it is convenient to express it. If alternating current through the circuit. so that the whole becomes Z It will = vT^ + 2 (Leo ~ 1/Cco) 2 ] be clear that by a suitable choice of values the inductive and capacitive reactances may be made to cancel. The capacitance. it is As already mentioned.
If e = E sin the mean value over a half cycle will be sir sin E — «" Cn I dd = 2E/tt Jo . 1. and hence if the current is varying.2.) and it may be noted that it applies to any form of fluctuating waveform.m. Illustrating Mean Value of a Sine Wavk each half cycle is followed by a corresponding half cycle in the reverse direction. As was shown in Section 1. the value of an alternating waveform may be expressed in terms of its mean rectified value. value E = J/V2 = conversely.. The meansquared value of a sine wave E sin 6 is ^ jga £2 1 f 2" E 2 sin 2 Odd run p — cos 20) dO (1 4 Hence the r.m.s. Each half cycle.25.m.f. 1. the energy developed during any cycle will be proportional to the mean value of the square of the current at successive instants. will have a mean value as shown in Fig.m. The average value of a sine wave (or any symmetrical waveform) is clearly zero since Fig. and the equivalent steady current would be one having a value equal to the squareroot of this "mean square.s.f. value of the current (or e. the peak value 0707$ E= Form Factor 14140 Alternatively. which is not dependent on the direction of the current.25." This equivalent value is thus called the rootmeansquare or r. and not merely to sine waves.FUNDAMENTALS 47 This can be done in terms of energy. however. the energy is proportional to the square of the current or e.
E/ \/2. will absorb a power of 1.m. the current is zero and. value. and the form factor. which is the <f> in the next chapter (page 74) that this ratio is cos <f>. In the intervening periods energy is alternately absorbed from and returned to the circuit. Power Factor power will only be developed in must multiply the current the resistance. Similarly. to calculate this one Power It is factor = iJ/V(R 2 +^ 2 ) shown where is power factor is sometimes expressed in this form. across the resistance. Clearly.s. If the current and voltage. so that the actual power is as nearly as possible equal to the apparent power. Thus at the instant when the voltage is a maximum. a pure inductance or capacitance (i. are quite different as will be seen shortly. Hence the apparent power EI must be multiplied by a power factor.000 watts or 1 kilowatt (kW).m.48 RADIO COMMUNICATION r. Current which is 90° out of phase is often referred to as "wattless" current. and this gives the power in watts. however.s. the phase angle between the voltage and current. therefore.2. and by that component of the voltage which is expended in overcoming the back e. the power factor can never exceed unity.e.s. Power As explained in Section 1. and the .s. is not as great. In circuits concerned with the generation or utilization of energy one endeavours to maintain the power factor large.m. As long as the current and voltage are in phase it is sufficient to multiply the r. but the average power consumed remains zero. in which case the r. In such cases the voltage and current are 90° out of phase. when the current is a maximum the voltage is zero. the power in a circuit is obtained by multiplying the voltage by the current. as just stated. The 111 ratio is between the equal to and mean values is called the form factor and 7T/2V2 = It must be remembered that these simple expressions only apply to pure sine waves. and mean values.f.m. Thus a circuit carrying a current of 10 amps and having a voltage across it of 100 volts. which is the ratio of the resistance to the total impedance of the circuit. for this reason. having no resistance) cannot absorb any power at all. are not in phase the power Indeed. Many of the waveforms encountered in practice are complex. values together. so that once again the power is zero. is The r. Hence In a practical circuit.m. therefore. the power is nothing.
Waveforms Containing Second Harmonic frequency of the other. are pure. Fig.26. The two waves are known as the fundamental and the second harmonic respectively. 1. Fig. one having only 10 per cent of the amplitude but twice the Fig. where three different waveforms are illustrated. The form of the wave again depends on the phase of the harmonic.27 illustrates the effect of a third harmonic. There are many circuits. Complex Waves. the ratio of the resistance to the reactance in the circuit. In such circumstances is small compared with Lea or l/Cco and we capacitors. Few practical waveforms. however. 1. 1. but in this . Fig. each containing the same amount of second harmonic. This is shown in Fig. Waveforms Containing Third Harmonic The harmonics are not necessarily in phase with the fundamental. 1. but of different phase.26.27.FUNDAMENTALS circuit. Harmonics may be present having any multiple of the fundamental frequency and of variable amplitude. which requires the power factor to be small. particularly those using which it is desirable to keep the effective resistance as low as possible. and the actual form of the wave will in fact depend upon the relative phase of the harmonics.26 shows a wave made up of two sine waves. in fact. 1. 49 value obtained when there is no inductance or capacitance in the or when the effect of one has been neutralized by the other. Harmonics Many of the waveforms encountered in practice are far from being pure sine waves. is to say. in R can write the power factor as R/X that .
This is characteristic of waves containing odd harmonics. Where a larger harmonic content is present the approximate treatment above is not accurate enough and it is necessary to analyse the voltage wave to determine the effective fundamental and harmonic voltages. as shown in Fig.30. which is only 28 per cent of the funda Hence an inductive circuit tends to suppress harmonics to an extent which increases as the "order" of the harmonic increases. The reactance of the circuit at the fundamental frequency will be a 2 V(200 + 314 ) = 3723 ohms and the current will thus be 269 mA.SO case half RADIO COMMUNICATION it will be seen that the waveform is symmetrical. whereas even harmonics produce asymmetry. as in the vowel sounds of speech illustrated in Fig.s. many circuits used for control or test purposes may be of a specifically nonsinusoidal form. is thus 076 mA. The voltage is 05 volt and the *n (a) Fig. as the reader can easily verify for himself. Suppose we have a voltage having an r.28. the r. They may be of a regularly recurring nature. The majority are complex in form and hence contain a high proportion of harmonics. Equipment to handle such waveforms must clearly be able to respond adequately not only to the fundamental but also to the .m. of the waveforms encountered in communications work are pure sine waves. the negative wave being an inverted copy of the positive half wave. 1. such as are illustrated in Few Figs. (b) Nonsinusoidal Waveforms Two Typical current mental. Let the funda mental frequency be 50 Hz. 1. 1.26.28 to 1. If the percentage of harmonics is small. The reactance of the circuit to the second harmonic will be \/(200 2 + 628 2 ) = 6591 ohms.s. known as transients. and which we know to contain 5 per cent of second harmonic. to be applied across a circuit containing 200 ohms resistance and an inductance of 1 henry. 1.37 or they may be single irregular pulses. In addition. value of 10 volts.m. while a capacitive circuit will tend to increase the percentage of harmonic. value of the voltage or current is only slightly affected (although the form factor may be appreciably different).
. i. 1. 6 X noting the actual value of the function at suitable points in the . Specifically. 1. cycle. .FUNDAMENTALS highest significant harmonic frequency. frequently used in oscillo graph work.f f Jo I 2* f(x) cos nx dx 2ir it h f(x) sin nx dx wave of Fig. . and so on. 1. 51 It can be shown that any waveform. . the square wave of Fig. +. Fourier Series The exact analysis of waveforms was developed by Fourier. 20 per cent fifth harmonic. who showed that any periodic function may be written in the form /(*) = «„ + «! cos x + a + b sin x + 6 x .) The triangular wave of Fig. f• . 2 cos 2x sin . the a fundamental and odd harmonics (since expression being BE 8$ / [ . 33J per cent third harmonic. . can be analysed into a fundamental component and a series of harmonics.28 (a) is not greatly dissimilar to that of Fig. sin cot 1 1 — ^ sin Scot + ^i sm ^ mi . .e. of amplitude inversely proportional to the order of the harmonic. if the period of /(#) is 2tt.28 (a) gives a of odd harmonics. however complex. 2 2x • • The values of the constants a ax are obtained by . 1. The equation is analysis for the square series The Fourier fundamental and a e = — IT (sin cot + J sin Scot f i sin 5cot 1. i r 2w f(x)dx 1 .29 illustrates a sawtooth wave. ~~ • • • ) Fig. may similarly be built up from symmetrical)..27 (a) containing a fundamental and an inphase third harmonic. For example. This is an asymmetrical wave and will thus contain .28 (b) it is f . The addition of suitably proportioned fifth and seventh harmonics will produce an even closer approximation as will be seen shortly.
RADIO COMMUNICATION It is. similar to an inverted wave of the form of Fig. Also since the e.. .5 + . cos 2ncot )] r\_/\_. the complete expression being =& _2 77 (sin cot + \ sin 2cot + J sin 3<wi + . + 1wo „ t< + (2» — 1)(2» + 1) /"HZ . but a certain proportion of odd harmonics will be required in order to produce the sharp transitions from maximum to minimum. it is necessary to add a constant d.26 (a). Sawtooth Wave The dotted waveform ia Fig. The mean value of the wave is thus IIn.30. 1. the earlier mean value is 22?/tt was stated on page 47). 1. being always above the zero line. Fig.52 even harmonics.0 1 f •)] (as Here..30. shows a sine wave in which the negative halfcycles are suppressed. as one would expect. 1.i + — cos 4to< 6. component (equal to half the peak amplitude). The equation to this is similar. 1.29. 1. 1.c. being _ cos 2cot a \JT i IT \l.f.30 (6) shows a rectified sine wave in which the negative halfcycles are reversed instead of merely being suppressed. » . does not reverse.* * I /vw\ \ ' * i (a) <*> Fig.26 (a) inverted A Fig. in fact.m.30 (a) further interesting case is the rectified sine wave of Fig. The equation i • to this is fll" 1 '2/1 cos 2cot  1 \ r= cos 4eo< 6.)] Fig. Rectified Sine Waves I/77 This wave only contains even harmonies plus a constant term of times the maximum value. 1.
the distance between the molecules increases so that a free electron can travel relatively long distances before being captured by an ionized molecule. the Fourier series is apt to be lengthy and it is easier to base the calculations on simple proportion as explained in Chapter 17 (page 715). for occasional molecules of gas to lose an electron. 1. wave contains a zerofrequency (d.FUNDAMENTALS Average Values of Complex Waves The r. the only current which will flow in a gas is a displacement current caused by a varying or alternating e.m. Conduction 1.s. so that equilibrium is quickly restored. . . provided that its Fourier components are known.) component. The molecules are of simple structure and do not readily part with an electron (though a notable exception is water vapour.1.m. r. and mean values of complex waves 53 are evaluated by adding the appropriate values for each of the components.m.m. and at normal pressure the rate of recombination is high.s. Normally. however. value If the + E sin 2cot + E sin Scot + VtK^i + E + ®% + tot 2 3 . Thus for a wave e =E 1 sin is the r.)] mean (rectified) value E + (E + E 2 +. so that water vapour is slightly conducting).c. 8 z • •)] and i.s. therefore. Each free electron drifts through the gas until it encounters an ionized molecule when the two will recombine.6.s. in which the electrons have some freedom. however. molecules each deficient in an electron. value is VW + Similarly the i(E 1i + E£ + is . 1 . The gas is then said to be ionized. As the pressure is reduced. . Hence in this case the r. values of this E . in Gases and Liquids gases are normally insulators. . the mean component are the same as its actual value.) These expressions apply to any form of complex wave.m. .e. As stated in Section Ionization It is possible.f. i.e. in which case the gas contains a number of free electrons and a corresponding number of ions. In the case of square waves or pulses. The average distance which an electron can travel is . however.
Tabus 1.m.f. while the positivelycharged ions drift towards the negative electrode. This is known as a metastable state and may last for an appreciable fraction of a second until a collision occurs with another atom. The electrons will move towards the positive electrode.1 Ionization Potentials Gas Ionization Potential Hydrogen Helium Nitrogen . two electrodes in a glass bulb filled with gas. giving up its excess energy in the form of light waves. as the pressure is reduced.f.m. Generally it then returns almost immediately to its original state. This can happen in two ways: (a) An electron may jump from one orbit to another (of higher energy level). which is called the anode (meaning literally "way to").1. the mean path increases so that the gas remains ionized for a longer period. . the resistance is not constant but varies with the e. applied across At any . Glow Discharge instant the current will be equal to the e. divided by the resistance in a gas.54 called the free RADIO COMMUNICATION mean free path and.f.m. There are. (b) An electron atom ionized.f. if the gas is ionized. however. and varies with different gases. will tend to displace some of the electrons in the gas molecules. then. may jump out of orbit altogether. called the cathode ("way from"). Mercury vapour 135/159* 245 145/167* 136/158* 215 157 104/19* * The two values are for the gas in atomic or molecular state respectively. leaving the The e. The electric field produced by the e.f. typical values being shown in Table 1. Consider the effect of an e. Oxygen Neon Argon . but as explained in Appendix 1 any actual displacement can only occur in jumps from one orbit to another. however. a conduction current will flow.m.f.m. certain orbits from which the electron cannot revert to normal. applied across them. If we have two electrodes in a glass bulb or other container with an e. required to produce this result is called the ionization potential.m.
f. The current increases very rapidly by what is called avalanche effect up to a limiting value where the gas is completely ionized. entirely different. . A notable exception. Practical usage falls into two distinct classes. This limiting value depends on the number of gas molecules available. solutions consisting of chemical in Appendix 1.f. which is sodium chloride. Electrolysis The conditions are particularly aqueous solutions. with chlorine two elements join forces the chlorine atom can absorb the isolated sodium electron. Thus common salt. the shape of the electrodes and the distance between them. i. producing still further ionization. The second usage is concerned with gases at atmospheric pressure. This is known as ioniza and is clearly cumulative.e.FUNDAMENTALS Ionization by Collision 55 As the e. Such devices. chemical due to what may be called economy of possibilities. however. such as oil. Any detailed discussion of gaseous conduction is beyond the scope of the present work. are discussed further in Chapter 5. falls into two distinct classes depending upon whether the substance is a pure liquid or a solution. In the first the gas is at low pressure in a suitable container.m. while any molten metal still behaves as a metal even though liquid in form. and depends on the gas pressure. its which has 7 electrons in outer shell. They are dielectrics and behave in the same manner as solid dielectrics. exist naturally in liquid form. so that between them they fill up the outer shell completely. the current being fairly small. with solutions.m. being capable of carrying a displacement current but not passing any appreciable conduction current. usually of glass. is a combination of sodium (which has only one electron As explained is combination between two elements in its outer shell as against a possible maximum If the of 8). is still further increased the velocity of the electrons becomes so high that any encounter with a gas molecule displaces an electron. Conduction in Liquids The behaviour of liquids under the influence of an e. of course. is mercury which is a liquid metal and behaves as such. and hence increases with the tion by collision pressure. in which case the current flows in the form of an arc or spark. usually known as glow tubes or gasfilled valves. salts dissolved in water. The potential at which this effect occurs is called the breakdown potential. Certain substances.
so that a current will flow. which will appear as bubbles at the cathode.f. Similarly the sodium.). This process is known as electrolysis and the liquid is called the electrolyte. and is called a chlorine ion (CI— ). Illtjstbatino Electrolysis field existing between the electrodes will cause the negative ions to move towards the positive electrode (the anode) and the positive ions to move towards the cathode. however. The electric Battery Cathode (negative) Anode (positive) Electrolyte \ " ' Fig.f.m. the chemical is a solution of a suitable metal salt (such as nickel sulphate) the nascent metal forms a metallic coating on the cathode.56 RADIO COMMUNICATION If the salt is dissolved in water.m. This has many practical implications.31. The atom will combine with another to form a chlorine molecule and bubbles of chlorine gas will appear at the surface of the anode (some of which will dissolve in the water). For example. When a CI— ion arrives at the anode it will give up its surplus electron and become a normal chlorine atom again. as shown in Fig. while similarly an Naf. releasing hydrogen.* Consider now a solution of salt in water with two electrodes connected to a source of e. and negative . called a nascent state. * It should be noted that in aqueous solutions both positive ions exist. It will be seen that the passage of current through an aqueous solution produces a breakdown of the dissolved chemical chlorine into its constituent parts. now deficient in an electron. The nascent sodium will combine with the water to form sodium hydroxide. if some useful. and there will thus be a flow of electrons through the connecting wires (including the source of e. 1. The neutral atoms released at the electrodes are in a very active condition. becomes a positively charged sodium ion (Na+). the sodium and chlorine atoms separate but the chlorine still retains the electron which it borrowed from the sodium. It will therefore have an excess of negative charge.31. this being the basis of electroplating. which they cannot maintain.ion arriving at the cathode will collect an electron and become neutral. some troublesome. whereas an ionized gas molecule is always positive. 1.
The equivalent weight of an element is the relative weight of the element which would combine with approximately 1 gramme of hydrogen (actually 1008 g as explained in Appendix 1).. Thus copper factor. 2... The mass of material liberated by one coulomb is called the electrochemical equivalent and can be deduced for any element by multiplying the equivalent weight by the above Some typical values are shown in the table... . .. . The amount of substance liberated at the electrodes is proportional to the equivalent (chemical) weight of the substance. The amount of electrolytic action which takes place is independent of the size of the electrodes but is proportional to the quantity of electricity which passes through the solution (i. who formulated two basic laws which are known by his name. the product of the current and the time). Electrochemical Equivalents Equivalent Electrochemical Equivalent Element Weight (mg/coulomb) Hydrogen Oxygen Sodium Nickel Chlorine Copper (monovalent) Copper (divalent) Silver . it is found that if a current of one ampere is passed through a solution of silver nitrate for one second the mass of silver deposited at the cathode is always 1118 mg..FUNDAMENTALS Faraday's Laws 57 The first quantitative study of electrolytic action was made in 1830 by Faraday. They are said to have different valencies.. Now..e.2. and hence in order to determine the quantity of material liberated by electrolytic action it is simply necessary to find (experimentally) the appropriate factor which will express the result in terms of the quantity of electricity involved. Note that some elements.. The equivalent weight of silver is 10788 so that the factor we are seeking is 1036 x 10* in SI units or 1036 X 10~ 2 if the result is expressed in mg/coulomb. can combine in more than one way. . Table 1. This relationship.. such as copper. in fact.. is so precise that it has been used as the practical standard of current. These are: 1. 1008 80 230 2935 3546 6357 3179 10788 001045 00829 02383 03041 03674 06588 03294 1118 .
1. in fact. If prongs will move to and fro some hundreds of times a second. one twice the upon the particular chemical combination used. The farther one goes from the source the greater the volume of air available so that the jostling becomes less pronounced. the Air molecules displaced by movement of prongs Fig. The two types essentially differently in others of wave. longitudinal and transverse. depending can have two equivalent weights. Longitudinal waves are of a material nature and operate by the of particles along the direction of travel of the disturbance. the most common example of this being sound waves. 1. Tuning Fobk When the prong moves outwards the molecules of air will be pushed away when it moves inwards it will create a partial vacuum which will be filled by molecules of air rushing back again.32. ripples on a pond. though similar in some respects. Such waves can be material. air. Vibrations and Waves The transmission of intelligence from one point to another is accomplished by means of waves produced by suitable forms of vibration. and this movement will displace the surrounding this is set in vibration. 1. behave and are best considered separately.58 RADIO COMMUNICATION other. e. But the molecules of air thus displaced will in their turn displace : other molecules farther away so that the disturbance is transmitted by this jostling of adjacent molecules for an appreciable distance. as one would expect. but are mainly electromagnetic in character.32.7. . In a transverse wave the movement takes place at right angles to the direction of travel. There are actually two types of wave. decreases as the square of the distance. Sound Waves Consider the sound produced by a tuning fork as in Fig.g. such as light movement or radio waves. The actual displacement.
g. Velocity of Propagation . is situation will then be as in Pig. the blast from an explosion) and as a result a transient sound may be produced. These in turn.33 (c) and so the disturbance travels outwards. The (6). 1. Wavelength The transmission of the sound obviously takes a finite time. 1. Meanwhile the "white" molecules will move back in the reverse and then forward again reaching the original position one cycle later. The human ear is designed to be very sensitive to slight differences of air pressure so that the tiny displacement of the air molecules may be detected at a considerable distance from the source and we say that we "hear" the sound.. Consider a section of the air (or other medium) a short distance away from the source of the sound as in Fig.«.»«. ABC (o) oo o •• oo o•• oo o •• (« ABC ooooo* ooooo* ooooo* ABC (e) Fig. The white circles represent molecules moving from left to right. important to note that the air as a whole does not move. but this is not a true sound wave. moving outwards with a constant It is velocity. during which time the wavefront will have travelled a certain distance depending on the velocity with which disturbances can travel in the particular medium called the velocity of propadirection — gation. having been displaced by the sound source.•»». Illustrating Travel of Sound Wave displace further molecules so that the wavefront will gradually travel to the right. 1. o • • • • o • • • • • o• • •• • ».33 clearly a repetition of the conditions at A.FUNDAMENTALS 59 This effect is called a sound wave.. the space between the source and the wavefront being composed of a series of compressions and rarefactions an equal distance apart. Certain physical disturbances may produce actual movement of large masses of air (e. . and at B there During the next cycle the wave front will travel a further similar distance as in Fig. and it is because of this that sound can be heard at a distance. while the black circles are molecules as yet unmoved. it is only the individual molecules which are momentarily displaced and so communicate the intelligence to neighbouring molecules.33. and they will then *• •*.. 1. will be displaced to the right but with a slight time lag because of their inertia.33 (a).
two points of maximum pressure in the same direction) is called the wavelength. . Sound waves can thus be transmitted through liquid or solid materials just as well as —in fact better than—in air. 1. (6) The frequency of the vibrations.g.34. depending as it does upon physical displacement of adjoining molecules. In fact we can say that Frequency X Wavelength = Velocity of propagation and this is a fundamental property of any wave motion. The arrangement constitutes what is called a travelling wave and the distance between any two similar points (e. the closer the molecules) the more rapidly will any disturbance be transmitted. 1.60 RADIO COMMUNICATION Wavelength have assumed that the molecules in Fig.34. corresponding to the maximum movement of the prong of the tuning fork. We Hence at any point in space the air pressure will gradually vary between these two maxima in opposite directions but the r Wavelength' Fig. clear that this distance will depend upon (a) The velocity of travel of the wave. can take place in any material. 1.e. and the greater the density (i.33 were at the condition of maximum displacement. The velocity of propagation depends on the medium in which the wave is transmitted for clearly the transmission. for if a complete cycle of vibration is completed in half the time the wavefront will only have Now it will be travelled half the distance. but it •will be clear that a similar process takes place over the whole cycle from maximum in one direction to maximum in the opposite direction and back again. Distance Illustrating Wavelength points of maximum (or any intermediate) pressure will travel out wards as illustrated in Fig.
000 metres/gee Frequency of Sound Waves Longitudinal vibrations of this type can occur at frequenoiesirom a few hertz up to tens of thousands of hertz. Velocity of Sotjnd at Normal Temperatures Air Water Steel 330 metres/sec 1.000 Hz. 1. particularly in youth.35. The ear is not uniformly sensitive.000 Frequency (Hz) Fig.000 Hz than at 25 Hz. The range of normal sound to which the human ear is sensitive is from about 15 to /OO 400 I. "sees" by emitting vibrations at a frequency of some 50. though some people. (Actually it emits these waves in a series of . Table 1. Relative Sensitivity of Human Ear 15. The top curve shows the threshold of pain. can hear frequencies of 20.000 times more sensitive around 2. The bat. Fig.000 10.35 illustrates the minimum audible sound level at different frequencies. 1.500 metres/sec 5.3. for example. Air pressures in excess of this will cause distress.OOO 4. but the figures indicate the relative values.000 Hz and detects the echoes from any obstacles by the reflections produced. from which it will be seen that the ear is some 5. There are vibrations beyond the limit of human hearing. just explained the exact value depends on density and temperature.FUNDAMENTALS As 61 Approximate values of the velocity of sound are shown below.000 Hz or more.
The actual intervals (with the names assigned to the different notes) are: Note .62 RADIO COMMUNICATION short bursts about 20. The small number pitch is determined by the fundamental frequency. depends on the nature of the vibration.000 Hz are used for a variety of purposes such as drilling... one of which is exactly twice the frequency of A the other.36. There is nearly always an admixture of harmonics or overtones. while as the frequency is increased the note becomes higher in pitch. of vibrations per second gives a low or deep note. are the result of air vibrations of this type. the actual frequencies being shown in Fig. These are other vibrations. There are two essential qualities in any sound. produce notes which are similar in quality but of different pitch.37 shows two simple vowel sounds. namely. Very few sounds consist of a pure sine wave. The term "octave" scale in ratios. however. but having a frequency which is an integral multiple thereof. a violin and a piano. the pitch and tone or timbre. cleaning. The flute note is nearly a pure sine wave.000 times a second and it is this pulse frequency which can be heard as the bat's "squeak. The same note played on a flute. Nature of Sound The various sounds with which we and music. Do 1 Re 9/8 Mi 5/4 Fa 4/3 Sol 3/2 La 5/3 Si Do 2 Relative frequency 15/8 Musical notes cover a range of some 8 or 9 octaves while speech covers about half this range. echosounding.. The quality or timbre of any note. It is found that two vibrations.000 to 50. Speech is similarly composed of vibrations which are rarely pure but which contain very many harmonics. are familiar. Fig. is derived from musical parlance for the major music consists of eight notes related by small (but uneven) the eighth note being of exactly twice the frequency. . 1. etc. will have quite a different quality though the pitch will be the same. The higherfrequency vibration is said to be the octave of the other. but even these sounds uttered by two different . including speech but these vibrations are of widely differing character. smaller in amplitude than the fundamental. but the violin and piano notes will contain many (different) harmonics which will give them entirely different and distinctive qualities. 1. for example.") In the industrial field supersonic vibrations in the range 20.
FUNDAMENTALS 63 people will not be identical.000 jM to^ Bin I Flute°T Strings* 4. More extensive examples of sound waves. 1. Vibrations per second 64. The quality of voice.000 Supersonics 32.000 Bats ^Piccoto] °\ 8.000 \ 16.000 Piano 1. showing their relative harmonic content. Renton (Pitman). 5ound 00 Fig Diagram of Representative Speech Waves Transients Not all for example.37. N.024 Speech 512 Music Brass 256 f f MiddleC 128 64 Double boss 32 16 Fig. are given in Chapter 15 of Telecommunication Principles by R. Y 16 ft organ pipe Atjdiofbequbncy Spectrum 3 WvVWvVWVWVW 1. in fact. depends on the nature of the harmonics produced. An object falling to the ground.36. produce a sudden sharp displacement of the air hear as a thud or bang. will which we sounds are repetitive.000 2. The beat of a drum or clash of .
. indeed.m. the translation into electric currents.38.64 RADIO COMMUNICATION cymbals in music produces a similar single pulse. one side of the container being flexible and hence able to yield to the varying air pressure of the sound wave. 1. but carbon exists in many forms. consists of a container filled with small granules of a suitable grade of carbon. This in turn varies the resistance of the carbon granules so that if an e. the harmonics in this case being large and of a high order (i. the relative speed with which the air pressure builds up. i. In particular it is found that if a number of pieces of carbon are in contact the electrical resistance depends on the pressure between them. such as the labials p and b. A carbon microphone. . i. They can be analysed into a fundamental and a series of harmonics. and when it is required to reproduce such sounds electrically the system has to be able to handle the extremely rapid changes required. The first process. from lampblack to diamonds. These single pulses are called transients. is a conductor of electricity. / Diagram or Cabbon Microphone device called a microphone which consists essentially of an arrangement in which variation of air pressure causes a change in electrical characteristics.e. the translation of sound waves into electrical currents and the subsequent reconversion into sound with which communication engineering is largely concerned. or dentals t and d are short explosive pulses. The more important aspect is the wavefront. while many of the sounds in speech. is accomplished by a Diaphragm Carbon Granules Container Carbon Disc n Mouthpiece Fig. Telephony It is. 1.f. Carbon. therefore. is applied across the microphone the current flowing will vary in accordance with the sound impinging on the diaphragm. many times the fundamental frequency).38. and the electrical performance varies very widely according to the form. thus providing an electric current of the same form as the sound wave. One of the simplest (and earliest) of such devices is the carbon microphone illustrated in Fig. though not a metal.e.e.
and providing for greater actual movement. Two coils of wire are mounted on iron pole pieces which are located very close to a flexible iron diaphragm. This is done by causing the current to move a diaphragm the vibrations of which set up air waves which will reproduce the original sound.39.39. These are discussed in more detail in Section 4. The air molecules . For special requirements different forms of microphone are used which utilize electromagnetic or piezoelectric effects.FUNDAMENTALS 65 For many purposes (e. moreover the air vibrations produced are relatively small. sound waves of considerable intensity can be produced. Reflection. Standing Waves If a sound wave travelling through the air meets a solid obstacle there will be a sharp change in the conditions. The electric translated can now be current into which the sound waves have been transmitted by wire or radio to a distant Case Diaphragm Winding Ebonite. Again.8. as discussed in Section 4. Such a device is called a loudspeaker. The varying currents through the coils produce varying attraction of the diaphragm which causes it to vibrate and reproduce the sound waves. By making the diaphragm larger. One of the simplest of such arrangements is the telephone receiver illustrated in Fig. earpiece Fio.7. and while basically the same in principle it incorporates a number of modifications in detail. Washer Telephone Earpiece point where it can be reconverted into sound. the normal house telephone) this simple arrangement is quite adequate but it only responds satisfactorily over a limited range of frequencies. this arrangement is only suitable over a limited frequency range. 1. 1 .g.
These points of cancellation will. In between the cancellation points. If the hand is moved slowly from side to side no unusual effect occurs but if it is jerked sharply sideways a ripple will travel along We A the rope. however. but (if the rope is slack) a rapid movement will produce a ripple which reaches the far end with considerable amplitude.66 will RADIO COMMUNICATION be unable to continue to move smoothly forward and will have to transmit their energy to other molecules around. and there will be points at which the two cancel each other out leaving only a small residual due to the fact that the reflected wave. The ripple which travels along the rope is a displacement of the appropriate portion at right angles to the direction of travel. and mainly backwards. be separated by exactly half a wavelength. in fact. The ripple is. This will produce a series of sound waves travelling back towards the source where they may arrive some seconds later to produce an echo. This arises because the inertia of the rope does not permit it to change its position suddenly so that the disturbance is communicated from one piece of rope to the next with a small time lag. Some small part of the energy will be absorbed and transmitted through the solid material but most of it will be reflected. Transverse Vibrations now come to the second type of vibration. of wave may be obtained by taking a length of about ten feet of heavy cord or thin rope and fixing the far end to a suitable anchorage. while in between there will be places where the waves reinforce each other to produce a maximum. a form of transverse vibration and if the . in fact. 1. or nodes. will be slightly less strong than the forward wave. 2. This is called a standing wave because. Two things will be noted from this experiment. although the forward and reflected waves are travelling through the air. there will. The nearend is held in the hand with the rope slack. in fact. If the hand is only moved slowly there is no ripple and the far end does not move appreciably. having travelled farther. On the return journey. the interference between them always produces the same intensity at any point in the path between the source and the reflector. the transverse vibration in which the displacement of the medium is at right angles simple demonstration of this type to the direction of the wave. be a gradual increase in the sound up to a maximum followed by a fall to zero again. they will meet the forwardgoing waves so that the air pressure at any point will be the difference between the two.
1. and the waves which are produced in this region are of a physical character which we know as heat and infrared waves. Ultimately a region is reached where the necessary movement of the electrons becomes comparable with molecular dimensions.e. It will depend on the tension in the rope and will probably not even be uniform along the length of the rope. These lines of force possess inertia just like a rope so that if the electron is caused to change its position rapidly. i. With still shorter wavelengths comparable with atomic dimensions we come into the region of visible light. This is similar to the longitudinal vibrations. produced by causing electric currents to travel up and down elevated structures called aerials. In fact there is a very wide range of transverse vibrations in nature which make use of electrical "ropes. As before. where it is explained that the waves are produced by sudden changes in the position of electrons. the wavelength is the distance between successive points of similar displacement e. As the wavelength becomes shorter we pass through the region of ultraviolet light. and in fact the same rule applies. while at still shorter . These changes may be relatively slow. It will also be noticed that the more rapid the movement the closer together the waves become.— FUNDAMENTALS hand 67 is moved rapidly from side to side a definite wave motion will be noted in the rope. The velocity in the crude experiment just described is not very definite. but this is unimportant since the experiment is merely to demonstrate the essential = between transverse and longitudinal vibrations.g. ripples will be produced in the lines of force associated with it. These will travel through free difference space in what are called electromagnetic waves. waves which are similar to those of light but beyond the limit of visibility of the human eye. viz: frequency x wavelength velocity. Since the "tension" in the lines of force is constant it is found that all these waves travel with a constant velocity c —= 3 X 10 8 metres/ second and hence the wavelength and frequency of electromagnetic waves are connected by the relationship this type. in which case we have the range of radio waves of gradually decreasing wavelength as the frequency is increased. as also is the vast range of natural Frequency (Hz) X Wavelength (metres) discussed =c=3 x 10 8 m/s The mechanism by which electromagnetic waves are produced is more fully in Chapter 3. the distance between the peaks in Fig." It has already been pointed out that an electron has associated with it an electric field which can be represented as a series of lines of force radiating from it.34. Radio waves are of phenomena from heat and light to Xrays and cosmic rays.
Electromagnetic Spectrum Physical Transverse Vibrations Transverse vibrations are also found in physical structures. mainly We .m. 1.4 eguei «w*l (Hz) £AJvfWvW/J</wj St/F vUltra violet Xrays —t r s \y. 1.f.°y i * Audio frequencies p * Radio Infra roys waves red waves Visible Cosmic roys light Fig. in various forms of circuit. This process is not used to any extent in practice. gamma rays and cosmic rays.s can be produced in practice. but because some part of the structure has to be fixed the vibrations only appear as standing waves.68 RAblO COMMUNICATION wavelengths we find Xrays. for example. which will be discussed in more detail at the appropriate time in the later chapters. There are six main methods. They are as follows: We have seen that a dielectric or insulated body 1. This will happen very quickly and the energy will be dissipated almost immediately except at the particular frequency for which the length of the string is equal to an exact wavelength or suitable fraction thereof.M. all of which are of the same basic character but are produced by very The complete range of special types of physical disturbance. » . 1. Friction. A violin or piano string. The vibrations will travel along the wire to one end. *o .8. and vibrations at this natural frequency can be produced relatively easily and may persist for an appreciable time.m. have seen the effects of e. where they will be reflected back to the other end and again reflected and so on.40. may be caused to vibrate in a direction at right angles to its length. 10" i — 10' I /O" I 1 IO~* 10"* IO' 6 ' IO' L_ I0 4 a /0~'° 4 IO~" id"' Wavelength (metres) 10* io" I0 B IO 10 IO 12 IO 14 10 " 10 m . .f. vibrations known 10" _i to science is shown in Fig. but a brief summary will be helpful here. This condition is known as (mechanical) resonance.F. may be charged by the physical removal of surface electrons by friction.40. but little has been said of the ways in which such e. Sources of E.
A machine in which the change of linkage is produced by mechanical rotation is called a an alternator.f. in paper mills the rapidly moving paper acquires a charge which can build up to a dangerous value and protective devices have to be installed to remove this charge.constantan which gives a p.FUNDAMENTALS it is 69 because the quantity of electricity so extracted is very small.f. By choosing suitable materials this thermal e. of self or mutual induction which by suitable design can be of almost any desired character. is electromagnetic induction. occasions where this electrochemical action is a hazard and precautions have to be taken to avoid or minimize its effect. are shown in Table 1.f. notably the production of very high potentials for particular requirements. There are also occasions where the production of static electricity by friction constitutes a hazard.4. Electromagnetic induction. The dynamo. generated is alternating. All substances possess a certain potential so that if two elements are brought into contact a potential difference between them.6. Typical values of e.f.d. Such arrangements are called batteries and are discussed in Section 4.d. 3. steady or varying. in which the linkage between a coil and a magnetic field is varied. can be 418 fiV per °C. By far the most common method of producing e. can be sufficient to be of practical use in which case heat energy is transformed into electrical energy. but employed for certain special purposes.m. principal forms of generator are discussed in Section 4. This produces an e.m. etc. In this case the electrical energy results from a chemical change in the structure of the materials used.m. The potential of any element depends on temperature. singly or as Such arrangements are called thermocouples and may be used an assembly of a number of elements connected in series .f. 4.m. Electrochemical e.f.s are potential differences they must be relative to some datum. but this does not mean that platinum must be one of the elements used. however.m.f.m. low or high voltage. for which platinum has been taken. The air friction on the skin of an aircraft will also cause it to become charged and tyres of conducting rubber are used to allow this charge to leak away when the aircraft lands. of exists solutions this p. 2. Hence if two metals are brought into contact and the temperature of the junction is increased the potential difference between the two metals will be changed. or if the e.m. For example.m. A common combination is copper. Since the e. With certain elements in suitable chemical made available for external use.5. Thermoelectric e.f. There are.
is directly proportional to the mechanical Appreciably greater response can be obtained from Rochelle but this is more fragile.f.f. is developed in a plane at right angles to the strain.f. This may take the form of a change of electrical resistance. but if the frequency of the e. across the crystal causing a mechanical deformation. P. The effect is normally very small.D. so that thermocouples are used mainly for measurement or control purposes since they provide a convenient means of converting temperatures into electrical terms. the e. however. This is known as piezoelectric action and is utilized for translating mechanical vibrations into electrical voltages.m.F.4 Thbbmoelectrio E. though small. The action is reversible. as in the case of selenium. or an actual emission of electrons from the material.f.f. generated is small. As will be seen. 470 650 74 160 344 to form a thermopile. an e.M... and this technique is stress. Photoelectric e.m.m. mechanical pressure or twisting it is found that a small e.m.m. With proper precautions and within suitable limits the e. . found to undergo a change of structure when exposed to light or other electromagnetic waves in this region. Piezoelectric effect. salt..m.f. against Platinum per "G (microvolts) Metal Aluminium Antimony Bismuth Copper 38 Iron Constantan . which thus behaves as a source of e.m. while a variety of other substances have now been found which exhibit the effect. The e. .... Certain elements and compounds are 5.. used in the construction of precision oscillators.70 RADIO COMMUNICATION Table 1.f.m. may be made proportional to the incident 5. coincides with the mechanical resonant frequency of the crystal the crystal can be maintained in a state of vibration of extremely constant frequency. Further details are given in Chapter If a crystal of quartz is subjected to 6.f. light.
and the various ways in which an applied e. Vectors have discussed the basic forms of e. indeed. for example. the currents phase. In the first case. This aspect has already been discussed briefly in Section 1. This is often called the vector sum. Moreover.f. Wk communication of any sort necessarily involves varying and voltages will be alternating and therefore the calculation of the combined effect is not a matter of simple addition or subtraction but will need to take account of the since quantities.f is opposed by the back e.5.f. it was shown.m. Practical radio engineering is concerned with the correct choice of circuitry to obtain the required currents and voltages at the requisite points. practical circuits are never simple arrangements of a single impedance operated on by a single e. that the voltage across an inductance and resistance in series is not simply the sum of the voltage across each but is \/(^£ 2 VB *). It is.m.m. inductance. For example.s developed.1. In general. if OF is the vector of 71 . where it was also explained how sine waves out of phase with each other may be represented by two vectors at a suitable angle. frequently helpful to consider a circuit in terms of the current and voltage vectors. or capacitance. and the present chapter deals with the methods principally used in such calculations. because it is the sum of two vectors VL and VR at right where + angles.m.f. The method by which a sine wave may be deduced from a uniformly rotating rod or vector was indicated in Chapter 1.Circuit Theory 2. There is always a combination of impedances in series and/or parallel and it is necessary to be able to calculate the currents and voltages in the various portions and to assess what the combined effect will be. consider the simple cases of pure resistance.
values.m.22 to 1. 2. they may equally well represent the r.1 (a). OA will represent the voltage across the resistance. inductance and portions is the same. is in phase with the current.1 (6) and (c) respectively)." by OB which * It should be noted that.72 RADIO COMMUNICATION the voltage. the length of 01 being proportional to VjR as shown in Fig.2 containing The current flowing through the two V IL(u L —V / 'A """/ O Fig. 2.s. however. then 01 will be the vector of the current. the voltage on the resistance. The total that voltage across the whole circuit will obviously be the sum of the voltages across the two separate portions.24 the veotors were made to represent the maximum values of the sine waves. Inductive Circuit is leading by 90° on 01 (which is the same as saying the current is lagging 90° behind the voltage). 01 behind the voltage.* = = +*" +" if U Co) (A) (e) Fig. which is equal to IB and is in phase with 01. 1. while the current in a capacitance will be 01 VCco leading by 90° (Figs. V/Lco and will be lagging 90° Similarly for an inductance. 2. 2. These two voltages. Voltage and Current Vectors Addition of Vectors resistance. if 01 is the current vector. however. . whereas in Figs.2. are out of phase and hence must be added "vectorially. 2. The voltage on the inductance will be equal to I Lay and will be represented Consider now a circuit as in Fig. while that on the inductance is leading by 90°.1. Hence.
To the distance OA VB we have to add a distance at right angles equal to OB. For example a current cannot be added to a voltage. Thus in Fig.3. but only to another current. and also. Use of Scale It should be observed that if the current and voltage vectors have been drawn to scale. 1.5 .CIRCUIT THEORY 73 sum is OV. The vector 00 is obtained by drawing from A a line AG parallel and equal to OB and joining OC. simple geometry this vector representing . 2.2 will actually represent the resultant voltage across the circuit in magnitude as well as phase. According to the wellknown laws of rightangled triangles OA i In other words. OF in Fig. + (Lea)* = Z* and Z= VfB* + (Lo») 2 ] which was the result obtained mathematically in Section (page 45). Illustrating n Addition op Vectobs two sine waves. representing VL which we can do by drawing a line from A at right angles. 2. By This construction can be used to add any two vectors. This will terminate at V which could be reached directly by the line V hence V represents the voltage across R and L in series.3. if OA and OB represent o Fig. provided that they can legitimately be added together. if (IB)* + (I La)) = I*(R* + 3 (Leo)*) Z is the impedance of the circuit. 2. of course. of length AV = OB. + AV = OV* i 0Fa = Now. the voltage OV = IZ so that OV2 = I*Z* R* Hence. then the sum of the two will be a third sine wave whose vector is OC. This simple method enables any two vectors to be added together provided they are of the same frequency. . whatever the angle between them.
The resultant of these w>irfTnflnflr>> 91 — two depends on whether VL or V c is the greater.4. greater than VL and the resultant is V x All as shown. This effect. since = .1/Cco) 2 ] It will be observed that if Leo is made exactly equal to \fCto the effects of the inductance and capacitance will cancel out and the circuit will behave as if it was a pure resistance.IjCco) Z = Vt^ 2 + {Leo . Inductance and is therefore Capacitance X= and the impedance (Leo . VL is the voltage across the inductance. which contains capacitance as well as inductance and resistance. In addition to this there is the voltage on the resistance VR the voltage across the whole circuit being V.4.  inductance. Here VL is equal and opposite to Vc . .5. and hence the resultant will be the numerical difference Fig. 2. Resonance Consider next the circuit shown in Pig. 2. the effect of the capacitance overwhelming that of the . CmotriT Containing between them. w T! — — j A . the resultant of VR and Vx The net effect therefore in the circuit considered */ is that the current leads on the voltage by a small angle. and is considered further later. so that the only voltage left in the circuit is VB the voltage on the resistance. only holds for the particular frequency which makes Lm is called resonance. Here if J is the current. It will be seen that the voltage on the capacitance is directly opposite to that on the inductance. which 1/Cto. and V c is the voltage across the capacitance. In this case V c is assumed V i. Actually. The combined reactance Resistance. The vector diagram of a resonant circuit is given in Fig. . 2.74 RADIO COMMUNICATION Moreover it will vector OV and the tan<£ be seen that the angle ^ between the voltage current vector 01 is given by while = AV/OA = ILw/IB = Lco/R = X\R cos $ = OAjOV = IRjIZ = R\Z.
7 (b) shows the two currents and the resultant current I with their relation to the voltage V.CIRCUIT THEORY 75 the resistance is part of the inductance. out of phase. Circuits of Resonant Circuit Parallel Circuits A different type of circuit is shown in Fig. which shows an inductance and capacitance in parallel. 2. In the bottom path the voltage is the resultant of the voltage on the resistive portion R it which is in phase with 72 an d that on the capacitor G.) These two resultant voltages. The primary fact here is that the voltages across the two portions are the same. 2. 2. and let the currents in the top and bottom parts of the circuit be I and 7 x 2 respectively. Vectob Diagbam Pig. as has been explained. is in phase with I. 2. Obviously a resonance condition could be . the voltage on the inductance will be VLB as shown. but the currents will be different and. The total current will be the (vector) sum of the currents in the two paths. Parai^lei.6. Now assume that the voltage across the whole is V. Fig. and hence the voltage will lag behind the current. must be the same. The total voltage V will therefore be the resultant of these two. Fig.5. \*% Fig. The voltage across the resistance Bt will be in phase with Ilt while the voltage across L will be 90° ahead. but the resultant of VLB and V c is VR which .6. Consider the top path first.7 (a) shows the voltage V and the currents Ix and 7 2 > lagging and leading respectively. It will be seen that I leads on V by a small angle. of course. 2. The component voltages across the resistances and the inductance and capacitance are also indicated. which is 90° behind. (The current in a capacitance leads on the voltage.
The resultant is then Fig. compared with the {. Vector Diagrams or Fig. 2. impedances the line joining the origin to the extremity.g. Two such may be added graphically by drawing lines of length proportional to the scalar value at the correct angle and constructing the usual parallelogram. as shown in series = =X The impedance would then be written 80/46°. Impedances are therefore sometimes specified in terms of scalar magnitude and phase angle. Thus. which may be measured for length and angle as before. in the case shown the resultant is 985/W. The voltage would lag behind the current so that the voltage vector will lie in the fourth quadrant. viz. e.76 RADIO COMMUNICATION obtained here. in Fig. The scalar impedance of the two in the phase 80 ohms and since B would be 566V2 angle would be 46°. 2. Leo R = in radio engineering and is therefore discussed in detail on page 90. For example. if we wish to add the vectors 80/46° and 40 /30° we draw a line equal and parallel to the 40 /30° vector from the end of the first vector. however.B% is small If the total resistance t reactances the condition for resonance is the same as for the series I/Gcd.7. . a special sign being used to indicate the quadrant in which the voltage vector lies. and taking the square root. consider a resistance of 566 ohms in series with a capacitor of reactance also 666 ohms. (a) (b) Fig.6 Circuit Vector Notation We have seen that the numerical or "scalar" value ofany impedance can be evaluated by adding together the squares of the resistance and reactance. This does not completely define the impedance. 2. 2. but the subject is one of particular importance case. If one of the two quantities has to be subtracted from the other.8. the current vector being assumed horizontal and to the right of the origin.8. the values of L and G being adjusted so that J is in phase with V. for it does not specify the phase angle between the resistive and reactive components.
Illustrating 4 tn Quadrant j\s* Nomenclature by Quadrants The "i" Operator An alternative method. Illustrating Concept ory=v'(— 1) in a given direction such as OA ~ l in Rg. 2. and "imaginary" If we have a quantity a which can be represented by a vector B 1" C+ j'<7~ *A jefo D Fig.CIRCUIT THEORY 77 then one of the vectors must be reversed as shown dotted in Fig. — 1 we represent it graphically by a vector in the opposite rw OG = —a we multiply this . then if quantity by direction.9. 2.9.8 and the resultant again worked out. which is in many ways involves the division of the expression into "real" parts. 2. 2nd Quadrant 1st Quadrant A\S1 Datum lint 3rd Quadrant as*\ Fig.8. more convenient. 2.
that it can have a meaning if we regard it as representing a quantity at right angles to the normal. Then its motion will be 2 2 represented by j x j j and OG = j a. That across the inductance is j Lai. The voltage across the resistance is IB as before. when using this form of notation it is possible to add together all the resistive components and all the reactive components. we use the positive sign. this position? How are we to represent the vector when it is in Let us represent this motion by j. Now. It will be clear. The square root of a negative quantity is therefore imaginary and is frequently spoken of as such.4 can be treated in the same way. is always positive. If we are depicting a vector which is leading. where it is at right angles to the to original position. This is particularly convenient in electrical engineering calculations where the voltages and currents in reactive components are actually in the From at right angles. therefore. Then the voltage reactive part of the vector developed across the whole circuit is IB + IjLoj = I(B + jLco) — j~. while that across the capacitance is minus sign being used because the reactive part is below the datum line. But we have seen that position OG = —a. \/(— 1). The voltage on the inductance is ILco. . 2. however. positive or negative. Thus the vector 80/46° in Fig. The voltage on the resistance is IB. namely that the total impedance is made up of resistance B and the reactance (La — \fCw). form 566 The minus sign is used because the — is below the datum line. 2. and therefore we can rewrite the expression V = I[B + j(Lm  l/Cm)] It should be noted that this gives us exactly the same result as we obtained with the other method. Consider the case of a resistance in series with an inductance. so that the vector in the OB is ja. With this notation. so that the reactive part will be above the datum line. so that j 2 = = — 1. we have a simple and convenient method of representing impedances both in magnitude and phase. the The circuit shown in Fig.8 would be written j'566.78 RADIO COMMUNICATION Let us suppose that the vector has rotated from the position OA OG through the position OB. and j = . Now suppose we rotate the vector a further 90° so that it comes into the position OG. at right angles to the resistive component. the usual mathematical standpoint \/{— 1) has no meaning since the square of any quantity.
scalar value of a vector quantity can be evaluated by simple The expression 912 j 366 represents two vectors at — right angles and we can add them together by adding the squares root. the scalar value of the vector. let us multiply (a + jb) by (c + jd): (« + fi){c + jd) = oc + jbc + jod + fbd But j = — 1 so that the expression — (ac — bd) + j (be + ad).CIRCUIT THEORY 79 Moreover the combined effect of several impedances may readily be assessed by separating the resistive and reactive components as just mentioned. to note that before converting a vector expression into scalar form all the reactive terms must be segregated from the resistive terms. applying the ordinary laws of algebra and remembering that j = V — 1. For example. is V912 2 which + ( 366) 2 = 985 is the same result as before. and multiply both top This leaves the value of the expression is 2 2 2 2 2 unaltered. however. in Fig. For example.8 would be written 566 J566 resultant would then be 912 j 366. The — — Modulus The geometry. 2 Thus the quantity 2 ]. ) + jb) + jc is not [a + j(b +c)] = 2 2 2 Rationalizing Suppose we wish to add two vector quantities a Proceed as follows: „ a. ^c+jd X (" + jb)(c + jd) + c+jd c+jd 1 _ We now and bottom by c (ac — bd) + j(bc + ad) + 1 resort to a mathematical trick — jd. 2. + jb and c+jd ^ Ja j. but the denominator (c +jd)(c —jd) = c —j d = c + d . Vectors may be multiplied or divided by the use of thiaj notation equally well. is called. 2 It is important. the two impedances 80/45° and 40/30° and 346 + j20. (a equal to Vt(a 2 VT« 2 + (6 + c) 8 ] + b + c It must be written = vta + b + c + 26c]. and taking the square or the "modulus" as it Thus.
f. In any mesh or part of a network the total (instantaneous) the (algebraic) This means that the total e. which 2. (It may be noted that the first law does not apply if the circuit is not continuous. 2. With complex networks the calculations may sometimes be simplified by adopting certain transformations discussed shortly. The two laws are 1. is impossible.f. which is a resistive and reactive 2.10. but the basic principle is unaltered. are used. however.m. by the use of cyclic currents. The behaviour of such networks can be analysed by assigning values i lt i 2 etc. The algebraic sum of the (instantaneous) currents meeting at any point of a network is zero. The procedure may be illustrated in simple terms by considering the circuit of Fig. leaving a simple vector in the form A + jB. in relation to simple ohmic circuits. present is exactly sum of the back e. which can be completely evaluated both as regards magnitude and phase angle.m. but they apply universally to any type of circuit if the instantaneous values of current and e.m. Network Theory Most practical circuits contain several impedances in various combinations./(l/Ceo).m.80 RADIO COMMUNICATION real quantity. If we assume a current t x in the mesh B^R^ . The numerator can thus be separated into components and each divided by ca + ^*. in working with these vector quantities that —j = 1/j.f.2. . is zero. l/jCco = —. In such a case the law has to be expressed in terms of the quantity e.f. The calculations may be simplified.s developed by the offset by product of the circuit impedances and the current flowing in them. For example. to the currents in selected portions or meshes and then expressing the relationships between these currents by the application of Kirchhoff 's laws. We can assign values i a ib and ie to the currents in Bv B s and B a respectively. of electricity.) By the application of these laws the performance of any circuit may be analysed to determine the current (or voltage) or the impedance at any part of the network. but the second law still applies. as in the charging of a capacitor from a battery. These laws were stated on page 10 . This is an obvious requirement in a closed circuit as otherwise there would be an accumulation of current at the point. It is convenient to remember. from which three equations could be written down and the values of the respective currents determined. Such arrangements are called netimrks and may be quite elaborate..
This is known as the principle of superposition. = h^i + = aB + i (*i (t s — H)B — iJBa 2 3 since ^ and ia now through B3 in opposite directions. networks in which the current in each element is proportional to the voltage across it.Ra/2 S = B B + B Ba + Ra R H i 3 ts t a s t.f. s distributed throughout the network is the sum of the currents (or voltages) which would exist at these points if the e. . be solved in the usual way.m. we only need two equations. This supplies two simultaneous equations which can. The current at any point (or the voltage between any two points) in a linear network due to the simultaneous action of a number of e.f. It does not apply to networks containing nonlinear elements such as an ironcored inductance or a rectifier or unilateral elements such as valves and transistors.s were acting separately. 2.m.e. but sources are therefore considered as delivering no any internal impedance which they possess must still be considered as part of the circuit. Note that the theorem only applies to linear networks.f. One of these is that the various sources of e.m. as shown. the full statement being as follows.f.CIRCUIT THEORY 81 and i a in the mesh Thus we can write B ZB S e .10. The remaining e. the calculations on more complex networks may often be simplified by the use of certain ancillary principles. in a network act independently of each other and hence the action of each may be considered separately. i. giving where = e(B + i* )/2 = e. after simplification. Simple T Netwobj: Principle of Superposition As just said.m. Fio.
e.f. known as the . Compensation Theorem network the calculations When it is desired to assess the effect of a change in some part of a may sometimes be simplified by use of the compensation theorem. %.m. ZR2 RfR/. (P) (A) Fig.f.m. words.82 RADIO COMMUNICATION Reciprocity Theorem reciprocity Another network theorem. and Ja Then clearly. by 11 simple calculation.m. we can assume that the current in the particular In other branch which is being modified is not changed.m. SEjlR. E in series with a network R R and R3 Let us assume these are all equal.f. As before. Illustrating Compensation Theorem all the other branches of the network is that which would be produced —IAZ in series with the changed branch. = 2EI3R = I = EJ3R 3 Suppose find that It R3 = Again by simple calculation we is reduced to R/2. Thus is for either position of e. theorem. acting at the second point will produce a similar current at the first certain current at point. equal to —IAZ. In Fig.f.m. states that if any source of e. if the impedance Z of a branch of a network carrying a current I is changed by AZ. i. which same and the circuit need only be analysed in one direction. 2. is the and current their ratio. an increase of EJ12R. This states that. called the transfer impedance. the effect on the currents in >/?. impedances which pass current equally in both directions in proportional fashion.11. by an e. the theorem is limited to networks containing linear bilateral impedances. . of A simple example will serve to show how this theorem can be applied. 2.f.m. then the same source of e. but that to compensate for the altered impedance we introduce an e.11(a) of three resistances v we have an e.f located at a given point produces a some other point. .
the i l X Z 2 condition for balance is that the products of the impedances of the opposite arms shall be equal. = i + B D = we get i]Zi = Z so that i^i^ = ZJZ^ Hence ZS \Z = Z fZ or Z Z = Z Z V . 2. Fig. part flowing through Z X Z 2 and the remainder through Z^Z3 There will also be some current through Zb unless the bridge is balanced. If the currents are i x and i 2 then i i (Z 1 i 2 (Z t Z3 ). R The additional current flowing in the network will be that due to the e. because the resistances were all made equal. but in a practical and less straightforward network it is often of value. Hence the additional current is E/6R.CIRCUIT THEORY 83 The compensation theorem circuit as in Fig. which in this instance is R/2.R/2 in series with R x and jR 2 in parallel. Z&) i Z and ix Z equal. .m.m.f. 2 3 making the potentials at the points % Hence ijiz Z %\Z % Substituting i 2 Z 3\ix for Z 2 in the first equation . current is of network often used is the "bridge" arrangeHere an e. which is the result already found. the use of the compensation theorem has not provided any advantage. This may flow via Zl or Zt depending on the conditions but we are mainly interested in the balance conditions. The feature of the network is 2.12. In this instance. Bridge Networks A particular form ment of Fig. 2.12. = + . .f. which will divide equally through Rx and The additional current in 2 Rr is thus E/12R. e acting on . divides. Bridge Network that if certain conditions are complied with the output current is zero and the bridge is said to be balanced.m. Clearly the current from the source of e. i2 2 In other words. Also if the bridge is balanced.f. is applied across two corners and taken from the other two. states that e 3 we could 3 rewrite the where = I Z = 1 R\2 = E/6.11 (*).
Similar bridges using impedance arms are used to measure reactances or impedances. The original network can be any fourterminal network.13 (6). FoUBTERMINAL NETWORK AND EQUIVALENT T NETWORK Let us apply this to the network of Pig.84 If the RADIO COMMUNICATION which arms are resistive we have the familiar Wheatstone Bridge Z x and 2 4 are fixed "ratio arms. and Z'oe The same three impedances for the equivalent T network may be written in terms of Z lt Z 2 and Zs so that. 2. " I 3/ 3o — AAA Zl Z2 . equating the two sets of impedances. we have . and the first step is to calculate the three impedances mentioned above. which we will call Zoe Zae ." Z2 an unknown resistance and Zs is varied until the bridge is balanced. 2. .z^z'oo] = ZM — Z = Z' . 2. 3. as for instance a bridge which will only balance at one particular frequency. Equivalent Networks It is often required to replace one form of network with another of simpler or more convenient form.12. Looking from the output end with the input opencircuited. Such transformations can be made by arranging that the impedances of the two networks are the same under the following three conditions: 1. Looking from the input end with the output opencircuited. . such as the bridge of Fig.13. while in special forms can be devised for particular requirements.13 (a) and derive the equivalent T network of Fig. when Z% can be calculated from the expression just derived. 2. 2. Looking from the input end with the output shortcircuited.z3 a oc . = z + %3 Z = Zt + Z2 Z I(Z2 + Zs Z'oe = ^2 + ^3 ^OC ae i 3 ) Rearranging these equations we obtain values of as under: Zlt Zit and Z z Z* Z^ Z% = VUZoc . o2 4o £ z3 (b) (a) FlQ.
Let the original star impedances be Z v Z 2 .. 2. 2. but the resulting expressions are more elaborate than for the equivalent T.— CIRCUIT THEORY 85 Sometimes one requires an equivalent ir network. Stab Network with Equivalent Mesh be replaced by equivalent impedances joining all the points. Then the mesh impedances between any two points is the product of the star impedances at those points multiplied by . Zs etc. can (<0 0*> Fia.15 the equivalent tt network of Fig.] = S(l/Z) For example. Za '. 2. in Fig.14 (a). 2.. impedance AC = Z Za[S(l/Z)] and the general impedance = Zm Z„[2(l/Z)] mesh impedance 1 Let us apply this to transform the T network of Fig.15 (b). [(1/ZJ + (1/Z 2 ) + (1/Z3 ) the 1 2 + . (a) into ylwHW^J Zt z2 A —twv— Zb i B V> c (a) Fio. StarMesh Transformation This is a transformation by which a network consisting of a number of impedances meeting at a point. AB = Z Z [S(1/Z)] Similarly. 2. It is often simpler to derive the equivalent T first and then deduce the equivalent it by use of the starmesh transformation. 2.14. as at Fig.14 (6). calculated as follows.15. This may be derived by applying the same basic rules. >2C c < b> l—c n T Network and Equivalent Network .
f. still with R 3 removed. The procedure will be clear if we consider a simple example. 2. operating on the impedance of the particular branch plus the impedance of the remainder of the network between the points in question. .e. the voltage across the points AB will be E= The impedance will eB 2 l(Bx +B 2) of the network at these points. Suppose we wish to know the current in R3 of Fig. . which states quite simply: The current in any branch of a network is that which would result from the application of an e. . Then by Thevenin's Theorem the current in B3 will be + = EI(B3 + Z)= eRaliBl + Ei) = eR * This is a speedy solution to a problem which. E equal to the e. + 1/Z3 X Z 2 + Z 2 Z3 + Z3 Z{\ Z1 Z2Z3 Zi%2 + Z2 Za + z3 zx ] **[« = Zt + Z3 2 \ z3 zx \z 2 Similarly Z b ==Z x and Ze =Z % + Z + Z Z \Z + Z + Z Z3\Z 2 X 3 a 2 X It may be more convenient to use admittance (= 1 /impedance) because with two conductors in parallel the admittances are simply added together whereas the impedance has to be evaluated by the usual expression 1/Z 1/Z1 1/Z2 If the admittances of the various branches are Ylt Y2 Y3 etc. = + . .). which is the mesh impedance between points A and 0. . Thevenin's There is Theorem one other transformation which is often of considerable use. i. (= 1/Z1( 1/Z2 1/ZS etc. though simple.m.10. be that of B1 and B 2 in parallel. B 1 B 2 /(B1 B 2 ) Z. available across the specified points when the branch is removed. is given by za = z1 z3 mz1 + 1/Z.86 RADIO COMMUNICATION Za . the mesh admittance is simply Ym Y„IT. This is Thevenin's Theorem.f. If R 3 is removed.(Y). would still require appreciably more calculation by ordinary methods.m.
As an example let us solve the expressions for the circuit of Fig. The equations. suitably rearranged. A X B2 —BA X 2 often occur in circuit Some i simplification can result if this is written in the form from A 1 B1 . i. 2.e. Matrices and Determinants Expressions of the form calculations.e = i R3 + (R + R = ) i 2 a x i2 2 3) . This is a simple example of a technique known as matrix algebra. B. whence x = (BXG 2 — C B )I(A B — B A X 2 The simplicity of this form will be apparent particularly if A.B . the specific expressions being called determinants. d which can be expanded by writing down the diagonal left to right and subtracting the diagonal from right to left.CIRCUIT THEORY 87 and in more complex networks the saving can be even more pronounced. the sequence ABCA maintained. The value of x (or y) can then be written down by treating items 1 and 3 (or 2 and 3) as algebraic fractions which can be rationalized by crossmultiplication. are ix (Bx + Rs . Thus BXCX BC 2 2 A X BX A 2B% X 2 X 2 ). Suppose we have two equations A xx + Bxy + Cx = A& + Biy + G = 2 In matrix notation. this can be written X \B1C1 \B2C2 is = y = 1 A 1 B1 A 2 B2 Note the symmetry of the middle term. Otherwise the sign will be wrong as explained later. and C are not simple numbers but complex expressions. It is a form of mathematical shorthand which often enables the solution of equations to be written down on sight.10.
) a 3) Whence *i H = (Bt e(B 2 + Ba B3 )(B2 + B3 + ) e(B2 ) B3* B B3 + X +B ^2^8 + ^s^i Similarly. eB. y *1 Ci <?2 l Ci B 2 B2 C3 — o2 c3 — A A A — A A ^i Bi B 2 a B3 A l B1 1 A 2 B2C A 3 B3C3 2 .88 RAO/0 COMMUNICATION Then —B3 (B2 +B — e (Bt 3) + B3 B 3 ) 1 (Hi + **) — B. let us take the equations sign. (#2 —B + ^. B2 G 2 X Note again the sequence of the second term to preserve the As an example. in a form involving a constant term D such that the equation = in each case. In the first two equations D — while in the third it is — Then e. BX B 2 + B2 B3 + B3 B1 which is the same as was obtained by Thevenin's Theorem. mentally. A ±x + B& + Gxz = A& + B$ + C# = A& + B& + Gaz = e We rewrite these. Thirdorder Determinants For a threevariable expression we obtain a thirdorder minant of the form deter A. B x Gt A 2 B2 C2 A 3 BB G3 This is handled by expanding minants. namely it into three secondorder deter B2G2 + A 2 B3G3 + B3C3 B1 C1 A B G.
AxBx AB X 2 2 2 2 any column (or row) by a constant multiplies the whole determinant by that factor. Interchanging two columns (or rows) of a determinant changes its sign. Thus if in the determinant originally cited on page 87 we change the columns. Thus (A t (A 2 + mB )B + mB )B 1 2 l 2 AxBx A 2 B2 which is often useful in effecting simplifications. the determinant determinants. the third A BX the remaining terms — A 2 B2 going out. Multiplying inter = B A . each constituent of any column or row comprises two terms may be expressed as the sum of two simple Thus (A x (A 2 + o )B + a )B 1 2 1 2 A B + A 2 B2 X X a2 B2 which again may be simply verified. If the constituents of any column or row are increased or diminished by equal multiples of another column or row the result is unaltered. simplify either directly or on expansion. If may easily verify. we get BxAx 2. . Thus mA B mA 2 B 1 1 2 AxBx A 2 B2 as the reader 3.A X B = . Similarly the first determinant can be expanded into Bi ca ca — +B 2 c* Ox —e + Ba Ci = —BjC2e + Bfitf Oa It is not practicable to discuss the subject further. 4. y.e e CIRCUIT THEORY which determines x. 89 and z. certain simple rules which may be stated. determinant resolves itself into X Usually the individual solutions will For example. There are. These are: 1. however.
2. The resistance represents the total resistance in the circuit. There is obviously some condition at which the voltages are not only in opposition but are also exactly equal. this means that the current in the circuit. Similarly. in the circuit is that developed across the resistance and the generator would only have to overcome this resistive voltage drop. and the capacitance as being lossfree. since it enables us to consider the generator.16. the two voltages are in direct opposition to one another.m. Let us assume that there is an alternating current of value / flowing through the circuit. and it was mentioned that this gives rise to a type of circuit having very wide applications.3. This circuit shows a an inductance.90 RADIO COMMUNICATION 2. 90° ahead of the current. the inductance. Tuned Circuits We in the first section of this chapter that inductive and capacitive reactances have opposite effects. Since the voltage across the inductance is exactly 90° ahead of the current and that across the capacitance is exactly 90° behind the current. saw resistance. Lumping all the resistances together in this manner is a convenient practice widely adopted in circuit design. the voltage across the capacitance is I/jwC or —jI/wC. 2. Sebies Resonant Circuit the losses in the capacitor (to which we shall refer later) and the internal resistance of the generator. Expressed in another way. and a capacitance. the minus sign indicating that this voltage is in the opposite direction to that across the inductance. all in series with a generator producing an alternating e. with the result that the total voltage across the inductance and capacitance together is zero. the resistance of the connecting leads.m.16. Under these conditions the only back e. Consider the circuit shown in Fig.f. made up of the resistance of the wire in the inductance coil. R VW II c Fig. for a given generator .f. The voltage developed across the inductance will be jcoLI.
i. As we get still further away from the resonant point the effect of the inductance becomes negligible and the circuit behaves almost in the same manner as a capacitance.CIRCUIT THEORY e. whereas in a superheterodyne receiver the inductances and capacitances in the i.f.. we make mL 1/eoO. the current which will flow is limited only by the resistance in the circuit since the inductive and capacitive reactances cancel out. . and both methods are used. so that the current merely falls off steadily with decreasing frequency. and the other is to vary the frequency.f. which may be rewritten in the form o) a LC 1. At resonance. and at a frequency only slightly removed from resonance the impedance is almost completely reactive. be limited only by the resistance in the being in fact ~ since (a>L E B+j(coL  l/caC) ~R E — l/coG) = 0. The impedance of the circuit is. the net reactance in quadrature with the resistance. = = = = Resonance Curves The behaviour of the plotted to obtain what circuit at different frequencies is may be as a resonance curve.m. amplifier are fixed and the frequency of the input signal is varied until the maximum response is obtained. will 91 circuit. the inductive reactance decreases but the capacitive reactance increases and the known two do not cancel. and as this is necessarily greater than the resistance alone the current will fall. One is to vary the value of the inductance or the capacitance. as we have seen. Let us suppose that we have a constant voltage but that the frequency is varied continuously over a region around the actual resonant point. The response falls off rapidly. Here it is the inductive reactance which becomes larger and the capacitive reactance which decreases. therefore. In the tuning circuit of a radio receiver the capacitance is usually varied until the condition of resonance is obtained (as indicated by the signal strength becoming a maximum). As we reduce the frequency. or both. A similar state of affairs occurs if we increase the frequency beyond the resonant point. Hence for resonance w l/\/(LC) and the resonant frequency is 1/2tt\/(LC) hertz.e. This condition is known as resonance and it clearly arises when we make the reactance of the inductance equal to the reactance of the capacitance. where L is in henrys and is in farads. / There are two ways of obtaining resonance in practice.
As the resistance is reduced.18.17. 2. as it were. and in fact if there were no resistance in the circuit the current would be infinite. the maximum current at the resonant point becomes larger. The sharpness or steepness of this curve is a property of considerable interest to the communication engineer.17.18. 2. approximately symmetrical in form.92 RADIO COMMUNICATION C co c2 Frequency Fig. 2. Frequency Fig. as shown in Fig. therefore. in which case the curves The c CD L I. is to increase the height of the peak. as in Fig. in fact. The effect of reduced resistance. —> Effect of Resistance on Resonance Cubve . *~ Resonance Cubve variation of response with frequency is. causing the curve to grow. In many cases it is more convenient to show the curves having the same maximum height. The skirts of the curve are determined by the type of circuit. 2. but the peak is controlled by the amount of resistance present.
At resonance we have seen that the reactances cancel The property round the circuit is out and the circuit adjusts itself so that l/wC)] / = EftR + j((oL  = EfR The voltage tor voltage coil E= across the inductance. In normal radio circuits its value is of the order of 100 to 300. having Q factors of 1. but it should be remembered that this apparent narrowing of the skirts of the curve only arises from the change of scale and that the real effect of reducing the resistance is to increase the peak value of the current. . In audiofrequency circuits the Q factor may be only 10 to 15.18 Redrawn on a Percentage Basis Magnification Factor of a resonant circuit whereby the current circulating a maximum at a given frequency has obvious applications to the tuning required in communication equipment. and hence the ratio of the voltage across the to that supplied by the generator is coL/R.: CIRCUIT THEORY 93 could be redrawn as shown in Fig.19. 2. and it will be clear that voltage developed either across the coil or the capacitor is many times greater than voltage across the generator. however. 2. is coLI. 2. The generaIB. 100 r requency Fig. This ratio is known as the Q factor.19. Sometimes. Curves or Fig.000 or more. one is more concerned with the voltage in the circuit. however. Circuits can be devised for special applications. but the usual order of Q is as stated above.
These two rules Actual current Current at resonance B B + j(a>L — B_ ljcoC) ~ B+ {llJa>C)(l . 2. the departure from resonance) in terms of a factor which involves the percentage of mistuning and the Q of the circuit. Obviously the greater the Q. merely particular cases of the general calculations regarding resonant circuits.e. and the phase angle is 635°.20.w*LC)  (1) . while the horizontal scale represents the amount of mistiming (i. as they are useful in dealing with resonant 1. and —> + tt/2 when coL ^> l/coC. so that = Q since It is = (I/CwyiB = 1/BCco = coLIR l/Cco= a>L at resonance. by making use of this Q factor. the mathematical expressions are given here. The vertical scale represents the ratio of the current at resonance to that at some frequency off resonance. <f> —> The phase angle in fact is arc tan (coL — l/coC^/i?. Phase Angle It will be clear that as the circuit departs from its resonant condition both the amplitude and the phase of the current will change.94 RADIO COMMUNICATION Since the voltage across the capacitor is equal and opposite to the voltage across the inductor at resonance it is clear that we should obtain the same answer if we calculated the Q factor in terms of the capacitance. of course.f. Such a curve is shown in Fig. and it will be seen that this result is obtained with the curve of Fig. the greater will be the ratio of the current at resonance to the current in a specified condition of mistime. At resonance the current is in phase with the e. are.20. The voltage across the capacitor IjCco.m. The may be quoted. 2. If the mistuning is 1/Q times the resonant frequency the current is reduced to 045 times the resonant value. For the sake of completeness. If the circuit is mistuned by an amount 1/2Q of the resonant frequency the current is reduced to 071 times the resonant value and the phase angle is 45°. 2. Below resonance the circuit is capacitive while above resonance it is inductive. possible to draw a universal resonance curve which applies to any circuit. following rules circuits. so that — it/2 when 1/coG ^> coL.
The Qd. — <w )/ft) 0l where l/jcoC and divide top and bottom by R. o CC 3 3 u _/nQ. curve of Fig. Then co = = <w (l + d). 2. the expression becomes = 1 x .20.20 is plotted in terms of a factor a u c C 10/ c V. TUO c w u u c3 01 06 o 8 c 04 2 20 15 10 06 05 10 15 20 00 = Fig. Q x C ycles off Resonance Resonant Frequency Universal Resonance Curve .co *LC(l + 1 2 <5) ] But l/i?Ctw = Q and co^LG = so that the expression simplifies to I __ jQ 1 (3) [~26S*\ 1+jQd m\ = This expression neglects any variation in the resistance of the circuit with frequency which will be small as long as d is small. (2) RG(oQ (l + d) 1.CIRCUIT THEORY 95 (co Now let the mistiming be represented by d w is If we now put —jjcoC for ///. [1 . 2. the resonant value.
will be the difference between the two.m. the essential difference being that the roles of the current and voltage are interchanged. 2. as quoted earlier. Thus.m.f. Here the coil and the capacitor are in parallel A Fig. 2. 7// = 1/(1 • • (4) 1/Q.f. The expression (3) then simplifies to I/I = (1 + d)IQd(2 + = 8) (numerically) This co/co may be we get = y.21. is the parallel resonant circuit in Fig.m.21. The current supplied by the generator therefore passes partly through the coil and partly through the capacitor.f. Q8 becomes^ 1. we have two currents.1) (5) " * This gives a negative answer when y 1. Parallel Resonant Circuit and the e.96 If 8 is RADIO COMMUNICATION small. while if 8 = i// = l/\/5 = 045. one lagging and the other leading. under the particular conditions which make these two . If 8 is appreciable. 2. Making these substiy = + — tutions h = y Q(y l)(y + = 1) y <2(y a . Parallel < Resonance The circuit of Fig. Clearly. is applied across the two. The current through the coil will lag 90° behind the voltage. Then y I written in terms of the actual frequency ratio 1 <S and 8 1. so that they are in opposition and the net current which has to be supplied by the e. This circuit is not markedly different from the other. as before. while the current through the capacitor will lead 90° ahead of the voltage.16 is called a series resonant circuit because the three elements are all in series with the source of e. the expression simplifies to + 2jQ8) = 1/V(1 + <W) It will be noted that if 3 = 1/2Q. however. circuit which is often encountered. indicating that the current changes sign when going through resonance. I/I = 1/ ^2.
As shown on page 98.ja>L)l(B* + . (b) Pabam^el Resonant Circuits having Resistance Here the presence of the resistance in series with the inductance causes the current in the coil to lag by an angle which is slightly less than 90°. as we have that if B just seen. which are the reciprocals of the impedances. The admittance of the capacitance is jcoC. R (a) Fig. and there is a small resistive component which has to be supplied by the generator. This. the network of Fig. Clearly the smaller the resistance. so is made zero the expression becomes infinite. The calculations on a parallel circuit such as this are more conveniently done in terms of the admittances. and therefore we can consider the circuits as of the form in Fig.22 (6). but in the majority of cases the resistance in the inductive arm is many times greater than the effective resistance due to losses in the capacitive arm. the larger does this expression become. 2. the inductive current does not entirely cancel out the capacitive current even at resonance. of course. In resistance in the coil and some loss in the represent by small resistances in series with each arm. Consequently. Rationawhile the admittance of the inductive arm is 1/(5 lizing by multiplying top and bottom by B —jmL.22 (6) behaves.22. at resonance. this becomes + (Bjcom^ + w ^) 2 The arms total admittance is the sum of the admittances of the £»*£*) two Y = jcoO + (B. is practice. 97 circuit and the has an infinite impedance. 2. 2. which we ideal case where there is no resistance. without being seriously in error. there is an some capacitor. jcoL).CIRCUIT THEORY currents equal. the net current is zero. as if it were a resistance of value L/CB. For a strictly accurate treatment it is necessary to show resistances in each arm.
at resonance. Parallel resonant circuits are often employed in circumstances where the effective circuit resistance is intentionally made large.e. in which case there will be two possible "resonant" conditions. cw 2 =» lfLC. but in practical circuits the second term B 2 /L 2 is usually negligible by comparison with 1/LO. = phase. Fig. The full lines represent the resonant condition for which co = \/{\jLG — B z /L 2 ) — 0865. The first is the is = condition for which the "feed" current supplied from the generator in phase with the e.f. for which.R* B + o> 2 £2 +j wC [ L B*+ co*L*\ is and the impedance = 1/F. in which case the difference is appreciable. If the Q of the circuit is below about 10 this additional term may be appreciable.m. i. The resultant current is in phase with the voltage. With the Substituting co* IjLC this reduces to values normally encountered in practice. as for the series case. The dotted lines show the conditions difference should be . that the term B 2 IL 2 in the expression for w is not necessarily negligible. however. which will occur at a slightly higher frequency given by to 2 1/LG but the current will not be quite in is — = . Normally these two conditions are indistinguishable.23 (a) shows the vector diagram for an arbitrary example in which V.98 RADIO COMMUNICATION Separating resistive and reactive terms Y= . Resonance occurs when the reactive term zero. 2. It should be noted. At resonance the effective impedance of the circuit is = 1/7= (fl 2 + co*L*)lB B+ LjCB. LJCB is very large (of the order of 10 s ohms) so that the first term may be omitted and the effective impedance at resonance (often called the dynamic resistance) simply L/CB. This condition is satisfied when eo 2 1/LG B 2jL 2 The second condition is that for minimum feed current (maximum impedance). Leo and l/Cco are taken as 1 and B = \. and has a value of 05 (= L/CB). in which case the expression becomes simply a> a l/LC. when wG which can be written = coLI(B* + 1 u?LF) i? 2 This will be seen to be slightly different from the series case. but the remembered.
and i. Vector Diagrams sob. HV(LC) Dynamic Resistance As already mentioned the resonant impedance L/CR is often dynamic resistance of the circuit. . This. as we shall see. Xn 10 = Q'8650).. (=0895) Fio. If this load is a parallelresonant circuit of the type just discussed it is clear that the impedance and hence the gain of the circuit would be a maximum at the resonant point and will fall off steeply on either side. is very useful in amplifier design.— CIRCUIT THEORY 99 is when w (0445) = l/\/(Z(7) = 1.20 to give the ratio of the actual circuit impedance to that at resonance. 2.f. and as the frequency is varied the impedance falls off rapidly in an exactly similar manner to the variation of current in called the a series resonant circuit. Most of the r.f. In m„ = Xj. circuits in communication equipment utilize resonant circuits of the type just described. 2.23 II (=1) (a). Indeed. In the audiofrequency portions the circuits are not usually tuned though some resonant action is often present. where the stage gain is dependent upon the impedance of the load. a Parallel Circuit with Appreciable Resistance (R = %XL ) Fullline vectors Dotted vectors —feed current = phase with voltage. This only applies at resonance. we can use the universal resonance curve of Kg. Here the resultant current slightly less and leads by 30°.
co*C*P*) Considering now the whole circuit. but for most purposes it is sufficient to assume that the effect of all the losses in the circuit is equivalent to a series resistance. There are various other causes of loss of energy.jwCP)(G . and writing P(l for 1 + o) C P 2 2 2 . 2. Sometimes the circuit is of such a nature that the loss can most easily be envisaged in the form of a resistance across the circuit. and then through G and P in Fig. a simple matter by applying the laws of a.100 RADIO COMMUNICATION Circuit Losses As explained in Chapter 1 it is not only the resistance of the conductor which absorbs energy in a circuit. Currents and Voltages in a Lossy Resonant Circuit across parallel. Z CP we have = P/(l +jcoCP) = P(l j(oCP)l(l 4. we can is the write i same so that.23 (6). It is. however.c. The current / supplied by the generator will flow through L and R. Consider the circuit of Fig. 2. A valve connected across a tuned circuit has this sort of effect. theory to convert parallel resistance into an equivalent series resistance. so that the total resistance in the circuit is made up of the ohmic resistance of the wire plus a small loss resistance due to the various causes already specified. Z=B + jmL + ZCP = R + jcoL + = [R + PjO] + ML .23 (6). if *„ and V Therefore ic = ljati = p P = i^jcoCP of these two currents.CP*/G] . The voltage C it and P are the respective currents. so that iv The current I is the sum ie I = + iP = C and P (l+jcoCP) is The impedance we have of in parallel Vjl and since V = »°„P. ij.
however.CPt/coWP*] = [S + LICF]+MLL] In other words the reactance vanishes and the effective resistance is increased —the normal tuning by L/CP—a simple effect result. The behaviour of the circuit can be deduced from first principles. It is necessary. We P If the circuit first place.c. At any instant we know. When this process has been completed another discharge will occur in the reverse direction from the first one and the whole cycle of events will be repeated. This field will not collapse immediately but will prolong the current in accordance with Lenz's Law. and in some way we charge the capacitor and then let it discharge through the circuit. and in modern transmission and reception the majority of problems will fall within this category. so that becomes simply co iC2 P2 for If also (o s LC = 1 we can write Z = [R + PIcoKJtpt] +ja)[L . The process will continue indefinitely. Natural Oscillations So far we have only considered resonance as a special case of ordinary a. is tuned the expression can be simplified. to be familiar with the behaviour of resonant circuits in the condition of free oscillation. In the any radiofrequency circuit coGP is usually much greater than unity. . theory. that the parallel resistance is equivalent to an increased series resistance. useful to remember. The rush of current through the coil will build up a magnetic field. inductance. by Kirchhoff 's second law. Current will continue to flow after the discharge is completed and the capacitor will charge up in the opposite direction. therefore.— CIRCUIT THEORY 101 see. that Ldifdt + Ri + qJC = Rdqjdt Rewriting this in terms of q gives Ld2q/dt2 + + q\G = . although the current at each successive oscillation will obviously grow weaker due to the dissipation of energy in the resistance of the circuit. while it also modifies the reactance of the circuit. just as a pendulum will ultimately come to rest due to air friction. Suppose we have a circuit comprising resistance. and capacitance in series.
the circuit ceases to oscillate as already stated. which clearly becomes larger as the resistance in the circuit increases.R /4Z = Decrement The frequency of the circuit.24. L and R are the inductance and resistance in the circuit. for the current becomes » — =jsmx. /S 1/&C) and sinh fi i^' If a 1/LC.24. 2. and = ^=J[ic _LC R*~ 4L*_ This current is of the form of an oscillation of gradually decreasing amplitude as shown in Fig. It of the oscillation is known as the natural frequency must not be confused with the expression for . so that the expression = \ e< R lZL * sin cot where E is the voltage to which the capacitor is charged. if R is such that 2 2 1/LC. /S becomes imaginary and equal On the other hand if a2 = > = vV — = < Amplitude ofcumeitt Fig. 2.102 RADIO COMMUNICATION is This a differential equation of which the solution is /»«) where a R/2L. t is the time since the discharge started. the current simply rising to a maximum and then falling to zero. In fact. but sinhja. Illustrating Decay of Current in Oscillatory Circuit to j\/[(llLC a2 )]. The decay of the current is controlled by the damping factor R/2L. /? is real and the circuit is nonoscillatory. . 1/LC.
the first maximum) is obtained by putting t ir\1m.e. is still used to indicate the effect of We say. and it is customary to quote the logarithm (to base e) of this quantity which is nR /C . * = 2L/R = The maximum current ^(LC). This is called critical damping and the discharge is said to be dead beat. is E^/iC/L).^ to imaXi (or strictly the ratio of any peak to the preceding peak in the same direction) is called the decrement of the circuit. It obviously equals e^"*'2'"'. which is the resonant frequency. so that the term in the brackets is The next peak obtained by putting = Snfeco. 1 Hence to obtain low decrement. The peak value of the current (i. resistance. so that the oscillation shall die away slowly. meaning that the effective resistance has been increased so that the circuit is not so sharply tuned as it was. and the expression reduces to l/2v\/LC. when the resistance is low. which means that the current rises to a maximum and then falls rapidly to zero without actually = reversing.f. may As we have seen.CIRCUIT THEORY parallel resonance. that certain forms of connection introduce damping into the circuit. but such methods are now obsolete. E nax *~ <oL The ratio of i. the second term becomes negligible. Then = = E coL Sin ir/2 g— R it 1 4: Leo H) may be omitted. giving t = 1. and is constant if the resistance is constant. for example. and is the same as / that obtained for maximum current in a series circuit with steady applied e. however.m. R must be small and the ratio GJL should be small also. which occurs when . the transition from nonoscillatory to oscillatory condition occurs when 5 2/4£ 2 1/LC. The term damping. however. Damping The earliest forms of radio communication were by means of damped waves of this sort. 103 As before.
. The basic action of a transformer was discussed in Section 1. There is a source of e. for the moment it is assumed that there is no capacitance present. Fig.f. 2. Simple Transformer the magnetic field from one links with the other. . 2.m. and . being.104 RADIO COMMUNICATION Energy in a Resonant Circuit The energy in a resonant circuit is partly electromagnetic and 2 i At the partly electrostatic. in the other. Coupled Circuits it is In many applications in communication engineering desired to transfer energy from one circuit to another. the energy is stored entirely in the capacitor. Let us assume that the current in the primary circuit is t lf and the current in the secondary circuit is i a . while a quarter of a cycle later. continues to change from one form to the other during the continuance of the oscillation.4. when the current is zero. where it was shown that. We require to know the currents and voltages in the two circuits. we assume arbitrary currents in the two circuits and write down the appropriate equations which we can then solve by reference to known conditions. there is said to be mutual inductance between them and any variation of the current in one coil will induce an e. The presence of the secondary circuit affects what happens in the primary and vice versa. connected across the inductance Lx which is "coupled" to a second inductance L 2 The resistances R x and R 2 represent the circuit resistances of the primary and secondary respectively. if two inductances are placed in such a position that l ^\ !** M Fio. Therefore. 2. at any instant.f. the two circuits often being quite separate from each other. The most common method of achieving this transfer is by use of a transformer.25 shows a simple transformer.4.m. \IA f JCv moment when the capacitor is completely discharged the energy is stored entirely in the inductor. Transformers. The technique of the treatment is interesting because the two circuits interact on one another.25.
)^ = e Therefore the effective primary impedance Zx since J 2 ' = e/n = R + jX + M*w*IZ x t 2 =— 1.J+4Xi ^. But Rationalizing this M»a>* ——— = Z2 we have M*m* R2 +jX2 if 2a> 2 R2 + jX 2 ~ Hence _ M*<»\R2 i? 22 2? x jX _ +Z 2) 2 2 x M 2 a> 2 "Z. while the reactance of the primary co 2 has been decreased by an amount M X 2 . R2 .jX == h+z^. 2. 2" (2?a _ jX * 2) ] Zx = ' + jX + (M*co*IZ 2 *)(R 2 jR2 . theory we can 'write h( R i i 2 + jx + 3 M <»H = e Z + jMwii = i) 2 where Z2 = R2 + jX 2 .^(JfW/Z.CIRCUIT THEORY 105 From ordinary a.c. In an exactly similar manner we can derive expressions showing the effect of the primary on the secondary and this gives us for the equivalent secondary impedance .X2J eo 2 This is a most interesting result which shows that the presence of the secondary circuit has increased the effective resistance of the primary by an amount circuit Jf 2 8 . Thus Substituting this in the i1 i2 = —jMcoiJZz first expression "we have (R1 +jXj) . These additional 2 amounts of resistance and reactance are said to be "reflected" from the secondary into the primary.
Here the windings are housed on a ferromagnetic core so that virtually all the primary flux links with the secondary and k approaches unity. The ratio e z je x is called the transformation ratio and.m. in which case only a part of the primary flux links with the secondary. namely (a) Loosecoupled transformers. where k 2 is = k*L ICR x . This reduces the effective value =M of k. These are usually (but not necessarily) air cored. 2. e x depending on the circuit conditions. A simpler circuit is that having only the secondary tuned as in Fig. (6) Closecoupled transformers. so that = M = k^(L Zx ' x are small by comparison.106 RADIO COMMUNICATION Transformation Ratio o>ix which may be more or less than The secondary e. But in radiofrequency transformers this simple relationship may not apply and there are. 2 Hence the effective primary impedance becomes called a coupledcircuit network. It should be noted that radiofrequency transformers are often wound on closed (ferrite) cores but the primary turns only form a small part of the primary circuit. is equal to the ratio of secondary to primary turns. e2 the primary e.f. two classes of transformers.m. but since B a and Z2 the secondary is tuned. Zx = ' i?i + co M IB + jX = <o il/ /jB t i 2 a x 2 approximately the coupling factor .26. TtnsnED Transformer When both windings are tuned the arrangement becomes what is which is discussed later. X = 2 = . other things being equal. It is customary therefore to tune one or both of the windings to the frequency of the signal.f. so that they behave like a loosecoupled system. in fact.26. 2. x L^j. obtain 4:C M Fig. Loosecoupled Transformers In a radiofrequency transformer the requirement is usually to maximum secondary current. The same basic formulae apply. and the coupling factor k is small. Such transformers are used in aerial circuits and interstage coupling networks. since both Rx and X Let us now write w 2 \JLZG.
