CIRCUITS FOR

AUDIO AMPLIFIERS
POWER AMPLIFIERS AND PRE-AMPLIFIERS FOR

MONAURAL AND STEREOPHONIC REPRODUCTION FROM

MICROPHONE, RADIO, TAPE AND PICK-UP SIGNALS

 First published-April 1959
Second edition-August 1960
(Reprinted with minor corrections-March 1962)
Reprinted - November 1962
Reprinted -October 1963
© Mullard Limited, 1963

This electronic edition in Adobe Acrobat format, is published in 2003 © Richard G. Smith

Mullard is the trade mark of Mullard

Limited and is registered in most o
the principal countries in the world.

PREPARED BY TECHNICAL INFORMATION DEPARTMENT

AND ISSUED BY MULLARD LIMITED, MULLARD HOUSE, TORRINGTON PLACE,
LONDON, W.C.1.(TP413)
 CONTENTS

CHAPTER PAGE

Preface v
Mullard Audio Publications vi
1 Amplifying Systems 1
2 Sources of Distortion in Recorded Sound 4
3 High-quality Amplification 16
4 General Notes on Construction and Assembly 26
5 Twenty-watt Amplifier 29
6 Ten-watt Amplifier 39
7 Three-watt Amplifier 53
8 Seven-watt, D.C./A.C. Amplifier 59
9 Two- and Three-valve Pre-amplifiers 67
10 Four-channel Input-mixing Pre-amplifier 82
11 Three-watt Tape Amplifier 88
12 Tape Pre-amplifier 102
13 Seven-watt Stereophonic Amplifier 113
14 Three-valve Stereophonic Amplifier 121
15 Stereophonic Pre-amplifier 127
 PREFACE

The first notable contribution of Mullard design engineers to the specialist world of high-quality sound reproduction
was the now-famous `Five-Ten' amplifier circuit which was introduced in the summer of 1954. In subsequent years, a
succession of designs for high-quality circuits has issued from the Mullard Applications Research Laboratory, and
these designs include circuits for power amplifiers and preamplifiers, circuits suitable for tape-recording equipment
and, most recently of all, circuits for stereophonic reproduction.

Many of these circuits have been described in articles in various magazines or have appeared in the form of booklets
or leaflets. The others have only recently been released by the laboratory, and have not yet been published. The object
of this book is to present the most up-to-date versions of the published circuits together with the new stereophonic
circuits in a way which will be useful and convenient to equipment manufacturers, service engineers and home
constructors. A list is given on page vi of the earlier Mullard publications which have been incorporated in this book.
These publications may be useful for purposes of comparison, but it should be appreciated that the information given
here is the latest available and consequently, supersedes that issued elsewhere.

Most of the material contained in this book is based on work undertaken at the Mullard Applications Research
Laboratory. The largest contributions have been made by R. S. Babbs, D. H. W. Busby, P. F. Dallosso, C. Hardcastle,
J. C. Latham and the late W. A. Ferguson.

While the circuits described in this book have been designed by Mullard engineers. it should be noted that Mullard
Limited do not manufacture or market equipment or kits of equipment based on the circuits. Mullard Limited will not
consent to the use of the 'Mullard' trademark in relation to equipment based on the circuits described, but will not
object to appropriate references to Mullard circuits, specifications or designs. Also, the information contained in this
book does not imply any authority or licence for the utilisation of any patented feature.
 M ULLARD AUDIO PUBLICATIONS

The dates given below are the dates of the first appearance of each publication.
The reference numbers are for the latest reprints of the publications.

5-valve, 10-watt, High Quality Amplifier Circuit August 1954 (out of print)

High Quality Sound Reproduction* August 1955 (out of print)

`3-3' Quality Amplifier Circuit August 1955(TP352)

Circuits for Tape Recorders August 1956 (TP430)(out of print)

Supplement to Circuits for Tape Recorders August 1957 (TP322)

7-watt D.C./A.C. High Quality Amplifier August 1957 (TP325)

Four-channel Input-mixing Amplifier August 1957 (TP319)

Two-valve Pre-amplifier August 1957 (TP339)

Circuit for self-contained Tape Amplifier August 1958 (TP356)

*The section of this book headed `Design for a 20-Watt High Quality Amplifier' was reprinted from articles which
were first published in Wireless World, May and June, 1955.
 CHAPTER 1