2. Coupled Circuit very weak the resonance unless the coupling between the circuits curve is severely distorted. 2. is X Let us refer to the circuit of Fig. and this is sometimes done instead of using a separate primary winding. Hence we can write L 2 \L X w 2 . the equivalent primary impedance just deduced reduces . in fact. The primary current. where n is the ratio of the number of turns on the secondary compared with the primary. will not be a x maximum because of the additional resistance which has been X =X = ' . As a corollary. = M = . 2 2 Hence for co we can write WL^wH 2 k 2 L a2 a> 2 ln 2 In particular. as if the primary were tapped across part of the circuit.27 and let us assume that both primary and secondary circuits are tuned so that 0. any impedance connected across the input will be equivalent to an impedance » a/Jfc 2 times as great across the secondary. The effect is. to Zt ' = k LJCB = 2 2 (k^n^LJCB^) The apparent input impedance is thus k2 /n2 of that across the secondary. To a first approximation the inductances of the two windings on the transformer will be proportional to the square of the number of turns. which is L2 ICB 2 and indicates that the apparent impedance of the primary is some fraction of the effective impedance of the whole (tuned) secondary. and Fig. Coupled Circuits we If both the primary and secondary of a transformer are tuned find that the interaction of the two circuits is very marked. It is sometimes convenient to express the various formulae in terms of the transformation ratio. however.CIRCUIT THEORY This is 107 very similar to the dynamic resistance of the secondary.27. x 2 is also zero.
1000 Frequency (kHz) Fig. as shown in Fig. one on each side of the true resonant point. 2. so that the circuits are no longer tuned. Z z rises rapidly. is less than before.108 RADIO COMMUNICATION from the secondary. The actual additional . the impedance of the secondary. however.28. and with quite a small coupling the peaks may be pronounced and widely separated. 2. If the coupling is increased the same effect begins to occur in the secondary current so that we have double humps both in the primary and the secondary resonance curves. Primary and Secondary Resonance Curves for Varying Degrees ov Coupling primary increase. reflected into the circuit resistance will be (M 2(o 2 IZ22 )E2 the frequency. This allows the primary current to close to the resonant point this effect and at frequencies will more than offset the increased impedance of the primary itself. The result therefore is that the primary current shows two maxima. . with the result that the equivalent resistance reflected into the If we vary .28.
Beyond this value double humps appear in the secondary current. when the two terms X =RR 2. can never exceed the value with critical coupling. simply Selectivity It is convenient to express At frequencies away from resonance the reactances are not zero. them in terms of the factor a = Qd. even at these peaks. Moreover. If = M/y/iL^) = Mco/Vi^co L2a>) Mat = V(^i^a)> h = llViQM = 1/Q if the circuits are identical. were applied to either circuit alone. Hence .m. < Zx = R (\ + J2Q 6) x 1 and Zz = R2 {1 + j2Q a d) . numerically. i. this becomes. + Z ti x 2i2 i t and i 2 are the primary and secondary two equations gives is 2 currents.f. X=jQRd(2 + 6)1(1 + d) ~j2QRd if d 1. which is the maximum possible. = —jefty/^Rt) (writing \/(#A) for Mco) R = R = R. Solving = jMcoeftZ Z + M eo At resonance X = X = and the expression simplifies to i = —jMa>el(R Rz + JPeo 2 t 2 ) . The maximum value of the secondary current is clearly —jMwefiR^ (making Rt R2 = M 2 co 2 ) If ej2R. in the denominator are a condition which is termed critical The coupling k factor . . the maximum value of the current. (7) This is a maximum equal so that M*ccP coupling. .e. It can be shown that up to this critical coupling the secondary resonance curve only shows one tuning point.27 we can e write and where these = Z —jMcoiz = jMmij. (6) l 2 2 2 x ) . and d is (co — w )/eo Then as was shown on page 95. x 2 onehalf the current which would flow if the e. . . where Q — Lco/R = 1/coCR. 2.CIRCUIT THEORY 109 Secondary Current From Fig.
I* was shown on page 96 that with a single circuit the response is 3dB down when d = 1/2Q. which is very nearly the square of the corresponding value (0316) for a single circuit. Now. This is assessed in terms of the bandwidth. as is to be expected since there are now two circuits in cascade. 2 . This is the "flattop" effect mentioned earlier. identical so that R t = = At resonance that at resonance becomes i 2 li = jMcoeHMW + . which is the total frequency spread between the points on the sides of the resonance curve at which the response falls by a specified amount. which corresponds to a reduction of l/\/2. Bandwidth In practice one to a is concerned with the ability of a circuit to respond band of frequencies around the resonant point in order to accommodate the sidebands which convey the intelligence in a modulated transmission. R 2 R. providing a bandwidth B =f /Q. though the bandwidth may be specified in terms of a smaller loss (e. d = (m — co )lco = (/ — /o)//o> on either side of the resonant point. . . . . more At twice the 3dB bandwidth i/i steeply than for a is 024 as against 045 for a single circuit. Hence the use of a doubletuned circuit increases the bandwidth for a given value oiQ by \/2 times. while at three times the 3dB width i/i is 011. 1 dB). as against the square with a single circuit. which requires Q<5 = l/V2 Thus d = 1IQV2.R (l +J2Q6)*] d = 0. so that the ratio of the current 2 2 off tune to = [MW + R critical i 2 \i (l+j2Qd)*]l(M*co* If we now apply coupling so that M . sion simplifies to = 1/(1 + 2jQd . (9) It will be noted that this involves the fourth power of <x (= Qd). (9) shows that for a doubletuned circuit the response falls tol/v'2ofthemaximumwhen4Q4<54 = 1. so that B = 2df Eqn.2QW) = 1/V(1 + 4Q*<5*) . we have *. and the bandwidth = 2df = V2folQ. Substituting these values in the expression for secondary current.g. as explained in Chapter 3.  Cascading Beyond this point the response falls off single circuit. Hence the bandwidth is twice (/ — /„).110 RADIO COMMUNICATION and assuming that the circuits are and Xx = X 2 = X. The usual criterion is a fall of 3 dB. 2 to 2 + R*) (8) = R the expres.
Q Values fob Specified Attenuations Attenuation (dB) Q Single circuit value Double circuit 3 15 1 folB 065/ /B 050/ /B 141/o/B l15/ /B /o/B Other Forms of Coupling The expressions so far developed apply to any form of circuit provided that the term is replaced by an equivalent reactance. 4. This is discussed more fully in Chapter 9.C .2. and if this is expressed in decibels the overall attenuation is the sum of the individual attenuations. amplifiers use a series of tuned stages and provided that there is no interaction (cf.1. but the circuit can be converted to the form of circuit (6) by using the transformations quoted in Section 2. (5) and (9) is the attenuation of the circuit for a given mistune. or 19 dB). Tabus 2. Because of this it may be necessary to calculate the appropriate Q value to provide less than 3dB attenuation at the stipulated bandwidth. Circuit (b) is a commoncapacitance coupled arrangement in Mm replaced by 1/Cco.29 shows four types of circuit in common use. Thus a single circuit at three times the 3dB width would have an attenuation of 316. (With a coupled circuit there is still a slight interaction so that the actual attenuation is 9. but Table 2. or 10 dB.: CIRCUIT THEORY 111 In practice r.f. G where : O'IG. Thus with three stages in cascade an overall attenuation of 3dB would require each stage to have an attenuation of only 1 dB. 2.1 shows the values of Q for attenuations of 1 and 15 dB. C=C s<74 + C3C + C. The denominator of eqns. Fig. page 413) the overall response is the product of those of the individual stages. An alternative form is shown in The coupling reactance here is not so easy to specify. is circuit at the following O'lC. Two such circuits in cascade would have an attenuation of 20 dB. If thia is done we arrive which Mm (c).
f. 2. The theory is simplified if we assume that both primary and secondary circuits are tuned to the same angular frequency co. e. (6) Commoncapacitance (c) Topcapacitance coupling. <«) » CO Fig. similar effects occur with certain important reservations. in a valve oscillator.29.m. Typical Bandpass Aerial Couplings coupling. («) Mutual coupling. If the coupling is weak. Let dco us then assume that the two frequencies of the system are co + and co — dco. The critical conditions to avoid this may best be arrived at by considering the frequencies at which the circuit will oscillate. As the coupling is increased the energy transfer increases to an optimum after which an unstable condition occurs in which the energy surges backwards and forwards from one circuit to the other Where the coupled producing a complex doublefrequency oscillation.g. (d) Mixed coupling Free Oscillation e. At the upper frequency X = L^co + dco) . circuits are not provided with a source of but are allowed to oscillate naturally.IjC^co + dco) = t LjC^co + dco)* .112 RADIO COMMUNICATION wm. the remainder being dissipated in primary losses.1 a^T+dco) . the energy in the primary is partially transferred to the secondary.
The resonant frequency rises accordingly and the current rapidly slides into the upper stable frequency corresponding to the positive value of dco. is unstable. M (/ = /„).2L M co )8co = i 2 2 This has three roots. . . Hence M*coa < RJLJLz M < (RJcoL^iL^) This requirement is equivalent to saying that the secondary voltamperes (i 2 2 £ 2 co) must be less than the primary voltamperes. co dco = co nearly. X = 2L da> X = 2L dco X = X — {M*co IZ *)X 1 2 2 ' t x circuit is selfoscillating it will choose 2 and if the 2 2 a frequency such that Xx = 0.«~w[^g] or If the circuit is to be monooscillatory the second two roots must be imaginary. ' „r 2 ^^+4^ = c M*co22L2dco ° i. Similarly if the frequency falls the Suppose / increases X X is circuit immediately takes up the lower value of oscillation. m Therefore . that if is greater than the critical value the frequency given by dco = 0. however. It should be noted. Closecoupled Transformers now come to the consideration of the second class of transformer. Here the arrangements are such that practically all the flux produced by the primary links with the secondary so that the coupling possible is We when the coils are wound This is only together on a core of magnetic virtually perfect. ' slightly. Similarly. stable conditions only obtain with one or other of these and two modes. the closecoupled type. Then 2 is no longer zero and t reduced. expression then simplifies to t 2 = and is small. SL^bco* + (2L1 R2i .CIRCUIT THEORY 113 f Now The LjCiCO since d<u 1. existence of three possible frequencies may come as a surprise to the reader who is accustomed to consider a coupled circuit as having only two modes of oscillation. which means that the expression under the root sign must be negative. which is The sometimes a more convenient way of stating the conditions. namely *• .e. The primary reactance .
does not appear in the formulae.114 RADIO COMMUNICATION material as discussed in Chapter 4 and even then some small leakage is inevitable though this can often be ignored.s are in direct proportion to the number of turns on the respective windings.30. 2. From basic theory This is But that since k is unity. Hence x e 2 /e x = wy/(L L^\a>L = =n ^/(LJLj) of the But the inductance of the windings is proportional to the square number of turns so that ^{L^LX ) = TJ^ = n. Hence «a/e i a fundamental relationship in a closedcore transformer. In such circumstances it is possible (and more convenient) to rewrite the various expressions in terms of the turns ratio or transformation ratio. + Ri+j{Q)L2 + x.f. which may not be known. 2. Let us now consider the effective impedance of the transformer with a load connected across the secondary as in Fig. We can write it = ejl^ca and k is x and e% = Mmi x Now M = ky/(L L x z) assumed to be unity.30 (6). Zl '=jcoL1 + M icoiIZ2 M2 = Lx La while Z2 — Bs co^L^s + j(coL + X a i z 2 ). 2. in which case the mutual inductance. *! ess* a a  <<8 & (A) s tjjj) CO v Fig. so Zx ' ~ = jcoL.) __ R slight j(aLvR z 2 — <oL X +j(coL + X t a) which by a rearrangement _ 3 mL S + JX ~ jcoLz + (Rz+jXJ i( * a) . namely that the primary and secondary e. Close coupled Transformer with Resistanceless Windings Fig.m.30 (a) represents a simple closecoupled transformer with no load across the secondary and in which the windings are assumed to have no resistance.
This means that some part of the flux produced by the primary does not link with the secondary.\n* >A + ZJn* This is the impedance of a load ZJn? in parallel with the primary inductance L x so that the equivalent circuit is as shown in Fig.31.Z. more detailed discussion will be given at the appropriate place. but as these differ in importance according to the circumstances in which the transformer is used. Leakage Inductance In a practical transformer the coupling is not perfect. but the condition may be satisfied (or . The small series inductance is called the leakage. 2. t and (B 2 + jX = 2) Zs .% L Fig. The ratio of the power developed . notably in Chapters 4 and 10. so that k is not unity though it may be nearly so (098 or even higher). R s feeds a load R L . circuit of Fig. this will not be the case. E having a source resistance maximum power in the load is obtained when RL = Rs In general. inductance and k % )Lv Strictly speaking the can be shown to be equal to (1 presence of leakage inductance affects the transformation ratio but — be ignored here.30 (c). The expression thus becomes _ jwLjZ. w^coii 1 + _ Z.f. 5<aLx . the approached) by introducing a suitable impedancechanging network between the source and the load. which also includes the resistance rx to represent the resistance of the primary winding.31. consisting simply of the primary inductance shunted by the reflected secondary load ZJn*. <.m. 2.31 in fact does not take into account all the factors. 2. the secondary load.CIRCUIT THEORY 115 Now L2 = n2 L . which has the same effect as a small r—•WWrfTffiV. 2. this may The simple Insertion Gain If an e. Transformer with Winding Resistance and Leakage separate inductance in series with a perfectlycoupled transformer as shown in Fig.
Another form of transfer function . the equivalent primary impedance B L 2 (neglecting losses) is BL jn which must equal Bs for correct matching.R^/m = (Bs + RL FI4n*B8* if BJn* = Bs This is true for any value of n. Transfer Function The ratio of the values of any particular function at the output and input of a network is called the transfer function. If this has a ratio n. as explained on page 627. s 2 2 2 ) t ) 2 2 2 t ) ra 2 2 t ) If the impedances are reactive. = . will not be a scalar quantity but a vector quantity of the form VJV^ \_0. The input voltage here is not the source e.000. If BL = 20.m. R s ) while 1^. . in general. 2 = 01 and A = 220 /(04 x 200 = 3025 as before. If some other form of network is introduced.200 /(40 X 200 = 3025. If one is interested in the relative voltages the transfer function is simply VmiIVin This. but the voltage actually developed across the input to the network. This is discussed further in Chapter 14 (page 625). Then the load power without the transformer . Thus if B s — 200 and B L = 2. . w = 10 and A = 2.amperes. Ej2 (if Rjnz In the example just given Vin Iin\n E/2nBs Hence the impedance transfer function is nBs If the matching is not correct and/or the impedances are reactive the values of V in and Imt must be calculated from the circuit constants. The most usual form of impedancechanging network is a trans' former. = PBL = E*BL/(BS + R L ? With the transformer in circuit the power is &Bj»HBa + RLln*f Hence the insertion gain A = (Bs + BL ln {Bs + .f. Use is also made in filter calculations oi&transfer coefficient which involves the ratio of the input and output volt. but the B s so condition for maximum power in the load is still that B L ' that the above calculations remain valid.116 RADIO COMMUNICATION in the load with this network inserted to that without it is called the insertion gain. is — = — = . indicating an insertion loss. such as an attenuator or filter. the current will be E\Z. A t may become fractional.  the impedance transfer which is the ratio of the input voltage to the output current.
Principles of Radio
Communication
essential requirements. First necessary to produce electromagnetic waves which, as explained in Chapter 1, are a particular kind of disturbance set up when electrons are caused to change their position rapidly. The first part of this chapter therefore deals with the mechanism by which these waves are produced and the manner in which they are propagated
it is
Radio communication involves three
through space.
The mere presence of these waves, however, does not communiany intelligence (apart from the fact of their presence) so that it is necessary to alter the waves in some way, in a predetermined manner which will convey the required information. We can, for example, start and stop the waves in accordance with the dots and dashes of the Morse code or, alternatively, we can produce variations
cate
corresponding to the sound waves of speech or music. This is called modulation and the second section of this chapter deals with the principal methods by which modulation can be achieved. Finally it is necessary at the distant point to provide equipment which will detect the presence of the waves and convert the minute energy in the waves into electric currents of practical magnitude. This receiving equipment, moreover, must not only detect the waves but must also convert the modulation into suitable variations of current so as to reproduce the original information in the required form. This process is known as detection or demodulation, and the third part of the chapter will deal with the basic principles of this
process.
3.1.
Production of Radio
Waves
An
which
as a disturbance from a sudden change in the condition of an electrical system. Sometimes this system is the atomic structure of a material which is disturbed by an increase of temperature, as in the case of
electromagnetic
arises
wave may be considered
117
118
light waves.
circuits
RADIO COMMUNICATION
Radio waves arise from disturbances in electrical which are of much greater physical dimensions than atomic structures. Hence the waves produced here are of much longer wavelength but the form of the disturbance is similar. A simple exposition of wave motion was given in Section 1.7,
but
we
shall
now
consider the process in greater detail, particularly
Fig. 3.1.
Lines or Force Radiating ibom an Electron
in respect of the electromagnetic waves used in radio communication. Let us first consider a single electron as in Fig. 3.1. This
electron has associated with in all directions
of force
it an electric field which surrounds it and which we usually represent as a series of lines radiating outwards in a straight line from the electron itself.
Fig. 3.2.
Kink Produced by Sudden Change of Position of the Electron
The intensity of the field is dependent on the relative concentration of these lines of force, which we saw in Section 1.3 to be inversely proportional to the square of the distance. Hence at a comparatively short distance the field strength becomes negligible. However, consider the effect on one of these imaginary fines of in Fig. 3.1., if the position of the electron is force, such as suddenly changed, as indicated in Fig. 3.2, the electric field cannot
AX
PRINCIPLES OF RADIO
COMMUNICATION
119
suddenly change its position. It will, in fact, take time for the information that the electron has changed its position to be communicated along the line of force and there will thus be a kink in the line. Before this kink the line of force radiates from the electron in its new position B. Beyond it, the line of force is still in its former condition corresponding to the original position A of the electron and the kink itself represents the change from one position to the other due to the sudden movement of the electron. This kink constitutes the electric wave. It travels outwards at a high, but not infinite velocity actually a velocity which is found to be the same for all electromagnetic waves, namely 3 X 10 8 m/sec, and the interesting fact is that, although the static field strength very quickly becomes negligible as the distance increases, the strength of the wave only falls off inversely as the distance so that the effect of the disturbance can be noticed at much greater distances.
—
Strength of Field
Let us examine this process more precisely. Consider an electron moving upwards with a velocity v. Suppose that for a short time dt the electron is subjected to an acceleration a, so that its velocity
Radius ct
Fig
Ili/ostbating Formation or
Radio Wave
becomes ».+ «&• Fig 3.3 represents the situation at a time t after the change of conditions. If the velocity had remained unchanged, the electron would have arrived at A where OA = vt but because
;
120
of the
RADIO COMMUNICATION
OB =
momentary acceleration it actually arrives at B, where + «.dt)t. Hence AB = aibt. However, as just said, the news of this changed position cannot
{v
be transmitted instantaneously but travels outwards with a velocity c, the velocity of light. Hence at the time t the field from the electron ct and will continue to will be undisturbed beyond a radius r radiate from A. On the other hand, within a slightly smaller radius c(t dt) the field radiates from the new position B. In between the two there is a sudden transition from one condition to the other, and it is this which constitutes the radio wave. Actually since the electric field is radiating in all directions the information regarding the change of state is contained within two dt), and the diagram rapidly expanding spheres of radii ct and c(t of Fig. 3.3 is a crosssection of these spheres. The portion we are interested in is the component DC, which can be resolved into two further components, one radial, represented by DE, and the other
—
—
—
tangential represented by EC. Now as was shown in Section 1.3 the field strength at a distance == qj^rs. This is the field strength /r s where r from a charge q is in a radial direction and hence the field strength along the line
E
E
DC
will
be
EJr 2 cos
cf>.
Resolving this radially and tangentially
we have
cos
<f>
Radial component
=
(^
is
/r 2 cos
c/>)
=E
/r 2
Hence the
radial
component
is
the same as for the steadystate
condition. There
now, however, a tangential component, given by
Tangential component
=
.
(E
dt
.
/r 2 cos
<f>)
sin
<f>
= (E lr
is
2
)
tan
<f>
Now CE
azimuth,"
= AB cos 6 = at
i.e.
if r is large,
DE
the "angle of the angle which AC makes with the horizontal, while is very nearly equal to the difference in radius of
cos 0, where 6
the two spheres
= ct — c{t — dt) — cdt.
tan<£
Hence
CE = ^=
t
ttt.dt. cos 6
j£
and
since r
=
ct,
so that
= r\c,
tan ^
= —
 J
cos
so that the tangential component
is
a,
E —
r*
•
or r
c*
•
cos 6
=E
r
va
;
.
cos
PRINCIPLES OF RADIO
It is this tangential
fact, constitutes
field
COMMUNICATION
is
it
121
component which
the electric wave, and strength of the wave is:
abnormal and which, in should be noted that the
(a) proportional to the acceleration a, so that the more rapidly the electrons are accelerated, or in other words the higher the frequency of the oscillation of the currents in the aerial, the greater is the field strength produced. (b) inversely proportional to the distance from the origin (as
opposed to the inverse square law of ordinary
(c)
;
electrostatics),
and
proportional to cos 6 i.e. it is greatest in the horizontal direction (assuming that the electron movement is vertical) and is zero in the vertical direction.
Aerials
such as to provide accomplished by feeding into the system an alternating current of high frequency usually many millions of hertz. Such a device is called an aerial, practical forms of which will be discussed later. With such a system carrying an alternating current the movement of the electrons in one direction is immediately followed by a corresponding reverse movement, backwards and forwards. Consequently if we take the crosssection of the whole of the space surrounding the aerial, and add together all the components of the field strength, we shall obtain a series of beanshaped loops of electric field travelling outwards and continually expanding in size with a velocity of 3 X 10 8 m/sec. This is represented in Fig. 3.4. The actual direction of the electric field at any point is at right angles to the line joining that point to the aerial.
this rapid electron
A practical form of radiator therefore must be
movement, which
is
—
Wavelength The strength of the field falls off as the as We have just seen. For the majority
angle of azimuth increases, of cases, only those waves which leave at a small angle to the horizontal are important. For small values of (up to 10° or so), cos 6 is practically unity. Hence near the ground level the field strength is uniform, so that the practical manifestation of a wireless wave is as shown in Fig. 3.5, which represents a succession of belts of electric field moving rapidly along the direction of travel. These belts are sinusoidal in form having a maximum intensity at one point. The strength then decreases to zero, reverses and reaches a maximum in the opposite direction and then reverses again so that we have alternate bands
122 of positive
RADIO COMMUNICATION
maximum
and negative field. The distance between two points of strength in the same direction is known as the wavelength
of the wave.
Fig. 3.4.
Form of Radiation from an Aerial
A A kkk A A
a A AA m A
>
a
u w
>
S.
mm
'
>
r
WavelengthFig. 3.5.
Form of Radiation Near Ground Level
Remembering that the disturbance travels outwards with a constant velocity it is clear that the more rapid the frequency of oscillation of the electrons generating the wave the closer will the
PRINCIPLES OF RADIO
COMMUNICATION
123
disturbances follow one upon the other. Hence the wavelength will be shorter and, in fact, the wavelength and frequency are related by the simple expression
A/=c
where A
is
the wavelength in metres,
in hertz,
/ is the frequency
c
=3
X
10 8 m/sec.
Magnetic Field Now, the movement of the electric field through space generates a magnetio field, and it is sometimes convenient to regard the disturbance from a magnetic point of view. The magnetic field produced will be at right angles to the electric field and also at right
angles to the direction of motion of the waves, so that it will be at right angles to the plane of the paper in Fig. 3.5, where it is repre
sented
by the
circles
crosses representing a field going into the paper
with the crosses and the dots in the centre, the and the dots a field
coming out of it.
be noted, however, that this magnetic field generated by the movement of the electric field and is, in fact, merely another way of regarding the same phenomenon. According to the convenience of the problem to be considered the wave may be regarded either as a moving belt of electric field or a moving belt of magnetic field, but not both. The condition is, indeed, quite different from that which exists in a circuit. In a simple resonant circuit the energy is first stored in the form of electrostatic energy in the capacitor, and one quarter cycle later is stored entirely in the form of magnetic energy in the coil. There is a phase lag of 90° between these two states and the energy is continually transformed from one form to the other. The conditions in an electromagnetic wave are essentially different from this, since the electric and magnetic fields are in phase.
is
It should particularly
Practical Aerials
Consider now the field strength produced length I carrying a current i.
I
by a
practical aerial of
The current
where v
is
i
•= qft — a = q It v
I
Hence v
the velocity of the electrons. (ljq)i and the acceleration a
=
= dvjdt = (Ijq)
124
RADIO COMMUNICATION
if
i
Now,
= I sin
cot,
a
= {llq)Ico cos mt
earlier
Substituting this in the expression for field strength deduced and writing q\4me for 9, we have
E
E.
=
Ilea
.
4nerc2
i
cos o cos mt
stitution
But we have seen that eo = 2tt/ and / = c/A. Making this suband inserting the appropriate values for c = 3 X 108 m/sec and e = l/(36rr X 109 ) we have (in free space)
Er =
1885
y cos
cos mt volts/metre
We have assumed that the current at any instant is the same throughout the whole length I of the aerial. In practice this is not true, and hence instead of the total length I we substitute a length h which is the length of a fictitious aerial in which the current is uniform throughout. This is known as the equivalent height of the aerial and is discussed in greater detail in Chapter 8 where various practical forms of aerial are considered. Moreover, in a practical aerial for long and medium waves it is customary to earth the midpoint. When this is done, it behaves like a free aerial of twice the actual (effective) height, as explained on page 363; and since we are interested only in the radiation at 1. The radiated field is then ground level, 0, and cos simply
=
=
Er =
Reception
377 Ih/Xr volts/metre
To
erected.
receive the signal at the distant point a receiving aerial is This is made similar in form to that at the transmitter
(though for long and medium waves it is not so large), and the passage across the aerial induces an e.m.f. equal to the product of the field strength and the effective height of the aerial. Thus, if at a given point the field strength of the wave is 10 microvolts/metre and the effective height of the aerial is 3 metres, the voltage induced will be 30 microvolts. This will produce a current which is made a maximum by tuning the aerial (i.e. adjusting the circuit to resonate at the frequency in question) and the resulting signal is then amplified as required. The technique is discussed in
electric field in its
detail in Section 3.3.
PRINCIPLES OF RADIO
COMMUNICATION
125
Propagation
These fundamental expressions refer to the generation of waves In practice, the transmitting aerial is located on the surface of the earth and because of the earth's curvature the waves will very soon leave the surface, as shown in Fig. 3.6, so that in the absence of other factors there would be no reception at a point B an appreciable distance away. The radiation from A is, of course, not confined to the direction at right angles to the transmitting aerial, so that there will be some radiation in the direction AB, but
in free space.
Fiq. 3.6.
Effect or Earth's Curvattjbe
this will
have to pass through the earth and its energy will be rapidly
absorbed.
Reception at considerable distances is, in fact, only possible because of reflection or refraction in the earth's atmosphere which causes the waves to return to the earth, as will be discussed shortly. There are, in fact, three main types of propagation, namely:
(a)
(b)
Ground waves, which travel over the surface of the earth, Sky waves, which are reflected back to the earth by the upper
atmosphere, Space waves, which travel in the space close to the earth for a limited distance.
(c)
Ground Waves
Consider first the ground wave. This would appear to be limited to the relatively short distance over which a line joining the transmitter and receiver does not pass through the earth to any serious
extent. In practice the range is rather more than this because of the phenomenon known as wave drag. The velocity with which an electromagnetic wave travels is proportional to l/y^e) whereat is the permeability and e is the permittivity = e er In free space e = e and the velocity is the
.
velocity of light,
c,
but if er is greater than unity, as
it is
with a solid
or liquid dielectric, the velocity is reduced. That portion of the wave which travels in the surface of the earth thus moves slower than the main part of the wave which causes the "feet" of the waves to drag
126
as
RADIO COMMUNICATION
in Fig. 3.7.
shown
is sufficient
The difference in velocity is only small but it to permit the waves to creep round the earth for an
appreciable distance.
earth, so that the field strength at a distance is less
however, absorbed from the wave by the losses in the than the value for free space, so that in fact
Energy
is,
Ed = AE
where
/d
Ed = E =
field
strength of ground
wave at a distance d from the
transmitter,
theoretical (freespace) field strength at unit distance
from the transmitter.
The value of A depends on the nature of the terrain, and also on the frequency. At low frequencies (up to about 1 MHz) the earth is mainly resistive and the losses are relatively small. As one would expect the losses are less over sea than over land because of the better conductivity of sea water. At higher frequencies, the earth behaves more as ». an inefficient dielectric and the attenuation is appreciably greater. There is extensive literature on the subject ^^^^.^^^_^ which cannot be discussed further here. It may
^° no*ed> however, that at distances exceeding 002A%, where X is the wavelength in use (in metres) and a is the conductivity (in S/cm), the factor A becomes practically proportional to 1/d, so that the received field strength is proportional to 1/d 2 and not to ljd as in the ideal case. For normal land a is of the order of 10 4 S/cm but may be as high as 45 X 10 2 for sea water.
Wave Dkag
Sky Waves
It will be clear that, because of the combined effect of the earth's curvature and losses, the range of the ground wave is limited. Transmission can be achieved over considerably greater distances, however, by means of the sky wave which is reflected from the upper atmosphere. If the waves strike this reflecting layer at too steep an angle they are not reflected, but below a certain critical angle, which varies with conditions as will be seen later, the waves are returned to earth, from which they are again reflected and can, therefore, travel onwards by successive reflections from the electrified layer and the earth's surface. At each reflection some energy is lost but
PRINCIPLES OF RADIO
this is usually small
COMMUNICATION
127
and the signal strength is still approximately subject to the inverse distance law. It is, indeed, quite common with short waves to detect signals which have been completely round the earth, and under some conditions a curious effect known as oneseventhsecond echo is obtained in which the normal signal is followed by another much fainter signal oneseventh of a second
This is the same signal which has travelled onwards completely round the earth and again operated the receiver, taking oneseventh
later.
Fig. 3.8. Iixtjstratinq Reflection
fbom Ionosphere
of a second in its transit and, of course, being much fainter due to the additional attenuation. As we have seen, the reflection is obtained as long as the angle of incidence is not too steep. There is, however, a blank patch over which the ground wave has ceased to operate and the reflected wave has not yet arrived because the angle at which the waves would have to reach the reflecting layer in order to come back within this specified distance would be too steep. This blank period is known as the skip distance. It may vary from a few hundred to several thousand miles depending upon the wavelength being used and the conditions at the time. The effect is illustrated in Fig. 3.8.
The Ionosphere
The reflections from the upper atmosphere arise because of the existence of layers of ionized gas and the region has, therefore, become known as the ionosphere. Its existence was first predicted by Kennelly and Heaviside independently in 1902, the height of the
suggested layer being of the order of 50 to 100 km above the earth's
surface.
As the height above the earth increases, the pressure of the gases constituting the atmosphere decreases. For a height of about 8 km we have the ordinary atmosphere or troposphere. Here we can
have varying degrees of temperature, pressure and humidity, with varying air currents and cloud formations. In general, the air
128
pressure
air is
is fairly
RADIO COMMUNICATION
high,
and except under thunder conditions the
not
electrified or ionized.
Above the known as the
atmosphere
troposphere there
stratosphere.
is
is a much more extensive region, In this there are no air currents and the at a more or less constant temperature, but at a
gradually decreasing pressure. The atmosphere, therefore, tends to separate itself into layers of constituent gases in order of density, which is the reason for the name stratosphere. Because of this the upper limits of this region contain light gases at a very low pressure and such gases are easily ionized. Under the influence of the ultraviolet rays in the sunlight these
gases in the upper regions are more or less permanently electrified, though the extent of the ionization depends upon the time of day. Experiments to determine the exact height of the ionosphere were undertaken by Sir Edward Appleton and Dr. Bamett, in 1924,* as a result of which it was shown that there are two layers of ionization, one at a height of around 110 to 130 km above the earth's surface
and the other
lower level
is
an average height of a little over 300 km. The as the E layer and the higher region as the F layer. These layers are much more well defined in darkness than in daylight. The presence of the sun's rays causes considerable ionization to take place and the electrified belt is more extensive and less well defined during the daytime. Indeed during daylight there is not only an increase in the ionization but also a penetration downwards nearer to the earth, resulting in the formation of an almost separate layer known as the D layer which is little more than 90 km up. Moreover, during daylight the F layer becomes very much more extended and appears to divide itself into two more or less well defined regions, known as the F x and F t layers, at heights of approximately 250 and 400 km.
at
known
At night the disappearance of the sun's rays causes the gases to recomhine into two more or less well defined layers, the E and F Therefore, the reflection which is layers as already explained. obtained is much more sharply defined at night than by day.
Mechanism of Reflection Let us now see what happens
into
if
an electromagnetic wave passes
such as the E or F layer. As already mentioned, the velocity of propagation of an electromagnetic wave is
an
electrified belt,
* These experiments were conducted by radiating short pulses of waves skywards, and noting the time taken for them to be returned a short distance away by reflection from the ionosphere. It was from this technique that radar
was evolved.
PRINCIPLES OF RADIO
COMMUNICATION
129
proportional to lJ\/(/ie), where p is the permeability and e is the permittivity of the medium. Ordinarily both these factors are unity, but when we reach an electrified medium the permittivity decreases so that the wave will move with an increased velocity. Now, the electric field in a wave is not confined to a small region but spreads over a considerable distance, particularly by the time it has reached a point some 100 km above the surface of the earth. Consequently the top of the wave reaches the electrified belt before the middle and bottom portions, and the top therefore begins to travel faster than the bottom so that a bending is produced as illustrated in Fig. 3.9, which causes the wave to be turned round
Fio. 3.9. Illustrating
Action of Ionized Layer
and directed back towards the earth again. The amount of bending produced depends upon the frequency of the wave, as also does the
attenuation or loss of energy in the bending process. The behaviour is calculable from ordinary optical laws, one of which, known as Snell's Law, states that the effective direction of travel at a point is such that
P
n
where n
is
sin
<f>
= sin
<f>
the refractive index of the medium and <f> and (f> are as shown in Fig. 3.9. Now, the refractive index, which is the square root of the relative permittivity, depends upon the degree of ionization, and can be expressed in the form
n
where /
is
=
V«r
=
\/(l
the frequency of the
+ 812V // wave and N is the number of free
2
2
)
electrons per cubic centimetre in the
medium.
sometimes called the point of
The top of the wave path,
Pm
,
130
reflection, occurs
RADIO COMMUNICATION
.
when = 90°, so that sin $ = 1, and n = sin ^ Hence the smaller the angle of incidence, the smaller is the refractive index required if the wave is to be reflected. With vertical incidence = 0), n must be zero, which occurs when/,, = 9N.
<f>
(<f>
If the frequency exceeds this value, the refractive index becomes imaginary and no reflection results. The transition occurs at a frequency of the order of 26 MHz (corresponding to a wavelength of 115 metres) and is accompanied by considerable absorption of energy because the time of mean free travel of the electrons in the ionized layer corresponds roughly with the period of oscillation of the wave. is less than unity. At angles of incidence less than 90°, sin Hence the critical frequency fc is greater than the value for vertical incidence, /„, and is, in fact, given by
<f>
fe
= A/cos
4>
= /„ sec
<f>
This is known as the secant law, from which for any given transcan be estimated) it is possible to determine mission (for which the maximum usable frequency (m.u.f.) with which satisfactory results can be expected. This depends, of course, on a knowledge of /„, which varies with time of day and year, but has been established by prolonged observations under normal conditions, from which suitable frequencies and angles of path can be determined to
<f>
specified requirements. Since the ionosphere is by no means uniform, either in height or properties, allowance has to be made in practice for a variety of secondary effects. For example, a longrange transmission may span regions of transition from light to darkness involving quite different ionospheric conditions, and it is usually necessary to change the frequency according to the time of day. These and other special
meet
requirements are discussed more fully in Chapter 13. Waves that are not sufficiently bent round to be returned to earth by the E layer will continue outwards into space until they encounter the F layer where they will receive further bending. The density of the electrons in the F layer is much less than in the E layer so that the additional bending produced is only slight. On the other hand, the extent of the F layer is rather greater than the E layer and the effect is that waves which would otherwise be lost have their reflection completed and are returned to the earth again. Within the range of usable frequencies the attenuation of the wave is very small so that the field strength is substantially that given by the expression for free space, and is inversely proportional to the
length of the path from the transmitter.
PRINCIPLES OF RADIO
COMMUNICATION
131
Fading
The lower surface of the ionosphere is not sharply defined and in consequence the plane of polarization of the wave may become
twisted. The electromagnetic waves which have so far been considered are vertically polarized, which means that the electric fields all lie substantially in the vertical plane. This need not necessarily be the case, however, because the electric fields can be in any direction at right angles to the direction of the propagation of the wave. If the transmitting aerial is arranged vertically then the electric fields in
wave will all be in the vertical plane and the receiving aerial must also be erected vertically. A receiving aerial arranged horizontally at right angles to the direction of travel of the wave will not receive any signal because there is no component of the electric
the
along the length of the aerial therefore, the electrons in the conductor cannot be set in motion. If the plane of polarization of the electric wave is changed, however, by reflection at the upper atmosphere, then its effect on the receiving aerial will be reduced, and if the plane of polarization is twisted through 90° so that the wave arrives horizontally polarized the receiving aerial will not respond at all. In practice some of the waves are twisted more than others and the conditions are continually changing, with the result that the wave in travelling down to earth again behaves as if its effective plane of polarization were gradually rotating. Consequently at one instant the receiving aerial will derive the full signal from the wave, while a short time later when the effective plane of polarization has rotated through 90° it will receive no signal at all. This gives rise to the effect known as fading whereby the strength of the received signal is not steady but continually varying. The fade may be slow or rapid, it may be complete or partial, and the effect will be familiar to every reader. Special aerial arrangements can be used to minimize the effect of fading, as discussed in Chapter 8, while the receiving equipment is usually provided with automatic gain control to provide still further compensation.
field
;
Space Waves The third type of propagation which applies principally to the very short wavelengths is that in which the received signal is derived partly by a direct transmission along the path between transmitter and receiver, and partly by a wave reflected from the ground. At low frequencies any waves travelling into the ground are absorbed
132
RADIO COMMUNICATION
but as the frequency increases an increasing proportion of the energy is reflected. The two paths are as shown in Fig. 3.10. The range of such waves is still limited to an approximately optical path, being in general about twice the length of the direct lineofsight path. The direct and reflected waves, however, will have slightly different path lengths and there will thus be a phase
Direct wave
Fig. 3.10.
Illustbating Space
Waves
difference
difference in
between them. From the geometry of the figure the path length d' is 2h^irfd. The phase difference is dependent on the ratio d'/X, being in fact 4nnhJi,\M. The sum of the two waves is proportional to the sine of the phase very nearly so that the angle, but since this is small sin 8
=
received field strength
is
E 4mhfrT _ kE 1' U ~M?
where EQ is the field strength in free space at unit distance, and k 4rrhfh r Hence the received field strength is proportional to 1/d 2 and not to 1/d. Reception can be obtained over much greater distances than the optical path under certain atmospheric conditions. The relative permittivity of the atmosphere near the ground is slightly greater than unity because of the presence of gas molecules (mainly water vapour, which has a high permittivity). The gas concentration,
=
.
however, decreases as the height increases, so that the permittivity decreases, producing a refraction of the rays similar to ionospheric
reflection,
which is the reason why the effective range is rather greater than the direct optical path. Normally the permittivity decreases linearly with increasing
height, but
when air masses of differing temperature and moisture content overlie each other, conditions often arise where, for a while, the permittivity decreases at a greater rate, which produces increased refraction and consequent increased range. The waves
PRINCIPLES OF RADIO
COMMUNICATION
133
can, in fact, follow the earth's curvature in a sort of duct, so long as the abnormal atmospheric conditions persist.
Classification of Radio
Waves
The range and behaviour of a radio wave is thus considerably dependent on the wavelength, which must be chosen to suit the
particular requirements.
The general classification and characteristics of radio waves modern use are given in Table 3.1 overleaf.
3.2.
in
Modulation
information can be communicated except by a change of conditions. For example, if a continuous wave were radiated the only information which could be received at a given point would be that the wave was present. In order to transmit some intelligence it is necessary to change the state in some way or other. The simplest way of doing this is to stop and restart the wave at appropriate intervals so that information can be transmitted by the use of some suitable code, but as we shall see, there are various other ways in which the conditions can be altered, such as, for example, altering the frequency. This change in the state or conditions of the wave is termed modulation and a wave which has been modified in this manner is called a modulated wave.
No
T/pes of Modulation The most obvious form of modulation is to start and stop the wave at appropriate intervals to produce the dots and dashes of
the Morse code. Alternatively, the amplitude may be varied in such a manner as to conform with the relatively slow vibrations of speech or music to transmit telephonic information. Both these types of modulation produce the effect by altering the general level or amplitude of the wave being transmitted: consequently this system is known as amplitude modulation. The wave of which the amplitude is varied is known as the carrier wave or often just the carrier, because its function is merely to carry from the transmitter to the receiver the changes of amplitude in which the real intelligence lies. When we have, at the receiver, made a note of the changes of amplitude, the carrier wave itself is no longer required and may be, and actually is, removed. One disadvantage of amplitude modulation is that the system does not respond readily to sudden changes. Just as a mechanical
CM fl e3£ £ M CO r( I * "5 sals' £ I* ? 8^3111.9 in <l £ nil ••§§ cp I I i £ HHHHOO HOHH a a lai ^4* CO CO O CO CO OOOOOO CQCOCOCOCOCO OOMOQCQ COO COO A A 14 En En 4 a WWBW •3 w (h PQPh 3 I I I I w liffll . ..II §111 ^s ogg&a arj age © d 3 *• 2.£1 .
so an oscillatory circuit will not change its frequency instantaneously.PRINCIPLES OF RADIO COMMUNICATION 135 flywheel possesses inertia which resists any sudden change in speed. provide the information as to the bearing and range of the target. but at slightly different frequencies in the keyup and keydown conditions. remaining constant for a short period and then falling to zero equally rapidly. These pulses of waves travel outwards from the aerial and are reflected from the target. This arose as a corollary to the development of pulse technique in radar. . the required intelligence is all contained in the change in frequency of signal and it is. The requirements are that the amplitude shall build up very rapidly to its value. however. and the time interval which elapses. This system. though more complex. has certain advantages. For radar it is necessary to generate very short bursts of waves which start and stop extremely rapidly and are separated by relatively long intervals. A somewhat similar form of modulation can be achieved by varying the phase of the carrier wave. and the direction from which the reception is obtained. With frequency modulation. Some early forms of transmitter therefore were arranged to radiate continuously. This is merely a variant of frequency modulation since any change in frequency is necessarily accompanied by a change of phase. Such pulses are a special form of amplitude modulation. but the technique has now been so developed that pulses can be generated lasting for only a fraction of a microsecond. This results in a very marked improvement in the signal/noise ratio. 144. therefore. The timeconstants of the circuits necessarily prevent an instantaneous rise and fall. Finally there is pulse modulation. Atmospheric disturbances and even locally generated "noise" in the receiver itself result in a fluctuation of the strength of the signal. At first sight it does not appear to be applicable to the transmission of speech or music but it will be c? tar that if the chang in frequency could be made proportional to the instantaneous amplitude of the speech we could modulate the frequency of the carrier wave instead of its amplitude. as is discussed on p. This method of communicating intelligence is known as frequency modiilation. the most important being the discrimination which it provides against random amplitude modulation. possible to pass the signals through a limiting stage immediately prior to the detector as a result of which the signal is of constant amplitude irrespective of any fluctuations which may have been introduced as a result of the interference. the receiver being tuned to discriminate between the "marking" and "spacing" waves.
These various forms of modulation will now be considered in detail." and this information is then translated into a code involving a series of pulses separated by varying time intervals. 3. (b) The changes in carrier amplitude must be faithful reproductions of the modulation frequency. as a result of the modulation. On the left is partial modulation where the amplitude of the carrier is altered but not completely reduced to zero. 3. It is clear that the carrier wave can be modulated to varying depths. it is found in practice that both the transmitter and receiver begin to introduce distortion of their own. (a) The modulation frequency must be only a small fraction of the carrier frequency. Indeed. The instantaneous value of the speech amplitude is "sampled.136 RADIO COMMUNICATION The development of this technique has led to the introduction of various forms of pulse modulation for the transmission of speech. Degrees of Amplitude Modulation different depths of modulation. due to nonlinearity of the circuit parameters.11 we have three Fig.11 represents an amplitudemodulated wave. when the depth of modulation . In the centre we have full modulation where the intensity is varied between zero and twice the normal. and on the right is a condition known as overmodulation where the maximum modulation is greater than the maximum amplitude of the carrier. which is not a permissible condition because it does not comply with (6) above.11. 3. Here the carrier wave is reduced to zero before the modulation wave has completed its cycle so that there is a blank portion during which no carrier is radiated. It will be two essential requirements. is just sufficient to reduce the carrier wave to zero the modulation is said to be complete and in Fig. Amplitude Modulation clear that there are Fig. When the variation in amplitude.
PRINCIPLES OF RADIO COMMUNICATION 137 exceeds about 80 per cent. These are known as sidebands. In addition we have two fresh oscillations. When we modulate the wave the peak value of the where carrier is subject to modulation. the modulation can be complete though even in this case it is often found convenient to leave the carrier oscillating faintly in the keyup position. These two frequencies are known as side frequencies and there is one such pair of side frequencies for every modulation frequency. The ratio of EJE is known as the depth of modulation and is often represented by the symbol m. Sidebands: Bandwidth The carrier wave can be represented by the expression e = E sin cot. E is the peak carrier amplitude and co is 2v times the carrier frequency. Consequently when the modulation is varying in frequency over a wide range there is a large number of these side frequencies extending over a range on either side of the carrier to an extent equal to the maximum modulation frequency. one at a frequency equal to the original carrier frequency minus the modulation frequency. One is the original carrier wave itself. and it is the presence of these sidebands which . With telegraph signals. a condition which we have seen is to be avoided. Clearly for full modulation E x must be equal to E If E x is less than E the modulation is partial while if it is greater than E we have overmodulation. and the other at the original carrier frequency plus the modulation frequency. Let us now expand this expression: . where we are concerned merely with starting and stopping the carrier. an additional variation at the frequency of the Hence the expression for an amplitudemodulated e wave is = (E +E 1 ainpt) sin cot Here p is the angular modulation frequency (= 2it times the actual modulation frequency) and the term Et represents the peak modulation. of course. so that it is usual to restrict the maximum modulation depth to between 80 per cent and 90 per cent. as a result of which the average modulation is only of the order of 30 per cent. c = (E + E = E sin == x sin^i) sin + cot cot cot Ex sin pt sin out X E sin + \E cos (co — p)t — \E X cos {oo + p)t It will be seen that the expression has been rewritten in the form of three separate sine waves.
necessary to allocate various transmissions into channels. in fact. and if the modulation is less than 100 per cent this discrepancy becomes even more noticeable. a telephony transmission operating on a frequency of 1 MHz and modulated with speech or music up to a maximum frequency of 5 kHz. The sidebands of the transmission will then extend over a range of 5 kHz above or below the carrier frequency so that the whole transmission will have a bandwidth of 10 kHz. which is the only part effective in communicating the intelligence. Consequently. Consider. that our modulation depth is 30 per cent so that m = 03.17). It is. 7. the effect will (in theory) be the same as if the carrier had been transmitted in the first place. If then the expression for the power may be written we write x in a Modulated Wave The power in a modulated sine wave + + = —E + PA E = mE P When oc # 2 (1 + m /2) 2 m = 1 the expression inside the brackets becomes 15 as we have just seen. If the carrier is then reintroduced at the receiving end. is only a little over 4 per cent of the total power. Suppose. referring to the expression deduced above it will be seen that the power in a modulated wave is proThus. however. these channels being wide enough in frequency to accommodate the whole range of the sidebands required.P?i) 2 E<? 100 per cent modulation {E x ) the total power is increased by 50 per cent but only onethird of the power is in the sidebands.138 RADIO COMMUNICATION makes it necessary to separate the frequencies of different transmitters by an appreciable amount so that there may be room in the frequency spectrum for the various communications existing simultaneously. for example. This can be achieved by the use of certain types of balanced modulator which are discussed in Chapter 7 (see Fig. with portional to E* (Py 2 (. Carrier Suppression It is clearly advantageous to remove the carrier wave.3. with considerable saving of power. and adjusted to an amplitude slightly greater than the peak sideband amplitude. twothirds being wasted in the carrier. Practical methods of producing the required modulation are discussed in Section 7. The expression in the brackets then becomes 1045 so that the power in the sidebands. Power is proportional to the square of the amplitude. is not practicable because . Complete carrier suppression. however.
(a) (<0 RkC . 3. 3.12. generate the carrier wave in the ordinary manner. Now. and reinforcing this at the receiving end. the vector 8t will be rotating relative to the carrier vector in the same direction. 3. 3.PRINCIPLES OF RADIO COMMUNICATION 139 the operation requires the reintroduced carrier to be of the same frequency and phase as that at the transmitter. position.12 (6) the condition of affairs is shown a fraction of a second later where the two side frequencies have each rotated 45°. If this carrier is modulated we have to add two further vectors each having an amplitude relative to the carrier proportional to the depth of modulation. kC R V) Fig. In Fig. The carrier frequency can be .12 is rotating but we are only looking at it in an instantaneous stationary . but the difficulty can be overcome by radiating a small proportion of the carrier only. in Fig. In effect.12. therefore. since one side frequency is higher than the carrier and the other is lower. They . These vectors are repre sented in the diagram by S x and S 2 Now. c S/AJAs. therefore. 3. the whole diagram in each of the positions shown in Fig. as will be discussed shortly. The resultant R therefore is of the same length as the carrier C. while the vector£2 will be rotating in the opposite direction. R . R c . which shows vector diagrams representing the carrier and the two sidebands. Their effects thus cancel each other out and the amplitude of the carrier is not altered.12 (a) the two side frequencies are in such a position that they are momentarily in opposition. This is impossible in practice. (A) Vector Diagram of Modulated Carrier with Sidebands nsr Phase represented by a vector C which will normally be considered as rotating and which will. R ffic . Effect of Carrier Phase Shift The reason for the phase requirement will be understood by reference to Fig.
though this is only true as long as the carrier is transmitted at its full amplitude. or the filtering arrangement may be such as to eliminate one of the sidebands only. leaving the carrier and the remaining sideband unaffected. Fig. In each of the figures the vector R represents the sum of the carrier G and the two side vectors Sx and S2 and it will be seen that in this case the effect of the side frequencies has very little influence on the amplitude of the resultant but merely causes it to swing in phase a little behind have. phase modulation but a negligible amount of amplitude modulation.13 is vector . 3. We Singlesideband Operation This is why it is essential to preserve the phase of the carrier and because this is too difficult it is not usual in practice to suppress the carrier completely. At (c) the two side frequencies have come into phase with the carrier and this is the condition of maximum amplitude. produced a little and a little ahead of the carrier. /f s^ (f) vs2 sSs)s2 R*~ ^ (h) N Vs. In other words the amplitude of the carrier has increased.13. This latter form of working is known as singlesideband working and relative to the modulation. results in . in fact. Singlesideband working also has the advantage that it can be received with a normal type of receiving circuit. 3. c< c ( c R s. some economy of power and also an appreciable economy of bandwidth since the total channel required is now only some 55 per cent of that required for a full doublesideband transmission. It may be reduced in amplitude to some small fraction of its normal value and then restored to its full value at the receiving end. but in this case the carrier rotated 90° relative to the modulation. (e) Fig. Vector Diagram of Modulated Carrier with Sidebands 90° out oir Phase shows the same sequence. c. The succeeding figures show the progressive conditions as the amplitude of the carrier falls to a minimum and resultant now produce B is then begins to increase again.140 RADIO COMMUNICATION a resultant which is added to the carrier so that the greater than the carrier vector.
sin coj. cos cojt is ±1. Then When cos e = A sin cot ± B cos cot if + B) and Now. The combined c signal may be written = A sin cot + B sin (co + cojt = A sin at + B [sin cos + cot to\t cojt cos cot sin coj] Periodically cos 1.4.5. Then (A e = A sin cot ± B sin cot = ± B) sin cot The amplitude of the {A corf signal will thus vary between (A — B). and sin cojt will pass through the values When = 0. With superheterodyne receivers this in more detail in Chapter 7. B\A is small (and equal to p). In practice these variations of amplitude and phase are usually of minor importance because is normally much less than A. etc. sin coj is ±1.h.v. then e = A sin (cot ± 2>7r/2) B approximately indicating a small (periodic) change of phase. 0. so that there is an amplitude modulation. as is explained in Section 9. Intermodulation should not be confused with crossmodulation. — 1.. including v.f. 0. radio and t. but it can also arise from interaction between a wanted signal and an undesired signal (usually of much smaller amplitude) operating on an adjacent frequency. use this vestigial sideband system. Intelmodulation Amplitude modulation can also result owing to interference between two oscillations of slightly differing frequencies. This effect is sometimes used deliberately in the technique known as heterodyning which is discussed in Section 9. = 0. In such circumstances the output from the subsequent detector stage is determined almost entirely by the modulation of the wanted signal.PRINCIPLES OF RAD/O COMMUNICATION 141 If the carrier is suppressed in part then special arrangements must be made at the receiver in order to amplify the carrier frequency more than the depth is side frequencies so that the normal modulation is automatically achieved so that many transmissions today. The method of suppressing the carrier and sidebands is considered restored. .
If we wish to use frequency modulation there are two factors which can be varied. The carrier would then be caused to swing above and below the mean frequency to the required extent at a rate depending upon the modulation frequency. i. say. These two possibilities can be utilized to convey the required intelligence. form of interference which arises if the amplifier and/or detector stages are not strictly linear.3 and 9. if we choose. in which the marking and spacing waves are transmore mitted with the same amplitude but different frequencies. With amplitude modulation this is quite a straightforward business for it is only necessary to vary the amplitude of the carrier wave in faithful reproduction of the variations in intensity and frequency A of the modulation itself. Frequency Modulation Let us now consider frequency modulation.6. For any intensity Anal/sis of Frequencymodulated Wave At first sight it might appear that the bandwidth of a frequency modulated wave was simply the maximum deviation of the frequency . as used for telegraph transmission.142 RADIO COMMUNICATION which is a different. The technique is discussed further in Sections 3. and sometimes more serious. for example. the frequency of the carrier will be made to swing between limits plus or minus 75 kHz about mean value.5. for full modulation. Reference has already been made to the simplest form. as discussed in Section 9. less than the maximum the deviation would be correspondingly reduced until with no modulation at all the carrier frequency would not be altered. The amount by which the frequency is changed is made proportional to the amplitude of the modulation so that. The reception of a frequency modulated wave requires a special form of receiver in which changes of frequency are converted into proportional changes of amplitude. both the strength and the pitch of the various tones are continually varying.e. One is the actual change of frequency from its mean value and the other is the rate at which the change is made. a maximum deviation of 75 kHz from the mean value. complicated arrangement is necessary to deal with speech and music because here the modulation intelligence to be transmitted is itself varying not only in amplitude but in frequency. To convey a modulation of 100 Hz. the carrier would be caused to vary between the prescribed limits 100 times a second.
co ± 2p. etc. + . a more significant parameter being the frequency deviation. The equation to a frequencymodulated wave can be written the form e =E sin {cot + m sin pt) where = 2w x carrier frequency p = 2tt X modulation frequency m = modulation index co deviation of carrier from mean value modulation frequency = A///m The modulation index The deviation A/ m can obviously vary in a since both its constituents complex manner. and wave contains components of co.] where A A lt A 2 . but this is not the case. are a particular form of constant. 1 {sin {co + p)t — sin (co — p)t} + 4 {sin {co + 2p)t + sin (co — 2p)t} + 2 . depending Thus the on the value of m. The equation can be expanded into the form . m m e = E[A sin cot + ^4 .f.. Hence it is not a convenient factor in design calculations. may be 20 or more. while the modulation frequency can have any value within the required (a. known as Bessel functions. . modulated wave.PRINCIPLES OF RADIO COMMUNICATION 143 on contains a either side of the carrier. co ± p. while with fuD modulation and a low modso that ulation frequency. Thus with a smallamplitude signal.) spectrum. co ± 3p so occupies a considerably greater bandwidth than an amplitudeIn practice it is not necessary to calculate the extent of these sidebands in detail because it can be shown that if sidefrequency amplitudes less than 1 per cent of the carrier are neglected the total bandwidth may be taken as 2( A/ fm ). Actually it number of highorder side frequencies which occupy a in band appreciably wider than the maximum deviation. as defined later. A/ may be less than fm is fractional. . is made proportional to the amplitude of the modulating signal and can thus vary from a (chosen) maximum down to a value approaching zero. depend on the modulation conditions.
e =E sin 0. and p — 2irfm fm being the modulation t . is then dd t /dt. by definition. by definition. by definition. . we can use m Equations for Frequency and Phase Modulation This distinction does not arise from any mysterious difference between the two systems. Clearly therefore either phase or frequency modulation. an important difference between the two systems. where 6 t is the total angular (phase) displacement at time t. In this case 6 is not constant. the modulation index of a frequencymodulated wave is A///TO This will be understood if we write the equation for the wave in the form . a frequencymodulated wave is one in which the instantaneous angular frequency co — co + Aa> cos pt = 27rA/ = hEm where Em is the maximum amplitude Aw modulating signal. with frequency modulation it is also inversely proportional to the modulation frequency. namely that whereas with phase modulation the modulation index is simply proportional to the amplitude of the modulating signal. The instantaneous angular frequency. so that f = fcof dt. = maximum. Aco is thus. Now. the same expression to denote and as explained in Chapter 7 it is often more convenient to modulate the phase rather than the frequency. cot.144 RADIO COMMUNICATION Phase Modulation As has already been mentioned. But co t = dd Hence e t fdt. but is itself a sine wave msinpt. a frequencymodulated wave is inherently phasemodulated and vice versa. but simply because. have seen that the We equation for a frequencymodulated wave e is =E sin (cot f m sin pt) = E sin (cot f 6) which to is E sin the expression for a sine wave with a phase angle 6 relative cot. t \ Q I sin pt I t . Aw < \ where = E sin Q = E sin / (o = E sin (coj + mf sin pt) m = Aco/p = A///„. however. There is.. a cosine term because with independent of fm The second term is the rate of increase of t is a t . where of the frequency.
is of the same form as the expression for a frequencymoduwave but the two will only be identical if mp = mf = A///OT Hence A/ = mjn = k'E^.PRINCIPLES OF RADIO COMMUNICATION US A phasemodulated wave. This is an oversimplification not correct.pt maximum amplitude of the modu where A= k'Em . a specific maxideviation is chosen (usually 75 kHz) to correspond with full modulation. so that t = cogt + A aia. In practice. . . = Actual deviation Actual modulation frequency . and for efficient operation this deviation should be several times the maximum modulation frequency. as stated above. . It is sometimes stated that a frequencymodulated integral of wave is the which is a phasemodulated wave. Deviation ratio Rated deviation Rated maximum modulation frequency This is a specific value of the modulation index. the deviation in a frequencymodulated mission is trans^ proportional to the amplitude of the modulating signal. . . is by definition a wave in which it is the phase angle 6 t and not the angular frequency. Hence if we wave becomes write mv for A. Em being the lating signal. from which is derived the . on the other hand. This is called the "rated" deviation. and the ratio of this to the maximum modulation frequency is called the mum _ . 6t = fcoj At FrequencyModulation Parameters As said earlier. . as before. Thus in a phasemodulated wave the frequency deviation is proportional both to the amplitude and frequency of the modulating This lated signal. the equation to a phasemodulated e =E sin (eo 2 + mv sinpi) . What is meant is that. corresponding to the rated maximum values of A/ and/m The actual deviation is then made proportional to the amplitude of the modulating signal. which is proportional to the modulation amplitude. . m.„ Modulation index .
3.14. which is arranged to be proportional to the instantaneous value of the modulating signal. the extent of the excursion being determined by the peak value of the modulation amplitude. One to transmit a succession of pulses of constant width having an amplitude controlled by the appropriate sample. The sampled information may be is utilized in several ways. Practical forms of frequencymodulated circuit are discussed in Chapter 7. of which the essential principle is illustrated in Fig. At the 12 V would produce a deviation of (12/20) X 75 maximum modulation frequency (15 kHz) the modulation index will then be 45/15 3. the deviation ratio would be 75/15 5. but this has no practical = = = significance. as in Fig. It is found that if the number of samples per second is greater than twice the highest frequency contained in the signal wave. the original wave can be successfully reconstituted from the succession of pulses. then a signal of 45 kHz.146 RADIO COMMUNICATION These definitions are as specified in BS. A second method is to transmit the pulses at a constant (fullcarrier) amplitude. Pulse Modulation Pulse modulation is a special form of amplitude modulation based on a sampling technique. 3. the bandwidth is 2(A/ + /m ). 204 (items 1402914032). This is known as pulseamplitude modulation. the concept of modulation depth does not apply. but to vary the duration in accordance with the . which also defines A/ as (Peak difference between instantaneous frequency of modulated wave and carrier frequency in a cycle of modulation Thus in a system having a rated deviation of 75 kHz and a maxi mum modulation frequency of 15 kHz. The frequency then swings on either side of the mean value at a rate determined by the modulation frequency. The instantaneous value of the wave is sampled at regular intervals and this information is then used to modulate the carrier. They utilize special forms of oscillator in which the frequency (or phase) is made to depend on the amplitude of a control voltage. say. As said earlier. If the maximum modulation amplitude is. Its equivalent would be the ratio of actual/rated deviation.14 (b). Since the amplitude of a frequencymodulated wave is constant. 20 V (which would be arranged to produce a deviation of 75 kHz).
14 (c). and is illustrated in Fig. modification of this method is to derive from the variablewidth pulses a series of short pulses of constant amplitude and width whose A (b) L (c) (d) Fio. . as shown in Fig. necessitates a wide bandwidth (of several MHz) so that the system is mainly used at ultrahigh positioning frequencies.14 (d). Pulse Modulation wave (a) Original (6) (c) Pulseamplitude modulation Pulsewidth modulation (d) Pulseposition modulation position is controlled by the initial by the width of the original pulses (and hence sampling information).PRINCIPLES OF RADIO COMMUNICATION is 147 sample information. Provided that the received pulse is sufficiently large. variations in amplitude do not affect the performance to any serious extent. Pulsewidth and pulseposition modulation possess the same advantages as frequency modulation in signal/noise ratio. 3. This may be achieved by a modified multicalled pulsewidth vibrator circuit as described in Chapter 17. This modulation. The steep wavefront of the pulses. however. The derivation of this variable is also discussed in Chapter 17. 3.14. This is called pulseposition modulation. 3.
a 30 per cent signal would be rated as 2 (286 per cent) and so on. to obviate any risk of interference. is extensively used to increase the amount of information which can be transmitted on a given carrier. which becomes less as the number of standard levels is increased. . and hence. it is possible to transmit a further series of pulses in these intervals. being of the order of a few microvolts.f. and so on. there is still 90 fiseo available in which five or six additional channels could be accommodated with a reasonable margin of separation. This type of coded information is not only economical but is in a suitable form for the operation of computers. such an aerial will respond equally well to signals from any source and it is therefore necessary to be able to select the wanted signal and reject any others. 15 ^asec. This in itself is no advantage. apart from any special directional or other properties. which is varied by the modulation to 10 /jsec. They can be used as a convenient method of transmitting intelligence over a land line. as explained in Appendix 3.s in the wire as they pass. This introduces a small error. which is known as multiplexing. Hence in the 3.s are normally very small. but by expressing the 8 possible values in a binary code only 3 digits are required. since the pulserepetition rate is constant.m. Reception For the reception of radio waves it is first necessary to provide an aerial which consists essentially of a length of wire so arranged that the electric fields in the travelling electric wave induce e. These e.148 RADIO COMMUNICATION Multiplexing With pulse modulation there can be relatively large idle periods between the pulses.m. Thus with a pulse rate of 10 kHz the time interval is 100 /isec. Here the input is sampled as before but the relative amplitudes are then classified in relation to a series of standard amplitudes. If the basic pulse width is 1 /isec. and this process. It may be noted that pulse modulation techniques are not confined to the modulation of a radiofrequency carrier. Thus if eight standards are chosen from 0100 per cent.f. say. The process is A called quantization. Pulsecode Modulation development of pulse modulation is what is known as pulsecode modulation. while a sevendigit code will handle 128 standard amplitudes. moreover.3. each 143 per cent apart. It is only necessary to delay the second series by. example chosen the information would be given by one of 8 possible amplitudes.
because of the resonant magnification. as will be seen from Fig.c.m.. current plus an a. 3. This is the basic process of detection for an amplitudemodulated wave. Now the mean value of a modulated carrier wave. however.m. Detector Circuit the signal which is done by passing the current through some form of nonlinear element which conducts more readily in one direction than the other. is at the frequency of the carrier wave and it is now necessary to extract from it the intelligence contained in the modulation. but if the signal is rectified so that the negative halfcycles are suppressed as shown in Fig. in fact. It is. is then applied to the detector circuit consisting of a rectifier in series with a telephone receiver.c.f.15.f. The picked up on the aerial and the induced e. amplifier.m. Fig. The most practical method of achieving this is to Fig. In practice the telephones would be incoming signal is .11. This signal.f.16 it will be seen that the mean value is no longer zero but has a positive value. of practicable magnitude (usually of the order of a volt) free of interfering signals. current corresponding exactly with the modulation.f.f. is transferred to a tuned secondary circuit which develops across the capacitor a substantially larger e.PRINCIPLES OF RADIO COMMUNICATION 149 This is accomplished by tuning the system to the particular frequency required by the use of suitable resonant circuits. This e. 3. is zero irrespective of the modulation. Except in the simplest cases. equivalent to a steady d. 3. a single circuit is insufficient and the signal is usually magnified by means of amplifiers which are themselves coupled by further tuned circuits so that a selective discrimination is obtained resulting in an e. Such a device is called a detector or rectify demodulator. amplified if necessary in a timed r.m.15 illustrates the essentials of a receiving circuit. the amplitude of which varies in accordance with the modulation. 3.
f. Illustrating Action of Discriminator Circuit capacitor in Fig.16. signal would be suitably amplified before being applied to the detector. Such an arrangement. 3. With frequency modulation it is necessary first to convert the Modulated wave Rectified wave Mean Fig.15 slightly off tune so that its operating point would be on the side of the resonance curve. 3. The device responsive to the variations in frequency is called a discriminator and may take various forms. however. One method would be to set the Fig.17. Any variation in the frequency of the carrier would then produce a corresponding variation in the voltage across the capacitor. is critical to operate and the more usual procedure is to use a loose. 3. while in all but the simplest type of receiver the r.150 RADIO COMMUNICATION replaced by a resistor across which the modulation would develop an audiofrequency voltage which would then be fed into an amplifying stage. Principle of / value Detection variations in the frequency of the carrier into amplitude variations and then to follow this with a rectifier as just described.coupled transformer having .
in both receiver and transmitter. Bandwidth It was shown in the previous section that the modulation of a carrier wave produces a range of side frequencies.5 both primary and secondary tuned.s a curve of the form shown in Fig. and 9. 3.17 (b) is obtained. as shown in Fig. These are briefly as follows: Amplitude modulation. and it will be seen that around the midfrequency point the output is proportional to the frequency. though for highfidelity reception a higher standard of performance may be desirable. For audio reception fm is of the order of 10 to 15 kHz. critical coupling. Pulse modulation. = = + B= 04/t where t is the time of rise of the pulse. 3.m. and that the relevant information is contained entirely within these sidebands. not less than l/\/2 times the maximum) at the extremes of the band. and so converts frequency variations into changes of amplitude. where A/ is the rated deviation and fm is the maximum modulation frequency. Hence it is essential that the circuits involved.6.e. The frequency scale has been expanded for convenience. B 2/m where fm is the maximum modulation frequency. Practical forms of detector circuit are discussed in Sections 9.17 (a). It is usually considered satisfactory if the overall response of the system is not more than 3dB down (i.f. The requirements are discussed in detail in Part II (and particularly in Chapter 9). B 2(A/ /m ). but it will be useful here to summarize the bandwidths involved in the different forms of transmission. shall respond adequately to a range of frequencies immediately adjacent to the carrier frequency. Frequency modulation. If the detector is now made to respond to the difference between the primary and secondary e. . but is roughly given by .PRINCIPLES OF RADIO COMMUNICATION 151 With such a circuit. as discussed in Chapter 17. to an extent dependent upon the bandwidth of the particular transmission. Here the bandwidth is dependent on the system used. using the maximum current in the primary does not occur at the same frequency as that in the secondary. as discussed in Chapter 16. but for television it is several MHz.
A.1. N. On the other hand there are many occasions on which resistance can be utilized as. and it will be seen that the metal which exhibits the highest conductivity is silver. To treat this subject exhaustively is quite beyond the scope of the present work. 4. This is too expensive for general use. though it is 152 . or specific resistance is The resistivity defined as the resistance of a cube of the material of unit dimensions.Practical Components The application of the basic principles discussed in the preceding chapters involves the use of components which exhibit the required properties to a known and controllable extent. Dummer. It is measured in ohmmetres and typical values are given in Table 4. but it is desirable to indicate the general methods employed. to use materials which have a large resistivity so that the required value of resistance can be produced in a reasonably small compass. and A p is the resistivity. W. The resistance of any conductor is given by the expression B = pl/A where I is is the length of conductor. in the various networks In such circumstances it is desirable associated with amplifiers. and/or Radio & Electronic Components (Vols. 16) by G. More detailed information will be found in Telecommunication Principles by R.1. Renton. Resistors In the conductors used to carry electric currents the inevitable resistance of the wires is a disadvantage and has to be kept to a minimum. for example. The reciprocal of the resistivity is known as the conductivity. et al. the crosssectional area.
depends upon With metallic conductors it is found that the resistance . An alloy of approximately 58 per cent copper and 42 per cent nickel has a somewhat lower resistivity but has the advantage of a very small temperature coefficient and is therefore suitable for the manufacture of resistors which are required to be constant in value under varying temperature conditions. where a high resistance is required use is made of materials of high resistivity.PRACTICAL COMPONENTS Table 4. An alloy of 80 per cent nickel with 20 per cent chromium has a high resistivity and is used for the elements of electric fires and heaters. usually in the form of a steel core for strength with aluminium conductors stranded round the core. Transformer steel Nickelchromium (80/20) Copper nickel (60/40) Carbon . Copper .1. Temperature Coefficient The resistance temperature. etc. . 10100 55 103 40 39 39 5520 10 negligible 48 4.5005. Aluminium is also a good conductor though it has only 63 per cent of the conductivity of copper. and electrical conductors are usually of copper in the form of wire or strip. Aluminium Iron (according to 16 17 27 grade) . It is suitable for this application because it does not oxidize too easily at the dull red temperatures at which heater elements usually operate. . Such conductors therefore are used for aerial lines such as are employed in highvoltage transmission. . but because it is so much lighter (its specific gravity being only 31 per cent that of copper) an aluminium conductor will have only half the resistance of an equivalent weight of copper conductor.000 02 to 05 employed occasionally. Constantan. of a conductor. . in general. covered where necessary with suitable insulation. 153 Resistivities and Temperature Coefficients Temperature Coefficient Material Resistivity {at 20°C) in ohmm x 10~ 8 per°G ( X 10*) Silver . Alloys such as this go by various trade names such as Eureka. For such purposes alloys of two or more metals are the most suitable. but it has the disadvantage of a large temperature coefficient. . On the other hand. The material in more general use is copper which will be seen to be nearly as good.
to construct resistors in the form of short rods. R 20 is the resistance at 20°C. With carbon. Carbon exists in many forms from lampblack to diamonds. at effects are often introduced. and this heat must be dissipated.e. It will also be noted that the temperature coefficient of a metal is positive so that the resistance increases as the temperature rises. This is a theoretical zero. Carbon Resistors For many purposes resistors are required having values of many thousands or even millions of ohms. The currentcarrying capacity of a resistor of this type is determined by its size. often admixed with suitable additives. This zero cannot actually be achieved in practice but many physical phenomena behave as if they originated at this point.1. so that the resistance of a conductor of this material does not vary appreciably with temperature. The temperature of . It will be seen to vary with the material and it will also be noted that with the coppernickel alloy the temperature coefficient is virtually zero. Energy will be absorbed by the resistor in the form of heat equal to PR. actually 273° below zero on the Celsius scale. they behave as if — they had virtually no resistance at all. In this case the resistance can be expressed in the form R = Ew(l + t at) where B is the resistance at a temperature of t°G. usually by radiation from the surface of the material. temperatures approaching absolute zero certain materials begin to exhibit what is called superconductivity i. To obtain such resistors with a wire conductor requires a very fine wire and a considerable length. though in practice as one approaches zero secondary As mentioned in Chapter 1. and t a is a constant for the particular material.154 at RADIO COMMUNICATION any given temperature is proportional to the temperature above absolute zero. The term a is called the temperature coefficient. the temperature coefficient is negative so that the resistance falls with increasing temperature. Hence it is more convenient to utilize the much higher resistivity of carbon. It is more usual to express the resistance in terms of its value at convenient standard temperature. usually taken as 20°C. on the other hand. being in fact the relative change of resistance per degree and typical values for this temperature coefficient are also given in Table 4. and it is possible by controlling the grade of carbon used.
known as highstability resistors. in fact.PRACTICAL COMPONENTS 155 the resistor will.1.wound resistors are used. The carbon rod itself is usually housed in a protective covering of moulded material. (d) Carbon rod in insulating Wirewound resistor. sleeve. usually in the form of a spiral. the actual mm mm mm fcJ <°> .). Thinfilm Resistors Greater stability can be obtained by the use of an alternative technique in which a thin film of carbon or metal is deposited on to a rod of insulating material. 4. The material is then cut away. Thin film (highstability) type.1.xm)j #0 ) Pig. so that the effect is cumulative. and they range from small units long and 25 8 diameter which are rated at oneeighth of a watt up to units between 25 and 50 long and proportionately fatter which will dissipate 3 or 4 watts. The resistance also is subject to variation after periods of use. Fig. 4. Because of the substantial negative temperature coefficient of carbon these resistors must not be allowed to become overheated or the resistance decreases rapidly with consequent increase of current. Wirewound Where Resistors precision is required or large wattage dissipations are necessary. Such resistors. illustrates the various types of resistor construction. maintain their resistance appreci ably more constant. (o) (c) Typical Resistob Constructions (6) Carbon rod resistor.£)*">> ») ( jm. value of the resistance being indicated by small coloured rings in accordance with an agreed code. The wire is of appropriately . rise until an equilibrium condition is obtained where the energy loss is all satisfactorily radiated as heat. so as to alter the effective length of the path until the required resistance value is obtained. wire. Practical resistors are therefore rated in terms of the wattage dissipation which they will handle.
Where the resistor has to dissipate appreciable heat it is usual to coat it with a vitreous enamel and the whole resistor is then baked in a furnace. sometimes loosely called a potentio meter. the device acts as a variable voltage divider. inductive. 4. of course. Carboncomposition elements may be graded by changing the dimensions and/or the composition. squarelaw. The element may be made to provide a specific nonlinear change of resistance with rotation of the slider (e.156 RADIO COMMUNICATION small diameter. This holds the wire rigid and prevents appreciable movement due to the successive expansion and contraction due to the heating and cooling. The wire is wound as a singlelayer solenoid on a suitable former and is then given a coat of protective lacquer.). 4. In a wirewound potentiometer the element is wound on a thin which is then bent into a circle and mounted in a suitable housing. This construction is only practicable for low resistance values. Resistors wound in this manner are. For many purposes the inductance is so small that the inductive reactance is quite negligible by comparison with the resistance. In these the usually made up in a circular form with a light rotating "slider" to make contact with any desired point along the element. More usually the element is made of carbon composition. By connecting the element across a source of voltage and taking the output between the slider and one end. flexible card of suitable insulating material .2. if this is not the case some form of noninductive winding has to be adopted. reciprocal law etc. The slider rests on the edge of the winding from which the insulation is removed. Fig. One of the methods used is what is known as the AyrtonPerry winding which consists of two singlelayer solenoids wound one on top of the other but in opposite directions as illustrated in Fig. AybtonPerry Noninductive Winding Variable Resistors For many purposes variable resistance element is resistors are required. However.2.g. often as little as nfouth of an inch. The grading is achieved in wirewound devices either by tapering the width of the card or by changing the gauge of wire at suitable points (though this obviously does not provide a smoothly varying law).
until the required value is obtained. If a. Thermistors are used in a variety of ways to compensate for changes in circuit conditions. but these need not be discussed here. if the coefficient is positive (ptc) the resistance increases with current. as in a varistor. These effects may be utilized in two ways: Varistors. Typical Thermistor Characteristics Thermistors. If the material has a negative temperature coefficient (ntc) the resistance decreases with current. 4. and such a device connected across a circuit will provide a measure of surge protection. but this is of secondary importance. 4. Nonlinear Resistors For certain requirements it is convenient to use resistors having a nonlinear currentvoltage relationship. Small wirewound preset resistors are made with the element on a thin card with a sliding contact which is moved along the card circuits require the value Many test to give the required performance. = 1000 Fig. They are responsive to changes in the ambient temperature. closer More elaborate forms of variable resistor are made to provide a and more accurate control of the resistance. though they are sometimes used as temperaturesensitive elements. This is a characteristic property of semiconductor materials. . Materials such a silicon carbide obey a law of the form i kif. These contain a small bead of semiconductor material so proportioned that the current flowing raises the temperature sufficiently to cause a significant change of resistance. of which the resistance depends on the current and also to a marked extent on the temperature.3 illustrates typical characteristics. 31 a 20 per cent increase in voltage will halve the effective resistance. is.PRACTICAL COMPONENTS 157 of a resistance to be adjusted on For this purpose preset resistors are used. say. which are made up as small versions of the normal variable resistor in a form convenient for mounting directly in the circuit. Fig. where x is between 3 and 5.3.
The increase is not important at low frequencies but The ratio of the it is quite considerable at radio frequencies. The effect is accentuated when the wire wound into a coil as discussed in Section 4. the . greatest at the centre of the wire. but in transmitters where considerable radiofrequency currents have to be handled. p = resistivity of conductor. The operating temperature of the wire and the pressure of the gas are so arranged that over a limited range of voltage the current remains constant. RADIO COMMUNICATION alternative form of nonlinear resistor is the barlamp. or ballast An Skin Effect This wire carrying an electric current produces a magnetic field. retter. is given by the expression BfIB9 = where d fi a dV(tflp) . The skin effect can also be minimized by using a rectangular strip for the conductor. which contains a filament of iron wire enclosed in a glass bulb containing hydrogen. = permeability of conductor.3 because the magnetic field inside the coil itself also tends to drive the current to the outside edge of the winding.f. in fact.158 Barretters.m. the seriousness of the effect increases as the diameter of the wire increases. If the current is varying. therefore. resistance at a frequency /to that at d. both inside as well as outside the wire and is. The effect. the size of the conductor has to be much larger than this and special constructions are adopted to minimize the skin effect. an e. As will be seen from the expression above. +b — diameter of conductor. In consequence of this the effective resistance of the conductor is increased. One method is to use hollow tube for the conductor since if the current is flowing mainly in the outside of the conductor there is no point in filling the inside with expensive copper.c. Still another method is to use stranded wire. For many purposes it is found that even at radio frequencies the increase in resistance is not too serious if the wire used is of relatively fine gauge (of the order of 0025 mm). of selfinduction is set up tending to oppose the flow of the current and this is again greatest at the centre of the wire. If the conductor is made up of a number of finewire conductors in parallel. is that the impedance of the wire progressively increases towards the centre so that the current tends to flow in field exists A the outer skin of the wire. is and a and b are constants.
the effective resistance will be reduced. Another device often used with highfrequency circuits is to silverplate the conductors. The fusing current is roughly proportional to d 312 .PRACTICAL COMPONENTS 159 wires being so arranged that each in turn comes to the surface of the conductor. Silver is a slightly better conductor than copper and therefore if the outer skin of the conductor. Hence the greater the current the higher the temperature until ultimately the conductor becomes white hot and melts. A sudden surge of current may cause such a rapid rise of temperature that the fuse explodes. thereby saving the rest of the circuit from damage. of course. the fuse wire can be housed inside a small tube of . These are short lengths of fine wire which are designed to reach a melting temperature at currents appreciably in excess of that for which the rest of the circuit has been designed. is expended as heat and the temperature of the wire rises to a value such that this energy can be satisfactorily radiated. so that a fresh length of wire can be inserted after the fault has been cleared. A normal conductor is. scattering molten metal around the vicinity. This may take the form of a removable unit of ceramic or other insulating material with contacts which plug into clips at each end. then the increase in resistance due to skin effect can be very considerably minimized. is of silver. When a wire carries a current the energy used in overcoming the resistance. it is The material : Fuses The radiation of energy from a body in the form of heat increases with the temperature. Then if the current becomes excessive due to a fault the fuse "blows" and breaks the circuit. so proportioned that its temperature rise is not excessive. therefore. to enclose the fuse in some form of protective housing. is difficult to use because it is necessary to clean the ends of each individual strand and solder them all together at the beginning and end of the conductor if any strand is not connected or is broken the eddy current loss induced in the dead strand may cause a considerable increase in the effective resistance. Alternatively. in which most of the current flows. but the actual law is modified by the form in which the fuse is made up. Wires such as this are called Litzendraht (or litz) conductors and are considerably used where important to obtain the lowest possible effective resistance. It is usual. running through a hole or channel in the ceramic holder. The fuse wire is then connected between these contacts. but this ability to produce melting is used to provide protection of circuits by what are known as fuses. which is I 2 Bt.
replaced will making an expendable unit which is by a new one when necessary. desirable to keep the physical size of the capacitor small. Practical Constructions The construction of a capacitor depends on the work it is intended The capacitance may be fixed or may be made variable at will. Such a device is called a capacitor or.4 (b) shows the effect of adding another plate to the first pair. a condenser. 4. say. Pig. The capacitance of such an arrangement is proportional to the area of the plates and basic idea of capacitance The i\(cv) H (b) ic) Fig. This may be done by using thin metal foil for the plates and separating them by a solid dielectric. This form of "cartridge" fuse can be designed to have a small time lag in operation so that it withstand momentary surges. a series of plates being mounted up in a bank and alternate plates being connected together. and . Capacitors was discussed in Chapter 1. . 4.4 (a). 4. of. Early small capacitors were made up in this way. five times normal current but will blow on a sustained current of two to three times normal. 4.160 RADIO COMMUNICATION glass or ceramic material. there is a limit to the separation between the plates and also to the size. Simp:lb Fixed Capacitob Fobmations inversely proportional to the distance between them.2. which simply consists of a pair of parallel plates. In communication engineering frequent use is made of devices designed to produce a known value of capacitance in convenient form.4. Here both sides of the middle plate are effective and double the capacitance is obtained. in older terms. 4. In receiving equipment and similar lowvoltage applications it is for. preferably one having a high permittivity such as mica.4. Hence the nearer the plates are together the larger the capacitance obviously. The process may be continued as in Pig. so that in order to increase the capacitance several pairs of plates are used. The simplest form of capacitor is shown in Fig. (c).
Calcium ceramic materials (e. which results in high stability with very small size to meet the requirements of modern miniaturized circuitry. while there are special ceramic materials which exhibit a marked negative coefficient (e. may be offset over quite a wide range of temperature. The main variation is thus that due to temperature. Alternate strips of foil and paper are rolled up together until the required capacitance is obtained. titanium dioxide. The voltage which the paper will withstand is limited and hence for high voltages two or more thicknesses of paper are used to each foil. The completed unit is then vacuumimpregnated at a temperature sufficient to drive out any moisture and allow the impregnant to fill any voids. Foil Capacitors For large capacitances a paper dielectric may be employed.PRACTICAL COMPONENTS where the capacitor is 161 required to cany appreciable current this construction is still used. including Capacitors can be monobloc units in which the element is encapsulated under pressure in an epoxy resin. uses silver deposited on mica.g. Silvermica capacitors exhibit a very small positive coefficient. Such capacitors have a temperature coefficient depending only on the expansion of the base. Formerly wax was used as . The capacitor is made up from thin strips of metal foil and specially prepared thin paper. which is almost invariably positive. a cheaper and more stable capacitor can be produced by depositing a metal coating on the dielectric. This minimizes the small changes of capacitance which can arise from slight mechanical movements of the plates due to temperature or other causes. the metal coating having sufficient elasticity to follow the changes in dimensions of the material on which it is deposited. and because of this intimate contact they are much less liable to erratic changes. For most purposes. the capacitance by about 00025 per cent per degree C. One such construction in common use. which 0068 per cent). has a coefficient of increasing — made up with a combination of two dielectrics so blended as to provide a zero temperature coefficient.g. such as the coils. porcelains) have a positive coefficient of the order of 001 per cent. There are many variants of this basic technique. after which it is sealed in a suitable case. which may be partially or completely airtight. known as the silver mica capacitor. while there are various alternative forms in which a ceramic base is used. however. while by deliberately using a suitable negative coefficient the variation in value of other parts of the circuit.
162 RADIO COMMUNICATION the impregnant. For smaller values the assembly is housed in a tube of metal or insulating board. end.f. which would render it useless as an r. 4. Even 01 uH. the conducting surface being deposited direct onto . but this expands with heat and so tends to deplete the inner layers of the winding.5. so connecting the sections in parallel. Pape Insulation Fig. In many instances the current which the capacitor has to handle is relatively small. 4. Noninductive Capacitors Where minimum inductance is required. Considerable improvement can be obtained by using thin sheets of plastic material. but even a small foil capacitor may resonate in the low megahertz region. Hence it is more usual to use petroleum jelly or a polyphenol resin. as shown in Fig. of the order of 001 to depending on the construction. for the dielectric. such as polystyrene. Illustrating Construction of Noninductive Capacitor some residual inductance is present. The individual foils clearly constitute what is in effect a "pancake" coil and so possess inductance. When the winding is complete. the foils can be arranged to overlap on alternate sides.5. Large capacitors (up to about 8 fiF) are normally housed in rectangular steel containers occupying about 40 cm 3 per microfarad. connections can be taken to these projecting edges. with the connections brought out as wires at each so. The effect is only important at high frequencies. though this is not usually of major importance. bypass. This permits the use of metallized foil in the construction. To minimize this (and also the resistance) connections are made by inserting a number of foil strips at several points as the winding proceeds. Paperinsulated capacitors do not have good longterm stability.
6 («). S pY Inwvoltage electrolytic 001 fii'' i. 025 /<!' [tokstyrene tubular. 4 Typical Fixed Capacitors (itelectrlo.lisi.6.Fio. £00 pK silver mica. 4. A Variable Cafacitoii 102 . 'I aS paper 50 jip electrolytic .(KM pi" silver coruiuir5. 4.00!) [>F silver mica: 16 t<T tubular eleetrolvtic. Ul fF jiiiper tubular Fifi.ivin: silver eeraniie.
T * Ifg: fe > V ~ Sill JeH Hg .I 3 ^vj Is! «« P* • K £ J 3 B c ™ i P vc r5 w ."03 E 1 B a S fi % •& g @ *B a I K I . . — .
partial breakdown of the dielectric film will occur. for certain applications. voltage across it. usually of aluminium. Electrolytic Capacitors Paper capacitors.voltage types for transmitting equipment. which consists of porous paper impregnated An . and this property is useful in many circumstances. was rolled up in cylindrical form and immersed in a metal can containing electrolyte. On the other hand. or is in the reverse direction. This reduces the and hence the overall size. to obtain very high capacitance in small volume in what are called electrolytic capacitors. but it has the advantage of being selfsealing. as will be seen later. however. except that one foil is formed with the oxide film as just described. however. however. electrolytic capacitor is thus limited in its application and it cannot be used on alternating currents. If a metal plate. be employed for smoothing out or bypassing alternating currents. The formed foil is the anode and the plain foil is the contacting electrode to the electrolytic cathode. W. A typical polycarbonate 10 //J? capacitor (63 V working) has a diameter of 165 mm and a length of 35 mm. It is more usual to use a "dry" construction. Here the capacitor is similar in form to a paper insulated capacitor. A. a very thin film of aluminium oxide forms on the surface of the plate. There is thus a wide range of foil capacitors from low. is liable to leak. This film forms a dielectric layer between the aluminium plate and the electrolyte so that the arrangement exhibits the properties of a capacitor. The maximum capacitance obtainable with a given anode surface area is inversely proportional to the voltage at which the capacitor is formed. corrugated to increase the surface area. particularly if the capacitor is allowed to get hot. Dummer (Pitman).voltage units for transistor circuitry to high. The oxide film is so thin (of the order of 10~ 6 cm) that the capacitance is far higher than could be obtained in the same space by ordinary methods. The electrolyte. Early forms of electrolytic capacitor were of the "wet" type in which an aluminium foil. an insulating foil of polyester or polycarbonate.PRACTICAL COMPONENTS 163 thickness.c. if the voltage applied to the capacitor in use exceeds the forming voltage. are inevitably bulky but it is possible. and ultimately dries out. provided there is a steady d. Further details may be found in Modern Electronic Components by G. It can. is immersed in a suitable solution such as aluminium borate and a current is passed through the assembly with the aluminium plate positive. since the film thickness is dependent on the forming voltage.