Amplifying Systems
1n the last ten to fifteen years, remarkable progress has been made in MONAURAL SYSTEMS
the field of sound recording and reproduction. The introduction of
high-fidelity disc and tape recordings and the improvement in the The choice of line-up is dictated to a large extent by the strength of
design of playback equipment have set new standards for the the signal to be reproduced. For example, the version of the 10W
discerning listener and have consequently created the need for circuit incorporating volume and tone controls can be used
systems providing amplification of the highest quality. Because of the satisfactorily without a pre-amplifier if input signals greater than
variety of signal sources in existence, the range of application of 500mV are available. (Such signals can be obtained from high-output
these systems has had to be wide. Therefore, preamplifying circuits crystal pick-up heads or f.m. radio tuner units.*) The 3W circuit with
which cater for pick-up and tapeplayback heads, f.m. tuner units and volume and tone controls can be used directly with sources which
microphones have assumed great importance. Furthermore, the give signal voltages greater than 100mV.
emergence of stereophonic recordings and pick-up heads has created
a demand for dual-channel amplification and this has been fulfilled in However, signal voltages from low-output crystal pick-ups, magnetic
equipment consisting of either self-contained stereophonic amplifiers pick-ups and tape-recording heads (which can have values of up to
or specially-designed stereophonic pre-amplifiers coupled to pairs of 500mV but are usually between about 2 and 100mV) are normally
conventional power amplifiers. This book contains descriptions and too low for direct use with the power amplifiers, and a preamplifier is
details of high quality amplifying and pre-amplifying circuits for necessary to increase the sensitivity of the system. In addition, discs
monaural and stereophonic applications which can be classified as and tapes are deliberately recorded in such a manner that the strength
shown in Table 1. Most of these circuits have been designed to be of signals derived from them will be a function of frequency (see
used with each other, and examples of various arrangements of the Chapter 2). Consequently, the strength of these input signals will
items of equipment are discussed in subsequent paragraphs of this usually vary in some way not related
chapter.
• The circuits described in this book can be used with the better quality f.m. tuner units
available commercially. No tuner unit is described here because it is felt that the inherent
difficulties of alignment cannot be overcome with the test equipment currently available
to the home constructor.

TABLE 1
Classification of Mullard Circuits

Power Amplifiers of General Application Power Amplifiers of Special Application

Five-valve, twenty-watt circuit (`5-20'); Seven-watt, d.c./a.c. circuit;
Five-valve, ten-watt circuit * (`5-10'); Three-watt tape circuit (Type A revised);
Three-valve, three-watt circuit * (`3-3'). Seven-watt stereophonic circuit;
Two-watt stereophonic circuit.
Pre-amplifiers of General Application
Three-valve circuit; Pre-amplifiers of Special Application
Two-valve circuit; Tape-equalising circuit (Type C revised);
Input-mixing circuit. Twin-channel stereophonic circuit.
* Two versions: one with, one without tone and volume controls.

to the original sound patterns and, to obtain realism from such The simple arrangement for the 10W and 3W amplifiers connected
recordings, some degree of compensation is required in reproducing directly to the input source is shown in Fig. 1. The use of
these sounds. This equalisation is best incorporated in a pre-amplifier, pre-amplifiers is indicated in Fig. 2.
so that the scope of application of power amplifiers can be increased
without any resultant fall in the standard of performance. The 2- and
3-valve pre-amplifier circuits described in Chapter 9 have been
designed to precede either the 3W, 10W or 20W amplifier.
The 3 W tape circuit comprises a self-contained record-playback
system for microphone and radio input sources. The circuit is also
suitable for replaying pre-recorded tapes. A programme-monitoring
signal is available when recording and, during playback, this source
will provide a low-level signal suitable for driving a power amplifier
if an output power higher than 3 W is required.

STEREOPHONIC SYSTEMS

For stereophonic reproduction, two separate amplifying channels are

The tape pre-amplifier provides recording facilities comparable with
required, one for each of the twin signals obtained from the
the self-contained unit but, on playback, the circuit only provides an
stereophonic signal source. In the simplest arrangement, the 2W or
equalised signal for driving a power amplifier. A simple arrangement
7W stereophonic amplifier is connected to the source as shown in
for the combination of tape pre-amplifier and power amplifier is
Fig. 6. If the stereophonic pre-amplifier is used, it should be coupled
shown in Fig. 3. This line-up is suitable also for the 3W tape
to two power amplifiers and loudspeaker systems as in Fig. 7.
amplifier when the low-level playback output is used.

For a complete record/monitor-playback system using pick-up input

sources, the tape pre-amplifier should be used in conjunction with the
line-up of Fig. 2. The new arrangement is shown in Fig. 4*. When
recording, the input signal is fed to the appropriate input channel of
the pre-amplifier and then from the programme-recording output of
the pre-amplifier to the input of the tape circuit. The programme can
be monitored by using the normal output of the preamplifier to drive
the power amplifier. For playback, the equalised output from the tape
pre-amplifier is fed via the Auxiliary input channel to the main
amplifying chain.

Stereophonic systems should be suitable for reproduction from

monaural sources, and `compatibility' switches are usually
incorporated for this purpose. For instance, the switching indicated in
the preamplifier stage of Fig. 7 selects the input channels* and also
sets the equipment for normal or reversed stereophonic reproduction
or for monaural reproduction using both power amplifiers and
loudspeaker systems.

* It is important that the tape deck in both these arrangements should always be switched
to the playback position before the connection from the tape output to the pre-amplifier
input is made. 1f the tape recorder is in the record position when the connection is made,
oscillation will occur.