A "dry" electrolytic capacitor is not selfsealing and. Such capacitors are not only smaller in size. but are extensively used in = sophisticated transistor circuitry. electrolytic capacitors have a great advantage in size for a given capacitance over all other forms of capacitor. mm Some typical fixed capacitors are illustrated in Fig. exceeded there is a momentary conduction current which in turn produces ionization resulting in a cumulative buildup of a current of such intensity that the dielectric may be punctured. A capacitor of this type will provide a capacitance of 5ftF. impregnated separator and plain foil is then rolled up and housed in a suitable container. after "solid" types made using a sintered pellet of tantalum powder. therefore. 4. A greatly improved performance is obtainable by the use of tantalum. The permissible voltage depends . more expensive than aluminium types and are therefore usually limited to lowvoltage applications. must not be subjected to voltage in excess of the forming voltage. The assembly of formed foil. A further increase in capacitance for a given size can be obtained by mildly etching the anode foil with acid before forming. improved stability and very low leakage current. Since the electrolyte is a paste it is able to make contact with the whole of the irregular surface area of the oxide film. but the power factor is somewhat increased. so that an irregular surface of oxide is formed having a much greater area than that of the plain foil. An etchedfoil capacitor can have a capacitance up to 8 times greater than a plainfoil capacitor of the same size. However. are rated to function at or below a specified working voltage. the oxide of which has a high permittivity (er 27). The conventional foil construction may be used. in a bead only 64 in diameter.164 RADIO COMMUNICATION with electrolyte (usually a paste of glycol and ammonium tetraborate). Nor should it be used under hot conditions for it is not truly dry and continued high temperature will cause rapid deterioration. the high leakage current and the variation in capacitance with time and temperature. Further disadvantages of electrolytic capacitors are the high power factor (about 10 times that of a paper capacitor). contact with the oxide being maintained by an admixture of manganese dioxide. 20 V working. therefore. Breakdown Voltage withstand a certain potential gradient. dielectric will only Any If this is All solid dielectric capacitors. but have a long storage They are life. This is oxidized which the electrolyte is removed. but are electrolytically.6.
Law. The variation of capacitance with angular rotation depends on the shape of the plates. is not known beforehand. so that if the plates are so shaped as to make G oc 2 kx y'(C') &2A. This will give an approximately linear relation between the shaft rotation and the wavelength to which the circuit is tuned. the maximum value depending on the number and distance apart of the plates. for the electric strength of an insulating material decreases rapidly as the temperature rises. such as: (a) Square Law. A k\/(LC). Capacitor ratings often include a "surge" voltage figure. A bank of plates is arranged shown in Fig. (b) 8. superheterodyne receivers use a capacitor for the oscillator tuning which provides a constant difference in frequency from the input signal. With plates of constant radius the capacitance variation is linear. (c) Logarithmic Law.F. as The second method is more common. A V V must be remembered that the important factor is the peak value of the voltage.L. and the other is to alter the effective area of the plates. though with electrolytic capacitors the margin is small (10 per cent or less). in general. The maximum capacitance depends on the size and number of plates and their relative spacing. This is known as a straightlinefrequency or S. but by suitable variation of the radius the capacitance change can be made to obey various other forms of law. The relationship is only approximate since G must include the stray circuit capacitance which. This gives a continuously variable capacitance. paperinsulated capacitor which can safely be used up to 750 at 60°F would be limited to 600 working at 75°F. Capacitors used in receiving equipment usually have a capacitance of the order of 200 to 600 pF. Capacitors. = = = . Variable Capacitors One There are two methods of varying the capacitance of a capacitor. will safely withstand momentary It voltages higher than normal. is to vary the distance between the plates.PRACTICAL COMPONENTS 165 on temperature.L. though it is more usual to obtain the effect required in a slightly different manner. Caz log 6.F. As shown in Section 2. 4. Law. the moving plates being capable of rotation into the spaces between the fixed plates.6 (a). This can be obtained by suitable shaping of the oscillator capacitor plates. As explained in Chapter 9. and the circuit must be such that the peak value of voltage under any conditions does not exceed this figure. (d) Constant A/ Law.3. . By making G oc 1/0 2 a linear relationship can be obtained between the frequency to which a circuit will tune and 6. in general.
with means for varying the distance between them by means of a screw. usually separated by a thin mica sheet. 4.166 RADIO COMMUNICATION Trimmers For many to a given value (adjusted locked. The various types are illustrated in Fig.9. Various forms of trimmers exist. (6) Capacitors comprising a cylindrical inner element on a screw which can be moved inside an outer tube to any desired extent. sometimes purposes a capacitance is by actual trial in using a solid dielectric. Fig. the principal ones being (a) Miniature versions of a normal variable capacitor.9. Typical Tkimmbks Power Factor Loss Angle . and no energy should be absorbed during the process. 4. With an air dielectric this condition is very . Such capacitors are required which can be set the circuit) and then known as trimmers or preset capacitors. (c) Capacitors comprising two flat plates. The ideal capacitor is simply a device for storing electrical energy and releasing it when required.
If we assume the resistance to be in series. which results in a loss of energy.PRACTICAL COMPONENTS 167 nearly obtained. Nevertheless. whichever is more convenient.2. Since in general the loss resistance is much smaller than the reactance. but with liquid or solid insulation there is a certain electrical friction occasioned by the movements of the electrons in the material. Hence current Power which factor = RCu> is the same as the loss angle in radians. The power factor.c. loss does exist and has to be allowed for. is in this case indistinguishable from RjX. which is equivalent to saying that the Q of a capacitor (which is the ratio of reactance to resistance. The effect of the dielectric loss can be represented by assuming that a practical capacitor is composed of an ideal capacitor associated with a loss resistance which may be treated as being in series or in shunt. Since this angle is very small it may be expressed simply as R/X(= RCm) radians (the radian being the measure of the ratio of an arc of a circle to its radius. 00004 0009 . Power Factor of Typical Dielectrics (average values) Relative Permittivity Material Power Factor Air 1 Mica Glass Bakelite . then by the normal a. 5 55 Mycalex 005 0002 6 8 25 65 2 Polystyrene Ceramics Paper 00002 0002 0005 . but by keeping this to a minimum the dielectric loss can be kept low. is nearly 90°. This current will not lead by 90° but by some smaller angle given by arc tan VC /VB arc tan (l/BCco). It will be seen that the power factor is in general very low. which is the angle ip by which the loss It is clear that + <f> = = <j> fails to lead by 90°. which is the ratio of resistance to impedance. Typical values of power factor for different dielectrics are given in Table 4. =l/RGco) is very high. laws the current / == V/(B l/jcoG). There is also some caused by the (ohmic) insulation resistance of the dielectric. as explained on page 42).2 below. we cannot support two sets of plates in midair without some solid dielectric at a suitable point. Table 4. and it is more usual to specify the loss angle of the capacitor.
from which point B oc V(f) is a third source of loss known as dielectric absorption.. of course. It may be noted also that the capacitor inevitably possesses some inductance. Hence in terms of parallel resistance. The material appears to absorb a certain abnormal extra charge.168 RADIO COMMUNICATION is If we represent the loss by a parallel resistance. Good oil. a further discharge will occur showing that the charge was not all dispersed on the first occasion. Synthetic materials. it is greatest at low (power) frequencies. but those used as insulators in capacitors tend to exhibit a constant power factor. and is found to be inversely proportional to the frequency. where Y is the admittance of G and B v in parallel. are not uniform in their behaviour. This is normally negligible but becomes of importance at high frequencies because of the increase due to skin effect. and grows less as the frequencies are There raised. Variation of Loss with Frequency Dielectrics. . glass has an efficiency of about 60 per cent. depending on its quality. The energy loss from this cause is comparatively small. The resistance of the plates and connections. is nearly as good. = V(llB p +ja)C) = 7[(1 + jcoOBJIBJ. the effect again being negligible until frequencies in the shortwave The effective series resistance of dielectric loss.e. but in general introduce appreciable loss. and mica 40 to 90 per cent. such as resinimpregnated paper. by their very nature. If a capacitor is discharged and then set aside and again shortcircuited a little later. The loss angle ip is then arc tan 7 B /7 C = 1IBpwG radians. Compared with air. vary widely in performance. the current VY. Factors Affecting Capacitor Performance a capacitor is not only due to but includes the actual resistance of the plates and leads. Air is the most efficient dielectric in this respect. is virtually constant until skin effect begins to be appreciable. i. Power factor = \\B vo/J we see By that BP = equating this to the expression for a series resistance l/w 2 (72 i?. Hence the equivalent loss resistance is inversely proportional to frequency. region are attained. free from moisture.
but not in direct proportion. It is measured in volts per unit distance. but since. (best ruby) Glass Bakelite Mica . perature which in turn still further increases the losses. As the distance between the plates is increased the breakdown voltage increases..000 . . 4. breakdown occurs is called the formerly the dielectric strength. 1 Electric Strengths for Plates mm Apart Dielectric Electric Strength (volte/mm) Air Oil . the relationship is not linear and also the breakdown potential is dependent on the shape or disposition of the electrodes. A Transmitting Capacitors Capacitors for transmitters usually employ air dielectric. In any case the plate spacing has to be sufficient to prevent breakdown at the voltage in use.000 8.PRACTICAL COMPONENTS 169 The effect is to increase the apparent capacitance to a value C = C/(l .. though liquid dielectrics such as paraffin are sometimes used...co*LC) The losses in an electrolytic capacitor are considerably higher than and rise rapidly with increasing temperature....000 30. figures.3. each section having only a portion of the full voltage across it. This rise with temperature is true of all dielectrics and becomes important at high frequencies when the increased losses cause a rise in temfor other forms. as has just been stated. Table 4. The gives some representative Table 4.000 7..3 potential gradient in which electric strength.g.. This necessitates a considerable reduction in the permissible working voltage so that the current through the capacitor may be kept within safe limits.000 60.. capacitor which has an electric strength sufficient to withstand 10 kV at low frequencies may not be able to handle more than a few hundred volts at 10 MHz. Hence transmitting capacitors are often made up in several sections is in series. it is important to specify the exact conditions as an indication of the order of potential gradient required to produce breakdown. e. if the distance doubled the breakdown voltage only increases about 175 times.
for the sulphur content of the material introduces considerable loss which in a transmitter may produce so much heat as to cause serious mechanical distortion. The paper is impregnated with Bakelite varnish or other special insulating varnish. however. This is a ceramic material containing mainly china clay (kaolin). Silvonite. One form of transmitting capacitor is made up by mounting a series of plates separated by air and supported at the corners on glass blocks. Such capacitors. 1 70 RADIO COMMUNICATION As voltages of from 10. and to use an oil or mica dielectric. or are rolled A trade names such as Keramot. Mica is considerably used for small parts. not only in capacitors but for general purposes. Dielectric Materials The material for insulation. particular example is a synthetic redcoloured material composed of rubber and mica. This not only A increases the capacitance but enables the plates to be assembled nearer together without any risk of the voltage sparking across.000 volts are quite common in transmitting practice this question requires proper attention. such as ebonite.. A material which is much used. Another method is to hang plates from rods of glass or other insulating material. Thin sheets of paper are either compressed together up to form rods of circular or other section. such as Paxolin or Tufnol. however. Porcelain is suitable for many applications. Paper products impregnated with suitable composition are widely used. This class of product goes by various trade names. various mica products formed by bonding powdered mica with suitable materials and forming sheets or slabs. Numerous other synthetic materials are continually being developed. etc. and feldspar.000 to 50. It is very good for medium radio frequencies. but its brittleness seriously limits its application. are considerably more costly. are not usually employed except at quite low frequencies. The dielectric is thus principally air. is a study in itself. C . which is ground mica bonded with lead borate C glass and moulded under pressure at about 675 It is a dull grey material which can be machined with special tools. known by various to form one thick slab. is Mycalex. the distance between the plates being flashover. and combines low losses with high electric strength (55 kV per mm) Hardrubber compounds. rigid made large enough to avoid any more common method is to build up the plates in a structure. There are. flint. and in some cases linen is employed to give greater mechanical strength. particularly for highfrequency work.
however. Here the winding is arranged to travel from one side of the coil to the other every turn. are wound one on top of the other spiral fashion. for some applications. the result being that a large number of air spaces is obtained throughout the coil. particularly for shortwave work. The ceramic insulators are undoubtedly the best from the point of view of dielectric properties. i. Here the successive turns. Such a coil is called a "solenoid. In the latter case the chief consideration is the obtaining of the requisite inductance in a fairly compact form.winding in which the traverse from side to side is so arranged that the wires lie close together without the large air spaces of a true honeycombweave.e.3. and then mixed with a small quantity of moisture to enable them to be moulded. This method. With this arrangement. results in a rather bulky coil and it is more usual to use a wave. but because of their fragility they are only used in positions where the highest efficiency is essential.PRACTICAL COMPONENTS which are 171 finely ground and mixed." Another form of inductor sometimes employed is the spiral or pancake variety. This gives a more compact Wavehaving a performance not significantly worse. which may be minimized by using a "honeycomb" winding. is also utilized. there is appreciable stray capacitance between the turns. For higher values of inductance multilayer coils are employed. They are then fired and. glazed to provide a smooth surface. Aircored Inductors 4. which is a purified talc sometimes known as soapstone. however. Similar ceramic compounds have been highly developed. Magnesium silicate is also used instead of clay (which is an aluminium silicate). It is also possible to use a dry casting process in which the powder is formed into the required shape under pressure without firing. coil wound coils have the advantage of considerable rigidity. instead of being side by side. For low values of inductance the coils are usually wound in single layers upon cylindrical formers of bakelized paper or ceramic material. while steatite. It is possible to arrange that the successive turns of each layer are spaced by a small distance from each other. Inductors are used in both transmitting and receiving equipment. when one layer is completed a further layer is wound on top. If this is done. being quite . the successive layers of the coil cross each other at an angle (giving a crisscross or honeycomb effect) and do not lie flat on top of one another.
to avoid losses due to eddy currents and skin effect. in other words. the mutual inductance between the coils. 172 RADIO COMMUNICATION selfsupporting.e. or copper tube. Both cylindrical and spiral coils are employed. The extent of such coupling. An intermediate form of construction is to use a former having a number of slots. Magnetic Coupling It has been seen that if one coil is placed near another which is carrying current. Where the coil is not required to carry highfrequency currents a simple multilayer winding may be used.7. They are often given a light dip in varnish and then baked at a low temperature. one turn several layers up may slip down and lie against a turn near the beginning of the coil.000 volts. i. and are usually wound with stranded wire. All these constructions are subject to minor temperature effects due to changes in the physical dimensions. If the coil is placed at right angles to the first coil no flux will flow through the coil but only across it. the self. and inductors in general exhibit a small positive temperature coefficient. The winding is arranged to fill each slot in turn Due to this sectionalizing. This may be wound on a former having cheeks at each end to retain the wire. The turns also are well spaced apart and are wound on a suitably insulated former owing to the high voltages involved. The inductance in a highpower transmitting station has a voltage across it careful insulation of the order of 50. there will be no linkage between the flux and the coil and no coupling will . This condition of affairs is referred to as a "coupling" between the two coils. Otherwise. Transmitting Inductors For transmitting units the coils have to be wound with much thicker conductors to carry the heavy currents. an arrangement which would not only create a high selfcapacitance. but multilayer coils are seldom used.capacitance is considerably minimized. so that the need for is at once apparent. it is affected by the variations of the magnetic field of the latter. 4. but might result in a breakdown of the insulation. depends on the proportion of the flux which links with the coil.. copper strip. This applies both to aircored coils and the ferrite assemblies described later. Typical coil constructions are illustrated in Fig. on which the inductance primarily depends. and there is usually an interleave of paper or other insulating material between layers in order to keep them apart.
will appear in any coil which links with the field. 4. produce any effect but if the magnetic field is varying.— PRACTICAL COMPONENTS result (Fig. and the mutual inductance varied by altering the distance between them and/or their This is relative orientation.10 (a)). the presence of this "stray" field is often an inconvenience. Coils with Closed Magnetic Circuits circuit.16. Illustrating Magnetic Coupling exists may be between any two coils in proximity.f. while this sometimes useful (as in an r. The type of inductor so far considered has an "open" magnetic The magnetic flux generated within the winding has to is complete its circuit through the surrounding medium and.aaoacoce SBBOOBOBB Fig.10 (&)). Hence coupling ^oggggooo jaoqonryyn. from the production of variable inductors as discussed later. This coupling will not. Here the is coil is wound on a ringshaped core and the magnetic field very largely .10. utilized in various ways in practice. induced e.m. (a) Toroidal Coils toroidal winding A was illustrated in Fig. transformer). 4. Moreover the indeterminate length of the magnetic path makes inductance calculations difficult. or by employing a closed core of magnetic material. This can be done in two ways.f. of itself. as in radiofrequency transformers. These disadvantages may be overcome by providing a closed magnetic circuit. It will be observed that the magnetic field is not confined within the coil but spreads appreciably around the outside. to the transfer of energy from one circuit to another. 4. 1. by using a toroidal winding. 173 As the secondary is turned on its axis it links with more and more of the flux until the maximum position is reached when the coils are in line (Fig.
have an e. At radio frequencies the eddy. The design and construction however. as explained in the next section. behaves as if it were an assembly of a large number of shortcircuited coils. where it can be minimized by increasing the resistivity of any metal which has to be in the field. Special winding methods are necessary with a toroid because the spool holding the wire has to be small enough to pass through the centre aperture of the core. if the field is varying. transformers.174 RADIO COMMUNICATION confined within the core so that not only is the stray field reduced but the length of the magnetic path is definite. (b) Magnetic Cores Inductors wound on closed magnetic cores are very frequently used in communication engineering. for any metal in a powerful magnetic field may have induced in it eddy currents of sufficient intensity to make the metal red hot.4. It is not large at low frequencies. in each of which currents will circulate depending on the resistivity of the material.current loss is appreciable with . This is known as the eddy current loss and can be represented by an additional resistance in series with the normal conductor resistance. Now. Any metal object in the vicinity of the field will. . however.f induced in it which will produce what are called eddy currents. The metal. causing an increase in the effective resistance and a decrease in the inductance. is proportional to the magnetic field and the frequency and hence the eddy currents will be proportional to Bf/p where p is the resistivity of the material. and the energy loss will be proportional to the square of the current. it was shown in Section 2.m. in fact. another effect of great importance resulting from the presence of the magnetic field. though it does enter into the calculations on power and a. For this reason it is desirable to keep metal objects out of the magnetic field as far as possible. Hence the presence of any metal near a coil carrying a varying current will absorb energy from the circuit. involves appreciably different technique and they are therefore discussed at length in Section 4. The effect thus becomes increasingly important as the frequency is increased.m.f.f. This is particularly important with transmitting inductors.4 that the presence of a secondary winding coupled to a coil affects the primary impedance. The induced e. Eddy Currents There is. Consequently this type of coil is only used in special circumstances.
)
PRACTICAL COMPONENTS
quite small fields
175
and there
is
the further effect that the inductance
of the
coil is
reduced by the eddy currents.
Shielding
There is, however, an important use to which such currents may be put. The eddy currents will themselves produce a magnetic field which we know, by Lenz's Law, will be in such a direction as to oppose the original field. If we enclose a coil completely in a metal box or can, the eddy currents induced in the can will produce a counter magnetic field. Moreover the strength of the eddy currents will automatically adjust itself so that this counter magnetic field
virtually cancels the initial field at that point.
Metal can
Coil
Fig. 4.11.
Shielded Con.
Thus
outside the can the effective magnetic field is reduced almost
to zero so that
any coupling between the
original coil
and any
Inside the can, of course, the field is also reduced, but only by an amount equivalent to its value at the radius of the screening can. This is considerably less than the field inside the coil itself, so that the only effect is a slight reduction in the inductance and a slight increase in the effective resistance. This process of shielding or screening is of great usefulness in the
coils is largely eliminated.
adjacent
design of receiving equipment since it permits components to be placed appreciably closer together than would otherwise be possible. To be effective the shield must be such as to permit substantial eddy currents to exist i.e. it must be of highconductivity material, such as copper or aluminium, and it must have a thickness several times the skindepth at the particular frequency in use (see Section 4. 1
;
.
Variable Inductances
For many purposes it is desirable to be able to vary the inductance of a circuit. This is not so easily contrived as a variable capacitance, and is usually limited to relatively small variations.
1
176
RADIO COMMUNICATION
is to alter the number of turns on the This will normally only provide discrete changes and is little used except in transmitter inductors where the capacitance is fixed, so that tuning must be provided by varying the inductance (cf. Fig. 7.1). For such purposes the coil is constructed with bare wire or tube, with clip connections which can be adjusted as required. With any form of tapped coil allowance must be made for possible "deadend losses". The unused portion of the coil will constitute with its own self capacitance a circuit which is closely coupled to
The most obvious method
coil.
Coil
Ferrite slug
Fig. 4.12.
Permeabilitytuned Coil
the active portion, and may produce undesirable effects. In some cases it is preferable to shortcircuit the unused turns. For receiving circuits it is useful to be able to vary the inductance within small limits to allow for stray circuit capacitances. This is usually achieved by altering the permeability of the magnetic circuit. With a single or multilayer solenoid this can be provided by fitting a small ferrite slug within the former, as shown in Fig. 4.12. A screwdriver adjustment allows the position of the slug to be adjusted and so provides the required (small) variation of inductance. The variation obtained by this means is only a few per cent, which is sufficient for the purpose. Somewhat greater (and more precise) control of the inductance is obtainable with a completely closed ferrite core, as illustrated in Fig. 4.31, and this form of construction is frequently employed. Certain special applications, usually associated with measuring equipment, require an inductance capable of being varied over an appreciable range. This can be arranged by providing two coils, so mounted that the coupling between them is variable. The inducL2 L x \ Z2 2k\/(L1 L<s ). Such a tance is then L x device is called a variometer, a simple form being illustrated in Fig. 4.13 (a). With this construction the coupling factor between stator and rotor is only about 05 so that the total variation obtainable is only about 3:1, but by using modified construction (e.g. spherical former) the coupling can be increased and a range of 10 or more can be obtained.
+
± 2M =
±
:
PRACTICAL COMPONENTS
177
Rotor
(a)
Stator
(b)
Shaped
Disc type
coil
eddycurrent
plates
Fig. 4.13.
Two Forms
of Variometer
An alternative method is to utilize the change of inductance caused by eddy currents. This is known as spade tuning, suitably shaped metal plates being brought into proximity to the coil, which is usually of the disc type as shown in Fig. 4.13 (b). The range of variation with this type of variometer is only small, of the order of 10 to 20 per cent.
Losses in Coils
The main source of loss in an inductor is the ohmic resistance of the wire. This we have seen to be more than the d.c. resistance because of the skin effect mentioned in Section 4.1. The efficiency of an inductor is rated in terms of the ratio of reactance to resistance which is known as the magnification factor Lco/B. Where the coil constitutes part of a tuned circuit or a Q filter network the object is to make this factor as high as practicable. Where the inductor is being used simply as an impedance (as in a choke) a high Q is relatively unimportant.
=
In addition to the purely ohmic losses there are dielectric losses occurring in the material of the former and the insulation on the wire itself, plus any eddycurrent losses as just discussed. have seen that the eddy current loss is proportional to the square of the frequency; the dielectric loss varies approximately as the cube of the frequency. The dielectric loss can be considered
We
178
RADIO COMMUNICATION
as due to a lossy capacitance connected across the coil and, as was shown in Section 4.2, the equivalent shunt resistance of a lossy
capacitance Bj, across a
Rv =
coil
l/fcuC where of inductance L
y>
is
A resistance is the loss angle. equivalent to a series resistance
R = t&&HR9 +3wL)
—
G>
a
£ a /Rj>
if
Rv
is
large
compared with jcoL. Hence the equivalent
dielectric is
(o
2
loss resistance
due to the
L2
(o
3
L*Cip
l/yxoG
Hence
for a given coil the
Q
will at first increase
with frequency,
though not directly, since the ohmic resistance will increase approximately as v7 due to skin effect. As the frequency is further increased dielectric losses begin to be appreciable and the Q begins
to
fall again. If the coil has a core of iron or ferrous material additional losses will be caused by the presence of the iron in the magnetic field.
These are discussed in the next section.
4.4. Ironcored Inductors and Transformers For many applications it is convenient to provide a substantially closed path for the magnetic flux in an inductor or transformer. This not only permits a better utilization of the material but also renders the arrangement more amenable to calculation in certain
respects.
The issue is complicated, however, by the fact that the permeability of ferrous material is not a constant factor, like the permittivity of a dielectric, but varies with the degree of magnetization.
The Magnetic
Circuit
As explained in Section 1.4, the conditions existing in a magnetic circuit are in some respects analogous to the flow of current in a
conducting
circuit,
the appropriate relationship being
^
where
<t> is the flux, the reluctance.
M
8
(m.m.f.),
M
=
is
the magnetomotive force
S
is
The reluctance, which is analogous to the resistance in a conducting circuit, is 8 l\Afi, where I and A are the length and crosssectional area of the magnetic path, and (x, is the permeability. In an aircored inductor, or indeed in any system having an
PRACTICAL COMPONENTS
179
"open" magnetic circuit, it is difficult to define I and A, but where a substantially continuous magnetic path can be provided for the
flux,
much more
precise calculations
become
possible.
These are the conditions which exist in ironcored inductors and transformers. Fig. 4.14 illustrates a simple magnetic circuit and the meaning of I and A is clear. The m.m.f., is, by definition, equal to IN, where I is the current in the coil and is the number of turns. The magnetizing force,
M
,
N
H,
is
the m.m.f. gradient, or m.m.f. per unit length.
Hence the
relevant fundamental expressions are:
= MjS M = IN H = IN
<&
/I
Mean core
length I
£ ~Core area A
Winding
Fio. 4.14. Ironcored
Inductor
The
/u /u r .
flux
<I>
= HAfjt,
<D
where
(i
is
the permeability and equals
7
Substituting the value of
fi Q
= 477/IO
r
.
we have
and
= 1267 HAfj, 10" webers 10« = 1267 INA(/i B = Q>fA = 1257 (IN}L)ii 10 teslas = 1257 Hfi 10«
r ll)
.
6
T
.
r
.
BH Curve; Saturation
Now, as already mentioned, the permeability of iron and magnetic materials generally is not constant but varies with the value of B, which in turn depends on H. One thus needs to know the conditions under which the iron is working, and, to some extent, its past
history.
Kg. 4.15 (a) shows the magnetization or BH curve for soft iron, and also the variation of fiT with H. It will be seen that the value
180
RADIO COMMUNICATION
of rises rapidly at first and then much more slowly. This subsequent falling off is known as saturation. Magnetism may be considered as the result of an orientation of the orbits of the electrons in the atoms each of which forms a small electromagnet. Normally the distribution is random and only a small magnetic effect results, known as residual magnetism or
B
3000
15 
'0
500
WOO
H
(At/m)
1500
Fia. 4.15(a). Variation or
B and
[i r
with
H
Eltctrolytic iron
IS
10
OS
SOO
1.000
1,500
H (AT/m)
Fig. 4.15
(6).
BH Curves for Various Materials
remanence. The magnetizing force causes the orbits to align themselves so that their magnetic effects all tend in the same direction. At first this action is rapid, but after the majority of the orbits have been correctly aligned, increase in produces little effect and in the limit the only increase produced will be that due
H
H
PRACTICAL COMPONENTS
to the magnetizing force
saturated.
itself.
is
181
Such a condition
The material would then be fully rarely attained in practice, but it is
easy to reach a condition where doubling the value of only produces a 5 per cent increase in B. It will be seen that fiT rises to a maximum and also then falls off rapidly. We shall see later that the effective value of pr is much below the static value of several thousand shown. Fig. 4.15 (6) shows similar BH curves for other materials.
H
Hysteresis
A further peculiarity is that, after iron has been magnetized, the removal of the magnetizing force does not result in complete disappearance of the magnetism. Even if the magnetizing force is only reduced slightly the magnetism will be greater than it was for the same value of previously. If the magnetizing force is taken over a cycle from a given positive value to a similar negative value and back again, the value of B will follow a curious path as shown
H
in Fig. 4.16. This
is
called hysteresis
and plays an important part
in
the design of iron circuits.
Fio. 4.16. Hysteresis
Curves
In an inductor carrying an alternating current the iron is taken through this cycle many times a second. It will be clear that the effect of hysteresis is that some of the energy originally applied to the iron is stored in the molecular structure and before we can magnetize the iron in the opposite direction we must dissipate this energy. There is thus a loss of energy every cycle, which is known
as the hysteresis loss.
iron
maximum flux density p to which the magnetized, rather more than in direct proportion but not as much as to the square, while it is also directly proportional to the number of cycles per second. The loss is, in fact, of the form kfBpia where A; is a constant depending on the material. It is found that the loss is reduced by incorporating from one to four
It is proportional to the
is
B
W=
182
RADIO COMMUNICATION
per cent of silicon in the iron, and such material, known by various trade names such as Stalloy, Silcor, etc., is normally used for transformers and chokes. It is put up in the form of thin sheets as explained later under the heading of Iron Loss.
Energy in a Magnetic Field We saw in Chapter 1 that the work done in establishing a magnetic field is stored in the field and it was shown that the value of this stored energy in an inductance carrying a current / is \LI*. We can rewrite this expression in terms which do not involve the inductance or the current, as such, and so obtain an expression in terms of the 2 field strength. We know that L = [iN All, where A and I are the circuit; also the magnetizing force area and length of the magnetic H = IN/l. Hence
\LI*
=
\
fiNUP/l
=
=
W*PPIP)Al
%/j.H 2
yt
= \pB*M
we have simply
j/j,
But Al
is
the volume of the magnetic circuit, so
Energy per unit volume
= \B
2
= \BH joules
(In these expressions, of course,
=
fi fi r )
Remanence; Permanent Magnets
It will be seen from Fig. 4.16 that if the magnetizing force, having been increased to a maximum, is now reduced to zero, the value of B, which follows the upper of the curves on the hysteresis is decreasing, will not fall to zero. This residual value loop when of B is called the remanence. To reduce it to zero a negative value will have to be applied, and this is called the coercive force. of For normal core material both these quantities should be kept low because the area inside the hysteresis loop represents energy absorbed by the iron; but conversely, if one wants to utilize this property of remanence the material should have a large BH loop. Carbonsteel is such a material, having a loop of the form shown in Fig. 4.16 (6). It requires a large coercive force, but after the removal of the magnetizing force it exhibits a large remanent magnetism. Such an arrangement is called a permanent magnet. Various materials have been developed which have a very large hysteresis loop for use as permanent magnets. The effectiveness of a material for this purpose is not determined solely by its remanence, but by the energy stored, which is proportional to the area of the hysteresis loop. This is controlled not only by B„ m but also by the both of which should be large. The performance coercive force e
H
H
H
,
PRACTICAL COMPONENTS
of a permanent magnet
4.17,
is
183
thus dependent on that portion of the
hysteresis loop lying between these
which
value of
B
two values, as shown in Fig. as the demagnetization curve. For any given the energy, as just stated, is \BH and it is customary to
is
known
xlO 3 10
30
20
10
20
40
H (AT/m)
Fig. 4.17.
BH
xl0 3
Demagnetization and Energy Product Curves for a Permanent Magnet
rate the effectiveness of a magnetic material in terms of its energy product BH. This is shown on the righthand side of Fig. 4.17 and will be seen to have a broad maximum at about 80% of rm This value of max is quoted as part of the specification of the material, some typical values being shown in Table 4.4. (The optimum value
B
.
BH
Table
4.4.
Permanent Magnetic Materials
Coercive
Energy
Material
Force
Bemanence
Product
At/m
Tungsten
steel
T
(BH)^
joules
(6% W, 07%
Cobalt steel (16% Co,
C)
.
5,200
.
105
2,400 4,800 35,000
9%Cr, 1%C)
Ni,
14,300
08
(24% Co, 14%
8%
Al,
3%
Cu)
.
43,500
125
184
RADIO COMMUNICATION
of B, for which
intersection
BH is a maximum, is given approximately by the between the demagnetization curve and a line from
BremH
e,
zero through
as
shown dotted
in the figure.)
Design of Magnet Systems
Permanent magnets are mainly used to provide a magnetic field Thus in a movingcoil meter (see Fig. 4.43) the iron circuit consists of a permanent magnet supplying the field, a softiron core with associated polepieces, and an annular gap in which the coil moves. The design of such a system involves the determination of the size of magnet necessary to provide the required
in various electronic devices.
flux in the air gap.
This is assessed from the demagnetization curve. Suppose, for example, that it is required to produce a flux density of 08 T long and 08 cm 2 crosssectional area. The in a (total) air gap 2 \Bg gA g lg This must equal the energy in the energy in the gap Bg and A m A g then magnet \B m m m If B m
=
mm = H = H AJ HJ m = Hg g =
. .
=
,
l
(Bgl/i
will
)l„
(1)
From
the curve of Fig. 4.17
it
be seen that to produce
B = 08T, H = 39,000 At/m.
*
Whence
o OD =3 28cm

= 08
X 2 X 10 3 X 10 7
4^x39,000
Now, the magnet crosssection may not be the same as that of 08 is the air gap, in which case m will not equal B g In fact, B m not the optimum for the material of Fig. 4.17, which occurs when B m 1. This higher value of B m can be obtained by reducing A m by 08, so that it becomes 064 cm 2 The length will then need to be increased, though not by 1/08 because of the more efficient 10T, 34,500 At/m so that condition. From the curve, if m
B
.
=
=
.
B =
2
H—
HJm = 34,500 and
l
m
= (08 X = 37 cm
steel
3
X
37
10 3
X
10')/4tt,
The volume of
3*28
is
X
064
—
237
cm 3
,
as
against
is
x
08
= 262 cm
for the original condition, so that there
a
saving of approximately 9 per cent. Actually some m.m.f. will be required to force the flux through the softiron pole pieces and the central cylinder. This can be allowed for by replacing lg by (lg l t lpr), where l { is the total length of the softiron material (assumed to have the same cross section as the air gap) and /ur is the relative permeability of the iron at the
+
PRACTICAL COMPONENTS
value of B„ (08 T).
185
le
and (it = 1,000, example chosen.
ljfir
— 50 mm This correction is small. If = 005, which is only 25 per cent of in the
l
g
It is therefore
lg
customary to allow a small arbitrary
to allow for the pole pieces.
increase of a few per cent in
Ironcored Inductors
Let us now relate these data to the design of ironcored inductors, often called chokes. For a given flux the inductance, assuming no leakage, is ON/I. have seen that
We
so that
'
0= 1267 INA/irfl L = 1257 JVM^/Z
.
.
10«
.
10~ 8 henrys
80 r
£60
<a
40
.£ 20
20
d.c.
40 (mA)
60
80
d.o.
Fig. 4.18. Illustrating Variation of
Inductance with
Current
Thus the inductance is proportional to the square of the turns, the area of the core and inversely proportional to the length of the core. It is, however, directly dependent on /nr which we have seen to be a very variable quantity depending entirely on the degree of magnetization. Thus the inductance of a choke carrying an alternating current is varying from instant to instant and we can only assess some average value. Moreover, a choke is often required to carry a steady d.c. (though this is not always the case). It thus has a steady magnetization with a ripple superposed. The permeability of the iron will vary with the amount of steady magnetization rather in the manner shown in Fig. 4.15 so that (except possibly for very low values of H) the inductance falls off steadily as the d.c. increases. This effect
,
is
known as saturation and is illustrated in Fig.
4.18.
Sometimes
this
186
effect is of use (as in
rectifier
RADIO COMMUNICATION
a "swinging" choke for use with a chokeinput system) but as a rule it is preferable to avoid this dependence of inductance upon the current. This is the more so since the permeability is affected not only by the d.c. but also by the a.c. component, the inductance tending to increase as the ripple becomes
larger.
Incremental Permeability
The
is
effective permeability, in fact, is
not the static value for the
if
particular value of
H.
It is
found that
an
a.c.
component of
H
superposed on a steady value the iron goes through a small hysteresis loop as shown in Fig. 4.19, and the effective permeability
Fig. 4.19. Illustrating
Incremental Permeability
this loop
of the iron
is
clearly
dependent on the average slope of
(the dotted line in Fig. 4.19).
It will be seen that this incremental permeability, as depends upon both dc and ac In general
it is called,
H
H
.
(i)
fij
(ii)
/j, t
H increases as H
decreases as
Fig. 4.20
de
increases;
increases.
ac
The curves of
sheet. Similar figures are published
show values of //< for typical siliconsteel by the makers for other materials
including the special highpermeability alloys such as Radiometal
and Mumetal.
PRACTICAL COMPONENTS
187
100
200
500 1000 2000 SOOO 10000
Fig. 4.20.
Ba.c. Peak Typical Values op Incremental Permeability
of Air Gap The effects of varying permeability may be offset to a considerable extent by including in the iron circuit a small air gap. Even a small
Use
(of 025 to 1 in the average choke) has a reluctance many times the iron, so that the performance is determined mainly by the air gap, which is of constant permeability. The flux is reduced because of the increased reluctance, but the permeability increases to an extent which may more than compensate for this, so that the inductance with an air gap is often greater, for a given d.c. through the winding, than with no gap.
gap
mm
shows a magnetic circuit having an and an air gap la The reluctances of these paths are IJAfiofii and IJAfi^ respec= <M, so that tively. The m.m.f.
Fig. 4.21
iron path
lt
.
Pig. 4.21.
M
Gapped Core
M =
t
<S>li IAfj,
[j, i
and
M = mjApt
2
Hence to obtain a given value of O the
is
M + ^gx
total ampereturns required
The
total m.m.f.
=M +M
x
2
may
be written in the form
so that the effect of the air gap
may
length of the magnetic path which is usually valid).
by
n$,a (assuming
be taken as increasing the A to be the same,
188
RADIO COMMUNICATION
number of turns and the current through remain unchanged the flux will be reduced in the ratio (tJa) If we vrish to maintain the flux at its original value the hl(h ampereturns must be correspondingly increased. In the case of a choke we do not wish to do this in fact, the object of introducing the gap is to reduce the flux but with some airgap circuits, such as the field of a loudspeaker "pot," we do wish to maintain the
It will be clear that if the
coil
the
+
—
—
flux,
and
this
method of calculation
applies equally to either usage.
Optimum Air Gap Now let the ratio
magnetic circuit is of the choke
l t (l
JJZ
= p.
Then the
effective length of the
+ ndp)HAn/if
We can thus write for the inductance
\2&An%
L
The
=
np
t
h( l
+ PilP)
h
'P
+
f*i
as
first term is constant, while the second depends on p. But reduce p (increasing air gap) (i t increases and vice versa. In fact, over a wide range of practical values we are not greatly in error by assuming the product /i (p to be constant. On this assumption we may write fi { kip. The second term in kjp). If we the expression for inductance then becomes l/(p plot this expression in terms of p we find that it rises to a maximum and then begins to fall off again, and we find (either graphically or by calculus) that this maximum occurs when p ft t This is another way of saying that the reluctances of the air and iron paths are equal, which the reader will recognize as a condition
we
=
+
.
=
internal
often found in electrical practice (e.g. maximum power and external resistances are equal). The value of
course, is the incremental permeability.
when
jj,
u of
The
best results therefore are obtained with a gap ratio approxilt
mately equal to ji^ With a choke having this would give a gap of
= 20 cm and
fi t
= 400
p
= p = 400 =
t
l{ /la
= 20/Jo
whence la = 05 mm. With such a gap the inductance, for a given Hdc> ^ould not differ greatly from the value at the same H& without a gap. In most cases it will be found to exceed the ungapped figure and, of course, the inductance with the gap is much more constant despite variations both of the d.c. and the a.c. ripple current through
the choke. It is not practicable to express this improvement in simple terms because the various factors are so interdependent. The simple
PRACTICAL COMPONENTS
189
result just obtained for the optimum gap only arises from the assumption that n tp k, which is by no means rigidly true. The practical designer, however, is able to make his calculations with comparative ease by the use of tables or curves drawn up partly from more extended theory and partly from empirical data.
=
Pull of
Electromagnet
gapped inductor is the relay or electromagnet. Here we are interested not in the inductance but in the mechanical force developed. This can be assessed as follows. Consider two magnetic surfaces, of area A, separated by a distance x. If the flux density is B, the energy per unit volume is ^B^J/jlq. If now the distance between the surfaces is increased by a small amount dx, the work done in separating them will be F6x. This must equal the increase in energy in the field. Hence
particular case of the
A
Fdx
=
(hB*l(i
)
x Adx
so that the magnetic pull will be
F=
Prom
lengths of the air
calculable.
iB^A/jUo
= \BHA newtons
the configuration of the magnet system and the relative and iron paths the flux density can be evaluated in terms of the m.m.f., and the magnetic pull is then immediately
Transformers
As explained in Section 2.4, a transformer consists essentially of an arrangement in which two coils are magnetically coupled to one
another. If a varying current is passed through one coil, known as the primary, the changing flux linking with the second coil, called the secondary, induces a voltage therein. In a perfect transformer all the magnetic flux produced by the primary should link with the secondary. This ideal state of affairs cannot be attained, but for power frequencies and audio frequencies it is possible to approach the ideal by arranging the coils around an iron core. Since the permeability of iron is many times greater than that of air the lines of magnetic force lie mainly within the iron, and if both coils are wound on the same core the greater part of the flux produced by the primary links with the secondary and vice
versa.
Consider a simple transformer as in Fig. 4.22, consisting of a
primary winding and a secondary winding both wound over the same iron core, with the primary connected to a source of alternating
190
e.m.f.
will
RADIO COMMUNICATION
The alternating current flowing through the primary winding produce an alternating magnetic flux. This flux linking with the turns of the secondary winding will induce e.m.f.s therein proportional to the rate of change of flux and the number of turns in the winding. If the flux is varying sinusoidally the secondary voltage will also be sinusoidal so that as long as we ensure that the
Fro. 4.22.
Diagram or Simple Tbansfobmeb
flux
wave is not distorted the transformer will deliver a sine wave output on its secondary. It is convenient to work from this assumption of an alternating flux as a starting point. This flux will, of course, induce voltages in both primary and secondary windings. Therefore, in the first place if we are to maintain the flux we must apply across the primary a voltage equal and opposite to the back e.m.f. induced in the primary. As long as we continue to do this the flux wave will be maintained and this flux linking with the secondary will induce a corresponding
voltage.
Let
$ = <&„ sin mt be the flux, where $>P is the peak flux.
Primary back
= e = —N d^Jdt = — N$>vm cos mt Secondary e.m.f. = e = — N%<$> v w cos mt
e.m.f.
x x 2
Thus the ratio of the e.m.f.s on the primary and secondary windings is equal to the turns ratio. In practice we design the transformer so that e x is equal to the applied voltage v. Now e x depends on the turns, the frequency of supply /(= eo/277) and the flux. If the first two factors are fixed the flux will adjust itself until c x v (numerically). We know that ®„ B^A, where Bj, is the maximum, or peak flux density; so we can write e 1 B^An^m cos mt. Hence the r.m.s. primary e.m.f. E x Bp 1 ml\/2. Writing
=
=
=
eo/\/2 as 2nf{y/2
— = 444/, we have E = 4448^^
x
AN
PRACTICAL COMPONENTS
191
We
This is the fundamental expression used in transformer design. choose a suitable value of Bj,, usually around 1 tesla, and then adjust the area of core A and the number of turns N,_ to make E x equal to the input voltage required (e.g. 240V). The secondary
turns N% then E^NJE^ subject to a correction for internal voltage drop as explained later.
Effect of
=
Secondary Load
if
What happens
we connect a
load across the secondary?
We
know, from Lenz's law, that the e.m.f.s of self and mutual induction are in such a direction as to oppose the change of flux which generates them. Hence if we connect a load across the secondary a current will flow = EJZ 2 which will be in such a direction as to reduce
the flux in the core. But once this happens, the primary e.m.f. E1 is no longer equal to V and hence more current will flow in the primary until the flux
restored to its original value. The secondary ampereturns producing the demagnetizing effect is I%N 2 and this must be counteracted by an equal and opposite (additional) primary ampereturns I t v Hence I1 1 ^^2, and the reflected load current
is
,
N
N =— I =  {NJNJI,
x
Magnetizing Current
There must be some primary current, even when there is no load on the secondary, in order to maintain the flux. This initial current is termed the noload or magnetizing current. It is nearly 90° out of phase with the primary voltage because the primary is almost wholly inductive. There is a small resistive component due to the winding resistance and the losses in the iron, as we shall see later, but in general this is small. The total primary current is thus the vector sum of the magnetizing current and the reflected load current which is —(NJNJIz and is thus 180° out of phase with the secondary current. Fig. 4.23 shows
tive
the simplified vector diagrams neglecting losses, for resistive, inducand capacitive secondary loads. The primary voltages and load currents are in opposition to the secondary quantities in each case. Under noload conditions there is no secondary current and the only primary current is the magnetizing current Im This produces in phase with Im and since this is common to both windings, a flux are induced lagging 90° behind Im The primary and e.m.f.s
.
O Ex
,
E%
.
voltage must be equal and opposite to
E x as shown.
24 illustrates the equivalent circuit of a transformer. has a certain resistance so that the voltage actually applied across the transformer is V x IxB^ The secondary e. that the connection of a load across the secondary of a transformer results in a secondary current which is reflected into the primary. because as soon as we begin to draw a current from the transformer the internal impedance of the device begins to take effect and we find that the output voltage is no longer in simple turns relation to the input voltage. 7/ lags behind V by the same amount as I2 lags behind E 2 while Ix lags slightly more because of Im With a capacitive load the reflected primary current leads on V.m. With an inductive load. 4. . I2 is in phase with E2 and Ix ' is 180° out of phase with E 2 and hence is in phase with V.f. With a resistive load. V. Fig. This is only part of the story.23. the ratio of the currents being in the inverse ratio of the number of turns. The total primary current I x is then the vector sum of Im and 7/ as shown.h<» = £. therefore. will be E2 I1 R1 )N2 /N1 and the (Vx — — — . I2 N2 'I N. c2 Resistive load c2 Inductive Capacitive load load Typical Tbansfobmeb Vector Diagrams Equivalent Circuit of Transformer We see. = Irr. Uf\ ^ N No load Fia. We commence with the voltage Vlt which is the input voltage. 4. but at a slightly lesser angle than I 2 because Im is lagging.192 RADIO COMMUNICATION . The primary winding. lagging slightly behind V. . however. however.
€. Equivalent Transpobmeb Regulation Because of these internal voltage drops. k»_ «. which are clearly dependent upon the current flowing. so that It will be seen.m.f. have seen that the object of winding the coils on an iron core is to arrange that all the flux produced by the primary shall link with the secondary and vice versa. the voltage output from the secondary of the transformer is not constant but falls steadily as the load increases. . but as the transformer increases in size it is obviously necessary to restrict the losses as otherwise the heat generated becomes considerable. Ei=VriiRi"£i<gE2=ir " o Circttit of Fia. As the size of transformer increases and the quality becomes better the regulation improves to something like 5 per cent.24. and vice versa. This object is not completely We maintained and a certain amount of the flux on the primary does not link with the secondary. I VWv 2"^2~h R2 "2 v.PRACTICAL COMPONENTS 193 secondary output voltage will be less than this amount because of the voltage drop on the secondary resistance. Leakage Inductance The resistance in the winding is not the only source of voltage drop. while with a large transformer the figure may be less than 1%. If the transformer is a small one it may be that these losses can be tolerated. 4. A small radio set transformer may have a regulation of 15 to 20 per cent. The fall in voltage from no load to full load expressed as a fraction of the fullload voltage. therefore. This figure varies with the size and quality of the transformer.s and is actually appreciably greater than F 2 /Fx to allow for the internal voltage drops. It is obviously desirable to keep the regulation good because the resistance of the windings not only causes a voltage drop but actually occasions heat losses which are wasteful. is known as the regulation of the transformer. that the turns ratio must be calculated on the internal e.
If J2 makes an angle ^ with V t .f. It is clear that there must be appreciable flux linking the primary which does not link the secondary. Strictly speaking the leakage inductance should be introduced in series with both primary and secondary .s into the circuit.m. The applies to the secondary. which may not be valid. generated by the main flux cutting the same turns there will be a subsidiary e. and the simplest way of representing this effect is to introduce into the equivalent circuit a small inductance as shown in Fig. 4.194 RADIO COMMUNICATION This is known as the leakage flux and the effect will be to introduce additional e. for example./. This has been done in Fig. the second term.circuit laws Rt = ' + h' = h + i? 2 l WW 2WW 1 With this arrangement Vl but V2 =B = E .f.25 where both the leakage and the winding resistance have been represented as being in the secondary circuit.m. In the primary. representing the voltage drop.' + jot. — R2" R2 + (^ff R > /vw aJfflT 1 Fig.m. 4. By the usual coupled. Effect of Leakage Inductance but it is usually adequate to represent the leakage flux by a leakage inductance in series with one winding only. We can improve matters by splitting the winding and putting * This assumes a resistive load.25. due to the leakage flux.(*. Fig. . Here the primary winding is placed on one limb of the core and the secondary winding is placed on the other. )* 2 Transformer Construction The amount of the leakage flux depends on the construction of the transformer. in addition to the main e. 4. 4.26 (a) shows a type of transformer which is bad.f. should be I t R t cos Unless 4> is <j>\Il%(o sin <f> large the difference is not serious.25.
Core and Shell Types or Transformer adopted. Here a threelimb core is S P Insulation Fig. for a high. 4. This type of construction. is . 4. This interleaving or sandwiching process clearly limits the stray flux and thus reduces the leakage inductance. and since the flux has two alternative paths it is clear that the outer limbs. but if still lower leakage is required still further precautions are required and it is necessary to interleave the windings in the manner shown in Fig.26. For small powers the arrangement of Figs. need only be half the thickness of the centre limb. Thus. has the advantage that only one set of windings is needed.26 (c). This type of transformer is used for medium and high powers and known as a coretype transformer. 4.26 (d). The sections may be in series or parallel as required. known as the shell type. the secondary being wound over the primary as shown in Fig. 4. as well as the top and bottom limbs.PRACTICAL COMPONENTS half the primary 195 on each limb of the core and winding half the secondary over each section of the primary. Leakage inductance is difficult to calculate and in most design work practical experience is taken as a guide.voltage transformer we might connect the primaries in parallel and the secondaries in series. Both these arrangements provide greatly reduced leakage inductance. Whichever connection is used we must ensure that the directions are correct so that the current at any instant is such as to produce a magnetic field which is in the same direction around the core.26 (c) and (d) is more usual.
while the material itself is more expensive and the stamping of the laminations is more costly owing to the higher precision required in the tools. The effect of hysteresis was discussed on page 181. The resistance of the windings causes heating. The losses are of two main forms. though sometimes 0020 in. Typical a Forms of Lamination thin sheets or strips. They are assembled with the joints U alternately at each core is end so that the effect of a substantially solid obtained. The cost is greatly increased. For higher frequencies or special circumstances requiring an unusually high value of Bv the stampings may be 0007 or even 0006 in.. In addition. Eddycurrent loss is proportional to B^f 2 and thus becomes increasingly important as the frequency is raised. 4. the core is built up of currents set up a a 2a a a a la a a Fio. therefore. In a practical transformer. This eddy. . h A +  W = ' . Otherwise burrs are formed at the . where it was shown that it produces a loss of the form 1 9 There is a further loss due to the circulating eddy kfBj.current loss would be quite serious if a solid core were used. Copper Loss.196 RADIO COMMUNICATION Transformer Losses practical transformer does not transfer all the primary energy to the secondary.27. The usual thickness for 50 Hz is 0014 in.27. which not only reduces hysteresis but has a high electrical resistance so that the eddy currents are still further reduced. insulated from each other by a thin facing of paper or paint on one side. the loss due to this being I^Bx /22 i?2 Iron Loss. thick. 4. we have the loss in the iron core due to the varying magnetization. It is also dependent on the thickness of the sheet. In addition. but built up of insulated thin laminae so that any eddy currents induced can only flow in restricted paths. is used. in the iron. for more stampings are required for a given area of core. For small transformers the laminations pieces or sometimes E and are stamped out in the form of T and I pieces as shown in Fig. some energy being lost in the process. the sheets are made of silicon steel.
. The total iron loss for typical material is shown in Fig. The measured loss is then p krf {. = . If there is reason to believe that it is of importance. so defeating the object of the use of thin sheets.28. Vx must be twice as great in the 50 Hz test. Hence if we use /x 50. 4. . 4. insulated joints must be provided where the bolts come through the clamps and similar precautions taken wherever practicable to reduce the number of complete conducting paths which might permit eddy currents. The steel manufacturers market material under various trade names in which the hysteresis and eddycurrent losses are of the same order at normal frequencies. = = .k^f found for two values of/ two equations are obtained which can be solved for &x and k 2 It is important that B p shall be the same throughout so that the applied voltage must be proportional to 25 and /2 frequency.28.) If it is desired to separate hysteresis and eddycurrent losses this test must be repeated at two different 2 If this is frequencies. (This is not VI because 7 the magnetizing current. Even with normal sheets the presence of burrs is to be avoided. being a direct result of the magnetization of the iron which does not change with load. It may thus be measured by finding the power taken by the transformer on no load. There is also a small eddycurrent loss due to circulating currents in the clamps used to hold the whole structure together but this is 2 2 Si 1 n 06 10 b p m 12 Pi Fig. Total Iron Loss fob Typical Transformer Sheet usually negligible in small transformers. is nearly 90° out of phase. Analysis of Losses The iron loss is present all the time.PRACTICAL COMPONENTS 197 edges and these pierce the insulation when the laminations are stacked together. .
The losses.28 and calculating the copper loss from the resistance of the windings. though it is approximately so over the initial portion of the curves. but this one in which the iron loss and copper not be the most economical design. in radio practice. the ratio of output watts/input around 80 to 90 per cent. as is shown in Fig. the efficiency as such is not as important as ensuring that the losses shall not cause undue temperature rise.29. a transformer For small transformers it is is . and if Bp is too large. Thus a transformer having dimensions 2 of 15 X 12 X 10 cm would have a surface area of 900 cm and thus about 45 watts. These tests require a wattmeter. The power taken by the transformer in this condition is the copper loss because Vx will be so small that p. with corresponding increase in iron loss.198 RADIO COMMUNICATION The copper loss varies with I2 and is taken at full load. Iron is cheaper than copper and it may pay. In many cases. either estimated or measured. say. but is influenced by the permissible iron distortion. In such an instance the copper loss might be only The most efficient design is may half the iron loss. 4. may be estimated with reasonable accuracy by calculation. 4. B and hence the iron loss. As a good practical rule it may be taken that a transformer can safely dissipate 005 watt per square cm of its surface. rising to 96 per cent for medium transformers and 99 per cent or even higher for very large transformers. to increase the area of the core 50 per cent. Iron Distortion The choice of the maximum flux density is not entirely determined by the losses. the secondary e. which may not be available. Too small a . will be negligible. giving serious distortion of the flux wave.16 shows that the B and is not linear. however. loss are equal. basing the iron loss on data similar to that of Fig.f. One method is to shortcircuit the secondary and to apply a reduced voltage to the primary of such a value that the secondary current is the required value. 4. is far H from It sinusoidal. This would permit the number of turns to be reduced with saving of copper. relationship between Reference to the magnetization curves of Fig. Hence it could be used could safely dissipate to handle a power of about 500 watts with an efficiency of 95 The efficiency of watts.m. per cent. mum customary to operate under conditions such that the maxivalue of B is on the "knee" of the BH curve. We have assumed that the flux is sinusoidal but in fact a sinusoidal magnetizing current will produce a flattopped flux wave.
say. however. slightly higher values of BP being permissible in a power transformer than in an audiofrequency transformer where negligible distortion is BP does not utilize the iron to maximum advantage.000 Hz are sometimes adopted is for special applications. Mognttizfng current Flux wove AA Fig. while America has adopted 60 Hz. Hence if the frequency is. From the fundamental transformer equation it will be seen that for a given maximum flux density in the iron the product ANf is constant for any particular value of voltage. The choice is thus a compromise. The iron loss.PRACTICAL COMPONENTS value of requiring 199 more turns on the winding. for 50 Hz (or both increased so that the product Conversely. so that some improvement is obtained by reducing the thickness of the laminations.29. It is the eddycurrent loss which preponderates. Secondary Illustrating Iron Distortion Frequency The standard supply frequency in Britain is 50 Hz. if the frequency is increased the volume of iron required AN reduced. 25 Hz the core area or the number of turns will need to be twice the value is doubled). which is usually operated at a frequency of 400 Hz. increases with frequency so that it is usually necessary to reduce the flux density somewhat. 4. Power transformers are thus normally required Effect of to operate at frequencies of this order. This is useful where weight is of importance. as in airborne equipment. material . Too large a value of Bp introduces increasing distortion. while frequencies as high as 4. required.
000 Hz. as discussed in Chapter 10. be necessary to sandwich the windings if good regulation is essential. component by a parallelfeed arrangement. the primary being designed midway between zero and the knee of the to produce a value of BH curve. or 0005 in. 10. many requirements a range of 50 to The same basic considerations apply. and its effect is to limit the available magnetization swing. 4. but this is to some extent offset by the smaller physical size which of itself reduces the leakage inductance. .30.000 Hz 20. the effect of leakage inductance increases directly as the frequency. This applies to both valve and transistor circuitry. stampings customarily used at 50 Hz. the iron circuit is designed to handle the maximum voltage swing required at the lowest frequency. the value of p being chosen with regard to the permissible iron distortion. Audiofrequency Transformers Transformers are used in audiofrequency amplifiers for interstage couplings and for matching the loudspeaker or other load to the output stage.200 RADIO COMMUNICATION 0007 in. or 0020 in. As already mentioned.c.9. as in Fig. though for is sufficient. in which the currents in the two halves of the primary are in opposition.) required. thick is frequently used in place of the 0014 in. however. The design of interstage transformers is affected by the fact that the primary will normally carry a steady direct current with the signal current superposed. (This limitation does not apply with a pushpull output stage. Such transformers have to operate over a range of frequency which may run from 20 to Statloy Radiometal Permalloy 50 75 100 H Fig. It may. B B This necessitates a larger core size than would otherwise be and it is often convenient to remove the d. Typical (At/m) BH Cttrvbs iob Coke Matbbials 10. as is discussed in later chapters.
5.30 and Table 4. 10 to 20 times greater than ordinary transformer steel. Moreover.000 A 5Cu 785 Ni: 216 Fe 25. so that although the material is more expensive its use is still economical. but the permeability was low. A . Attempts were made to employ closedcore techniques by using irondust cores constructed of finelydivided iron particles bonded together under pressure with a suitable binding medium.PRACTICAL COMPONENTS 201 The physical size. Table 4. Characteristics of some of these materials are shown in Fig.000 50 24 195 08 16 100. which minimizes the leakage inductance. Highpbbmeability Materials Coercive force Material Composition Permeability (%) (max) Saturation induction (T) (At/m) Stalloy Mumetal Badiometal Permalloy 96Fe:4Si 76Ni:17Fe 5Cu:2Cr 50Fe:45Ni: 7. These are chemical compounds which exhibit appreciable permeability and at the same time have a high resistance so that the eddycurrent losses are reduced. Ferrites considerable improvement resulted from the development of materials known as ferrites. Certain nickel alloys exhibit permeabilities. Consequently for many years coils and transformers for radio frequencies were air cored. Their performance is critically dependent on the heat treatment during manufacture and they should not be subjected to mechanical stress during assembly.5.000 90. of the order of 1020 only.000 150.000 24 4 8 11 15 Perminvar 45Ni:25Co: 30 Fe Radiofrequency Transformers At frequencies above about 50 kHz the iron losses with conventional materials become prohibitive. can be considerably reduced by the use of special core materials. at low values of H. and had to be enclosed in screening cans to limit the radiation as explained on page 175. 4. however. the number of turns is reduced. This permits the required performance to be achieved with a much smaller core size.
whereas the nickelzinc ferrites have lower permeability but much higher resistivity and can be used up to 100 MHz. they crystallize in a cubic system which is not only mechanically strong but still magnetically homogeneous. The broad division between the two types is that the manganesezinc ferrites have high permeability (6001. Fig. usually known as magnetite or lodestone) by an atom of a divalent metal such as copper. factor. the resulting flux is core will remain in this state indefinitely. On the left are various forms of transformer core. or "hard" as in a permanent magnet. usually magnesiummanganese ferrite and copperferrite.f.e. is applied to an and the inductor wound on such a core.400°C) and allowing them to cool. while some ring and rod forms are shown on the right. high stability Storage Ferrites Certain ferrites. 163. 0001 . 4. manganese.000) and low losses below about 05 MHz. having low remanence). 4. may be made to have substantially rectangular hysteresis loops.202 RADIO COMMUNICATION In simple terms these ferrites are produced by replacing one of the iron atoms in ferric oxide. and these cores are used for a variety of purposes either in place of conventional magnetic materials or in applications utilizing special unique properties.8. Magnetically soft ferrites are most often found in the form of cylindrical "pot" cores for highquality inductors. In the centre is a selection of storage cores. transformer in more detail. If a magnetizing pulse manganese +B . Fe 3 4 (which is a magnetic iron ore. These are the manganesezinc ferrites and the nickelzinc ferrites. thus producing a compound of the form MFe 2 4 It is found that such a material is . Fig. and the properties of these two types depend upon the proportions of the principal oxides used in the composition and the manufacturing processes employed. There are two basic types of magneticallysoft ferrites in common use today. strongly magnetic but that its resistance is many thousands of times higher than pure iron. Ferrite materials may be magnetically "soft" (i. picturetube deflection yokes and rod aerials for domestic radio receivers. When a magnetizing pulse \H. cores can be produced in a variety of shapes from simple rods to completely closed structures. By sintering these materials at a high temperature (1. nickel or zinc. illustrates some typical ferrite constructions. combining high Q and the facility of simple inductance adjustment. opposite p.31 shows the construction of a typical r. With these materials. but they are also used in large quantities for television linescan transformers.
31. This ferrite has a much higher coercive force than metal magnets and can thus be used to make shorter magnets which have a greatly improved resistance to demagnetizing fields.F. Ferbitb R. as loudspeaker and headphone magnets and in many other applications where magnets are required to have high coercive force and high resistivity combined with small size and light .PRACTICAL COMPONENTS 203 and again remain applied the flux will change to Thus such a device has two stable states and can be used to store information in a binary code. 4. / ?/////////7tt\ ~ S°jL forme. the hysteresis loop is not truly rectangular but the ratio between the maximum flux B max and the residual flux Br is of the order of 095. Transformeb The connections are brought out to the tags in the base plate through slots in the sides of the ferrite core pieces (not shown). Barium ferrite magnets are used in synchronous motors. Threephase Transformers Power transformers are often required to work on threephase systems such as are described on page 209. indication of the state of the core. Construction op Typicai. so the ratio of flux change is adequate to give a clear is —H now —B in this condition indefinitely. Adjustable core for fine tuning' Conical spring washer Container Windings 2g ^\s^s^\\\\v\Sg^:\v\\\\\v\<\^\^T Insulating base plate carrying connection tags Fig. ferrites and small toroidal cores of these form the elements of many computing circuits. weight. Such transformers may . In practice. Magnetically hard ferrites are also made usually consisting of barium ferrite BaFe 12 19 which differs from the magnetically soft ferrites in having a hexagonal crystal structure.
The back e.5. 4. however.32.32. in the opposite direction. .m. that the opposition to the red phase voltage is Yj2. 4. therefore. so that the flux in the second and third limbs is exactly what is half the value of the red phase voltage and It will be seen from the vector diagram. the three phases are accommodated on a special core having three limbs of equal crosssection. being driven by a suitable prime mover such as a petrol is By engine. Alternatively. Generators. Motors and Meters far the greatest source of power for communications equipment the electricity supply mains. the generator may be located at the site or in the equipment itself.f. Such a core is illustrated in Pig. for mobile equipment. Often. The area of each limb is therefore designed in accordance with the same principles as for a singlephase transformer and the windings are calculated accordingly. Here electricity is generated by dynamos or alternators and is then distributed throughout the country by transmission lines and ultimately brought into the user's premises in a convenient form. and similarly for the blue phase. induced in R B' "St* ft Y/2 Y B Fig. Let us assume that the red phase is at its maximum voltage.204 RADIO COMMUNICATION be made up as three separate singlephase transformers connected in star or delta as required. 4. This will produce full flux in the middle limb and this flux will complete its circuit by passing half through each of the other two limbs. however. component of the yellow phase voltage in is required. Thbbephasb Cobb these limbs.
leaving only a small clearance to permit the central rotor to revolve freely. The speed is determined by the frequency required. v = V sin (at. It is then possible to include two coils. and v = velocity at which the conductor moves at right angles If the coil is rotating. where to is the angular this will velocity of the rotation.34.s. 4.f.p. Small alternators are made up in this form.m.PRACTICAL COMPONENTS 205 in field The simplest form of generator is a coil rotating as shown in Fig. either in series or parallel as required.34. Still better utilization of the rotor or armature is obtained by distributing the windings evenly round the the whole periphery in a series of coils housed in slots and suitably 4. The frequency generated is then pn.m. this rotor. so that to increase the e.f. = length of the coil.f. induced is a magnetic BINv where B= I field strength.s induced in them and they may therefore be connected together.33. Diagrammatic View op Alternator produce a sinusoidal alternating e. and is therefore fixed.33. Increased field strength is obtained by filling the space between the polepieces with iron. each occupying only half the rotor periphery. where p is the number of pairs of poles and n is the rotor speed in r. If a fourpole stator is used. The coil is then housed in a slot in shown in Fig. Both coils will have similar e. to the flux. proportional to the field strength. 4. Hence an arrangement such as Fig.33 does not make the best use of the available space. the length of the coil and the speed of rotation. as through a complete flux reversal in a quarter of a revolution only.m.f. the rotor passes connected to form a continuous coil. 4. The poles on the .m. 4. as shown in Fig. N = number of turns in the coil. The simple arrangement of Fig. The e. we must increase B or I or both.
which maintain the flux. 4. are held by springs.c. and at a low voltage. output from the machine or from a d.s in the armature windings. The d. will rise and substantially perfect regulation can be obtained.f. the ratio of the voltage drop at full load to the actual (full load) voltage is called the regulation and this may vary from 10 to 20 per cent for small machines down to a few per cent only with a large machine. Diagram of SalientPole Alternator a. Fig. generator mounted on the same shaft. be less . Regulation When an will alternator is supplying current the terminal voltage than the generated e. usually of carbon. because of the voltage drop on the internal impedance.f.m. in which case the generated e. which the are housed in slots in the stator. in such a way as to provide alternate N and S poles at the periphery. so that the sliprings and brush gear are simpler. The rotor then carries windings which are energized with d.m. energized by field windings carrying d. for the field windings has to be supplied either by rectifying part of Field Coifs Stator Armature or Rotor Path of Magnetic Flux indicated by Dotted Lines. In larger machines it is more convenient to reverse the arrangement and use a fixed armature with a rotating field. and the rotation of this field induces e.34. This is done because the field current is only relatively small.206 stator RADIO COMMUNICATION may be permanent magnets but are more usually electromagnets. As with a transformer. Armature Coils embedded in Slots m the Rotor.c.f.c.c. The connections to the rotor are brought out to slip rings against which brushes.c. In some cases the field current is arranged to increase slightly as the main alternator load increases.m.
The is similarly subjected to varying flux though only thus also laminated. where n is the speed. Fig. A d. but we can say that the e. so the armature is built up of insulated stampings to reduce eddy. Generators is required to be unidirectional instead of alternating necessary to include some form of rotary switch. There will thus be iron losses set up as with a transformer.current losses. and k is a factor determined by the number of poles on the stator and the particular form of armature winding adopted.36.C. 4. generator is arranged in this way. which thus takes the form shown in Fig. but to obtain a more regular output.35 the connections to the coil will be reversed every half cycle so that the output will be in the form of a series of half sine waves. As with a transformer the optimum arrangement is usually that for which the iron and copper losses are equal. The main loss is in the armature since this carries the full load current.35. z is the number of armature conductors.m. generated is knz<&. In practice the coils are not separate but are joined together to form a continuous winding. The commutator will not be a simple twosection affair but will consist of a series of insulated segments it is If the output which may number several hundred in a large machine. The split slipring is called a commutator. Output with Commutator Multisection Commutator D.36. 4. but there is also some loss in the field coils which generates additional heat.c. Simple 4. though the thickness of the stampings can be somewhat greater. i i i Fig. field system is partially and i . Copper losses appear in the windings. each connected in turn to the output. while the material itself is silicon steel to minimize hysteresis loss. a series of coils is arranged round the periphery of the rotor. There are various ways of doing this which need not be discussed here. all in the same direction.f. i i v i i i * . $ is the flux. 4. If the coil in a simple alternator is connected to a split slipring as shown in Fig.PRACTICAL COMPONENTS 207 Alternator Losses The flux in an alternator is continually rotating and hence at any point the iron circuit will be subject to a varying flux. .
37 (a). the generator may be made to have any desired characteristic. 4.33 in which.38 (a) Fig. t The z^ rTfflSV^. Fig. one in shunt and the other in series. be modified on the lines already discussed . being connected in series as in Fig.C.f. is Phase Phase Phase ABC (6) (a) Fig. the iron circuit of a d. of course.m..m. In the first illustrated in Fig. Each of these will have induced in it a sinusoidal e. 4. generator can be energized by the machine There are two possible arrangements. 4.37.37 (6). When the machine supplies current its voltage will fall because of the internal voltage drop. Generators load current. 4. can be designed to be energized by the itself.c.s will be spaced electrically by 120°.38.37 (c) shows a compound winding in which there are two sets of field windings.c. causing the voltage on load to rise. By suitably proportioning these two. the field is connected across the output. in which case the flux will increase as the load current increases. 4. 4. The field coils.— 208 RADIO COMMUNICATION field coils in a d. there are three separate coils symmetrically spaced. As with an alternator. Field Coil Connections »or D. as shown in Fig. This in turn will reduce the field current and the voltage variation will be accentuated. Diagram or Simple Threephase Alternator Threephase Generators shows a modification of the simple alternator of a single rotating coil. 4. instead of Fig. j •* 1 i load \~\ I Co) Load \\ W Looa \~\ 1 «0 Fig. however. but because of their angular displacement the three e.f. Such an arrangement is called a threephase alternator.38 (6). 4. Its practical form would. generator laminated to minimize losses due to the varying flux.
Modern electricity supply systems are all threephase. but the provision of three separate phases in this manner has a number of advantages. while it simplifies the construction of electric motors and other devices. 4. is \/3 times the phase voltage. usually connected to earth. may be connected in two ways.39 (a) one end of each phase is connected to a common neutral point. This gives the resultant OB' at an angle of 30° to AB' 2 OA cos 30° OA. The normal domestic supply is single phase and is taken from one of the phases and neutral. which. (a) Fig. Domestic apparatus is thus designed to operate from 240 volts .PRACTICAL COMPONENTS 209 for singlephase alternators. 4. The line current. It permits better utilization of the material not only in the generator itself in the transformers will and cables of the but also distribution system. and yellow phase respectively. To subtract OB from OA. blue. In the star connection shown in Fig. steppedup in transformers to 132 or 275 kV and distributed on the overhead "grid" lines to suitable points where it is stepped down to a convenient voltage for industrial or domestic use. both of which are used according to convenience. The high voltage is used for the transmission because the current is correspondingly reduced and hence the losses in the transmission line are minimized. as be seen. The voltage between the phases is then the vector difference between the individual phase voltages. add a vector OB. 4. is clearly the same as the phase current. =— = = however. the voltage being transformed down to a value which is considered safe. The standard in Great Britain is 240 volts which means a voltage between phases of 240 \/3 = 415 volts. The current is generated at a voltage of 33 or 66 kV. (b) Stab and Delta Connections Threephase Relationships The three phases. whence OB' WZ)OA. customarily referred to as the red. as is shown in Fig.40.39.
38 and the vector diagram of Fig. which is equal and opposite to OA. The local distribution systems are arranged so that. 4. if required. Vector Relationships in Threephase System Transformers may be designed for either form of connection and may have a deltaconnected primary with starconnected secondary. is supplying a normal threephase load the sum of the voltages (or currents) in the three phases is zero. 4. so there is always some unbalance current in the neutral. with a phase .40. The line and phase voltages are the same. The sum of OB and OG is OD. 4. B > Fio.39 (b).voltage of 200. .40. while industrial a 415volt threephase supply. being used to power into mechanical rotation. In practice the need to supply singlephase systems for domestic use introduces a lack of balance. but the line current is the v/3 by two adjoining phases and is thus sum of the currents taken times the phase current. 4. It is only when the three phases are combined at the local distribution point that the cancellation of currents in the neutral occurs. This will be clear from Fig. voltage In America and certain continental countries the domestic is only 115.connected system carries no current. Here one end of each phase is joined to the beginning of the next. The alternative delta connection is illustrated in Fig. If we take a An electric translate electrical * Note that the neutral wire at the actual point of supply has to carry the same current as the line wire. When a threephase system Electric Motors motor is the converse of a generator. but this cannot be performed exactly.* or vice versa. This connection is mainly used for motors and industrial devices. no part of the system being earthed. the loads on the three phases are equal so that the system remains balanced. as far as possible.210 RADIO COMMUNICATION motors and similar devices run from single phase. This means that the neutral wire of a balanced star.
and this is one of its principal advantages. Hence the speed falls off as the load increases. having less turns but capable of carrying the full load current. voltage across the terminals a current •will flow in the field coils which will establish a magnetic flux and the current which flows in the armature will produce a mechanical force equal to BINl.c. Hence the speed is inversely proportional to the number of armature conductors and the flux. The field system of a d.c. and this is useful in a motor which is required to develop a high torque at low speeds. is quite small. as explained in Section 1 .C. is proportional to the speed. however.4. and a back e. Since the back e. so that the maximum current is always flowing in that portion of the armature which is in the maximum field and continuous rotation will result. the speed will automatically adjust itself such that the current flowing is sufficient to develop the mechanical power required. the difference between the applied voltage and the back e. which will cause the armature to rotate. Motors Actually. operating into the (effective) armature resistance.c. As it does so the connections will be changed by the commutator. It was shown when discussing d.m. but the torque developed is proportional to the square of the current since the flux has also increased. motor may be connected across the input as with a shuntfield generator. generator and apply a d. in which case the speed for a given setting of field current is constant except for the small drop in speed necessary to allow the armature to draw sufficient current to supply the torque required. will be induced in the armature proportional to the speed. when the motor is running. As soon as this rotation has been set up the machine will behave like a generator.PRACTICAL COMPONENTS 211 simple d.m. should never be run with no load because the field then is so weak that the speed may become dangerously high. Alternatively the field winding may be connected in series with the armature.m. V being the applied voltage. motor by varying the strength of the field.f. generators that E = knz<l> so that we can say that the speed n = iJ/fcO = Vfhz^ very nearly. It is therefore a simple matter to control the speed of a d.f. . Hence the armature current will be determined by the difference between the applied voltage and the back e.c. Speed of D. A series motor.f. With this arrangement the flux is proportional to the armature current and is therefore small when the motor is running light and increases as the load (and hence the current) increases. such as a traction motor.f.c..m.
there is at any instant a torque produced by the interaction of the armature current and the flux. forming a shortcircuited secondary in which currents will be induced by the flux. The rotor is provided with a number of conductors which are all joined to a common end ring. Motors There is a wide variety of a. In all but the smallest motors therefore it is necessary to introduce a series resistance when first switching on. the resulting torque is always in the same direction and is the mean of the instantaneous torques throughout the cycle. This simple arrangement is not efficient and is usually only employed for small motors designed to operate on either d. and hence if the full voltage is applied the current will be limited only by the armature resistance. motor there is initially no back e. The speed of the rotor is virtually that of the rotating flux though there must be a slight difference in order that the rotor conductors shall cut the flux. It is possible to operate a d. which is normally quite low. Because of the phase difference between the currents each of the field windings will receive maximum current in turn and the flux produced will thus move round the stator. The interaction between these currents and the flux will produce a torque causing the rotor to rotate.c. sign together motor is the induction motor which operated from a threephase supply. starter A. Such a device is called a motor and usually takes the form of an arm which is moved slowly over a series of studs which connect gradually decreasing resistance in the circuit. or a. Consider three field windings spaced equally round a stator and supplied with threephase current.c. so that the procedure has to be repeated each time the motor is started.c. motors and reference can only be made to the more important.C.m.f. For larger machines certain modifications are required which are beyond the present scope. commutator motor on an a.c. for although the current is alternating.212 RADIO COMMUNICATION Motor Starters When starting a d. supply. supply. The greater this difference. Induction Motors The more usual form of for industrial purposes is .c. and since the flux itself is rotating it will drag the rotor round with it. and. since both of these change when the current reverses. This type of motor is often called a squirrelcage motor because of the cagelike structure of the rotor conductors.c.c. At full travel the arm is held over by an electromagnet which releases it if the circuit is broken for any reason. the greater a. This is gradually cut out in steps until the motor is up to speed.
The necessary rotating flux can be produced by using two stator windings spaced 90° out of phase. The threephase stator may also be circuit is shown in Fig. It is not Motor stator is wound (a) (fi> Fig. as shown in Fig. 4. but between 60° and 75° is practicable and this is sufficient to produce a rotating field. Thus with a 2pole motor and a 50 Hz supply the speed is 3. Neither arrangement is as efficient as true 3 phase operation and splitphase working is normally only used in small machines.m. 4. . however. Splitphase If the for a twophase supply the current in the second phase may be fed through a capacitor which will produce a leading current and thus provide a phase difference.m. Splitphase Field Connections possible to obtain a full 90° difference. This is the principle of the twophase motor which is often used in control systems. less the slight drop necessary to supply the torque.p. This small difference in speed is called the slip.000 r. the slip begins to increase disproportionately and the motor stops. 4.. A connected for splitphase operation.41 (b). The speed is actually 60f/p r. Singlephase Induction Motors A singlephase supply will not produce a rotating field but it is possible to run an induction motor off a singlephase supply by two methods. where / is the frequency of the supply and p is the number of pairs of poles. so that the speed automatically adjusts itself to a value such that the current is sufficient to provide the torque required.p. and at full load is of the order of 5 per cent.41 (a). As the load increases.PRACTICAL COMPONENTS 213 the induced current. An induction motor is thus a substantially constantspeed machine with relatively small overload capacity (and a poor starting torque).41.
This causes a distortion of the flux path and the rotor takes up a position such as to reduce this distortion to a minimum. The poles thus develop maximum flux every half cycle. which occurs when the rotor revolves at synchronous speed. This is provided with a solid steel rotor in which magnetism is induced by the stator flux. and by arranging a similar (but not identical) series of poles on the rotor it will run at a speed which brings the poles into alignment at the appropriate instant. A type of synchronous motor used in small clock or timing mechanisms employs a stator provided with a series of poles and energized by a singlephase a. The arrangement is not selfstarting. The stator may be two. Fig. it being necessary to run the rotor up to speed by some external means when it will pull into synchronism. Here a portion of the stator pole pieces is enclosed by a shortcircuited winding. 4. which produce a secondary flux which lags behind the main flux by 90°.c.42. This is sometimes useful. *"~~ ~]~> In practice the shading poles cannot be located the full 90° away from the main w^~Main ES *" a jjSjT^ winding p ie8) DUt a sufficient displacement can be obtained for satisfactory operation. Currents are induced in these coils. the rotor will and will run at a truly constant speed determined only by the frequency of the current supplying the stator flux. is If the rotor of an induction motor align itself with the rotating stator flux N Stepping Motors which The toothedrotor construction can be utilized to provide a device will rotate in discrete steps on the application of suitable . though the torque developed is less than with an induction motor. Shaded Pole Synchronous Motors magnetized.or threephase.42. supply.214 RADIO COMMUNICATION Shadedpole Motor An alternative arrangement is shown in Fig. However. Another form of small synchronous machine is the hysteresis motor. if therotor is provided with an additional squirrel cage or similar winding it will run up to speed as an induction motor and will then pull into synchronism provided the slip is small. Shading coil Again the system suffers in efficiency and is only used for small motors. If the rotor is magnetized with alternate and S poles the arrangement can be made of selfstarting. 4. By disposing these "shading" poles between the main poles the effect of a twophase winding is obtained and a rotating field results.
43 illustrates the essentials of a movingcoil meter. To obtain a linear relationship the magnetic flux must be constant and always at right angles to the coil.43. 4. This is achieved by including a cylinder of soft iron inside the coil. . Such instruments are of two main types. It Fig. Essentials of Movingcoil Meter consists of a permanent magnet with a If current is cylindrical air gap in which passed through the coil a torque is produced. Typical torques range from 0004 to 004 newtonmetres according to size.PRACTICAL COMPONENTS 215 (squaretopped) pulses to the stator. This is used for electrical repeating in control or telemetry systems. as in an electric motor. as shown. The movement is restrained by a spiral spring which develops an increasing torque as the coil rotates. The stepping angle is usually 7£° or 15° with a maximum pulse rate of a few hundred per second. Hence the coil takes up a position proportional to the current through it. Spiral hair springs serve to provide the a coil is pivoted. forces produced by the interaction of magnetic or electric fields are utilized to provide indications of the strength of voltages and currents. Meters The mechanical movingcoil and movingiron. leaving only a small annular gap in which the coil rotates. 4. The coil is of many turns of fine wire wound on a light former and held in jewelled pivots. Coil Spiral spring Fio. and if a light pointer is attached to the coil it will indicate the current against a suitable scale.
though instruments are available requiring and The system is only 60 /j. The movement produced by such an arrangement. the sensitivity is appreciably less. is not linear and improved performance. however. and the power consumption is greater.d. a current of 5 to 10 being required for fullscale deflection. 4.216 restraining torque coil. Eig.A for f.44 illustrates a movingiron type of meter. 4. No permanent magnet ment is is required. so that a movingiron instru On the other hand. For measuring larger currents the meter is shunted with a resistor which bypasses the greater part of the current. is obtained by introducing a second fixed armature. This requirement is discussed more fully in Chapter 15. both in sensitivity and linearity. except over the first 10° or so. mA . Essentials of Repulsiontype Movingikon Meter armature which will endeavour to align itself with the flux. This may be of importance.44. while for voltage measurement a suitable series resistance is included such that the voltage to be measured passes the appropriate current through the meter. Different current ranges are obtained in a movingiron meter by suitable choice of the number of turns in the solenoid so that the use of a shunt is not necessary.s. For voltage measurements a series resistor is used as with the movingcoil type. Since both are similarly magnetized by the current in the coil there is a repulsive force between them and this can be arranged to give a movement of the pointer nearly proportional to the current. In its simplest form this consists of a solenoid in which is pivoted a light softiron Pointer Fixed armature Coil Moving armature Fig. RADIO COMMUNICATION also to lead the current into and out of the usually designed to provide a fullscale deflection with either 1 or 5 mA. lighter and cheaper than the movingcoil type. for it is an axiom in measurement that the measuring instrument must not appreciably influence the quantity being measured.
m.m. meter. plates the electrostatic attraction will cause the rotate and so provide an indication.c. into a pointer movement.s. known as rectifier meters.c. The force is . so that the meter can be used on a.m. This will be 09 times the r. a movingcoil meter will read the mean value. value. Meters 217 The movingiron meter has the advantage that the deflection is independent of the direction of the current. If an e. The reading is proportional to the square of the voltage.c. which is 2/n 0637 times the peak. value and that the indication is only correct on a pure sine wave.m. One method employs a linkage which converts the mechanical expansion of a fine wire due to heating.m. If the current is rectified.c. and d. With a movingcoil meter the deflection is dependent upon the direction of the current and hence such a meter by itself will not give any indication on a. Since the magnetic repulsion is proportional to the square of the current the reading obtained is a true r. but it should always be remembered that they do not read the true r.f. proportional to the heat developed. Such meters are sluggish and inaccurate. This usage is discussed further in Chapter 15. and then apply this to a sensitive d. which can be made to indicate in terms of r.c.c. moving plate to small so that such meters can only be used to measure voltages of 1.s. Such instruments. however.c.PRACTICAL COMPONENTS A.s.s.000 volts or more. reading and the same scale applies for both a. For certain applications use effect of Electrostatic Meters is sometimes used is the a light metal plate carried on a pivot which allows it to rotate between two fixed plates in a manner similar to a variable capacitor. so that it consists of a series of half sinewaves all in the same direction. but this can be allowed for in the calibration.m. = Thermal Meters is sometimes made of the heating a current. is applied across the One other type of instrument which It consists of electrostatic meter. value. as well as d. or d.C. An alternative method is to use a thermocouple which produces an e.f. Assemblies of small copperoxide or other semiconductor rectifiers are made specifically for this purpose and are included inside the meter itself. are in considerable use.c. Hence it mil read a. and with modern materials and constructions is equally accurate on both types of current up to frequencies of a few hundred hertz.
and are used where considerations of portability or convenience predominate. the zinc reacting with the electrolyte to form zinc chloride and releasing ammonia at the negative (zinc) electrode and hydrogen at a positive (carbon) electrode. This gives up some of its oxygen to combine with the hydrogen.m. A Dry Cells To avoid the use of a liquid electrolyte. forming water. are immersed in a solution of ammonium chloride (sal ammoniac). There are other forms of primary cell. the intervening space being filled ammoniac and glycerine. The ammonia dissolves in the water but the hydrogen is not soluble and forms a film round the carbon which rapidly prevents the passage of further current. using different chemicals and depolarizers but the principle remains the same and they need not be discussed here.f. known &» primary and secondary cells respectively. of zinc and carbon respectively. the Leclanche cell can be made up in a "dry" form. the top being sealed with pitch. One of the most common types is the Leclanche cell in which two rods.m. The zinc electrode is steadily consumed as long as current is passing and ultimately both the zinc and the electrolyte have to be replaced. Here the carbon rod with its surrounding depolarizer housed in a zinc can. This is done by surrounding the carbon rod with manganese dioxide. Such cells can be made in small and compact form and can is with a jelly of sal . if the cell is connected to an external circuit.6. leaving a lower (and unstable) oxide of manganese which during the periods when the cell is not in use absorbs oxygen from the atmosphere and reverts to its normal state. current will flow.f. Primary Cells primary cell is one in which the e.218 RADIO COMMUNICATION 4. This arrangement exhibits a potential difference of approximately 14 volts between the carbon and the zinc (the carbon being positive) and. The manganese dioxide depolarizer is usually mixed with graphite to improve the conductivity and is held round the carbon rod in a porous pot or bag. The effectiveness of the action depends on the amount of depolarizing agent available. but a typical "wet" cell will give currents of the order of one ampere. An electrolytic action then takes place. They are of two main types. Batteries Batteries are sources of electrochemical e. is produced by the interaction of suitable chemicals which are themselves expended in the process. This action is known as polarization and to make effective use of the cell the hydrogen must be removed.
A modified form of standard cell known as the Clark cell uses a zinc amalgam instead of 143 V.m.f. Techniques have now been devised whereby this action can be reversed. The e.f. This uses a zincmercury combination which can be made up as a dry cell having certain advantages.f. is about 09 V. notably absence of deterioration when not in use. in potentiometer circuits (See Chapter 15). A zinc anode is employed with a selfdepolarizing cathode of mercuric oxide mixed with graphite.m. It is used as a standard of e. It delivers a very constant e.f. A group of electrodes in a large cell can supply continuous currents of the order of tens of amperes. 219 provide any required e. in the process. It delivers (b) Mallory cell.f. Fuel Cells A current passed through an aqueous solution will cause the water to dissociate into its constituent elements by the electrolytic action described on page 55. cadmium for the cathode.m. from 1 to hundred The internal resistance of such a construction is necessarily higher than with the wet type.PRACTICAL COMPONENTS be assembled in several series to volts. and becomes progressively higher in use so that the life of the dry battery is determined more by this rise in internal resistance than the dissolving of the zinc.m. Such a device is called a fuel cell and is illustrated in simple form in Fig. The process requires an interchange of electrons and hence only takes place when there is an external circuit. The e. . so that hydrogen and oxygen can be induced to combine electrolytically to form water.f. 4. Other Forms of Primary Cell There are various other types of cell of which only two are of importance in radio engineering. Two electrodes of porous nickel are immersed in a weak solution of potassium hydroxide Oxygen supplied to electrode A combines with water to form hydroxy! ions (OH). At B these recombine with hydrogen to form water again.45. . on load is between 13 and 135 V. and develop an e.m.m. This uses a cadmium cathode and a mercury anode in a solution of cadmium sulphate with mercurous sulphate depolarizer. but if this is provided appreciable current can flow. of 10186 V provided that it is only required to supply negligible current. There is also some chemical action taking place even when the battery is not in use so that it does not have an unlimited shelf life. These are (a) Weston standard cell. The electrolyte is potassium hydroxide.
which combines with the hydrogen in the sulphuric acid to form water. becoming as high as 25 to 26 volts when fully charged. Since the chemical action during discharge releases water it will the positive plate is Thus it is possible to put * Actually both plates become transformed into lead sulphate. so that if current is passed in the opposite direction the original conditions can be restored. Here the plates are of lead and lead peroxide in a solution of dilute sulphuric acid. back into the cell the energy which has been extracted from it. and the arrangement is therefore called an accumulator. though this falls to 22 volts when the charging e. The e. of the cell itself increases as it becomes charged. Section through a Fuel Cell Secondary Cells There are certain chemical combinations in which the changes due to the passage of current are reversible.m.* If current is passed through the cell in the opposite direction. but the process is complex and need not be discussed here. greater than that of the cell itself so that current is passed in the reverse direction.45. If current is drawn from the cell the positive plate gives up its oxygen. the most common type being the leadacid cell. but the principle is basically as described.220 RADIO COMMUNICATION The electrodes have to incorporate certain catalysts to enable the action to take place. while the sulphate ions combine with the lead to form lead sulphate. + Overflow Oxygen is :koh\ Porous nickel Hydrogen Fig.m. 4.m. the process is reversed and reconverted to lead peroxide.f. however.m. is removed.f. Such a device is called a secondary cell.f. The reaction produces an excess of water so that the strength of the solution has to be periodically increased. On discharge this drops rapidly to 2 volts and then falls slowly as the cell discharges to about 19 volts. The e. . available is initially about 22 volts per cell.f. at which point the discharge should be stopped and the cell recharged. To recharge a cell it is necessary to apply an e.
The e. In the charged condition the positive plate becomes oxidized. Silverzinc cells show considerable economy of weight. The e. the specific gravity falling by about 10 per cent during discharge. as the cell discharges. The initial value depends on the type of cell. A later type of alkaline cell uses electrodes of silver and zinc. The current permissible depends on the size and construction.f. An 80 Ah car battery will supply 20 to 30 amps for short periods while a large stationary battery may be required to supply several hundred amps. This can be used to determine the state of the battery. with either type is approximately 12 volts and as the chemical changes due to the passage of current are reversible. the oxygen is transferred to the zinc. the spaces of which are filled with a paste of lead oxide. Secondary cells are rated in terms of their amperehour capacity. Thus a cell of 80 Ah capacity will supply 8 amps for 10 hours.PRACTICAL COMPONENTS 221 be seen that the specific gravity of the acid decreases in the process.m. being of the order of 121 with a large stationary cell to 125 with a car battery. before requiring to be recharged. which does not enter into the reaction. or 4 amps for 20 hours. whereas for silverzinc cells the figure is about 80 Ah/kg. The plates are then "formed" by passing current through the cell which reduces the paste on the negative plate to spongy lead and forms lead peroxide on the positive plate. It is also customary to include porous separators of insulating material between the plates to resist the mechanical deformation which might otherwise occur due to the force between the plates when passing a heavy current.f. The plates of an accumulator are made in the form of a grid of lead alloy. If the current is only taken intermittently the charge will last rather longer. while. though the smaller e.f. on load is approximately 15 V.m. Leadacid types develop approximately 20 amperehours per kg. Alkaline Cells An alternative form of secondary cell uses plates of nickel hydroxide and iron potash). The advantages of this type of cell are reduced weight and greater robustness. A leadacid cell must not be allowed to remain in a discharged condition.m. necessitates a larger number of cells to obtain a given voltage. for the lead sulphate tends to change its character and will no longer recombine in the proper manner when recharging is attempted. with the lead cell . and pro rata. The electrolyte is potassium hydroxide. in an electrolyte of dilute potassium hydroxide (caustic is In some cases cadmium used instead of iron for the negative plate.
7. If one of the blocks is connected to (or is itself in the form of) a diaphragm any variations in air pressure impinging on the diaphragm will cause a corresponding variation in the resistance.46. A microphone is The Carbon Microphone The simplest type of microphone depends for its operation on the fact that the resistance between two surfaces of carbon depends on the pressure between them. This effect can be utilized to create a BRASS GRANULE " CHAMBER BREATHING HOLE FELT PAD WHITE CELLULOSE' ENAMEL' BAKELITE' WASHER MICA DISC Fig. Microphones a device for turning sound energy into electrical energy. The Mallory cell is also available in rechargeable form. . 4. Diagram or Carbon Microphone (By courtesy of the Post Office) microphone by mounting two carbon blocks in a suitable housing and filling the intervening space with finelydivided carbon granules. a typical construction being illustrated in Fig. the current will vary accordingly.46.222 RAD/O COMMUNICATION Nickel. the ideal microphone being one which reproduces the waveform of the sound exactly. 4.cadmium cells are now available in small sealed units. Microphones may be divided into two main groups: a pressure microphone is one in which the electrical response is caused by variations in pressure of the sound wave. whilst in a velocity microphone the electrical response corresponds to the particle velocity resulting from the propagation of the sound wave through an acoustic medium. This type of microphone is used in telephone handsets. Hence if a current is passed through the microphone from a battery. 4. thus providing the convenience of a dry cell with the facility of rechargeability.
. directly.PRACTICAL COMPONENTS 223 Such a microphone requires a source of low d.f. Brief descriptions of these methods are given below. sensitivity The directly actuated or soundcell type has lower but natter frequency characteristics and is almost nondiaphragmoperated type. One method 70. is relatively insensitive. Benton (Pitman). voltage to pass current through the system. though having higher sensitivity. Hence appreciable amplification first is required. The arrangement. giving an output of a few millivolts only. and because of the high impedance the amplifier.f. Bochelle salt is the most commonly used material owing to its high piezoelectric sensitivity. Crystal recordplayer pickups use a Bochelle salt crystal which is deformed by the motion of the needle in the record grooves (see page 510) and this produces a varying e. is to use the piezoelectric effect described on page whereby mechanical pressure can be converted into Alternatively the air vibrations can be caused to produce variations in capacitance or inductance. N. has a less uniform response and is more directional directional. is developed by the variations in current due to the changing capacitance.f. but the sensitivity is good. however. and suffers considerable nonlinear distortion for other than small displacements.f. generated by a piezoelectric crystal when it is deformed. A potential of several hundred volts is applied across the plates through a high resistance.m. Capacitor Microphones A capacitor microphone operates by variations in capacitance between a thin stretched diaphragm and a rigid plate parallel to it. Such a system must be heavily damped to avoid serious resonance but can then be made to have a response which is nearly flat from 30Hztol0kHz.c. Crystal Microphones A crystal microphone depends for its operation on the e. The noise level is high and the frequency response is seriously peaked. Where greater fidelity is required a more sophisticated system is required. whilst the at high frequency.m.m.m. mounted stage is made up as a preas close as possible to the microphone itself. For more detailed information of the reader may refer to Telecommunication Principles by R. e. The crystal microphone may be actuated directly or through a diaphragm. across which e.
This resistance may take the form of a finite pipe damped with tufts of felt. Both these limitations. in essence. fitted with a horn or trumpet to concentrate the sound.000 Hz. whilst directradiation loudspeakers are those in which the cone or diaphragm is directly coupled to the air. the pole pieces serve as two sides of the ribbon. struction the response can be extended to 6010. Loudspeakers . Horn loudspeakers are those in which the diaphragm is coupled to the air by means of a horn. sound wave the ribbon is driven from its equilibrium position and the motion of the ribbon in the magnetic field induces a voltage between the two ends of the ribbon which is proportional to the particle velocity in the sound wave. a small version of the movingcoil speaker described on page 226. This simple arrangement is no longer used because of its limited range of frequency response.39). while the horn introduces undesirable resonances. In addition to supplying the magnetic flux. The freeribbon microphone is strongly bidirectional but reproduces transients faithfully and is accepted as a baffle for acoustically separating the Under the influence of a the best type for high fidelity. The earliest forms of loudspeaker were merely enlarged versions of the simple telephone earpiece described earlier (Fig. Pressure Ribbon Microphones The pressure ribbon microphone consists of a light metallic ribbon suspended in a magnetic field. can be overcome with a properly designed horn and drive. freely accessible to the atmosphere on one side and terminated in an acoustic resistance on the other. 1. Velocity Ribbon Microphones The freeribbon microphone consists essentially of a loosely stretched ribbon suspended in the air gap between two pole pieces.000 Hz. as explained in the next paragraph. The resonance of a freeribbon microphone is usually below the audible limit and the frequency response can be very flat over the range 3015. by more elaborate conflat over the operating band. The frequency response in the simplest construction is somewhat limited but is substantially However.224 RADIO COMMUNICATION Movingcoil Microphones The movingcoil microphone is. 4.8 Loudspeakers may be divided into two principal groups. Very high quality microphones may be made in this way and nondirectional microphones with a flat response up to 15 kHz have been produced. however.
producing a quite unacceptable frequencydoubling action. in general. it is necessary to design the horn to follow an exponential or logarithmic law. it is possible with an exponential horn to obtain satisfactory radiation at frequencies as low as 20 Hz. Hence. so that the area at any point equals . and the currents in the coils cause variations in this pull. Such a horn will give practically uniform radiation down to a In order that the vibrations of the diaphragm into sound energy. The pole pieces in a telephone mounted on a permanent magnet so that there is a constant pull on the diaphragm. The ordinary conical horn is not found satisfactory as phase differences are set up throughout its length which cause vibrations in certain parts of the column of air to interfere with those in other parts and restrict the effective radiation of sound. where A is the initial area and x is the distance along the horn. If this is to be avoided. In other words. may Design of Horn may be converted necessary to set in motion a column of air. be 10 s times b) but the pull is now dependent on b and not b 2 If this were not so there would be a pull on the diaphragm every half cycle. the larger and longer the horn. Hence with all loudspeakers a permanent field is provided and to obtain the best sensitivity this is made as high as practicable.4 e te . One way of doing this is to mount a horn over the orifice of the loudspeaker and this acts as an acoustic transformer. . The design of the horn must be carried out carefully. converting the smallamplitude highpressure air pulsations at the diaphragm itself into lowpressure highvolume pulsations at the far end. This provides a marked increase in sensitivity because the pull on the diaphragm depends on the square of the magnetic flux density. and. the total variation in the pull on the diaphragm receiver are is proportional to (B + b) — (B — by = 4Bb 2 Hence not only do we obtain greatly increased sensitivity (for B . if the permanent magnet contributes a flux density B and the currents in the coils cause this to vary by a small amount b.PRACTICAL COMPONENTS 225 One point may be noted here. This cutoff depends entirely upon the rate of expansion of the horn. the lower the cutoff frequency. it is point where the horn suddenly ceases to load the diaphragm. the air currents fall out of phase very rapidly and a "cutoff" is obtained. Whereas a loudspeaker with an ordinary horn would give a very poor response below 200 Hz.
This led to the developin ment of the movingcoil speaker. illustrated in Fig. However to be fully effective at low frequencies a considerable movement of the diaphragm is required.226 RADIO COMMUNICATION Movingcoil Speakers Another way of achieving the desired results.47. The former is technically preferable owing to the fact that higher field strength can be obtained. or a somewhat similar arrangement may be constructed with a permanent magnet. namely the setting motion of an appreciable column of air.47. is to increase the size of the diaphragm. Here. this led to the development of cone speakers in which a simple reed mechanism similar to that of a telephone earpiece is coupled to a cone of paper. and most modern speakers use permanentmagnet fields. buckram. or some similar material. or more. in the Barrit Diaphragm Fltxibl* surround Fig. will set up a force tending to move the coil out of the gap. Any current flowing through the coil. 4. . 4. the direction of motion being dependent upon the direction of the current. This magnetic field may be produced by passing a current round the core of the field system as shown. Hence. having a diameter of 20 cm. but great progress has been made with permanent magnet systems. It will be clear from the figure that the magnetic field flows across the gap and therefore cuts the turns of the coil at right angles. Diagram op Movingcoil Speakeb centre of the diaphragm is mounted a small coil which is capable of lateral movement in and out of an air gap in a powerful magnetic field. therefore.
to the back. Now. consisting of a stout piece of wood. the wavelength of the sound is very much greater than the diameter of the diaphragm. a partial vacuum is set up behind the diaphragm. The air vibrations are set up. but does assist to a large extent in producing the proper radiation.47. at the same time. This was found to be an interference between the air waves produced by the front and the back of the diaphragm. cases the baffle is not a plain piece of wood but is made form of a box. but they are not radiated into the room. it was found at first that the radiation of the low frequencies was disappointingly small. in order to radiate a uniform amount of sound energy. Use of Baffle At these very low frequencies. All that happens. it can be shown that. Otherwise. therefore. In many in the . and therefore air waves produced by the front and back of the diaphragm are sensibly in phase. but. it is capable of moving a considerable distance. The back of the box. however. Since the diaphragm is free to move and has very little restoring force. and consequently the sound vibrations are radiated into the room. This does not have quite the same effect. or other suitable material. Clearly. and this will be transmitted to the diaphragm. and it was not for some time that the real cause was discovered. and usually consists of a large square piece of wood with a hole in the centre through which the diaphragm operates. a baffle is used. if we move a diaphragm a short distance a large number of times per second we produce the same agitation of the air as in moving the diaphragm a large distance a relatively small number of times per second. If the diaphragm moves forward it pushes air away from it. arranged as a continuation of the diaphragm. resonance will be set up at certain frequencies and the reproduction will sound unnaturally lowpitched. must either be open or provided with large slots covered with light gauze so that air may have free passage in and out of the box. To obviate this difficulty. This is the principle on which the movingcoil speaker works. The sound waves coming off the front of the diaphragm are now unable to interfere with those coming off the back. however. 4.PRACTICAL COMPONENTS 227 an alternating current will cause a continuous oscillation of the coil. where there is a partial vacuum. the vibration of the diaphragm at low frequencies being of considerable amplitude. where it has been displaced. setting up the desired radiation of sound waves. the motion of diaphragm should be inversely proportional to the frequency. Even so. This is illustrated in Fig. is that air rushes round the edge of the diaphragm from the front.
it will be seen that there is an appreciable difference in the length of the paths to the point X. however. but breaks up into a number of subsidiary portions. If the diaphragm of the loudspeaker is comparable with the wavelength of the sound being reproduced. then it is clear that these two sound waves will arrive out of phase and will interfere with one another. each of which vibrates in its own way. and frequencies above this. . Illustrating Phase Difference with Large Diaphragm vibrate as one unit. Even so. off the axis of the dia4. The radiation will reach this point from various points of the diaphragm. and music being flat. and if we consider the two extremes A and B.228 RADIO COMMUNICATION Phase Difference Difficulty still arises with the large diaphragm type of loudspeaker in the radiation of the upper frequencies. speech lacking crispness. therefore. and the design of the diaphragm is a matter of great difficulty. The modern tendency is to use diaphragms of comparatively soft material which are unable to break up into "paper" resonances at the upper frequencies. being transmitted principally along the axis. are not radiated uniformly. as shown in Fig. the diaphragm is not usually a flat disc but is made conical in shape. but are strongly directional. The rest of the diaphragm is thus largely ineffective.500 to 2. and the result of these two factors is a cutoff in the upper frequencies. Let us suppose that we have a simple oscillating diaphragm AB. Such a system is not a good reproducer of the upper frequencies and must therefore be fed with an increased supply of energy in the upper registers in order to maintain a balance of tone. This arrangement. at the upper frequencies the diaphragm cannot X \ '• » Direction v of displacement Fig. is preferable to a diaphragm which breaks up into resonances and accentuates certain upper frequencies unduly. 4.000 Hz. Consider the radiation at a point phragm.48. however. This condition of affairs will be obtained at about 1.48. so that it is capable of moving as a whole and not flexing too much. This also tends to give the diaphragm greater rigidity. To minimize this.
and a good transmission the diaphragm. The art has become highly specialized because at the higher frequencies reflections from the walls of the room have a marked influence.49. and shown shaded in the figure. An interesting point. and an exponential horn is fitted over the diaphragm in order to transform the energy to lowpressure high. however. and if the horn is correctly designed to obey an exponential law.coil system is similar to that already shown. If the distances from any part of the diaphragm to the throat of the horn are within a quarter wavelength at the (highest) frequency concerned. but instead of being wl&g ^^ iw?w Phase Connector ^Exponential Metal Diaphragm Morn Fig. This is a small conicalshaped obstruction. the function of which is to overcome the difficulty already mentioned of unequal radiation from sound middle and edge of the diaphragm would be out of phase. no serious cancellation occurs. specifically designed to handle the upper registers. these sound waves are transmitted down the horn by successive reflections without serious loss. and would therefore not be radiated. the radiations from the centre of the diaphragm are forced to take a longer route. .49. however. the waves coming off the is obtainable. 4. and careful siting of the speakers is necessary. is the phase corrector placed in front of the diaphragm. It will be seen that the moving. This gives a considerable increase in response to the upper frequencies. Tweeters Where really A Public Address Horn Speakers Public address speakers utilize a movingcoil drive with a horn radiator as shown in Fig. If this phase corrector were not there.volume fluctuations as already explained. conventional movingcoil unit is employed for the low and middle registers with small additional units. Mechanism of Coiidriven Horn Speaker attached to a large diaphragm a comparatiuely small diaphragm of about 5 cm in diameter is used.PRACTICAL COMPONENTS 229 highfidelity reproduction is required two or more speakers are used. known as "tweeters". 4. With the corrector in place.
300 52. successful entirely The The thermionic valve. Thoriated tungsten 45 34 26 18 13) 10/ Caesium Strontium oxide .000 (approx. consisting of spasmodic jumps from one energy level to another.1 Electron Emission When a metal is heated the electrons in the atoms take up fresh orbits at greater distances. and although this is being replaced to an increasing extent by solidstate devices there are many applications in which valve techniques continue to be used.000 15. 162 0001 50. while others are so violently agitated that they energy. escape Table Material 5. 5. or vacuum tube. corresponding to a condition of greater The process is neither uniform nor continuous. leave the confines of the atom from the material altogether. Barium oxide * . and even.) Thermionic Valves development of the art of radiocommunication was due to the evolution of the thermionic valve. is essentially a device in which current is produced by the emission of electrons from a cathode under conditions of complete or partial vacuum.1 Typical Emission Constants Work Function {electronvolts)* A 30 60 60 3 6 Carbon Tungsten 40 . Thorium . An electronvolt (eV) is the energy required to raise the potential of an electron by one volt. if conditions permit.500 21.400 38. 230 .900 30. Some of the electrons fall back into more normal orbits emitting light or heat waves in the process. this current then being varied by voltages applied to suitable control electrodes.
* Since the action only takes place if the anode is at a positive potential the device acts as a nonreturn valve and it was from this property that the generic name "valve" became used in British parlance.. As we shall see. will obviously be largely determined by the work function of the material and also by the temperature. it is possible to introduce other electrodes into the valve structure.1. and the simple diode is only used for certain limited applications. But they are repelled by further emitted electrons. which they in turn repel. known some typical values being given in Table 5.m. The amount of the electron emission. are subjected to a positive electric field by introducing another electrode connected to a source of (positive) e.1. The net result will be that a current will flow across the space from cathode to anode and return to the cathode via the external circuit. and A and b are constants for the particular material.f. Space Charge The emitted electrons shoot out from the heated cathode and then fall back again. varies with different materials. proportional to the work function. The distinction is only of importance when considering the actual design of a valve. The result is a cloud of electrons surrounding the cathode to form what is called a space charge and a condition of equilibrium is obtained. which exercise control functions of various kinds. = AT*e bIT 6 being where T is the absolute temperature (= °C + 273). Dushman which says that the emission in amps/cm 2 of emitting surface is J.THERMIONIC VALVES The energy which an electron has to acquire before It is 231 it can escape as the ivork function. and the Americans have preferred to call such devices vacuum tubes or. however. With the addition of these further electrodes the action becomes far more than that of a simple nonreturn valve. since the higher the temperature the greater the disturbance of the electron orbits. If the electrons. * The effective source of the electrons is thus not the cathode itself but some region within the space charge which is known as the virtual cathode. . electrons will be attracted to this anode which will destroy the equilibrium in the space charge so that further electrons will be emitted from the cathode. as it is called. Values for A and b are given in Table 5. The relationship is customarily expressed in a form first postulated by S.
however. the free gas is first exhausted by very efficient pumps." i. Other forms of "getter" are employed in special circumstances. and its behaviour may be reproduced by a similar valve operating under identical circuit conditions. a small piece of metallic magnesium being attached to the electrodes. where they will deliver up their charge. This expels all the particles of gas that are likely to escape in the normal operation of the valve. A given voltage on the anode will then always produce a definite emission current. therefore. This augmented current. 5. In these circumstances the electrons will occasionally meet an atom of gas and the force of the collision may produce ionization as explained in Section 1.charged ions be attracted to the cathode. reserving the term tubes for cathoderay tubes.e. therefore.6. Cleaning up the residual gas is further assisted by introducing into the valve some active agent known as a getter. every possible particle of gas is exhausted from the bulb. ing on the . that the glass and the metal of the anode (and any other electrodes) hold between their own molecules a certain number of molecules of gas. and these will ultimately come forth and the valve will become soft. Practical Forms of Valve The whole assembly is enclosed in an evacuated bulb. To get rid of this "occluded" gas. Magnesium is one such agent. depend amount of ionization which has taken place. however. When the valve is made hot the magnesium volatilizes and combines chemically with any residual gas which is left. partly to prevent the cathode from oxidation (burning) but also because any gas at normal pressure is an insulator and so would prevent the electrons from travelling from the cathode to the anode. the bulb is warmed and the electrodes are made red hot by electrical means. while the remaining magnesium is deposited over part of the glass of the bulb. however. The air is evacuated. The products of such combustion are drawn off in pumps.232 shortly. Occasionally a small amount of gas at a greatly reduced pressure is introduced.2. usually of glass. the modern valve is made "hard. it is found. tubes. RADIO COMMUNICATION We shall. These positively. the result being that the current flowing across the gap will be increased (since a positive charge flowing in one direction is equivalent will to a negative charge flowing in the other). and moreover the relatively heavy ions are liable to damage the cathode. Hence except for certain types of highcurrent rectifier. adhere to British convention. is obviously variable. In the manufacture of the modern valve.
Contact Potential The third point of interest in the curves of Fig. Such a valve is said to be spacecharge limited.e. (b) At a certain point the anode current ceases to obey this law and reaches a constant value so that further increase in anode volts = '/iiaw mA 5 Contact potential S 6 1 Anodt voltt TO Fig.T.T. battery to apply the positive potential to the anode.1 in which there is an L. Several interesting points will be noted.1. Normally a diode (or any other form of valve) is so designed as to provide far more electron emission than is required. Actually the relationship obeys what is called the threehalves law.e. the current i E%.8.T. i. This is because. battery for heating the filament and an H. the current increasing more rapidly as the voltage is increased.T. since the increase in filament current will produce a higher temperature and hence a greater supply of electrons. there is still 5. The figure also shows the characteristics of the diode. or L. Simple Diode Circuit and (idealized) Characteristics produces no increase in current. It will be noted that the saturation current increases with the filament current as one would expect. i. the manner in which the anode current varies as the H. 5.T.1 is the fact that a small current even when the H.THERMIONIC VALVES 233 The Diode A simple valve containing a cathode and an anode as just described is called a diode and is illustrated diagrammatically in Fig. there is a small . This is called saturation and indicates that all the electrons emitted from the cathode have been drawn to the anode. (a) The relationship between anode voltage and anode current is not linear. battery voltage is zero. in which case the anode current follows the threehalves law throughout the range of working conditions. as explained in Section 1. voltages are varied. 5.
therefore only of importance around the The Triode The usefulness of the thermionic valve. Such a valve is called a triode. without itself taking any appreciable current. If the grid is negative with respect to the cathode it will tend to repel the electrons and prevent them from reaching the anode. of itself. As long as the grid is kept negative it exercises its control with a negligible consumption of power. If the grid potential becomes positive. The simplest such modification consists in the inclusion of a third electrode in the form of a spiral or wiremesh grid between the cathode and the anode. and a sufficient negative grid potential will cut off the anode current completely. The presence of the grid. can be considerably increased by the addition of further electrodes. which shows the variation of the emission as the anode voltage is varied. In order to reduce the current to zero it is necessary to apply a small negative potential to the anode in order to potential as it is called. because of the small area which it presents. as the anode voltage is increased. Negative voltage on the grid .) Characteristics Let us now consider the characteristics of a triode. the anode current rises with increasing rapidity. even though the anode is at a positive potential. however. 2. it begins to draw some small current itself. It will be seen that. of the order of one volt. 5. and while normally it is still only a small fraction of the anode current. This is known as grid current. (Certain secondary effects arise at high frequencies but these will be discussed in later chapters. The two most important are 1. has little influence on the electron stream. offset this contact The contact is potential is quite small. and region of zero current. Fig. it is usually to be avoided because it means that the grid begins to take power. The anodecurrent/gridvoltage curve. with various fixed negative potentials on the grid. which shows the variation of emission with the voltage on the grid.2 shows these characteristics for a typical small valve. The anodecurrent/anodevoltage curve.: 234 RADIO COMMUNICATION naturallyexisting difference of potential between dissimilar elements. at various fixed values of anode potential. however. but if it is connected to a suitable source of potential it can either accelerate or decelerate the electron emission from the cathode.
as the grid voltage is made more negative.g. and it will be seen that variations of grid potential around this bias point produce proportional variations of anode current.t. This is the effective resistance of the valve to varying current and is obtained by observing the change in anode current which is produced by a given small change in anode voltage. 100 200 Volts IS W Anode Grid Volts SAO Fig. The grid current in any case is only a few microamperes since most of the electrons shoot through the spaces in the grid to the anode. 5. voltage has to be applied. . As already explained. It will be observed that.c. individual curves obey an approximate threehalves law. as with a diode. the valve is not normally used with positive voltage on the grid. with a small negative voltage on the grid. e. The first of these is the anode a.THERMIONIC VALVES The 23S delays the process so that more h. the anode current is ultimately reduced to zero.2. Valve Parameters The performance of a valve can be assessed without reference to the full characteristics by the specification of certain parameters. but the point A at which the grid current starts to flow is important for certain uses to which the valve can be put. This is known as grid bias. but the slope gradually decreases as the grid voltage is made more negative. resistance. Characteristics ot Small Triode The valve would be used. 4\ volts with the curves shown. as discussed in the next section. and will be referred to later. The second series of curves shows the variation of anode current with grid voltage.
and this is known as the mutual dia \dvg conductance of the valve. Hence A — .e. there being two principal types. and having. second characteristic figure which we require to know is the relative effect of the grid voltage on the anode current. having a lower work function than tungsten. in the neighbourhood of 1. the change of anode current in mA divided by the change in grid voltage. Cathodes The earliest forms of valve utilized a filament of fine tungsten wire which was raised to a white heat by the passage of current. Under suitable conditions of heat treatment during manufacture this thorium forms a very thin layer on the outside of the filament and. gives the necessary electron emission at a dull yellow heat. a * Since the Siemens is the unit of conductance. Amplification factor = = dva Jdvg = gmra ft With modern valves the a. Modern valves use different cathode materials.c. In the first type a small percentage of thorium is mixed with the tungsten.236 RADIO COMMUNICATION It is actually the slope of the valve characteristic for the particular conditions under consideration. i.000 or more for tetrodes. requiring much less filament current. To determine this we observe the change in anode current produced by a given small change in grid voltage and divide the former by the latter. The amplification factor ranges from 2 or 3 for power valves up to 1. Mutual conductance is usually specified in milliamperes per volt (mA/V)*. Whereas formerly it used to be less than 1 it is now possible to make valves having mutual conductances of over 40. usually written as gm A final parameter which is useful is the amplification factor. resistance ra = dvajdia = — The slope must be measured under conditions corresponding to the working conditions or. . As we are here dividing a current by a voltage the result is not a resistance but a conductance. at some accepted standard which is usually v a 100 and vg 0. and it will be clear that this figure is obtained by multiplying both the previous parameters together. This is the change in anode voltage produced by a given change in grid voltage. so that the a. incidentally. ImA/V = lmS.c.000 resistance may range from something several ohms for a power output valve up to hundred thousand ohms for the tetrodes and pentodes discussed shortly. for purposes of comparison. This figure has improved considerably over the years..
except that the return circuit connections are made to the cathode. because the filament does not tend to crystallize is and is thus not so brittle. be operated direct from alternating current. strontium. Through the centre of this tube runs the heater (Fig. however. The valve behaves in exactly the same way as a batteryheated type. consists of a wire of platinum. and this in its turn gives off electrons. but is a small tube which is coated with electronemitting oxides. which is supplied with current from the mains. The other form of cathode.3). may Tetrodes To obtain a high amplification factor it is necessary for the grid wires to be relatively close together. etc. which is independent of and insulated from the heater. and for this special types of cathode have been developed. This is connected to a positive . or tungsten. fluctuations in the supply. the heater current is obtained from the electric light mains. it possesses appreciable heat inertia so that its temperature is ^^^ 3 Illustrating Construction or Indirectly not affected by the comparatively rapid Heated Valve Such a valve. on which is deposited a coating of what are known as "rare earths" such as oxides of barium. Heater. The most common heater supply is 63 volts with heater current of 03 or 06 amps. 5. or small valves to be run off batteries. nickel. Mre'l relatively large Owing to the mass \ of the cathode. the voltage being transformed down to a suitable small value for the purpose. therefore. Directlyheated cathodes are used for many purposes. which now more common. These have an even lower work function so that it is only necessary to heat the filament to dull red to obtain the required emission. For the majority of uses. In a tetrode a screen grid is interposed between the anode and the control grid. mainly in valves required to deliver appreciable power. Such a grid has a greater effect on the electrons passing through it but it also increases the spacecharge effect. The heat generated by this filament causes the cathode to warm Tube coated with oxides up. In these valves the cathode is not a simple filament of wire.THERMIONIC VALVES 237 much longer life.
This provides good mutual conductance with a high amplification factor. where the electrons shooting through the screen acquire sufficient velocity to displace some of the electrons in the material of the anode. however. due to an important secondary effect. This can be avoided by interposing an earthed screen between anode and grid. while the electron flow can be controlled by varying the voltage on the inner grid in just the same manner as with a triode. This is because the total current drawn from the cathode is determined mainly by the voltage on the screen. As form of saturation sets in. As the anode voltage is increased further. It is indeed. leaving the voltage on the screen at its normal value. producing what is called secondary emission. is the voltage on the anode increases from zero. An even more important advantage arises from the avoidance of capacitance coupling between the anode and grid of the valve which permits some of the (amplified) energy in the anode circuit to be fed back into the input (grid) circuit and so causes instability. different. but when the anode voltage is about 4/3 times the screen voltage the current begins to limit and a The anodecurrent/anodevoltage characteristic.4. which is nearer to . the anode current at first increases normally.238 RADIO COMMUNICATION anode potential. so that we have the curious result that the anode current decreases while the anode voltage is increasing. it becomes comparable with the screen voltage and there is no longer sufficient difference of potential for the screen to capture the secondaryemission electrons. a triode except that usually has a much steeper however. The electrons emitted by the anode are attracted back to the screen because its potential is higher than that of the anode. The anode current then rises rapidly. The subject is discussed more fully in Chapters 7 and 9. and this from the filament. The total anode current ceases to increase and actually falls. usually about onethird of the attracts electrons effect of the control grid. 5. By the time the electrons reach this second grid they are travelling so fast that they shoot right through it and reach the anode without difficulty. possible for the secondary emission to exceed the emission from the cathode so that over a small range of anode voltage the anode current actually becomes negative as shown in one of the curves in Fig. Secondary Emission The anodecurrent/gridvoltage is characteristic of a tetrode valve it similar to that of slope. which is effectively provided by this second grid. A point is reached. so overcoming the repelling potential.
4 thus in a triode. The effect of this suppressor grid is to remove the kink in the characteristic completely so that a pentode characteristic is of the acteristic unsuitable for . the screen itself draws appreciable current. Unlike the control grid. however. by the grid voltage exactly as it is shows a family of curves for different values of grid voltage.4. to an extent depending upon the relative values of anode and screen voltage. the sum of the anode and screen currents is constant so that the variation of screen current with anode voltage is as shown dotted in Fig. This third. Pentodes its The presence of the secondaryemission kink in a tetrode usefulness because it makes an appreciable part of the normal use.4. In fact. which is what primarily determines the current. 5. can be overcome by the introduction of another grid between the screen and the anode. suppressor grid is in the form of a wide spiral. even when the screen is at a higher potential than the anode. 5. This screen voltage is constant so that beyond the "knee" the tetrode becomes very nearly a constantcurrent device. and serves to introduce a small negative field close to the anode. normally connected to cathode. Tetrode Characteristics The total current is controlled Fig. This has very little influence on the rapidly moving electrons passing from the screen to the anode. limits char This limitation. 5. V/5 Screen current Screen volts =ZOO i/OO 200 volts 300 Anode Fia.THERMIONIC VALVES 239 the cathode and has the major influence on the electric field surrounding the cathode. but it serves to repel the secondary electrons emitted from the anode and thus prevents them from reaching the screen.
and at the same time inhibiting the emission of secondary electrons. This will be seen to retain all the desirable characteristics of the tetrode with the further advantage that there is no unusable portion of the characteristics until the anode voltage falls to quite low values. 5. Cathode Fio. 5. Here the anode is located at the critical distance with deflector plates at zero potential between the screen and anode as shown in Fig.5. 5. . Beam Tetrode A variant of this arrangement is the beam tetrode. Typical Pentode Characteristics It is possible to obtain a similar suppression of the negative emission by a suitable choice of the relative distances of the screen and anode from the cathode but the anode diameter has to be larger than normal. Such valves are known as criticaldistance tetrodes. 5. Deflector plates Anode 9.6.5. 300 Anode Fig.6. which is frequently used for power output valves. 100 200 Volts.240 RADIO COMMUNICATION form shown in Fig. They produce a negative electric field which has the twofold effect of concentrating the main electron stream into a beam.
it practically instantaneously. Variablemu Valves For use in amplifiers where it is desired to vary the amplification by suitable adjustment of the operating conditions. while restoration of the normal suppressor grid voltage will bring it back into its full operating condition it is so that available. although this grid is normally connected to the cathode. there is a type of valve called a variablemu valve. not available for use. which is achieved by winding the control grid with a nonuniform spiral.THERMIONIC VALVES Suppressor Control 241 It should be noted that the suppressor grid may be used for additional control purposes. This. and in many valves it is connected internally. A particularly useful application of this technique is where a valve is to be brought in and out of use quickly without disturbing its other functions. has 'Anode Cathode 6r!d 5 6rid Volts Fio. If the suppressor grid is externally may be found that some control of the anode current can be obtained by altering the voltage on the suppressor grid. The normal signal may be applied to the usual control grid while some subsidiary signal may be applied to the suppressor grid. This is sometimes convenient. 5.7. This means that the control exercised by the grid varies over the length of the cathode so that when the closemesh portion has cut off the emission there is still current available from the openmesh portion. particularly where two controls are required. Applying a large negative voltage to the suppressor grid may completely cut the valve off and prevent it from functioning at all. Illustrating Cutaway Portion in Centre or Grid to Give Variablemu Action and the Type of Characteristic Obtained 10 a variable amplification factor. It is found that if the spacing of the turns on the grid wires is logarithmically . however. as its name implies.
This mixer grid exercises a control on the anode current in a similar manner to the control grid. with a construction separate mixer grid in between them.7. 5.\ Cathode *£ grid r=r Heater Fig.8. 5. This is similar to a tetrode in but the screen comprises two concentric grids. Diagram oir Tbiodehbxode Here two valves are included in the same envelope. The glass foot is then sealed into a bulb and a small tube is provided either in the bulb or in the foot. This effect is known as micraphony because the valve is in fact acting as a microphone in that it will translate mechanical vibrations into electrical e. Typical Construction In a practical valve the various electrodes are supported on stems mounted in a glass foot.. A similar effect may be obtained rather more cheaply by omitting some of the grid wires as shown in Fig. Any movement of the electrodes relative to each other will influence the electron stream and therefore give rise to spurious signals. Mixer Valves In superheterodyne receivers (and certain other applications) there a requirement for a valve which provides an anode current change proportional to the product of two control voltages. while there is usually a support of mica or other insulating material at the top of the assembly in order to keep it rigid.f. With such a valve the mutual conductance (and hence the amplification) varies is smoothly with the grid bias so that a gradual control obtained instead of a sudden cutoff. which can be connected . 5. but the presence of the additional screen around it reduces interaction between the circuits. special valve can be done in various ways. as indicated in Fig. the most usual being by means of a known as a hexode.s. 242 RADIO COMMUNICATION graded an exponential variation of anode current with grid voltage is produced.m.8. This is t Signol anode Mixer grid Screen Signal gridOscillator 'i '". a triode for generating the local oscillation and a hexode for providing the mixing. i anode Oscillator VT^.
THERMIONIC VALVES
to the
sealed off
243
pumps so that the air can be evacuated. This tube is then when the vacuum is complete, and the remaining gas is then removed by raising the temperature of the electrodes by some
suitable
means so that the getter fires as explained on page 232. The connections to the various electrodes are then brought out
to a suitable base which plugs into a socket so that the valve may be removed and replaced as and when necessary. With many
modern
receiving valves the base
is
made
part of the foot
itself,
a
Cathode Mica support
Control,
grid Screen grid
Suppressor grid
Anode
Heater
Connecting pinssealed through glass
(o)
Diode
Fig. 5.9.
(b) Triode
(c)
R.F Pentode
Representative Valve Constructions
series of short pins being located on the glass of the foot at the correct spacing, while for very small constructions the connections to the electrodes are brought out on thin semiflexible wires which are
actually soldered directly into the circuit. Fig. 5.9 illustrates a variety of modern valve constructions. Multiple valves are often used incorporating two or more assemblies round a common cathode. Typical examples are the doublediode
triode,
or the triodehexode
5.3.
shown
in Fig. 5.8.
Basic Valve
Usage
The manner
which valves are used in practical circuits is discussed in Part II, but it is appropriate here to review briefly the basic principles involved. Valves are used in three main ways: as amplifiers, rectifiers and oscillators.
in
244
RADIO COMMUNICATION
of Valve as Amplifier
Use
have seen that the variation of potential on the control grid of a valve will produce corresponding variations in the anode current. To enable this variation of current to be utilized it is necessary to include in the anode circuit a suitable load, across which the variations in current will develop useful voltage, or in which useful power can be produced.
Fig. 5.10 shows a simple triode circuit in which an impedance Z has been included in the anode circuit. If a signal e.m.f. e„ is applied between grid and cathode there will be a variation in anode current
We
J e9
W
Fig. 5.10.
(l>)
(c)
Triode Amplifier
equal to ge a
alternatively
,
where g
is
the mutual conductance of the valve;
we can say that this variation of anode current is that which would be produced by an e.m.f. fieg operating in the anode circuit, where [i is the amplification factor. We can therefore draw an equivalent circuit using these relationships, which can be done in two ways according to whether one uses the voltage or the current
as the primary variable. Fig. 5.10 (6) shows the voltage form. The effective anode voltage /jifig will pass current through a circuit consisting of the internal resistance of the valve (r a ) and the load impedance (Z) in series.
Then
clearly
ia
= l* Wa +
a
Z)
The output voltage
and the amplification
V
=
Zi a
= ———
Z)
Z
.
fie a
A=
ixZ\{ra
+
Z) is the (Note that if Z is not a pure resistance the term (ra vector sum.) Fig. 5.10 (c) shows the alternative (current) form. The anode current ia g m e a flows through ra and Z in parallel, and the voltage
+
=
THERMIONIC VALVES
thus given by parallel, so that
is
24S
ra
V
ia
multiplied
by the impedance of
and
Z in
„
ra
Z
and
A = gmra ZI(r +
tt
Z)
= ^Zj(ra + Z)
since g m ra [i. Thus either approach gives the same answer. In a practical circuit, as will be discussed in later chapters, there
=
are other factors to be allowed for, notably the capacitance between the various electrodes in the valve, but their effect may be assessed by suitable modification of the equivalent circuit. Either form may be used depending on which is more convenient. The same type of equivalent circuit would be used for a tetrode or pentode because the additional electrodes are at a steady potential and do not affect the basic operation of the valve. It will be noted that the amplification, or stage gain, depends on the ratio of the external load impedance to the total impedance, so that when possible the load impedance is made high compared with the
is practicable with triodes where ra is relatively low, but with tetrodes or pentodes ra is very high and, in general, Z
valve resistance. This
is
only a small fraction of ra On the other hand ju is also high which more than compensates for this apparent inefficiency. In fact where ra is large compared with Z the expression for stage gain above becomes simply A gm Z.
.
—
Phase Relations
Because of the varying anode current the anode potential will also In fact it is clear from the circuit of Fig. 5.1(a) that the anode voltage will be the steady h.t. voltage plus or minus the voltage developed across the load Z. If the grid voltage becomes more positive, the anode current will increase, the voltage drop i a Z will increase, and the anode voltage will fall, while if the grid goes negative, the anode current will fall and the anode voltage will
vary.
rise.
Hence the anode voltage variations will be in the opposite direction to the grid voltage variations, being in fact 180° out of phase, if the anode load is a pure resistance. If the anode load is reactive the
anode voltage will be (180 + </>)" out of phase, where phase angle of the load and may be anything between ±90°.
<f>
is
the
246
RADIO COMMUNICATION
Choice of Operating Conditions
It will be clear that, in use, the valve does not operate over the static characteristics of Figs. 5.2 which are taken with zero anode
load, because every change in anode current causes a corresponding To assess the performance under variation in anode voltage. working conditions it is necessary to superpose on the appropriate static characteristics a load line representing the load impedance. Consider the characteristics of a triode as shown in Fig. 5.11. The standing condition with zero signal applied to the grid is at O,
200
Anode
Fig. 5.11.
voltage (y)
400
Showing Use of Load Link
13 mA. Suppose 200 and vg 10, at which point ia with va the grid voltage is now reduced to zero. Then the current will lie curve. At the same time the anode somewhere on the va voltage will have fallen because of the increased voltage drop produced by the increased anode current and we have to find a
=
=—
=
=
condition which satisfies both requirements. To do this we draw B, a line having a slope proportional to through the point is such a line representing where R is the anode load. The line a load of 13,500 ohms for it will be seen that if it is produced it cuts respectively, corresponding to the axes at 375 volts and 275
—
XOT
mA
R = 375/00275 =
will fall to
13,500.
so that the anode volts curve at This line cuts the vg 100 and the current will rise to 20 mA. Similarly if the 20 the valve will adjust its operating grid voltage is increased to 6 mA. 295 and ia condition to the point Y where va
=
X
—
=
=
THERMIONIC VALVES
247
The peak anode voltage swing is thus 195 volts which has been produced by a grid swing of 20 volts, giving a stage gain of 975. It will be noted that the positive anode swing is only 95 volts as against 100 for the negative swing so that there is a small amount of distortion, though this would be quite acceptable for most requirements. If, however, we use a load line such as POQ corresponding to a load of 4,500 ohms the anode swing is +50 and —75, a quite marked discrepancy, which would not be acceptable. The higher the load, the less the slope of the load line and with a triode it is found that, in general, the higher the load the less the distortion. With tetrodes and pentodes, however, the reverse is the case, as is explained in Chapter 10. The design of practical amplifiers thus involves a choice of operating conditions which gives the best approximation to the particular requirements. It will be noted that the anode voltage can rise appreciably above the standing value. If the anode load is a high resistance, as is often the case, the h.t. supply must be high enough to allow for this. Thus in the example just given the standing current is 13 mA which with 13,500 ohms anode load will cause a voltage drop of 175 volts so that the h.t. supply would need to be 375 volts, which is, in fact, the point where the line XOY cuts the axis. If the h.t. voltage were less, say 250 volts, the conditions would need to be recalculated and the input signal would have to be limited. This limitation only applies with a circuit having a resistive anode load. If the stage is transformercoupled there is only a negligible steady voltage drop and the standing anode voltage is practically the full h.t. supply voltage. The voltage on the (inductive) transformer primary winding, however, can be either positive or negative according to the change of current through it, so that the actual anode voltage can swing up above the steady
h.t. value.
It will be noted that in this example the operations take place around a mean grid voltage of —10 V. The circuit has to be arranged to provide this standing (nosignal) bias by some convenient method, as is described in Part II.
Load
is
Ellipse
This load is a pure resistance. With a load containing reactance, there will be phase displacement which will cause the load characteristic to become an ellipse. This may run the valve into undesirable conditions during certain parts of its swing, but the effect is considered more fully in Chapter 10 (Fig. 10.17). only true
We have assumed a straight line for the load characteristic.
when the
248
RADIO COMMUNICATION
as Rectifier
Use of Valve
As
anode
said earlier, a simple diode valve will only conduct when its This rectifying action can be is positive to the cathode.
utilized in various ways.
Consider the circuit of Fig. 5.12
(a).
An
alternating e.m.f. applied to the input will produce pulses of current on the positive half cycles which will charge the capacitor G. In the
absence of any load R, the capacitor voltage peak value of the applied a.c.
will build
up
to the
o r
C=
R<
W
Fig. 5.12.
(A)
Diode with Capacitor
during
If a load
(rather
is
present
it
will partially discharge the capacitor
more than) the negative halfcycle, but the charge will be replenished by a pulse of current through the diode during that
portion of the positive halfcycle when the instantaneous value of the applied e.m.f. exceeds the capacitor voltage. The output voltage will thus vary as shown in Fig. 5.12 (6). Such a circuit (or its solidstate equivalent) may be used in the detector stage of a receiver, as explained in Chapter 9. Fig. 5.12 (c) shows a modification of the circuit in which the diode is connected across the resistor. The action is similar: when the signal e.m.f. is positive, current will flow through the capacitor and the diode (which will have a much lower resistance than the resistor and will thus virtually shortcircuit it), while during the negative halfcycle the diode will not conduct and the charge Fig. 5.13. Leaky will leak away through R. Note that with this grid Rectifier arrangement the polarity of the charge on the capacitor is such as to make negative the side connected to the anode. Fig. 5.13 shows what is known as a leakygrid detector. Here a triode is used with a capacitor in the grid lead and a resistor between grid and cathode. Thus the grid acts like the diode anode in Fig. 5.12 (c) and acquires a negative potential. This will cause the anode current to fall and so develop an amplified signal across the
THERMIONIC VALVES
anode
resistor.
rectifier
249
The arrangement is a convenient combination of and amplifier but is now mainly of academic interest.
Anode
Rectification
is
If a valve
characteristic
biased at a point towards the bottom of
its iave
Fig. 5.2), rectification will be produced by virtue of the curvature of the characteristic, since a positive grid swing
(cf.
produce a greater change of anode current than a corresponding negative swing. valve in this condition is called an anodebend detector, but this usage is virtually obsolete. However, if the grid swing is small the mean anode current is very nearly proportional to the square of the gridvoltage swing. This squarelaw relationship is useful for certain applications, one advantage being that it is not primarily dependent on the exact shape of the valve characteristic and can therefore be used for measurement.* A similar action can be obtained with semiconductor diodes which are more usually employed today.
will
A
Valve Oscillators
If a circuit
is set
in oscillation the currents will ultimately die
circuit. If at each oscillation, however, a small e.m.f. could be introduced, sufficient to bring the current up to its original value, the circuit would continue to oscillate indefinitely. Owing to its amplifying properties, a valve can readily be arranged to supply this e.m.f. and so can be made to generate continuous oscillations. Fig. 5.14 shows the simplest way of doing this. LtC is an
away due to the losses in the
oscillating circuit
e.m.f. will
and due to the
oscillations therein
an alternating
be impressed on the grid of the valve. This e.m.f. will cause magnified oscillations to take place in the anode circuit, which will flow through the coil L2 If L2 is suitably coupled to Llt an e.m.f. will be induced in Lt which will be in the same direction
.
may appear strange in view of the fact that the emission characof a valve obeys a threehalves law. However, the characteristic may be represented mathematically in the form
* This
teristic
ia
=
ge
+
g'e*l2\
+
J/V/3!
+
.
..
where
e
g
= —
(e„
+
ej/t)
dijde,
are the derivatives dg\de a d*g/de,,*, etc.
,
and
g', g", etc,
value of ia when e„ varies about its steady value. Provided the valve is not biased to cutoff and the signal is never large enough to reduce the anode current to zero, then the mean value of the first and third terms does not change but the second term represents an increase in a 2 i„ (since e is always positive) proportional to e
are interested in the
.
We
mean
250
RADIO COMMUNICATION
as the e.m.f. already existing there, and will therefore restore the currents to their original value, as described above, and the circuit
The coil L2 must obviously be will oscillate indefinitely. coupled in the right direction or the e.m.f. induced will be opposite to the e.m.f. existing in L x and will stop the oscillation instead of maintaining it; also the coupling must be greater than a certain critical value, or else the energy fed back will not be sufficient to make up the losses in the oscillating circuit. A useful rule may be cited in this connection. If the coil £2 i s wound in a given direction from anode to H.T. +, the coil Lx must
LjG
Fig. 5.14.
Valve Oscillator
if it is
be in the same direction from cathode to grid
sustain the oscillation.
It will be noted that a capacitor
desired to
Gt
has been included in the lead
action similar to that of the leaky grid detector this capacitor will acquire a charge which will make the grid negative. This will reduce the conductance of the valve and limit the anode current. The valve will thus adjust itself to a suitable operating
to the grid.
By an
point on its characteristic such that there is just sufficient feedback to sustain the oscillation. Various modifications of this arrangement are possible. The tuning capacitance may be connected across the anode coil instead of the grid coil, if preferred, while it is possible for both grid and anode circuits to be tuned. In such a case the coupling required to maintain oscillation is very small and it will usually be found that the internal capacitance between anode and grid of the valve is sufficient to set up and sustain an oscillation. This internal feedback, in fact, introduces problems in the design of r.f. amplifiers, and precautions have to be taken to prevent unwanted oscillation. The action is discussed in detail in Section 9.2, where it is shown that the feedback is positive when the anode circuit is inductive. Hence if both grid and anode circuits are tuned, oscillation will occur at a frequency slightly above the resonant point, at which the anode circuit exhibits an inductive reactance.
THERMIONIC VALVES
Another
circuit, often
251
used because of
is
its simplicity, is
shown
in
employed. The oscillatory circuit is connected directly across the grid and anode of the valve. The cathode is connected to an intermediate point on the coil, at or near the centre. The potential of the anode end of the timed circuit is clearly 180° out of phase, relative to the cathode, with the grid end. But the anode voltage variations produced by the e.m.f. applied to the grid are also 180° out of phase so that they reinforce
Fig. 5.15 (a) in
HfTchoke
which a direct coupling
HX*
»—O
O § +H.T(?)
i
1
(*)
Fig. 5.15.
Hartley and Colpitis Oscillators
the tuned circuit e.m.f. and oscillation results. This circuit is known as the Hartley circuit, named after its originator. Fig. 5.15 (6) shows a modification known as a Colpitts oscillator in which the tapping is taken on the capacitance. There are other forms of oscillator, some of which do not use conventional tuned circuits. The subject is discussed in more detail
in Chapter 12.
5.4. Gasfilled
Tubes
been pointed out that in an ordinary valve all gas is removed. For certain applications, however, a small amount of inert gas is deliberately introduced. Consider a triode in which a small amount of gas is present. If the grid is at a certain potential, depending upon the anode voltage, a current will begin to flow. The electrons in the cathodeanode space will encounter molecules of gas which they will ionize, releasing further electrons which in their turn will ionize further molecules and so there will be a cumulative buildup of current producing the avalanche effect mentioned in
It has already
Section
1.6.
This means that once the current has started to flow it will no longer be under the control of the grid potential but will actually rise to a very high value, limited only by the resistance of the circuit, and moreover this current will continue to flow as long as there is voltage on the anode. It can only be stopped by reducing the
252
RADIO COMMUNICATION
less
anode potential substantially to zero (actually to a value
than
the ionization potential of the gas). The electrons in the gas will then recombine so that after a sufficient time interval (which is normally about one millisecond) the gas is again nonconducting and increasing the anode potential will not produce any current until the critical voltage is reached. The ratio of anode/grid voltage at the "firing" point is called the control ratio. Thus with a control ratio of 20 and an anode voltage of 100, no current will flow if the negative grid voltage is greater than 5 volts. Conversely if the grid is set to —5 volts no current will flow until the anode voltage exceeds 100. It will be seen that such a valve acts as a highly sensitive relay. It passes no current as long as the grid voltage is more negative than the critical value but as soon as this value is reached a large current flows which can only be stopped by cutting off the anode voltage. There are many circuits in which this sort of behaviour is convenient, as will be seen in later chapters. Valves of this kind are known as gasfilled triodes, or thyratrons. The gas used is
usually neon or argon, though sometimes mercury vapour is used. One difficulty with this class of valve is that the heavy ions which drift towards the cathode tend to destroy the emitting surface of the cathode and therefore precautions have to be taken, both in the manufacture and in the conditions of use, to minimize the damage
caused in this way. This type of valve is now obsolete, having been superseded by a range of much more convenient semiconductor devices known as thyristors which are discussed in Chapter 6.
HighCurrent Rectifiers
The presence of gas within a valve permits a considerably increased current to flow. Hence extensive use has been made of diodes containing a suitable gas filling usually mercury vapour as rectifiers in power supply circuits such as are discussed in Chapter
—
11.
Such valves will pass currents from a few amperes in small sizes to hundreds of amperes in larger sizes. The main disadvantage is that the heavy ions can produce destructive bombardment of the cathodes. This can be minimized by special constructions, but it is necessary to delay the application of the anode voltage for a suitable time of the order of 30 seconds to permit stable internal conditions
—
—
to
become established. Such valves are again obsolescent, since the requirements can be more conveniently met with solidstate devices.
THERMIONIC VALVES
Coldcathode Tubes
253
A special type of gasfilled device is the coldcathode tube, which does not require a heated cathode. This contains an inert gas at low pressure but the conditions of operation are rather different from an ordinary valve. Current flows, in fact, as a result of a glow discharge as explained in Chapter 1 .6, and hence differs from emission current in two respects: (a) Current will not flow until the voltage on the anode exceeds
a certain value.
(b)
The current
is
not unidirectional but can flow with either
positive or negative anode potential.
The simplest form of the device is the familiar neon tube commonly employed as an indicator lamp. This will glow when the
applied e.m.f. (in either direction) exceeds the striking voltage, which is of the order of 80 to 130 volts according to the construction, It may be operated from the standard a.c. supply if a suitable series resistor is incorporated to limit the current once the tube has struck. By making the cathode of much smaller surface area than the anode the conductance can be made greater when the anode is positive, and the effect may be augmented by coating the cathode
with suitable emissive material, but the tube will still conduct with the anode negative and any appreciable current in this reverse direction may damage the cathode surface. With this type of tube, therefore, it is necessary to take precautions to prevent the anode
from becoming negative in
Stabilizes Tubes
use.
The cold cathode diode is frequently used as a stabilizing device because of the constant voltage drop which it displays. A typical characteristic is shown in Fig. 5.16. The tube will not conduct appreciably until the anode voltage is sufficient to produce ionization by collision, when the current will rise rapidly to the saturation value. Once the tube has struck, however, the current can be maintained with a somewhat lower anode potential, and moreover, as will be seen from Fig. 5.16 (b), there is a region where a very small change in potential occurs over a wide range of current. In practice such a tube would be used in a circuit such as Fig. 5.16 (a). Provided the circuit conditions are such that when the exceeds the critical tube is not conducting the voltage across value, the tube will strike. The additional current will cause an additional voltage drop on R t which will reduce FAB to some point in Fig. 5.20 (b). The total current will now be in the region
AB
XX
254
RADIO COMMUNICATION
divided between R 2 and the tube, and any variation in the load current through R 2 will be compensated by a corresponding change in the current in the tube. But since the voltage drop across the tube is substantially constant irrespective of the current (within the range
— Striking
voltage
Current
(.a)
<M
Fig. 5.16. Illustrating
Use or Stabilizer Tttbe
XX) the voltage across B2 is held constant. Such tubes are therefore
usually
known
as stabilizer tubes.
Trigger Tubes
If a third electrode is introduced between anode and cathode the point at which the tube becomes conducting may be controlled by varying the potential of this "trigger" electrode, so providing a relay action similar to that of a thyratron. Once started, the current continues to flow until the anode voltage falls below the
ionization potential.
Current will flow when the potentials of the trigger and anode exceed a certain value in either direction, but to avoid damage to the cathode, current should only be allowed to flow when both anode
and
trigger electrodes are positive to the cathode.
Indicator Tubes
With suitable modification a coldcathode tube can be arranged to display a series of numbers or characters. typical device uses a series of cathodes fashioned in the shape of the number or character required. These are assembled in a stack about apart, so
A
1mm
arranged as to produce the minimum optical interference, so that when any cathode is energized its character is clearly displayed. The cathodes are normally held at such a potential relative to the (common) anode that no discharge occurs, but by reducing the potential of any given cathode to zero a discharge takes place to this cathode causing it to glow clearly. As a variant the indication can be provided by a series of (seven) coplanar segments. By activating the appropriate segments a stylized figure or letter is displayed. (This type of tube is also made with a heated cathode,
THERMIONIC VALVES
255
the segments being coated with fluorescent material which glows green when actuated.) An alternative arrangement uses a series of cathodes in a single matrix, such as 7 rows of 5. Then by earthing the appropriate cathodes a suitable representation of the required figure or letter can be produced. This type of tube, of course, requires more elaborate control circuitry.
The use of such tubes
is
discussed further in Chapter 15.
An
earlier
form of tube which combines the functions of display
and counting is known as a dekatron. This carries a ring of 10 cathodes
arranged around a central anode. By arranging a series of "guide" electrodes between the cathodes the discharge can be arranged to step from one cathode to the next every time an input pulse is supplied. This device is also discussed in Chapter 15, but developments in integrated circuitry have made it more convenient to perform any counting operations in the control circuits, and to display the results on a simple indicator tube.
5.5.
Photocells
As mentioned
in Chapter 1 certain materials can absorb the energy
in light waves, causing
some of the electrons to jump to orbits of higher energy level. Under suitable conditions electrons may actually leave the material in a similar manner to thermionic emission. Whether this happens or not depends on the wavelength of the light and the work function of the material. Electrons will, in fact, be emitted provided the wavelength is less than l24/<£ (im, where <f> is the work function. Only a small range of materials have values of <f> such that this threshold wavelength is within the visible spectrum, the most com18 so that A 0685 /im which mon being csesium for which <j> is in the red region of the spectrum. All the alkali metals, however, have low work functions and so exhibit photoelectric effects. For best results the photosensitive material should be in a very thin layer and the usual practice is to deposit a film of caesium on silveroxide or antimony, but various
—
=
other combinations are possible. The wavelength at which the maximum photoemission is obtained depends on the materials used. There is always a pronounced peak in the wavelength/
sensitivity curve.
Vacuum
To
Cells
the emission
it is
utilize
at a positive potential, as in a valve.
necessary to provide a suitable anode A practical photocell thus
256
consists of
RADIO COMMUNICATION
an evacuated glass bulb, containing a photosensitive cathode either in the form of a plate or deposited directly on the inner surface of the glass, while a simple anode in the form of a
Cathode deposited on inner wall of
bulb
•Anode
Via. 5.17.
Photocell
5.17.
is located a suitable distance away as in Fig. The anode is made in this simple form to avoid obstructing the light falling on the cathode. The characteristics of a vacuum photocell are shown in Fig. 5.18. With a given illumination the current gradually increases as the
straight wire or loop
is increased until all the emitted electrons are captured and saturation occurs. As the illumination is increased similar curves are obtained of progressively higher saturation level.
anode potential
Output
§
u
o
Load line
Anode potential
Fig. 5.18.
Characteristics of
Vacuum Photocell
The resulting characteristics are similar to those of a pentode with illumination taking the place of grid voltage, and if a high resistance is connected in the anode circuit, as shown, an e.m.f. will be developed across the resistor proportional to the illumination. The extent of this e.m.f. can readily be assessed by drawing a load line through the operating point as shown in Fig. 5.18. It should be noted that the currents obtained are of the order of microamperes so that load resistances of several megohms are customarily used.
THERMIONIC VALVES
Gasfilled Cells
257
The anode current for a given illumination may be increased by including a small amount of gas in the cell. The primary photoelectric electrons then produce ionization by collision •which results in an increased current. This process, however, must be strictly limited. When the anode
Breakdown
area
Anode voltage
Fro. 5.19. Charactebistics of
Gas filled Photocell
voltage reaches a certain value dependent on the illumination, avalanche effect occurs and a continuous discharge takes place which will destroy the cathode if it is allowed to persist. Moreover, even in the safe region the current for a given condition is not precise, and when the illumination is removed the gas takes an appreciable fraction of a millisecond to deionize. Gasfilled cells, therefore, are
mainly used as lowfrequency relays or occasionally for sound reproduction, under suitably controlled conditions. They are not typical characteristic is shown in suitable for measurement.
A
Fig. 5.19.
Electron Multipliers
The sensitivity of a vacuum photocell may be increased some thousands of times by utilizing secondary emission. A series of anodes is used at progressively higher potentials, these anodes being coated with a material which has a low work function and is thus relatively easily persuaded to emit secondary electrons as a result of bombardment by primary electrons. Thus each anode in turn acts as a fresh cathode giving out three or four times as many electrons
as
it
receives.
16.
Further details of this technique will be found in Chapter
Photoconducting and Barrierlayer Cells There is a further class of photocell which depends for its action on the properties of semiconductors. For example, selenium changes itB resistance under the influence of light and can thus be
258
RADIO COMMUNICATION
used as a photocell, while if a thin film of a noble metal is deposited on the surface of a semiconductor a transfer of electrons takes place in the presence of light across the barrier between the two materials which actually generates an e.m.f. These devices are a particular application of semiconductor phenomena and are discussed in the next chapter. Because of their smaller size and more rugged construction they have largely replaced the thermionic type of
photocell.
;
Semiconductors
6.1. Semiconductor Physics Reference has been made in Chapter 1 to materials known
as
semiconductors, which behave partly as insulators and partly as conductors. Many semiconducting materials are known, the most widely used being germanium and silicon, though there are various compounds such as copperoxide and indium arsenide which exhibit semiconducting properties.
Intrinsic
Semiconductors
Broadly speaking, semiconductors are crystalline substances, having their atoms arranged in an orderly pattern known as a lattice. Germanium and silicon both occur in Group 4 of the periodic table, having four electrons in the outer or valence shell. This results in a crystal having a diamond structure with each atom equidistant from four neighbouring atoms and each valence electron being shared between the parent atom and one of the neighbouring atoms. Two electrons, one from each neighbouring atom, thus form
what
is
known
as a covalent bond.
is
We see, therefore, that a perfect
crystal of
germanium
current carriers.
an insulator because there are no available In practice, even at ordinary temperatures, a few
electrons acquire sufficient thermal energy to break their bonds, thus becoming available as current carriers. Therefore there is a small
semiconductor.
degree of conductivity and the material is known as an intrinsic This conduction clearly increases with temperature and is one of the main limitations to the use of semiconductors.
Impurity Semiconductors
Conducting properties are also conferred by the presence of certain Consider the effect of the addition to pure germanium of a very small quantity of atomic phosphorus, antimony, or arsenic these atoms are pentavalent, having five electrons in their outer or
impurities.
259
Quantum theory postulates that electrons in matter can exist only at specified levels of potential energy corresponding to the distance of To their orbits from the nucleus. it is called an ntype semiconductor. will normally be completely surrounded by ger manium atoms. Impurity semiconductors can also be formed by the introduction of small quantities of a trivalent impurity such as boron.260 RADIO COMMUNICATION valence shells and are comparable in size with germanium atoms. The material thus becomes an impurity semiconductor. Energy Levels appreciate the behaviour of semiconductors fully. say. particularly from the standpoint of energy levels. Since it takes energy to move an electron from an inner orbit to one that is more remote from the nucleus the electrons of an atom will prefer the inner orbits : it is. It is convenient to regard this moving hole as a current carrier and the resulting material is called a ptype semiconductor. This vacancy where an electron would normally be in a perfect lattice is called a positive hole: when an electron moves to. or indium. however. it leaves a corresponding vacancy behind and the hole appears to move to the right. gallium. and owing to the relatively small number of impurity atoms. but the fifth electron is held by no such binding force and becomes available for conduction in the same way as a free electron in an intrinsic semiconductor. Now that the concept of positive holes has been introduced it may be mentioned that in intrinsic semiconductors every time a valence bond is broken thermally not only is an electron released for conduction but a positive hole is also created. the left to fill this hole. The impurity atoms are called acceptors since they accept electrons from the germanium atoms. Thus in an impurity semiconductors there will always be a small concentration of carriers of opposite sign to those produced by the impurity action. and since the current carriers are negative. The impurity atoms are called donors since they donate free electrons. so that positive and negative current carriers exist in equal numbers. . so that one of the electrons of a neighbouring atom is left without a partner. The four nearest germanium atoms then form covalent bonds with four of the five valence electrons of the impurity. Such an atom can hence occupy a position normally filled by a germanium atom. In this case there are only three valence electrons of the impurity to form covalent bonds with the germanium. it is desirable to consider the phenomenon of electrical conduction in greater detail. These carriers are known as minority carriers and those produced by the impurity as majority carriers.
Therefore as further electrons are accommodated in an atom they have to occupy orbits of progressively higher energy level.1. 078 eV for germanium and 112 eV for silicon. In insulators the lower band is filled while the upper one is empty and separated from the lower by a gap of several electronvolts: the probability of an electron gaining enough energy at ordinary temperatures to jump the gap to the higher level is very small and thus the conductivity is very low. conduction then . Consequently in a solid.1.SEMICONDUCTORS 261 however.Donor level level Conduction •^Acceptor bond Valence Filled bond Insulator valence bond Semiconductor bond Conductor (<0 Fig. A comband means that there are no free electrons whereas a band that is partially filled denotes that free electrons are present. Energy Bands in Various Types of Material maximum number pletely filled that it is permitted to accommodate. Since the electrons always fill the lower energy bands first the conductivity of a solid is determined by the situation existing in the highest level energy band that contains electrons and by the relationship of this band to the next possible band. so that even at ordinary temperatures some electrons can gain enough thermal energy to reach the upper energy band. 6. where the concentration of atoms is high. It is also found that when two similar atoms are brought close together there is an interaction or coupling between the orbits of their electrons that causes each individual energy to be split into two slightly different levels. Conduction of electricity is possible only when there is present an energy band that contains some electrons but not the theoretical J Unfilled bond t •a: Unfilled or conduction bond X t ^ Filled or _^. The various situations that can exist are shown in Pig. 6. a fundamental law of nature that no more than two electrons in an atom may occupy the same energy level. In semiconductortype material the gap between the energy bands is much smaller. interaction between atoms produces a large number of energy levels so close together as to form a band that can be regarded as essentially continuous.
The simplest form of semiconductor device is the diode. As a result the nregion acquires a positive charge and the pregion a negative .2. Consider the electrical properties of such a junction. The addition of impurity atoms to the semiconductor provides new permissible energy levels.262 RADIO COMMUNICATION takes place due to the fact that the upper energy band now has some electrons in it and also because the lower energy band is no longer completely filled. In good conductors the valence and conduction bands overlap. they exhibit substantially unidirectional conductivity. The two types of junction so formed are known as "grown" and "alloyed" junctions and have substantially the same electrical properties. In consequence the gap is reduced to a small fraction of an electronvolt. 6. The Junction Diode Consider a piece of semiconducting material in which there are both ptype and ntype regions. but they all basically depend on the provision of a boundary between p. and it is upon the electrical properties of this junction that modern semiconductor electronics is based. which is basically a unidirectional conductor like the thermionic diode and has many applications in communications engineering. they tend to diffuse across the boundary between the two regions. as detailed in materials. since their current carriers are predominantly electrons or holes. allowing free movement of electrons and high conductivity. Such a semiconductor is called a junction diode and may be constructed by one of two principal techniques. Various refinements of these processes are adopted to meet Appendix 4. allowing conduction due to their presence in the conduction band or their absence from the valence band. just below the conduction band with an ntype impurity and just above the valence band for a ptype impurity. The junction may either be formed by the introduction of the appropriate impurities at different times during the growth of the crystal or by fusing a small quantity of ptype impurity into ntype material or vice versa.and ntype specific requirements. so that much smaller increases in energy permit electrons to cross the relative gaps. The boundary between such regions is called a pw junction. Since the majority carriers (positive holes in ptype material and electrons in ntype material) are free to wander at random. Semiconductor Diodes The essential feature of semiconductors is that.
i. these minority carriers being carriers in intrinsic semiconductors. which is of the order of 015 formed by the same thermal action that produces V rapidly in an approximately k(v vf ) approximately. If. If a potential is applied across a junction so as to make the ptype side positive and the ntype side negative. In fact.2. Characteristics or reverse current a Typical Junction Diode is greatly magnified across the junction from p to n and negative electrons from n to p. the existence of this ionized spacecharge region gives asymmetrical conducting and the transfer of an electron all the way across the junction requires the expenditure of a certain amount of energy which is termed the barrier energy or barrier height. dijdv is the forward voltage drop or "hopoff"). 6. electrons in ptype theoretically be material or holes in ntype material. Thereafter it rises = — reverse leakage current i . positive holes will move properties to the junction Reverse voltage K The Forward voltage Breakdown voltage Fio. where vf exponential manner (i. forming an ionized depletion layer. The basic form of characteristic is shown in Fig. 6. These charges are confined to the boundary.SEMICONDUCTORS charge. a perfect diode would then be formed by the junction. In practice there is a small reverse current due to the presence of minority carriers.e. The forward current does not begin to flow to an appreciable extent until the applied voltage exceeds a small amount known as the forward for germanium and voltage drop.2. the potential is applied the other way round there will no current flow because there are no available current carriers of the right polarity. with zero volts across the diode there is a very small 06V for silicon. 263 and an equilibrium is reached.e. The characteristic can thus be expressed . however.
a point is reached where some of the covalent bonds are broken. so that the action is reversible. the current falls to its original small value. Germanium diodes will = — 1). At a certain voltage. If the voltage is reduced. while above it i increases rapidly. It may be noted that this condition can also arise if the forward current is excessive. however. or thermionic devices. which are now obsolescent. For small power applications it is sufficient to limit the current. will withstand reverse voltages of the order of 1. though having a slightly greater forward voltage drop. 6. When the withstand a reverse voltage of several hundred volts with junction temperatures up to 70 to 90°C. but for larger powers it is necessary to mount the material in a relatively massive housing which can be bolted to a metal heat sink having a large radiating area. With such precautions quite small semiconductor elements will handle currents of several hundred amperes. This is a much better performance than can be obtained with intrinsic semiconductors like copperoxide or selenium. which destroys the material.2. As the reverse voltage is increased the reverse current reaches a saturation level and thereafter only increases slowly. where = — = v = V— vf . usually within the range 5 to 15 volts. which increases the current further. This causes the temperature of the junction to rise. Zener Diodes Special forms of silicon diode can be constructed in which the avalanche effect occurs at voltages well below the normal breakdown voltage. so that with any semiconductor device suitable arrangements have to be made to limit the temperature of the junction. i (slightly) negative. an avalanche effect occurs producing a very sharp rise in current. partly by providing adequate heat conduction and partly by limiting the current.000 and can be used at temperatures up to 150°C.264 in the RADIO COMMUNICATION form i = i (e** Below this value i is 0. called the Zener . This is known as the Zener effect and the point where the abrupt transition occurs. Silicon types. but because of the low voltage and leakage current there is little temperature rise and thermal runaway does not occur (provided that the permissible dissipation is not exceeded). leading to a condition known as thermal runaway. Ultimately. causing the release of additional carriers and a cumulative (avalanche) effect occurs resulting in a rapid increase in current. These use a heavily doped pn junction with a very small (normal) reverse current. 1) (1 applied voltage V 0. as in Fig. The reverse current is due to the presence of minority carriers in the material and is thus very much smaller than the forward current.
11. Typical Capacitance Curve fob a Reversebiased pn Diode . The material of the contact is usually a tungsten wire. This consists of a springy pointed wire in contact with the surface of a piece of impurity semiconductor material. for instance.15). point. is 265 precise and consistent. while the semiconductor used is either silicon or germanium. Zener diodes are frequently used in stabilized supply circuits as described in Chapter 11 (Figs.3.contact diode can be regarded as a special case of the junction diode. the capacitance of these diodes is less than 1 pF. Ptype silicon pointcontact diodes are extensively used as detectors and mixers at microwave frequencies. in rectifier. Diode Capacitance In the forward direction the diode impedance is purely resistive.SEMICONDUCTORS voltage. A reverse bias across the junction. Thus a verysmallarea pn junction is formed under the metal point. is The which carefully controlled pulses of current are passed through the thought to produce. The semiconductor used can be either p. a small pregion directly under a metal point in contact with an ntype semiconductor. however.or ntype. Because of the small contact area. The Pointcontact Diode An alternative form of diode is the pointcontact diode. The forming process used in the manufacture.14 and 11. forms an ionized depletion layer so that the device holds a small charge and behaves 15 10 5 Voltage (volts) Fio. 6. where their higher forward resistance more readily matches that of the associated circuit. so that it may be used as a reference voltage.
This effect can be used to provide a controllable variation of frequency in an oscillator for f. = . They have. Applications of Diodes Semiconductor diodes are normally used for the purposes of and detection. hV 1' 2 Hence V jidt {Ijco) cos cot then Q that the device has a squarelaw characteristic. = = = = .c. or a. Chakactbbistics of (a) Semiconductor Diodes Backward diode Tunnel diode (6) .4 (a). while the forward region possesses a negativeslope region as shown in Fig. These will be discussed more fully in their appropriate context. and is also designed to have a low internal loss at high frequencies.variation law. it being sufficient here to note some of the more important of these special forms. The value of this capacitance is an inverse function of the bias. Backward diode 7 (a) (b) Fig. however. 6. {I}hco) % cos 2 cot. Fig. = uses.f.266 RADIO COMMUNICATION as a small capacitor shunted by a high resistance. Tunnel Diode This uses a very thin wafer of material with a high concentration of impurity. In one form the capacitance is given by G kV~ in Then Q kVv% If the diode current is i / sin cot. It was discovered / hi Normal If diode /A*.4. being of the form G kVlln where n is between 2 and 3. In particular. 6. With this construction the reversebias region does not exhibit the usual high resistance. it can be used as an efficient frequency doubler. 6.3 shows a typical = characteristic. the important advantage that by suitable choice of material (as explained in Appendix 4) they can be made to exhibit certain special properties. since cos 2 cot = \{\ + cos 2cot). which has various so = . as with the now obsolete thermionic diode. circuits.m. rectification Varactor Diode This is a modification of the normal diode which follows a specific capacitance.
It is described more fully 6. The junction triode transistor is a logical development from the junction diode and consists essentially of two junction diodes The usefulness of a semiconductor element arranged physically so that the forward current of one influences the reverse current of the other. the reversebias region has a higher conductance than the forward region. which will be seen to be a sandwich of pnp material.3. Switching Diodes Computer applications involve circuits which are required to change rapidly from an "off" to an "on" state. in Chapter 13 (page 609). Such an arrangement is shown in Fig. electrons can transfer from states of high to low mobility. which is equivalent to a negative resistance characteristic and hence can be used to generate oscillations. in 1958 267 and has obvious uses it in generating dynatroritype oscillations. Such a device has certain special applications. Such a device is known as a transistor. The portions E and B constitute one junction diode biased in the forward .4 (6). notably as a microwave mixer. for which can be used at frequencies up to several after J. by passing current through ntype gallium arsenide at high field strength. PIN Diode This is a highresistance diode which is used as a controllable variable resistor rather than a rectifier. as shown in Fig. This produces a decrease in average electron velocity with field strength.SEMICONDUCTORS by Esaki GHz. The frequency is dependent on the thickness of the layer and is in the range 1100 GHz. The Transistor would clearly be greatly increased if the conductivity could be made responsive to some control electrode. 6. who discovered that. 6. With a normal diode (or transistor) there is a small time lag but with suitable construction this can be reduced to less than 1 nanosec (10~ 9 sec). B. Gunn Diode This is named Backward Diode By arranging the Zener point of a diode to coincide with the origin. This type of usage is discussed further in Chapter 18. Gunn.5. analogous to the grid in a thermionic valve.
however. as will be seen shortly. so that a change in emitter current will E Fig. which is called the emitter. the minority carriers diffuse through the base to the junction BC. in effect. The majority . the power gain will be considerable. This factor a is a fundamental property of the transistor and is a measure of the current transfer through the device. 6. Current Transfer Ratio Under these conditions nearly will all the holes emitted into the base reach the collector. In particular. We can now see that a transistor has amplifying properties since it can be regarded as two diodes backtoback. called the current gain. Consider EB. the base is made only a few thousandths of an inch thick and of lower conductivity than the emitter.5. but incorrectly. 6. in which case the current gain has to be these . If the base region were thick or of high conductivity these minority carriers would quickly combine with the majority carriers already present in the region when. which is called the base. the behaviour of the mobile charges at the junction carriers in the portion E. minority carriers. a transistor is often used with the input applied to the base. the transistor is not necessarily used in the manner shown in Fig. the connections being such that the emitterbase circuit is of low resistance while the basecollector circuit is of high resistance. a. which is hence called the collector.268 direction RADIO COMMUNICATION and the portions first B and C constitute another diode biased in the reverse direction. will be swept across the junction EB into the region B. where they become.5. and is typically just less than unity for a junction transistor. but in fact. thus. It is sometimes loosely. Here they come under the influence of the strong electric field due to the reverse bias across the junction and are swept across into the portion C. since the current gain is only just less than unity. The ratio of two changes is called the current transfer ratio. B C The pnp Junction Transistor cause nearly as great a change in collector current.
SEMICONDUCTORS 269 expressed differently.5 constitutes what is called a pnp transistor.5. the circuit being otherwise unaltered except that the battery voltages would be reversed. For an npn transistor the arrow is directed away from the base as in Fig. 6. This is known as an npn transistor and behaves in a similar manner. The arrangement of Fig. which only involve one type of carrier.6 illustrates the arrangement corresponding to that of Fig. The conventional symbol for the input is the transistor will be noted. the base being denoted by a short (thick) line with the collector and emitter drawn from it as shown. 6. In this case the majority carriers are electrons which have a somewhat better mobility than holes.6. are merely adaptations of the basic parameter at which remains the fundamental and most important property of the transistor. with a pnp transistor is directed towards the base. A similar action can be obtained with a sandwich of ptype material between two ntype materials. 6. Any such modifications.6 (6). 6. devices of this type are known as bipolar transistors. Fig. and therefore provide a better highfrequency performance. Because the operation involves both types of charge carrier. . applied to the emitter with the output taken from a The base is thus common to both input and output so that the arrangement is known as the commonbase or groundedbase mode. except that the battery voltages have to be reversed. where Input (a) (6) Commonbase circuit with pnp transistor Graphical symbol for an npn transistor suitable load in the collector circuit. as distinct from the fieldeffect types described in Section 6. Practical Forms of Circuit There are various ways of using a transistor in practice. however. The emitter is distinguished by an arrow which.
Since i e = a. in fact. Hence whereas in a pentode valve the anode current is controlled by the grid voltage.7.i e the current transfer ratio in this mode is a. They will be seen to be similar to those for a pentode valve.) collector resistance. The collector current collector voltage over the IJmA) . 6. the equivalent control in a transistor is exercised by the emitter current. 4Q 20 . Ie =8mA I =6mA e l ~4mA e 60 . 6. as said earlier. in fact. 6. which is the reason for the amplifying action. transistor. In particular (a) it will be noted that: i is almost independent of the whole working range down to zero.80 .7. 1 I 2mA e • 1 1 +/ O 2 4 6 8 /exponttetA Fig. which typically has a value of the order of 098. V (yolts) c Collector Current Curves for Germanium Junction pnp Transistor in Commonbase Connection (b) The almost horizontal slope of the characteristics indicates a very high (a. This means that the transistor can be operated (for small signals) on very low battery voltages. since i e ai e . but it is still very large compared with the input resistance. (c) The factor which mainly influences i c is the emitter current ie the two currents are. . . = device.c.10) relate to a pnp For an npn type they would be similar but the values of V c would be positive. nearly the same. In practice the effective resistance is modified by the circuit conditions. These characteristics (and those of Fig. it is essential to regard it as a currentoperated . For a proper understanding of transistor behaviour.no RADIO COMMUNICATION The characteristics of a typical transistor connected in this manner are shown in Fig.
the small proportion lost due to recombination) flowing in the base. to any It is transistor.c. being of the order of 5 ^A for a smallsignal germanium transistor and virtually negligible with silicon types. 6. 6. in relation to the collector voltage..8.8. difference is insignificant . Cukbent Kelationshifs in Commonbase Mode flows through the base in opposition to that action. as due to the true transistor shown in Fig. current Ib is only important in its effect on the biasing arrangements. depending upon the voltage to which the currents are referred. The d. 6.SEMICONDUCTORS Current Relationships The currents in the different elements (a. 6. 6. It is frequently convenient to develop a bias voltage by inserting a resistor in the base connection. . la.8 very small.The current Ico also currents. Hence in some diagrams i e and t 4 are shown in the opposite direction to those in Fig.J unless (1 — «)7 co e is Hence the directions of the currents is for d. The collector leakage current with the commonbase mode of Fig. so that the base current Ib = (1 The Ico . It may be noted that. 6. which is called the collector leakage current.c.8. /„„. Hence ie ih ie This is a fundamental relationship which applies = + .8 (6). The emitter current ie divides between the collector and the base. the collector carries a reverse current.e. collector current /. .c. not a static relationship because of the presence of leakage In particular there is a current due to the minority carriers in the collector. the greater part (= aie ) flowing in the collector. This is a matter of convenience. when Ib may be comparable with (a) apply reversed. + te V U (a) IP* b ^/ (b) +Ic = 0Lle +ICo b M =0a)IeIco Fio. conditions unless /„ shown in Fig.) 271 are shown in Fig.c.a)/.is thus cd. the remainder (i.6 is very small. which will normally develop a bias voltage in such a direction as to reduce the effective emitterbase potential. In practice the direction of the d.
9. with the 45 Fig.. 6. but influence on the performance expressed in terms of base .10. the smaller values of control current (i b ) indicate a much higher input resistance.6 has certain disadvantages in and for many purposes the alternative arrangement shown is in Fig. Here the input is taken to the base. the controlling factor and since ic is nearly equal to ie ) is now the base current %(= ie quite small usually between 2 and 3 per cent of ie With i e ib is this arrangement the current transfer ratio is no longer dic ldi e (= a). the knee of the curves now appears at a small negative potential. If a which means that a transistor is some 50 times more sensitive to changes in base current than to changes in emitter current. = It is customary to use the symbol oc' or /? for this modified factor 098. 6. the arrangement being emitter mode. but this is still only a fraction of a volt. With this circuit only voltage divider emitter potential.V 272 RADIO COMMUNICATION Commonemitter Mode practice The form of circuit shown in Fig. on the other hand. — . — e ditjdi. 6. = di jdi e dia ldi e = 'a = oc/(l 1 Hence the current transfer ratio dic \dih — a). . but the slope of the "horizontal" portion increased. 6. Now since ib R^^ as the commonis required. applicable to the commonemitter mode. The characteristics of a transistor in commonemitter mode are = shown in Fig. indicating a lower output resistance. Commonbmitteb Circuit emitter earthed. Finally. It will be seen that they are similar to those for the commonbase circuit. Here. the being adjusted to provide a suitable base known one battery — .9 used. i e —i e we can write but is dijdib. is Collector Leakage Current The collector leakage current its is the same as in the commonbase mode. a' 49. although the basic action is the same.
whence than Ico/a. 6. of collector leakage is much less marked with silicon transistors..SEMICONDUCTORS 60pA Ibl20fjA 273 2 4 V (volts) c (. /. Characteristics or ja Typical Germanium pnp Transistor in a Commonemitter Connection Solid curves 25"C Dotted curves 45°C With silicon types the variation of current is considerably less with temperature current is considerably greater. The fundamental relationship is But Ie =I c \ Ih from which the expression . 6.10 in which there is a small collector current even when the base is opencircuited (i b 0). This is I' co and will be seen to increase proportionally with V e The expression above may be written in the form I'eo = — . may be written Ic = U(\  a) + a'/» In terms of the base current therefore the collector leakage a). which is some 50 times greater than Im Its effect /co/(l can be seen in the characteristics of Fig. as said earlier. be seen that if 7C (the true transistor current) is less the base current will become negative. tending to a value when Jc effect = 0. 8 Fig.10. The . it will /'„„. . Ih = (/.„ is usually very small.I'JIol'. with which. = .
The voltage required to As produce these currents. I6 == = — . — in Fig. In these circumstances. be the maximum permissible swing. 6. This would. as shown the predominant factor.4.274 RADIO COMMUNICATION Working Point In both commonbase and commonemitter modes the steady (d. so that a 9 volt battery would suffice. current is of the form / With the commonemitter connection the value of /„ is increased by reason of the collector leakage current Im which in fact becomes To/a.11. Thus in Fig. A transistor will thus only provide a linear output. with a swing from 07 to 75 volts.10 are plotted for various values of i e and i h respectively. in fact.e. as is discussed in Section 6. in which the input signal is arranged to produce appropriate variations in input current. however. is not directly proportional to the current. and the characteristics in Figs. with the input providing variations of input voltage. so that with the commonbase connection the input I {e kv 1). a forwardbiased diode.) potentials on the electrodes need to be suitably chosen to accommodate the necessary current swings.10 a suitable working point would be 0. as discussed earlier (page 264). An The basic .9." Similarly at B the collector current is (nearly) zero and the transistor resistive load of is said to be "cut off. Effect of Temperature mechanism of semiconduction is essentially dependent on temperature. This brings the working value of collector voltage to 39 volts.1. 6.c." Input Characteristics said earlier. when used under currentdrive conditions. the line AOB representing a 1250 ohms in the collector circuit of Fig. so that with Vb = 0. The point A coincides with the knee of the curves so that any increase in i b (along OA) produces no further increase in i c and the transistor is said to be "bottomed. The emitterbase junction is. with a steady base current of 60 fiA. in fact. the controlling factor in a transistor is the input current. 6. Hence the performance of a transistor is considerably affected by the temperature of the junction both in respect of the ambient temperature and the local heating resulting from the current in the device itself. It is nevertheless often required to operate under voltagedrive conditions. 6. due allowance has to be made for the variation of input impedance. as was explained in Section 6. an output proportional to the input. i.7 and 6. which increases rapidly beyond the hopoff point in an exponential manner.
. so that current does not flow until the base voltage exceeds between 06 and 07 V. Silicon transistors are much better in this respect and can be safely used with junction temperatures as high as 150°C. and for safety the maximum temperature of the junction should not exceed 80°C. particularly in largesignal (output) stages. but may have some significance in in Fig. I'eo at 45°C could be some eight times its value at 25°C. as one would expect since the collector leakage is due to the intrinsic minority carriers in the material. with a smallsignal germanium transistor. temperature. For example. 6. It will be evident.10 has the effect of raising all the characteristics. This has negligible influence on the performance as an amplifier. The only operational difference with silicon types is the slightly greater "hopoff".11. Frequency more detailed discussion of the various circuit arrangements appears in the following sections. At temperatures approaching 100°C the leakage current may entirely swamp the transistor action. including a third Effect of A (commoncollector) mode which is sometimes used. This affects the input characteristic. it where switching circuits. and not widely varying.SEMICONDUCTORS 275 important part of transistor design is thus concerned with the limitation of temperature rise. 6. as shown dotted •WO T=25'C 50 T ° hoi* Sp* WO V b (rnV) 150 (b) Fig. with consequent alteration of the working point and limitation of the available swing. but the leakage currents !«. increasing by about 05 per cent per degree C. Transistor (a) Common Base Input Characteristics (b) Common Emitter It also affects the output characteristics of Fig. and I'm are considerably increased as the temperature rises. The current transfer ratio is only slightly affected. This restricts the use of germanium transistors to conditions of reasonable.11. as shown by the dotted lines. 6.
(Note that this is dependent on the basic parameter a not the ratio of the standing currents IJIe . are appreciably better than the pnp type at high frequencies. and the base.f. accelerates the holes towards the collector.) This is not independent of frequency. that whatever the mode. which is > discussed in Section 6. or by direct measurement. but they are still subject to limitations. In addition to this transittime effect. The normal transistor has a high collector resistance and high reverse potential.m. In the first place. As said earlier. The effect may be minimized by making the base region very thin. and is usually given = the symbol a As the frequency is increased. the performance is limited by the internal capacitances. and then providing an accelerating or "drift" field across the junction. but this necessarily limits the A . is not easy to measure. The alpha cutoff. Thus in a pnp transistor the ptype emitter pellet would be doped with both p. however. the diffusion of the current carriers across the base is a relatively slow process so that at high frequencies only a proportion of the carriers reach the collector before the polarity of the applied e. The assembly is heated to a high temperature for a carefullycontrolled time during which the ntype additive penetrates the base more deeply than the ptype and so forms a graded layer which.276 RADIO COMMUNICATION however. in use. the performance is fundamentally dijdig. Alloydiffusion Transistors As just said. particularly that between the collector .5 (page 288). but falls off as the frequency is increased. and its use has certain other disadvantages in practice. This is known as the alloydiffusion process. or fa in common emitter mode. and these various internal factors combine to produce a progressive reduction in the current gain as the frequency is The frequency at which the current transfer ratio increased. reverses.e. various internal effects begin to appear which introduce limitations. modification of the alloydiffusion process is what is called an epitaxial construction. npn transistors. falls to l/\/2 times its zero frequency value (i. so that it has been largely replaced by an alternative parameter /x (or fT ). in which the carriers are electrons (which have greater mobility).and ntype additives. The value at low frequencies (approaching zero) may be determined from the static characteristics. 3 dB down) is called the alpha cutoff frequency and is given the symbol fa in commonbase mode. the highfrequency limitations in a transistor are due mainly to the relatively slow diffusion of holes (or electrons) through the base material.