To feed several inputs into an amplifying system at the same time, the
main pre-amplifier in Fig. 2 should be replaced by the input-mixing
pre-amplifier. Recording, monitoring and playback are possible with
the new arrangement shown in Fig. 5.* The output impedance of the * Facilities are provided for stereophonic radio signals in anticipation of stereophonic
mixer is low so that the inputs of the tape pre-amplifier, and the transmissions.
power amplifier can be connected simultaneously to the mixer
without affecting the quality of reproduction.
POWER SUPPLIES

Facilities are provided in the 10W and 20W amplifier circuits for
obtaining power supplies for a singlechannel pre-amplifier or f.m.
radio tuner unit. For stereophonic applications, the h.t. supply for
each channel of the pre-amplifier can be taken from the power
amplifier used with 'that channel.
The h.t. supply facilities for pre-amplifiers are also available in the
3W circuit. Those for f.m. tuner units are not immediately so, but the A separate power stage is described for the tape pre-amplifier, and
modifications required are not extensive. The current drain when an this can be retained when the preamplifier is used with any of the
f.m. tuner is coupled to the amplifier is increased considerably and it power amplifiers. Physically, however, the extra unit may prove an
will be necessary to choose a rectifier and mains transformer which encumbrance and consequently it may be preferable to use the supply
will supply this current. Extra stages of decoupling will be required to stage of the main amplifier to provide the needs of the tape
prevent excessive hum voltage being picked up in the sensitive early pre-amplifier.
stages of the tuner.

NOTE: The word monaural is at present commonly used to depict conventional 'sound reproduction in which a single-track
recording and a single pick-up head (or tape playback head) are used. The word monophonic may often be encountered as a
synonym of monaural and may eventually supersede it. Monaural means, literally, of one ear and a monaural effect is that of
listening with one ear, when no impressions of space or displacement are conveyed to the listener. Monophonic means of one
sound source so that monophonic reproductions again lack any impressions of spread or position. Monaural has been adopted
in this book simply because it is currently more popular: its use should not be taken to imply any etymological superiority of
the word over monophonic
 CHAPTER 2

Sources of Distortion in Recorded Sound

Some distortion is always present in the sound reproduced from a distortion will be reduced considerably, added to which there will be
gramophone record or magnetic tape. It is inevitable that part of this a better distribution of high frequencies from the smaller source.
distortion will be distortion recorded on the disc or tape itself, but the
loudspeaker, pick-up and recording head also make substantial Dual systems using twin direct radiators are available in two forms:
contributions. A carefully constructed amplifier is far from being the either with two completely separate units, each mounted in its own
weakest link in the chain: in fact, it is improbable that the limit to the enclosure or, alternatively, with the high-frequency unit mounted on
quality of reproduction which can be obtained will be set by the some form of modified frame, making it an integral part of the
performance of the amplifier. High-quality reproduction will only be low-frequency unit.
achieved if careful attention is also paid to the sound input source and
to the loudspeaker. Greater improvement may be obtained by using three loudspeaker
units together with a three-way cross-over network. This extends the
Even if perfect associated equipment could be obtained, however, it frequency range and reduces the distortion still further, as well as
would not be possible to eliminate distortion entirely. There would providing a wider range of distribution patterns.
still be some distortion - a comparatively small amount it is true
which is bound to occur when discs or tape are used. The function of Normally, suitable cross-over networks are recommended by the
this chapter is to give brief accounts of the main items of associated loudspeaker manufacturers, and it is usually best to adopt these
equipment needed for disc or tape reproduction and also to discuss recommendations. However, an example of a two-way network
some of the causes of distortion, indicating the methods by which having a very compact design and unusually low losses is shown in
imperfections are minimised. Fig. 1. Two Mullard Ferroxcube pot-core inductors, type LA23, are
used in conjunction with the special metallised-paper capacitors now
available, and a maximum rate of attenuation of 12dB per octave can
LOUDSPEAKERS be obtained from the constant-resistance circuit.
The choice of a suitable loudspeaker is very important for both disc
and tape reproduction. While it would be wrong to assume that price
is the only guide available, normally -the higher grades of
loudspeaker will justify their extra cost by giving correspondingly
better reproduction. The choice should never be made hastily. It is not
sufficient to listen; to one or two records because it is very easy to
construct a system which will sound impressive at first hearing.
Extended periods of listening to music are desirable to discover if
mental fatigue is likely to be caused by the system. Furthermore,
because it is easier to judge speaker quality by listening to
comparatively simple sounds rather than to full orchestral recordings,
the system should also be tested by assessing the fidelity with which The values of L and C in the diagram are given in terms of the
good-quality broadcast speech is reproduced. loudspeaker impedance R and the crossover frequency f by the
following equations:
A simple, direct-radiator loudspeaker - that is, a single-cone unit
handling the entire range of frequencies- can give excellent results L = R√2
when mounted in a suitable enclosure. It is important to remember 2∏f
that this enclosure should always be that specified by the speaker
manufacturer since it is not possible to design a suitable cabinet C = √2
without knowing accurately the basic performance of the loudspeaker 4∏fR
unit. One of the main disadvantages of all types of simple direct
radiator is that the high-frequency response is very directional. This The approximate number of turns required on the inductor for a given
fact, however, can to some extent be overcome by using the inductance L is 53√L, where L is in millihenries.
loudspeaker in conjunction with some form of reflector. The reflector
can be built into the cabinet, or the corner of a room can be used for In practice, the impedance R of the speaker will already be
the purpose. The effect of the reflector is to distribute the determined, and it will be convenient to obtain the value of
high-frequency components in a more random manner so that, with capacitance to give the required cross-over frequency from the
the adequate treble boost provided by the amplifier, good results will appropriate section of Table 1. The corresponding value of
be obtained. inductance, and the approximate number of turns to give this value,
can then also be obtained from the table.
There is no doubt, however, that even the simplest of dual systems is
much better than a single unit. Such a dual system may consist of two One advantage of horn-loaded loudspeakers over direct-radiator types
direct radiators- for example, a 12 in, unit for the bass-end of the is that the effective damping is increased, and, consequently, there is
spectrum and a 5 in. unit for the middle and treble components- used improved reproduction of transients. Some units incorporate a
in conjunction with a suitably designed cross-over network. It must
be remembered that, while the effective frequency response may not
be very much increased by this type of dual system, various forms of
 combination of a direct-radiator bass speaker and a horn-loaded treble
speaker. They enable a higher conversion efficiency to be obtained in
the treble region of the spectrum, thus ensuring a more even
distribution for a given frequency response. In other words, the total
energy output of the high-frequency unit is maintained up to a higher
point in the musical scale. Although it is possible to use a completely
separate horn-loaded tweeter, at least two manufacturers use a
double-magnet system which permits the treble unit to be mounted
concentrically with the bass unit. Thus, phase distortion at the
cross-over frequency, which is produced in other systems by the
distance between the two units, is also avoided. Also, the overall
frequency response can be made smoother with the concentric
system.

The most complicated form of dual loudspeaker system is that which

is horn-loaded over its entire range. This may be achieved by using
separate horn assemblies, each associated with a loudspeaker unit
covering part of the frequency spectrum. Alternatively, loudspeakers
of the concentric type can be used. Such systems are always
expensive but they have certain advantages: the electro-acoustic
conversion efficiency is high, the acoustic damping for the entire
audio range is much improved over that of any simple direct-radiator
system, and the larger size of bass horn gives better realism.

DISC EQUIPMENT

Magnetic recording heads are normally used for cutting discs: the
stylus is attached either to the moving iron armature or to the moving
RECORDING CHARACTERISTICS
coil. The velocity at which the stylus vibrates (measured as the stylus
passes its equilibrium position) is directly proportional to the
Cutting Styli
recording-signal strength, so that, for a constant voltage, the velocity
is constant, and the amplitude or width of the groove is inversely
When recordings are made on discs, the recording head is used to
proportional to the signal frequency. This is known as
convert the signal voltage (or voltages, in stereophony) from the
`constant-velocity' recording.
recording amplifier into vibrations of the cutting stylus. The stylus
cuts the groove in the disc during recording, and the vibrations are
If the amplitude at high frequencies is chosen to give an acceptable
reproduced in this groove as programme modulation. With monaural
signal-to-noise ratio, the amplitude at low frequencies will be
recordings, the single programme modulation appears as an identical*
excessive. Distortion will be high and too much space will be needed
trace in both walls of the groove. With stereophonic recordings,
between adjacent rings of the groove to ensure that breakthrough
however, the twin signals give rise to a more complex trace in which
from one ring to another does not occur. Thus, some restriction of the
each wall of the groove carries one signal.
amplitude is desirable at these low frequencies and, to achieve this,
bass signal voltages from the recording amplifier are attenuated
The cutting edge of a monaural stylus is rectangular and it cuts with a
before being recorded. Recordings in which the maximum width of
lateral movement so that the width of the recording groove is
the groove is restricted to some limit are known as
constant*. A stereophonic cutting stylus is triangular and has two
`constant-amplitude' recordings.
cutting edges which are set symmetrically about the vertical at
right-angles to each other. Each signal causes a cutting edge to move
Because of the difficulty in maintaining a satisfactory signal-to-noise
in a direction perpendicular to its length, so that the resultant
ratio if a constant-velocity characteristic is used at high frequencies, it
movement of the stylus is not a simple one from side to side as with a
is normal to boost treble signals from the recording amplifier before
monaural stylus, but is two-dimensional in the plane of cutting edges.
recording them. Consequently, a recording characteristic indicating
Thus, whereas a monaural stylus needs only freedom of lateral
the variation of recording voltage with frequency can be divided into
movement, a stereophonic stylus must have vertical compliance as
three sections: the bass section, showing voltage attenuation; the
well.
middle section, showing constant-velocity recording; the treble
section, showing voltage boost.
Characteristics

Except for the duality of stereophonic discs, the characteristics of

monaural and stereophonic recordings are fundamentally the same.
Therefore, it is sufficient in these simple comments on recording
characteristics to refer only to monaural recordings. The remarks can
be applied, with suitable allowance for their two-fold properties, to
stereophonic discs.
 shown in Fig. 3. In 1955, however, the majority of these companies
agreed to adopt the characteristics of the Recording Industries'
Association of America (RIAA), which are drawn in Fig. 4.

An imaginary recording characteristic is drawn in Fig. 2. The

frequencies at which the characteristic changes from one section to
another are called the `turnover' frequencies. The slope of the bass Consequently, there is no need, with up-to-date recordings, to adapt
and treble sections of the characteristic depends on the degree of the pick-up head and record to each other by using separate
attenuation or boost applied to the recording signal. If recordings with equalising networks for each brand of record. Instead, one network
a constant amplitude of modulation are required, the stylus velocity (usually in the form of a frequency-selective feedback network) can
must be proportional to the frequency and hence the signal voltage be incorporated in the amplifier. Of course, with the two
must also be proportional to the frequency. Thus, the signal voltage characteristics for standard and microgroove records, separate
must be halved or doubled as the frequency is halved or doubled and networks will still be required for these.
this gives a characteristic slope of 6dB per octave.