12 and will be seen to display certain Avalanche Effect The collector of a ^XV^ffl positive J«=0 Fig. as shown. By providing a thin siliconoxide layer over the epitaxial layer the leakage may be still further reduced. To overcome this.12.4. depending on the relative connections of the electrodes. but once the action has started the base loses control until a substantial positive current is injected into the base. and may be summarized as follows: (a) Commonbase mode. Provided that precautions are taken to limit the current so that thermal runaway does not Occur. Further details of these and other forms of construction are given in Appendix 4. 6. These are illustrated in Fig. The characteristics. 6.SEMICONDUCTORS 277 current which can be passed. this effect can be used to produce large controlled current pulses. as discussed in Chapter 18. the main portion of the collector is made from lowresistance material. This maintains the advantages of high reverse potential (and low leakage) with a lower effective collector permitting greater current without saturation. 6. Illustrating * Avalanche Effect effect interesting effects. The characteristics in this region are illustrated in Fig. The onset of the depends upon the base current.13. This is known as planar resistance. turn back on themselves. transistor is a reversedbiased diode so that if the voltage across it becomes excessive it will run into the avalanche region as explained on page 264. 6. with the base common . on which a thin layer of highresistance material is deposited. in fact. Here the input is fed to the emitter with the output taken from the collector. Base and emitter regions are then formed in the epitaxial layer by the conventional alloydiffusion process. Transistor Performance There are three basic circuit configurations for a transistor. epitaxial construction.
with a (voltage) gain of slightly less than unity. of varying complexity according to the accuracy of the representation required. It has a high input impedance and a low output impedance and is thus useful as an impedance changer. but its highfrequency performance is not as good as with the commonbase mode. 6. (b) Commonemitter mode.13 (a). with the emitter earthed (Fig. Input Output Inpui Output (CL) (b) Input Output Fig. It is convenient to obtain a general appraisal of the performance of the three configurations by the use of relatively simple equivalent circuits. Illustrating the Thbkk Basic Transistor Modes Here the input is applied to the (c) Commoncollector mode. This is an arrangement similar to the cathodefollower in valve circuits. This has a substantially higher input impedance and is therefore much more frequently adopted.278 RADIO COMMUNICATION to both as shown in Fig.13 (6)). 6. Here the input is fed to the base.13 (c). The arrangement is characterized by a very low input impedance and high output impedance. after which certain aspects can be considered in greater detail by more sophisticated analysis. .13. The performance of a transistor in any of these modes can be assessed by constructing equivalent circuits. It is also useful as a current amplifier. and is sometimes called an emitterfollower. 6. 6. base with the output taken from the emitter as shown in Fig.
. part being dissipated internally and part appearing B R usefully in the load RL . 6. The source e.14 (6). but as many of the later equivalent circuits involve the currentgenerator concept it is preferable to use this from the . Equivalent Circuit fob Transistor in Commonbase Mode s and L are the source and load impedances respectively. 6. and the stage gain. 6. (a) Fig.i e ) +r — b (i (2) From these expressions we can deduce the input and output impedances.f. 6.c. .i)r b (1) *e ) «. where it is represented by an infiniteimpedance current generator ocv This current will be distributed through the total network across the points AB.SEMICONDUCTORS 279 Commonbase Mode Fig. The resistors re rb and re represent the effective a. outset.14 shows the equivalent circuits for the commonbase connection of Fig. e will produce an input current ie as shown.13 (a). Alternatively the current generator cd e may be replaced by a zeroimpedance voltage generator <xrc i e in series with re as shown in Fig. base and collector respectively. resistances of the emitter. and the greater part of this will be transferred to the collector.m.14. From Kirchhoff's laws e we can e )ie write (i e = (B s + r + = iRL + r — e (i .
v operating into Z \. can write v If — /(e) and hence /(e) = i(Z + B L ).+rt ) i Rl + rb +r c ~.f. i e may be t it> (Bs written in terms of i.» (4) This may be rewritten rb as + ar„ Rs + re +r = Bli b L + r * + r.m. now we can From From (1) above e = + i (Bs + r. so that we .m. In practice r6 (l several times r. may be regarded as being produced by some internal e. obtain from the two Kirchhoff relationships an expression in this form. e.m.. e or c to denote the relevant mode.f. + r * h )i e — ir b (3). whence r.a) b ft . ^ + r.. The input impedance Z = t efi t It is customary to add a second suffix. it will disclose the required value of Z (and /(e) if required). r6+arc r. i. + rJ 4 . (3) we can write Z = r ' + ^[ 1 r + r + RL\ If If BL = BL = oo. it — a) is clear 10 and 20 to 1) may is of the same order as re and since rb is that a considerable variation in Z ft (between be produced by changes in B L Output Impedance To deduce the output impedance a slightly indirect approach is required.f.280 RADIO COMMUNICATION Input Impedance Prom So that (2). will bear some relationship to the input e. i(B L +r +r = b e) b) e = \(BS +r +r e —— — i2 s b t i e (rb + arc ) (3) jJ. 6. Using this nomenclature and substituting the values in eqn. The output current. 0.BL This internal e. Z = Zn = ib rt re +r + r (l .
+ r 1 '•i If . which from A„h * .15.13 The equivalent circuit for the commonemitter connection (Fig. Z = r (l — R8 = oo. ij.r (r + <*r e b c) b b <xrc ) ^s since r + U + r. Commonemitter Mode 6.) SEMICONDUCTORS This is 281 )i in the required form /(c) = (RL + Z (rb <xre b so that Z* =r + b re + )rb R8 + r. . which is a characteristic of the commonbase mode. rb 700 Q and For a typical smallsignal transistor re = = rc = 125 MQ. the current generator should be . 6. indicating that the output is in phase with the input. 6. This will be seen to be similar to Fig. since the primary is variable now the base current.14 with rt and rb interchanged. However. so that again variations in Rs can produce large changes in Zob Av = Stage Gain If RL is small compared with re the voltage gain (4) iR^e. then 87 Zib = 265 Q.) very nearly. 26 If Rs = Q and RL = 2k£l. (6)) is shown in Fig.a) 6 <^ rc This expression is positive. It is useful to evaluate these expressions with typical values. but it must be remembered that the output will probably be fed into a second transistor which will 2 kQ is not have a low input resistance and a value of R L unrepresentative = . 125 Q. Z = Bb kQ and A* = 38 (The value of R L may seem low.^^"** e =r R8 = 0. Zob &b c a. above is _ — (*« (»•» + re + r6 )(i? i + )R L + r + r .(l .
r.(<Wo e e) b (since aV' c = «.Ye . K = hK 1 — <*) and <**« =i _g *» In terms of base current.ie re e The Kirchhoff equations are = . (The symbol ft is often used for «'. i(B L + r' e +r = e) ib («. Equivalent Circuit for Transistor in Commonemitter Mode of reversed polarity (because i„ is in the opposite direction to i e ). <x. = — . therefore. is a' oc/(l Hence the current generator is shown as a'fy and a). by similar reasoning to that for the commonbase mode. then if the e.(» = r (l — a) and a' = a/(l — a) b e) e (5) (6) e (i i„) e Input Impedance From (6).m.) o B RL Fio. ij.L b <$> oWW rb A VWSArrc (la. = a ie — = ie • i e (l — a).(l L e b .f. Writing r' e is to for the modified value.) (7) whence. 6. across the points re a/a' rc (l whence r'c a). Z = it re r + r + R + r + rJLl .282 RADIO COMMUNICATION . <t'ib r' be the same. the current transfer ratio. expressed in terms of ib instead of ie so that Now.a.re ) czre + r + B + a»v . = AB — e where r' e = ^(Bg + r + r — ir = iSL + r' + a'h) + r. — .a) L »".) The effective value of re also needs to be modified.a) r.16.
) + r (ar . .r e b e c . with a lower output resistance.r.a) + i?s = 0. + re + r' c )(B .a) + If If cur. Z„ = 25 kO and A = 38 ve provides a much higher input resistance. Stage Gain An — » (arc (BL r„ is r' c .. Thus the common emitter connection has the advantage that the input and output resistances are not seriously affected by the load or source resistances. we find that on = — = = = Z = ie 127 kii. so that the extreme variation of Output Impedance By the same procedure as for the commonbase calculation it is found that „ ZM =re c (l r r + r {l — a) + B S+ r + r . the second term is of the same order as the first so that the maximum variation is about 2:1. p c . . e. . Z Bs = oo. so that it is preferred for Hence the commonemitter mode . Zu BL = 0.SEMICONDUCTORS since re <^ re 283 and is usually small compared with rc (l — a). Z it is rb is several times usually only about 2 to 1. Z oe oe c ar. The negative sign indicates that there is a phase reversal with the commonemitter mode. smce re <r e = r (l .(l commonbase mode. but r' e a) Rs 125 k£i with BL 2 kQ as before.a) = r (l . Again using typical values (which will be the same as for the 25 k£2) and assuming r..r )R L 8 + r + r. =r +r Z„ = r + a'r e b b «* e In practice.rc /(r ft + rj In practice again. . but depends a' rather than a.. If If #£=00. e) In practice compared with small compared with rb and rc and B L is small The expression can then be simplified to ) A ve ~ *'BL H(B 8 + rb + «VJ This is similar to that for the commonbase mode.. t (care e) e b e ~ r .
+ BL + r d ~ c a) r e is small compared with R L is small compared with r (l — «) and BL The input impedance then becomes Z = rb +RL l(lOL)^ a. we find that Ac In practice = »» + r. The gain is appreciably affected by R s Making R 8 = in the above example increases A m to 74. Equivalent Circuit for Transistor in Commoncollector Mode = so that i/h = hlh = (h + h)lh = hlh + 1 — a) But *c/* = a/(l Hence the current transfer ratio a" = 1/(1 — a) = 1 + a' The current transfer ratio is i\ib . as before. 6. Similarly the output impedance 7oc If _ ~ **« + . but this is not a preferred condition as is explained shortly. would be approximately 50 kQ. 6. Commoncollector Mode The equivalent circuit for the commoncollector mode in Fig.\ and the collector resistance is modified to rc (l— a)."B L c .16. (l»)(*g 1 + rt ) c Rs —* oo (current drive) R s = (voltage drive) Rs is small but finite + r»)lr Z ca r (l — a) Z = r + r (l — a) Z = r + (»» + i2s )(l — «) + (RS oc c oc e 6 oc e . is shown generator is Since the primary variable is again it the current a. o b —vw— r r wv*— Va'i fc re il >* ^ Fig. many applications. ic which with R L = 1000 Q.16.o 284 RADIO COMMUNICATION . But here i ie . 6 From the Kirchhoff equations.
Mutual Conductance This is <x. so that the current will not be A proportional to the signal e. In particular. = . «". are not constant. where g m Since in many circuits the emitter voltage is not equal to the input G mv it where O m is voltage v ( the expression may be written ie (See page 292. the source resistance.f. as is discussed in Section .f. as is frequently the case. with no voltage gain. B s should be large compared with Z in so that the input current is not seriously affected by variations in Z in This is essential with a commonbase circuit. .) (calculably) somewhat less than g m This parameter is called the mutual conductance (or forward The collector current may = = . is high so that the mode is useful where pure current gain is required. and the circuits should be arranged accordingly. 6. however. . However. the emitter resistance is inversely proportional to the emitter current. is unity and in a practical a •Ann _ Rl BL (1 + rh \r e ) + (1 . and to a somewhat less extent in the commonemittor mode. but depend upon the operating conditions. . where voltage drive conditions exist. .m.. and e is the voltage across the emitter. Now. having a high input resistance and a low output resistance.i„ where i e is the current in the emitter. in the commonbase mode. the input impedance is largely dependent on r e and is therefore by no means signal e. the current transfer ratio.5. On the other hand. Since a transistor is basically a current amplifier it should be operated under conditions where the controlling factor is the input current. which is known as current drive.a) This connection is thus principally used for impedance changing. applied to the input of a transistor will constant. Current and Voltage Drive The evaluation of the foregoing expressions is clearly dependent on the values or r e r b and rc These. ie g m e.a) ~ RL + RL (1 . With a commonemitter circuit the input impedance is less dependent on r e and satisfactory operation can usually be obtained if R s is approximately equal to Z in .m.SEMICONDUCTORS assuming 285 circuit. The maximum stage gain B L ^> re and rb is . be expressed in terms of the input voltage in the form dijdv. thus encounter a varying impedance.
.17. and is analogous to the corresponding parameter in valve technique. The action is illustrated by the simple equivalent circuit of Fig. by the operating conditions. the values of these parameters are not constant. This current divides in the inverse ratio of and L the proportion actually flowing in L being R I(R l)Hence the current iL Gmv i l(R R£} and the useful output = R R .5. 6. 6. while the output impedance 1 %h H^ 6 Fig. in which the equal.17. R = R + +R power The input power is v^fR it so that the power gain This is a maximum when RL = R„. in which case (Ap) max = iO m R R L 2 { With practical values this is of the order of 10. there is a substantial power gain. Power Gain Since the emitter and collector currents of a transistor are nearly is many times greater than the input impedance.000 (40 dB). Equivalent Circuit Illustrating Power Gain generator supplies an input power v?\Rit which appears in the output circuit as a current generator O mVi operating into R and RL in parallel.286 RADIO COMMUNICATION transconductance) of the transistor. Transistor Parameters Design calculations in transistor circuits are necessarily based on the parameters r e r„ and re (or certain fourterminal forms derived therefrom. though of considerably different value being of the order of 40 mA/V. . 6. but are considerably influenced . However. as described later in this section). Whether it can be fully utilized depends on the nature of the load R L This is discussed further in Chapters 9 and 10.
SEMICONDUCTORS 287 There are three main factors which influence the performance.c. value. one can say that d. the frequency and the internal feedback path.6. but it is necessary to stabilize the operating current. namely the operating current. be seen later. or less ranging from about 500 Q for a small alloy transistor to 5 transistor. value depends on the materials and construction of the range 50 kO to 5MQ. first increasing at as /„ increases. varies The current transfer ratio a„ is not greatly affected by /„. as will ~ = . Moreover. The additional suffix indicates that this is the zerofrequency or I e very nearly. as such. The but its collector resistance re0 is also inversely proportional to /„. It is thus not die ldv e <x /r e0 Hence for consistent operation is proportional to I e constant. The base resistance r b0 is not appreciably dependent on I e having a roughly constant value determined by the construction. though their effective value is modified by the internal feedback . Q for a large power type. being generally in the . and the a. value is even less. It by about 5 per cent above or below the value at !„ = 1 mA. . Thus a 55°C rise in temperature only produces a 20 per cent rise in T). Hence r e0 is inversely proportional to the emitter current but increases with the temperature of the junction (though to a secondary extent because T is in 25 ohms. this has a value r e0 = kT/el. as explained in The mutual conductance gm0 = = . It can be shown that for any type of transistor. but falling off afterwards. This is obviously quite low.c. Effect of Operating Current The emitter resistance is the forward resistance of the emitterbase diode. since Ic r e0 25// where I is the mean working current in milliamperes. irrespective of size or type. = Boltzman's constant (= 1380 X lO23 e = Charge on an electron (= 1602 x 10" T = Temperature (K = °C + 273) /„ = Emitter current At a temperature of 20°C and a current of 1 mA this works out at ) where k 1 *) is a useful figure to remember. . which kelvins. Variation of a with Frequency The parameters r e r„ and rc are not affected by frequency. Section 6.
in fact. In this region any free electrons will have been . This is a complex relationship having real and quadrature components and the frequency ft is that at which the real part of the 05. = Internal Feedback The third factor affecting the circuit parameters is the influence of the internal capacitances and feedback paths within the transistor itself. known as the transition frequency. This is more easily measured than /„. 3 dB down). the highest frequency at which the transistor can be used as an amplifier (in commonemitter mode). expression This point corresponds very nearly with the frequency. but because the base is very thin the majority will be swept on until they come under the influence of the collector field. and for any given transistor and mode there is a critical frequency known as the alpha cutoff. however.288 RADIO COMMUNICATION paths. . the current carriers from the emitter do not all reach the collector. a' or a" according to the mode) falls to l/\/2 zerofrequency value (i. In a pnp transistor some of the holes from the emitter. At frequencies not too greatly different from the transition frequency.3 (page 276) where it was shown that. is not easy to measure. because of the finite time required for the current carriers to diffuse through the base. The incidence of this effect depends upon the construction and the connection mode.e. markedly frequencydependent. As said earlier. fT at which /? (= <x/(l a)) falls to unity. This was mentioned in Section 6. is The current transfer ratio. and indicates. however. so that it has been largely replaced by an alternative parameter fv Basically the variation in current gain is given to a close approximation by the expression times its The alpha a=a where a is <f> eJW+''A > is the zerofrequency current transfer ratio. = = . cutoff. To appreciate this slightly more detailed equivalent circuits are required. The parameter fT is therefore sometimes called the gainbandwidth product. h ///a and a constant depending on the construction of the transistor. and its use has certain other disadvantages in practice. at which the current transfer ratio (a. — . the effective value of a decreases as the frequency is raised. a. /? can be regarded as being proportional to l//r Thus at a frequency 05fT (8 2. as discussed shortly. in passing through the base will combine with free electrons in the base material.
of this generator is fiV' e where fi is the voltage feedback factor (having a value of the order of 1/2000). 6. a small voltage feedback from the collector to the emitter and this is represented by a zeroimpedance generator in series with the shortcircuit across the end of the line. it is not in a convenient form for practical calculations. 6.m.SEMICONDUCTORS drawn 289 into the collector by the action of the field so that no recombination is possible and the holes continue to move into the collector material.18 illustrates the basic factors involved. 6. in fact. however.* Finally. however.18. that the width of the depletion layer depends on the collector voltage. (This is usually small. Practical w Forms of Equivalent Circuit While the equivalent circuit of Fig.18. It was said. shunted by the capacitor Cd to represent the capacitance of the depletion layer. in a practical transistor it is not possible to make contact . There is. These holes. Mors Commute Equivajueut Circuit reaching the depletion layer are swept into the collector so that the line is effectively shortcircuited. and its width is obviously dependent on the collector voltage. of the order of 75 ohms). as shown in Fig.f. shunted by capacitances representing the emitterbase capacitances. This barrier region is called the depletion layer. appear in the collector as a current so that a separate (infinite impedance) current generator is inserted in the collector portion to represent the current. The e. the holes oc Fig. . The base can thus be represented by a transmission line having series elements representing the forward resistance and shunt elements containing resistance representing the effect of recombination.) used in valve technique. with the active portion of the base material so that a resistance r has to be included to represent the resistance between the active part of the base and the actual base connection. At the end of the line. * This factor must not be confused with the voltage amplification factor ft(= dejde.
various more amenable it can be derived from in Fig. One such transformation is shown become r. the values of which now by suitable simplifications. KEH WW* die —WW— rb Ct>. where the various impedances have been combined into a T network comparable with that of Fig. 6. Practical Forms or Equivalent Circuit (a) Commonbase mode (6) Commonemitter mode . be lh ^ rGDi (a) ruu bb' b > rb l0aj ' cb'e c ce e (b) Fio. The internal feedback generator fiV' e has been eliminated as such.19.14. 6.19 (a).* 290 RADIO COMMUNICATION circuits However. rb = r *o — (1 . while both input and output impedances are modified by the capacitances c b  e and C Vc . its effect being allowed for by suitable modifications to the parameters. 6.«o)^co = f*rc0 r = re0 (\ fi)^ rc0 a = («o — <")/(! — *o e <") The significant modifications are thus that r e is appreciably less than re0 and that an additional resistance r iV has been included in the base.
v SEMICONDUCTORS 291 Fig. V > w . The relevant transformations are rv. The values of the elements are unchanged except that.20.. An network shown in Fig. . e o —— r f « — re/M (SOkfL) oc 9m vb'e( ' :cc (10SpF) (WOOpF) (2sn) * (25 Ma) y\ rbb'(7sn) bo(a) C b 'c (105 pF) ob rbt) f(7sn) ioAAA rb > (2SM/Df c cye (woo pF) oC fee 9m y b'e (so kn) eo (b) Fig. makes use of the fourterminal. 6.f. Fig.15). and is alternative transformation 6.oA" U'c = r e0 l(l — «. is — Equivalentn Circuits rr. since the current generator is shown as xi b the effective value of rc modified to rc (l a) as explained earlier (c. It will be noted. 6. E(jtnvAEEKT7r Cracurrs (a) Commonbase mode (6) Commonemitter mode °e extensively employed in the practical treatments discussed in Part II. . or This form is often more amenable to calculation. particularly with commonemitter circuits.)}* With this form of circuit it is more convenient to represent the current generator in terms of the input e. that signal by reason of the effective input b e is less than the applied In consequence. as was said earlier. however. i.e.19 (6) shows the circuit rearranged for the commonemitter mode. as gm V b e where gm is the forward transconductance. 6. the the voltage drop on r ' .m.f.20. = rj(l — «) r„ = r.
but more usually they are specified by the transistor manufacturer.Rsoo .) = Forward (current) transfer characteristic {di^/dij) with B L = = Output conductance {di /dv with R s = oo = Reverse (voltage) transfer characteristic 2 2) (dvjdv 2 ) with . 6. which are vji v and w 2 /i 2 respectively.18) is represented by an appropriate impedance between the collector and the emitter or base. with gm = 39mA/V. 6.) be noted that in these circuits the effect of the internal = 367 voltage feedback (the generator [iV' e in Fig. the calculations can be simplified if the transistor is re Four Terminal Parameters The T parameters employed garded simply as a fourterminal network. = g m j{l + rw lr but as co increases the capacitance term becomes increasingly important. (The approximate calculation which may be close enough for practical purposes. Fotjbtebminai. Typical values for a smallsignal transistor are shown in brackets against the various elements.20 (6). These impedances can be evaluated from an equivalent circuit of the form of Fig.21. as follows: hf hf h hT = Input impedance (dvjdij) with R L = (to a.  ) approximation the capacitance term may be neglected. however. In practice. 6.c.„) b e ).20. and assuming /= 500kHz. +U *«!** V 2 <2 $*L i Fig. Hybrid Parameters For many purposes the relevant information can conveniently be obtained from four parameters. Using the values shown in Fig. the modified value being Om As a so that first = g m Zv.21. as shown in Fig. so far are useful in providing a physical appreciation of the behaviour of a transistor. the Om value of Qm becomes makes Om It may 358 mA/V. (Blackbox) Network For example. 6.292 RADIO COMMUNICATION effective transconductance is slightly reduced.l(rw + Z = 9 mryel(rw + rVe + J0iGVerwrVe b . 6. two parameters of immediate practical interest are the input and output impedances.
c. ^parameters e c (1  a)r t h ie = r b + pr K.1. while hft is the equivalent parameter for the commonemitter mode (= a' or /J).a)] h. They are also dependent to some extent on the operating conditions. since they are easily measurable. Typical ( yparameters are shown in Table 6.(l + h {e c c /?)] The Tparameters can be derived from the Aparameters by the following conversions re rb re = *« — (1 — a)K*lhb = Kblhb = \jh oh y and z Parameters Parameters can also be derived from the fourterminal network having the dimensions of impedance (zparameters) or admittance ^parameters) the latter being the more useful in practice. = l/r (l . Thus hn is the current transfer ratio with commonbase connection (= —a). As with h parameters. It will be seen that whereas A.SEMICONDUCTORS 293 The first pair are evaluated with the output shortcircuited (to a.1 shows the relationships between these parameters and the relevant T parameters. + r„(\ .c = (1 + P) h re = r /[r + r. and also on the frequency. a second suffix is added to indicate the relevant mode.2.a) = a' . which are usually specified. Table 6./(l — a) = l/[r. the input admittance y tt is simply ijvi. = . Thus by shortcircuiting the a. hf and hr are ratios. output of a commonemitter transistor with a large capacitor. h n = a K> = r„lr e 6. so that v 0. Table hit = r.c.. Fparameters are frequently used in r. fi h ot = r b + r. Hence it is customary to add a further subscript indicating the circuit to which they apply.a) hr = P h rc = (1 + frjr. = <x/(l . so that unless otherwise stated they should be regarded as relating to the lowfrequency (sometimes called static) conditions.) and the second pair with the input opencircuited.and h have the dimensions of resistance or conductance. calculations.f. The values of these parameters depend on the circuit configuration. + K» = 1/r. Hence they constitute a hybrid system (which is why the symbol h is used).
cireuite d f yr gr cT <j> T \ Feedback admittance Feedback conductance Feedback capacitance Phase angle of feedback admittance' j r dlort . even a momentary excess with a transistor will destroy it A and irreparably. cf <t> \ I Qut r t short . cirouited y„ g c <f>„ Output admittance Output conductance Output capacitance Phase angle of output admittance Transfer admittance Transfer conductance Transfer capacitance Phase angle of transfer admittance \ ' j r 8hort . 6. as said. It is thus essential. 6.21 and. are more convenient than the T parameters which are rarely quoted as such.294 RADIO COMMUNICATION Table yi 9i cf (f>t 6. Limiting Values is consideration of paramount importance in transistor circuitry that the rated maximum voltages and currents must not be exceeded. to ensure that under the worst possible conditions the limiting values are not exceeded. ^parameters I ' Input Input Input Phase admittance conductance capacitance angle of input admittance Qut r t short . voltage limits specified are V CBO = V CEO = VEBO = Collectorbase voltage with emitter opencircuited Collectoremitter voltage with base open. but there are certain fundamental transistor circuitry . j rouit e d All these parameters relate to the basic fourterminal network of Fig.circuited Emitterbase voltage with collector opencircuited The last of these is particularly important because it is.2. only a few volts even though the rated collector voltage may be 50 or more. and the circuit conditions have to be arranged so that these limiting values are not exceeded under any conditions. The principal instantly. The manufacturers therefore specify a series of absolute maximum ratings. when choosing a transistor for a particular application. circulted yt g.6. in general. Whereas with a valve overloads may do no permanent damage. Practical Transistor Circuits The design of is considered in detail in the appropriate chapters in Part II.
= — D. A suitable working point could be established by feeding the base from the supply through a resistor. 6. which is by far the most usual.23 in which the base is fed — . these variations can be of major importance. and still be stable with a highlimit sample. so that with a transistor type having a nominal hfe (/?) of 80. the actual value may lie between 40 and 120. Hence the working point can change considerably. In practice the production spread can be as much as 3:1. C. The circuit must be designed to give the required gain with a lowlimit sample. With the commonemitter mode. the value being chosen to provide the required conditions.SEMICONDUCTORS 295 requirements which can usefully be summarized here. . The leakage current /„„ increases with temperature. 6. being in fact IJet' 1^. and the gain. It was shown earlier that with this mode the stage gain is approximately fiR L But since /S <x/(l a) it only requires a small variation in a to produce considerable change in fi. The base current. 6. A significant feature of semiconductor devices is the allowance which has to be made for variations in the parameters. as in Fig. is influenced by any collector leakage current which may be flowing.22 (6). as was explained on page 273). the variation with germanium transistors being considerable. Illustrating Choice or 00 Working Point Any change in collector current will not only shift the working point but will also modify the input and output impedances. however. 6.22 (a). smallsignal amplifier is designed to work in a Class condition about a mean working point as shown in Fig.22. while fl is also affected by temperature and the actual value of /„. A A input 00 Fig. Stabilization A further factor of importance is that the performance of a transistor stage is substantially affected by variations in the collector current. and a more stable arrangement is shown in Fig.
Any increase in collector current will cause Ve (and hence V h ) to fall. i2 2 and this arrangement is often used to stabilize battery receivers. 6. it can also compensate for variations due to temperature changes. A convenient method is to connect a diode across. with a battery supply. 6.24. so reducing at signal frequencies. potential. The circuit will again produce negative signal feedback which may be avoided by decoupling the feedback resistor as at (6). Stabilization that its value increases as the current thereby holding V b constant. or in series with. which a capacitor across it Output Input Fig.296 RADIO COMMUNICATION through a potential divider circuit. may fall considerably below the measure of compensation is possible by making nominal value.C. which. and is mounted close to it. 6.24.23 nonlinear so falls. Here the base potenderived from the collector. If this diode is of similar material to that of the transistor. Any increase in collector current decreases the baseemitter I h and hence Ie R 9 also produces a feedback may be obviated if desired by connecting as shown. tial is An alternative A Out Out Fig. Blt E 2 with a resistor R z in the emitter . . Circuit Providing Better Stabilization circuit is shown in Fig. Neither circuit will compensate for changes in the supply voltage. tending to hold /„ constant. JR 2 hi Fig. Alternative Methods or D.23. 6.
) collector load is appreciable. resistor 6. the increased current will cause an increased voltage drop so that the collector voltage will fall and the transistor will ultimately "bottom.SEMICONDUCTORS The output shown in Fig. as is discussed in Chapters 9 and 10. The collector current is the true current /„ plus the leakage current la. R ^^ Thermal Runaway This question of thermal stability is an important factor in the design of transistor circuits. If the (d. This is known as the .25. This results in a cumulative effect known as thermal runaway which can destroy the transistor. providing a further increase in temperature. Transformercoupled Transistor Circuit have to be differformer primary. mainly in the collector portion owing to its higher resistance. begins to rise and this causes a rapid increase in /„. or a loudspeaker. as The effective a. collector load is then n2 R. and hence the internal temperature. 6.25. but to maintain adequate thermal stability." It can be shown that if the collector voltage is never more than half the supply voltage this limiting action will be sufficient to prevent thermal runaway.c. In this case the collector voltage will be virtually that of the battery since there will only be a small voltage drop on the trans RL may be replaced by a transformer.. and this is usual with transistor circuits. in either case it would have a low value so that to provide a reasonable collector load (of 10 kQ or more) a stepdown ratio would be needed. As the collector current increases. The effect may be limited in two ways. 297 where R is be the input impedance of a second stage. the secondary load and n is the stepdown ratio. the dissipation. The current in a transistor necessarily dissipates a certain amount of power.. R could s«0" Fio.c. and the voltage divider a Rb ently proportioned not only to provide the required working point.
voltage may well be in stage. it is necessary to assess the thermal loop gain. 8I.I.(7 . being mounted in good physical contact with a structure of relatively large surface area and good thermal conductivity so that the heat is dissipated externally and the temperature rise is thereby limited. and if the loop gain is greater than unity. and is generally used in smallsignal Alternatively (or in addition) the transistor may be provided with a heat sink.) . The system. as in an output stage. the conditions are unstable and thermal runaway will occur. by providing an adequate heat sink to conduct the heat away and so limit the rise in temperature.voltage" principle. where V is the supply (battery) voltage and R is the total d.R) = If the collector current 71. . Hence = . where is collector dissipation. in fact. = — Pe = 1. but where large currents have to be handled. if this is impracticable or undesirable. This limits the thermal conduction since mica is a poor conductor of heat. + 8Pe = + 81. stages. Fig.I.2I. and neglecting 8Ie2 8P.298 RADIO COMMUNICATION "halfsupply. the now 7(1. The factor . 6. the transistors are mounted in relatively massive metal casings which may be bolted direct to a chassis or heat sink. The problem is not serious in smallsignal stages. = 81.26 illustrates a typical arrangement.\8T. 6 depends on /„ and on the thermal conductivity of the system so that it may be kept within bounds by limiting /„ or. is less than \V. the rise in junction temperature per unit This must be less than unity. resistance in the collector and emitter circuits.c. as it is the halfsupply.*R 81. a transformer. the power dissipation in the collector Pe Ie Vce where 7ee is the collectoremitter voltage 7 Ie R. forms a thermal amplifier. and for safety about 05.R) — \7 which . but the overall dissipation is usually adequate.)*R By subtraction.coupled which is Q8P. + . This If I is criterion R clearly zero if IC R mentioned above.(I. If electrical insulation is required a thin mica washer is interposed. Now. increases by a small amount dissipation will increase to P.(7 .
Steps must be taken to avoid thermal runaway. This usually Where large basic circuitry involves a transformercoupled input stage of the form of Fig. . Collector Fig.25. 6.26. as shown in Fig. connection Emitter connection . as just described. Construction or Power Transistor Largesignal Stages currents are required. while the input current must be correspondingly increased. the same is used with appropriately larger transistors. as in an output stage. as is discussed more fully in Chapter 10. Output Input Fig. the standing current may be reduced. by using two transistors in a Class B connection. Mica washer Base . 6. 6. Class B Output Stack Variablegain Circuits The gain of a transistor stage is normally controlled by replacing one of the circuit resistors with a variable voltage divider and thereby controlling the proportion of the signal which is transferred to the succeeding stage.27. 6.SEMICONDUCTORS Collector^ 299 1 — ""* r— Transistor 4i Mica washer Chassis ASSSW Insulating a Bolt bush Emitter and base connections . As with valve circuits. with corresponding increase in efficiency.27.
As a result. 6. RADIO COMMUNICATION however (as in a.c. 6. the normal working point would be at O. being in the form of pulses which become of increasingly shorter duration as Ie is increased.28. as shown in Fig.g. If B t is 1 kQ and the battery voltage is 9. and if a (decoupled) series resistor is included in the collector circuit. The i bv b characteristic is approximately exponential in form. The conditions are illustrated in Fig. with Ic 9 2 7 V.29 (6) which shows some representative characteristics. as in shown in Fig. Ve is reduced to 1 V. similar to that with a varimu valve. Hence a given value of input signal (±50 the figure) will produce considerably different excursions of i b depending on the bias voltage V b and since the output in the collector is dependent upon the variation of base current. 6. circuits) where it is desired This can be done by varying the potential of the base and thereby the standing base current. the operating point on the characteristics moves towards the knee of the curves so An that the output swing is limited. circuits involves increasing the base potential. This form of control is called reversebias control. \ "0 T ' 500 Vb (mV) ' 1000 Fig. Illustrating Reversebias Control Provided that the input signal is small an exponential control of the gain can be obtained. and the output swing A'O'B' for the same load and i b swing is only 3 V. 6. = mA = — = . mV . the gain can be controlled by altering the bias.f.28. This increases the collector current.29. alternative method applicable principally to r. to control the gain electrically. The output is obviously distorted. but the range is limited to about 30 dB (316 times). 2 and hence Vc The load line AOB then produces an output swing of 14 V.300 Occasions arise. this increased current causes a reduction in the effective collector potential. If now /„ is increased to 8 mA.
160 301 V. (volts) (b) Fig. Commonbase Amplifiers shown so far have all utilized the commonemitter by far the most convenient for the majority of applications. This is known as forwardbias control and can provide a range of up to 40 dB.24 (a). often called an emitter follower analogous to the cathode follower in valve technique. Emitter Followers The third (commoncollector) mode. which is high output impedance but a very low input impedance. 6. A practical disadvantage is that a separate supply of reverse polarity is required for the emitter bias.f. is . Illustrating Forwardbias Control but when the collector circuit is tuned this is tolerable. It is also used in wideband r. being in fact an example of the Class C operation described on page 323. It has the advantage that it will accept large input swings which is useful in a. This is sometimes useful in circuits which have a very low source impedance. amplifiers where its better h. circuits. whereas with the commonemitter mode there is a phase reversal. such as a movingcoil microphone. The alternative commonbase configuration was discussed on page 279. and this is sometimes useful.29. of the order of 25 ohms. where it was shown to exhibit a relatively circuits The mode.SEMICONDUCTORS . though this can sometimes be obviated by an arrangement such as in Fig.f. 6.g. performance is an advantage. Here the base is held at a small negative potential (assuming a pnp transistor so that the emitter is positive to the base as required).c. A feature of the commonbase circuit is that the output is in phase with the input.
but has a relatively high input impedance.30. 6. so that . which if is r h a). 6. = + Bi=P(V„V b )II a With Vcc — 6 V. drop V e is then \ Vcc and the Input Fig. R2 = 3k£2 and a silicon transistor having p = 50. The base current is Ie lfi. approximately 015 V for germanium for silicon. it results in a reduction of gain. this circuit has a gain of slightly less amplifier. Rt would be 115kQ. A practical circuit is shown in Fig. Emitter Follower collector current /„ where V is and 0*7 V ^V^R^ The base potential must be V e V the hopoff voltage. 6. It may be used as a buffer stage between a highimpedance source such as a photocell and a commonemitter frequently used as an impedance changer.302 RADIO COMMUNICATION As shown on page 285. which becomes somewhat less than the expected value. The load impedance R L will be less than R% because of the impedance of the following stage. Hence it must not be assumed that the input impedance of an emitter follower is always high. conditions must take this into account. The emitter follower differs from a valve cathodefollower in one R L is low. signal. than unity. The gain is not appreciably affected. but the input impedance.30. and the a. +R — important respect. The base potential is controlled by the resistance Rt which may be calculated as follows.31. In general. This is not normally of great significance unless the input signal is appreciably dependent on the input impedance of the transistor. the volt To obtain the maximum swing of the output across JB S should be half the supply voltage. namely that the input impedance is not independent of the output. Hence under operating conditions some portion of the output signal is fed back to the input. falls off considerably L l(l as shown in Fig. .c.
32. 6. Darlington Circuits . With the circuit shown. illustrated in Fig.SEMICONDUCTORS Darlington Circuit 303 Simple combinations of two transistors are often used to obtain a modified performance. One of the most popular is the Darlington pair. which is of use where the load resistance is too low to be conveniently handled by a normal circuit. . the input resistance of Tr% is approximately /?i? t and since this constitutes the emitter load of Trx the input resistance of Tr x is (PR L Thus. 6. Typical Variation of Input Kesistance of Emitter Follower (jS 50) an = Fig. This is used in a variety of ways.32 (a) shows the arrangement as an impedance changer. 6. . 6.32. or superalpha circuit. 100 10000 Fig.31.  Output Input Input lb) )Tr 2 ist Input Output (c) Fig.
which with silicon transistors may be inconvenient. dotted.33. a load as low as 10 using transistors having /? an input resistance of approximately 50k£2. 6.34. . Tr% which is thus operating in the commonbase mode. disadvantage of the circuit of Fig.33. Cascading Various other forms of cascade connection are used. the collector current of Trx can be suitably increased by including a resistor from the supply line as shown = Q A . 6. 6.304 RADIO COMMUNICATION would provide 70. One such arrangement is the cascode circuit shown in Fig. 6.32 (c). Fig. 6.32 (b) the arrangement behaves as a single stage having a very high current gain /Sj^. transistor It . but permits the output voltage (and hence the output signal) to be approximately twice that which could be handled by either transistor alone. This may be overcome by using complementary transistors. so that the emitter load of Tr t would be several hundred ohms. and the input resistance (^Se) would still be reasonably high. Tr2 must have a considerably higher collector current than Trv If this is not convenient.In th*8 case R L would be several kilohms. thus contributes nothing to the overall (current) gain. its emitter amplifier with Here Trx is a normal commonoutput fed to the emitter of a second Output Output Input Cascode Amplifier m m < Beanstalk Amplifier Fia.32 (a) is that the hopoff voltages of the transistors are added. as shown in Fig. It should be noted that in arrangements (a) and (c) the collector current of Tr x is the base current of Tr2 Hence if 2% is to be working under suitable conditions to provide an adequate value of /?. 6. as shown in Fig. the transistor lending itself readily to such circuitry. If Tr 2 is connected as a common emitter stage.
to a.coupled oscillator. and the circuit is sometimes used in r. Fig. The transistors must obviously be of a type having similar nominal collector currents since they both carry the same current.Voltage .35 shows a simple transformer. The transformer ratio is designed to match the highimpedance collector circuit into the lowimpedance base. Feedback and phaseshift transistor operation. The output is taken from the voltage divider R x in the collector circuit. as with a valve. the overall voltage gain is low.SEMICONDUCTORS 305 Because Trx is feeding into a very low load (the input impedance of Tr2 which is approximately 25 Q). as in the beanstalk amplifier of Fig. 6. while the primary is tuned with the capacitance G x to obtain the frequency required.34. Transformercoupled Transistor Oscillator circuits may be used. This is a useful method of extending the output voltage. Fig.35. Transistor Inverters Transistor circuits can be devised to convert d. The process may be continued indefinitely. 6. and/or to replace rotary converters or vibrators when a low. power.c. as is oscillators can also be adapted for discussed further in Chapter 12. amplifiers where stability is more important than gain.c. 6. it is only occasionally used.f. but for this reason the internal feedback through Trx is negligible. Transistor Oscillators By arranging a circuit having a loop gain greater than unity a transistor will maintain continuous oscillations. . but since transistors are now made to handle well over 100 volts. while R 2 is chosen to and any of the customary provide a suitable operating point.
306 d. so that the switching time should be kept as short as possible. then becomes zero and the base voltage will fall to zero. Pushpull Selfoscillating Inverter primary inductance L will at first be practically constant and the di = primary current will rise according to the relationship initially will V =y . the Lood Fia. To achieve high efficiency in a transistor inverter it is necessary that the transistors are either bottomed or cut off. RADIO COMMUNICATION supply is required to be converted to a higher voltage. at the feedback windings F x and F 2 The polarity of these will be such as to maintain conduction in Tr x and to cut off Tr 2 When the core material begins to saturate. the primary inductance will fall and the collector current in Trx will rise even faster until it reaches . collector current . while the reliability is greatly enhanced since no moving parts are involved.c. which constant voltage will be induced be virtually linear. reversing the polarities of the voltages at Fx and F2 so that Tr% will begin to conduct. 6. but have a higher efficiency. The then begins to fall. Such arrangements are not only smaller and cheaper. in each case power dissipation is low.  A a value a'/ 6 di dt .36.36. During switching between these states the power dissipation will rise. A transformercoupled selfoscillating inverter is shown in Fig. This process will then continue as an oscillation. 6. Consider that the circuit is in operation with Trx conducting.
. i. 6. as distinct from the valve.t. As will be seen.37 illustrates such effect transistor (f. so producing a device which is voltage controlled. a region in which no charge carriers are available. so that devices of this type are known as unipolar transistors. Fig. the depletion region spreads until it finally inhibits the passage of current through the slab.e. 6.).e. As the potential of this electrode is increased.SEMICONDUCTORS 307 The output waveform at the secondary will be substantially square and can be rectified and smoothed to produce a d. The Transistor as a Switch Because transistors in general. As with diodes.c. A pellet of ptype material is now introduced in the middle of the slab. in fact. the transition is not instantaneous.36 is a useful way of starting oscillations. The diode will not conduct initially and a current will flow through the resistor to the bases of both transistors causing the transistor of higher oc' to conduct first. When the action has started. to simple lowvoltage logic switching. The addition of the resistor R and diode D shown in Fig. but with suitable manufacturing technique the 9 transition time can be reduced to less than one nanosecond ( 10 sec) Very considerable use is. an arrangement.7. 6. The techniques warrant more than brief mention and are discussed extensively in Chapter 18. can be controlled by using an auxiliary electrode to vary the electric field within the material. The conductance of semiconductor material. which is voltage operated. have an abrupt transition from the nonconducting (reversebiased) state to the conducting state they lend themselves to a variety of switching applications. ranging from the control of very large currents by means of special devices known as thyristors. however. made of switching circuitry. This is negatively biased and so produces in its vicinity a depletion area. but this will result in only a small voltage drop. output. base current for the conducting transistor can flow through D. often employing multipleemitter gates. the mechanism only involves one type of charge carrier. and silicon types in particular. which is A slab of ntype material is called a fieldprovided with ohmic connections at the two ends so that a voltage applied across the slab will produce a current by the normal transfer of free electrons through the material. FieldEffect Transistors a current The essential feature of the bipolar transistor is that it is operated device.
With the nchannel type just described a negative gate voltage is required. the input resistance of the gate is very high of the order of 104 MQ. 6. the slab.308 RADIO COMMUNICATION through. with a shunt capacitance of only a few picofarads.38. 6. The gate voltage at which the drain current is reduced to zero is called the pinchoff voltage. and the device is normally operated with a standing gate bias of about \Vf The symbols for jugfets are shown in Fig. 20r vGS =0 05 10 15 Drainsource voltage V os Fig.i Characteristics of hChannel Jugfet . . which will be seen to be similar to those of a pentode. or and has characteristics of the form shown in Fig. called the drain current Hence the current can be controlled by varying the gate voltage. — gate and drain. + Source — Gate Drain ntype material T z Depletion sPtype material area Fig.41. Such a device is known as a junctiongate fieldeffect transistor. 6. The three electrodes are called the source. 6. and since the gateslab junction forms a reversebiased diode. Basic Structure op a Junctiongate Field effect Transistor (Jugfet) jugfet.37.
If the gate potential is zero it exercises no influence and the drain current is simply the (negligible) leakage current of one of the pn junctions. This permits the construction of devices having insulated gates. If the two separate ntype regions are replaced with a very thin continuous layer on top of the ptype base material. 6. the field between gate and base still controls the current. which is sometimes more convenient.S The gate of a jugfet is not isolated from the rest of the circuit.39. pchannel types are available. with ntype gate material. so permitting the flow of a drain current which is controlled by the gate potential as with the junction type. Stbuctube of Igfet (Enhancement Type) a slab of ptype material being used with nchannel regions under the source and drain connections.E. as with the jugfet.40. Such devices are known as insulatedgate f. and increases as V a becomes positive. which slightly modified. as shown in Fig. electrons are attracted to the surface.SEMICONDUCTORS 309 Complementary. Insulatedgate F.s. However.e. In the symbols the distinction is indicated by the direction of the arrow in the gate lead. but because this layer is very thin. With the arrangement described the drain current is virtually zero with V a 0. the whole being covered with a thin insulating film.t. igfets or mosts (metaloxidesemiconductor transistors). it is a simple matter with silicon material to form an insulating layer of silicon dioxide (or nitride) on the surface. 6. = = . as explained in Appendix 4. as in Fig. 6. which forms with the base a small capacitor. Gate (aluminium) Insulating (oxide) layer Source Drain ptype material ntype material Fia. drain current will flow even when Va 0. but if the gate potential is positive to the base. It is therefore said to be working in the enhancement mode. in which case a positive gate voltage is required.39. An aluminium coating is then formed over the insulating layer between the nregions. The construction is .T.
310 is RADIO COMMUNICATION decreased by a negative bias and in the limit reduced to zero. The device is therefore said to be working in the depletion mode. is indicated by breaks in the symbol as shown. The fact that with the enhancement type the drain current is zero when V g 0. 6.channel oJ^ Igfets: enhancement type Fig. up a voltage sufficient to destroy the insulating Hence they are supplied with the leads twisted together. the materials may be interchanged to provide complementary types. Graphical Symbols for Fieldeffect Transistors O. so that the circuit is virtually opencircuited. Drain <S. because of the negligible Gate Source Drain Insulating Slayer— \Thin nlayer Ptype material / Base Fig. 6. Under such conditions it acquires a static charge which. ® g ««« @=f s ^ pchannel Igfets: depletion type ^t n. An important proviso with any type of igfet is that the gate must never be allowed to remain opencircuited. incorporate a gateprotection diode which obviates the difficulty. Substrate D. 6. Further details of the construction of igfets are given in = Appendix 4. Source As before.41. can build layer. however. . Gate B. 4£= . the graphical symbols being indicated in Fig. Later forms.40.41. Structure of Igfet (Depiction Type) leakage.
y. Fig.c. + Input Output Input m of r ( Output """"G °UtpUt Commonsource Commondrain Commongate Fio. but whereas mode. Configubations possible configurations as on the circuit mode. These are y yt t = Input admittance with output shortcircuited to a.c.e.T. The mutual conductance gm is the real part of yf The reactive part only becomes significant at high frequencies so that for many purposes yf may be considered equal to gm These parameters are not constant but depend on the circuit conditions. they are frequencydependent. = Reverse (feedback) admittance with input shortcircuited to a.SEMICONDUCTORS F.T. F. so that the gain simplifies to gm Z.e. F. 6. there being three shown in Fig. A typical f. = 25 mA/V.f. and the suffixes «.E.c. pentode.c. so that the appropriate circuitry tends towards valve techniques. For example.e. commondrain and commongate modes respectively. r„ == 100 kii and + .43 shows a simple circuit in the commonsource Z). d and g are added to denote the commonsource. = Forward transfer admittance with output shortcircuited to a.s are customarily expressed in fourterminal form. this is not necessarily the case with an f. Circuits These are comparable with the parameters of an r.42. has values = 10 12 .t. 6.42. about 50 MHz. usually in terms of admittances (yparameters) which is most convenient for design. though the effect is usually small below . 6.E.. yis represents the They are also dependent commonsource input admittance.E.T. Thus since they include the internal capacitances.t. .t. 311 Parameters The relevant parameters for f. The voltage gain would be gm Zr \(r with a valve r is many times Z. y yr = Output admittance with input shortcircuited to a.
The commonsource connection provides a phase reversal. as with a valve. as already mentioned. 6. The commondrain connection is. using the appropriate f.f. 6. .t.312 RADIO COMMUNICATION further difference is that the value of yu (~ grm ) is not constant. Simple F. Fig.e. Barrierlayer cells. in effect a cathode follower while the commongate mode is equivalent to the groundedgrid Hh & Circuit c±3 Output Fro.h. Photosensitive Semiconductors is We by have seen that the conductivity of a semiconductor heat. affected Certain semiconductor materials are also responsive to energy in the visible or infrared range of the spectrum. There are three main forms of photosensitive semiconductor.44. Phototransistors.8. and performance calculations be made accordingly. being approximately Z/gm . Typical Variation of g m with Drain Current valve circuit discussed on page 589. as with a valve or commonemitter transistor. Typical circuits are discussed in Chapter 13.T. but depends on the drain current I d as shown in Fig. however. parameters. applications where the impedance of the feed lines is low in any case.43. but this can be tolerated in v. and the arrangement has the advantage that the feedback through the internal capacitance is negligible.44. namely: (a) (b) (c) Variableresistance devices. and this forms the basis for the use of semiconductors as photosensitive devices. the input impedance of the commongate mode is low.E. It also falls off with frequency. As with the groundedgrid valve. may 6. whereas with the other two modes the output and input are in phase. 6. the performance is similar to that of a valve. A Subject to these reservations.
The cell is then annealed to convert the selenium to its pure metallic form and mm suitably mounted. which has previously been provided with two intermeshed grids of a suitable inert metal to serve as connections.45.46.f.f.SEMICONDUCTORS Selenium Cells 313 The most common variableresistance device is the selenium cell.m. known as the barrierlayer cell. The response is not linear.f. which makes it very convenient for light meters . 6. is formed by depositing a very thin layer of noble metal. being of the form shown in Fig.m. The presence of light waves produces an increase in the number of these 300 2>200 £ 'tOOV SO Fig. The deposited layer is so thin as to be transparent and it is found that with such an arrangement the energy acquired by the electrons under the impact of the light appears as an e. such as gold. onto the surface of a semiconductor. causing the resistance of the material to decrease. 6. Hence if a suitably constructed selenium cell is connected across a source of e. Babbieblayeb Cells The second type of cell. and requires no external supply. Selenium is an intrinsic semiconductor and hence its conductivity depends on the existence of the relatively small number of electrons which acquire sufficient energy to break their covalent bonds. the current which flows will vary with the illumination. across the boundary between the noble metal and the semiconductor. This type of cell thus generates its own e. Selenium cells are made up by depositing a thin layer of a suitable selenium compound (about 025 thick) onto a groundglass plate.m. 100 /SO 200 Illumination (lux) Response Ctjbve ov Selenium Cell free electrons.
If. holeelectron pairs will be liberated and more current will flow owing to the increased minority carrier concentration. i: ^oo too d // /. the actual ratio depending on I' m which should therefore be kept as low as possible. Response of Barrierlayer Cell with Different Load Resistances Phototransistobs The third type of semiconductor photocell is a development of normal transistor technique.314 RADIO COMMUNICATION similar devices. Such a device is called a phototransistor and may be used with simple circuits of the type illustrated in Fig. the small current in the absence of illumination being called the dark current. Such a device is known as a photodiode. the illuminated junction forms part of a transistor. The base is open circuited so that the collector current is simply /'. Hence . If a pn diode is biased in the reverse direction under conditions of darkness. substantial current amplification is obtained. /' /y <^ AVSO /^ "SOOn '/. and 400 "CjOO 5ft fSOa. The cell is essentially a current generator and hence works best into a lowresistance load. little current will flow owing to a scarcity of carriers of suitable polarity. if such a junction is now illuminated.46.10 with I b expressed in terms of the incident light. As the load is increased the output falls off as indicated in Fig. 6. 6. The characteristics of a phototransistor are in fact similar to those of Fig. If the base is now illuminated a current Iph will appear in the base which will be multiplied a' times by the normal (commonemitter) amplification.47. however. 6.. . A ratio of light/dark current of several hundred may be obtained with a phototransistor.„.OOOn 100 Illumination (lux) Fig.46. 6.
47. it will draw current through the collector resistance. this must have a high input resistance. It is best therefore to use Output (o) lb) (c) Fig. but this is not usually necessary. .e. but some larger value.48.SEMICONDUCTORS the output this is is 315 not greatly affected by the collector voltage as long as beyond the knee. 6.48 Photometric Units The which incident illumination is is measured in lux. with other photometric units. 6. 6. coupled. in Appendix 2. for the first amplifier stage as in (c).t an emitter follower or an Fig. Hence if the output is fed to an amplifying stage. so that the dark current is no longer simply /'„. 6. Photocell Amplifier Circuits f.c. otherwise. Some phototransistors incorporate a Darlington amplifier as part of the (monolithic) construction. and the light/dark ratio will be reduced. as in Fig. The dark current can sometimes be reduced by connecting the base to earth through a suitable resistor. thereby improving the light/dark ratio. Phototransistor Circuit Photocell circuitry is usually d. the significance of discussed. as shown dotted.48. Output Fio.
.
Part II Practical Applications .
.
7.1. tuned with the inductance Lx (called the aerial tuning inductance V <j i Fig. item (c) being dealt with in Chapter 8. The modifications required for transistor operation can then be discussed at appropriate points. 7. at medium and long waves. behaves substantially as a capacitance. For medium and high powers. Simple Transmitter Circuit 319 .1 which will be seen to comprise a tuned circuit consisting of an aerial which. though for small powers transistor operation is more efficient.The Radio Transmitter The (a) radio transmitter consists essentially of three parts means of producing an oscillating current of the required A frequency and magnitude. The present chapter is concerned with items (a) and (6). 7. Valve Transmitter Circuits The essentials of a transmitter are shown in Fig. (6) A method of modulating the current so that the required intelligence can be communicated. valve techniques predominate. (c) An aerial system wherein these currents produce the electromagnetic waves required.1. Since the requirements are basically the same it is convenient to discuss valve circuitry first.
which builds up to a point where the oscillation is a maximum.T. It is fed therefore through the inductance L 3 which is called a radiofrequency choke. During the next half cycle this charge leaks away through the resistance R. a steady current which passes through L 3 and an alternating component which passes through the capacitor C a into the tuned circuit Lfi^ Oscillator Operating Conditions anode connection to the tuned circuit This is to permit the effective impedance to be matched to the valve as explained later.). The network C3 R in the grid circuit is for the purpose of automatically adjusting the bias. This is effectively in the anode circuit of a valve while the coil i 2 is coupled to Lt in such a manner as to supply the grid with e. grid current flows which charges the capacitor Ga and thus establishes a negative bias. of the right phase and magnitude to sustain continuous oscillation. The remedy is to reduce the time constant of the grid circuit as already explained. The varying anode current is thus separated into two components. The time constant of the will circuit must not be too large or the bias not be able to vary sufficiently rapidly." rise to what is known as "grid tick" or "squegThe frequency with which the oscillations stop and start may vary from one every few seconds to several hundreds a second.I. The H. but if the charge on the capacitor does not fall at the same rate the oscillation will actually cease for a short period until the bias has fallen sufficiently. giving ging.m.320 RADIO COMMUNICATION or A. Suppose the grid bias increases beyond the optimum point.T. supply to the anode is not connected in series with the tuned circuit. and the conditions adjust themselves so that the pulse of grid current each cycle is just sufficient to make up the charge lost through the resistor. Any further increase in bias then reduces the grid swing so that the pulse of grid current is no longer sufficient to maintain the charge and the grid bias falls slightly to restore the It will be noted that the is made variable. . Under these conditions. The amplitude of the oscillation will begin to decrease at a rate depending on the decrement of the tuned circuit. because this is connected to earth. The inductance of La is such that its reactance is large compared with the tunedcircuit impedance. the oscillation will be continually starting and stopping.f. When the grid swings positive. This action automatically provides the optimum bias. optimum condition.
and the final bank may consist of a number of watercooled valves in parallel.s.2. supplies the power to the When Banks of such valves in parallel are used.T. and the swing would be r. B. 3.m. In particular. swing would be F/V2 The alternating power is then ±  [F/V2][/max/2V2] = IF/max . A valve can be used in three main conditions. Class B. Class A. Class zero A Operation A condition the anode current swings between nearly and twice the normal steady value as shown in Fig. A further advantage is that variations in the constants of the aerial and associated circuits do not affect the frequency of the oscillation. the required oscillation is generated by a lowpower oscillator. or to modulate it with speech or music for radio telephony. The output from this oscillator is aerial. is only suitable for a very simple lowpower transmitter and various modifications would be introduced in a practical transmitter. The bias is adjusted to cutoff so that anode current 2. except in the simplest cases.s. as follows: Class A. If this were such as to cause the anode voltage to fall to zero at the instant of maximum anode current. where V is the steady (H. particularly with short waves. Highefficiency Amplifiers. It is a comparatively easy matter to start and stop such an oscillation. If the mean value is I the peak current swing J max = 2J (very nearly) and the r.) value. This is most important. only flows during the positive halfcycles of signal. current is imax/2\/2The anode voltage swing will depend upon the effective anode In the Class . handling only a few watts of oscillating power.1. Class C. impedance. then fed to an amplifier which considerable power output is required several stages of amplification are employed. using valves capable of dissipating more power at each successive stage.m. however.THE RADIO TRANSMITTER Drive Oscillator System 321 The circuit of Fig. The valve is biased beyond cutoff so that the anode 1. but only have a minor effect on its amplitude. the anode voltage V. 7. 7. The valve is biased so that the anode current never falls to zero and the mean value remains unaltered whether there is a signal on the grid or not. and C Operation It is important that these amplifying stages shall convert as much as possible of the input power into oscillating energy. current pulses occupy less than half a cycle.
Voltage and Cubbent Relations with Class Operation A practice neither the anode voltage nor the anode current is reduced to zero so that the voltage and current swings are less than the theoretical maxima and the efficiency rarely exceeds 40 per cent.c.3. This is normal with power output and transmitting valves. over the negative half cycle no anode current will flow Here the valve Imean Fig.2. With Class B Opebation the Valve Cutoff is Biased to if the anode circuit is tuned the resonant action will maintain the oscillation so that the anode voltage will continue to vary but sinusoidally. Note that the grid is shown as running appreciably positive on the positive peak. 7. as shown in Fig.3. the anode voltage will be a faithful copy of the grid voltage. The power supplied to the valve efficiency is VI = ^Vlm&x. 7.322 RADIO COMMUNICATION d. In Grid Swing Fig. Class B Operation Greater efficiency is obtainable by operating in a Class B condition. is biased to cutoff. Thus. although anode current only flows every halfcycle. 7. is so that the maximum with Class A operation 50 per cent. Over the positive half cycle of grid swing the anode current varies in proportion. .
anode current flows in short "rfsww Grid Swing Fia. If the valve conducts for a complete halfcondition) 6 is 180°. Class C Operation Still better efficiencies are obtainable by increasing the grid bias beyond the cutoff point as in Fig.4. but since we are only using every alternate half wave the average value in a Class B circuit is Jmax/w. thereby considerably increasing the peak anode current. Where a modulated wave is being handled Class B working must be used. When smaller proportion of the total time so that the mean value is less. As before. but the d. For Class C conditions is usually cycle (Class B around 100° to 150°. 0/18077. Under these conditions. but efficiencies of 60 to 70 per cent are attainable. The mean anode current can then be assessed in terms of 0. the alternating power is JFimax. so that the circuit can only be used for maintaining or amplifying a steady oscillation. and in fact. 7. 7. The anode voltage is no longer directly proportional to the grid voltage. sometimes called flick impulsing.THE RADIO TRANSMITTER 323 which are designed to permit short excursions into the region of grid current. The anode current with Class B working is thus a series of half The average value of a sine wave is 2/n times the maximum value.4. and efficiencies of 80 to 85 per cent are obtained. Angle of Flow The proportion of the is cycle during which the anode is conducting called the angle of flow. Again this is sine waves. to a reasonable approximation.c. 7 /imax = = . power is now Vlmax/irHence the theoretical efficiency is 7r/4 = 785 per cent. not fully realizable in practice. Class C Operating Conditions pulses of a high peak value. With 180°) we have seen that the mean anode Class B working (0 is less than 180° the current flows for a current is imax/ff.
which that for maximum power output under Class load impedance approximately equal to that of the valve. The point where this load line cuts the limiting edge represents the maximum peak We anode swing. for instance. Now. it will be clear that the lower we make the load (and hence the more vertical the load line) the greater the peak anode current obtained.324 RADIO COMMUNICATION Amplifier Operating Conditions The adjustment of a Class B or a Class C transmitter depends upon the matching of the external impedance in the anode circuit to the valve. the grid is Fio. but the conditions are appreciably different from those of a Class A circuit.5 illustrates a typical series of characteristics of a transmitting valve. 7. can draw a load line from the operating voltage V at such an angle that it corresponds with the load in the anode circuit. since a change of 400 volts corresponds to a change of 200 milliamperes. 7. most of the current will go to the grid. the optimum impedance is lower than conditions. The line shown. Hence beyond this point. represents 2. Characteristics or Typical Lowpower Transmitting grid voltage becomes comparable with that allowed to go positive. but if the on the anode.5 is known as the limiting edge.000 ohms.5. In general. little further increase in anode current results. Since we are concerned with obtaining the largest A Emission Limit 400 600 Volts. Anode Valve possible anode current. 7. determined by the . require a Fig. however. a limit to this. and the thick line on the characteristic in Pig. There is.
The ratio Imaxfl. as discussed in the next paragraph. Performance Calculations Having determined the optimum load we can assess the operating conditions more closely. with a shorter pulse of anode current a.T. value E down to a value m i n as shown E .c. in general. is approximately equal to half the peak anode current 7max» but in a Class C amplifier the ratio is less. 7 . Transmitting valves are usually designed to withstand a peak emission of six to ten times the average value. series of smaller components at harmonics of the funda(c) mental frequency. and secondly. This is than the safe steady anode current. and the methods adopted to radiate this heat effectively are discussed later in the chapter. This asymmetrical waveform can be represented by (a) (b) A d. and. this peak anode current occurs at the point when the anode voltage is at its lowest. Nevertheless. the anode current is a series of pulses followed by an idle period of zero current. a limit. Although the maximum anode current occurs at periods of low anode voltage. An component. It will be noted that the anode voltage is still not reduced to zero. component at the fundamental having a maximum A the peak current will necessarily be greater. voltage. Firstly. The anode voltage swings from the H. since to maintain the same value of I. This energy is dispersed in the form of heat. it is only momentary. is 6 »max/«« 80 375 100 30 120 27 140 24 160 22 180 20 Given this ratio we can then estimate the efficiency. The anode current depends upon the design of the valve. but the overriding consideration is the dissipation. known as the emission limit.c. though the relationship is not The actual ratio depends upon circuit conditions but given to a fair approximation by the following table quite linear.. value I. As we have seen. It is the first two components in which we are primarily interested.T.THE RADIO TRANSMITTER 325 considerably higher maximum safe emission of the valve. for two reasons. does exist. signal frequency. is clearly dependent on the angle of flow. but we need first to know the signalfrequency component of the anode current. In a Class B amplifier I. there is nevertheless an appreciable quantity of power to be dissipated at the anode. the anode voltage swing is about 80 per cent of the H. and the load must be such that it cuts the limiting edge of the characteristic at or below the emission limit.
This must be such that the peak input signal is just sufficient to swing the anode current up to the limiting value. CiBctTiT of Lowpower Transmitter tor 150500 kHz Keying is effected by breaking the earthy aide of the H.T. . The difference between P and P. This may be determined from the characteristics. E power will be (E. the operating conditions must be suitably amended (e.s. as said earlier.6. The network across the key is to absorb the energy when the current is suddenly broken. is dissipated as heat. 7.5 to be correct working load. r. by limiting the peak current). is the power lost in the valve.g. can Anode Tap Having found the Pig. . cos 46/(1 . 7. as explained The input power on page 323.m. RADIO COMMUNICATION Hence the peak signal voltage.B. so that Output power is En — Emtn The x.cos id) it where Ea is the peak positive grid swing if this is not known. or it may be estimated with sufficient accuracy for practical purposes from the expression Ei{a = *. and this must . supplies. (E — ^mm)// it is which will be seen from necessary to "match" the tuned Primary oscillator IflOOM* ATI Fia. whence the operating efficiency P.ly/2)(I~I^2). jE? i not exceed the permitted rating for the valve.5. which. and O. — \(E — m n)/» P is EJ where I = (e/1807r)Ima = P. If it does.IP is readily calculable. The final calculation required is that of the grid bias. s. . be taken as 80 per cent of the minimum anode voltage.//* + E.326 in Fig 7. and so eliminate sparking.
the effective resistance of is Rt + (M.T. where k is the coupling factor and n ^(LJLi). as is usual. 2 s fc Z /» .6 shows the circuit of a simple lowpower telegraphy transmitter utilizing these principles. It should be noted that when. Such stages are known is Output Where the required output single valve. This is done by tapping the anode connection across a suitable portion of the inductance. Then. to adjust its effective impedance to the required value. as would be required if the valves were in parallel. several valves may as driver stages. Fig. The advantage of the method is that the peak current demand from the supply is only that of one valve instead of twice as much. i. Energy Transfer It will be seen that the then aerial is inductively coupled to the tuned anode which circuit of the final amplifier stage. Since Z 2 LJCR. During the first halfcycle of the input the first valve operates in the normal way. the first of which is not tuned but uses a simple inductance for its anode load. the grid current which flows on the peaks imposes a load on the preceding stage which has therefore to be designed to supply a (small) power. 7. 7. so that each valve supplies power to the load in turn. connection.16 illustrates a pushpull output stage. is the resistance of the aerial circuit. Keying is achieved by making or breaking the negative H. Fig. an output stage is driven slightly into the positive grid region. In such cases it is often convenient to arrange the valves in a pushpuU circuit. The valves feed a common load through a transformer or tap which reflects into each the same load as it would require if operating alone. Zl 2 tPLJCR.e. while the second incorporates a tuned anode circuit with a matching tap as just described. The arrangement is used considerably in audiofrequency circuits and is more fully discussed in Chapter 10. During the next halfcycle the conditions are reversed. = = = = PushPull more than can be provided by a be used in parallel.ca^lZt )Sa where M is the mutual inductance and R. with a resistancecapacitance network across the key to suppress sparking. The master oscillator is followed by two triode amplifiers. and the final result obtained by trial. while the second is driven negative and hence is inoperative.THE RADIO TRANSMITTER 327 circuit. this can be written Zt ' In practice a rough preliminary calculation is made to determine the approximate tap position. including the radiation . The valves are arranged in pairs with a centretapped input circuit. as was shown in Section 2.4.
By suitable adjustment of this capacitance any internal feedback may be counteracted. It is therefore important to avoid any feedback between the output and input whether round the whole system or in the individual stages. but there remains the coupling via the internal capacitance between anode and grid of the valves themselves. themselves. and more symmetrical. but the frequency of the oscillation is unaffected since this Some broadening is entirely determined by the master oscillator. which coupling is permissible. with results in an energy transfer of 85 to 90 per cent. M is transferred to the aerial. at which condition 71 per cent of the primary power ' M . This could be done in Fig.15 and 7. in fact. 7. This is discussed in Chapter 9 (page 413). 7. different. Since the 2 2 co IR 2 secondary is tuned Z a = R 2 so that R x = B x + The coupling between the primary and aerial circuits is subject to the limitations discussed in Chapter 2. one end feeding the grid and the other connected to the anode through a small neutralizing capacitor which is adjusted to obtain stable operation (which will occur when the capacitance is of the same order as the internal valve capacitance). Accidental coupling through mutual inductance or stray capacitance can be avoided by proper attention to layout. arrangement is adopted in the circuits of Figs. M Neutralizing It is essential to ensure that the r. Tighter coupling produces double tuning points and the circuit is liable to change spasmodically from one frequency to the other. an advantage in handling the sidebands of a modulated transmission. oscillate or even exhibit any tendency to do so. Hence tighter i 2 co equal to 2 or 3 times R^^. A .f. of the top of the aerial tuning curve is. If the circuit were selfoscillating the coupling would be limited to the critical coupling 2 w 2 = R^ILi. RADIO COMMUNICATION which is the apparent resistance due to the power expended in radiation as explained in the next chapter. With a driven circuit this limitation does not apply.328 resistance. amplifying stages do not.6 by connecting a small capacitor from the anode of the final valve to the grid of the preceding stage. Increasing the coupling will cause the aerial resonance curve to become doublehumped. where it is shown that the effect may be overcome by arranging to neutralize the energy fed back through the valve by an equal and opposite amount of energy introduced into the circuit in the opposite direction. where it will be seen that the tuned circuits in the intermediate stages are centretapped.16.
*» 400 600 Volts Anode Fig. . £. so causing interference. 7. therefore. 7.2. Any variation of frequency will cause the modulation sidebands to encroach upon transmissions occupying the adjacent frequency channel. 7. is to use the highest value of grid leak which can be employed without Under these conditions.THE RADIO TRANSMITTER Screened Valves 329 The alternative is to use a tetrode or pentode instead of a triode and this is frequently done in the early stages of a power amplifier though triodes are still used in many instances for the final stages. the intermittent operation (squegging). The coils and capacitors of the master oscillating circuit should be minimally dependent on temperature and other atmospheric conditions. Hence it is important to minimize these variations as much as possible.7. The next requirement is to minimise the variations caused by changes in the valve impedances since these are connected across part of the tuned circuit. Characteristics op Positivedbive Tetbode The characteristics of a tetrode power amplifier intended for positive drive are shown in Fig. It can be shown that stable operation is obtained if either the anode resistance or the grid resistance of the valve can be maintained constant. grid resistance of the valve is approximately half that of the grid leak. One method of maintaining stability. and in general a tetrode or pentode has a greater efficiency. Frequency Stability Modern conditions require that the frequency of transmitters shall be both accurate and constant. The limiting effect with positive grid volts is not apparent here. This is discussed more fully in Chapter 10. and therefore tends to remain constant. and influence the frequency to a small extent.7.
A R series resonance between L and C. In practice these two frequencies are quite close. 7. is to use some form of electromechanical oscillator as the primary source of frequency. which makes use of the piezoelectric effect mentioned in Chapter 1.9 (a) shows a] crystalcontrolled valve oscillator. an e. It depends on the dimensions of the crystal slice and is generally in the megahertz region. This is .f. when it frequency becomes appreciable. while the elasticity is represented by the capacitance C. If a thin slice of quartz is subjected to strain. is developed across it. corresponding to a Q of 20. in which the inductance L simulates the mechanical inertia. In either case the resonance is extremely sharp. either tension or compression. because a change of frequency then produces a rapid change of phase. 7. of mechanical resonance of the crystal slice. differing only by about 1 per cent.m. represented in electrical terms by the equivalent circuit of Fig. The actual operative mode depends on the circuit.8. It will be seen that there are two resonant conditions.m. The other is a parallelresonant condition determined by Clt at which the effective impedance is very high. Crystal Control The more usual method of obtaining the requisite stability. Pig. however. with the resistance representing the molecular friction.000 or more.330 RADIO COMMUNICATION circuit with a high Q is an advantage. and this exercises A a correcting influence on the frequency. 7. the loss due to molecular friction. at which the impedance is simply B. The most commonly used device is the quartz crystal. Connection plates Crystal slice i Fig. The resonant frequency is high.8. Equivalent Circuit op Quartz Crystal Conversely an alternating e. One is the possible. though by special constructions resonant frequencies as low as 200 Hz are It will be evident that such devices can be used as a quartz crystal can be source of extremely constant frequency.f. applied across the crystal will cause The effect is very small except at the it to expand or contract.
7. precise operation is obtained. 5. if the crystal has several possible resonances (close together) as explained in the next section. The problem of frequency stability is even more important with transistor oscillators since the parameters of solidstate devices are markedlytemperaturedependent.10showsa crystalcontrolled . Transistor Crystal Oscillator here by feedback through the anodegrid capacitance of the valve. by tuning the anode circuit to a frequency slightly higher than the desired parallelresonant frequency of the crystal. Two Forms of Crystal Oscillator Circuit the circuit has the disadvantage that.9. and to the extent that. 7. Only approximate tuning of the anode circuit is therefore required.15) in which the tuning capacitances are the anodecathode and gridcathode capacitances of the valve. the anode circuit is inductive (as explained on page 415). The crystal operates at a frequency just sufficiently below resonance to permit it to exhibit an inductive reactance.9 (b) which is effectively a tunedanode tunedgrid arrangement.10. If the crystal has only one mode the anode tuning may be omitted. Hence. Fig.7. 7. Oscillation is maintained Fig.THE RADIO TRANSMITTER effectively 331 a Colpitts circuit (Fig. An alternative circuit is shown in Fig. it will select the mode having the highest Q. which is only positive when. but i—fffl5W5>— + < it f "v Tuning capacitor ^ (sometimes omitted) Fig.
The simplest mode of oscillation which can be produced in quartz is longitudinal. and the axis along the length of the crystal the . 7.332 transistor circuit. the axis at right angles to this the Yaxis. slices are normally cut from the crystal with sides nearly parallel to the Yaxis. This is known as an Xcut crystal since its face is normal to the Xaxis.730/< hertz. 7.11 (a). optimum temperature characteristics are obtained from crystal slices cut with the length at an angle of +5° from the Faxis. / (a) (b) Fig. presents a high impedance to any but its (series) resonant frequency. which is Crystal Modes way since the piezo The crystal slice has to be cut in a particular electric and elastic constants are dependent on the orientation of ir */. so that the feedback is only operative at this point. however. Illustrating Quartz Crystal Slices within the crystal. resulting in a preciselydetermined frequency of oscillation. and these slices are the slice section as . where t is the thickness of the slice in millimetres.11. In fact. The axis joining two opposite corners is called the Xaxis. RADIO COMMUNICATION This contains a tuned circuit in the collector maintained in oscillation by feedback to the base via the coil Lv The crystal. Quartz crystals are hexagonal in crossshown in Fig.—». and for this application.Zaxis. The resonant frequency with this cut is approximately 2.
which occur at a distance 0224 times the length from each end of the crystal. they have largely been replaced for this purpose by rotated Fcut crystals operating in one ofthe shear modes in which squares in the crystal are distorted into rhombi. and this can be reduced to ±10 parts in 10 s by careful cutting.THE RADIO TRANSMITTER 333 as +5° Xcut bars. Such crystals are used in oscillators down to 2 kHz. TovA crystals operating in the thicknessshear mode are made to resonate at fundamental frequencies up to DT(— 52°) A 20 MHz. and some of the more commonly used cuts are shown diagrammatically in Fig. The temperature coefficient of frequency is very dependent on the wit crystals can have a normal type and angle of cut. The angles of cut are chosen to give the best combination of temperature coefficient.11 (b). activity and known harmonic distortion.BTcut crystals can be made to produce fundamental resonances at even higher frequencies. the inner oven being maintained at a slightly higher temperature than the outer. Here the crystal is held by four wires mounted at the nodal points. These ovens are very often double. and the outer being kept at a temperature higher than the highest ambient temperature ever likely to be encountered in AT — . as described later. or ET(+66°) cut crystals operating in the faceshear mode. while even lower frequencies can be obtained by adopting a bimorph construction in which two thin plates of crystal are bonded together. Oscillations at lower frequencies can be obtained by operating Xcut crystals in the thicknessflexure mode. while the AT{+35° 15') and BT(— 49°) cuts are used for higher frequencies. for frequencies above 20 MHz overtone crystals are generally used. An alternative Xcut at —18° to the Faxis is frequently used in crystal filters because it produces a very pure frequency spectrum. 7. Higherfrequency oscillators can be made using Xcut crystals in the thicknesslongitudinal mode. These bimorph or duplex crystals operated in the thicknessflexure mode can be used in oscillators at frequencies as low as 200 Hz. while . but owing to the poor temperature coefficient of Xcut plates. However. Such crystals are used for oscillators in the range 40200 kHz. These crystals are cut from planes which are rotated by various angles from the ZSbxia. but typical tolerance of ±25 parts in 10 6 from 20°C to +70°. Oscillators in the range 150950 kHz normally use CT(+38°). Most other crystals have a parabolic frequencytemperature characteristic with the zero point normally between 30 and 50°C but giving a rather worse tolerance that the ^Tcut over a wide frequency range. Where very close tolerance is required the crystal is housed in a temperaturecontrolled oven.
Since one of the anodes receives an impulse every halfcycle. circuit capacitance.12. The valves. utilized to "pull" the fre Frequency Multiplication At frequencies above about 25 MHz the quartz slice becomes so thin as to be too fragile for convenience.F. The input circuit is tuned to the fundamental frequency and the anode circuit to twice this frequency. since any spurious modes of vibration relatively close unevenness can result in together. In many cases the crystal is supported by these connecting wires within a glass envelope. This frequency is then doubled one R.C *Moj * h. The most usual is a variant of the pushpull arrangement. from which connections are taken at a nodal point. 7. the assembly being then suitably encapsulated. Fig. are biased to cut . Crystal Mounting ground to a high precision.334 RADIO COMMUNICATION the particular climate. which occurs at twice the fundamental frequency. The cut of the crystal is then arranged so that the zerotemperaturecoefficient point of the parabola occurs as near as possible to the oven temperature. as shown in Fig. it will be seen that the anode circuit is satisfactorily energized. off.12. Fbbquency Doubling Cibcttit or more times by suitable circuits. which may be evacuated to reduce molecular The crystal slice has to be friction and. and the crystal may prefer to operate in one of these spurious modes in preference to the intended one. Hence it is usual to form the electrodes by depositing a suitable metal film on the face of the slice. of course. It is therefore customary at the higher frequencies to generate a frequency which is a submultiple of the value required.so increase the Q. Any grease or scratches on the surface of the slice can also cause erratic behaviour. The parallelresonant frequency is affected to a minor extent by the This is sometimes quency to the exact value required.t. 7.
this frequency the crystals become unwieldy and fragile. both of . It is more usual for such frequencies. the circuit can be arranged to provide an output frequency several times the input frequency.f. so avoiding the necessity for subsequent frequency ing. Alternatively. By suitable choice of dimensions and mount can be designed to operate in one of these harmonic modes. Overtone Crystals An alternative arrangement for generating highfrequency tions is oscilla to use an overtone crystal. AT Electromechanical Oscillators Crystal oscillators are not normally used below about 50 kHz. Alternatively. Oscillators are con(or BT) cut structed using third. to divide down from a higher frequency as explained in Chapter 10.m. transmitters as explained on page 354. The fork is mounted on a rigid base with two coils located near the ends of the tines. though crystals are made for frequencies as low as 200 Hz. quartz plates multiplication. fifth and seventh overtone crystals to resonate at frequencies between 15 and 200 MHz. in the grid coil which causes amplified currents to appear in the anode coil and if these are in the right phase the vibration will be sustained. So far we have considered the crystal as vibrating in its fundamental mode.THE RADIO TRANSMITTER 335 Several stages of frequency doubling are often used to provide four or eight times the initial frequency. The vibration induces e. however. and this is often done in f. in its fundamental It is important to note the difference between operating a crystal mode and electrically selecting a harmonic.m. though with decreased amplitude. The frequency is thus dependent almost entirely on the mechanical constants of the fork and is only influenced to a minor degree by the maintaining circuit. as shown in Pig. A tuning fork constitutes a mechanical resonant circuit and can be maintained in a state of oscillation by a valve in similar fashion to an electrical tuned circuit. and actually driving the crystal to vibrate at a mechanical overtone which is never exactly integrally related to the fundamental because of the necessarily imperfect mechanical structure. for these lower frequencies use may be made of Below tuning forks or magnetostriction oscillators. 7. but it will be appreciated that it can also vibrate in a harmonic mode.13. The primary frequencydetermining factors are the length of the tines and the elasticity of the material.
and is given by v/l. Otherwise it will vibrate at twice the applied frequency. will provide amplified currents in the anode coil and hence build up the oscillation. Certain materials. 7. so proportioned as to produce virtually 1 *(MF* r jms. which. show an appreciable change in length when subjected to a magnetic field. By using a transistor instead of a valve for the maintaining circuit. housed in a long solenoid carrying a. polarizing current is also applied to prevent the magnetic field from reversing. this variation can be limited to about +10 parts per million per degree Celsius. irrespective of the circuit conditions. and connecting these in the anode and grid circuit of a valve. With such forks the influence of the maintaining circuitry becomes appreciable. A small current in the anode coil will set the rod vibrating. Circuit of Valvemaintained Tuning Fork zero overall coefficient. if correctly phased. but with suitable precautions the overall stability can be within a few parts per million. will thus vibrate longitudinally at the frequency of the a. A rod of the material. The rod will vibrate as a half.336 RADIO COMMUNICATION which vary with temperature.c.13.c.wave resonator. complete units can be produced only a few cubic centimetres in volume. Fig. as is now more usual. An oscillator can be constructed by coupling two coils to the rod. By the use of special alloy steels. which is comparable with a quartz crystal. so that the centre point is a node and may be rigidly clamped. An alternative form of electromechanical oscillator makes use of the effect known as magnetostriction. where v is the velocity of sound in the material and I the length of the rod. Frequencies ranging from one to several hundred kHz can . (provided a d. Still greater accuracy is attainable by the use of "slice" forks made up of a sandwich of two materials having positive and negative temperature coefficients. The frequency of vibration is determined almost entirely by the rod. since the magnetostriction effect is independent of the direction of the magnetism). and this will induce voltages in the grid coil. however. such as cobaltsteel or nickel. a. each extending from near the centre to the end.
output RJ: input Fig. the upper limit occurring when the skin the material prevents the magnetic field from penetrating to an adequate depth. This may be done easily enough with a driven transmitter by simply rendering one of the amplifier valves inoperative in the keyup position. With a telegraphy transmitter this is done by signals in terms of the starting and stopping the oscillation at suitable intervals to provide Morse code or other system. Alternatively the frequency may be varied by a small amount.3. Choke Modulation is With amplitude modulation the most usual and convenient method what is known as choke modulation. to vary its character in such a way as to communicate the intelligence required. The stability is comparable with that obtained with a tuning fork.THE RADIO TRANSMITTER be produced by effect in 337 this means. With a selfoscillating transmitter the oscillation itself must be started and stopped. Modulation We have now to consider the methods used to modulate the signal. 7. so providing "marking" and "spacing" waves of different frequency. though it is usually more convenient not to stop the oscillation completely but to reduce the amplitude to a small value in the keyup position. which also supplies the anode of the modulator valve as shown in Fig. The fundamental aspects of modulation were discussed in Section 3.2. We are concerned here with the methods adopted to produce the required modulation. Variations of the grid voltage of the modulator cause Modulottd J. i. Chokemoduxation Circuit . With a telephony transmitter the amplitude or frequency of the oscillation has to be varied in conformity with the audiofrequency signal.e. 7. The anode supply to the modulating amplifier is obtained through a lowfrequency choke. 7.14.14.
and it is essential that the amplitude of the oscillations in the anode circuit shall be strictly proportional to the excitation applied to the grid. therefore. and since these tend to vary with the anode voltage. Moreover. Fig. This is undesirable. The power stages following the modulation cannot be Class C operated because such a system is not linear. A Class B amplifier or oscillator under suitable conditions may fulfil these requirements. The modulated output is then amplified by two neutralized 250W valves. amplifier is and since the anode of the modulating anode of the modulator. the instantaneous power radiated is four times the carrier power. Modulation. A and then a 50W amplifier on which the modulation is provided. is seldom employed on the actual oscillating valve. The oscillator is isolated by passing the signal through a buffer amplifier before reaching the modulation circuit. amplifier must be operated under such conditions that the oscillating current is directly proportional to the anode voltage.f. The modulated signal is then passed through a further amplifier which drives the output valve. 7. This has to be supplied by the modulator.16 shows a simplified circuit of a 50kW transmitter. Hence for the final stages Class B amplification is used. Fig. but it is equally possible that it will not do so and this point must be verified. it follows that an ordinary selfoscillating valve modulated as described will vary in frequency as well as amplitude during the process. to obtain true fidelity of modulation it is necessary for the impedance of the modulating choke to be substantially constant over the full range of modulation frequencies.338 RADIO COMMUNICATION tied to the variations in the anode voltage. Hence. and in practice this is usually more convenient. Since the frequency of the oscillations generated by a valve is dependent to some extent on the constants of the valve.15 and 7. which must therefore be considerably larger than the oscillator valve itself.15 shows the circuit of a mediumpower telephony transmitter. As explained in Section 3.2. 7. A crystal oscillator is followed by a buffer amplifier which feeds the modulator stage. this voltage varies in the same manner and so modulates the amplitude of the highfrequency oscillations as required. where a small percentage variation in the frequency may produce a larger change in the received signal than the whole of the modulation. with 100 per cent modulation the instantaneous maximum value of the oscillations is twice the mean value of the carrier. as shown in Figs 7.16. which then feed crystal oscillator feeds a buffer amplifier . The r. Alternatively several modulator valves may be used in parallel. particularly at high frequencies.
tdoh .
Simplified Circuit . 7.16.Fig.
of A 50kW Tbansmittkb .
17. is the ring modulator shown in Pig. with the result that the carrier voltage does not appear in the secondary. It is.2 to singlesideband This may be achieved by passing the signals from the modulator stage through a bandpass filter which suppresses all frequencies except one sideband and some or all of the carrier. Carson Balanced Modulator Circuit The modified in the signal is then passed through the remaining amplifiers normal way. leaving only the sidebands. so that the grids of the valves are in opposite phase.T. 7. The carrier is impressed across the points P.2 this is not. one of the bestknown being the balanced modulator due to Carson. A pushpull arrangement is employed and the carrier and modulating voltage are introduced separately. 7. The positive half cycles flow through the rectifiers A. a practicable form of transmission). Balanced Modulators Reference has already been made in Section 3. by itself.18. operation. however. A centretapped output transformer is used. whereas the modulation is introduced in pushpull manner. Special forms of modulator have been devised to do this. H. The former is introduced so as to affect the grids of the two valves in the same phase.+ Modulation a Si— %Corr/tr > Fig. Q and flows through the two halves of the transalternative carrier An former windings in opposition.342 RADIO COMMUNICATION finally feed six watercooled two watercooled valves and then valves in parallel pushpull. whereas the modulation combines with the carrier to give sidebands in the ordinary way. B and the negative half cycles through G.17. D. While . 7. Ring Modulator form of modulator which suppresses both the and the modulation frequency. which is illustrated in Pig. sometimes desired to remove the carrier altogether (although as was shown in Section 3.
The modulation is impressed across Tx and current flows alternately through A. These side tones do not cancel out so that the desired sidebands appear in the output while the carrier D is suppressed.b. is carrier. Only one sideband is .THE RADIO TRANSMITTER 343 the rectifiers A.) technique. B are working. in changed over by the SingleSideband Operation Most modern highpower transmitters. Jln^nJLTu Ring Modulator Circuit No carrier voltage appears in the output due to the symmetry of the arrangement. use singlesideband (s. producing side frequencies in T a by combination with the carrier.g. Modulation Input § Sideband Output Carrier. particularly in the H. a switch of which the connections are provided that the carrier voltage exceeds the modulation voltage (as it always does in practice). Modulation Sideband * Output Fio. band. D are operative A and B are cut off. Centretapped resistances may be used in place of T x and Tt . The arrangement. modulation voltages) are impressed across them. B and C. provided these additional voltages do not exceed the carrier voltage. 7. fact. Similarly D when the rectifiers G.18. C and will be subject to a negative voltage and therefore will remain nonconducting even if further voltages (e.s.F.
as indicated diagrammatically in Fig. say.19. 7. together with a small "pilot" carrier (since complete suppression of the carrier is not a practicable arrangement). The process is usually carried out in three stages. These are sufficiently widely separated for the normal aerial tuning to select the upper band and suppress the lower one. resulting in a transmission containing a single sideband with a amount of carrier. At the receiving end the carrier is restored to its proper amplitude either by the use of circuits which accentuate the carrier frequency vestigial . The Speech amplifier 016kHz 100106 kHz 1st balanced modulator 1st filter i 31 to 3106 MHz filter Output 23 to 23106MHz Final 100kHz carrier 2nd balanced modulator 2nd 20MHz skirt of modulator '(reduced by aerial tuning) 3MHz carrier Final carrier 17 16694 to (not MHz 20MHz Operation accepted by aerial) Fio. this filter will have a passband of 100 to 106kHz. This has the advantages of avoiding the waste of power which results when the carrier is transmitted at full strength. 7.19. containing the two sidebands and a vestigial carrier. A small amount of carrier is reintroduced and the combined output is then applied to a second balanced modulator with a carrier frequency of 3 MHz.344 RADIO COMMUNICATION transmitted. If this is. Single Sideband output. is amount of then passed through a bandpass filter which selects one sideband only. 20 MHz the output frequency bands will be 23 to 23106 and 16894 to 17 MHz. Assuming an initial modulation band of 6 kHz. The speech signal is first applied to a balanced modulator with a carrier frequency of 100 kHz. and also reducing the overall bandwidth which increases the signal/noise ratio. This output is then used to modulate a final carrier at a frequency determined by the actual transmission frequency chosen. and choosing the upper sideband. the output from which is then passed through a second filter accepting 31 to 3* 106 MHz.
as is explained in Chapter 16 (cf. This involves the transmission of a succession of pulses of which the amplitude. If the carrier is still being radiated during reception periods it can induce currents in nearby structures.) component to synchronize a Vestigialsideband Transmission modification of the singlesideband system which is more suitable for broadcast transmissions is what is known as vestigial A sideband operation. width or separation is varied to communicate the information required. This can easily be arranged by using the speech signal to control the carrier. the remaining time being taken up by pauses or listening to the reply. Pulse Modulation Reference was made in Chapter 3 to the pulse modulation technique which has certain advantages in commercial or military communication links. This form of transmission is standard practice for television transmissions. or by using the carrier local oscillator. The control circuitry is provided with a slight delay to prolong the carrier for a short period after the cessation of the speech to avoid clipping between syllables or during legitimate pauses. but which is cut off when the speech ceases. It is found that with a normal twoway telephone conversation. which is transmitted at full amplitude as long as speech is present. Hence considerable saving of power results if the carrier is only transmitted during for actual speech. full Quiescent Carrier Systems method of saving power which is more applicable small power transmitters is what is known as suppressed carrier operation. Here the full carrier is transmitted with one sideband but only a small portion of the other. while still providing appreciable reduction in the overall bandwidth. at which point the local oscillator will lock onto the pilot carrier. In the latter case it is only necessary to vary the frequency of the local oscillator around the correct nominal value until intelligible speech is obtained. This is basically a particular . This avoids the necessity to reintroduce or reinforce the carrier at the receiving end. which can reradiate spurious noise into the receiver.9). The system has the further advantage of reducing noise. speech is only transmitted on either channel alternative An to medium and about 13 per cent of the total time. Fig. (See Chapter 12.THE RADIO TRANSMITTER in relation to the 345 modulation frequencies. 16. particularly on board ship.
the tops may be squared off by a limiter. the carrier only radiates full power for a fraction of the time. For this purpose the speech frequencies before modulation are combined with a fairly high frequency. so that the lowest speech frequency is now the highest modulation frequency. Such a modulation is quite unintelligible to the ordinary listener. It is shown there that the bandwidth required is approximately 04/t. If the amplitude of the pulses is appreciably greater than the noise level. usually of the order of 1 //sec. and privacy is thus assured. The requirements for the satisfactory handling of such waveforms are discussed in Chapter 17. the circuitry has to be able to handle much higher frequencies than the normal speech range. The pulse duration is very brief. while to make doubly sure the speech is often inverted. This is known as multiplexing. These frequencies are now heterodyned with another frequency lower than the lowest frequency in the band. and if a sharp edge is to be preserved the carrier must be able to change from zero to full amplitude in something like onetenth of this time. The difficulty has largely been overcome with modern shortwave circuits which employ beam transmission and reception. and the fact that ordinary broadcast telephone signals could be easily picked up militated at first against the use of radio telephony. and we are left with a modulation much the same as the original as regards range. It is translated at the receiving end by going through the reverse processes. any modulating signal has superposed on it a proportion of background noise which interferes with the clarity of the reception. which also describes the methods used to provide variable width or separation of the pulses. as discussed on page 418. thus virtually eliminating the noise. The difference tone is used. The lower sideband only is chosen. A further advantage of pulse modulation is the improvement in signal/noise ratio. . Inversion and Scrambling Privacy is one of the principal requirements of a commercial telephone service. but completely inverted. in the requisite order. producing sidebands in the ordinary course of events. If this is taken as 01 fisec the bandwidth is 4 MHz. In general.346 RADIO COMMUNICATION form of amplitude modulation. Because of the brief duration of the pulses. and permits 8 or 10 channels to be handled simultaneously. and the idle period may be occupied with a succession of suitably delayed additional pulses. all other frequencies being filtered out. but if the pulses are to be clearly defined. where t is the time of rise.
materials are used which will run at a higher temperature. in order that the glass shall not soften.000 to 49. Another method of increasing radiation is to blacken the anode. Valves for Transmitters Transmitting valves are similar to those employed for receiving purposes. would be selected. Since a large bulb has to be employed. this ultimately sets a limit to the size of the valve.900 Hz. since this silica material will run at a higher temperature than glass. the construction being such that if the filament burns out the base of the valve can be removed. In order to increase the peak dissipation at the anode. Still another method is to use carbon for the anode. This would be arranged to modulate an oscillator of. which is done by selecting the required bands with filters and heterodyning them to different degrees so that when recombined they occur in a different order. ranging from 47. say. 50 kHz. In practice both methods are employed either separately or together and speech broken up in this general manner is said to be scrambled.000 Hz giving difference frequencies of 5.000 to 7. Molybdenum is one such material. the ratio of peak emission to mean anode current being of the order of 10 to 1. since a black body radiates heat better than a bright one. and valves of this type are usually made demountable.THE RADIO TRANSMITTER 347 As a typical example we may take a frequency range of 1003. from the anode and to a smaller extent The valves are designed to give a high peak emission. except that they are larger in size and are designed to dissipate heat very readily from the grid. is very expensive. envelope. a new A .000 Hz.000 Hz. but completely inverted. This is often done in two stages. Heat dissipation is by direct radiation from the anode to the glass and thence by ordinary convection. and anodes made with this will run at a cherry red. For example. and subsequently made to heterodyne another frequency to reduce the actual order of frequency. Another method is to transpose certain bands of frequency. Forced draught cooling is sometimes employed. The lower sideband.900 Hz which would give the original scale of frequencies ranging from 100 to 3. the other frequencies being suppressed. the first stage would use an oscillator of 42. and this if necessary could again be heterodyned with an oscillator of 4. The next step was in the direction of silica valves.900 Hz. however.
but these are discussed more fully in Chapter 13. The anode is usually formed with large cooling fins to increase the radiation.C. For medium powers this is sufficient in itself. and this is usually done by connecting the supply to the anode through a long length of rubber hose suitably coiled up. so that a simple changeover of connections is all that is required.. In many instances. Both channels are tuned up and the valves kept alight. These are combined in a special transformer known as a hybrid coil.B. and by throwing over a switch the current may be routed through the alternative stage. This enables continuity of service to be maintained. Certain specialized developments are necessitated by shortwave working. powers the anode itself is immersed in a water jacket through which a supply of cooling water circulates to dissipate the heat. as in the transmitters of the B. 7. The grid and filament assemblies are mounted on a glass foot as usual.. the water circuit has to be suitably insulated. Valves of this type are in daily use handling powers of 25 and 50 kilowatts each. etc. the base refined the whole reexhausted. Terminal Arrangements The transmitting apparatus steel itself is usually housed in cubicles of frame construction. and this is sealed into the anode. is a four.wire system having a pair of leads for the transmitted speech and another pair for the received signals. and The most successful technique is to use cooledanode valves. up and welded to the envelope. including the valves. The radio channel .4. modulation. each stage in the chain is duplicated. for example. Here the anode itself is made the outer container of the valve. so arranged that access to the "live" parts cannot be obtained without disconnecting the supply. on which are duplicated the principal circuit meters and a monitor of the output. but controls of input. This is an important proviso in view of the high voltages involved. so that the control engineer can see the situation at a glance. Land Line Connection A special circuit is necessary to link a radio channel to a land line. are effected from a remote control desk. The main controls are on the transmitter. and considerable care has to be taken to make the operation safe. Since the anode is at a potential of several kV. and banks of such valves handling radiofrequency outputs of 500 kilowatts or more are For still larger employed in modern technique.348 RADIO COMMUNICATION filament fitted.
Illustrating Connection of Radio to Land Line and cannot force its way back owing to the unidirectional characterisof any valve amplifier. but as one of the transformers is reversed the voltages are in opposition and cancel out. These currents also induce voltages in A and B. induces voltages in windings A and G. 7. Speech from the subscriber Fig. tic Singing interaction between the transmitter and receiver will give to a continuous oscillation or sing. .THE RADIO TRANSMITTER 349 which enables both signals to be carried over the same pair of wires as is done with ordinary telephony. The former passes to the transmitting repeater (which is merely an amplifier) and this in turn modulates the transmitter. and also in the balancing network. Currents are induced in the land line. while the other is an allround sing from the local transmitter to the distant receiver and Any rise back through the distant transmitter to the local receiver. 7.20 illustrates the system. Speech picked up on the receiver is amplified by the receiving repeater and appears in windings G and D. where they travel on to the subscriber. one being a half sing when energy is fed back from the local transmitter to the local receiver. because it is in the output of the receiving repeater From Home Subscriber Fig. no trouble should be experienced. which is an artificial line having the same characteristics as the actual line. the received speech is prevented from operating the transmitter. There are two forms encountered in practice.20. Thus. The first difficulty can be overcome by adequate design of the local receiver. The voltage in winding C can produce no effect. Given sufficient wavelength separation and a selective circuit.
This normally carries no current. is transmitter. the receiver is cut off.21. speaking. and hence the field strength continually varying in accordance with the . so therefore. This may be done with voiceoperated relays.c. the receiver takes command. Generally speaking. Volume Compression and Expansion The power radiated by a at the receiver. difficult to circuit. singing is more prevent with highgain receivers. involves careful adjustment of the balancing network which is connected across one side of the hybrid coil to be equivalent to the land line. In the grid return of one of the amplifier valves is a high resistance which is arranged to form the anode load of a second valve operating as an anodebend rectifier. This. If the distant subscriber speaks. suppressors are included in the arranged that. but when speech voltages are applied to its grid a voltage is developed across the anode load which applies a large negative bias to the first valve and so renders it inoperative. 7. Alternatively electrical methods may be used. one such system being shown in Fig. is The speech freapplied to a suitable HT Fig. as one might expect. of course. as soon as the local subscriber commences to speak. quency is rectified and the resulting d.350 RADIO COMMUNICATION sing depends The allround on accuracy of the hybrid balance. 7. Echo Sotfbessor Circuit relay which changes over the connections.21. provided the home subscriber is not Echo Suppressors As a general rule.
A hexode can also be arranged to provide variable lator grid a. the use of two valves being preferable to cancel out any second harmonic distortion which might arise from the nonlinear characteristic. and it is only occasionally that the level rises to full volume.h. The main special requirement in a f. voltage proportional to the instantaneous speech level. which develops a d. 7. and by applying this to the nonlinear amplifier the modulation may be compressed or expanded (according to the direction in which the control voltage is applied) to any desired extent within reason.f. For example.5. and at the receiver the modulation is expanded in the same ratio to restore it to normal. so that the available power is only in use for a small portion of the total time. transmitter is the ability to vary the carrier frequency in accordance with the modulation. gain by applying speech input to the moduand the control voltage to the signal grid. as shown in Fig.c. : 7. An immediate improvement would result if the speech waves could be compressed so that the difference between normal and peak values was reduced. The anode of the reactance valve will then be subjected to the r. two variablemu valves in pushpull would provide a gain proportional to the grid bias. Control voltage is obtained from an amplifier followed by a detector.22. band. Commercial longdistance channels. With .m. of normal speech sounds are of low intensity. Moreover a large proportion. and f. therefore. Transmitters frequencymodulated transmission the procedure is basically the same as with amplitude modulation except for the method of modulation.THE RADIO TRANSMITTER 351 modulation. Both devices are similar but act in opposite directions.m. the average level would be doubled and the power radiated : would increase four times. but the special techniques used at such frequencies apply both to a.f voltage a . which is a portmanteau word denoting a compressor (at the transmitter) and an expander (at the receiver).M.m. There are certain differences due to the fact that frequency modulation is generally only employed in transmissions operating in and above the v. F. If this compression were in the ratio of 2 1 only. utilize a device called a compandor. They comprise an audiofrequency amplifier of which the gain can be controlled by the application of a suitable voltage. transmissions and are discussed in Chapter 13.f. One method of accomplishing this is to connect a reactance valve across the tuned circuit of the oscillator. In practice a compression of about 2 1 suffices.
Such an arrangement clearly provides the modulation required. We can vary g by altering the potential of any of the grids of the reactance valve. 7. whichever is the most convenient. for the change of oscillator frequency is proportional to the input to the control valve (i. .352 ea . the instantaneous amplitude of the speech or music). however. where Z is the effective impedance of the valve. The arrangement will also work if C and changed. RADIO COMMUNICATION This will pass a current through the network GR. but this current will be eJZ. = Output —CWT* RFC + Modulation^ Fig. Reactance valve Modulator that if we vary g we shall change the effective capacitance and so alter the tuning of the oscillator. phasemodulated wave is also frequencymodulated.22. 7.2. and A vice versa. Hence Z l/jcaCRg. as was pointed out in Section 3. namely that the oscillator must be of a type in which the frequency is determined by the tuned circuit values. while the rate at which the frequency changes is determined by the frequency of the modulation. R Phase Modulation The arrangement of Pig. Fortunately this difficulty can be surmounted by using phase modulation which. and if 1/coC is larger than R this current will be very nearly jcoCe a and the voltage at the grid will be jcoCMe a This will produce an anode current jcoGRge a where g is the mutual conductance of the valve. in which case the valve looks like a variable inductance.e. so much . is an equivalent process. . has one serious disadvantage. and by suitable choice of valve and components we can obtain the requisite variation are interof frequency.22. In other words the valve "looks like" a small capacitance CRg. But we have seen that for modern conditions this is inadequate and the main frequencydetermining element has to be a quartz crystal or equivalent device.
. as explained in Section 3. deriving the primary frequency from an independent oscillator. E z will be developed . The result is a phasemodulated wave.23 the modulating signal must be passed through a frequencyinverting network.23 shows one way of doing this. 7. Balanced modulator Armstrong Phask Modulator through a limiter. These are then combined with a fresh carrier derived from the original oscillator but shifted by 90°. .24. E x across the tuned circuit will suffer a 90° phase shift. therefore. who was the "father" of frequency modulation. A rather simpler form of phase modulator is shown in Fig. . The voltage across the capacitor is then inversely proportional to m. Armstrong. 7. The initial carrier frequency is passed through a balanced modulator which removes the carrier and leaves only the sidebands. By making use of this effect. the modulation does not produce any appreciable change of amplitude but only causes the phase of the carrier to swing around its mean position. Fig. due to E. which can be crystalcontrolled if desired. is applied to the grid of a valve having a lowQ tuned circuit in its anode. 7.2 that. The initial carrier. which is then passed Carrier 90°phase shift — Modulation Limiter — Frequency inverter Fig.THE RADIO TRANSMITTER 353 Now. it was also shown in Section 3. if the carrier in an amplitudemodulated wave is 90° out of phase. which is an amplifier so designed that with any input exceeding a certain value the output is constant.23. Hence if a frequencymodulated output is required from the circuit of Fig.m.f. An e. we can modulate in a separate stage. A simple RC series circuit will do this if R !>l/a>(7. which can be supplied from a crystalcontrolled oscillator.2. With phase modulation. the frequency deviation is proportional to the amplitude and frequency of the signal. Voltage will be developed across this circuit by direct coupling through the gridanode capacitance. This removes any small changes of amplitude which may have been introduced during the modulation process. H. whereas with frequency modulation the frequency deviation is independent of frequency. This capacitance in series with the dynamic resistance LjCR will constitute a phaseadvance network so that the voltage. however. 7.
Current pulses of the form of Fig.354 RADIO COMMUNICATION across the tuned circuit. where n is the multiplication ratio.m. Thus to bring a deviation of ±25 Hz to ±75 kHz a multiplication of 3. This difficulty is overcome by introducing.f. a heterodyne stage. — . 7. will have the effect of swinging the phase of the resultant signal across the anode load. say. say.4. The arrangement is thus a Class C amplifier of the type shown in Fig.4 contain appreciable quantities of loworder harmonics. 7.24. 7. would give a final carrier frequency of 600 MHz. For example. at some suitable point in the process.f. It may be that the multiplication required to obtain the full deviation is such that the final carrier frequency is too high. will be in phase with the input and can be controlled by the modulation. A frequency multiplier is simply a tuned r. This disadvantage is overcome by generating the initial oscillation at a relatively low frequency of.f. Simple Phase Modulator Frequency Multiplication A limitation with phase modulation is that the frequency deviation produced is very small of the order of 25 Hz as against 75 kHz usually employed for frequency modulation. 200 kHz and then passing this through frequency multipliers which multiply both the carrier and the deviation.m. amplifier in which the anode circuit is tuned to a frequency which is a harmonic of the input signal. by the normal amplifying action.24 (6) that variation in the amplitude of this inphase e. from 2 to 5 times being usual. Since the amplitude of the harmonic falls off with increasing values of n the process can only be used with relatively small multiplications. This further e. and it will be seen from Fig. and by suitable adjustment of the operating conditions the particular harmonic required can be emphasized. however. the initial carrier of 200 kHz ± 25 Hz could be mixed with an oscillation of. but the anode current pulse only occurs once every n cycles.000 is necessary which. Output E varies between t these limits (<0 (b) Pig. with an initial carrier of 200 kHz. 7.
It will be seen to comprise a modulated oscillator feeding an output stage. as explained in Section 9. will disregard interference produced by atmospheric disturbances and/or circuit noise which causes amplitude variations. reduces the carrier frequency but not the deviation. involving a primary oscillator. The heterodyne process.THE RADIO TRANSMITTER 355 resulting in a final carrier of only 210 MHz. 7.6. which would produced a beat frequency of 70 kHz 25 Hz. ± Preemphasis is that the receiver. which restores the modulation to its original value but with considerably reduced interference. transmitters may be constructed using transistors in preference to valves. deemphasis network is then included in the receiver after the detector. The design of a transistor transmitter is basically the same as for For small powers and portable valve circuitry. the only difference being that the circuitry has to be modified to accommodate the different matching necessitated by the lowimpedance currentoperated behaviour of the transistor. and an output Each of these has to fulfil more specifically in Chapters 9. as a result of which interference does enter the system to an extent which increases with the modulation frequency. but permits a greater overall efficiency by reason of the saving of the power required to heat the valve cathodes. This was outlined in Chapter 6 and is considered amplifier stage. but inverse. similar. 10 and 12. This preemphasis network should not be confused with the "distorter" network in Armstrong modulation. The advantage of frequency modulation is which A 7. as explained in the next section. in fact.25 . This not only provides an economy in size and weight. transmitter. shows a typical circuit for a 4watt mediumfrequency a. Transistor Transmitters units. This is simply a network which increases the amplitude of the modulationfrequency components above about 1 kHz to an increasing extent. All these functions can be performed equally well by equivalent solidstate circuitry.m.4. The primary oscillator produces oscillations in the circuit Lfi1 Fig. To counteract this it is usual to accentuate the upper modulation frequencies at the transmitter by including a preemphasis network. However. designed to respond only to changes in carrier frequency. which is a network giving linear inversion of the output with frequency over the whole range. 130 kHz. a modulation the various requirements already discussed. certain secondary effects occur.
This will provide true modulation since the collector voltage will vary above and below the mean value in accordance with the modulation signal. this may be obtained by including a quartz crystal in the feedback circuit as shown within the dotted rectangle. may . Lowpower Transistor Transmitter this amplifier voltage.14.25. This presents a high impedance to any but the (sharplytuned) resonant frequency so that feedback only occurs at the crystal frequency. for small modulation depths up to about 40 per cent. It is therefore simpler. but it must also have an adequate inductance to deal with the lowest modulation frequency. as in the choke modulation circuit of Fig. The output of 72/ Alternative modulation Fig.356 RADIO COMMUNICATION . modulation amplifier which is tuned to the oscillator frequency. To avoid undue d. 7. Moreover. as was explained on page 338).c. If a high stability is required. 7. . to modulate the signal by varying the base potential. be modulated by varying the effective supply This can be done by introducing a transformer in the collector lead of Tr2 as shown. as described in Chapter 9. since it is "looking into" a low impedance it must be fed from a modulating amplifier capable of delivering adequate power (four times the carrier power for full modulation.f. This is followed by an r. by feedback from collector to base of Trx via the coupling coil L% The operating conditions and matching requirements are determined in accordance with the principles discussed in Chapter 12. voltage drop the secondary of the transformer must be of low resistance.
. by biasing the transistors to cut off as shown in Fig. However. Negative excursions of base current in Tr 4 then produce variations of collector current proportional to the input.THE RADIO TRANSMITTER 357 As explained in Chapter 6 (page 299) the gain of a transistor amplifier be varied by such means. 7. considerable saving of power can be obtained if the supply current can be made proportional to the output. is However. Output Stage The modulated output from Tr2 is then fed to an output stage comprising a driver stage Tr a (described later) feeding two transistors in Class B pushpull. as explained on page 322. so that the mean carrier power is reduced. this transistor will be operative during the quiescent halfcycle of TrA so that energy is supplied to the load over the complete cycle and the effective power output is doubled. as required for faithful modulation. On the positive halfcycle Tr4 is inoperative. The method also does not may provide true modulation since variation of the base potential can only produce a reduction in output. . by providing a second transistor. the method is satisfactory. Since the full power output is only required 3 05 Fio. 7. for small modulation depths. but because the collector circuit is tuned a sinusoidal output will be maintained over the complete cycle. and simpler. Idealized Pushpuix Characteristics with 100 per cent modulation whereas the average modulation depth is only about 30 per cent.26. This can be done.26. but except for small variations the output is subject to increasing distortion. as said. 2V6 of which the base supplied with signals in antiphase as shown.
Chapter 10. and they must be mounted direct on to the chassis to provide an adequate heat sink. with the circuit shown. If /? is taken as 150 the peak base currents will be of the order of 6 mA. and the mean carrier power would be onequarter of this.f. The 1 ohm resistor in the (common) emitter connection is a further precaution against thermal runaway. 7. The inductance of the secondary winding = Z w \Q w <o. as explained on page 338. 100 fiK. a swing of approximately 55 V. The output transistors are of the relatively massive construction illustrated in Fig. Hence the peak drive power will be some 330 mW. Driver Stage It remains to design a suitable input stage to drive the The design of such stages. stage the load is determined by the effective collector This form of Class and is A impedance Zw = Qw La>. This is inconveniently small. the effective radiation resistance of a simple (quarterwave) aerial is of the order of 40 ohms. The effective Q of the aerial circuit will be low because of the radiation resistance. the output transistors can swing over a range of 11 volts (allowing 1 V for the knee of the characteristics) with a peak current of 1 A.000 pF. Briefly. Care must be taken to limit the operating temperature to within the prescribed limits.26. this is the condition for 100 per cent modulation. which will require. which for 500 kHz would be approximately 1. and the stage must be designed to supply a power of this order. This tertiary circuit can have a high Q (in excess of 200) which will have a negligible influence on the overall working Q so that the matching will not be affected. and this must be such that. This provides a peak power output (from the pair) of 11 W. and a figure of 15 would be representative. base current swings of 1//? will be required. as explained in the next chapter. tuned with an appropriate capacitance. is considered in output pair. . in order to drive the output transistors up to a peak current of 1 A. or 375 W. so that to match this to the collectors a stepup ratio of approximately 2:1 will be required (for each half of the transformer).26 ranging from cutoff to the knee of the characteristic. The effective load for each transistor must be 11/1 = 11 ohms. In an r. Now. However. load discussed in detail in Chapter 10 (page 501). with the transistor shown. again. say.358 RADIO COMMUNICATION amplifiers B pushpull operation is widely used in a. 6. Hence a tertiary winding is provided having an inductance of. line is calculated which provides voltage and current swings as shown in Fig. which for a frequency of 500kHz would be 085 ^H.f.
27 shows a frequencymodulated oscillator using a varactor diode to perform the function of the reactance valve of Fig.THE RADIO TRANSMITTER 359 Because the collector circuit is tuned. The applications are not limited to medium frequencies. as was explained in Chapter 6. For this reason. the capacitance of which is varied in accordance with the a.f. oscillation being produced by feedback from collector to emitter through the capacitor C 2 The frequency is controlled by C lt in parallel with which is the variablecapacitance diode D. as is discussed in Chapter 13. this stage could also be operated in Class B. . Fig. modulation. transistor circuits are often used as preamplifiers for highpower transmitters. It is evident The RFC Modulation Fig. been discussed in broad terms to individual circuitry would be based on the principles discussed in Chapters 9 and 10. 7. Transistor Frequency Modulation that since no cathode heating power is required the overall efficiency can be high. approaching 70 per cent. . 7. For example.27. and with appropriate modifications transistor techniques can be used well into the megahertz region.22. The oscillator uses a transistor in commonbase mode. Solidstate techniques in fact can be used for most of the applications formerly performed with valves. 7. but in view of the small peak current the saving would be small. the final output stages using valves. Applications of Transistor Transmitters This typical design has only indicate the general approach.
Hence. it becomes necessary to erect a mast or tower of some kind to support a vertical wire into which the current may be fed. Similarly at the receiving end a suitable structure is required to extract the maximum energy from the waves at that point. since a wire possesses inductance and capacitance of its own. there will obviously be some particular frequency at which the inductance and capacitance of the wire resonate so that the aerial itself constitutes an oscillating circuit.8 Aerials and Feeders Having generated radiofrequency oscillations in a radio transmitter.1. Finally it is often required to locate the transmitter or the receiver at some distance from the aerial so that it is necessary to design feeders which will carry the currents from one point to another with as little loss as possible. To do this it is necessary to cause the current to flow in an extended structure known as an aerial or it is antenna. The passage of the electrons up and down this wire will produce the desired radiation. This occurs when the length of the wire is about 80 per cent of the wavelength of the radiation which would be generated by current of that particular frequency. and it will be clear that the larger we can make this ripple the stronger will be the wave produced. Natural Wavelength Now. These three aspects of aerial systems will be considered separately. and for most transmissions these points are situated vertically apart. then necessary to cause these oscillations to radiate wireless waves as effectively as possible. 360 . is to cause the electric current to flow between two points as far apart as possible. The Transmitting Aerial The concept of a wave as a ripple in a line of force has already been propounded in Chapter 3. Our aim. therefore. 8.
but are more or less uniformly distributed over the whole length. Similarly if the small "elemental" capacitors discharge and then recharge together. must clearly be zero. two halves. Consequently. The voltage will thus gradually increase as we go from the centre to the ends since the voltages on the individual elements will all add together. but will vary from place to place along the wire. so that the radiations would cancel If we want to radiate we must limit the length to a half wavelength as shown in Fig.1. 8. however. It may be considered as being made up of a large centre of capacitor with a corresponding element located at number of small elements symmetrical about the the wire. The current at the end of the X/2 \(approx. Distribution of Cubbent and Voltage in a Wire Oscillating Naturally wire current. the current is in the opposite direction out. the current and voltage will not be uniform.1 (a). Dipoie Aerial aerial is called a dipoie. where we have zero current at each end and a maximum Such an in the middle. because there is no other path for the and since the arrangement is symmetrical the current at the centre will also be zero.) V Voltage (c) Current (b) (a. Full wave Aerial Fig. the distribution being sinusoidal in form as shown in Fig. In such an in the aerial.1 (b). The inductance and capacitance are not concentrated. increasing centre. 8. 8.) Half. in reverse direction the currents will all add from nothing at the ends to a maximum in the .AERIALS AND FEEDERS is 361 An aerial in such a condition said to be operating at its natural wavelength. and we can consider each element as forming a an equal distance from the centre.wave Serial.
.
2 shows the change in amplitude of the voltage and current at successive instants during the oscillation. Even so. Earthed Aerials For long waves where the length of even a half wavelength would be impracticable it is convenient to connect the bottom end of the aerial to ground as shown in Fig. the remainder of the aerial being supplied by the image in the earth itself. This enables the aerial length to be reduced by half. for it brings the point of zero voltage and maximum current actually at the ground. depends very considerably on the wavelength.4 the capacitance of the upper portion to ground is increased and the . 8. The voltage distribution is similar in character but 90° out of phase both in time and space. It will be seen that the strength of the current at any point in the wire passes from a maximum through zero to a maximum in the opposite direction and back again. This clearly gives a convenient arrangement. so that we have a quarter wavelength above the ground. while at any instant the distribution of the current at different points of the wire increases from zero at the end to a maximum in the middle. We then find that the current in the aerial induces currents in the earth. it is only feasible to use an aerial oscillating naturally for wavelengths of a few hundred metres. It must be remembered that this is a space distribution along the wire quite distinct from the change in the value of the current or voltage with time. and it is easily shown that the distribution along the wire is sinusoidal in form.3. if the aerial is provided with a horizontal top. (b) Free aerials. which produce the effect of an image of the aerial in the ground. and we can divide aerials into two main classes (a) Earthed aerials.1 (6) that the circuit is oscillating. suitable for short waves (though sometimes these shortwave aerials also are earthed at one end). indeed. The practical form of the aerial. Fig. Beyond this it is necessary to increase the inductance or capacitance. as shown in Fig. For example. 8. Practical Form of Aerial Dipole aerials such as this are used to some extent where the wavelength is of the order of a few metres but are not very practicable for long waves several hundred metres long. The aerial would be coupled through a coil or by other suitable means to the closed oscillating circuit of the transmitter.: AERIALS AND FEEDERS 363 Thus the voltage and current distributions are of the form shown and (c). 8. 8. due to the fact in Fig. suitable for medium and long wavelengths.
A type of aerial occasionally employed for the shorter medium waves. and a loading inductance included at the base to tune the aerial to the required frequency. ranging from 150 to 250 metres. Illus. as shown in the figure. Such a capacitor is usually called a is flowing in the properties. but it has the disadvantage that the maximum current coil. 8. . then. The capacitance is no longer uniformly distributed. 8. a compromise between the two methods is used for medium and long waves. Fig. nor is the earth point at zero potential. Flattop Aerial wave Aerial portion is nearly uniform (which is an advantage.5. It is then no longer practicable to feed in current at the current node. Alternatively the aerial may be loaded by including an inductance at the earth end. is shown in Fig. and to adjust these so that the total effective length of the aerial is threequarters of a wavelength. again becomes a point of maximum current and zero voltage. It is necessary to insert a capacitance in the aerial circuit as well as an inductance. This again reduces the length of aerial wire required to tune to the required frequency. m "shortening" capacitor. and the length of the aerial in this case becomes greater than the quarter waveCapacitor length.5. usually from one to three times the actual height. The earth. 8.4. which has poor radiating In practice.large effective height. the aerial being provided with a flat top. being almost entirely due to the flat top and the current in the vertical I \W/i///W\ \ I I mnnintin)})) QuarterFig. 8.3. 8. In order to maintain a Fig.5.364 RADIO COMMUNICATION than resonant frequency can be attained with a height considerably less A/4. masts 50100 high may trating use OF "Shortening" have to be employed. as is shown later when discussing effective height). therefore. as illustrated in Fig.
w. Flo 8 6 An . Such arrangements are discussed more fully in Section 8. Certain aerial arrangements have a natural tendency to do this. and if it is comparable in strength with the ground ray at any given point serious fading will obviously be obtained. before the ground ray has really died out.6. the radiation from the horizontal portion interacts with that from the vertical portion and strengthens the radiation in one direction while diminishing 8. as indicated in the w.4 and Chapter 13. figure.mS. The broadcast engineer therefore only considers his station to be effective over the area served by the ground ray. l „ Akbiaj k however. interference begins to creep in from highangle radiation which leaves the transmitter in an upward direction and is sharply reflected by the ionosphere so that it returns to earth in a comparatively short distance (of the order of 150 km with ordinary broadcast wavelengths) This highangle radiation suffers from fading and distortion . With a flattopped aerial. NonFading Aerial A special type of directional radiation is sometimes employed with aerials for broadcasting transmitters to minimize fading. . would thus tend to radiate more strongly upwards and away from the horizontal top.. in which the currents are so disposed as to produce a concentration of the radiation in one direction and a cancellation in the other. but there are it is when desirable to concentrate the radiation more strongly in one direction. both of which will increase the ratio of ground wave to sky wave at a given distance from the transmitter and consequently will increase the service area of the station. in the same manner as the longer range indirect ray.. Special forms of aerial can be devised which minimize this highangle radiation and at the same time accentuate the horizontal radiation./w. it in the other. Unfortunately. which is liable to fading and distortion as explained in Chapter 3. and effective directional Partially Directional characteristics can only be obtained by the use of aerial arrays.m/m/s . such as that shown in Fig. for example. This is only a partial directivity. Beyond the immediate neighbourhood of the station the transmission is accomplished almost entirely by the indirect ray reflected from the ionosphere../mm.AERIALS AND FEEDERS 365 Directional Radiation The radiation from a simple dipole or a quarterwave aerial all is uniform in occasions directions around the axis of the wire. An "L" aerial.
Experiments show that an optimum condition is obtained when the length of the aerial is about 06A.366 RADIO COMMUNICATION Consider a simple vertical wire half a wavelength long. But the base wave will take half a period to travel to the top of the aerial. which is the criterion which concerns the engineer whose object is to produce the maximum horizontal radiation for a given power. Moreover. by which time the current at the top will be flowing in the opposite direction and will generate a wave that will cancel out the wave which arrives from the base.7 shows the improvement in conditions with increasing height. It will be appreciated that this technique does not eliminate fading but reduces its severity by increasing the range over which the ground wave predominates. It will be seen that with a height of 06A there is an increase of nearly 50 per cent in the horizontal radiation. Fig. page 616. I. The vertical axis is the horizontal field strength divided by the square root of the power input." Proc. . 8. entitled "High Quality Radio Broadcasting. It is built in a gradually increasing section up to about 025 A after which the section decreases again. At the same instant a similar wave is being generated by current at the top. The usual procedure is to make the mast itself the aerial.E. On the other hand. Waves leaving from intermediate points on the aerial will not completely cancel out. the ground wave will suffer if the aerial is made much more than half a wavelength long because the radiations in the horizontal plane are in phase and hence the waves Fig. The base of the mast has to be insulated because it is not at earth potential. Consider a wave generated by the current at the base of the aerial and travelling in an upward direction.7. if we make the aerial still more than half a wavelength long the current in the top portion will begin to emit reverse radiation. but it will be clear that the upward radiation will be distinctly limited. Vol. 22. Each portion of the wire will radiate energy according to the current at that point. The ultimate section is often a simple vertical rod on top of the mast proper. Illustrating Relative from the top portion will oppose Increase in Horizontal Radiathose from the bottom portion due tion with Halfwave Aerial to the reversal of the current.R.. 8. For further information refer to an article by Ballantine.
It is. if the aerial is loaded with inductance the current varies nearly uniformly from zero at the top to a maximum at the bottom so that the effective height is approximately 05 times the actual height. but this will not be just the conductor resistance because power is absorbed in radiating the electric waves. because in general it is not practicable to make hjX more than a small fraction. in fact the height of an equivalent aerial in which the current is uniform throughout. to arrive at an expression for the total power radiated. which comes to 16(hr2h*I2/A. so that since the power is I 2 Rr we can say that .AERIMS AND FEEDERS Effective 367 Height in Chapter 3 that the field strength at a distance r aerial is given We saw from an earthed transmitting by the expression ET = 377/A/Ar volts/metre in which / is the aerial current. and Z is the impedance.2 where / is the r. On the other hand. by integrating the values through the whole of the hemisphere surrounding the aerial. A is the wavelength and h is the effective height. Rr = IQOttWIX* = 1.m. Thirdly. This power may be assumed to be dissipated in an imaginary radiation resistance. which is 2[tt or 064. With a longwave aerial having a large horizontal top portion the majority of the capacitance is provided by the flat top and the current in the down lead is nearly uniform.m. this limitation does not apply and practically all the . A is the wavelength.580A 2 /A 2 In a long. Since the aerial is tuned. In such circumstances the effective height may be between 08 and 09 times the actual height. value of the aerial current. where V is the induced e.f.wave aerial this radiation resistance accounts for about 50 per cent of the total. From the expression for the field strength it is possible. in a halfwave or quarterwave aerial the current distribution is sinusoidal. Radiation Resistance The current in a transmitting aerial is given by V\Z. Z is simply the effective resistance R. lying somewhere between 05 and 08 times the actual height. however. because the current in the aerial is not uniform. This is not the actual height. Hence the effective height varies for each individual aerial. At shorter wavelengths. . and h is the effective height of the aerial.s. in which case the effective height may be taken as the actual height multiplied by the average value of the current.
Experience shows that. Earth losses are important from the viewpoint of overall efficiency.3 has an effective height of (2/7r)(^/4). and attempts have been made to reduce them to negligible proportions by using a counterpoise or earth screen. For this purpose the transmitting site is usually located on ground which has a good conductivity. At long and medium waves where h\X (and hence Rr ) is small. so that for any given structure there is an optimum wavelength. The downlead in Fig 8. just as a network of wires in the aerial system has an efficiency nearly as great as would be obtained by a solid sheet. . with consequent increase in For example. so that the ground may act as an efficient mirror as already described. the quarterwave aerial of Fig. and the earth connection is made by burying a series of wires or pipes running from the transmitting building. so an earth screen of this nature constitutes an efficient mirror with less loss than would be obtained from a conventional earth connection. These factors vary with frequency. depends upon the wavelength. and aerial losses generally. 8. It is not necessary to discuss this in greater detail here. This consists of a network of wires a few feet above the earth running under the aerial and for a considerable distance each side. the aerial losses can absorb an appreciable proportion of the available power and it is necessary to take into account not only the conductor loss (including the earth system) but also the dielectric loss caused by the presence of buildings. and vice versa. These may be concentrated immediately under the aerial system or may run in all directions. and substituting this in the above expression gives Br = 40 ohms. though it may be noted that with aerials operating on long and medium wavelengths it is customary to minimize the conductor resistance by using several wires in parallel. trees and other objects in the electric field of the aerial. Aerial Losses The importance of earth resistance. the deciding factor being largely one of cost.19 illustrates this type of construction.368 resistance efficiency. is RADIO COMMUNICATION radiation resistance. spaced round a ring to reduce skin effect. whereas conductor resistance and losses will only amount to a few ohms at most. The Earth System Efficient radiation with an earthed aerial depends to a large extent upon having a good earth connection.
m.m.f.2. arise Field Strength ftV per m Remarks 10. as is often done for commercial reception . but this may be amplified to any desired extent provided that the received signal is appreciably greater than the spurious e.m.f.m. is of the order of millivolts. (But much weaker signals can be handled by parametric amplifiers as discussed in Chapter 19. all of which are inducing e.. Fair signal requiring good tuning circuits and high amplification. will be induced equal to the product of the field strength and the (effective) height of the wire. or may be produced by atmospheric disturbances or locallygenerated interference. Such background may from random e.f. per metre so that the e. as discussed in the next chapter. Will be subject to serious interference. rather than volts. induced is quite small. liable to interference but capable of good service. Very 10 The smallest commercial signal. The table indicates the order of field strength encountered in practice. The actual field strength.) It will be seen that as the field strength falls off the influence of the background noise becomes increasingly greater.f. Its effect may be minimized by timing the aerial and the subsequent amplifying circuits to the frequency of the received signal. directional aerials. We have seen that the transmitting aerial radiates an electric field which travels outwards and gradually decreases in strength as the distance from the transmitter increases. AERIALS 8. Will require full resources of selective amplification.s customarily known as background noise. well audible above local interference. in practice. Good signal requiring a measure of h.m. etc.s generated internally in the receiving equipment. so that if we set up a wire in the path of the wave an e.f.000 (10 1. which also serves to discriminate between the required signal and the many unwanted signals. This electric field is actually a potential gradient. as we saw in Chapter 1.s in the aerial simultaneously.000 (1 mV) mV) 100 good signal. amplification. Still further discrimination may be obtained by giving the receiving aerial directional properties. AND FEEDERS 369 The Receiving Aerial The requirements for a receiving aerial are somewhat simpler than those for the transmitter.f.
developed across the secondary but reduces the effective aerial constants reflected into the secondary and so permits a wider tuning range to be obtained coupling with a given tuning capacitor.f. Two Types of Inductive Aerial Coupling ratio is permissible and it is usual to provide a small stepup which not only increases the e. Since the secondary 2 s cd IR. The equivalent circuit is shown in Fig. but this is a specialized technique which is discussed in more detail in Chapter 13. a few metres in length. particularly as it is often required to receive signals over a wide range of wavelengths. The e. tuned.f. 0. which has therefore been neglected. additional inductance or capacitance. in practice. Aerial Tuning Circuits aerial will contain inductance and capacitance. it is customary to use half. it is more convenient to couple the aerial to a tuned secondary circuit as shown in Fig.wave aerials or dipoles. since the aerial length is only a small fraction of a wavelength.m. therefore. induced will depend upon its effective height closely associated which. a tight Such an it is Fiq. It also depends upon the wavelength of the signal. and the equivalent primary resistance is 2 which is. often arranged in special directive arrays. or both.8. Since the aerial is not tuned. 8. will be required to tune it.8. 8. With long or medium waves it is neither necessary nor convenient to make the length of the aerial comparable with the wavelength. However.370 RADIO COMMUNICATION It is clear. At short waves. but since operating well below its natural wavelength. is X = M . sometimes with a horizontal top portion.m. will be nearly the same as the actual height. however. large compared with the aerial resistance. 8. At long and medium wavelengths it is sufficient to use a vertical wire.9. that the design of the receiving aerial is with the subsequent receiving equipment.
is and vz the voltage across the secondary . given by v2 since <o 2 = Mat ..10. It is loosely coupled to the •Ac 8. is not constant but increases with frequency.1/coCJ a M 2 ft>» + jX^R . Circuit Obtaining Uniform Aerial Stepup secondary and the energy is transferred partly by inductive coupling and partly directly to the top of the secondary through G . Fig. 8. more voltage is induced in L from Lx but . 8.m. The direction of Lx is such that these two couplings are in opposition. At low frequencies the energy transfer is nearly all inductive since G is small and offers a large reactance to the currents. The inductance of L x is higher than usual.e. AERIALS AND FEEDERS r'TSWS 371 L R Fig.ll<oC ea a) La includes the primary inductance of the transformer = Mcoii coil. Hence the effective stepup is M/BC MjC MW/B + j(a)L . .f. This usually gives a stepup of the order of five or ten to one. co X R. the reactance of the aerial wLa l/a)0o output is obtained when the two terms in the denomi2 2 nator are equal i.Fig.10 shows such a circuit.9. and aerial circuits are sometimes arranged to combine both forms of coupling in order to obtain a uniform transfer of energy over a range of frequency. A capacitance coupling would give where X x is = — Maximum — M =X the reverse effect. Equivalent Aerial Circuit Then where *i = MW/B +j{a>La . The secondary e. As the frequency rises. (Lm/B)^ = (MjBC)^ = v2 e \jLO. so that it would tune with the aerial capaci. tance just above the highest wavelength to be for received.
8. however. 8. though with a diode detector some loading is present. if a high input resistance fieldeffect transistor. which with typical values is several hundred ohms. plus the equivalent resistance due to the input resistance of the amplifier (or detector) to which the circuit is connected.11. as discussed in = current. It is Fig. with transistor circuits. of course. With valve circuits this is normally so high as to produce negligible loading. as . in which case the technique is similar to that for a valve. so that the requirement is for maximum secondary possible. a resistance P across a tuned circuit is equivalent (at resonance) to a series resistance of L/CP. i. LjGPn 2 The use of a step down in this manner will reduce the output voltage. so that the transistor must be fed from a tertiary winding (or an equivalent tapping) such that the reflected load Pn? does not reduce the circuit Q too much. Transistor Aerial Circuits must In the foregoing expressions the secondary resistance z include the reflected aerial resistance RJn which is usually negligible.e. The arrangement is illustrated in Fig. but it must be remembered that the transistor is a currentoperated device. Ferrite It is Rod Aerials transistor receivers to use a different type of aerial consisting of a short length of Ferrite rod \ to 1£ cm in more usual with . This is no longer true. the reflected load reduces the Q by onehalf.372 RADIO COMMUNICATION flows through more current secondary G in opposition and so keeps the total voltage fairly constant. This would reduce the Q of the circuit disastrously.11. Now. to use a Chapter 6. R explained on p. but with the bipolar transistor the input resistance is only a few thousand ohms. where the input resistance with bipolar transistors is quite low. Tbansistob Aerial Coupling is essential. and it may easily be shown that optimum results are obtained when R. 453.
The critical coupling is obtained when the effective resistance introduced into the primary by the secondary circuit coupled to it is equal to the initial primary resistance. With critical coupling the secondary resonance curve shows only one peak.f. as mentioned above. can be tuned over the frequency band required. and the coupling must be kept below a certain limiting value for satisfactory operation.m.12.4. Receiving circuits are sometimes operated in this condition.12. and e. while that of the primary shows two peaks very close together. with the appropriate variable capacitor. This coil is designed to have an Ferrite rodj Fig. 8. the (stepdown) ratio being chosen to provide an effective Q equal to half the unloaded Q. and so behave similarly to the frame aerials discussed in Section 8. The output is taken from a secondary winding as shown in Fig. Feebtte Abbial inductance which. the peaks being 5 to 10 kHz apart so that the sidebands are received = . When this is done the system becomes a species of coupled circuit. As the coupling is increased the primary peaks move farther apart. 8. which occurs when Mot \Z(R1 Ra ) (see Chapter 2). Bandpass Coupling In order to obtain increased selectivity the aerial and the secondary are sometimes separately tuned. If this is aligned approximately in the direction of the wave. Under these conditions the maximum current is obtained in the secondary and the maximum energy transferred from the primary. Ferrite rod aerials are directive. the magnetic component of the travelling wave is concentrated in the higherpermeability material of the rod. and a double hump also appears in the secondary resonance curve. is induced in a coil wound round the rod. producing maximum signal when aligned with the direction of the wave.AERIALS AND FEEDERS 373 diameter.
The interfering currents at the source generate waves which are radiated and picked up by the receiver. whereas with the ordinary peaked resonance curve the frequencies 5 to 10 kHz off tune are appreciably attenuated. including the effect of the aerial.374 RADIO COMMUNICATION at a strength comparable with the carrier. These take the form of very rapidly damped waves. is that of local interference. Mains radiation. is transmitted in three ways: (a) (b) (c) Direct radiation. and this form of disturbance is most troublesome in industrial . This gives rise to a loss of the upper frequencies. The range of such interference is limited to 10 or 20 metres. Beyond the limit of the Fig. or "top cut.13. irrespective of the frequency to which it is tuned (just as the dropping of a weight near a piano will set all The interference the strings vibrating). small preset "trimmers" are connected in parallel with the main tuning capacitance as shown. Direct conduction. Baud pass Aerial CraotiiT peaks the current falls very sharply. The first is selfexplanatory. particularly in broadcast receiving. 8. 8." which is minimized by bandpass tuning. Much modern electrical apparatus produces electrical disturbances. giving a better selectivity than could be obtained with one circuit alone. and will thus produce interference in any receiver in the vicinity. but to allow for the differing selfcapacitance of the two circuits. particularly with a sharply tuned circuit. A typical circuit is shown in Fig. which contain components covering a very wide range of frequencies. Antiinterference Devices One of the principal troubles in modern reception.13. The two circuits are tuned together with a twogang capacitor.
until it earthed.) [ Incoming Mains r (b) Z7pF Earth "SflMF 6 Earth Fig. It usually arises from inadequate smoothing in the h. The second arrangement preferable as the earth terminal is Mains always dead. going out on one main and returning on the other. Both forms are encountered and the interference may be conducted in this way several miles from the source. One remedy is to filter the mains at the point where they enter the building by connecting a capaciis in Fig.14. The interference Incoming Mains  <5 3_ » (a. is to fit suitable suppressors at the actual source of the interference. while of course all these methods only prevent the .15. The radiofrequency oscillations generated in the motor are then short circuited to earth through the bypass capacitors and prevented from travelling along the mains. It is thus mainly a matter of design. AND FEEDERS 375 re frigerators though lifts and even household appliances such as may be annoying. Ninety per cent of the interference is of class (c). It may travel either symmetrically. In some cases it is necessary to insert radiofrequency chokes in the leads as well as the bypass capacitors. to the point where the receiver is located and is then reradiated on to the receiving aerial.t. The most satisfactory remedy. Suppression Circuits for Filtering Mainsborne Interference conducted by the mains. or asymmetrically. supply portion of the set or from bad layout or poor earthing. 8.AERIALS districts.14. going out on both mains together with the earth as a return. Direct conduction is usually lowfrequency in character and is only occasionally encountered in the case of sets operated from the electric mains. 8. 8. can often be prevented from causing interference by connecting two capacitors across the brushes and earthing the centre point as shown in Fig. acting as a radio transmission line. With the first arrangei — FlG 815 J {Arm) Field is is ment the centre point is live. The sparking at the brushes of a motor. of Suppressor Circuit Fitted to Motor course. for example.
The only remedy against this latter trouble is to enclose the whole machine in a shielding box of metal sheet or gauze. 8. The fairly local direct radiation from interfering plant is thus prevented from affecting the greater part of the aerial.376 RADIO COMMUNICATION radiation of the disturbance from the mains and do not affect the direct radiation from the offending machine.f. 8. and. Trams and trolley buses are also troublesome. Shielded Leadin (Fig. It can only be silenced by enclosing the equipment in an earthed screened room and filtering all outgoing mains and telephone wires inside apparatus the screen. Shielded Leadin Cable is very troublesome.16. The subject are one of some complexity and numerous special for remedies available particular circuits.16) Apart from the elimination of the disturbance at the source some measure of relief can be obtained by using a shielded downlead for the aerial. and is this is not always practicable. The aerial is provided with a short horizontal top and the downlead itself is enclosed in metal braiding. although there is some . Electromedical jlnsulator Aerial Mast Earthed covering Fio. and it is frequently the practice to fit special h. for it actually uses radiofrequency currents and is thus a small transmitter of very bad noise. chokes in the trolley arm leads and also in severe cases to connect capacitors from the overhead lines to earth every hundred metres or so.
Let us break the feeder up into small sections and consider the inductance and capacitance concentrated in each section. 8. so that voltage will appear across V similar to the input voltage but slightly delayed. To avoid undue Feeders and Transmission Lines In a simple transmitting system it is possible to lead the aerial straight into the transmitting building.c. and it is obviously better if the aerial systems can be kept well spaced from one another to avoid any possibility of interaction. likewise uniformly distributed along the wires. The capacitor will build up a charge. This. FlO.17. the overall result is a marked improvement. as shown.AERIALS loss in AND FEEDERS 377 the effective signal strength received from the required signal. again with a slight time lag.17. The design of such transformers is discussed in the next section. for some type of feeder is essential in the modern shortwave transmitting station. This current will lag slightly behind the voltage. . therefore. however. attenuation of the signal due to the capacitance between the downlead and the earthed covering it is necessary to include matching transformers at each end.3. There will also be a capacitance between them. voltage. so that the system looks like Fig. Suppose we consider two wires of indefinite length connected at the sending end to a source of a. The wires will each have a small selfinductance which will be uniformly distributed along their lengths. is not always convenient. discuss briefly the technique of the transmission of energy through feeder wires. 8. Consequently. IlXUSTRATINO FEEDER AS A SERIES OF SECTIONS is in much the same way as power carried at low frequencies for industrial purposes. We will. The development of this feeder technique received a considerable impetus following the introduction of short waves. because of the inductance in the wires. and to situate the aerial tuning inductance close to the transmitter. The voltage at the sending end will cause a current to flow into the first capacitor. it is necessary to carry the energy from the transmitting building to the aerial through a transmission line. 8. One may have several transmitters to be housed in the same building.
Loss of current due to leakage and capacitance. There is clearly one particular value of resistance which accepts energy just as fast as it is being supplied by the feeder. The wave will. 2. This will feed current into G s and so the voltage will be transmitted along the feeder until we reach the far end. and with this Energy is passed unhindered critical value no reflection occurs. in turn. and the treatment is beyond the scope of the present work. but the result is surprisingly simple. C2 . we have a partial absorption and partial reflection. will feed current into the next section. and the energy from the preceding section of the feeder has nowhere to go except back along the line. . but not enough. The last capacitor is unable to discharge through the load and will therefore force current back along the feeder. The conditions are fulfilled if the resistance is made equal to \Z(L/C). If the resistance is too low the capacitor discharges too fast.* or of a given length of line which comes to the same thing. This involves calculus. Reflection What happens here depends upon the value of the resistance. The discharge of a capacitor through a resistance takes time. and if the resistance will not discharge the capacitor fast enough. where L and C are the inductance and the capacitance of the line. Here the feeder is terminated in an impedance. which it proceeds to do. Thus. If the resistance is high but not infinite. and will commence to charge up. and again we have partial reflection. for here the last capacitor is not able to accept any charge.— 378 RADIO COMMUNICATION This voltage. the excess charge can only go back along the feeder. which we will assume to be a resistance (since we nearly always terminate a feeder in a tuned circuit which will have no reactance at resonance). Similar reflection occurs if we shortcircuit the far end. it will accept some current from the last section. and the larger this capacitance the greater the loss. be reflected. In a transmission line we have to consider two forms of loss Loss of voltage due to resistance and inductance. with very little loss in transit. right along the feeder and is absorbed at the far end. The leakage current across the insulation * 1. and energy will commence to surge back to the transmitting end. Suppose it is infinitely high. The shunt current is dependent on the capacitance. in fact. Characteristic Impedance What is this value of resistance ? It can be deduced by finding the rate at which the capacitors charge and equating this to the rate at which the end section will discharge through the load.
but the characteristic impedance is usually about 70 ohms. therefore. but by tapping down the coil we can make the effective primary impedance what we wish. Those readers who desire a more mathematical treatment should refer to High Frequency Alternating Currents. and if we terminate it with a resistance of equal value. and is of the form Z= At radio frequencies La> (?.AERIALS since the total inductance length. AND FEEDERS 379 and capacitance each depend upon the This is called the characteristic or surge impedance. This is usually very high. The input is feeder. is The capacitance between them io5^j 3 36 loge (d/r) rrr\ ^F r per metre Hence the characteristic impedance <y/(Z/C) = 276 log10 (d/r) ohms. which is to ensure that the termination is correct. the line resistance. so that the maximum efficiency is obtained. The principle is the same. which is the reciprocal of the insulation resistance. provides A a very flexible we can use for any frequency. will Matching the Line arrangement. and we do this by matching the line to the output by highfrequency transformers. and C. The impedance of a parallel circuit at resonance is L/CR.18 shows a termination using a feeder of this type. by Mcllwain and Brainerd. Hence. the larger the leakance the greater the current. Both the input and the output are tapped across the respective tuned circuits. Fig. or by tapping the feeder across the tuned circuit in the usual way. It is the impedance of the line itself when the far end is open. The inductance of a pair of parallel wires was shown in Chapter 1 to be 02 loge (d/r) fiS. All that is necessary may be expressed in terms of the leakance. the leakance. For shortwave work tubular feeders are often used. one wire running inside the other. The complete expression. and an average value of 600 ohms be found to give results of the right order. involves both R. . Over a wide range of practical values for d/r this gives an impedance of the order of 500 to 700 ohms. irrespective of the length of the line. therefore. 8. which is a similar variation to that of the capacitance. energy will be accepted from the line without reflection. per metre where d is the spacing of the wires and r the radius. y/[{R + j(oL)l{Q + j<oO)\ is is much greater than R and Ceo much greater than so that the expression simplifies to \/(L/G).
Fig. where the anode feed current is only large enough to maintain the losses. which another Fro. The building .18. It should be noted that the feeder only carries a current necessary to transmit the requisite power. 8. while the current oscillating in the tuned "fly way wheel" circuit is much greater. insulator Fig.19 illustrates a typical feeder arrangement.380 RADIO COMMUNICATION is matched to the feeder because. The receiving end is matched to avoid reflection. whereas the oscillating current in the circuits at either end is much greater than this. Downlead at a Typical Highpoweb Broadcasting Station of saying that the maximum energy transfer is obtained at this point as well. due to the usual resonant action. energy transfer into the feeder when obtained when the Hence we obtain maximum it is suitably matched at the is transmitting end. 8.19. as so often the case in is communiinternal cation work. the maximum efficiency and external impedances are equal. Feedeb House with Aekiai. 8.V. The process is the same as in the valve oscillator or amplifier. Matching the Input and Output by Auto tbansfobmers Cage downlead from Aerial Parallel wire Aerial tuning feeder house Transmitter building i 77/777rm/ H.
But this reflected wave will interfere with the forward wave so that the voltage at any intermediate point will be the (vector) sum of the e.m. the voltage will vary in accordance with the impressed e. every part of the line in turn goes through the complete cycle of the e. there will be a progressive delay along the length of the line. as explained on page 389).f. while in between they reinforce one another. If we apply an alternating e.wave input) but. This wave travels with a finite velocity very nearly equal to that of light (though if there is appreciable loss in the line the velocity is reduced. to the input of a feeder. Tuned Feeders.AERIALS AND FEEDERS 381 in the foreground houses the aerial tuning inductance. The feeder is split into two. applied to the input.m. At the distant end. The wave reaching the far end will be reflected and will thus travel back along the line towards the sending end. so producing what is called a standing wave.f. and in this case special junction arrangements have to be made. In shortwave transmitters. Consider a feeder open at the far end so that reflection occurs. Standing Waves Let us see what happens if the feeder is not correctly terminated. . but since. at the input end. The system is. Because of this finite velocity of travel the voltage (and current) are not the same at all points in the line at any given moment. At any instant the voltage and current will be distributed sinusoidally along the wire (assuming sine. but will go through the cycle a fraction of a second later. and each half into two again. When this happens we find that there are places where the two waves cancel each other out. since the wave is travelling.f. from the main building only Travelling It is Waves important to have a clear understanding of the voltage and current distribution in a transmission line. having been transmitted along the line by a travelling wave. At each point a matching transformer is arranged so that the impedance of the line viewed from either direction always appears to be the characteristic impedance ^(L/O). as we have seen.s due to the forward wave and the reflected wave.m. which has to handle a current of several hundred amperes. it is often necessary to supply a number of aerials from the same feeder. and so on. this voltage will be transmitted along the line by the process just described. there is a slight time lag in the transmission.m.f. while the feeder carries ten or twenty amperes and is thus quite light in construction. therefore.
A. but the amplitude of which is varying sinusoidally along A/./A/ = 2E cos 2ttx\1 77/2. 2ttxjX an even multiple of 7r/2. but as there are two wires. actually equal to 2x/v. we have alternate maxima and minima of both current and voltage.f. at a.e. We can assess the effect mathematically relatively simply. which started The that e r resultant e. and a maximum when .382 in fact. The reflected wave voltage due to the forward wave is ef will have had to travel a small additional distance. however. 2nx\l.e T = E sin + E sin — dt) . at x due to the reflected wave (neglecting any = = reflection loss) is the same as would result from a forward wave dt).dt)] = 2E cos (cox/v) sin co(t — x/v) This will be seen to represent a wave which is varying sinusoidally with time. The variation is only appreciable. 8. . At radio frequencies the attenuation is very small and the standing waves are of practically constant amplitude. The points of zero voltage are called nodes and are separated by half a wavelength. is thus dt earlier. because the attenuation of the forward wave is less. The actual voltage at each successive maximum point will increase as we approach the sending end. i. the currents in which are equal and opposite at any given section. is is an odd multiple of (equal to 2E) . .f.20 shows the distribution of voltage and current on an opencircuited feeder which is an even multiple of half a wavelength long. But v so that the amplitude may be written = 2E This is cos 2n/a. being proportional to 2E cos cox\v. when .e. while the reflected wave is more attenuated (having had farther to travel). called antinodes. 3A/4. which will take a time dt Hence the e. Thus. midway between the voltage nodes. The current nodes occur at points of maximum voltage. nothing but RADIO COMMUNICATION an elongated oscillating circuit..m. rather like the simple aerial discussed in the previous section. when x = A/2. in lines where the resistance and leakance are comparable with the relevant reactances. zero . . 2x. Let us assume that at some point at a distance x from the receiving end the E sin cot. when is x = A/4. so (at = E sin co(t — ef 4. where v is the wave velocity.m. 3A/2.dt)] cos \[tot — = 2E sin £M + — \dt) = 2E cos \codt sin co(t o>(t co(t co(t . as they may be in telephone lines because of the relatively low frequencies involved. Fig. i. the length of the line. little or no radiation takes place.
As distinct from — Assuming no resistance. as before. it would be supplying no voltage* and maximum current.20. be used to supply an aerial on shortwave systems. As we travel back towards the input end the voltage and current vary in the manner just described. If it were A/4 to the right in Fig. however. the voltage at the far end being greater than the generator If the feeder length is as if sometimes utilized. but since the two wires of the feeder are running parallel. and is sometimes used for measuring the wavelength of a transmitter. 8.21). is only used where the run is short not more than about one wavelength. this effect is and Tuned feeders may also aerial {see Fig. by merely extending the end of the feeder to form the voltage. By finding the current nodes and measuring the distance between them. Current' ^Voltage Standing Waves on a Tuned Feeder Fig. 8. Voltampere Requirements is not an exact multiple of A/2 the situation the generator were located at some intermediate point along the line.AERIALS AND FEEDERS 383 Such a device is known as a tuned feeder or Lecher wire. an accurate estimation of the wavelength can be arrived at. The voltage and current supplied by the generator are then determined by its position. and at any other point it would supply something less than the maximum. and all that we have done is to increase the length of the aerial very considerably. the feeder does not radiate as explained above. The nodes are usually located by means of a shorting bar carrying a flashlamp or other indicating device which is moved along the feeder.20. giving nodes of either current or voltage every quarter wavelength until the generator is reached. A feeder can thus provide a stepup action like an ordinary tuned circuit. Actually some voltage would be required to make up for losses. This form of tuned feeder. all the actual radiation coming from the aerial itself. The voltage at the far end is still a maximum and the current at this point is clearly zero. 8. The current and voltage relations in the system can now continue without any hindrance. maximum brilliance indicating a current antinode. * .
8.f. The actual impedance (neglecting resistance) can be shown to be R B cot ojVLCx for a feeder with the far end open. and a. when may be noted also that Z = x = A A/8. y/(LG) — 1/v where v is the wave velocity. As shown later. . We can then rewrite the above expressions in the form Z =R =R =R tan 2ttx/X = B cot 2we/A cot 2irfxfc tan 2irfa\c when Zr = when Zr = oo Thus Z will be seen to vary from zero to infinity according to the familiar tangent law. oscillating current. the velocity of light. Tuned Feeder Supplying a Hamwave Aerial Input Impedance of Tuned Feeder The input impedance of a tuned feeder clearly varies between zero at a voltage node and infinity at a current node.22 where an untuned transmission line is being used to feed a half.21. transmission line is substantially equal to c. One example is shown at (a) in Fig. and this Fio.wave aerial. R being the characteristic impedance = ^(L/C). the pattern repeating every wavelength. which in an r.384 RADIO COMMUNICATION line. the distance from the commencement of the feeder. It Z . the arrangement depending upon conditions. and tan (oVLGx when the far end is shortcircuited. The subject is further discussed in Chapter 13. the characteristic impedance. Matching with a Feeder feeder may thus be used to match two different impedances. the untuned transmission full the resonant feeder has to cany the so that the losses are greater restricts its usefulness to short runs. 8.
where Zx and Z% are the impedances of the aerial and transmission line respectively. other value of is Zr the ratio of the reflected e.22.f to the forward e. no reflection occurs if ZT = Z and it can be shown that for any . but this is only correct in the two extreme cases of open. Matching Feeders more likely in practice. As we have seen. b \bT A Quarterwave r Variable shorting bar X/4 matching feeder J_L A/4 fl line (<0 ^ ^ Transmission Transmission line 0) Fig. 8.or shortcircuit termination.m. Another method is to use a quarterwave feeder in series with the aerial as shown at (b). given by p=[Z . With intermediate values of ZT some of the energy may be accepted by the termination so that the reflection is only partial. and the transmission line is tapped across at the point C at which the impedance.Z )/(Z + Zr T ) Z is the characteristic impedance of the line and Zr the terminal impedance. as determined by the expression above (or. The same device may be used to match any two impedances provided they are not too widely different. (It should be noted where . Reflection Coefficient So far we have assumed that any voltage at the far end of a mismatched line is completely reflected.m. The impedance of the feeder thus rises from zero at A to infinity at B. This ratio is called the reflection coefficient. Under these conditions. by trial and error) is equal to the characteristic impedance of the transmission line. it can be shown that no reflection occurs if the characteristic impedance of the matching line Zm = y/{Zx Z^.AERIAIS AND FEEDERS 385 Here the aerial is fed with voltage through a short length of tuned feeder made equal to an odd multiple of A/4 and having its bottom end shortcircuited.f.
At radio frequencies. Losses in Feeders = (1 + P )l(l . however. but when p is less than unity the variation is obviously not so great and the performance of a feeder is sometimes assessed in terms of its standing. This may be expressed in terms of the reflection coefficient as V. though small.P = ZJZ ) r We have neglected resistance throughout. in a twowire feeder. effect. . the expressions deduced are not appreciably affected with normal lines. the voltage standingwave ratio (V. being subject to the full variation of potential.S. one line is earthy.) or.S. more precisely. for example.wave ratio (S. Standingwave Ratio When the reflection coefficient is unity the standing waves vary in amplitude from zero to 2E.R. by earthing the midpoint of the transformer winding to which the feeder is connected or by some similar form of symmetrical connection. In a large commercial station the matter receives some consideration and the feeders are designed to give the best efficiency. This can be arranged. In practice. as in Fig. may induce voltages in nearby wires and feeders giving rise to crosstalk.).20. accounts for some loss. which is the ratio of maximum to minimum amplitudes of the standing waves. the other.) and must be expressed as such.R. radiation loss decreases as the spacing reduced.386 that i.wire is of the wires = to about 75 ohms.e.W. and a spacing which makes Z With concentric feeders radiation loss is negligible and the problem is one of conductor loss. feeder. RADIO COMMUNICATION in the form Z and Zr are vector quantities ZT = Br + jXr .R. The presence of losses will cause the current to be less than the theoretical value. This is the order of spacing used in coaxial television cable and gives a value of Z equal With an open. which is least if the radii of the external tube and central conductor have a ratio of 36:1. 8. but even this is only serious on long tuned feeders for the reasons already stated. resistance and other losses are inevitably present while radiation. Balanced Feeders If.W. To avoid this a balanced arrangement is often used such that as the potential on one feeder increases that on the other decreases by an equal amount. but resistance loss increases due to proximity 600 ohms is about the best.W.
however. L. is the voltage at the input or sending end and y a factor im known We (see as the propagation coefficient. in the form v v. The shunt admittance is Y = G + jcoC \/(ZY). lines in qualita quantitative analysis involves mathematical treatment beyond the scope of this book. V(2 Y) = VKR + jcoLHG + jwG)] + If we expand this and collect the resistive and reactive terms together we may write it in the form a jfi. A By considering the line as composed of a number of small elements each receiving energy from its predecessor and passing it on to its successor. and C represent quantities per unit length. a V ei"°. .— AERIALS AND FEEDERS 387 Transmission Equations We have discussed the properties of transmission tive terms. from its initial Assuming a reflection. we can derive an expression for the voltage at any distance x from the sending end. G. where a and /? are somewhat complex terms involving the four quantities R. but it is desirable to state briefly the main expressions dealing with wave propagation. we correct termination.er* = Fse<«+#>" . the voltage at a distance = V. = Note that Now. so that there is no the second term vanishes. called the forward wave. can evaluate y in terms of the line constants. It is then found that y since we are dealing with distributed constants the symbols B. andC Hence we can write the expression for x from the sending end. L. The complete expression is of the form v = F. G. . footnote to page 378). so that for most practical cases can write v = Vfiw is where V. is The series pedance Z =R f jcoL. which has a value V at the (2) VT e r receiving end but decreases as it travels back towards the sending end. and hence increases with the distance x.e''* + is Vrey* the Thus the voltage at a distance x (1) sum of two voltages which decreases exponentially value Vs and vx called the reflected voave.
and equal to the voltage E. A. becomes a multiple of 2ir. the evaluation of a and is found that. This changing phase may or may not be of importance. even in a properly terminated line which does not exhibit standing waves of voltage or current.RG ~ o>*LC)] ( phase distance x shift varies with 1 is independent of frequency but /? is not. The actual phase shift at a tan fix.* It « = %By/(CIL) + GV(L/C)] = i[RIZ + GZ ] The expression for /} is clumsy. In shortwave feeders where various aerials have to be supplied with current in the correct phase. provided Lm ^> B.e *" (cos fix — j sin (cos fix /fa. Phase Displacement Hence. accompanied by a progressive phase by the second term involving /?. of exponential form. this means that the numerical value always the same. Otherwise it is immaterial. In other words there will be standing waves of phase. As already stated. but the phase will be continually changing.) is the phasechange coefficient. determined by the term e **. the length of the feeder has to be exactly chosen. in the form P Clearly a is = 2 v/[« . is Since cos A of the voltage — j sin A = = may * Because of the form \/{p + jq) quite V signeasily. If we consider a feeder having no losses. . there is a progressive phase displacement along the line. /J is cumbersome. so that the the frequency.tfx can be written — j sin fix) so that v = F. of wavelength X 1ir\fi. the voltage at a distance x becomes simply Ex = where /? .B(cos fix — j sin /fa. e.) Thus the wave is subject to a gradual decay. but it may conveniently be written in terms of a. Hence displacement determined the factor a is called the attenuation coefficient. while /? is known as the phasechange coefficient.388 RADIO COMMUNICATION Now. The expression \/(ZY) can be written in the But a is not equal to p nor /? to q as the reader quickly verify for himself. making a complete rotation through 2tt every time /fa.
Then as we progress back along the line the comVr and Vf Vr bined voltage will vary rhythmically between Vf At intermediate points the voltage will be the vector sum depending + . upon the phase displacement between Vf and Vr We must not assume that the two vectors are necessarily in phase at the receiving end. — . the wavelength A = 2irj(i. to a progressive phase lag of f} radians/metre. hence a. they will only be so provided that both Z and Zx are purely resistive. for instance. In fact. so that v = 2irflfi o/fi. which is the 378) are such that l/\/{LC) velocity of electric waves in free space. and O are often where the frequencies involved are much lower. are displaced by an angle 6 depending on the (vector) relationship between Zx and Za The effect of this is that the standing waves . if the line is not terminated correctly the term Freyx does not disappear and at any point the voltage is the sum of the two terms. This is true of lossfree lines in general. and the velocity of = R transmission Effect of is appreciably less than c.AERtMS AND FEEDERS Velocity of Propagation 389 Since the phase changes by one complete cycle every time fix becomes a multiple of 2tt. Mismatching stated above. If R and are sufficiently small to be negligible by comparison with u>L and wO. but only as long as the above assumptions are justified: in telephone lines. so that the line may be considered as lossfree. becomes zero and /? reduces simply to co\/(JJC) = = . It is important to realize that this is the vector sum because the forward voltage Vf and the reflected voltage Vr are not As have seen that the forward wave is subject necessarily in phase. not negligible by comparison with coL and mC. can represent the effect by two vectors Vf and VT rotating at We We the same speed but in opposite directions (remembering that the rotation here represents travel along the line). and the relative effect is not altered if we consider the forwardwave vector as stationary with the reflectedwave vector rotating backwards at twice the speed. then Vf and V. Hence for a lossfree parallelwire line v c. v = l/y^XiC) metres/sec = For a simple parallelwire feeder the values of L and C (see page 3 X 10 8 metres/sec. The velocity of propagation along the line can thus be deduced from the fundamental relation v A X /. If the load is reactive. the reflected wave is subject to a similar phase lag travelling in the reverse direction. .
23. and thence produce a combined diagram as . and the relationships are somewhat involved. 8.r 2L Inductive ZL Capocltlvt 2. the terminal conditions are (a) (6) VL IL =Vf + V and — If — I (the reflected r.390 RADIO COMMUNICATION do not start from the termination but build up as if they had started from a point beyond the end of the line.23. If V L and I L are the voltage and current in the load. 8. so that the first node is not zl/4 back along the line. An Z — oe (open circuit) ***(»4> . being in opposition to Vr ). The effect is illustrated in Fig. = .24. The actual voltage developed at the termination and the phase of the standing waves depend upon both the magnitude and the phase angle of the load. Now. r current /. 8.0 Fig. Distribution of Standing Waves on Mismatched Feeder appreciation of the conditions can be obtained from the vector diagrams of Fig. Vf IfZ Hence we can use the same vector to represent both these voltages.
a capacitive load. the line and a rotation through 2n represents a complete cycle. (a) Vector Diagrams for Different Loads Z L inductive > Z. Vr as rotating relative to Vt we shall Vt — VT and Vf f. This graphical construction* provides a rapid and convenient excellent discussion of this vector aspect of reflection is given in Wireless Communication by A. W. It .Vr already menwe are moving back along The rotation is clockwise because Fig. (London.24. Vector Diagram of Mismatched Termination in the standing and we can deduce from this the position of the nodes waves. which displays the terminal conditions very shows the relative magnitude and phase of VL Vf and Vr and also the phase angle between VL and IL The angle between Vf and Vr depends on the terminal conditions clearly. Fig. occurs in a line length of A/2. R. 8.25 shows two other possible conditions for ZL Z and for . 39* shown in Fig.24. > . Ladner and C. which as we have seen. 1950). 8.24 shows the conditions for an inductive load less than Z Fig. Fig. Stoner. Chapman & Hall. (6) Z L capacitive .25. 5th ed. 8. * An Short Wave . If we regard the vector obtain the fluctuation between tioned. 8.AER/ALS AND FEEDERS . 8.
we know Vf Vr and (f. ). the voltage at a distance I from the termination Hence when is depends upon the angle 6 between Vf and Vr known. = nn and a minimum when = 0. the phase angle of the load is where a = arc tan V r =a+y sin 0/(F. Vi = Vt —V r (cos f + j sin ip) where This y> = {^irl\X + 0). xp is a maximum when 8.Z L ) whence the reflection coefficient p = Vr \Vf = (Z . Conversely.BL f + XL *\ The expression for the phase angle between Vf and Vr some but it can conveniently be expressed as is cumber =a+ where a /5 and /S = arc sin XL/(Z + ZL = arc sin XL /(Z — ZL if . As stated earlier. When Z is resistive may be written a (as it usually is in radiofrequency lines) this _ \/[(Z. ).392 RADIO COMMUNICATION solution to terminal problems for which the corresponding mathematical expressions are often complex. 0.ZL)I(Z + ZL) as previously stated. xp Directional Aerials The simple aerials so far considered are not appreciably directional but receive signals equally well from any direction. It is often required. = \I(Z .4. + V cos ^)> r f and y — arc tan Vr sin 0/( V — Vr cos 6). which can be achieved by a combination of aerials so spaced as to produce an appreciable time lag between the voltage induced by . to use systems which are directional in character. however. From the geometry of the figure we can write V. . the voltage at a distance I from the receiving end is given by .
f. due to the fact that the wave reaches one side of the loop before the other.AERIALS AND FEEDERS 393 the waves at the two ends of the system. Consider the effect of an electromagnetic wave travelling from (see Fig.s will be slightly out of phase. will co jfe" oppose that induced in AB. _t e.f. While r> ABCD AB A travelling from to occurs because the D no effect wave does not move across either of the wires AD B BC reaching DC.m.f. depends on the phase difference.m. 8. induced. therein.m.26. There will be a small resultant e. the e.f. was set up a very short time before that in DC.f. will again be induced in the same direction as before. therefore. It will right to left across the loop of wire first and will reach the side a ^ao induce an e. and hence the two e.27 at an angle The time dt which the wave will take to pass across the loop is (d/c) cos 6 where d is the distance direction of the incoming wave. By suitable phasing of the currents set up in the aerial this time lag can be caused to accept a signal coming from one direction and to reject one coming from the opposite direction.m.s as before will The no phase difference resultant. The e. and they will be in between them in this cancel out completely without leaving any e. . and takes the position shown at (b). while a second zero position will be obtained when the loop is again at right angles. induced will gradually vary from a maximum (when the loop is pointing directly towards the transmitter) to zero (when the loop is at right angles). however. Since k ab DC. Polar Diagram with the Consider a singleturn loop as in Fig. is opposition. constitutes the opposite side of the loop. as the loop is rotated. and the CD / X? two will tend to cancel out. 8. Loop Aerial r a n i v.f. Suppose now that the loop is turned through an angle of 90°. Consequently.m. this e. therefore. Further rotation will cause the e.f.m..m.26). Actually a they will not quite do so because the Fig.f.f. On c i ' " I I ' ' — > m AB • the passage of the wave. Here the wave will affect the two sides AB and DC simultaneously.m. however.m.m. 8. an e. to increase again to a maximum when the loop is again in line with the transmitter. but there position. due to ' or BC.f.
required.f. 8.m.f. induced in it in exactly the same way.m. If we plot the e.dcosS* Fig.f. this reduces to — e = Eh sin codt ca Ehcodt (because codt is small) = Eh(2irclX){dlc) cos = 2irEh(dlX) cos Thus the loop e.m.27.m. and the total e. 8. is dependent on the dimensions relative to the wavelength and on the cosine of the angle 0. The field is strength of the wave will E sin cot so that the resultant e. will be the sum of the e. ^ > in terms of Q Fig.394 across the loop RADIO COMMUNICATION and be c is the velocity of the wave. however. each of which will have e.f.m. <k\ Direction of wave [+. as shown in a "figure of eight" having two pointing directly towards the transmitter is at right angles to the direction of the is called is we obtain what is it which for a simple loop maxima when the loop and two minima when oncoming wave. the total inductance must obviously be such as to tune conveniently to the wavelength Direction Finding The directional properties of a frame may be employed to obtain the bearing of a given station. Several turns may be wound. t 0.m. Practical Forms of Loop So far only one turn has been considered.s in the several turns.f. Eh[sm. The actual number of turns which may be employed with a given area of frame is limited by the inductance. m(t + dt) — sin cot] where h If is we put the height of the loop. which increases rapidly as the turns are increased.f. The station is tuned in the usual .27. Polar Diagram of Loop a polar diagram. A loop with several turns is called a frame aerial.
the frame leads and connections must be made as symmetrical as possible.f.m. zero. errors from this source. in this position the frame will be at right angles to the direction of the transmitting station. Under these conditions. and to .m. since the e. but quite false.f. When the main frame is at zero there will still be a signal picked up on the auxiliary frame. is very small. and this will be many times greater than the small e. Robinson System Other errors arise due to the width of the frame itself. Effect of Frame Width . It was claimed that it was easier to match two small signals to equality than to determine a zero accurately.f. One of the chief sources of error is the tendency of the frame as a whole to behave like a simple vertical wire aerial. as far as possible. H. instead of a sharp zero. but the method is now mainly of academic interest.AERIALS AND FEEDERS 395 manner. Signals received in this mode are independent of the direction of the transmitting station. as such. giving a small phase difference which will still provide a small e. the turns nearest the direction from which the signal is coming will be affected slightly before the remainder. i.m. near the zero point. and tne frame is then rotated until no signals are heard. and so destroy the zero. In an endeavour to overcome this E. This additional frame is connected in series with the main frame through a switch which permits it to be connected in either direction. but becomes appreciable when the e.m. The effect will thus be to give a minimum signal only. from the main frame is zero.f. The effect is normally small.e. If the frame is slightly rotated beyond the zero the small e. when the frame is at right angles to the transmitter. induced in the frame. As a further precaution the middle point of the frame itself is very often connected direct to earth. even in the zero position. With such a frame. induced in the frame acting normally may cancel out that due to the "width effect" and provide a crisp. Robinson suggested the use of a main frame with an auxiliary frame at right angles. and the true zero is therefore located by the fact that no difference in signal is observed as the auxiliary frame is reversed.f. the residual signal will be the same whichever way the auxiliary frame is connected. The sharpness of the bearing depends on obtaining a good zero.m. due to width effect. Crossedloop System size of Another method of minimizing the difficulty is to increase the the frame so that it becomes a single turn or loop. To avoid.
the rotation of the coupling coil producing the same effect as the rotation of the whole system. .396 rotate it electrically. it picks up the Fig. if the receiver is coupled to the second loop. In intermediate positions the coupling coil picks up some e.29) may Signals comparable with those obtained on an ordinary aerial be obtained with this system. The fixed coils are connected one in series with each loop. BelmniTosi Aerial e. from each loop and intermediate directional effects are obtained. Hence rotating the coupling coil through 90° is equivalent to rotating the aerial system through the same angle. The coupling between the aerials and the secondary is usually made tight so that the tuning of the secondary also tunes the aerials. It is also capable of being erected very symmetrically. The large (triangular) loops are erected at right angles to each loops are coupled to the receiver by a special arrange ment. The arrangement is known as a angles. i. having two fixed coils mounted at right angles to each other and a single rotating coil inside. and none from the other.m. receives from a direction at right angles. 8. (Fig.f. which is at right The system. When the coupling coil is in the same plane as one of the fixed coils. the system has the directional properties of this loop. radiogoniometer. therefore.f. Similarly. 8. from that loop. so that the zeros obtained are quite sharp.28.m. 8. is RADIO COMMUNICATION This is known as the crossedloop or Bellini Tosi system and illustrated in Fig. has the directional properties of the loop to which the coil is coupled. Two other.e. and the moving coil is connected to the receiver.28.
since the signal is induced. often as much as 180° in a few seconds. As with width effect. This effect was shown by Eckersley to be due to polarization of the waves reflected known Fio. . These reflected waves arrive twisted so that the electric fields Under such waves are coming downwards at an angle. however. At sunrise or sunset. After dark. Eckersley pointed out that the errors produced by this effect were worst when the frame was at right angles to the wave.AERIALS AND FEEDERS D. but did not affect the direction of maximum reception. the apparent bearing may vary considerably. 8." It is found that by day the bearings obtained on given fixed stations are reasonably reliable (within 2°). this may be cancelled out by the true signal in some other and quite false position of the frame giving wildly inaccurate bearings. work 397 Night Effect A phenomenon which considerably affects is what is as "night effect. conditions become steadier but are still unreliable. the top of the frame is affected before the bottom and a strong arrive partly or wholly horizontal instead of vertical. conditions. which is more or less permanently electrified as was shown in Chapter 3. Radioooniometeb from the upper atmosphere.29.F.
although a frame has two maximum positions.f. while if the frame is turned through 180° the two will help and a maximum e.f. Illustrating Heartshaped Balance If the aerial current can be position and will oppose it in the other. in between the maximum and zero. as unless they are.30. Fig. Now. If the frame and aerial are thus coupled in opposition to a second circuit. will made equal result. the result will be that the effects of the aerial and frame cancel out in one position.398 RADIO COMMUNICATION Heartshaped Balance position. The resistance is for the purpose of reducing the aerial current until it is equal to that of the frame (while it also serves to ensure that the currents in the aerial and frame are in phase. Intermediate positions will give values of e. 8. Both and an aerial are employed in this scheme. . while that of an aerial is a circle (since it receives equally well in all directions). in This led to the suggestion of a method which the frame is used in its known as the heartshaped maximum Direction of Transmitter ^Combination of frame and aeria/. the polar diagram for such an arrangement being in Fig. to the frame current. is a figure of eight. 8. no balance can be obtained) The secondary circuit is loosely coupled to a third tuned . the current in the frame must clearly flow in the opposite direction in the second position.m. the tuning of this circuit also tunes the frame and aerial owing to the tight coupling.30. the frame will help the aerial in one balance. The polar diaa frame gram of a frame. The zero will be seen to occur when the frame is pointing towards the transmitting shown station so that night effect Fig. shows one method of producing this balance. which explains the name given to the method. it will be remembered. The frame and aerial are untuned but are tightly coupled (in opposite directions) to the secondary circuit.m. It will be seen to be a heartshaped figure. since the side which received the wave last now receives it first. 8.31 is obviated.
since it is only possible to determine the general direction of the signal and not on which side of the receiving station the transmitter lies. 8. If the frame is rotated through 180 degrees a similar zero point will be obtained. as under conditions the minimum. The middle point of the frame is earthed as previously explained to reduce the capacitance effect to earth. is not sharp. With frame is possible. In order to obtain still greater accuracy. a small vertical . however. is still employed to determine the sense of the bearing. Tight Coupling To . quite a small small value and practicable direction finding is possible. Circuit for Producing Heartshaped Balance Sense Finding This arrangement. The loop or frame is enclosed in a metal tube. however. although correct. This. effectively screens the frame and prevents it from acting as an aerial. By this means the vertical component is reduced to quite a suitable precautions. although it does not prevent it from picking up signals as a frame. which is. however.AERIALS AND FEEDERS 399 circuit connected to the receiver. many Screened Loops It is not essential to use large loops and a radiogoniometer.31. is not completely successful. The ordinary frame is ambiguous in this respect. Receiver loose Coupling Fig. however. therefore. split at one point so that it does not form a completely closed loop. and this may be rotated mechanically in the ordinary manner. The heartshaped balance. but the heartshaped balance is definite in that only one zero point is obtained (when the frame is pointing towards the station).
aerial is fitted close to the frame. to obtain a more compact arrangement. provided that the general superstructure around the equipment is alternatively positive fact. mounted at right angles. and the net e. is that suggested by Adcock. The aerial is also employed in conjunction with the frame to obtain the heartshaped balance for determining the sense of the bearing as already described. In essence this system consists of four vertical aerials. 8. . and arranged as shown diagrammatically in Pig. so that any voltage induced in the top one is immediately cancelled out by a corresponding voltage induced in the bottom one. situated 90° apart. and the only method of finding out the extent of this variation is by actual calibration of the apparatus on site. Two such screened frame aerials are sometimes used. being degrees.m. pick up energy from the wireless waves and reradiate this energy. left in the aerial is simply that induced in the vertical portion. Calibration Any directionfinding apparatus set up on board ship or even on shore is usually calibrated by visual observation or some similar system. with a goniometer. but instead of using frame aerials.400 RADIO COMMUNICATION connected to the receiver through a coupling coil which is adjusted to provide a reverse e. and these leads are run close together. however. The error. sufficient to cancel the vertical component from the frame.f.28. so that the wave passing the frame is not the pure wave from the transmitter but is distorted by the addition of these small extraneous radiations. It will be seen to be somewhat similar to the crossedloop arrangement. but. The apparent direction of the wave will.f. varies and negative each 90 from a maximum not altered. accurate. the calibration. 8.m.32. be changed. and will. Two leads are taken from the midpoint of each to the radiogoniometer. is found to remain Adcock System The most satisfactory system of direction finding. only vertical aerials are employed which are not responsive to other than normal waves. once obtained. The error is generally found to be of a "quadrantal" nature. Each aerial is a simple vertical wire. Massive metal objects or wires in the vicinity can. There are also likely to be isolated objects introducing errors of their own. instead of the large triangular loops illustrated in Fig. therefore. in in one direction to a maximum in the other direction rather in the manner of sine waves. in order to determine the effect of local obstacles.
Otherwise the actual intersection is indefinite since a small error in the direction of one or both bearings may produce a large error in the station. SI . A modification of the system which is frequently used for obtaining bearings from aircraft Sigutilizes a cathoderay tube.AERIALS AND FEEDERS 401 The connections to the goniometer from one aerial. If three bearings can be obtained the accuracy is greatly increased.m. even at night.f. X Y Position Finding A single bearing only gives the direction of the transmitting If a ship or aircraft requires its position it is necessary to obtain several bearings from a number of stations. apparent position. The same applies to the two other aerials at right angles. This is essential with an aircraft which can change its position by several miles in less than a minute. in exactly the same manner as in a loop.832 aerial Arrangements for Adcock System of coupling these through a goniometer they are applied to the and plates of a cathoderay tube. stations now utilize this system. is due simply to the spacing of aerial is the aerials.f. ." The stations asked for the bearing should be chosen so that the bearings intersect at an angle between 60 and 120 degrees. This is called a "fix. the intersection of the two lines will indicate the position of the transmitter.F. induced opposed to that from the diametrically opposite so that the effective e. It is possible with this arrangement to receive bearings which are accurate to within plus or minus 1°. and hence the bearing can be read off directly and instantly. are such that the e.m. and by coupling the receiver to the field coils with a goniometer the orientation of the Goniometer/ * whole system may be varied. nals are derived from two Adcock aerials at right angles but instead FlG. If two stations some distance from one another are asked for bearings and these are then plotted on a map. and practically all land D. This produces a line of light on the tube face at an angle depending on the relative strength of the two signals.
It operates by virtue of the fact that the lower portion of the vertical field in a wireless f rT \ \ y >/? Utilized in the is To Receiver V////////////////////SS//////SS//////////?/////// Fig. 8. slightly due to the resistance of the earth. wavelength to produce a reinforcement of the signals from one Such arrays are frequently employed with shortwave reception.402 RADIO COMMUNICATION three lines should all meet at a point. due to inevitable small errors in the bearings. it is often desired to provide a receiving aerial with directional properties. often called a "cocked hat. Signals from the reverse direction produce the same result but on arrival at the point A they are absorbed by the resistance R. Aerial Arrays Apart from direction finding. Meanwhile the advancing wave is inducing further changes in the successive elements of the wire and the effect is cumulative the wire is several wavelengths long a considerable signal at the far end. Showing how Wave Dbag Beverage Aerial wave drags Fig.33. W. Beverage. but they form a small triangle. Actually they do not. and are discussed in detail in Chapter 13. 8. . which is made equal to the characteristic impedance of the line and thus absorbs without reflection. It consists of a long horizontal wire a few feet off the ground pointing in the approximate direction from which the signal is coming. is thus a horizontal component which induces voltage in the The voltage in the first element will travel along the wire to the far end. Signals from any other direction so that builds if up clearly cannot build up to the same extent. This can be achieved by suitable arrays of aerials which are spaced in relation to the direction. as shown in There wire. The Beverage Aerial A different form of directional aerial suitable for fixedstation work was developed by C.33." and the true position is some The where within this triangle.