Actual recording characteristics differ considerably from the

maginary one of Fig. 2. The straight lines of the characteristic cannot
be achieved in practice, and actual characteristics are more like the
continuous curve shown in Fig. 2.

The bass and treble turnover frequencies in actual curves are defined
as those frequencies at which the ratios of the recorded voltages to
the true signal voltage are 1:√2 and √2:1 (that is, - 3dB and + 3dB)
respectively. The true signal voltage is assumed, for reference
purposes, to be the level of voltage recorded at 1kc/s.

In most practical characteristics, the middle horizontal section is

discarded and only a point of inflection at 1kc/s retained. The bass
and treble sections of practical curves are not produced indefinitely
with a constant slope, but are flattened at the limits of the range of
audible frequencies.
PICK-UP HEADS
PLAYBACK CHARACTERISTICS
The principle on which a gramophone pick-up head operates can be
If a magnetic pick-up head is used to reproduce the sound recorded one of several, but the commonest types of component belong to
on discs, the output voltage from the head will be proportional to the either the constant-velocity or the constant-amplitude group.
velocity of vibration of the needle. Consequently, the output from a Magnetic pick-up heads (moving-iron and moving-coil) have
magnetic head used with recordings made to characteristics of the basically constant-velocity characteristics, the output voltage being
type shown in Fig. 2 will increase with frequency in the bass and proportional to the velocity of vibration of the pick-up stylus.
treble regions. (This will not be so with crystal pick-up heads as the Piezo-electric (crystal) heads possess constant-amplitude
output of these is proportional to the amplitude of modulation.) Thus, characteristics, the output from these varying with the amplitude of
correction or `equalisation' will be required during playback modulation in the recording groove.
amplification to restore the true level of the signal voltage, and this
equalisation should be the converse of the attenuation and boost As the RIAA recording characteristics (and, similarly, most of the
applied while recording. commercial characteristics before 1955) do not possess purely
constant-velocity or constant-amplitude properties, some degree of
The recording characteristics used by the different recording equalisation will be needed with either group of pickup head, and the
companies before 1955 followed, to a greater or lesser extent, the degree will be different for each group. However, simple
curve shown in Fig. 2, but the differences were large enough to make resistance-loading of lower medium-output crystal heads will convert
equalisation a matter depending greatly on the records to be played. their characteristics into very near approximations of
Many companies issued their own equalisation characteristics, and constant-velocity characteristics without prejudicing the standards
corrective networks had to be designed with these in mind. Examples demanded of high-quality equipment. Consequently, this method of
of the playback characteristics of the major recording companies are loading is often adopted in high-quality amplifiers to give a simpler
switching arrangement of the equalising feedback networks. With it, frequency. A typical output voltage from a high-output crystal head
the networks provided for standard and microgroove records using would be about 1V at 1kc/s. The output at a frequency of 300c/s
magnetic heads can also be used for medium- or low-output crystal would therefore be about 3V. It is thus evident that attenuation is
heads. desirable at the lower frequencies to provide a balanced output and to
prevent the input valve from being over-driven. The highest-quality
Magnetic Pick-up Heads crystal heads have outputs considerably lower than this, a typical
value being about 0-5V at 1kc/s.
Both types of magnetic pick-up head -that is, moving-iron and
moving-coil - work on the principle of the dynamo: the change in The impedance of crystal pick-up heads is equivalent to a
magnetic flux through a conductor generates a voltage across the capacitance, the value of which is of the order of 1000pF. The
conductor. The change of flux in the heads results from the optimum load resistance of a crystal head will depend on the method
movement of either the magnetic core or the conducting coil, and this of loading. If it is loaded so that its characteristic is similar to that of
movement is caused by the vibration of the stylus (attached to either a magnetic pick-up head, the load resistance is of the order of 100kΩ.
the core or the coil) in the recording groove. If the characteristic is not modified, then an input impedance of the
order of megohms is desirable.
The output voltage from either type of head is proportional to the rate
of change of flux through the conductor and thus to the velocity of Stereophonic Pick-up Heads
the stylus (measured at its equilibrium position). Output voltages
which are quoted for comparing magnetic pick-up heads should Pick-up heads for use with stereophonic equipment can be of either
therefore be given with reference to the stylus velocity. This the magnetic (moving-iron) or the crystal type. The construction of
reference can be made directly by quoting the velocity; alternatively these pick-up heads, however, will differ considerably from that of
the ratio (expressed in decibels) of the actual output voltage to that monaural heads. They must be capable of responding to the
which would be obtained at some standard velocity (1cm/sec) should two-dimensional modulation of the recording groove and of
be given. Hence, an output voltage could be given as, say, 100mV at translating the response into two separate signals, whereas monaural
3.16cm/sec or 100mV at +10dB, and both expressions would be heads have only to follow a lateral displacement and convert it into a
equivalent. single signal. Consequently, there must be two signal generators in
each stereophonic head and these must have very similar
Moving-iron magnetic pick-up heads can be divided into three characteristics. They will therefore both be of the same type - both
classes, governed by their outputs (measured at a stylus velocity of crystal or both magnetic. Also, the pick-up stylus will require a
3.16cm/sec): vertical compliance comparable with its lateral compliance so that the
tip of the stylus can follow the contours of the recording groove
Low-output heads: Voltages below 20mV faithfully and provide balanced stimuli for the twin generators.
Medium-output heads: Voltages between 20 and 100mV
High-output heads: Voltages above 100mV. The output voltages obtained from these generators are very much
lower than the voltages derived from monaural heads of the same
The basic output of a moving-coil pick-up head is of the order of type. Moving-iron, stereophonic heads will produce voltages of the
millivolts. The lower level of output results from the desirability of order of millivolts for each signal, and crystal heads will give signals
lightness of the pick-up unit and hence the restriction on the number of the order of 100mV. The impedance of each section of a
of turns of the coil. stereophonic head will be of the same order as that of the same type
of monaural heads-about 1 kilohm for magnetic heads and the
Both types of head are suitable for high-quality equipment, although equivalent of about 1000pF for crystal heads.
the intrinsically low output of the moving-coil type necessitates the
use of a high-ratio pick-up transformer between it and the amplifier. DISCS
The impedance of magnetic heads is of the order of kilohms at signal
frequency of 1kc/s, and the required loading (that is, the input Records are normally made for replaying at three different speeds:
impedance of the amplifier) for the rated output of the heads is of the 78, 45 or 33⅓ revs/min. The first or `standard' type of disc is pressed
order of 100Ω. Their loading should be independent of frequency, in a hard non-pliable material called shellac. Record wear with this
otherwise the input signal to the amplifier will also vary with material is relatively high even under good playing conditions. Wear
frequency. with vinyl, the pliable material used for the 45 and 33⅓ revs/min,
'microgroove' type of disc, is very slight under comparable
conditions. The main trouble with vinyl records is the tendency to
Crystal Pick-up Heads collect dust through electrostatic attraction.