8. D £>_.Amp. L. Certain difficulties are experienced in applying this system to . fading is produced by irregular reflection in the ionosphere. arising in the latter case from a circularly polarized in wave which the plane of polarization is rotating continuously. for no amount of increased amplification can compensate for a complete fadeout.Amp.Amp.F. but this is not entirely satisfactory. 'wft?. in which case aerial.34.F. H. H. while.F. we should obtain a reasonably uniform signal. V V \J/ Vffl?.34. The arrangement is illustrated in Fig. in addition. that the reception at different localities is not the same at the same instant. the plane of polarization is twisted so that a wave may which arrive horizontally polarized.AERIALS AND FEEDERS 403 Diversity Reception Reference may be made to a particular form of receiving aerial is intended to overcome fading.Amp. In particular. 8. When the signal has faded to vanishing point in one locality it may be quite strong only a short distance away. H. however. Consequently. Fig. W7W.F. if two or more receiving aerials are erected a few wavelengths apart and each one individually tuned to the required signal. the increasing amplification always brings with it an increased background noise. and the output of each mixed subsequent to rectification. it will have no effect on the ordinary vertical receiving it This variation in the plane of polarization may be spasmodic or may be regular. ZZDL Arrangements joe Diversity Reception Attempts may be made to minimize fading by the use of automatic volume control. When one aerial is receiving practically nothing there will be some signal in one at least of the others. As explained on page 131. and in practice such a combination usually provides satisfactory reception. It is found.
f. . squarelaw detectors are used. The aerials are spaced a few wavelengths apart and the signals are brought away by means of r. The system is reasonably satisfactory in practice and is used to a considerable extent. To overcome this. Since the squarelaw detector operates much more effectively on a strong signal than on a weak one. feeders of the type discussed in the previous section. because the modulation is not always in phase on the three receivers. together with an automatic volumecontrol device following the detectors. this arrangement ensures that the greater part of the output obtained from the system comes from the particular aerial which is receiving the best signals at that moment.404 RADIO COMMUNICATION telephone reception. which operates on all three receivers simultaneously.
F. as far as possible. (c) Amplification of the audiofrequency signal to the required extent. and its application to a suitable soundreproducing or recording device. This involves the use of the amplifying and tuning techniques already discussed in basic terms in earlier chapters.. with the selection of the wanted signal in preference to unwanted signals. amplifier — — Output to load aerial Fig. of whatever sort. signal so that the intelligence conalso involves the selection of the required signal tained in the modulation can be extracted. These are (a) Amplification of the r. 9. This function Input — Selective from — amplifier _. in particular. but there are still occasions where valves are used. so that tuned amplifiers are customarily employed. Three separate functions are required.1.f.f. Essentials of a Radio Receiver and the rejection of unwanted signals. It is then shown how these requirements are applied to the two different types of circuitry. 405 . (b) Demodulation of the r. a fundamental approach has been adopted. (  Detector (demodulator) A. 9. The basic principles are the same for both techniques so that. as illustrated in Fig.: The Radio Receiver The radio receiver is concerned with the amplification of the signals picked up on the receiving aerial and. signals induced in the aerial to an extent sufficient to permit efficient demodulation.1. These functions will be considered separately. The development of transistor techniques has rendered the use of valves obsolete in many cases.
Receiver and it is customary to change the frequency of the incoming signal to an intermediate frequency for which a highlyselective fixedtuned amplifier may be used. amplification is necessary. to a second amplifier A 2 the output from which feeds the detector (demodulator) D. For the majority of requirements considerable r.1. This may be carried out at the frequency of the incoming signal. amplifier. These requirements are to some extent conflicting. influences the performance of the tuned circuits. tuned aerial circuit feeding an amplifying device A x which is coupled. It comprises a . This is called a superheterodyne receiver.r. 9. Basically.15.2.2 is a amplifier. 3. in turn. so that a practical compromise has to be chosen.f. However. the need to readjust must be capable of being tuned to the is signal. it employs certain special techniques which are discussed practical difficulties in Section 9.f. the device receives a certain input which is magnified by its internal action and develops at its output an amplified signal Fig. D Twostage R.f.f. and while it is merely a special form of r. this is only suitable for the reception of powerful local signals where considerations of selectivity do not present any problem. because the amplifying device. as in Pig. in which case both the aerial and the interstage couplings Such an arrangement and is convenient for certain requirements. known as a tunedradiofrequency (t. We are concerned here with the amplifier stages which are required to provide maximum amplification of the desired signal with a maximum rejection of unwanted signals. For broadcast reception.406 9.F. where the required signals cover a band of frequencies. diagrammatic representation of a twostage r.) amplifier all the tuning circuits when selecting another station introduces M Fig. RADIO COMMUNICATION Radiofrequency Amplification The simplest form of receiver requires only a tuned aerial circuit feeding a detector. . and the condition for maximum gain is not the same as for maximum selectivity. whether it be valve or transistor. however. 9. particularly with commercial stations (including marine and aircraft services) which operate on a fixed frequency.4. The design of the aerial circuit has already been discussed (page 370).
As shown in Chapters 5 and 6. = + = = Variation of Gain with Frequency If the circuit is tuned with a variable capacitor the impedance (LjCE). and. as in the case of a simple valve oscillator. The e. by a small top capacitance coupling. If this feedback is sufficient in magnitude and in the right direction. 9. ^X ^ . . induced in the secondary is then partly inductive.r J e ° „ Fig. stages are required. The capacitive energy transfer may be achieved in other ways. e. which falls as the frequency increases.f. Mixedcoupled . if the feedback is in the reverse direction the gain of the stage will be reduced.f. i. provided at a fixed frequency by using the superheterodyne principle o< to k. but occasionally § § (y\ ^ S t. mixedcoupled transformer. Hence the stage gain is simply proportional to Z.3. r is the internal resistance and n is the amplification factor. When r is large compared with Z. ? In such cases it is possible to "J* obtain a uniform sain by using a „ „ „ . continuous oscillation will result. the gain may be written in the form A v gZ. In practice. . i. . Conversely.Z/(r Z) where Z is the external load. where g is the forward transconductance dioujde^.g. and can theoretically be made very large. There are two principal methods by which feedback takes place. amplification. Currents are induced from one circuit to the other. as is usual with the valves or transistors used for r. 9. owing to the very high amplification which can be obtained (a voltage gain of 100150 is quite normal in a .3. and by correct proportioning a constant voltage transfer can be arranged. as in Tbansfobmek Fig. which rises with frequency.r. With modern receivers the main amplification is M ~^. described later.f. and partly capacitive. and hence the gain. This limitation is discussed later. '  .THE RADIO RECEIVER 407 of which the form and magnitude depend on the relative external and internal impedances.m. Stability It is important in any amplifying stage to avoid any unauthorized feedback of energy from the anode to the grid circuit. will vary with frequency. the basic stage gain can be written in the form A„ /J. it is limited by internal (and external) feedbacks through which a small proportion of the output is transferred to the input and may cause continuous oscillation. l . One is due to direct magnetic or electrostatic coupling between portions of the circuit or the wiring.
f. The magnetic fields generated by the coils induce eddycurrents in the material of the shielding. but provided the screen is not too close to the coil the loss of energy is small. so that a coil has a lower inductance inside the can than out of it. These eddy. this modern feedback has only to be very slight in order to produce a marked effect. as is more usual with transistor circuitry. currents. Selectivity As said earlier. any electrostatic screening.2 the valve or transistor itself. otherwise coupling will be introduced between the circuits. the field outside the can is negligibly small. by Lenz's law.currents in turn produce magnetic fields of their own which. the design of the interstage couplings has to allow for the requirements of selectivity as well as gain. The magnetic field set up by the shield reduces the effective magnetic field within the coil. and 9. It is. as close as possible to the lowpotential end of the supply.408 RADIO COMMUNICATION receiver). must not be allowed to carry any r. A third source of instability is the internal feedback within This is discussed in Sections 9. These shielding boxes are made of copper or aluminium. One common screen for all the circuits. It is important that each circuit shall have its own screen. still necessary to avoid feedback due to common impedances in the supply lines. the field is suitably confined and there is negligible interaction.3. A separate heavy gauge busbar should be run between the appropriate points. The chassis should not be used as an earth line. including the chassis. The process obviously involves a loss of energy. Consequently. Shielding To minimize this interaction it is customary to enclose the tuning coDs in metal shields or cans. With coils or transformers housed in completely closed magnetic circuits. as circulating currents in the screen may induce voltages from one circuit to another and so defeat the whole object of the screening. since the eddy currents circulating in the material of the can must absorb power. are in opposition to the original field. and operate as follows. even if it is partitioned off. however. this busbar being connected at one point only to the chassis. For the same reason. The selectivity of a circuit is the measure of the frequency discrimination exercised by the tuning. This is discussed more fully on page 416. is unsuitable. It may be expressed in terms .
which involves a lower individual value of Q. the bandwidth is twice this deviation. Hence. so that 28 /„. If the shunt resistance is P. in general. the for the presence of the circuit impedances. in general. if the overall response is to be 3dB down. With two stages this gives 065fr /Q. It was shown on page 111 that. whence Q8 = \. in practice the circuit requirements are determined by considerations not so much of selectivity as of bandwidth. expressions the relative response can be estimated and it will be noted that the attenuation is dependent upon Q (\fR)\/(LIC). so that the required Q = f /B. the information in a modulated wave is entirely contained in a series of side frequencies extending over a small range on either side of the carrier. be in the form of a resistance across the tuned circuit and will thus increase the effective resistance. 05/r /Q. so that if the information is to be properly received. effective series resistance is increased by LjCP. the tuned circuits must maintain an adequate response over the full range of these sidebands. As explained in Chapter 3. if circuit P is the unloaded parallel impedance of the (= LjGR). Hence 8 = 1/2Q. but in practice several employed in cascade. Hence. the higher Bandwidth From these = the L/C ratio the better the selectivity. This is The bandwidth the criterion for a single circuit. This was shown in Chapter 2 (page 95) to be where = 1/(1 + J2Q8) = 1/V(1 + 4<W 8 = (f—f )lfo and Q = L(o jR. if the attenuation is expressed in dB. The ratio 8 above relates to the deviation on either side of the resonant frequency. then for n stages the criterion is that for each individual stage the response at the specified bandwidth must be only 3/w dB down. or "loaded. while with three i? If the circuit is transformercoupled it is necessary to calculate circuits are usually B= = . ///„ Q In this expression Q is the working." value allowing These will. Alternatively. the individual circuits must produce less attenuation. The overall response is then (7// )"> where n is the number of stages.This requires width is not more than 3dB down (I/I B= 1 — = = + 4Q*8* = 2. However. B = 28 /„ =f IQ. It is customarily accepted that adequate 2(/ /„) reception is obtained if the attenuation over the stipulated bandl/\/2 ). then Qw =QP I(P + P) for a given value of 8.THE RADIO RECEIVER 409 of the ratio of the voltage (or current) for a given deviation to that at resonance.
This again must be modified if more than one stage is used. as shown on page 110. chosen to meet the specified bandwidth requirements with the total number of stages involved.f.F.i s2 ) develops across it a voltage equal to the bias required. The cathode resistor is shunted by a small capacitor (of the order of 01 fiF) so that the signalfrequency component of the cathode current is bypassed and the only e. 1/V(1 For a single stage the 3dB bandwidth is then \/2/r/G. 9.t.f. a transformer with both primary and secondary tuned) then with critical coupling and identical circuits I/I 4Q4 <54 ). It will be shown in the sections which follow that the design of an This must be r. so that the grid is biased negative to the required extent. while the anode is fed from the h.t. the criteria for two and three stages being l15/r /<2 and/r/Q respectively. Valve Amplifiers The simplest form of valveoperated r. c2 To R2 Fig.f. The output is passed to the succeeding stage through the isolating capacitor C 2 The valve is biased to its correct operating point by means of the .f.coupled circuits (i. amplifier uses a tunedanode circuit as shown in Fig. which will reduce the stage gain. The input is applied to the grid. line through a resistor R 1 the value being chosen to drop the voltage to the correct rated value again. succeeding sta$e Tunedanode Circuit cathode resistor R 2 which is so chosen that the total cathode current {i a f. line through an inductance tuned with a (variable) capacitor. .410 RADIO COMMUNICATION the equivalent single circuit using the normal coupled. R.circuit laws. developed across R 2 is that due to the steady current. 9.4. (If this is not done the signal currents will produce a negative feedback e. . In the case of two similar loose. The cathode is thus positive to the earth line.m. stage is determined from the Q value required.) The screen is fed from the h.e. . — + 9. as is discussed on page 479.m.2.4.
It will in which case ra is large compared with Z. this will increase the effective tuned circuit resistance by LjCra Hence the true gain is = . where g is the mutual conductance of the valve.F. as shown on page 101. the screen bypassed to earth through a capacitor. as will be seen later. it is customary. In many instances the valve is designed for operation with the screen at full h. potential so that R t and C4 are not required. must not be taken as that for the tuned circuit alone (= LJGR) but must allow for the shunting action of the internal impedance of the valve. . and the effective = QtfuHZ + ra is ) bandwidth Bw =frlQv. ) be more convenient to express the relationships in terms of impedance. If the unloaded impedance L/GR Z then Z a Z rJ(Z r a ). to use pentodes in r. Transformers be seen that the shunting effect of the valve impedance reduces the effective Q of the tuned circuit.5. stages. The value of the load impedance Z. Alternatively it may = = . Such a circuit will give a gain A v gZ. + The effective Q is Q. 9. gLIC(R ) a ).f. Neglecting any secondary effects.t. this is ra and.THE RADIO RECEIVER to eliminate the effect of is 411 any signalfrequency currents. The effect is not normally of major importance because. + L\Cra = gL\{GR + Ljr and the working Q (= Qw = a)Lj(R + L/CVJ. however. in order to maintain adequate stability.=MZo + ra )IQ ra Fio. Simple HiGHrnEQtrENCY Transformer and Equivalent Circuit R.
if ra is comparable with. With such conditions t becomes fractional. The effective primary impedance is then it ra +Mm 2 2 IR and the primary current h The voltage induced across the secondary coil = «. transformer.412 RADIO COMMUNICATION However. and the voltage (L<ojR)e 2 ~ LW Mo R MLco 2fi Rra 2 ra + M a) IR 2 2 ft6' + M u> 2 e„ The first term of this expression represents the gain of the stage. M . so that M = k\/(LLj)./(»» + MWIR) in the secondary is is Mmix . the internal and external impedances are equal. however.f. MWjR In this condition. is less than this because the primary flux does not all In practice. link with the secondary. Hence Optimum where stepup = k\/(LICRra ) = ky/(Zfr a ) Z is the dynamic impedance.5. LjCR. which is a useful criterion to remember. as shown in Fig. maximum The criterion for gain can thus be rewritten Rra since a>* =M 2 co* = WL^w* = WLJC = 1/LG. almost invariably uses pentode valves in which r is anything from 05 megohm upwards. to allow for which we introduce a coupling M factor. . or appreciably smaller than Z. of the secondary circuit. k. Hence Lx = CRr \k a 2 Now. The mutual inductance between two windings perfectly coupled without leakage is \/(LLt ). and by differentiating this it can be shown that the maximum gain 2 2 = Zv This can be written ra = results when a> = Rra In other words. 9. where L and Lx are the inductances. may be better to use an r. With a triode the optimum value of t is usually between 2 and 3. Modern equipment. the stepup ratio t is approximately y/(L\L^ and the optimum value of this is obtained by substituting the value of Lx just obtained. the effective Q of the secondary is reduced to onehalf the unloaded Q.
THE RADIO RECEIVER 413 For example. In a valve circuit this can take place through the internal capacitance between anode and grid of the valve itself. is very symmetrical. and it is overcome by the use of neutralizing circuits in which energy is fed from the output to the input. as shown on page 415. r. due to Rice. representing a stepdown. but where considerations of selectivity are important the use of an r.f. but suffers from the disadvantage positive produce continuous oscillation. this refinement is rarely necessary because typical pentodes have a mutual conductance appreciably greater than unity.f. becomes 036. The first circuit.000 ohms and the circuit has a dynamic impedance of 100.6 shows two satisfactory forms of circuit. through a special circuit. of 28 to 1. this would give a stage gain of 1085. Assuming g = ImA/V (fi = 500). transformer may be desirable. such as would be covered by the normal tuning operation.000 ohms (which is about the average for commercial receivers) the optimum ratio. The magnitude and direction of this internal feedback depends upon the nature of the anode impedance. When the anode circuit is nearly Fig. if r = 500. Neutralizing Mention was made earlier to the possibility of undesirable feedback between output and input of an r. In practice.f. that only half the full voltage is applied to the grid of the valve. so that a tuned anode circuit will develop all the gain that the circuit can handle before instability occurs due to stray couplings. with k = 08.6. Two Forms is of Neutralized Circuit in tune the feedback and may easily be sufficient to In a triode valve this is a very serious difficulty. 9. amplifier. 9. . in opposition to that which passes through the valve. If these circuits are symmetrically designed. the adjustment remains adequate over a wide range of frequency. Fig. as against 835 with a tuned anode circuit.
It is essential that the primary and neutralizing winding shall be very tightly coupled. negligible proportions in the valve itself. .7 shows the equivalent circuit of a valve. theory it can be proved that Input resistance R. By ordinary a.f. This is that valve if feedback is to be avoided the stray capacitance external to the must be reduced to a similarly low value. Fig.* who showed that the effective input impedance depended on the anode load. There is. but connected in opposite phase. however.414 RADIO COMMUNICATION In the second. It must be remembered. Miller Effect The feedback of energy from anode to grid of a valve through the internal capacitance was first analysed by John M. and the modern receiver uses tetrodes. which requires careful screening and attention to layout. 351. Modifications of these circuits are used. or more usually pentodes. however.6 Gg — G + Cga (l + A cos 6) go Bureau of Standards Bulletin. exactly similar to the primary winding. this being a simple network except for the fact that there is the amplified voltage [ie g in the anode circuit. This cancellation only applies for one particular frequency so that the method is of limited application but is used in amplifiers operating at fixed frequencies such as in radio transmitters.c. and the usual practice is to wind one over the other. a third method. Such valves have internal anodegrid capacitances of O01 pF or less. a separate neutralizing winding is used. known as coil neutralization. No. stages. in which an inductance is connected between anode and grid. due to Hazeltine. If the reactance of this inductance is equal to that of the anodegrid capacitance it will clearly provide a feedback in antiphase and so eliminate any tendency to oscillate. 9. " Input capacitance * = A —~y„ and sm. Miller. and is discussed in Chapter 13. Feedback can also be eliminated by modifying the circuit to maintain the grid at earth potential with both anode and cathode at varying signal potentials. usually employing one or other of these two basic forms. in the r. as explained in Chapter 5. Screened Valves The alternative remedy is to reduce the anode—grid capacitance to done by introducing an earthed screening grid between anode and grid. This groundedgrid technique is not convenient at normal frequencies but is occasionally used in shortwave receivers.
Cos 6 remains positive (for 6 never exceeds —tt/2) Cgc . . so that there is no tive. sin 6 and the input resistance is infinite. the damping is reduced and self. and hence an r. At resonance. the input circuit is subjected to additional damping. Equivauent Circuit of Tbiode If the anode load is inductive (at the frequency under consideration) positive and hence the input resistance is negative. The input capacitance is obviously always greater than the "geometric" capacitance of the valve. the damping is increased.. With a capacitive anode load. sin 6 is The shunt = .c. Cgc + and Gg is thus always more than Gga when cos 6 is nearly zero. stage with the circuits dead in tune is not unstable at that frequency (but it is at lower frequencies.f. i.e. Cg would be very high since cos 6 1.f. but the condition is an impracticable one and need not be analysed further. and 6 is the angle by which the voltage on the load leads the a. is that as the anode circuit is tuned the input capacitance rises to a maximum and falls away again. 9. amplifiers. which make the anode load inductive)." resistance also varies. anode voltage jueg Let us examine these equations further.oscillation may occur. = resistance is positive. Cga and Cgc are the gridanode and gridcathode capacitances respectively. THE RADIO RECEIVER where 415 the stage gain. If the anode circuit is capacinegative and the grid resistance is positive. sin 6 is negative and the input 6 is . This effect distorts the tuning of the input circuit and is Thus in an amplifier the effect known as "pulling.7.f. Cos 6 is fairly small and G is g equal to C gc plus two or three times Cga If the anode load is resistive. 6 is zero and the input resistance is infinite. This is the cause of instability in h. is . This condition can only be obtained by a tuned anode circuit. C gc while the input resistance may be positive or negative according to the sign of the denominator A sin 6. A c w =}= Ih 'off C SC i Fig. the limiting value being r. i. If it is positive it means that extra damping is introduced into the circuit.e. If negative.
while below resonance the response is maintained beyond the true resonance point by the negative input resistance." Proc.* but about 10 MHz. and it is sometimes necessary to run some leads in screened cable. The r. "Input resistance of vacuum tubes as ultra high frequency amplifiers.416 effect RADIO COMMUNICATION on the circuit other than the additional capacitance referred to above. Ferris. Since substantial overall gain may be developed it is essential to eliminate accidental feedback from output to input. A filter comprising an r. though Gga is of the order of 001 pF only. lation a second effect which occurs when the time of the oscilcomparable with the time taken by the electrons to move between the electrodes in the valve. as shown in Fig.9. The layout of the components must be such as to avoid stray coupling.f currents are thus * Ferris.8. for example. if sufficient. This results in a serious decrease There is is in the effective input resistance. though it may still be appreciable with pentodes because. continuous selfoscillation. whence it will be transferred to the grid of the second valve. If the anode circuit is inductive the input resistance becomes negative. R. This difficulty is overcome by decoupling the various leads. an intermediate frequency as A more important source of feedback is the presence of common impedances. p. The effect was first described by W.. the anode currents of both valves have to run through the common impedance Z (which is inherent in the circuit). Inst. Decoupling For simple receivers a single stage may suffice. either at the signal frequency or. Rod. but usually several stages are used. R. effects combine to produce a distortion of the tuning frequencies above resonance the additional damping reduces the response. 9. Engrs. which could provide a feedback path. causing either an increase or decrease in amplification according to the phase. more usually. causing a reduction in damping which. 9. at in the superheterodyne technique described later.f. . The result is a flattopped resonance curve slightly displaced from the true resonant point. The voltage developed across this impedance by the anode current of the second valve is thus automatically introduced into the anode circuit of the first valve. whether inductive or capacitive. choke or resistance with a capacitor bypass is inserted in each lead. At stage gain can be quite large. In Fig. It is only troublesome at frequencies above referred to in more detail in Chapter 13. 24. the will cause The two curve. W. The effect is normally only serious with triode valves. is Multistage Amplifiers. 82.
THE RADIO RECEIVER 417 bypassed to earth without going through the battery or powersupply unit. i. which will be reintroduced via the . the circuit gives a small but nearly equal response to any frequency within quite a wide range. A still more important requirement. but equally a matter of layout. perhaps in the earthy or lowpotential side where it might be assumed unimportant. 9. The currents in the last stage will set up voltages due to the impedance of this common lead.8. for the use of a high L/C is ratio ensures a fairly steep "skirt" to the resonance curve. 9.9. then two effects will occur. Highfbequency Decoupling Circuit Residual Signal The layout of an amplifier is also of vital importance from the point of view of selectivity. Suppose we have two circuits. which means the reduction in strength likely to be obtained on a signal relatively close in frequency to (only a few kHz away from) the wanted station. From the knowledge of the constants of the circuits it is possible to calculate the resonance curves (allowing for valve damping and similar factors) and arrive at an estimate of what is called the adjacent channel selectivity. however. this may not necessarily be the case. ^Wffv Fig. in the Cause The Common Impedance Z Hightension Lead will Coupling between the Valves Pig. This is partly a matter of circuit design.e. separated by one or more amplifying valves. and even with good circuits Once the relatively large resonant gain has been lost. at frequencies 20 kHz or more off resonance. With a powerful local station close at hand. the residual voltage even well away from the tuning point may be quite large and it is necessary therefore to ensure that each circuit continues to exercise its full selecting action. each containing a short length of common lead. is that the selecting action shall continue to be progressive.
then some common coupling is present and it should be located and removed. Where this does not apply. it is possible to increase the number of stages and to obtain thereby an improved performance.418 RADIO COMMUNICATION first circuit into an earlier portion of the amplifier. Atmospheric Disturbances or Interference. and they appear indiscriminately over the whole frequency spectrum. They are in the form of damped waves of so short a duration as to be practically instantaneous. 1. This will produce feedback which may either increase the gain (and in the limit produce instability) or reduce the gain and cause the set to be less sensitive ' than it should be. The only satisfactory method of combating them is to use a combination of sharp selectivity and directional methods. . mainly from considerations of cost. The second and even more pernicious effect is that residual signals from a powerful station. Such matters require careful attention if an amplifier is to behave A good test is to see whether the amplifier "cascades" as it should. partly filtered by the first stage. The commercial receiving station is located away from such sources and adequate suppression is fitted to any machinery which has to be in the vicinity. With broadcast technique this is not always possible. Interference from electrical plant is a local disturbance and is more troublesome in broadcasting than in commercial practice. or to use frequency modulation.6. It is necessary to limit the amplification per stage because there is a overall amplification beyond which useful results cannot be obtained. principally in the direction of increased selectivity. in which case the required information is largely independent of the amplitude. The cause of background noise is threefold. This limit is set by the noise level in the amplifier. Atmospherics are maximum random waves which emanate from natural sources. and it is clearly necessary that the signaltonoise ratio shall be such as to enable the signal to be clearly distinguished. as explained in Section 9. and . which manifests itself as an indeterminate background which is amplified at the same time as the signal. properly. so that the remote channel selectivity suffers seriously. If the gains of two adjacent stages are A and A the gain x of the two together should be AA X If it is not. Noise Level In broadcast technique the number of stages is limited. have a direct entry into the later stage without having to go through the filtering action of the intervening stages.
electrons v* where e is i a is = 2keFMJg i g is is is A k and F is the charge on an electron. Shot Noise. is determined. and as a result of these various factors F increases to a figure of the order of 025 to 03.THE RADIO RECEIVER 419 additional devices have to be used at the receiver to minimize the disturbance. the anode current. Both this and the shot noise are only important in the first stage of the amplifier where they are followed by the full amplification of the remaining stages. Thermal Agitation. acting as a reservoir as already described. This is a noise arising from uneven emission of the in a valve. the noise is random and is distributed over the whole frequency spectrum. The equivalent grid noise due to shot effect can be shown to be given by 2. and F has a value of about 005. although the steady anode and screen currents bear a fixed relation for any given operating conditions. With a temperaturelimited cathode F 1. the mutual conductance. the frequency band of the receiver. — . In the case of thermal agitation. As is to be expected. produces five or six times the noise of a triode. a factor dependent on the operating conditions. from a noise point of view. With a pentode valve the presence of the screen overcomes the space charge to a large extent. as explained on page 445. but with a valve such as a triode the space charge. the impedance across the grid and cathode of the first valve develops a noise which is proportional to the effective resistance and to the absolute temperature. In other words. the instantaneous values are fluctuating slightly. while the random distribution of electrons between anode and screen further increases the noise. If adequate emission is provided from the cathode. a constant. The criterion of goodness. however. by the ratio 2 /i The higher this can be made the less will be the noise. A screened valve. this is continually surrounded by a space charge which acts as a reservoir so that the effect is reduced considerably. arises from the movement of the electrons in the material of the conductors. while with a frequency changer the conditions are still worse. This 3. for a given class of valve. sometimes called Johnson noise. so that fif . considerably reduces the noise. therefore.
we must have a signal of 50 or 60 microvolts. The uniformlydistributed is sometimes called "white" noise. in which case the formula reduces simply to a tuned circuit behaves as a high resistance that this dynamic resistance should be employed in the formula. The magnitude of the effect may be easily assessed. a tuned circuit having an effective resistance of 025 megohm at a temperature of 300 K. followed by an amplifier with a 10kilohertz bandwidth would develop 64 microvolts. therefore. Q. This is a general formula which assumes that the resistive component of input impedance is not constant but varies with frequency. T = absolute temperature = 273 + temp. This is defined as the ratio of the total noise power fed into the load to that due to the thermal noise in the source. by analogy with white light. For example. R = resistance component of input impedance. Thus N=j where \ut 2 dfJj'A%^df /2 A is the stage gain at frequency / within the band — fv . which includes all the frequencies in the visible spectrum. It should be noted that = LfCR and Noise Factor To distinguish between the noise generated by the valve circuitry and that due to the source impedance it is customary to specify the noise factor of the valve. To provide a satisfactory signaltonoise ratio. The magnitude of the voltage produced can be calculated from the formula to fulfil random noise «n * = 4JcT J ' Bdf where k = Boltzmann's constant = 1380 x 10 23 J/K. which accounts for the statement made on page 369. In the more usual case the resistive component is constant.420 RADIO COMMUNICATION a receiver with a wide bandwidth inherently possesses more background noise than a sharply selective receiver. that the smallest commercial signal is of the order of 10 microvolts per metre. in °C. which is a further argument for restriction of the bandwidth to the minimum necessary the particular requirements. limits of the receiver and fx and /2 are the frequency Hz.
the noise factor becomes simply N= 10loBm vaul 'l4kTRA* receiver. and this leads to the need for some means of controlling the amplification of the r. A most satisfactory way of achieving this is by using what are known as variablemu valves. the effective amplification is under easy and simple control by merely altering the grid bias applied to the system.f. so that in the end the interference would be eliminated. the valve should obey an exponential law so that the rate of change of slope at any point is proportional to the slope itself at that point. and therefore a compromise is adopted. give decreased amplification as the grid bias is increased. there is still a "hole" through which some electrons can flow. Normally we should pass the signals through still further tuned stages at each of which the ratio of wanted to unwanted signal would improve. v in vmt is = is usually not frequencydependent. Variablemu Valves The signal at the output end of the has already been mentioned. and this gives rise to a pernicious effect known as crossmodulation. These are provided with a speciallydesigned grid so that the slope decreases progressively as the grid bias is increased. the strength of the interfering signal may well be large enough to swing the valve over an extended . of course. when the major part of the electron emission is cut off by the negative voltage on the grid. and if also substantially independent of frequency over the band. By suitable design a smooth variation of slope can be obtained. however. This results in a valve having a rather low maximum conductance. As said above. Theoretically. by omitting some of the wires in the centre of the grid. Let us assume that we have an interfering signal several times as strong as the wanted signal.THE RADIO RECEIVER The factor is often expressed in logarithmic units. The necessity for keeping the voltage applied to the detector within certain limits Gain Control . the effective amplification factor is quite small. Consequently. and since the grid at this point is a widemesh one. In the early stages. either in the form of audiofrequency output to a loudspeaker or telegraphic signals to a recorder. must obviously be capable of control. Then. but the variation is not uniform. CrossModulation Any ordinary valve will. stages. in 421 which case 2V(dB) 101og10 iy.
000 ohms. are of limited application and the majority of modern circuits use the variablemu technique. If e = Ey sin coj + E a sin cojt. it is sufficient to assume a linear relationship and ignore the subsequent terms. involves e» = E 3 sin 3 cot = JiS 3 (3 sin cot — sin Scot).c. A variablemu valve is so designed. In a variablemu valve this effect is minimized by avoiding sudden changes in the slope. however. adequate for most practical purposes. pentode may be anode load. etc. the effective gain of the stage will depend on the strength of the interfering signal so that the output of the wanted signal will be controlled by the modulation of the interfering signal in addition to its own legitimate modulation.f. component and a second harmonic. In a simple treatment. amplifier is required to operate in an untuned condition. which is in parallel with the load. firstly by the best practicable preselection in the tuned circuits prior to the first valve. . It must therefore. For example a capacitance of 13 pF at 6 MHz has a reactance of 2. The gain is gB (g being the mutual conductance and B the anode load) up to a frequency for which the stray circuit capacitance. Other methods of radiofrequency gain control are sometimes used. then e a = E" sin 2 cot = £E 2 (1 — cos 2cot). the arrangement being simply a = ge + g'e'12 ! + gr"e 3 /3 ! + .f. If e = E sin cot. once introduced. thus introducing the crossmodulation. If over this portion the slope of the valve varies considerably.f. Aperiodic Amplifiers In certain circumstances an r. becomes comparable with B. so that the second term introduces a d. which thus contains not only a third harmonic but also a component of fundamental frequency which adds to that resulting from the first (linear) term. particularly with the bias run back.422 RADIO COMMUNICATION portion of the characteristic. . such as variation of screen voltage. and secondly by ensuring that the valve will handle a really large input without serious distortion. alteration of aerial coupling.be avoided at the start.* This crossmodulation. but in certain conditions these have to be taken into account. The third term. the curve being approximately exponential in character so that the rate of change of slope is always proportional to the slope itself. In general. These. . the third term contains fundamental components at both frequencies. technique discussed in Chapter 10. If this is not to exercise appreciable shunting action the anode resistor would * As previously mentioned on page 249. the valve characteristic can be represented by the expression » For this purpose an r. cannot be removed by subsequent tuning. this requires a very low value of load but with modern valves appreciable gain is still possible. used with a resistive special example of the a. however. because it contains a term at the frequency of the wanted signal.
on the other hand. A transistor. the term 4Q 2 d 2 must be unity. if this the value of is the bandwidth required 6 MHz and G 25 pF. is the resonant frequency./ =f IQ =/ /27r/ CP = P may be 1/2ttCA calculated. television frequencies. so that the gain is gP. . With an EF91 valve having 76 the gain would be 81. Its effective series resistance is then L\CP and the Q of the circuit Lco/R Lm](L\CP) 2wCPf where/. the impedance at the midfrequency is simply P. P= = 1. 423 need to be limited to about 750 ohms which.f. amplifiers for Here it is essential that the gain shall be reasonably uniform over an appreciable deviation from the mean.THE RADIO RECEIVER slope of 7. ±3 MHz at a mean frequency of 50 MHz. circuit when  bandwidth A From = 2<5.F. g = 9. R. In such circumstances the damping of the parallel resistance P far outweighs any other resistance in the circuit and the network behaves like a paralleltuned circuit having an equivalent resistance simply equal to P. is \\2mCP For example. Wideband Amplifiers An alternative requirement is to be found in r. Hence the general principles outlined in Section 9. is a lowimpedance currentoperated device so that the basic requirement is a matter of obtaining the optimum power transfer. was shown on page 96 that the relative response of a off tune is 1/V(1 + 4Q 2 d*). i.070 ohms As already explained. V/V = 1/V2 For this to obtain. Now 6 = (/ — /„)//„ which is the relative deviation on each side of the tuning point. Transistor Amplifiers In a valve amplifier the design is based on considerations of voltage gain.e. e. The total deviation is thus 2d and the total it Now. = = .3. and the usual procedure is to make the anode load an inductance of such a value as to tune to the mean frequency with the stray circuit capacitance. so that Q = 1/25. This requires a very broadlytuned circuit. The bandwidth of a network is usually assessed as the point where the response is 3 dB down. with a valve having a would give a gain of 525. where <5 = (/ — /„)//„. and then to shunt this with a low resistance to broaden the tuning.1 have to be applied in a somewhat different manner.g.
Hence the primary inductance is increased to a more . . Fio.f. Fig. A simplified equivalent circuit is shown in Fig. requiring a large tuning capacitance. 9. The transistor Trt is Fig.f. With such cores the coupling factor is very nearly unity.10 convenient value. = . and since fairly high ratios (between 5 and 10 to 1) are involved these transformers are usually housed in Ferrite cores of the type described in Chapter 4.Rz = n 2 Ri. Z is the dynamic impedance of the tuned circuit LfGR. 9. Gm e . 9. while R t is the reflected impedance of the input impedance of the following stage. but must be replaced by r. The transistor Trx is represented as a current generator Gme developing power in the network RyZR^ G m is the effective transconductance di mt lde in the significance of which is discussed later. 9.F.11.f. ratio it is more convenient to tune the primary winding the primary inductance calculated from purely matching considerations is usually inconveniently low. 9. R t is the output impedance of the transistor. Simple Transistor R. which simplifies the calculations. the collector being connected to an appropriate tap to maintain the matching. Stage coupled to !Y a through a transformer of a suitable ratio to match the relatively high output impedance of Trt into the low input impedance of Tr % Since this requires an appreciable stepdown even so.'zfa /CR .10. stage. Simplified Equivalent Circuit or Fig. For the present these can be assumed .10 illustrates a basic r.11. 2? a is thus n 2 B in where n is the stepdown ratio of the transformer. stage. transformers to obtain proper matching. the use of simple tuned circuits is no longer practicable in an r.424 RADIO COMMUNICATION In particular.
and Rt R 2 R the voltage across the network %IR. The optimum conditions are obtained when the external internal loads are equal. which occurs when •«!• This is an important result. (2)) is %IR(Q Q w)jQ = = t . = —  The power is proportional to F 2 . however. = .circuit capacitance. for it would seem at first sight that the optimum result would be obtained by matching the effective load (R 2 and Z in parallel) to the resistance R t but in fact the requirement is for maximum power to be transferred to Tr 2 (i. The working impedance Similarly Z = Q X. hence This Za = QwX = ^ + ^Q§z/( ^ + 2 ^ as 2 )Q 1 (D may be rewritten in terms of Q and Qw Zw = R^Qo . R t). The working voltage is IZ m which (from eqn. Zw = CJf. maximum power in R 2 which is obtained when R 2 . = Insertion Loss From these expressions we can deduce the loss introduced by the presence of the load Z due to the tuned circuit. and the power actually available externally is appreciably less than half the total. the external load is made up of JR 2 and Z in parallel so that a proportion of the available power is dissipated in the tuned circuit impedance Z.THE RADIO RECEIVER to be all resistive. but the condition for maximum power transfer (i. .e. If Z is infinite. so that Actual power output f& O A * Output with perfect coil . If there were no coil loss (Z = oo) the optimum condition would result when R1 R 2 In a practical case the tuned circuit absorbs some of the power. in which case half the total and power is and half is available externally. 'o J a «2 = maximum when dP/dJJ 2 = 0. where X = Leo.QW )I(R + R The power in R V /R = (IZw f\R 2 is 2 2 (2) whence This P= is — I2 L (Rx +R (3) 2 )Q. 425 iJ t the capacitive components of and R2 con stituting part of the tuned. But Zw is Rv Z and R 2 in parallel.e. In this case. maximum power in R 2 ) is still that R t R 2 This may be proved as follows: dissipated internally = .
2 so that the power gain A v = OJBL Bin Knowing the ^ ^J (^ + RJ circuit values. gives = Effective Gain The current generated in the transistor divides between the and external loads B t and B L as shown in Fig.Qw ) dB Thus if (5) which is actually a more convenient form. is composed of B 2 in parallel with the internal ) ) .426 RADIO COMMUNICATION In logarithmic form this 20 lofo (Q may be written ) . Illustrating Division of Current is is and the proportion of the developed in B 2 is Z]{Z + total external Rz\ power which Hence the power in B 2 actually The input power is e /i? <n . tuned circuit impedance Z. namely though strictly speaking 20 log10 {Q I(Q . 9. The load BL however.QJQ loss.12. 201og10 05 = 20(16990) == 20(03010) = 6 dB (5) the minus sign denoting a loss. so that the current will divide further. 9. dB (4) This is sometimes called the insertion the loss is the inverse of this.12. The proportion in B L is i L = iBil{R l + B L = G meB 1 l(B 1 + B L and the external power is thus i L z B L = O mi e^B L{B l l{B 1 + B L )f.QW)IQ = 20 log 10 (1 . The inverted expression the loss directly as 20 log 10 2 6 dB. the gain can be evaluated from these . expression (4) becomes Qw = \Q Q. i=Gm e Fig.
but can be translated into practical terms by an autotransformer arrangement as described below. = J^iJjiJg.10 3 /(4 . = \R^ if Rt = R 2 This makes Qw = iQo> givm8 an insertion loss of 6 dB. It was shown on page 110 that the bandwidth of a single tuned circuit B = fjQ. whence L = Z„layQ w This determines the value of inductance required. 6 is = 7100 = 385 dB The actual gain will be 6 dB down on this. Zw = ^Rv The inductance L = RJlcoQn .000/800) = . The output capaciTrx would be about 40 pF while the reflected capacitance of . namely 325 dB = 1780. however. . There is no point in making it too large since this will only increase the insertion loss.THE RADIO RECEIVER expressions. R L then = R 2 and yl„(max) . and hoe (= R x ) = 29 kQ. a figure of 250 = 1/Lco 2 then becomes 2450 pF. If the gains are expressed in dB. Hence Q w = f/B. would have Gm = 35 mA/V. . where / is the resonant frequency. transistor n The maximum gain {A p ) max = ^(29.f. . 10 3 x 08 10 3 The inductance is RJicoQ^ Assuming a bandwidth of 9 kHz and a resonant frequency of 500 kHz. The maximum gain is obtained when Z is infinite. is then The capacitance required is then I/Loj 2 This is usually inconveniently large. so that L = 29. The actual gain is then (6) A. Z = \RX as just stated. and would involve needless expense. =i 10« x 29 . 2n 500 . It is customary to make Z equal to the combined impedance of R± and R2 in parallel. A v = \A V Design Procedure The first step in the design is to determine Q w This is determined by the bandwidth requirements. . . A work out the values for a practical example. The value of Q is arbitrary. (5). 427 In practice. The working impedance Zw = LcoQ„. Q w = 56. If R x = R % and . X 35 2 . 56) = 415 ^H is The capacitance C high. The stepdown ratio to make R2 = R% will then be It will be useful to typical r.QJIQ Y the square root of the power gain. The voltage gain = is lOm R1 R2 [(Q . which too tance of pF being more practicable. h ie (= R2 ) = 800 Q. 10 3 . it is simpler to work out the maximum gain and then calculate the insertion loss from eqn.
would be necessary to use stranded Effective Transconductance wire. 9. The wire used must be such that the resistance of the coil (at 500 kHz) 112.13 (a). leaving 2385 pF to be supplied as the reflected capacitance of 250 pF across the full 31. 9.11 is modified in practice because of internal feedback within the transistor. for which it . These two strays will thus contribute 65 pF to the total. . this is somewhat less than the intrinsic transconductance gm by reason of the finite resistance '(. The coil would be tapped at is 11 ohms to provide Q 140/31 45 turns. 9. and the total primary. core 415 The wire used must be such that the effective resistance (including For the example quoted this would be losses) = LJCZ. 10«/(2450 10. where of turns. so that for 400 fiH. Illustrating Internal Feedback .12 X 145 10 3 ) = 117 ohms. Internal Feedback The performance of the simple arises circuit of Fig. the number of turns required is 140. The gain calculations involve the transconductance Om As explained in Chapter 6 (page 292).' in the connection to the base. as shown in Fig.!. T (a) tb> W'e <W> *> T Fig. while the secondary winding would be 45/6 7£ = = N N = = = turns. In practice the difference is not great. Such cores are very consistent in production and the makers specify the number of turns per microhenry. . (?„ being of the order of . This from the inevitable small capacitance between collector and base.13.428 RADIO COMMUNICATION Tr 2 input would be about 25 pF (as is discussed shortly). these windings will be housed in a Ferrite core assembly. A typical core is the number would provide an inductance of 2 /50 fiH. Hence the tap ratio is \/(2385/250) 400 fiH. as a result of which a small fraction —runi A z. 095srm . inductance is 415 X 31 2 As said earlier.
The feedback can. aCT fim Zx Z% If the input < 1 and output will occur at the circuits are tuned.. nevertheless. In either case the maximum feedback occurs at the frequencies for which the reactance and (parallel) resistance are equal. Below resonance the voltage at will lead on ic so that if leads by more than 90°.. A will result. produce appreciable distortion of the tuning.. + . Above resonance the feedback current will lead by less than 90°.. since below resonance the gain will be increased while above resonance it will be decreased. producing a phase angle of 45° (90° total). but since Zt tion of this voltage developed at B is Zx \(Zf The voltage is much greater than Z x this can be simplified to Z X \ZS at B is thus O^tZtZJZf. and if this is greater than v t oscillation . where Z' is the impedance shunted by the feedback path Zt Z v However. Zf is much higher than Z 2 so that Z' may be taken as equal to Z2 The proporZx ). the voltage at is in phase with i c and since l/a)Cre is much greater than Zv the feedback current if and hence the voltage at B. If Z x and Z 2 are resistive. 9. For stability this must not exceed v t Hence. maximum feedback frequency for which the phase angle (of each . This is actually a small capacitance shunted by a high resistance. writing \\ofire for Zt the criterion for stability is Z% + . 9. will lead by 90° and will not affect the signal. It may be that even at this frequency the loop gain is less than unity so that oscillation does not occur. The equivalent circuit is shown in Fig. producing a lopsided resonance curve with a peak at a frequency slightly below the true resonant point. if the magnitude and phase of the feedback are such as to increase the overall gain. the voltage at B will also lead on it so that the combined effect produces a component at B in phase with the input. . This produces a slight modification of the values of B t and iJ 2 but the more important effect is that. . resulting in negative feedback. the circuit may oscillate continuously. = A . where Zx and Zi represent the input and output impedance and Zt is the impedance of the feedback path. If the loads are tuned circuits the impedances will only be resistive at the resonant frequency. but for practical purposes the parallel resistance may be ignored so that Zt 1/coC.THE RADIO RECEJVER 429 of the output voltage is transferred back to the input. Stability Factor In Fig. thus having a component in opposition to the input.13 (6) the voltage at A is Gmv t Z'.13 (b).
For complete compensation both capacitive and resistive components of the feedback should be neutralized. the made for tolerances in the parameters — minimum requirement. RADIO COMMUNICATION This occurs when Z 1 = R /\/2 1 and Z2 = ii 2 /\/2. This can be replaced by ordinary transformations to an equivalent series form as has been done in Fig. Using the values 15 pF the criterion in the example worked previously. With early transistors Cre was of the order of 10 pF. but with a modern r.h. Zf is actually Cre shunted by a high resistance of the order of megohms. types have Cre as low as 015 pF.14. Modified Form of Fig. . Hence the stability criterion can be written wCreQm Rx R 2 This is < 2 allowance must be as the production spread so that the criterion is taken as 2JK.f. where is a factor of safety called the stability factor. To enable the full gain to be developed it is necessary to counteract !fen:/ Fio.R2 must be less than 05.g. transistor the capacitance is only some 1 to 15 pF. 9. which should be at least 2 and is usually 4 or 5. to 1 or 15 MHz) would introduce the risk of instability. C 3 is then slightly less than Cre .430 circuit) is 45°. K = Neutralizing A circuit which is inherently stable is said to be stability limited. a)Ct eCrm R 1 . but such circuits in general do not make the best use of the transistor. Thus with a stability factor of 4. In practice known — . which would be acceptable with a stability factor of 4. and special v. with C re works out at 0385.14. 9.f.13 with Addition of Neutralizing Circuit the effect of the internal feedback by some suitable external circuit which provides a compensating feedback in antiphase in a manner similar to that adopted with valve techniques. but any appreciable increase in frequency (e. 9.
9. ® (a) (b) Fig. Such a circuit is said to be unilateralized. Cn must then be Cre%/w 2 Still another arrangement is the bridge circuit of Fig. R = . The then be neutralized by an external circuit taken from the secondary of the coupling transformer (which must be connected to provide the required antiphase)." This was discussed on page 418 and was shown to arise partly from small irregularities in the electron flow (shot noise) and partly from thermal agitation of the electrons in the circuit impedances. 9. the neutralizing capacitor may be fed from the primary. Noise Level Transistor amplifiers suffer from the same limitations as valve background "noise. Thus. However.10 to obtain the correct antiphase. 9. The impedances involved in transistor circuits are much lower than with valves but the amplifiers in respect of . as shown in Fig.THE RADIO RECEIVER while 431 internal feedback can R3 is only of the order of 10 kQ. Two Fobms of Neutralizing Circuit are not affected in technique by the internal feedback. and for the majority of circuits the resistance is t omitted.15.15 (a). Since the transformer will normally be arranged with a stepdown ratio the neutralizing circuit impedance must be correspondingly reduced to equate the internal and external feedback.15 (6). in which Cn CreCJC2 . . adequate neutralization being effected merely by a correct choice of Cf Various alternative arrangements are possible. = = which means that the input and output impedances \n2 in. so that Cf nC 3 and Rt B sjn. improvements have rendered this complete compensation of secondary importance. It has the advantage that there is no transfer of energy between output and input in either direction. 9. by rearranging Fig.
This second factor also increases with the collector voltage so that to keep the noise level low I e and Ve should be kept small. and «. It is also dependent on the source impedance. 9. The "excess noise" component is an additional factor which appears to be connected with leakage current. 9. . The low input impedance of a transistor. falling to a broad minimum when $ in and thereafter rising again.432 RADIO COMMUNICATION currents are correspondingly greater. mA 75 fy noise Ie = 1mA 5 fee 01 1 10 100 ' woo f(kHz) Fio. but these are factors in the design of the The noise is directly proportional to (absolute) temperature. however. so that the input noise is much the same. namely intrinsic noise and excess noise. and to the bandwidth of the circuit. It is most important at low frequencies and in fact decreases in a linear manner as the B =R frequency increases. Transistor noise appears to have two principal constituents. Fig. introduces appreciable thermal noise within the transistor itself (mainly in r ) in addition to the shot noise which is inevitably w present. it is sometimes called the 1// component for there As the frequency continues to increase. Variation of Noise Factor with Frequency division of the current between base and collector.16 shows the type of variation encountered. Intrinsic noise arises from the normal transistor action and depends mainly upon the emitter the noise is roughly constant but then Up to about 1 begins to increase proportionally. a becomes complex and is an increase in the partition effect so that the intrinsic noise begins to increase rapidly. It is also desirable to keep rb transistor. This is mainly due to a partition effect (similar to that in tetrode or pentode valves) arising from the current.16.' small. this reason.
m. If the factors are not greatly dependent on frequency then (using the expression for v n * on page 420) the noise factor is simply to the source A . Illustrating Noise Level in Cascade Amplifier Thus if Sin is. say. 50 dB.s. the additional noise generated in the stage will increase the relative noise level. N = 10 log 10 v^lticTRA* where k = Boltzmann's constant absolute temperature T= R= v source impedance = r.17. . If A. and S^ = signal/noise — Smt If S is expressed in dB. then N = 8 in . 9. The noise factor for the stage is then 4 dB. Input vs +vn Fig.m.s. noise output voltage A = voltage gain. so that Smt is less than 8 in becoming perhaps 46 dB. noise output voltage. It is the first stage of an amplifier which is of major importance. Hence N= where v is 10 log 10 [ V dfj j f *A*vn * dfj dB the r. vn is the r. . This is defined as the ratio of the total noise power fed into the load of the transistor to that portion which is due to the thermal noise of the source impedance.m.THE RADIO RECEIVER Noise Factor 433 Since the total noise is produced partly by the (external) input impedance and partly by the transistor it is desirable to separate the two by specifying the noise factor of the circuit.s. = signal/noise ratio of the source. This may be expressed in another way by denning the noise factor as N = SiJS^ where Sin ratio of the output. and/or v and vn are frequency dependent they must be expressed in terms of/. Let the input signal be v s and the input noise be vn . Consider the twostage amplifier illustrated diagrammatically in Fig. noise due and is the voltage gain. 9.17.
Hence. The signal/noise ratio at this point will then be .f. The advantage of this procedure is that the tuned circuits employed are all fixedtuned (except the aerial and oscillator). and here it will be observed that the operative factor is A/4.. particularly at the higher frequencies. highefficiency circuits can be used giving high gain and improved selectivity without the troubles attendant on the use of ganged r. and is suitably amplified until the voltage is sufficient for the particular detector in use. so that to maintain a high signal/noise ratio the gain of the first stage should be high. In particular. A modulated wave is then produced having a frequency equal to the difference between the incoming oscillation and the local oscillation. 9. This difference frequency is then accepted by the use of tuned circuits. stages.f./(»„ + AJAJ x) clearly less The signal output than S from the second stage will be A^A^o.f. signals as such has certain limitations. = A (AyVn + A + A = A A (vn + \IAj_ + A /4jJ = v~ftvn + ^ \A + ^jA A 2 x) 2 x 2 2 2) l l x 2) very little different from S x because A 2 is divided by A x In other words the internal noise of stages subsequent to the first contributes very little to the overall signal/noise ratio. and it is more usual to arrange for the amplification to be effected at an intermediate frequency by using the superheterodyne principle. while the noise output will be V2n whence This is S2 . the amplification in a t. while A should be minimized by using a small value of I e (less than 1 mA). Hence it is the first stage which is of significance. The Superheterodyne Receiver Amplification of the incoming r.434 RADIO COMMUNICATION so that S v s /vn Assuming that the signal and noise voltages are amplified equally (which is not strictly correct) the respective A l7 where A x outputs from the first stage will be AjV s and xvn is the additional noise generated internally in the stage. The procedure is to convert the incoming signal to a different (lower) frequency by mixing it with a suitable local oscillation.4. stage varies with the setting of the tuning capacitor since the dynamic impedance L/CB . + A = «. with a low value of source impedance (= B in for optimum and A2 performance). — A + 8± which is = Af)s \(A xvn .r.
amplifier tuned to this new frequency. by a detector and audiofrequency output in the The local oscillator tuner in such a manner as to provide a constant frequency difference between the incoming signal and the local oscillation.f.f.F. The incoming signal is first passed through an r. Amplifiers A valveoperated i.6.F. followed R. it is convenient to review briefly the requirements of the i.f.THE RADIO RECEIVER 435 becomes less as G is increased. or frequency changer. The basic principle remains the same but the circuitry has to be adapted to the higher frequency. is mixer ganged with the r. i / tuner f Frequency changer I. 9. stage. This difficulty is obviated in a superheterodyne receiver. thus providing the required fixed intermediate frequency. of 465 or 470 kHz is customarily adopted and the details which follow will be mainly concerned with frequencies of this order. practice except that because of the fixed frequency it is practicable to use bandpass . The essential components of a superheterodyne receiver (commonly known as a superhet) are shown in Fig. to accommodate which an i. stage follows normal r. Valve I.f. of 107 MHz is usual. as is discussed in Section 9. It is evident that the mixer.f. Superheterodyne Receiver stage.f. stages since to some extent the technique is the same for both circuits.18.18. however.F. For frequencymodulated transmissions in the v. of the order of 40 MHz is necessary as is discussed in Chapter 16.f. is a vital part of the system. which gives uniform gain and selectivity.f.h. Detector and a. Still higher frequencies are involved in television receivers.f. It is then passed through an i. amplifier. an important advantage. The signal is then passed to a mixer stage in which the carrier frequency is translated into a chosen (lower) intermediate frequency. band an i.f.f. Before discussing this in detail. amplifier amplifier 7777 / / I Local oscillator ^/— 1 J Fig. which require a video bandwidth of 6 MHz. 9. which is usually a simple aerial tuning circuit but may include an r. For ordinary broadcast reception an i.f.
This means that the Q of the secondary is no longer .f. The design procedure is thus first to determine the overall Qw necessary to meet the bandwidth requirements.19. and since Q value for L to be determined such that with the particular construc = tion proposed the resistance (at the i. Z x is the effective impedance of the primary. and . 9. as illustrated Z t ).type trimmers operated through holes in top of con + Magnetic slug cores' adjustable from side Coils Colls Fig. with critical coupling between them. R.) is of the right order. so improving the selectivity. The coils are usually housed in the same screening can. tuned by small preset capacitors. The coupling transformers thus have both windings tuned.5. Transformer Constructions allowing for the transferred secondary impedance. = .F. With critical coupling.436 RADIO COMMUNICATION tuning. because the detector load is not negligible. The individual circuits must then be designed to have an unloaded Lco/R this enables a Q equal to twice this value. which feeds the detector. 9. where g is the mutual conThe stage gain will be gZx \(ra ductance of the valve. by adjustable ferrite slugs. or in Fig. pentode ra is large compared with Z1 so that the gain is simply gZ v However. and ra the internal resistance. assuming identical circuits. finally the capacitance can be evaluated 1/ico 2 Z is then L\GR (= Q Lco). and Z w \Z Q from which the gain can be calculated. a slight modification is required. as is explained in Section 9. Typical Valve I. For the final stage.19.f. which with critical coupling is half that of either circuit alone. so that the coil separation can then be arranged to provide the required mutual Mm = inductance from C = M = R/m. With a normal r. as discussed earlier. Compress/on.
stage is given in the section on f.f. but a modified value of Qa would be evaluated to allow for the load across it.f. this may not provide adequate selectivity as was discussed on page 417. It is customary therefore to use several i.THE RADIO RECEIVER 437 equal to that of the primary. . Both circuits would be designed to have the same Q as before. with an additional stage or stages if the selectivity is insufficient. by an equivalent capacitance coupling as shown in Fig. but in the majority of popular receivers only a singleiuned circuit is used. The design of a singletuned i.5 (page 454). An example of a doubletuned i.f. the third being tuned to/ (but with a modified Q w to allow for the detector load as explained above). Stagger Tuning While a single circuit may comply with the bandwidth requirements. the final stage feeding the detector requires modified treatment. and n is the number of stages. As with a valve i. when if the circuits are properly cascaded the offtune response is n (sx ) times that of a single circuit. which is slightly higher than before. 9.20. If this is not adequate. factor k It is = #//„. stage previously discussed.f. with two circuits each mistimed by <5/2 on either side of f the response curve is the same as for a coupled circuit with a coupling .. For more sophisticated receivers doubletuned stages are employed.f. by the use of a suitable tapping or coupling coil.29 (page 112). and this would be used in calculating k„it With this modification the effective primary . where s x is the relative response of a single circuit x kHz off tune. receivers (page 470). but with a singletuned circuit this simply involves different matching. It can be shown that.f. in which case k„ it ll\Z{QiQ 2 ). stage is thus similar to that of the r.F. 2. but since the coils are normally housed in closed ferrite cores the necessary coupling has to be piovided externally. the successive tuned circuits may be adjusted to resonate at frequencies slightly displaced from the midfrequency /„. as illustrated in Fig. Transistor I. = impedance will still be \Z Q . Amplifiers Similar procedures can be adopted with transistor i. This is called stagger tuning and has the effect of providing an approximation to a bandpass circuit. or if this is inconvenient. as is discussed in detail in Section 9. stages. amplifiers.m. customary to apply stagger tuning to the first two stages.
It is desirable. carrier frequency is converted to the chosen intermediate frequency. the r. but it is important to ensure that the waveform is as pure as possible. Consider first the additive process. for which any convenient circuit may be used. Harmonics are to be avoided because they will heterodyne with than the wanted one and introduce audible whistles. 9. The second is to modulate one with the other. The sum of two sine waves sin co t and sin co a t is 2 sin f (eo + a> B )t cos \{u> — <x> s )t which is a sine wave of the mean frequency modulated at what appears to be half the difference frequency. In either case a local oscillator is required.f. and we shall therefore first discuss the basic requirements and then show how these are achieved with the two types of circuit. The first is to which produces a similar result but has certain advantages. to use a powerful local station or a small multiple thereof. . Theory of Mixing There are two methods by which two oscillations can be mixed. particularly if they happen to coincide with the frequency of a stations other therefore. This involves the process of combining.f. which is the device through which the incoming r. or mixing. This simple expression.438 RADIO COMMUNICATION I S/2 I 6/2 1Or % a. good oscillator circuit having a high Q. Illustrating Stagger Tuning amplifier is derived from the frequency changer. The circuitry is necessarily somewhat different according to whether valves or transistors are used. Frequency Changers The input to the i.f. signal with a local oscillation. 1/1/2 (0 Frequency — » Fig.20. add them together by the heterodyning process described earlier.
The amplitude of this difference term is Es which means that any modulation on the original signal will be faithfully reproduced. if is E„ is large com called heterodyne With electronic mixing.<a )t] In other words.co . This amplification. .o> s (ai. which is essential. and E Hence pared with E s there is an amplifying action. It may be noted that. does not provide a correct representation of the action.' —w = 1{E + E cos . Illustrating Addition of Two Signals or Different Frequencies a> s E the vector Es will be rotating at an angular frequency and the resultant amplitude at any instant is to E.co )t}*]l + sin 2 S s 8 )fl* (remembering that cos2 6 If = 1) E s is small compared with E this reduces to ) E ' t = E [l + (EJE cos (co s . which it is necessary to take into account the relative phase of the oscillations. 9.a> )tf + {E sin = [E* + E* + 2E E cos (o> . This will be understood by reference to the vector diagram of which E and Es represent the local and signal oscillations respectively. essentially different.21 in Fig. Now. both vectors will be rotating at their respective angular frequencies. 2EJ!S cos (co s . . giving an output proportional to (E s ') 2 the modulation term is . B )t which is proportional to both E.THE RADIO RECEIVER for 439 however. s (o>. If these two are added together the resultant amplitude will be the vector sum of the two components.21. the mechanism is The incoming signal is made to modulate the of the and the output current is proportional to the product two component frequencies. In such circumstances two . 9. but relative Fig. if a squarelaw detector is used. the envelope of the resultant wave in amplitude at a frequency equal to the difference is varying between the two component frequencies. local oscillation on the other hand.
440 RADIO COMMUNICATION distinct frequencies are produced.T.V. The signal is applied to the first grid of the hexode portion while the local oscillation developed by the triode portion is connected (internally) to the third modulates the current produced by the signal. 1 A. It will be noted that the amplitude is proportional to EJJlg. Since electronic mixing does not require rectification the hexode can operate as a normal amplifier and will develop substantial gain. of course. Line J megohm Triodehexode Frequency Chancier Fig. Assuming inputs of E sin co t and Es sin coj the resultant voltage is E„ES sin mjt sin wj. and will also provide an improved signal/noise ratio because of its higher effective mutual conductance. and the other to the sum.22. being disregarded and attention concentrated on the difference frequency. 9. the latter. grids are screens. it (mixer) grid where The second and fourth . so that again there is an inherent amplification in the process.)<] difference oscillations have been produced and the be extracted directly without the necessity for rectification. but with the increasing use of transistors it is only necessary to refer to one. 9.+ fj Stages WW Fig. one equal to the difference in frequency. Thus two distinct term may Triodehexode Mixer During the reign of valve technology many forms of mixer were devised. connected to a suitable positive potential as in a normal tetrode. Hence this type of circuit is now accepted as standard.C. namely the triodehexode circuit shown in •H. = £ ^g [cos (<w — <o — cos (eo + s )t «>.22. which uses electronic mixing.
9. which is justifiable since the valve is made with characteristics similar to those of an ordinary r. It will be noted that in the expression for electronic mixing on page 440 the output contains a factor of £.) frequency. This indicates that the conversion conductance is only half the equivalent mutual conductance. .f.23.f.f.23. This. which is utilized by including in the collector a transformer tuned to the difference (i. and it is known as the conversion conductance. and in practice. assumes that the anode impedance is small compared with the internal resistance of the valve. The gain of a frequencychanger stage may be determined in the same way as an ordinary r. where c is the conversion conductance and Z is the anode impedance at the conductance is at beat frequency divided intermediate frequency. Selfoscii^ating Transistor Mixer applied to the base. while the collector is coupled back to the emitter through a secondary circuit tuned to the oscillator frequency. transformer Aeria. pentode. 9. of course.on the signal grid. Transistor Frequency Changers The oscillating simplest form of transistor frequency changer uses a selfmixer circuit such as is shown in Fig. the gain will be twice as great as that which is obtained when the valve is operating as a frequency changer. being equal to cZ.THE RADIO RECEIVER 441 Conversion Conductance In frequency changers a term analogous to the customary mutual usually quoted.f. This is the change in anode current by the change in grid voltage at signal frequency. pentode. if a frequency changer is supplied with i.1 circuit Fig. The emitter will thus carry currents at both signal and oscillator frequencies so that an additive mixing is obtained. The signal is IF.
V = 2LcoQ Q IZ w If the i. 100 and Q = 120. the 55. so that a substantial stepdown ratio is needed.000 £2. stages. Oscillator design is considered more fully in Section 12. because the collector load will be tuned to the i. = = Input Transformer The input impedance to the signal (into which the aerial transformer must be matched) can be estimated from the expression for Z t (commonemitter mode) with RL 0. however.2. If the relevant impedances Z and Z f are 30 kQ and 1.f. With a typical core an inductance of 175 /jH would be obtained with 72 turns. and will have a greatly reduced impedance (approaching zero) at signal frequency. By the usual circuit laws it can be shown that this ratio is given by = = = . Hence a compromise has to be adopted. r^ The actual 175 fiR. With this reservation the design of the input transformer is in accordance with the procedure already discussed in Section 8.f. The \/(30/l) ratio of collector /emitter windings will be n combined loading of the collector and emitter is then that of Z and \Z In this case. In a frequency changer. thereby minimizing circuit damping.1 and it will suffice here to note the values in a typical frequency changer. however.f. and the lowest signal frequency is 550 kHz. in which case the primaries will be 11 and 2 turns respectively. so that the effective output impedance of the transistor varies widely. The coupling windings themselves are so proportioned as to match the collector/ emitter impedances. stages through a coupling transformer matched to the transistor as in a normal r. L will be 66. is (Q . The requirement is that the transistor loading shall only have a small influence on the tuned circuit.020 kHz. and it is customary to make the coupling transformer the same as in the succeeding i./OS0 will be 1. . it is not desired to n 2 Z t in parallel match this to the tuned circuit. stage.Qw ) 470 kHz. which permits the oscillation to be maintained with only a loose coupling to the transistor. = Output Transformer The output is fed to the i.f. it is not possible to offset the internal feedback by unilateralization because of the range of frequency involved. as shown. Taking Q w turns depend upon the core used. In practice this provides a relatively low effective load in the collector circuit which masks the variations in Zn .f.442 RADIO COMMUNICATION The oscillator tuned circuit is designed to cover the frequency band required (as explained on page 447) and has a high Q. If the tuning capacitance is 140 pF.
24. The circuit thus produces bursts of oscillation at a repetition rate (usually audible) depending in Chapter 12. but there is a further and more serious feedback path through the stray capacitance between the oscillator and aerial circuits. distorted * Squegging is an unstable condition in which a rapid increase in amplitude shifts the operating conditions so that the transistor is momentarily paralysed. 9. 9.* The stray feedback increases with frequency and will ultimately cause the oscillator to pull into step with the signal so that no i. FrEQTTENCY CHANGER ClRCTJIT WITH SEPARATE Oscillator amplitude of the oscillation.23 will produce a current through the internal feedback path in the transistor which will tend to reduce the amplitude of the oscillation.THE RADIO RECEIVER Effect of Stray 443 Feedback The oscillatory voltage at the collector in Fig. If the increase is is large the waveform and squegging may result.23. 9. This internal feedback is usually small and may be neglected. The voltage on the tuned circuit is some 6 times that on the collector and the feedback current reaching the base is still further increased by the amount of the stepdown ratio of the aerial transformer. The effect of this feedback depends upon the phase of the oscillator windings and may thus cause either an increase or decrease in the : Aerial jg circuit g Oscillaior tune Fig. as shown dotted in Fig. (The action is discussed further .) on the time constants. Since this is of the order of 10 1 the total feedback is some 60 or more times as serious as that in the transistor itself. The internal capacitance is of the order of 2 pF so that a stray capacitance as low as 1 pF can produce 30 times as much feedback as that due to the transistor alone. is produced.f.
Oe = g m Z\{rw A\vw s + Z) *) = 9mKrw\rVe + 1 + ja)C t<jf b e rbb . It will be noted that in this circuit the an earthed screen between the still r. If the collector current is to be parallel rVe \{\ ./(^i)r. = = w+ + . 9. The signal is amplified in a buffer stage before being fed to the frequency changer. In such circumstances the oscillator and mixer must be kept separate. suitable for frequencies of the order of 100 MHz./..f. and oscillator stages use the commonbase mode. is negligible." which is discussed in Chapter 12.This modified value of g m is called the conversion conductance./. Since the aerial and oscillator tuning capacitors are usually ganged there is appreciable capacitance between them which must be minimized by providing sections. Such a circuit. the pulling may be present at high frequencies. 9. expressed as Gcv (= g mVi). stage is approximately g m Z. The gain of a mixer stage (at the intermediate frequency) \9m"i Fig. Even so. so that the expression becomes m /(l + rw lrb e ).24. Thus in the simplified equivalent circuit of Fig. which is also fed through a small capacitance from an entirely separate oscillator.25. where Z is the impedance of rVt and CV e in v s Z\(r v1 ja>Cb e rb e ). where Z is the load and g m is the effective mutual conductance = dic ldv 1 {see Fig. which is essential to obtain a sufficiently high cutoff frequency. is shown in Fig.f. Conversion Conductance The gain of an r.444 RADIO COMMUNICATION This "pulling. 9.25 the collector current is g m v 1 and Z). Simplified Fbequency Changer Equivalent Circuit may be similarly assessed if the value of g m is (d* c )<. ge In a transistor the value of ge is not a definite proportion of g m (as in a valve) but depends on the frequency. Cv C 2 and G 3 would be ganged.25). which is the same as the expression for the normal transconductance G m .] . > . At low frequencies the reactive term in the denominator . 9. can only be avoided by reducing the stray capacitance.
it will be clear that the factor ijg2 = 2keFAia /g 2 is increased because the value of g is approximately halved. and often a valve is deliberately designed to have more shot noise in order to permit improvement in these other respects. Firstly the value of g m a//v e falls off with frequency. equally well be lower. = 3 dB down on g m in . which causes increased "pulling" between the circuits.f. stage becoming the predominating factor. Gc could be with representative values shows that if/osc ~ . In a transistor frequency changer the noise factor is somewhat increased by reason of the increased emitter current.THE RADIO RECEIVER 445 (see page 292). while the factor is found to be of the order 05 to 06. stages. this term becomes appreciable and the expression is better written in the form Ge =* 9ml(l + jfoCvfw) This is frequency. therefore.dependent for two reasons. but this involves long cathodes and consequent increase in interelectrode capacitances. Referring again to the basic expression v„2 quoted on page 419. However. The local oscillator is adjusted to differ from the wanted signal by the intermediate frequency and is usually arranged higher than the signal frequency for convenience of ganging. the signaltonoise ratio of the frequency changer becomes of minor importance. and a simple calculation 15 MHz.f. The most obvious form of false response is that known as image or second liability channel interference. Undesired Response Probably the greatest defect of the superheterodyne system is its to produce false signals. that for any setting of . which is higher than the signal frequency. Noise Frequency Changers A frequency changer is subject to increased shot noise. there is a further falling off due to the increasing effect of the input capacitance CVe It must be remembered that eo is 2tt times the oscillator frequency. but is not greatly different from that for normal amplifier operation. particularly in a valve. This arises from the large number of combinations which will affect the frequency changer. Up to a point the designer still aims to keep g2 jia high. however. and it follows. if the stage is fed with an amplified signal from one or F more r. the Johnson noise on the input to the r. . however. being 3 dB down at the cutoff frequency /a In addition. It could. As the frequency increases.
With low intermediate frequencies this is not always easy. The criterion is actually the relative values of the intermediate frequency and the signal frequency. Only one of these is the wanted signal.446 RADIO COMMUNICATION the oscillator frequency there will be two signal frequencies.f. is relatively high to eliminate second. Ideally. This is adequate to avoid second channel interference on medium and long waves but is not sufficient for short waves where the interference is present.channel interference. followed by a second frequency changer which converts the signal to a lower frequency at which the main amplification can be more easily obtained.. must be capable of discriminating between the wanted signal and another signal separated by twice the intermediate frequency. and each station will be found to show two tuning points. to avoid interference the preselecting circuit. and strong local stations will invariably give rise to these beats. Broadcast receivers usually use at least two tuned circuits prior to the frequency changer. and since it is also handling the carrier of the wanted station (converted to intermediate frequency). Whistles is that which gives of the tuning scale.f. These whistles arise from interaction between the oscillator or one of its harmonics and some received oscillation which differs by the i. and as the signal frequency becomes higher it becomes desirable to increase the intermediate frequency correspondingly. This oscillation will be accepted by the i. The first i. Another form of interference is obtained if two strong local stations differ in frequency by an amount equal to the intermediate frequency chosen. even though this second signal may be many hundred times stronger.f. one lower and the other higher. amplifier. which in turn will beat with the carrier of the station being received and will produce an audible beat. The only remedy is to improve the quality of the tuning prior to the frequency changer so that the An even more annoying form of interference rise to whistles at certain parts . Commercial receivers often use two frequency changers. the tuning circuit prior to the frequency changer. each of which will produce the required intermediate frequency. often associated with a radiofrequency amplifying stage to improve the signaltonoise ratio. Hence.e. no signal should reach the frequency changer at any frequency other than that required. the two oscillations will beat and produce an audible whistle. i. but in practice the preselection is by no means perfect. These will beat to produce an oscillation of the order of the i. of course.f. plus or minus some small amount.
or within a restricted band.f. A simple calculation of frequency differences. however. for a relatively short length of common earth lead will form a direct coupling from beginning to end of the chain and seriously reduce the selectivity. Ganging in Superheterodyne Receivers The usual practice is to tune all the signalfrequency circuits. For example. The choice of earthing points is important. together. it is not practicable to make an ideal choice and adequate preselection is the only remedy. then whistles can arise since the second detector is bound to produce some measure of harmonics of the intermediate frequency. while the oscillator must range from 1.f. including the oscillator. The oscillator circuit. signals will find their way to the signal input of the frequency changer still which (a) by direct pickup on the signal frequency circuit of the frequency changer (or the oscillator circuit) (b) by leakage from the aerial circuit to the frequency changer via stray capacitances or conductive paths which will shortcircuit the selecting action of the tuned circuits. The ganging of the bandpass and/or r. say. it is possible to choose the intermediate frequency so that the beats which would be produced are outside the audible range. 550 to 1.000 to 1. Another form of whistle arises from harmonics of the intermediate frequency. will soon show what interferences are likely to occur.950 kHz. THE RAblO RECEIVER residual voltage circuit is off 447 comes through the chain when the tune is very small.500 kHz. owing to the wide range and the different conditions under which it has to be used so that the "local" stations are different at different parts of the country. With receivers working on a fixed frequency. using an intermediate frequency of 450 kHz the medium. stages is simple and is effected by the methods described on page 374. and if any coupling exists between the second detector and the input of the set. Reducing the inductance of the oscillator coil will be necessary. but we are still . is more troublesome owing to the somewhat different frequency range to be covered.. Otherwise. In any case very careful screening and layout are essential on the stages prior to the mixer. as explained on page 417. taking into account not only fundamentals but harmonics up to the fifth. With a broadcast receiver.wave signal frequency range will be from. If reception is attempted at a frequency corresponding to a harmonic of the i.
448 left RADIO COMMUNICATION with the fact that the requisite range of the capacitor will be too great.500 kHz a capacitance range of 745 will be required. to 1. and a value can be so chosen as to give correct capacitance at top and bottom of the scale (though with a different oscillator inductance). of course. the intermediate "tracking" may . and <7„ are the trimmer and padder capacitances respectively. it will give 715 kHz at the maximum. As before. a combination of the two is used a third point being chosen in the middle of the scale at which it is arbitrarily assumed that the tracking is correct. whereas we require Thus to cover a range of 550 (l. This can be done by including a series or "padding" capacitor in the oscillator section of the gang capacitor which will reduce the effective maximum capacitance to a value sufficient to tune the particular oscillator inductance to 1. be wave range and are usually incorporated in the coil system and changed over by the switching which alters the coils. Such a capacitance will be fairly large and will thus have little effect on the minimum.C. Hence three equations can be written down in terms of L. at the top. If we make the oscillator inductance such that it will tune with the minimum capacitance to 1. . the padder and the trimmer. so that we can calculate the (a) (6) LG product The for the three frequencies.I(C.950 kHz.000 kHz. and C„. in practice.500/550) a = 1. One arrangement is shown in Fig. effective capacitance at each point is given by C^C + C.* altered for each The padding and trimming capacitances must. An alternative method is to add a large "trimming" capacitor in parallel with the oscillator section.26. Hence. C.000 kHz. + G t P) where C. 9. we The signal frequency to which the circuit is tuned. C The oscillator frequency at each of the three points must be the signal frequency plus the intermediate frequency in use. We then have three "spoton" frequencies and three variables the oscillator inductance. but we have no guarantee that the capacitance will be correct at intermediate points on the tuning scale. a little out. be incorrect and actually both methods produce errors in opposite directions. Clearly the capacitance range must be restricted. and we can find definite values for each which Elsewhere the tracking is still will fulfil the three conditions. but the method can be made to give very satisfactory — results. scale. middle and bottom of the note . The capacitance of the (oscillator) capacitor. while C is known. This again restricts the capacitance range. * For the three points chosen.
speciallyshaped vanes may be used on the oscillator section of the capacitor. Moreover the drift may continue almost unabated for some hours. . it is customary to arrange that the gain of the r. for example.f. but certain materials incorporating oxides of titanium have a negative coefficient of 006 to 007 per cent. padders and trimmers must still be used. trimmer Tracking Circuit fob Superheterodyne Oscillator can only be used with a specified i. Frequency Drift A most important practical consideration is that of frequency drift. shows a change of +02 per cent per degree C. An advantage of transistor operation is that the internally generated heat is usually negligible. synthetic Flutter As explained in the next section. so that the circuit is substantially free from frequency drift. Bakelized linen. M.W paddtr _ / Oscillator ^"^ tuning ^*«capacitor LW trimmer Fig. By allocating a suitable proportion of the circuit capacitance to this negative. This is due to change in the relative permittivity of the insulation of the timing capacitor. and on one waveband. 9. W. the technique being known as automatic gain (or volume) control. Ebonite and high grade compounds are subject to changes of the order of +002 per cent.THE RADIO RECEIVER 449 Alternatively. but such a capacitor M.26. As the parts of the receiver warm up. amplifier is controlled by the strength of the signal. Ordinary ceramic materials have a positive temperature coefficient between 001 and 002 per cent. the oscillator frequency tends to change and with bad design this drift is serious.coefficient material the positive coefficient not only of the remaining capacitance but also of the inductance in the circuit may be offset. necessitating constant retuning. which is made to have a smaller maximum capacitance than the other sections.f. For any other wave range.
If the frequencies are very close the interaction appears as a sub audible flutter and/or an interference between the wanted and unwanted modulations which is known as "sideband splash. causing alteration of the bias on the frequency changer (via the a. input. It is found in practice that.f.f. input due to modulation) but also from considerations of selectivity. If the frequency drifts the detector voltage falls.5. Diode Detectors The linear simplest form of detector is the diode. Fig. The r. 9.) output should be proportional to the r.f. removing the a. In this case an audible whistle appears in the detector output (together with the desired modulation). If this is above the audible limit the form of the modulation envelope is not seriously modified. This rejection does not occur if the carrier frequencies differ by an amount within the audible range. output may follow faithfully any variations in the r.e. and the tuning circuits prior to the detector must therefore be such that this requirement is met.g.g.c. Where the expense is justified they may be combated by rejector circuits tuned to the interfering station and/or the use of directional aerials." Such effects cannot be eliminated by normal tuning methods.f. If a second (weaker) signal is present. and the original state is resumed. so that the detector only responds to the modulink in A very important — lation of the desired (stronger) signal. voltage across the tuned circuit will produce a current through the diode which will charge capacitor C x and has a high rectification efficiency. not only on the score of freedom from distortion (so that the a.27 (a) .450 RADIO COMMUNICATION If this control voltage causes a change in the oscillator frequency an effect known as frequency flutter arises. The Detector Stage any radio receiver is the detector.c. The detector voltage rises again. a linear detector produces an output corresponding to the variations in amplitude of the modulation envelope. if the desired carrier has a mean amplitude more than twice that of the unwanted carrier.f. the detector ignores the interfering modulation completely. 9. which is practically shows a simple diode circuit. As explained in Chapter 3. the modulation envelope is subject to a superposed fluctuation determined mainly by the difference between the two carrier frequencies.) which pulls the frequency back. The detector voltage is thus in a continual state of fluctuation causing a disagreeable fluttering of the speech or music. This is particularly important for telephony reception. the rectified (a. This should normally obey a linear law i.
amplifier (or sometimes direct to the output stage) and this is usually done through an isolating capacitor G to permit the necessary bias to be applied .f. The a. 1l (a) Fig. . Fig.m. To the steady d.f.7). so that if the signal is modulated the e.27 (b) shows an alternative form of circuit in which the diode and the capacitor are interchanged.m. voltage with a superposed alternating component due to the modulation. In many cases t is a voltage divider. In either case the e. the impedance is simply R v The alternating component. however. across R t is a steady d. across x is equal to the peak amplitude of the r. 9. Typical values with a valve circuit would be 100 pF and 025 MO. ryi . \^L oatP ut o i. 451 to a voltage approximately equal to the peak value of the applied During the remainder of the cycle this charge will leak away through the resistor J? x until the next peak. 9.c. Hence Gl l should not be greater than 1/4/. The e. input. The value of t must not be too low or it will impose too much load on the input. Tig. The timeconstant G X X must be such x that the charge on the capacitor can vary at least as rapidly as the R R R highest modulation frequency.c.f. 1. while C1 must be large compared with Gd to ensure that the voltage drop on C a is negligible. has a parallel path through G and R R 2 with C 2 in parallel.f. which would handle up to 10 kHz. when the diode will again conduct and the charge will be restored. as in Fig. however. independently.m. 9. Assuming 80 per cent modulation this requires the capacitor to Cd AF . RF input o J.m. to provide a convenient control of the output. which requires a CR product of approximately 2 (cf. It is clear.f.THE RADIO RECEIVER e.27. and the circuit presents different impedances to the two components.f. With the lowerimpedance transistor circuitry. across follows the modulation.f. values of 001 jiF and 5 kO R R would be suitable. that these output arrangements modify the behaviour of the circuit. (b) Simple Diode Detector discharge to 20 per cent of its peak voltage in a time of one halfcycle. output across R t will be fed to the a.31.
c.c. If Z is the effective impedance of the circuit to a. corresponding to a mean (negative) potential V. Special networks have been suggested to obviate the difficulty.c.c. this peak value cannot exceed the d. The maximum modulation depth that Z is simply i? 1 R a /(J? 1 m = — + thus #2)R2KR1 At higher frequencies Z is less than the value just quoted so that the permissible modulation falls off still more. Hence we Z\RV e/Rlt whence m can write mejZ At low frequencies the capacitances exercise a negligible effect so R z ). By making R% three or four times R 1 we can approach full modulation. 9. although the a. Now. and the peak value of this will be mejZ. Excursions on either side of this will then cause the operating point to shift on to a corresponding characteristic. will produce relatively more current than the d. current. Then the is the depth of modulation component will be me sin pt where modulation and p is the modulation frequency. this being the ratio of the a. load. impedance is lower than R x the current is greater.m.c. and in fact the a. will be less than this as just explained. voltage across R x to the modulation voltage across the whole circuit. which cuts the zero axis before the full modulation is developed. so that distortion is liable to be introduced. corresponding to a load line such as XPY. A possible remedy is to apply a small positive bias to the diode as shown at Q. component. therefore. The operating conditions of a diode may be represented by characteristics of the form of Kg. but the usual practice is to make R 2 large and admit a small distortion on the peaks. The maximum modulation which can be handled without distortion can be arrived at as follows. With a signal of varying amplitude the mean operating point would be as at P.f. is . It is worth noting that up to the critical value the efficiency of rectification is very high because.27 will charge the capacitor C t to a voltage nearly equal to the peak value of the applied e. modulation voltage is the same as it would be if the circuit only contained the r). however. m— + + . load R v The a.28. 9.4S2 RADIO COMMUNICATION The alternating component. the modulation current will be (me/Z) sin pt. one resistance Rv The rectification efficiency is then R1 I(R1 where r is the diode resistance. Let the mean value of the capacitor voltage be e. The point P will be determined by drawing a load line OA having a slope corresponding to the d.c.c. but some distortion is unavoidable. The application of a signal to the circuits of Fig. as shown. which is sometimes done in highfidelity equipment. This has the effect of shifting the diode characteristic to the left so that a family of curves can be drawn for different values of input signal.c.
Dykamio Diode Charaotebistics Input Resistance The input resistance of a detector has an important bearing on the design of the circuit. special methods have to be adopted to estimate the input resistance. Since it is only a unidirectional conductor. load.c.28. Steady Voltage on Diode Fio. The power absorbed is then !». the small distortion present being accepted. S is where B x is the diode load resistance and /} is the rectification efficiency. The presence of a shunt resistance R v across a tuned . usually about 07 to 08 so that the effective input resistance is approximately §R V Design of Input Circuit This finite impedance must be allowed for in the design of the input circuit.THE RADIO RECEIVER 453 but more usually the circuitry is arranged so that the a.c. where the peak voltage and i mean the average diode current.„„„. Thus Power absorbed = ^jR 1 But & = 2E\ m ^ so that Power = = JE7 a RJ2P Hence the (j is effective input resistance Rxfifi. Now. A simple method is to assume that the diode only conducts on the peaks of the signal. load is not seriously lower than the d. 9.
Some improvement can be obtained by tapping the diode across part of =R = . R p would 160 kQ.29. Assuming Q 150. be about f resulting in an appreciable loss of gain and selectivity. Transistor Detector Stage With transistor operation there is an additional consideration.29. secondary having a Q of 150 are shown 9. In the case of a valve circuit with a diode resistance of 025 MQ. This appreciably improves where t is the stepdown and. cent increase in Qw . for small values of t. which is not the same as the criterion for maximum power gain. The requirement here is for maximum power in the diode circuit. namely the necessity to develop an adequate voltage across the diode. and the rectification efficiency fi is taken as 075. ratio. It will be seen that a stepdown of 15: 1 gives a 10 per cent increase in gain and 50 per n Pig. Consider a transistor having an output resistance hoe of 32kii operating from a 9V supply with a standing current of 25 mA.f. which for reasonable efficiency requires a signal input of at least 3 V. gives a slight improvement 15 20 t Step down ratio Fio. allowing for a small emitter .4S4 circuit reduces the RADIO COMMUNICATION Q in the ratio Q W IQB v \(Q Lm v ). 9. Q w would be 49. The maximum collector voltage swing. +R R= the coil. in which case Rv is replaced by Rj 2 Qw in overall gain. Q =150 Effect of Detector Tap to = 310" L = 750 aiH Typical results with an i. The gain is assumed to be proportional to Q w jt.
10 75 voltage Comparison of Typical Diode Characteristics it can supply sufficient current. diode load will be somewhat less than 3kO because of the shunting effect of the succeeding a. Values of 5 to 10 kQ are commonly available. which is inadequate.c. stage.f. = be satisfactory. The diode conducts when the applied e.30. The output would then be 163 X 10~ 3 X 310 3 x 07 == 343 V.THE RADIO RECEIVER bias.27) the form of the diode characteristic is irrelevant. which would least The 1mA.m. provided m . The a. With a diode efficiency of 70 per cent this will develop an output voltage of 082 x 10. and so follows the modulation. Squarelaw Detectors With a linear detector of the type so far considered (Fig. Assuming a diode load of 5 kO the effective impedance will be approximately twothirds of this. This would require a transformer ratio of ^(8/3) 163. the capacitor voltage is maintained at the instantaneous peak value of the carrier. The transformer ratio will then be = \/(S2jS) 327. 9.000 025 mA. as just explained. i. Hence it is desirable to follow the detector with a transistor having a relatively high input resistance. 455 would be ±8 V. Since a typical receiving diode will pass a current in excess of 50 A there is no difficulty in meeting this requirement. 9. = = permissible current swing in the transistor. 3k£2.3 x 310 3 X 07 = 172 V (peak). law 3 I JyI lL i/ /J' //I //sT s^oermaniu Thermionic 3 taw) / 2 ( 01 1 J OS Forward Fig. is at to obtain which the effective load should be 8kO. however. so that a primary current swing of 025 mA will produce a diode current swing of 025 X 327 082 mA.f.e. Hence the transformer ratio for a detector stage is generally only about half that for an amplifying stage. exceeds the voltage on the capacitor Clt 700  /Silicon 3 c Qj 30 10 3 ' " Exponential / . If the load is matched for maximum gain the current swing will be 8/32. the only reservation being that with the silicon so that.
if the signal and local oscillator e. At these frequencies the conventional form of frequency changer is inadequate but.f: stages &(01ftF) WW stages Rt (jMa) Automatic Gain Control Fig. 9. The d. 9. = Automatic Gain Control The range of signal strengths varies widely. signal. such usage thus providing the required i. * B Em Ih loA. However. so that the diode will not conduct appreciably until this voltage is exceed. A semiconductor ae 6*. and the subsequent circuit is arranged to respond to the actual diode current.f.f. component of the detector output to control the gain of the r. if the capacitor is omitted.f.f. stages. Hence it gain in inverse proportion to the signal.30. despite the fact that the characteristic obeys a nominal threehalves law.f. voltage across R . is One as a mixer at high frequencies in the microwave region as discussed in Chapter 13. but if the voltage at A is fed through a desirable to control the as in Fig. to be handled by a practical receiver Moreover the signal from any particular (distant) + earlier — To r.456 diodes RADIO COMMUNICATION 06 V.31. The time constant CR is chosen so that this voltage can follow the variations in amplitude due to the modulation. as commonly employed there is a hopoff voltage of the order of shown in Fig. the output can be arranged to be proportional to the square of the input voltage. This condition can be partially achieved by utilizing the d. r.c. the output contains a term proportional to the product of the two inputs at the difference frequency.m. signal input. 9.c. It was shown on page 249 that with small inputs a thermionic diode provides a squarelaw output. diode obeys an approximately exponential law of the form i but by similar reasoning it can be shown that for small inputs the output follows a square law. Basis or station is is liable to considerable variation due to fading. as was shown on page 439. and this possibility is utilized for certain special applications.f. is proportional and nearly equal to the peak value of the r.s are applied together to the input of a squarelaw detector.31.
is provided with a small permanent negative voltage so that rectification does not take place until the peak value of the applied signal has exceeded this delay voltage. Thus for a pnp type reverse a. Fig.f.f. In a valve receiver a negative control voltage is required which can be used to increase the bias on one or more of the r. amplifying valves and so reduce the gain. requires a positive control voltage. and so provide the necessary control.c.G. as explained in Chapter 5. It is important that the time constant of these a. 9. filter circuits should not be too large. feedback and hence possible instability. With transistor circuitry the diode may be arranged to provide either positive or negative control voltage according to the requirements. signal and will ignore the modulation.g. while an npn type will require a negative control voltage. the voltage at B will be simply proportional to the peak r.g. The operation is otherwise the same as already described.c. reducing the base voltage is methods being called forward by reducing the baseemitter voltage Control which operates by known as reverse a. a. In some cases it is desirable to arrange that the automatic gain control does not come into operation until a certain minimum strength of signal has been reached. . or As was explained in Chapter 6 (page 300) the gain of a transistor can be varied either by increasing the collector current. one of which provides the signal frequency voltages and the other. Sudden variations in strength. may momentarily paralyse the receiver because a large voltage is built up on the capacitors which takes an appreciable time to leak away.THE RADIO RECEIVER 457 second network E 1G 1 which has a long time constant.32 shows a circuit of this type.g. provides the a. Delayed A. is tied in parallel with the first.c.f.g.c. This is accomplished by using two diodes.C. The network R^! also effectively suppresses any r. component which would otherwise produce r. the alternative The polarity depends upon the type of transistor.c. which voltage. Otherwise there is an appreciable time lag between the variations in the carrier and the correcting action. This second diode. over a range of several hundred to one.f. however. such as are obtained when tuning in. By using variablemu valves a smooth control of the gain can be obtained without distortion. The direction of the control voltage depends upon the connection of the diode..g.
t W\A input T J. A.33 show.000 10. as the curves in Fig. Plain A. Circuit fob Providing Delayed A. If plain a.000. without a. delayed a.458 RADIO COMMUNICATION Signal diode XControl diode P\ R. It will be noted that the gain on weak signals is appreciably reduced.e.F. Showing the Effect of Automatic Gain Control on Receiver Output once and cannot therefore be allowed to be too fierce in its operation or the output would never reach its normal value. giving an almost level output and a much better control ratio.c.g. this becomes operative at The aim of the designer large control voltage for No " a. Control Ratio is to obtain a good control ratio.33.000 Voltage input to aerial (fiV) Fio.C.g. i.000 1. The full curve represents the output voltage in terms of input.G.g.g. is used. To some extent.gc.c.c. automatically provides an improvement.. 1. Delayed a.g.F. 9. The third curve shows delayed a. which does not operate until the output is nearly up to the full amount and then comes into play with full action.c. 9.32.g.f. .c.000 100. The chain dotted curve shows a typical performance without delay.c. ' output To r. 9. a a small change in signal strength. Delay voltage stages Fig.
automatically cuts down the source of the a.c. 4 volts delay.34.C.c. Fio. say. as already explained.c. 9. Thus with a delayed system requiring 5 volts radio frequency to load the detector.g. so that it is impossible to obtain a truly level characteristic.c. of course.c. Quiet A. the action of the a. but it is more usual to employ a separate muting diode for the purPig.G.g. diode. stages) may be accomplished by suitable circuit arrangements of the normal network.34 shows such a circuit.g. voltage.c.g. amplifier is used for the a.c. 9. the signal diode is released and the circuit functions normally. The process is known as muting. This amplifier is not controlled by the a. With the normal system.THE RADIO RECEIVER 459 It may be that the voltage at which the detector delivers full output is insufficient to give a sharp enough control. One method is to overbias the output stage and allow the +H. Another method is to put negative bias on the signal diode so that it normally does not operate. which would be quite inadequate.f. . is left operating normally with a small delay if necessary. The release of the signal diode (or the a.f. however.g. As a development of this system the circuit is sometimes arranged to give no audible output until the carrier reaches a predetermined value. voltage itself.c. we could use.g. Vx is a doublediodetriode. In commercial practice a separate i. so that the full increase of signal is allowed to develop a.c. circuit to "knock off" the bias when the carrier voltage exceeds the delay voltage.g.g.T. The a. supplied from some point subsequent to the frequency changer.g. On the arrival of a signal sufficiently strong to operate the a. Muting Circuit fob Interstation Noise Suppression a. This gives only 14 volts to provide a.
This is because the normal increase in strength as the tuning point is approached is counteracted by the automatic gain control coming into operation. knocking off the bias on the signal diode.c. It is. Alternatively. principles of a.460 RADIO COMMUNICATION the first diode being used for a.C. but a satisfactory measure of control is obtainable by altering the base potential.G.g. or cathoderay tuning indicator. to avoid which tuning indicators are sometimes fitted. As the bias on the r.c. possible for an inexpert user to adjust the receiver quite incorrectly.g. are equally applicable to transistor cirThe gain of a transistor r. negative potential is applied to the grid of V 2 The anode current R . Transistor A. stage cannot be controlled with the same smoothness as is possible with a variablemu valve. at the tuning point) the anode current reaches its lowest value. and will provide a control range of some 30 dB (316 times). Alternatively a magic eye. and at the point where the carrier is a maximum (i.c. but the leak in this case is returned to earth so that it is held negative to the cathode of Y x by the voltage drops on B t and R 2 V 2 is a highcurrent highslope valve.e. may be used in which the height or spread of a fluorescent glow is controlled by the carrier voltage and so indicates the correct tuning point.f. and therefore made negative to the cathode by the drop in the resistor v which provides bias for the triode section of the valve in the usual way.g. Tuning Indicators In a receiver fitted with automatic gain control.35 illustrates the collector current. therefore. the base potential . potential. As soon as the a. Hence. Fig. The second (signal) diode is coupled through a small capacitance to isolate it from the first as regards d. the strength of the signal does not vary appreciably as the station is tuned in. diode operates.g. but has the disadvantage of reducing the signalhandling capacity of the stage. a meter in the anode circuit can be used to indicate the exact tuning point. valves is run back the anode current decreases. rapidly decreases. As said earlier. 9. may be increased. which again and if a suitable circuit.c. This increases the collector circuit includes a suitable (decoupled) resistor the collector voltage is reduced. the control may be arranged to decrease the baseemitter potential.c.f. which thus comes into operation. This is called reverse a. Circuits The cuitry. .
operating on different stages. and has the advantage that the signalhandling capacity is increased.C. The connections are such that the d.THE RADIO RECEIVER 461 reduces the gain as explained in Chapter 6 (page 300). A.c.f.G. input C. stages Fio.c. as used with a damping From frequency changer Fig. D D . This is called forward a. voltage provided by x reduces the base (and . When a signal is present. duction of a damping diode across the shown in Fig.35. Circuit with Damping Diode diode Z) 2 across it. 9. Transistor Amplifier with A. sometimes both are used together. however.g. 9.g.c.C.36. 9. Damping Diode The control ratio may be substantially increased by the introfirst is i. potential across this diode is the difference between the emitter bias on Tr l and a suitable portion of that on Tr 2 which in the absence of a signal provides a reverse bias on 2 so that it is nonconducting and does not damp the tuned winding. Either method may be used according to circumstances. the a. A separate winding transformer. while the range of control is rather greater.36.G.
(automatic frequency control) or a. can be largely counteracted. 9. 780 (1959).c. 9. stage is reduced. Two i.c. 9. but reference may be made to a system providing a linear control over a range of 90 dB which was described by Willis and Richardson Proc. If the frequency of this can be automatically adjusted to give correct tuning the effects of slightly incorrect setting and. and has the advantage that the effective load of the first i.g.f. and muting. as in valve technique. p. with and without the damping diode. This result is accomplished in the first place by using a discriminator circuit as shown on the left of Fig.t. which enables it to handle larger inputs. Fig. damping _L _1_ 40 20 60 Relative signal input(dB) Fig.E.37.38.c.f.C. The system is principally applied to superheterodyne receivers where the fine control of the tuning a function of one circuit the oscillator.g. The method is known as a. transformers are used.c.E.462 emitter) voltage RADIO COMMUNICATION on Trx until at the predetermined level the diode begins to conduct and damps the tuned winding to an increasing extent. 3 3 With diode . The arrangement thus provides a form of delayed a.f.g.g. 106B. Supplements 1518. I. to control the tuning of a receiver.c. These are beyond the scope of the present discussion. frequency drift.37 illustrates the D 2 effect of the a.c. more important. sharply tuned to peak a few kHz above and below the true is A — circuit . Automatic Frequency Control development of some importance is the application of a. § 10 •_ S*~Simple Ag. (automatic tuning control).. Illustrating Effect or Damping Diode Various more sophisticated forms of control have been devised providing amplified a.
and no "control" voltage will be developed. An equivalent arrangement using transistor circuitry is shown in Fig. both 2\ and Ta will develop the same voltage. Consequently R1 will develop more voltage than R 2 . A positive control voltage will cause the gain to increase. Discriminator Control valve Oscillator Fig. The effective gridcathode capacitance of the valve will change accordingly. by suitable choice of anode load. an inverted Miller effect may be used to vary the effective anodecathode reactance of a valve. 9. a similar but opposite control voltage will be produced. the higher tuned transformer. as in the frequencymodulated circuits discussed on page 352.38 the control is applied to the suppressor grid which is an effective method of altering the gain. will develop more voltage. 9.THE RADIO RECEIVER 463 tuning point. Discriminator Circuit fob Obtaining Automatic Frequency Control and there will be a resultant voltage applied to the control valve. and as this is in parallel with the oscillator tuning circuit the oscillator frequency will be modified in such a direction as to counteract the drift. In Fig. This control voltage may be utilized in various ways. The currents in the diode loads R and R will be x 2 equal and opposite and will offset one another.f to rise. If the oscillator setting is correct so that the intermediate frequency generated lies accurately in between these two tunes. 9. If the oscillator drifts in the opposite direction.38.39. to produce this result. It uses a slightly different form of discriminator (actually of the FosterSeeley type described in the next section). while a negative voltage will cause a decrease. but the . each being partially off tune. The effective gain of the control valve must be arranged. say Tx . while T will 2 develop less. The most usual is to employ the Miller effect of a valve. Alternatively. which depends on the effective amplification. If the oscillator now drifts causing the i.
F. namely that any deviation in frequency causes a variation in the d. on the other hand.f. At the detector stage.M. F. The a frequencymodulated signal is contained entirely in the variation of frequency. however. This output is applied to a varactor diode connected across the oscillator circuit of the mixer. actually an important part of the process.39. 9. Varactor diodes have the property. it is necessary to convert the variations of frequency of the carrier into corresponding variations in amplitude. however.i 464 principle is RADIO COMMUNICATION the same. This is not a practicable arrangement. output above or below the mean value. Transistor A. so that the output is dependent on the frequency. Receivers For the reception of frequencymodulated signals the technique prior to the detector is the same as for amplitude modulation. stage slightly. any This limiter is intelligence in .F. 9.C. of changing their capacitance in inverse x I. Circuit proportion to the voltage across them. 9. is almost entirely amplitudemodulated so that the received signal will be a mixture of the two.c. One obvious way of doing this is to mistune the last i. r Nl— X s/>// Oscillator input nm^— Discriminator Fig.6. If this signal is passed through a limiting stage which gives a constant output irrespective of the input (above a certain minimum level). At the same time it is customary to pass the signals through a limiter stage to remove any spurious amplitude variations which Li miters may have been introduced. for it requires critical setting and it is more usual to employ a discriminator circuit similar to that used for the automatic frequency control described earlier (Fig. Hence any deviation in frequency can be arranged to produce a correcting change in oscillator frequency. Atmospheric and local interference.38). as explained in Chapter 6.
f. pentode operated with low anode and screen voltages. as shown in Fig. It is a simple matter to arrange the conditions such that any signal greater than a given minimum produces no increase in output. distorted. 9. becoming flattened.41. 9. Under such conditions anode.current saturation occurs as the grid potential approaches zero. while if the grid runs appreciably negative the anode current is cut off. 465 amplitude modulation (and hence the interference) will be removed Such a limiter is easily provided by a shortbase r. These secondary circuits are tuned to frequencies slightly above and below the intermediate D/ser/m/notor 1 Output 1 <? T Limlttr Fig. which introduces certain secondary effects.40.THE RADIO RECEIVER leaving the frequency modulation unaffected. Discriminators The output from this limiter stage is then coupled to a double secondary circuit. of course.41. A corresponding transistor circuit would be designed to handle only a small input. any increase in signal causing bottoming and cutoff on the respective peaks. The rectified outputs are connected in opposition as shown in Fig.40. 9. which will be seen to produce a substantially linear Fief. Discriminator Curve . The waveform of the carrier is. 9. discussed later. Doubi»e tuned Discriminator frequency and the voltage developed across each is rectified with a diode.
Here a single centretapped secondary is used. 9. 9. so that the secondary voltage is 90° out of phase with the primary. Z2 is not resistive. 9. Phasedifference Discriminator IThis circuit is often called the FosterSeeley discriminator.c. Seeley. as the frequency deviates from . through a capacitor. as shown in Pig.42 Circuit diodes are the vector sum of the halfsecondary voltage plus the primary voltage in quadrature. Because of the symmetry of the system E d and E' a are (numerically) equal so that the d. from its originators D. 9. Radio Engre. voltages developed across B x and B\ cancel out. Vector Diagram of Fig. 289. W. The diode voltages are then not equal and there is a resultant signal.] is i u the voltage across the primary ja)Li v The secondary current i 2 is jcoMi 1 IZ 2 and the voltage be (m^ML 2 fZ 2 )i v Now. 9. Z 2 is resistive. 9. In fact.42. Foster and S. Inst. Proc. however.42. E.43. p. but the centre point is also fed with signal from the primary. so that the secondaiy voltage will not be exactly 90° out of phase but will be as shown in Fig.43 (a). 25. if the across the secondary is jcoL 2 i 2 secondary is tuned. and actually more efficient. A is somewhat simpler. Hence the voltages across the If the current in the primary will = (a) 0>) Fig.43 (b).466 RADIO COMMUNICATION variation of amplitude with change of frequency over the operative range. At any other frequency. Output Fig. arrangement the phasedifference discriminator shown in Fig.
A reduction in input causes an increase in Q which results in a corresponding tion.c. Because of its symmetrical form. but a difference output when these voltages differ. Ratio Detector to that of Fig. but has one of the diodes reversed. the d. This will be seen to be similar k Output Fig. R 2 is earthed. This form of discriminator has a higher output than the doubletuned type. that the a. rejection is improved. but this is compensated by other advanand the centre point of In particular the network R'JDR' % provides a reservoir action which tends to maintain the voltage across R z constant irrespective of changes in the input signal due to unwanted amplitude modulatages. 9. it tends to ignore amplitude variations. by reason of the frequency deviation. output an exactly similar manner to that of the doubletuned discriminator as in Fig. This causes a small phase shift which has the effect of reducing the difference voltage. The diode load network is shunted by the network G 2 2 (via the resistors R' 2 ) R The arrangement thus forms a bridge network giving zero output when the voltages across iJ x and B\ are equal. This action is assisted by the fact that an increase in input causes the diodes to draw more current which reduces the Q of the secondary.44. 9.THE RADIO RECEIVER varies in 467 the midfrequency to which the circuits are tuned.42. so reverse correction. is the ratio detector illustrated in Fig. and by choosing R' 2 suitably the variation due to change in amplitude can be substantially eliminated. 9.41.m. The difference output is only half that which is obtained with the arrangement of Fig. but in practice this still necessary to precede it with a limiter. 9. 9. is insufficient and it is Ratio Detector A modified form of circuit which is somewhat simpler to produce.42.44. and provides much better elimination of amplitude modulation. .
Locked Oscillator Discriminator A detector circuit which operates on a different principle is the arrangement known as the lockedoscillator discriminator illustrated in Fig.468 RADIO COMMUNICATION Since ii 2 provides a d. Lockedoscillator Discriminator operating at nearly identical frequencies the weaker oscillator will tend to synchronize with the stronger (the "pulling" effect referred to in Chapter 12). 9.f. In the circuit of Fig. 9. Because of its inherent of detector is stability now almost and high a. RC.c.m. Transistorized Ratio Detector the circuit will virtually disregard amplitude variations in the in coming signal.45.46. stage while the suppressor grid is connected to a second circuit tuned to the mean frequency of the incoming frequency modulated signal. The incoming signal modulates the electron . 9. 9.46 the signal grid is supplied from the final i. of the order of 01 sec Frequency modulated input Fig. 9. This is based on the fact that if two oscillators are II Output Incoming signal Fig. By making the time constant of the load. rejection this type universally adopted.45 shows a transistorized version of a ratio detector. Fig. discharge path for the capacitors Cfi'^ the resistors R t R\ may be omitted in a practical circuit.46.
Even when the interference is relatively small.* Preemphasis The use of a limiter to signal does not result in the elimination of all interference. Avina and T. p.THE RADIO RECEIVER 469 stream in the valve. therefore. Review. so that if the interference is greater than the signal the true signal is swamped. This means that the higher the (audio) frequency of the noise the greater the amplitude at which it is reproduced. This is not important in itself but it can introduce a certain amount of spurious frequency modulation as a byproduct to an extent which depends upon the strength of the interference and its frequency separation from the wanted signal. however.m. a direct relation between the spurious frequency deviation produced by the interference and the frequency separation. but it copes adequately with the deviations in a practical f. it cannot be entirely eliminated by limiting because. transmission. however. Since two oscillators will only lock over a limited range the circuit is only effective for small deviations. although the amplitude variations are levelled out. as previously mentioned in discussing f. remove any amplitude modulation in the For one thing the capture effect already referred to in the previous section (page 450) still applies. The effect is aggravated because in a normal modulation the amplitude of the higherfrequency components tends to be less than at the lower frequencies. If the frequency of the incoming signal. and the suppressor. Detector" by J. deviates from that to which L %C % is tuned the phase angle of the suppressor current is no longer exactly 90° but becomes greater or less depending on the direction of the deviation. The suppressor current (= dq/dt) will thus vary in quadrature with the input and so provides a negative reactance effect which maintains the circuit L 2C 2 in oscillation. transmitters * For further details refer to "A Locked Oscillator Quadrature Grid F. 648.M. There is.C. in fact. To counteract this effect. the interference will remain as a phase modulation of the frequencymodulated signal. it is customary to accentuate the high frequencies in the modulation at the transmitter. Brady. 16. The modulation is passed through a preemphasis network. and since the anode current is controlled by the vector sum of the grid and suppressor voltages there will be a variation in the anode current which is directly proportional to the frequency deviation. B. . being in this stream. acquires a charge which varies accordingly. A.m.
which begins to operate at about 1 . However. They are usually designed to have a time constant of the order of 100 fis.f.47. If Cre 6 is R^ = This is 1/(4tt X 107 less X 10" X 10~ 12 X 35 X 10. page 111). cannot be matched to the transistors in the normal way because of the greater internal feedback arising from the higher frequency involved. 9. it was shown on page 455 that. usually immediately following the discriminator. 9. aCrfi^yR^ 05. if p is 15 kHz. With critical coupling this requires the circuit Q to hef /B 107/02 53 (cf. B 180 kHz. where A is the frequency deviation and p is the highest modulation frequency. than the product R^R^ for the transistor is not possible. stages are usually employed. work I. Deemphasis Fbeqtjencycorrecting Networks (page 355).470 RADIO COMMUNICATION Input 1 input Output sir Output Pre emphasis Fig.m. Amplifier is The bandwidth of a frequencymodulated transmission greater than with a. These networks are simple RC circuits as shown in Fig. then with Om = 35 mA/V. The feedback can be minimized by using ^l^a = = = — — transistors having a specially taken as 02 pF. however. using doubletuned transformers to preserve the frequency response over the required bandwidth.m. This clearly demands a much higher intermediate frequency than the 470 kHz used with a. and for certain practical reasons a frequency of 107 MHz has become standard. so that. much being approximately 2(A+#). To obtain the necessary selectivity three i. It was shown earlier (page 429) that with a stability factor of 4. transmissions.000 Hz and produces an accentuation (or attenuation) of 20 dB at about 20 kHz. and in practice a figure of 200 kHz is usually allowed.ec.F.. to obtain maximum output from a typical very much so that normal matching . The usual deviation is 75 kHz. while at the receiver a corresponding deemphasis netis included.47. X 02 low internal capacitance. The circuits. so that l/^wCre^m.s ) ~ 10 itself.
R x can be 8 kQ. 9. Cx must thus be 0(1 + w)/«. . 9.THE RADIO RECEIVER 471 transistor.10 3/(2t7 X 107 X 10* X 53) (=Qw I<dZ w)~ 100 VF. since an inductance of the order of 2 fiEL would only require a ~ *c2 Fig. However. If n is the stepdown ratio.48. which with the values above makes C t = 1125 pF andC2 = 900 pF. Amplipibb Stage few turns. The secondary stepdown ratio will be \/(8. 107 MHz I.000/125) = 8. as is more convenient to use a capacitive tapping and shown in Fig.F. The stability requirement can then be met by making R 2 = 125 Q.48. and C ZJwQ w = 8. C 2 = nC v while G x and C 2 in series must equal the tuning capacitance C. it it is difficult to arrange the correct tapping point. To provide a load Zw = 8 kQ the inductance required will be 225 [M.
The early stages are concerned with voltage amplification. with appropriate modifications. technique which any type of amplifier. 10. which is approximately gZ where g is the effective transconductance of the device and Z is the load. with suitable crossreferences to factors which are common to both.f.f. Strictly speaking. one important aspect of a. and results in the transfer of some part of is. amplifiers are usually required to handle equally effectively all the frequencies within the required spectrum. which is currentoperated.f. audiofrequency signals are those which fall within the range of audible sound (approximately 16 to 16.10 Audiofrequency Technique Audiofrequency technique is relatively lowfrequency signals which do not concerned with the handling of fall within the radio frequency range already discussed.000 Hz).) and the supersonic range up to 100 kHz or more. and r. Feedback There however. requires a different approach to a bipolar transistor. being voltageoperated. but is still subject to the same basic requirements. technique is used. The later stages are concerned with the delivery of power to the reproducing device in use. applies to 472 . so that the two systems are best discussed separately. but the techniques are applicable. A valve. whereas r. This may be accidental or deliberate. a.1. to signals in the subaudio range down to zero frequency (d. which requires a slightly different approach. amplifiers are concerned with a single (adjustable) frequency. The interpretation of the requirements depends on the type of amplifying device. The main difference between a.f.f. namely feedback. circuits is that.f.f. for which a blend of a. while at the upper end of the spectrum are the video frequencies up to some 6 MHz involved in television practice. and r.c. so that the design is concerned with the provision of a load which is either constant or varies in a specified manner over the range of frequency to be handled.
which and negative feedback which will reduce the effective gain. It can also occur as a result of e. in general.1. will be modified to V '.iV = AI(lpA) t the fundamental feedback relationship. Their values depend upon the circuit configuration and are.f. The factor /? is called the reverse transfer ratio. The effect has been discussed in relation to r.f.m. Feedback may also occur through the internal capacitance of the but this is not usually troublesome at audio frequencies. 10. and it is a simple matter to determine whether it is likely to be significant in any particular a.1. the effective input voltage becomes Vt + fiV The output voltage . we can write Hence the modified gain A'=V. input is V u the normal output V will be A V t If now a proportion /? of the output is fed back to the input. This . Amplitibb with Voltage Feedback . of the feedback increases the gain (and 473 the output energy to the input.'. stage. circuits in Chapter 9. = A{Vt + fiV is is ') so that AV = t (1 —0A)V. developed across any impedances in the supply lines which are common to several stages. either around an individual stage The result depends on the relative phase and may vary between positive feedback may result in oscillation). Basic Feedback Conditions Let us deduce a generalized expression for an amplifier with feedback.AUDIOFREQUENCY TECHNIQUE or the amplifier as a whole. circuit. 10. frequency dependent.f. as shown in Fig. If the amplifier gain is A and the signal Vi Amplifier % t Load Input Vf Feedback network Fio. Accidental feedback may occur through stray coupling (usually capacitive) between output and input wiring or components so that attention to layout is important. so V. and /L4 is called the feedback factor. This was referred to in the previous Chapter (page 416) and may be avoided by the provision of a bypass capacitor to earth having a negligible impedance at the signal frequency.
In fact. the magnitude and phase of the feedback factor may be determined over any desired frequency range. the modified im pedance B ' becomes jR /(1 + ftA). but tends to a limit 1//5. Effect of Negative Feedback on Circuit Impedances Voltage feedback tends to maintain the output voltage constant. the actual feedback voltage appearing at the input is determined by the relative circuit impedances. are obtained at the expense of overall gain since to obtain a given output the input signal has to be appropriately increased. which is equivalent to reducing the effective internal impedance. providing considerably improved stability. conditions established. The arrangement also reduces distortion since if there is initially a harmonic component FA in the output. The feedback network is often made deliberately frequencydependent to provide a specific form of amplifier response.474 If /? RADIO COMMUNICATION is positive (so that the feedback voltage is is in phase with the overall gain increased. It may be applied either in series with the input signal (series feedback) or in parallel therewith (shunt feedback). is V t) oscillation will result. A' = . of by the feedback by the same factor (1 course. The feedback may be a suitable proportion of the output voltage (voltage feedback) or it may be made proportional to the output current {current feedback). In practice the implementation of the technique involves a number of subsidiary factors. is and if (1 — /£4) is greater than unity. and the stability + . Both A and ft are subject to phase change with varying frequency. thus to be avoided unless deliberate oscillation Negative Feedback If j8 is negative. Current feedback. Finally. In fact. which is independent of A. These advantages. Such a condition required. conditions are reversed. By plotting the vector of ftA against phase angle in what is called a Nyquist diagram. which is equivalent to an increase in output resistance. if R is the normal output impedance. and if the overall change is more than 180° the sign of the term f}A will be reversed and positive feedback will result.4/(1 + pA) = 1//9 if pA > 1 This means that the overall performance is virtually independent of the amplifier parameters. tends to maintain the output current constant. on the other hand. producing a tendency to instability. so that the feedback factor flA has to be assessed individually. In this case B ' R (l pA). = + . The overall gain is reduced. this will be reduced pA).
V current through CB lt the proportion of V developed across Bx is Bi («! + = llJCco) VW + Bi 1/C 2 «> 2 ) . Here the load a resistance B which is constructed (usually of carbon composition) so that it is substantially noninductive over the frequency range involved.2.2. 10. in which valve and transistor operation will be conseries or = + = + sidered separately.f. where next stage is that across B lt so that the gain is v[il(B v B B R + B„ ~B = BBJ(R + B if B^ B . Mr Fig. however. the resistancecoupled stage is illustrated in Fig.AUDIOFREQUENCY TECHNIQUE There whether is 475 a similar influence on the input impedance depending on shunt feedback is employed. A. Where the feedback (of either voltage or current form) is introduced in series with the signal. The practical applications of feedback will be seen in the sections which follow. Its effective value. Resistance coupied Stage reactance of C is negligibly high by comparison with B and the reactance of C negligibly low compared with R v The voltage fieg passes current through the anode resistance r in series with and 1 in parallel. Valve Amplifiers amplifier is The simplest form of a.(l /Li). 10. t) At low frequencies the reactance of G is not negligible and the If i is the voltage Vt is thus only a part of the voltage developed. Consider first the gain at a medium frequency such that the .F.2. the effective input impedance becomes B/ i?. 10. The voltage transferred to the r). With shunt feedback the input resistance is reduced becoming R/ BJ(1 PA). is modified at high frequencies by the stray circuit capacitance represented by G while at low frequencies the performance is affected by the network OBx through which the signal is transferred to the succeeding stage.
(The higher a> or G the less R must be. product GR X to be 00065 which makes VJV. An alternative criterion is R^Cm 2. (It 'will be clear that the fundamental requirement is that the reactance of G shall be low compared with R x and so we can either increase C or R v Thus it is the product of the two which is important.) Miller Effect. The amplification is then reduced by the lowering of the effective impedance. and this soon limits the C includes the anodecathode capacitance of the valve in question.Z) taking due account of the vectorial nature of Z. for the input capacitance of a valve is not merely its static or geometric capacitance but is increased by the Miller effect. of 30 and cos 6 08.500 ohms. we have Gg 7 182 pF. and three stages would give a lead of 774 degrees.) = = Phase Angle This criterion is by no means adequate for highquality ampli where several stages are in use. The loss in each stage is cumulative. particularly Highfrequency Loss At the upper frequencies the capacitance G becomes troublesome and shunts the resistance R. effectiveness of the circuit. is the change of phase which occurs.. however.476 RADIO COMMUNICATION It is easy to calculate the relative values of C and B 1 to make Pi/^o a given percentage. fication. 10. as will be clear from Fig. which would have a reactance 7(1 24) at 10 kHz of only 87. the value of R necessary to maintain 90 per cent of the full amplification is given by RC m = J. From = + = + + = = + = . There are many circuits in which correct phase relationship is of vital importance. sothat with three stages VJV would only be073.2. and the input capacitance of the succeeding valve. the stray circuit capacitances. A. R and G in parallel provide an impedance Z = R/(l +jRG m) and the gain is fiZI(r \. Use of Screened Valves the formula quoted in the previous chapter the input capacitance Cg Cge Cga (l A cos 0). If GR = 00065 this angle is 258 degrees. and hence the voltage V lt is by no means in phase with V but leads by an angle tan 1 (l/RGco). In such circumstances the product CR must be increased to reduce the phase angle to the required value. Assuming an effective gain. for the current /. which provides 90 per cent of the full gain at frequency <w/27r. This latter capacitance may entirely swamp the previous ones. and a criterion often quoted is for the 09 at 50 Hz. with Cge and Gga 7 pF each. Under given conditions of C and co. A still more serious defect.
6.000 ohms to be used without undue loss of highfrequency response. stage. must be kept low. This permits anode loads of 50. which means a limited gain per The remedy is to use pentode valves which have gridanode R pF or less. the phase shift per stage. the Resistance = — = .000 to 100. 10. This tunes with the stray capacitance and so maintains a high effective anode impedance at the upper frequencies. from which it High Frequencies in can foe shown that if L C R 2 the response 8 ' S rises to y/2 times the normal at the frequency coupled Ai^lctier may be Minimized by for which co 2 L C 1 and then falls off sharply. Once again the effect is cumulative and if the phase shift at the end of an wstage amplifier is not to exceed a degrees.3. capacitances. It is possible to maintain the upper frequencies up to a point by including a choke in series with the resistance as shown in Fig. and this is the current >j "»' »/£ by the valve. 10. however. Wiring an Induc jj l Q 5 GR2 the response i s practically r ° tanch in Series with . . gains of several hundred are easily obtained with modern valves. must not exceed oc/w. Hence V Flo 10 3 Vkctor Diagram o* on jmb by an angle tan1 RC a>.2 Circuit at High Frequencies giving a phase shift which increases with R. This supplied leads HJ angle 6 is the phase shift in the Miller effect formulae already quoted. Loss ox usual circuit laws (see page 97). Highfrequency Phase Displacement Phase as will shift occurs at for a similar reason. The effective impedance of the combination can be calculated from the Fig. Miller effect with such valves is though the static capacitance Cgc still remains. 10.4. The total current through jR qJ and C in parallel.t < j as shown. Fig. . current through R is in phase with / V V . G and w.AUDIOFREQUENCY TECHNIQUE Hence. if effective 477 is amplification of the upper frequencies to be maintained.4. level up to the resonant point. . The high frequency as well as at low frequencies across R results in a lag. leads on V^°\ z*r '"'"^ ^_^ . plus the anodecathode capacitance of the preceding valve and the straycapacitances of 001 negligible.. 10. The presence of C be seen from Fig. and since with a pentode the stage gain is gR.
10.478 RADIO COMMUNICATION Direct Coupling Where very low amplifier frequencies are to be handled. the bottom end being connected +240V to a point of negative potential. a directcoupled may be used. down to zero (d. In the former connection.5.t.t. It is also liable to drift. though sometimes a deliberate phase shift is permitted. but it is usually easy to obtain Fig. 10. it Other Several stages may be used in cascade to produce increased overall The individual stages are designed as already described but is important to ensure that (a) The overall phase shift is tolerable. forms of circuit are discussed in Chapter 17. The method has the advantage that all the cathodes may be at the same potential is there (and all earthy). By suitable choice of iJj and a B . to avoid the heavy positive grid voltage which would otherwise result. An alternative arrangement is shown in Fig. but it suffers from the same disadvantages as the ordinary resistancecoupled amplifier at the high frequencies. Here the anode of the first valve is connected either directly or through a resistance to the grid of the next valve.5. (See page 416.c). particularly that arising from coupling in the h. as in the phaseshift oscillator described in Chapter 12. This method is of limited application since it involves an increase in the overall h. the cathode of the second valve must also be raised to a slightly more positive potential. 10. since small changes of current through the network A directcoupled amplifier will respond to frequencies may upset the somewhat delicate balance of the voltages.^i ^2) so that . Normally this will be kept small. Here the voltage developed at the anode of Vx is applied across resistances R x and R 2 in series. the potential of the point A may be made slightly negative with respect to the cathode of V 2 The signal voltage is divided in the ratio B 2 l(.6 common impedance . voltage. Directcoupled Amplifier more gain from the valve than is actually needed. line. Commonimpedance coupling has been eliminated by the decoupling circuit in the anode circuit of the Fig. Multistage Amplifiers gain.) shows a twostage amplifier in which deliberate negative feedback has been introduced. (6) Stray feedback is avoided. + appreciable loss of gain. therefore.
Fig.m. + (* + ««)]. . results from the omission of the usual bypass capacitor can be utilized in various ways. Twostage Amplifier with Negative Feedback transferred to the cathode of V t which will thus move in opposition to the gridcathode signal input. HX 12. When the grid of V x goes positive the voltage at the anode of F 2 will also go positive. with consequent reduction + in gain (and. Consider the circuit of Fig. 10. while a further feedback is provided from output to input.AUDIOFREQUENCY TECHNIQUE first stage. "R Stage gain = a RI(e — a R = i g i c) R. and some part of this e.000a Output I2. If i a is the a. this reduces to /x)] A= This is (iRI[ra + R + R (l + c denominator the same as the normal expression except that R in the is increased by R c (l (i). v in #/[(e e /«'a ) = e„ — a R —R i c] e Substituting the expression above for ia .6.7 (a).= Wlr.200 so that the gain 1//? will be 61.f.c. 10. improvement in linearity). component of the anode Cathode Feedback The negative feedback which current. At medium frequencies (3 is = 200/12.000fi. we can write ia ». possibly. 479 A measure of individual negative feedback has been provided in each stage by the omission of bypass capacitors across the cathode resistors.
and an appreciable fraction of the output voltage appears across the cathode resistor.t. But since the anode is already negative with respect to earth (h. there is a further loss of output because the anode resistance is only a fraction of the total external load. just short of unity be seen that the maximum gain is and may not be more than a . It will be clear that an increased anode current will cause an increased voltage drop on both anode and cathode resistors. this A= and in the limit. +) it will — become still more negative.T.7 (b). as in Fig.480 RADIO COMMUNICATION resistor be increased beyond the value required tapped across a portion only.(l + /*)] Cathodefollower Circuit If we reduce the anode resistance to zero. Development op the Cathodefouower Circuit voltages in opposition one from the anode and the other from the cathode. 10.(l + (i)] A= R (l c f 1). This effect is sometimes utilized to obtain two output for bias.7 reduces to (c).+ H.+ The cathode may HX+ Fia. 10. Under such conditions. as in Fig.T. so that anode and cathode (signal) potentials The voltage move in opposite directions. across the cathode resistor is i a R e so that the gain at the cathode becomes A= fiRc l[r„ + R+ B. the grid leak being H.7. apart from the reduction in gain due to feedback. 10. nl(n Such a circuit + much greater than ra we have simply It will is called a cathode follower. The cathode is positive to earth and will increase its positive potential. if fiR e l[ra /*) is + B.
A. If.F. operates not as an amplifier but as an impedance changer. on the other hand. Transformers Transformer coupling is used in the simpler types of circuit.1) which. in fact. the cathode resistance may be in the form at a voltage divider of quite low resistance such that the capacitance of the succeeding stage does not affect the operation. which is often of great convenience. The ordinary transformer possesses appreciable leakage inductance.AUDIOFREQUENCY TECHNIQUE fraction if 481 BC {1 + fi) is of the same order as ra . and the when yu^>l. a highresistance voltage divider cannot be used to feed the grid of a valve if the frequency is high because the parallel reactance of the valve input and stray capacitance mil be small. The upper frequencies with a transformer are entirely limited by the selfcapacitance in the circuit. the input is applied to the grid of a cathode follower. ra j/i + 1). This may be represented by showing a perfect transformer in series with a small leakage inductance to . By proper design a resonant effect may again be called into play to assist in maintaining a good response curve. becomes simply = ljgm . The arrangement. however. Miller effect is not present. Amplification of the bass frequencies is here dependent on maintaining a high effective primary impedance This involves using a transformer with a high inductance obtained either by using a large iron circuit (so that the saturation introduced by the heavy anode current shall not be serious) of a highpermeability material with a parallelrelative to the valve resistance. including the effective gridcathode capacitance of the succeeding valve. the magnetic field of the primary being incompletely linked with the secondary. The input impedance to the cathode follower. is very high being simply the gridcathode capacitance (including strays). For example. usually following triode valves. The expression for gain may be rewritten in the form A _ fi I* R . +l ra l(p+l) +R e Thus the effective internal effective amplification factor is fij(fi impedance is ra [(ju f. The circuit is also often used as a buffer stage to prevent one circuit from interacting with another. feed circuit. This largely shortcircuits the bottom portion of the voltage divider so that it ceases to provide adequate control.
000 Hz and thus to maintain the amplification of the upper frequencies. Fig.8. The second is between the leakage inductance I and the self. so that a relatively large iron circuit is required.T. and it will be seen that there are two possibilities of Fig. under which conditions the primary inductance acts as an infinite impedance and may be ignored. therefore.9.000 inductance this second resonance can be made to occur between and 10. Autotransformer Arrangement 5. The capacitance G can be arranged to resonate with L x at around .c.capacitance G which occurs at a high frequency. This permits the use of small and compact transformers on highpermeability cores. To avoid this a parallelfeed circuit may be used as shown in Fig.capacitance G in the circuit. H. The equivalent circuit of a transformer. Equivalent Circuit of Audio frequency Transformer The first of these occurs between the main primary inductance L and the self. and is usually arranged to fall between 50 and 100 Hz in order to maintain good bass response. is as shown in Fig. in which case we can represent the transformer itself by a simple choke. By ordinary transformer laws the quantities may all be referred to the primary. The presence of the steady d.482 RADIO COMMUNICATION represent that portion of the magnetic flux which does not link with the secondary.8. 10. 10. 10. the primary being conveniently (but not necessarily) in the form of a tapping on the secondary.9. By suitable design of the leakage resonance. component of the anode current reduces the effective inductance. 10.
Fig. In Fig. giving a definite rise in the amplification. and the distances OA and OA' should be equal if the load point is represented by AA' is correct. 10. Actually they are far from equal. and it will be The working chosen in this case corresponding to 200 volts on the anode and 12 J volts grid bias.10. showing that considerable distortion is occurring. 10. designed to work into the optimum load for maximum undistorted output. 10. Illustrating the Estect op Winding Resonances The Output Stage The choice of the valves to be used in the amplifier depends on the voltages to be handled.11 shows typical characteristics for a triode. first of all. Fig. The methods of determining this vary with the circumstances and the degree of accuracy required.10 illustrates the effect. 10. assuming the valve to be fully limits of grid swing will be loaded. so that the and 25. and for absolutely distortionless working the points where this load line cuts the limiting characteristics should correspond to equal excursions of anode voltage and current from the mean value. The variations of anode voltage and current may be determined by drawing a line through the point having a slope corresponding to the effective resistance of the load. the mean bias is 12J volts. It is usual to work back from the output stage which is. i ^ — K> & OlO> CD Q o S o oop O o o ooo o o Frequency (hertz) O) ^OiOj Co CJ o ca o cj <a o o «a o o s o o o Fig. for example. and from the ordinary laws of resonance it will be clear that the voltage across L t can exceed the input voltage supplied by the valve.11. is called second harmonic distortion because the form of wave is the same as would . This particular distortion. wherein one half of the wave is flattened and the other peaked. as well to review the process briefly.AUDIOFREQUENCY TECHNIQUE 483 100 Hz. The most general method is to adopt the graphical tactics outlined in Chapter 5.
484 result if a pure RADIO COMMUNICATION wave were mixed with a (smaller) wave of twice the frequency as shown in Fig.11. zoo Anode vo/ts 300 Illustrating Choice op Optimum Load fob a Tbiode The figures on the curves refer to the grid voltage. 10.12 has been shown displaced above the zero line for the fundamental. Hence the ratio of the positive to negative peaks is (F 2dV). Illustbating Flattening of Wavk due to Second Habmonic . allow for this the second harmonic in Fig. 10. 10. It will be seen that the positive peak of the combined wave is V 2dV and that of the negative wave is V 2dV. Hence if OA'/OA is (5 per cent distortion) this is 11/09 — + +  = = Composite Wave Fundamental 'AD. Now let V be the maximum amplitude of the fundamental and d V that of the harmonic.e. OA'/OA in Fig. 10. i. To 100 Fig. 10.11.12. It will be clear from Fig. 10.11 that a nonlinear operation such as this will be accompanied by an increase in the mean anode current (because the current swing OA' is greater than OA).12. 2dV)l(V If d 005 122.C T Component ofHarmonic Fig.
so that a more efficient valve results. Hence the power is .AUDIOFREQUENCY TECHNIQUE not greater than 122 distortion 485 we shall not introduce more than 5 per cent The this criterion is often used by circuit designers. so that the valve is operating less efficiently. load line A'OA does not fulfil this requirement. 10. line at 1100 V.) 300 Anode vb/ts Fig. m. It will be noted that the \XB . At the same time the screen current is appreciably reduced by aligning the screen and control grid. find with a triode that as rises to falls Hence.m. * XB . and criticaldistance or beam tetrodes.* does.s. XB'. but B'OB and anode current is no longer reduced to zero at each swing. current. we we increase the load the power output falls off again. value. Charactebistics of a Typical Otjtptjt Tetbode have the anode located farther from the screen and so disposed that the secondaryemission kink is removed. is the is peak voltage swing XB Similarly XB' XB'I(2^2)* = iXB 2\/2 = 2\/2 times the r.13. a maximum and then to a minimum and then Output Tetrodes and Pentodes similar though slightly modified procedure pentodes. while the distortion begins to rise again. A is adopted with latter valves The Clcuts 30V. times the r. slope of the load line the power output falls off still further while the distortion at first becomes less and then begins to increase again. The optimum distortion condition does not coincide with that for maximum power output. due to the tendency of the characteristics to As we reduce the become flatter with high grid bias. s. A steep load line corresponds to a low load resistance. while the power output (which is XB') is also less than before. .
and a typical set of characteristics is shown in Fig. These are shown in the table below. 10. Note that with third harmonic the wave is symmetrical. and 15 mA respectively for successive increments of 5 volts. 10.14. This. 10.14. 35. This will give a wave with a marked third harmonic. as shown in Fig. changes of 30. for AO OA'. for there is a marked thirdharmonic component which can be detected by noting the increments in anode current for equal increments of grid voltage. the valves behave similarly to pentodes. Illustrating Influence of Third Harmonic on Waveform . is not the best condition. 10. however. Table oi Anode Currents in Load Line Qrid Bias Fio. This is true of any odd harmonic. on each side. Fundamental Composite Wave & Harmonic Fig.486 RADIO COMMUNICATION In use.13. which shows the change in current per 5 volts grid change. The line AOA' represents a condition giving no second harmonic distortion at all.13 = A A' Increments Load Line BB' h 30 25 55 h 52 Increments 15 18 70 35 105 70 30 20 15 100 30 135 135 35 30 35 170 10 165  35 5 35 200 15 205 35 215 240 It will be seen then working from the middle (—15 V) we have.
so that we have very little third = harmonic. three parameters as under positive — negative peak current current Then evaluate peak B = 1 + a/0 C — 141/3/negative peak current .15. 10.15. if we reduce the load slightly to that shown by BOB' we introduce about 5 per cent second harmonic. but over most of the line equal increments of grid voltage give equal anode current changes. 10. As we increase the load the secondharmonic content falls to zero and then rises again (this time due to a flattening of the top of the curve instead of the bottom). f} = current swing for a negative grid swing of the same value. The third harmonic behaves similarly but o Fig. This behaviour is characteristic of pentodes and tetrodes. 10. First determine condition is It is difficult to locate the best load a = current swing for a positive grid swing of 071 times the peak grid swing. sooo 10000 15000 Anode Load (ohms) Illustrating the Effect of Load on Harmonics to zero much earlier and we usually find that at the point of zero second harmonic the third has risen alarmingly. The best falls between the two as shown in Fig. from the characteristics but curves similar to Fig.AUDIOFREQUENCY TECHNIQUE 487 On the other hand.15 are usually available from the makers. for OB 1 26 OB'. As an approximation the following procedure may be used.
a first a. voltage. Even if no damage results. giving rise to the same trouble if a large high (audio) frequency input is applied. with the actual effective anode voltage. but falls rapidly above a few thousand hertz thereby bringing the limiting resistance into play. sometimes with a capacitor in series. This causes the load line to swing round. X 100 Impedance Limiting It should be noted that tetrodes or pentodes should not be run with high loads. which may easily cause a breakdown of the insulation. If not. one a. amplifying stage is desirable. of providing without distortion an anode swing equal to the voltage required. rises rapidly in the upper frequencies. the capacitance being chosen such that its reactance is high at normal frequencies. Some commercial receivers use this system.f. however. To avoid this. valve must be employed capable of providing this voltageDue to the coupling impedance in the anode circuit the anode voltage will. Valves are made today which will give several watts out with only a few volts peak signal input. and shows that if the full 15 volts grid swing is applied the anode voltage will swing to a voltage of over one thousand. 10. This may be the detector itself if an amplifying detector is used.488 RADIO COMMUNICATION Then Percentage 2nd harmonic = = 1 A— . Choice of Valves Having decided upon the valve to obtain the required output one knows the grid input required. If sufficient voltage can be obtained direct from the detector this is all that is necessary. particularly if resistance coupling is used. very bad distortion will be produced. 1 R ~ X 100 R^.t. The line COG' (Fig. In general.13) represents such a load. and the valve must be capable. a limiting load is often connected in parallel. 4 Percentage 3rd harmonic ^ _ 5(7 . be less than the h. This involves as a corollary that to provide this anode swing the grid swing required . while fairly constant over the middle registers. A safety load of several times the normal load should therefore be connected permanently across the output or else a safety spark i gap incorporated. The average loudspeaker impedance. in general. which can easily be obtained from a diode detector.f.
i. It 489 limits of the grid bias must be remembered has to be made signal likely that.16. Two Tetrodes Arranged for Operation (Positive Drive) Class AB2 current which is the same whether the signal is large or small. allowance peak condition. As explained in the chapter on transmitters. in a design of this type. since the periods when anything like full modulation is in operation are only a small fraction of the total time. this form of operation consists in biasing the valves to such a point that the anode current is nearly zero.16. being biased B Operation The ordinary lowfrequency Fig. With an audiofrequency amplifier this is not permissible because we are dealing with a variety of frequencies and any undue resonance cannot be tolerated. Positive halfcycles of grid voltage then cause the anode current to swing over the operative portion of the characteristic. A typical circuit is shown in Fig. 10. for the Class amplifier operates as a Class A approximately to the middle of the straight portion of the characteristic. The outputs are combined in the normal manner.e. to use two valves in pushpull. This involves a large steady anode amplifier. In a transmitter the anode voltage maintains itself over this portion of the cycle by virtue of the tuned circuit in the anode. . one of which handles one halfcycle and the other the next. This has led to the introduction of Class B amplifiers for audiofrequency purposes. Particularly with batteryoperated receivers. It is necessary. the strongest possible to be handled under full modulation. while the negative halfcycles produce no appreciable effect. this involves a waste of current. 10. therefore.AUDIOFREQUENCY TECHNIQUE on the input to the valve must be within the used.
The various arrangements are thus as follows Class A: Valve biased to its mean current. as discussed on page 321.s. is from —25 to +5 on one valve and from 25 to —65 on the other. because a linear output is required. amplifiers.13. supply to the anodes fails the screens will draw excessive current and the valves will be damaged. With the conditions shown P f300. As the valve runs towards zero its partner runs negative. Under Class AB2 conditions 60 watts output may be obtained with only 3 per cent distortion. The peak power output is thus represented by the triangle XYZ. (This is rarely used in a. the power supplied for the first halfcycle of input. During the next half. this = = . = shows two tetrodes arranged to operate in this Class The output of the two valves in normal Class A pushpull is 145 watts with 9 per cent total distortion.m.cycle of input the second valve becomes operative and supplies a similar amount of power. This is we have just seen.103 594 watts.t. The operating condition of each valve is shown at XZ in Fig. The nomenclature with pushpull output stages is similar to that for transmitting valves. for if the main h. It is important to keep the screen voltage constant.f. The total swing. and as power is provided almost entirely by one valve of the pair. being derived from each valve in turn. which does not vary with the signal.) Class AB1: Valve biased to a small standing current. therefore. Such a valve absorbs power on its input and has therefore to be fed from a small power valve operating through a driver transformer which is designed in the same way as an output transformer to feed into the load provided by the grid current drive. The valve normally operates at 25 volts bias and runs up to +5 volts. between 5 and 25 per cent of the peak current. Hence the power \YX x YZ is the total power output.490 RADIO COMMUNICATION For Class B operation the valves are often designed to run into the positive grid region. 10. except that.396. 10. Class AB2: As Class AB1 but with the grid running positive on the peaks. Class C working is not used. YX is the peak current swing and YZ the peak voltage swing so that the r. but a separate power unit should not be used. with no signal Class B: Valve biased to cutoff so that ia and peaks to a maximum on positive grid swings. — \YX.16 AB2 (positive drive) condition. power is Fig.YZ.
voltage The performance of a Class B pair can be expressed in terms of the and current swings without actually plotting a load line is necessary to assess the probable distortion). the load load corresponds to 396/03 1. This is frequently done with transistor circuitry. Output Transformer correct load into the valves. line The output transformer must be designed so that it reflects the With the example just given. as discussed on page 501. and the operating conditions must be chosen so that this dissipation does not exceed the safe limit for the valves in use.320 ohms. in which case the operation approaches a true Class B condition. the bias may be increased. It will be seen that the anodeanode load is not twice the figure for one valve alone but four . and optimum loads are customarily expressed in this form.t. is very nearly the same as the peak current /. = = times. though it is of greatest importance in . If E is the h.AUDIOFREQUENCY TECHNIQUE 491 It will be seen that the arrangement makes a more effective use of the valve for.320) is ZX = 600 ohms this = stepdown will be 2(1/0675) 296:1. Maximum Power Output safe limit applies to The requirement that the internal dissipation shall not exceed the any valve. voltage and Es is the peak voltage swing (though this The peak current swing ( YX) (7Z)then = \E 1 Mean anode current = 1\tj (per valve) = 2//tt (total) Input power = 2E<J}ir Efficiency = inEJE Primary load resistance (aa) = 4E„// Power output (two valves) S The difference between the input and output power is dissipated in the valves as heat. If the secondary will require the secondary turns to be 0675 times the half primary. whereas the total current swing is not greatly different from that which could be obtained in a Class A condition.280 ohms. This is equivalent to saying that the effective anodeanode load for the two valves should be 600 296 2 5. the voltage swing is practically twice as great. Hence the overall v/(600/l. If it is desired to reduce the standing current.
and while this is usual there are cases where the load is partially reactive. The current will then lag 145° behind the voltage. At the point (AO = 122 OB) and W distortion criterion W W mA . Let us assume. 10. These conditions.17 for a maximum dissipation of 10 watts. Illustrating Dissipation Limit must be used to comply with the 5 per cent still lie within the 10 dissipation boundary. In such circumstances the load line becomes an ellipse and the conditions are entirely changed. that the load is inductive and that at the working frequency Leo \B.W dissipation vA i/„=5 —Vg / \ N v \V\r SO ^ \ \^ V\ \ ^^ \\\ "IV 1 1 1 Vg = 10 ^Vg = 15 XN^ '4 WO 200 240 300 V (volts) a Load Ellipse K Vg=20 400 S00 and Fig. moreover. 10.492 RADIO COMMUNICATION output stages. The internal dissipation is the product of the instantaneous anode voltage and current. a load of 4250 ohms — 100 Vg=0 \f^ ^£^10. If the standing bias is to be 10 V. The manufacturers state the maximum permissible dissipation and the working conditions must be chosen so that this limit is not exceeded at any part of the operating cycle. It will be seen that this gives a working point of 240 V and 40 which is only 96 but as Va falls the load line becomes tangential to the 10 dissipation line. and it is possible to plot a curve of the characteristics showing the maximum permitted anode current for varying anode voltages as shown in Fig. Hence it is not correct to assume that the standing conditions represent the maximum dissipation. for example. only apply for a resistive load.17. = .
Hence the presence of the inductance would result in a considerable reduction in the output. Transistor.AUDIOFREQUENCY TECHNIQUE where the will a. — = A.c. voltage of 240 10. is . 493 component of the voltage swing is zero the current and will not go through zero (i. It will be seen that this exceeds the maximum dissipation over a considerable portion of the cycle and to avoid overloading.F. By plotting the instantaneous values of (a. but the circuitry equally applicable to npn types.3.18. The question is considered further in the discussion of transistor output stages on page 501. 10. the standing current of 40 mA) until the voltage swing is 45 V (i.17. an anode be — ilmax 45 195 V). Transistor Amplifiers The considerations which govern audiofrequency transistor circuitry are not basically different from those discussed in the previous J T MT C3 t (a) Fia. RC Amplifier section. The equivalent circuit * Throughout this section pnp transistors are shewn.e. 10. but in practice a more conservative rating is usually necessary. but since the bipolar transistor is a currentoperated device the impedances involved are very much lower. 10.) voltage and current over the cycle it will be found that the load line is an ellipse as shown dotted in Fig.18 illustrates an RC coupled transistor amplifier using the commonemitter mode which. the working point would need to be moved appreciably to the left (lower V a and /„) while the effective slope would need to be reduced to avoid distortion.e. The conditions are slightly modified when two valves are being used as a Glass B pair. is the most suitable for low and medium frequencies*.c. Here each valve is only in operation for half the complete cycle so that it is permissible to use a load line which crosses the dissipation line for part of the swing. Thus Fig. Theoretically this should permit the use of a dissipation line corresponding to twice the normal rating. as said earlier. using a positive supply voltage.
where X — 1/coCj^ reactance of Cx this reduces to 2) When the Hence the Gm R 1 R 2 \(R l ratio of the output at +R low frequencies to that at mid frequencies becomes If Rx = kR 2 + i? and X = nR (i?1 2 )/(i?1 + i? 2 +jZ) may ] 2.f. R = R + . At low frequencies the reactance of Gx will begin to become appreciable. Z is the impedance of R in parallel with C R x 2 so that X offsets the attenuation of the If we assume . .18 (6). 10. The resistance R2 however. a compensating effect. as explained later. that the effective load Z is small compared with R (as it is in practice) the stage gain is approximately Gm Z.494 is RADIO COMMUNICATION in Fig. the proportion actually applied to Tr 2 being R 2 I(R 2 + 1/jcoGj). however. R 2 1000 O and R = 30 kii the stage gain would be 34. At medium frequencies the reactances of C x and C2 may be neglected so that the effective load Z RjR^KR^^ J. be seen that the load of the network R 1G1 R 2G 2 where 2 and C 2 are the input resistance and capacitance of the following stage. 10. There is. is not such a serious limitation as might at first appear. v 2= vx R 1 (R 2 +jX) R2 R 1 + R 2 +jX' R 2 +jX = G mR 1 R 2 l(R 1 + R 2 +jX) m is negligible. CX R 2 network. 47 kQ) in order to provide correct operating conditions.2 ) and the stage gain is Gm ZR l(Z R ). Fig. R 2 is normally small compared with R x so that the effective load is determined mainly by R 2 But if the reactance of G 1 becomes appreciable this shunting action is reduced so that more gain is obtained from the transistor and this partially This = . Taking Gm = 35 mA/V. R byR 2 . so that v 2 is less than the full amount. Actually x would be several thousand ohms (e. Now.1). the expression 2 be written in the form (i + t)M(i + *) + «2 . but is the input resistance of the succeeding transistor Tr2 This is of the order of 1000 ohms so that there is little advantage in making R x much larger than this.g. is not an external component as is the case in a valve circuit (c. but the gain is determined principally it will shown from which is in the collector circuit of Trt made up .
R 2C co 036. It was shown in Chapter 6 that the voltage gain in common. which is RJO^co).e. the loss will be 10 per cent when n = 28 i. so that using the approximate 6 dB per octave rule one would expect the performance to begin to deteriorate at about 15 kHz. If k = 5.) is <x'. will normally be about 5 times R 2 The base voltage may then be set by suitable proportioning of R a and R b (Fig. However. This falls to unity at the cutoff frequency fls which for a smallsignal a. then Gm a'r t ). (Fig. + = + . = Choice of Working Point The operating conditions must be chosen to provide the required current swing. With the commonbase mode Gm a/re and since a does gm not fall off as rapidly as a. the performance deteriorates much earlier than this due to the falling off of the mutual conductance G m This is not the intrinsic mutual conductance g m but is a derived parameter which decreases rapidly with frequency. Nevertheless. At the highfrequency end of the scale the reactance of Cx is negligible but C a begins to shunt B 2 The effective gain thus becomes Gm Rv ftl R1 Sa l(R1 R a ).f.' (being 3 dB down at/a approximately l'2/x for alloydiffused transistors and 2/x for driftfield types). whence the . which is roughly a. written as Gm Z. which can be done by drawing a load line across the characteristics as explained in Chapter 6 /„. where Rp 90 per cent of the midfrequency gain when RpG 2 co = \. where it was also shown that to avoid the danger of thermal runaway the working point should be such that the voltage drop on R t is not less than half the battery voltage (the halfsupply voltage principle). and while this can be done with the emitter grounded a more stable condition is obtained with a resistor J? 3 in the emitter circuit as was explained in Section 6.Z/(r6 a're ). With the values previously assumed (R 1 5 kQ. 10.10). . (neglecting the source resistance R. emitter If this mode is — + + = = .AUDIOFREQUENCY TECHNIQUE 495 from which the relative bass cut may be assessed at any frequency. the commonemitter mode is usually preferred.'l{r b proportional to a'. collector resistance Note that this is not the load resistance.18) to give the appropriate base current. transistor is of the order of 500 kHz. 6. 1 This is an interesting result which shows that the bass response is maintained at appreciably lower frequencies than with a valve (for which 10 per cent cut results when B 2C1 w 2).5. = = = = . JB 2 1 kQ) a capacitor of 115 /jJ? would maintain 90 per cent response at 50 Hz. the highfrequency response is much better. because of its other advantages. This determines Rt = and (V batterv — V )II e Vc and e. With typical values this occurs at a frequency approaching 1 MHz.
as shown in the equivalent circuit of Fig. the resistance x must include the decoupling resistor. To (external) power gain the equivalent primary must be equal to the output impedance R of Trv Tr 2 .496 RADIO COMMUNICATION R 3 must be bypassed with a capacitor G s to avoid signalfrequency feedback.19 illustrates a transformercoupled transistor stage. In determining the working point. 6^ . 416). The connected to the base of Tr% with the earthy end taken to the tap on the primary is in the collector circuit. 10. This equivalent primary impedance R t will be shunted by the primary inductance and the stray capacitance G (including the reflected input capacitance of Tr 2 ). 10. while the secondary is FlO. + 1/R. R Transformer Coupling Fig. . + IIjcoL) Hence im = GWVK1 + RiW + RflcoW] . Such a circuit obtain the maximum impedance Z v This can be achieved at midfrequencies by adopting a (stepdown) transformation ratio of i/{R IR t ). At low frequencies the effect of C can be will divide through R R x and L.19 (b).19. where R { is the input impedance of would be designed on a powertransfer basis. voltage divider Ea R b this slight rearrangement being necessary to prevent the negligible resistance of the secondary from shortcircuiting R b. (l/jy/fl/*. they may need to be decoupled to avoid commonimpedance feedback (see p. ignored so that the current The portion of interest is that through Rlt which is in the ratio. TRANSFORMERCOUPLED TRANSISTOR AmPLIITER . 10. the value being chosen so that ljC z (o is not more than about onetenth of jB 8 If several stages are being used.
transformer coupling is essential.5 (Fig. is at medium frequencies) the second term voltage gain is »i/»i l 9 When <oL ^> Rt and the negligible If is Hi 95. the peak input current may be of the order of 2030 mA.e.20. which is the most usual. which requires a transformer coupling from the preceding (driver) stage. and assuming 10 per cent drop at As the frequency is reduced . A Class A output stage is illustrated in Fig.800 Hz. Since this involves appreciable internal power dissipation a more massive construction is adopted as described in Section 6. Transformer coupling offers little advantages for voltage amplification because two RC stages in cascade will provide greater amplification at less cost. using the commonemitter mode. plus the reflected input capacitance of 2V 2 Oin jn2 In a practical circuit this would be of the order of 30 1000/55 2 63 pF. The circuitry is similar to that already described with due allow ance for the larger currents involved.26) and the transistor is mounted on the chassis or some form of heat sink to conduct the heat and prevent an undue rise in temperature. Ra R b must be bypassed with a . When current or power gain is required. 10. = R = 30 kii. Output Stages For the output stage the transistor is designed to handle large currents (of the order of amps rather than milliamps). Hence there would be a slight resonant rise in the output similar to that illustrated in Fig.AUDIOFREQUENCY TECHNIQUE and the voltage across the secondary. L 96 H. . (i. With the values quoted.10 (though not in this case due to = = . however. suitable capacitor. as in a driver stage. + H = leakage inductance) followed by a rapid cutoff. Rt = =R = = 50 Hz. This is made up of the output capacitance of Trx including circuit strays. 497 V2 = i^R^jn. and to avoid undue variation of the working point the bottom portion of the bias voltage divider. Since the peak collector current may be 1 A or more. which wouldresonate with an inductance of 96 at a frequency of 12. If R R t the gain falls to 90 per cent of its midfrequency value when a)L 71 per cent with a>L %RV The primary v and to l/\/2 inductance may thus be calculated in terms of the permissible bass cut. the stage gain so that coL becomes comparable with the current through R t (and hence the gain) falls. At high frequencies the effect of C becomes important. 6. 10. (? = GmB B jn(B + Ri) m = 35 mA/V and w = 55.
10. as usual. 10. Characteristics or Typical Output Transistor The figures against the curves refer to I b . by a resistor in the emitter but because of the large current this resistance is only of the order of 1 ohm or less. The output load can be determined from the characteristics as in Fig.498 RADIO COMMUNICATION stabilized. . as shown by Bt in the figure. This is no disadvantage in an output stage and in fact The conditions are circuit.20. • 120mA i__ ___^ 80mA \a12SW BOmA SOmA 40 mA ^ v 30mA 20mA 10mA Fig. To bypass this at signal frequencies would require quite impractical values of capacitance so that it is customary to omit the capacitor and accept the resulting negative feedback. 12 2:45:1 R 33kC\ 3»l% Input 82 kCl > 820n Fig. Class A Output Staoe additional negative feedback is often provided.21. 10. A supply voltage of 12 V has been assumed. which represent a typical power transistor.21. which allowing for the voltage drop on the output transformer and the emitter resistor would leave 3 .
Allowing 1 ohm for these two together an additional 50 mV peak will be required. and the power output is 425(13/14) 405 per cent.. which have been assumed to be 05 ohm each.22) it will be seen that the requires a base voltage of 480 mV. mA making 530 mV peak swing. so that mean current of 25 to reduce i.s. = = . From the input characteristic (Fig. requiring an input of the order of 5 mV depending on the effective O m of Trv B a would be adjusted to give a standing current of 09 A in Tr 2 . 17 . 10. The reflected secondary impedance should thus be 4 W. (OB A = second harmonic. = characteristics./2v/2) = 7^/8^2^X166/8 = 425 W.) If the secondary load resistance is 3 is ohms net (after deducting the secondary which should only be a small fraction of an ohm) the output transformer ratio V(13/3) It = 208 sufficient to remains to design the preceding driver stage which must be swing the base current of Tr2 over the required range. power is — = — V = (Fs /2V2)(/. giving a dissipation of 99 W. Part of the load is made up of the primary resistance of the transformer and the emitter resistance. These are the overall peak 01 current swing is 165 swings.) 05 while the peak 22 The peak voltage swing is 225 155 A.m. giving less than 3 per cent ably low distortion. so that the r. to zero a peak swing of this order is required. (The maximum giving an overall efficiency of 4/99 efficiency in Class A is 50 per cent.AUDIOFREQUENCY TECHNIQUE approximately 11 499 V for VM . which would be obtained if \YBI This 2VB and I. is clearly not attainable in practice because of the knee on the — = = = . and a load of 14 ohms about this working point would provide a reason11 OA. If we assume a net figure of 9 V we can safely assume a peak swing of 8 V in the collector of Tr lt which will require a transformer ratio of 8/053 = 15. which could be provided by a normal smallsignal transistor with a standing current of 2 mA. The supply voltage to Trt will be less than 12 because of the drop on the decoupling resistor B. To this must be added voltage drop on the emitter resistor R e plus the secondary resistance of the driver transformer. 2I giving a power of 4FB/ /8 V. 14 1 13 ohms. To allow a margin of safety a standing current of 09 could be chosen. The primary current swing will then be 50/15 = 33 mA.
000 V (mV) be Fig. 10.21 of 165 mA this in both directions. will ic of Tr2 is produced by a decrease in i in Tr e v then drive Tr± to cutoff. . Because of the nonlinear input characteristic the positive swing only requires an increase of 120 in voltage of if. and prevent overdriving 2V2 The nonlinearity will also be responsible for introducing a certain amount of distortion. heat dissipation were only just adequate for normal operation. It is desirable therefore to arrange the phasing of the driver transformer so that an increase in Excess input . Input Characteristic for Transistor oi Fig.22. a small increase in input voltage can produce a relatively large increase in ib. but the voltage developed automati Nevertheless. 10. cally adjusts itself to the changing load. . This is partially offset by the negative feedback produced by the unbypassed emitter resistance B 3 but it may be still further reduced by applying additional negative feedback. Tr x thus operates with a peak current swing mA mV 1. 500 RADIO COMMUNICATION the corresponding base voltage divider of Trx being chosen to give a standing current of 2 mA. and may known as overdriving cause a disproportionate increase in ie This is and could produce thermal runaway if the . from 25 to 50 but by this very fact the input impedance (and hence the effective load on TrJ is correspondingly reduced and the drive voltage is reduced accordingly..
20 is such as to produce a negative feedback 750 mV. which requires a slightly higher standing current of around 25 mA. Thus in the example above.700/35)/33 = 148. = the required ratio is (l. More efficient performance can be obtained by using two transistors in pushpull and operating them in Class B conditions.20. The transformer ratio may then be calculated to meet the requirements of the driver stage.20) to limit the gain if the temperature rises. the required basecurrent swing is Ic /x'. This would make the load line horizontal requiring a (theoretically) infinite voltage swing. In essence the basis of the arrangement is that the two transistors . while a' is 35. collector voltage. General Precautions an output stage to ensure that the arrangements no undue rise in temperature can occur. Ic is 17 A. 10.AUDIOFREQUENCY TECHNIQUE This 501 resistor may be done around the output stage alone by a between collector and base as shown dotted as Rt in Fig. If the total swing of the collector current in the output stage is /„. of 3 dB the drive voltage must be increased to 530\/2 The driver transformer ratio then becomes 107 and the current swing required from Tr x is 45 mA. 10. This kind of operation was discussed on page 489 in relation to valve circuits and the same general principles apply. and temperaturesensitive circuit elements are sometimes introduced (e. In particular the load should never be disconnected when the circuit is in operation. Thus if Rf in Fig. An approximate design may be derived without reference to the characteristics. Assuming a total driver current swing of 33 mA as before. The other major precaution is concerned with the maximum The load must be such that the peak negative voltage swing does not exceed the rated limits. The circuit should therefore be designed so that overdriving and/or changes in bias conditions do not lead to undue peak currents. or by feedback round the whole circuit from the secondary of the output transformer to the base of Trv If this is done the drive conditions must be suitably modified. Class B Stages The disadvantages of a Class A stage are the limited efficiency (50 per cent maximum) and the high standing current (approximately \lc ). a rise in temperature causes a rapid rise in collector leakage current lm which can cause a further rise. 10. Among other things. which would destroy the transistor. It is important in for heat conduction are adequate so that . leading ultimately to thermal runaway. across R b in Fig.g.
23. On the first halfwave Tr v is driven to its peak current. Input is supplied to the pair symmetrically as shown in Fig. Then Trx supplies the power for the first halfwave. 10. indicated by the Fio. . In practice true Class B operation such as this is not used. Illustrating Action of Class B Output . 10. 10.502 RADIO COMMUNICATION are biased nearly to cutoff. While two transistors (of the same type) may have similar characteristics over the normal range of operation (and are often selected as matched pairs) this does not hold good for very small values of ib so that appreciable mismatching can occur around the point 0. 10. while the second half is supplied by Tr 2 the two being combined in the centretapped output transformer of Fig. Stage shaded area. while Tr 2 remains inoperative. The operating point is the battery voltage but the base voltage (on each transistor) is such that the collector current is zero. The behaviour is illustrated by the WW Output Fig.24. On the next halfwave.24 in which the to Trt and the dotted curves Tr 2 . Pushpull Output Stage fullline characteristics load line relate to AOB in Fig. 10. conditions are reversed so that each transistor delivers up onehalf of the power in turn.23.23.
with an appreciably lower load and larger peak currents. . the mean d. Fig.c.25. 29/3 so that the efficiency is 74 per cent. so that from i b to ib 120 the current swing /„ is 29 A. S y\ _— \ V\ T N '" 1 H J= 5 "v\ — 4 ™ \ . 10. 10. = = wj^yj if = bn'Vo'I* I v but this is not attainable in practice as be clear from Fig. This is a maximum of 785 per cent v h) The d. 02 A. mean of the the product of the battery voltage V and This is not the standing current I but the sinusoidal swing from zero to I v 2I pJtt (for the complete cycle). Chabactebistics of Fig. power is current.21 Transistor in Class B . The line AOB represents a load of 4 ohms through a working point of 12 V. To avoid i. power 2V I1>Itt and the efficiency is 1 V. Hence d. .\ \ *^\ \ X^xo t»i° — & V • c ?*<— ' 10 75 (y) 20 *B VB r 25 Fig.21. 10. (Because of the smaller standing current the voltage drop on the primary and emitter resistances can be ignored. Class B working permits greater utilization of the transistor. = = mA = — = (VJVWbIV*) = which is W. 10.e. 10.. VP = V and Ia = will = — 3.25. 1 With two transistors in pushpull the combined output is this the transistors are operated in Class AB1 conditions. This results in a smooth transition from one to the other.25 illustrates the use of Class B operation with the characteristics of Fig.c. AUDIOFREQUENCY TECHNIQUE which produces what is 503 known as crossover distortion.c.en™ a n ?. while the voltage swing V„ 125 115 V. with a small standing current about 5 per cent of /„. = 165 W over four times that obtainable from a single transistor in Class A.) This load causes the transistor to peak up to 3 A. With the values shown Vs /V 115/12 and Isjlj. I.Sir\ — 120 m A — 100 mA 80mA .
the the transformer primary. = = = = = mA R mV = = mA Dissipation limit. which requires a ratio of equivalent secondary load 107 per half. The 85 current swing will then be 120/14 so that the driver will need to operate with a standing current of the order of 10 mA. If R e 325 £i. Because of the symmetrical nature of the output stage less decoupling of the driver stage is required. The collector leakage current normally only a few per cent of Im produces further losses. The mean current of from Fig.22 requires Vbe 60 would require 650 mV. This is approximately transistor for the load line hVBh AOB. This is /„. 10.504 RADIO COMMUNICATION As said. which 750 mV. Over the halfcycle the mean voltage swing is VJrr. The load of 4 ohms will include Re and the resistance of half 05 O and ^Bj. With the load AOB this requires 120 mA. The driver transformer is designed in the same manner as for a Class A stage. Each half of the secondary in turn has to supply a current of i b peak. making 2I total. Hence in the quiescent state with no input a current I will be drawn by both transistors.cycle. so that allowing as before for the drop on e and the secondary resistance a swing of some 700 will be required. 10. With the example chosen the quiescent current is 04 A rising to 192 A at full drive. The load line AOB in Fig. When signal is applied this will rise to a maximum of 2I v JTr (drawn by each transistor in turn) at full drive.25 is well within the dissipation and the dissipation is still further reduced by the fact that the is only in operation for half the cycle. 025 D. and may often be omitted altogether. and the mean current IJn. This will require a somewhat larger transistor than for a Class A stage. If we assume that a peak voltage swing of 10 V is possible the transformer ratio per half will be 10/0*7 14:1. the mean current is not I but 2/„/77. but it increases rapidly with . However this is not the total loss. There is appreciable dissipation in the transistor due to the presence of the standing current I during the idle half. Hence the mean power dissipation is VJJn* cs: VJJlO which with the load AOB is only 33 W. still = £ X is 12 X 02 = 12 W total will This an appreciable fraction of the and any increase in I due to a change in the bias conditions increase the loss further. or 214 overall (assuming a 3 O secondary \/(325/3) load as before).
which is a measure of the rise in junction temperature relative to the transistor casing. including the heat sink. Finally there is 505 which the disspation due to the base current.25. AOB = This we assume /„„ = 01 A.g. the maximum ambient temperature t. This depends on the efficiency of the arrangements. Applying these values to the load AOB in Fig. while 6 X can vary between 02°C/W with the transistor in direct contact with the chassis to 05°C/W if a = = = mica insulating washer is used. the total dissipation 01 x 115 X 3 + £(12 x 02) + 12 x 01 = Ambient Temperature is transistor.w(d h + e. 10. = t. The relationship between power dissipation and junction temperature may be expressed in the form pd = where t (Tt . but t is a function of the transistor itself. 90°C and t Typical values for a power transistor are T} 4°C/W. but this is normally negligible. 6 h and X are within the control of the designer. Hence the total dissipation is approximately will be Vb (mean) X (2/tt)/6 .+ F^ for the load 57 W. for dissipating the heat. . well below the permissible limit of 12 \ However this limit is the absolute W for the particular maximum and it is necessary to examine the permissible ambient temperature conditions to ensure the stability of the design. OIF^ + If PV.57(22 + 05 + 4) = 90 . It is made as low as possible by the transistor manufacturer.. with the transistor mounted on the chassis) A 22°C/W. With a good heat sink (e.AUDIOFREQUENCY TECHNIQUE temperature.Ta )/(e h + et +e t) T = maximum temperature of junction Ta = ambient temperature (°C) Qh (°C) = heat sink coefficient = rise of temperature of the heat sink in °C per watt X = transfer tivity dt = coefficient depending on the thermal conducbetween the heat sink and the transistor transistor thermal coefficient.385 = 515°C t) Hence this arrangement would be inherently safe. + e = 90 .
1 = 18 /. Here 025 A. A possible arrangement is shown by the load line ACD. materials usually employed). . that the transistors of Fig. in fact. 10. Input Output Fig. As a precaution.506 RADIO COMMUNICATION Highoutput Operation It is evident. at an efficiency of 75 per cent.26. The dissipation is 01(18 X 3) + 4(18 X 025) + 18 X 01 = 845 W Applying this to the temperature equation above we find that the maximum ambient temperature is 33°C. Modified Form op Pushfull Circuit Precautions are still necessary to avoid temperature rise (which can cause a rapid increase in Im ) which might be produced by overdriving or by a change in the bias arrangements causing an increase in I which has a material influence on the dissipation as shown earlier.c. 18 and /„ VB = while f = 19 . which is a safe enough margin for normal requirements. Fig. 10.26 shows an arrangement in which the collectors are at .25 would handle a greater power output if the battery voltage were increased. s = It will be seen that this load line crosses the dissipation line over an appreciable part of its length but this is permissible because the transistor is only operating over half the cycle. Temperaturesensitive elements may also be incorporated in the biasing circuits to reduce the gain if the temperature rises abnormally. = 3 — 015 = 285 giving P = X 285 X 18 = 265 W. the emitter resistor Be may be wound with copper or other wire having a positive temperature coefficient (as against the zero t. There are many modifications of the basic pushpull circuit. 10.
Transistors in the are available. for operation at voltages well in excess of 100 V if required.27.27 By 2V Input Fig. Thus in Fig. Because even a relatively small transistor will handle currents of the order of amperes the required power output can usually be obtained with collector voltages of the order of 50 V. shows a socalled transformerless Class B output stage. If the voltage at A (derived from the . Fig. transistors are available having similar characteristics with either pnp or npn construction. 10. however.28. Complementary Output Pair such complementary pairs (preferably matched) the need for an input transformer can be obviated. By using ^ Input "• t® Output Fig. Complementary Operation As said in Chapter 6. 10. The loudspeaker impedance must be designed to provide the requisite load.28.AUDIOFREQUENCY TECHNIQUE 507 battery potential. Tra and Tr s are a complementary pair. Asymmetrical Class B Stage Feedback arrangements may be incorporated in Class B stages same manner as in Class A amplifiers. 10. 10. using a centretapped supply the output transformer may be eliminated but a driver transformer is still required. an arrangement which is not only simpler but provides slightly better thermal conductivity. which permits the transistors to be mounted direct on the chassis.
* * Tobey. as in Fig. 565. Tr2 conducts while Tr s is cut off. "Transistor audio power amplifiers. 67.. is known as the Tobey and Dinsdale circuit and is widely used in highfidelity and publicaddress equipment. K. which is reproduced by courtesy of the Editor of Wireless World.29. and Dinsdale. while when A goes negative. This arrangement 15 is suitable for small outputs." . Wireless World. Amplifier features discussed in the preceding pages. The resistors R^ and R 2 must be arranged so that with no signal the point A stands midway between the positive and negative rails. Hence the potential of the point B moves towards the positive rail.508 RADIO COMMUNICATION input transistor Trt ) goes positive.29. Tobey and Dinsdat. This arrangement. In such cases a normal pushpull output would be used with a symmetrical input derived from a complementary input pair.. which is a typical circuit incorporating many of the larger outputs —45V Input R2 Load C4 OOlfiF Fig. p.f. B moves towards the negative rail. 10. Watt For Amplifier matched pairs are not readily available. 10. J.
AUDIOFREQUENCY TECHNIQUE 509 A feature of the circuit is that it is direct. R . = !<. the diode serving to compensate for variations in the supply t voltage as discussed in Chapter 6. = 10. by making 10 and lz equal and employing transistors having equal values of a'. so that the overall gain is 9. . B R R . in considerable use employs magnetic wire or tape on which the A .poweroutputis(FC(. so that there is no need for any ironcored components despite the fact that it will deliver a power output in excess of 15 W.m. . The output of 2V2 is applied to the bases of the complementary pair Tr z Trt Tr a is an emitter follower. . Any variation in the quiescent potential of the point P is amplified by Trx and the subsequent stages. With a supply voltage of 45 V the peak output voltage is 225 V. D E . signalfrequency feedback is provided from the output The effective value of t is modified to the emitter of Trv via 16 by the impedance of Trx but with the values shown the effective feedback factor is approximately 1/9. Sound Recording Sound recording may broadly be divided into three sections according to the medium employed. stabilizer to maintain the midpoint of the output stage accurately equal to half the supply voltage. while 2V4 is a commonemitter stage. The current gain of both stages is approximately a'. providing a feedback which holds P constant. Tr± is an npn signalfrequency amplifier which feeds the hase of the pnp transistor Trt However. so that the peak input required is approximately 25 V. The base of Trx is held at a steady potential by the voltage divider R^R % while its emitter is fed from the point P through B 6 (decoupled by C2 ). and with the values shown the total and Rfooa harmonic distortion can be less than 1 per cent. 9 is adjusted to provide a small (symmetrical) unbalance of Tr s Trt which sets the standing current of Tr6 and Trt to a small value sufficient to avoid crossover distortion.coupled throughout. Finally.R od which with Vcc 150 is 17W. Mechanical groove recording on cylinder or disc was the first method used. followed later by soundthird method now onfilm recording as used in cinema technique.c. 45 V Ther./2' V/2) 2/. The capacitor C 4 limits the feedback at high frequencies (outside the pass band). so that. which is essential for maximum undistorted output.4. because of phase shift within the amplifier. which. it also serves as a d. It will be noted that the decoupling capacitance G 3 is taken to the point P. the output transistors Tr6 and 2V6 receive equal inputs in antiphase. . might become positive and so cause instability. which is the zerosignal point of the output network. .s. .
from background noise (scratch). it is The techniques of Disc Recording Recording of sound on disc is carried out by cutting a spiral groove in the disc and then modulating the contours of the groove in accordance with the recorded sound. and a suitable correcting network may be needed as explained in Chapter 14.p.m. A more detailed exposition will be found in Radio and Electronics. dynamic (moving coil). the grooves are narrower. and it is much more durable than the early shellac discs which were very brittle.m. the main consideration being lightness of weight to minimize wear on the record. A typical piezoelectric pickup construction is illustrated in of the order of 05 to 2 volts so that preampUfiers are not required. known as E. Chapter 7 (Pitman). and only possible here to discuss the basic principles. is Pig. The modulation may be carried out in the vertical plane ("hill and dale") or in the horizontal plane (lateral recording) the latter system is now almost universally employed.P.p. discs. The recording may be cut directly on a lacquercoated disc for direct playback.30. 10. In the 45 and 33J r.m. but these have now been superseded by vinyl discs running at 45 or 33 J r. all three methods are highly specialized. Vol. The output from such pickups .p. (longplaying) recordings.. and relative freedom of movement in a lateral direction with stiffness in the vertical direction.510 RADIO COMMUNICATION imposed and subsequently reproduced by signal variations are suitable magnetic pickup heads. Vinyl has the advantage of greatly reduced signal/noise . discs the needle or stylus was usually of steel (or sometimes fibre to minimise background noise). or piezoelectric crystal type. The pickup used to translate the mechanical motion of the needle into an electrical waveform is generally of the electromagnetic (moving iron). The principles of operation are fundamentally those of the corresponding types of microphone. Early recordings used shellac discs rotating at 78 r.p. and it is usual ratio being virtually free also flexible so that the discs are to use sapphiretipped needles. 2. the impedance is predominantly capacitive so that any load across the pickup reduces the response at the lower frequencies.P.m. but in the gramophone record industry the original is made of wax from which a metal stamper is produced and used as a press tool to make a large quantity of copies in a plastic material. (extended playing) and L. However. With the early 78 r.
These are recorded on the disc by cutting the modulation on one side of the groove only. 20 Pickup Stereophonic Recording Normal sounds have a directional quality because of the small phase differences between the arrival of the air waves at the two ears. The excursions of each stylus are controlled only by the modulation of its appropriate side of the groove. fed independently with signals from two similarly spaced microphones at the recording studio. and do not come into contact with the opposite side at all. and then delivers two separate outputs which are fed into separate amplifiers . The two sides of the groove can thus carry entirely separate signals. having two styli set at 90°. Two loudspeakers are used. 10. This quality is not reproduced with a normal recording. The pickUp is lowered into the groove in the normal way. Section through Acos GP. They are predominantly inductive so that any resistance across the device reduces the highfrequency response. A special pickup is used.30. Crystal Element Case Filled with Plastic Gell Electrodes Secured to Contacts by Conductive Cement Cantilever Sapphire Contacts Pivot Guara Fig.AUDIOFREQUENCY TECHNIQUE S11 Movingiron types of pickup are appreciably less sensitive and generally require an additional amplifying stage. each being 45° to the direction of the groove. but can be introduced by stereophonic techniques.
512
RADIO COMMUNICATION
and loudspeakers, thus reproducing the stereophonic quality of the
original.
Alternatively, the two channels may be recorded on film or magnetic tape, or provided by two separate radio transmissions.
Sound on Film
Just as in disc recording the groove may be modulated in the a photographic track may be modulated by either the variable area or variable density method. In the density method the track image is of constant width and its density is modulated to conform to the signal voltages, while in the area system the track is of constant density but occupies only a portion of the available track area. The modulation is then displaced laterally so that the image on the film corresponds to an oscillogram of the signal waveform. In the reproduction of both systems the sound track is illuminated evenly from a fine slit at right angles to the direction of motion of the film. The movement of the film thus modulates the light beam and the light modulations are translated to electrical energy by a photoelectric cell. With correct adjustment the systems are capable of yielding equally good results and both are in general use. The sound track is usually located some distance ahead of the picture frame to which it refers. This is necessary because the picture is jerked past the gate, frame by frame, whilst the sound track must be run smoothly past the slit. Consequently the two actions must take place some distance removed from each other.
lateral or vertical direction so
Magnetic Tape Recording
In magnetic tape systems the signal is caused to produce variations magnetism of a thin iron wire or tape. The signal is applied to a recording head of toroidal form with a small air gap, as shown in Fig. 10.31. The tape or wire is drawn across this air gap so that the magnetic field is completed through a small longitudinal portion of the tape. Thus at any instant the small element of tape across the air gap becomes magnetized, and if the material has suitable remanence the tape as it moves past the recording head becomes a series of very small bar magnets having magnetizations proportional to the
in the
instantaneous signals.
The signal may be reproduced by drawing the tape at the same steady speed over the poles of a similar reproducing head; voltages are then induced in this coil which are proportional to the signal
waveform.
AUDIOFREQUENCY TECHNIQUE
The material used
for the magnetic tape can be
513
made
either
by
coating a base with magnetic material or by impregnation. Steel wire can be used, but is less satisfactory. typical coated tape would consist of a base of cellulose acetate coated with iron oxide, while an impregnated tape would consist of equal proportions of vinyl plastic and magnetic material. In a typical tape recorder the tape is passed over an erasing head before reaching the recording head. The erasing head is fed with
A
HtoH
Fio. 10.31.
Magnetic Rkoobdino Head
sufficient current at about 60 kHz to carry each element of the tape to saturation repeatedly in each direction. The result is that the magnetic domains are then aligned in a completely random manner and the result is equivalent to demagnetization. In fact this method
not completely efficient but the residual is usually 60 to 80 dB below the level of the peak signal. In the basic recording system described above the residual magnetism after the tape leaves the recording head would be far from being proportional to the signal current owing to the pronounced curvature of the BH curve (see Chapter 4). This difficulty is overcome by using h.f. bias at about 60 kHz on the recording head. The signal is applied as modulation on the bias so that for any value of signal the tape is driven through several complete cycles of magnetization by the bias signal and the residual magnetism follows the signal modulation quite closely. The same h.f. oscillator is used for the bias as is used for erasing, though at a suitably reduced amplitude. The most notable advantage of magnetic recording is that the signal can easily be erased and the tape used over again many times. The recording can be reproduced many times without loss of quality and tape can be cut and stuck together again very simply during editing. Wire recorders can be pocket sized and are very useful for
is
514
field
RADIO COMMUNICATION
work in broadcasting and as portable dictaphones; the technical quality is substantially inferior to tape but this can be tolerated
for these applications.
The commercial disadvantage of magnetic tape recording as opposed to disc recording is the comparative difficulty of obtaining a large number of copies of a magnetic record. For special purposes tape recorders can be designed to accommodate a number of separate tracks side by side on the same tape. One obvious application is the stereophonic recording mentioned earlier, the signals from the two microphones being recorded side by side on the one tape, and reproduced by suitably located pickup
heads.
The upperfrequency response with magnetic tape recording is determined by the width of the gap in the recording (or playback) head, and the speed at which the tape moves over the head. The gap cannot be made too small or the magnetic flux would prefer to flow through the gap rather than the tape. A further problem is that, if the elements on the tape are too close together, they tend to demagnetize adjacent elements, while there is at high frequencies a certain skin effect which prevents the flux from penetrating the tape adequately. In practice the minimum gap width is of the order of 025 to 05 mils and since two elements are involved in a complete cycle the maximum number of cycles is between 1000 and 2000 per inch. The standard tape speed is 1\ in/sec, so that allowing for the effects just mentioned the maximum, frequency is of the order of 10 kHz. For high fidelity a higher speed (e.g. 15 in/sec) may be used. For speech or mediumquality music a speed of 3f in/sec is often used. Considerable care has to be taken to maintain the tape speed constant, for obvious reasons. Since the e.m.f. is determined by the rate of change of flux it is clear that with a constant tape speed the output will be proportional to the frequency of the recording. This is true up to about 1 kHz, after which the effects mentioned above cause a progressive limitation and the output begins to fall rapidly. Hence again there is usually a need to incorporate suitable equalizing networks.
Video Tape
Although the frequencies involved in a television modulation run into the megahertz region, as explained in Chapter 16, the number of picture elements in any one line is only of the order of
1000. Hence it is possible to arrange to record television signals on a tape one or two inches wide and arrange that the recording and
AUDIOFREQUENCY TECHNIQUE
515
reproducing heads are moved across the tape, at right angles to the normal direction of travel. The head must be arranged to traverse the tape laterally in one line time, i.e. 064 millisec with a 625line transmission, but by mounting a series of heads on a rotating arm this can be achieved with practicable speeds of rotation. The axial travel of the tape is then determined by the frame speed required, and by varying this,