A voltage is developed between the faces of a piezoelectric crystal Standard records are made with diameters of 10 or 12 inches and
when the crystal is strained. The voltage is proportional to the strain. provide playing times of about 4 or 5 minutes. `Extended-play' or 45
In a crystal pick-up head, the playback stylus is attached rigidly to revs/min discs have diameters of 7 inches and playing times of about
some piezo-electric material, and a strain is produced in the material 10 minutes. `Long-playing' or 33⅓ revs/min records are made with
by the vibration of the stylus as it follows the modulation of the diameters of 10 and 12 inches and play for up to about 30 minutes.
recording groove. The amplitude of this modulation governs the Stereophonic discs are produced in `Extended-play' and
output voltage of the head. `Long-playing' forms.

The rated output voltage of a crystal pick-up head is always quoted at Several types of pick-up stylus are available (particularly for use with
a particular frequency. This is essential because, for comparison standard monaural records), but the most common for both standard
purposes, a recording made to a constant-velocity characteristic is and microgroove discs is the permanent or jewel-tipped stylus.
used and the amplitude of modulation - and, therefore, the output
voltage - of such a recording is inversely proportional to the
Diamond styli are the hardest wearing, but sapphire types are very PINCH EFFECT
popular.
The sine-wave groove of Fig. 5 illustrates the so-called `pinch effect'
TRACING DISTORTION which results from the varying width of a modulated groove on a
monaural recording. Because of this varying width, the pick-up stylus
One form of distortion which occurs with monaural disc recordings is will have to ride up and down as it traverses the groove (Fig. 6). At
that resulting from the difference in shape of the recording groove the end of a microgroove recording, the maximum vertical movement
and the rounded playback stylus. The way this distortion arises can may be as much as 11% of the maximum lateral displacement. The
best be seen by considering the width of the recording groove. output of some pick-up heads varies slightly with this vertical
movement so that the pinch effect can be an additional source of
When no sound is recorded, the groove is un-modulated and the distortion. As the needle has to move up and down twice during each
cutting face of the stylus is at right angles to the length of the groove. cycle, the distortion is of even harmonic order.
The width of the groove is uniform and is the full width of the cutting
edge. With a signal, however, the stylus will be displaced from the NEEDLE-SCRATCH
un-modulated position and the cutting face is at an acute angle to the
direction of the groove. Thus the width of the groove is not uniform. As the pick-up stylus traces the groove, the fine particles in the
This is illustrated in Fig. 5, which shows a sine-wave groove, very material of the disc cause the stylus to make small irregular
much magnified. movements. The voltage set up by these movements is amplified and
is heard as needle-scratch. The noise lies mainly between 2 and
10kc/s

Needle-scratch will be particularly bad when the diameter of the

needle is too small for the groove (Fig. 6). For good reproduction the
stylus should rest on the two sides of the groove. If it rests right in the
trough, it may ride up the walls, thus giving excessive scratch and
additional distortion. If it rests on top of the groove only, it cannot
follow the groove correctly, and there will then be a tendency for the
stylus to `skate' across the record.

The width of the groove in the direction a-a always corresponds to

the full width of the cutting edge of the stylus. The true width of the
groove - that is, the width at right-angles to its length - therefore
depends on the angle between the modulated groove and the direction
of the un-modulated groove, and will be smaller for greater values of TRACKING ERROR
this angle. The true width of the groove is the same as the width of
the cutting edge only at the peaks of the sine wave. Another important difference between the making and replaying of
discs is that the pick-up stylus as it traverses the record does not
If distortion is to be avoided, the sideways movement of the playback follow the same path as the cutting stylus. The cutting stylus is set to
stylus should reproduce exactly the movement of the cutting stylus. move inwards in a straight line along a radius of the disc, whereas the
The circles in Fig. 5 represent the point of the pick-up stylus as it pick-up stylus must necessarily be mounted on a pivoted tone arm,
rests in the groove during playback, and the centres of these circles and must cross the record in an arc of a circle. The playback stylus is
should therefore lie on the line traced by the centre of the cutting thus not always held at right-angles to the groove and the output will
stylus. The figure shows that this condition is not fulfilled: the dotted vary slightly. This effect is greater for large records, and can be
line traced by the centre of the pick-up stylus is not the same as the reduced by using a. longer pick-up arm.
chain line traced by the centre of the cutting edge. The amount of TURNTABLE DRIVING SYSTEMS
distortion resulting from this lack of coincidence depends on the
difference between these two lines, measured parallel to a-a. There is It is essential that the mechanical driving systems of disc-playing
no distortion at the peaks of the sine wave, nor in the un-modulated equipment should be accurately made, especially with stereophonic
position. equipment. Rotation of the turntable must be extremely steady. These
requirements must necessarily be very severe if the mechanical
Tracing distortion consists of odd harmonics. The distortion increases system is to match the performance expected from the acoustical and
as the modulation becomes more spiky. Thus, it increases with the electronic equipment.
amplitude of modulation - one reason why constant-amplitude
recording is necessary at low frequencies. It also increases with In a high-quality installation in which the loudspeaker is separate
frequency, which is a reason for restricting the amount of treble boost from the main chassis, the system can be mounted solidly to the
used when recording. The distortion is also greater if the modulation motor board, and the motor board itself floated on rubber to ensure
is compressed, as it will be near the centre of the record. that no mechanical vibration is transmitted to the cabinet.

An unsuitable system can spoil the loudspeaker output in three

objectionable ways: `wow', `flutter' and `rumble'.
Wow and flutter are produced by variations in the speed of the when a long note is being played. A sufficiently steady speed is
turntable. Those variations occurring at low frequencies are termed usually achieved by providing a heavy turntable, or some other
wow; flutter results from faster fluctuations. Slow variations can be moving part, to act as a flywheel. Care should be taken to see that the
caused by spindles and other rotating parts being loosely mounted or speed-changing mechanism is not subject to undue wear. This can
off centre, so that the wow resulting is heard at regular intervals. lead to wow or flutter after much use.
Unevenness of the drive, changes of friction (catching or slipping)
and record slip can give rise to intermittent wow. Small irregularities Rumble results because the mechanical parts run noisily. These parts
are normally experienced when driving from a motor. Sudden should run sufficiently quietly for no objectionable rumbling to be
changes in speed of as little as 0-5 will be clearly noticeable as flutter heard at full gain when there is no other input signal.

TAPE EQUIPMENT
the erase-head core should be considerably larger than in the core of
the combined head. The erase gap is usually about fifteen thousandths
RECORDING AND PLAYBACK HEADS of an inch wide. Normally there is no auxiliary gap in the core of the
erase head.
Generally speaking, tape decks (with the exception of some of those
used with the so-called `professional' equipment) combine the duties TAPES
of recording and reproduction in a common head. The requirements
of the recording and the playback heads are fundamentally the same Modern magnetic tapes consist, generally, of a non-magnetic base
and, as they are not required simultaneously, economy is served by (paper, cellulose acetate or p.v.c., for example) coated with a
combining them. magnetic material. The base obviously needs to have strength and
suppleness. The magnetic material (often red or black iron oxides) is
Modern heads are usually constructed of two semi-circular stacks of a very finely divided powder mixed in some binding substance
high-permeability laminations of about a half to one inch in diameter. (lacquer, for instance), and the coating applied to the base material
The stacks are assembled symmetrically with a `head gap' (against has to be very smooth. Unevenness in the surface of the tape will tend
which the tape passes) of about five ten-thousandths of an inch or to cause broken contact with the heads, and consequently an
less, and an auxiliary gap, diametrically opposite the head gap, of undulating level of recording. The magnetic material needs a high
about ten times that width. coercivity to prevent, in particular, very large demagnetisation losses
at treble frequencies, and a high remanence to give a good level of
Both gaps are filled with non-magnetic, metallic `shims' to prevent recording.
any accumulation in the gaps of magnetic material which may be
rubbed off the tape by friction, and to maintain a linear relationship Tapes are usually supplied in 5- or 7-inch reels, giving playing times,
between the flux at the gaps and the current through the exciting coils depending on the transport speed and the thickness of the tape, of
of the head. between about 10 and 90 minutes. The standard width of the tape is a
quarter of an inch, which normally allows for two (and sometimes
The impedance of these coils, which are placed symmetrically about four) adjacent recording paths. With stereophonic tapes, adjacent
each core, should, for recording, be low compared with the source paths are used for the twin signals comprising the programme. The
impedance. This ensures, as far as is possible, a current which is total playing time of such tapes is therefore reduced by half.
independent of the frequency of the signal. For the playback process,
however, maximum output voltages are to be gained from TAPE TRANSPORT SYSTEMS
high-impedance windings. Any conflict between these requirements
can be resolved by feeding the output from low-impedance coils into The basic requirement of the transport system is that a steady speed
a step-up transformer. should be imparted to the tape. Momentary changes in this speed will
cause `wow' or `flutter'. (Wow results from slow variations; flutter
Stereophonic playback heads consist fundamentally of two monaural results from fast fluctuations: the dividing line is arbitrary.)
heads mounted one on top of the other. Each head forms part of a
complete playback system, the input signal for which is derived from The actual speed of transport has a considerable bearing on the
one of the tracks of specially recorded stereophonic tape. Facilities performance of the recording apparatus. The higher the speed, the
for domestic stereophonic recording are seldom provided because of better the performance at the higher frequencies. This is closely allied
the difficulties to be encountered and the advanced technique to the size of the gap in the recording head. But, of course,
required for such an operation. economically, great speeds are a disadvantage: the playing time of the
tape is obviously reduced. Most transport systems are standardised
for playing speeds of either 1⅞, 3¾, 7½ or 15 inches of tape per
ERASE HEADS second, with possibly, a choice of speeds.
Although similarities exist between the erase head and the recording AMPLITUDE DISTORTION
head, they differ in construction in two principal respects: Firstly,
appreciable power (about 2 to 4W) is required for efficient erasure, The non-linear relationship between the magnetism residing on the
while that necessary for the recording head is usually about 1mW. tape and the magnetic field inducing it, will cause considerable
Hence the core of the erase head must be of a material having a amplitude distortion in the recording.
higher point of magnetic saturation than the material in the recording
head. Secondly, the longer the tape is under the influence of the
erasing field, the more complete will be the cleaning. Thus the gap in
 of upper sections of the characteristics, and the `clipped' reproduction
would contain a high percentage of odd harmonics of the original.

The curve OB3B5 in Fig. 7 shows a typical relationship between the

magnetisation produced in a magnetic material and the increasing
magnetic field producing it. A field strength of H2 units, for example,
would produce a degree of magnetisation corresponding to the point The middle sections of the arms of the transfer characteristics are,
B2, provided the material was unmagnetised initially, and also however, approximately linear, and if the variation in the magnetic
provided the field strength was not, at any time during its general field strengths are confined to these portions, then the recording will
increase from zero to H2, caused to diminish. The levels of be a relatively undistorted replica of the original sound. If a constant
magnetisation corresponding to B1, B2, etc., are those obtaining when current is fed into the recording head together with the signal current,
the field strengths are actually H1, H2, etc. These do not represent the effect is to `lift' the variations in the magnetising field above the
residual magnetism. If the fields are reduced from H1, H2, etc., to lower curvature of the characteristic (Fig. 8). The d.c. bias induces a
zero, the induced magnetism decreases along the paths B1Rl, B2R2, constant degree of magnetisation in the tape on which the audio
etc., respectively, and the points R1, R2, etc., denote the degrees of variations are superimposed. Only the variations of the residual
residual magnetisation, or remanence, induced in the magnetic magnetism will appear on replaying the recording, so that the bias
material by maximum applied field. strengths of H1, H2, etc. will not be translated into sound (in fact, some noise does result from
the biasing remanence). But because of the limitation imposed by the
The curve OC3C5 plotted in Fig. 7 gives a typical `transfer' saturation curvature of the characteristic, the fact that the bias
characteristic for a magnetic material; that is, it shows the intensity of magnetism is added to the audio magnetism means that the amplitude
residual magnetism (R1, R2, etc.) resulting from any given maximum of the signal must be restricted quite considerably. The
magnetising g field (H1, H2, etc.). Complete transfer characteristics signal-to-noise ratio is thus lowered when d.c. bias is used.
for positive and negative magnetising fields are shown in both Figs. 8
and 9. This method of preventing amplitude distortion by introducing a
constant biasing current into the recording head together with the
signal current has has been superseded by a method in which the
direct current is replaced by a high-frequency, alternating current.
The result is a recording relatively free from harmonic distortion in
which a good signal-to-noise ratio is maintained. No complete
explanation of the mechanism of a.c. biasing has yet been accepted
generally, but some idea of the process can be obtained from Fig. 9.

A feature of this method of biasing is the fact that no residual

magnetism is induced in the tape by the a.c. signal, provided, of
course, that these oscillations are free from even harmonic distortion.
As any point on the tape passes the gap in the recording head, it is
subjected to a rapidly alternating magnetic field, the strength of
which increases as the point approaches the gap and dies away as the
point recedes. Such a process causes no remanence in the tape. Only
the audio variations, which are superimposed on the a.c. field, cause
any residual magnetism, and, because of the bias, this magnetism
depends linearly on the audio signal. Thus, the tape is not loaded with
It is obvious from these characteristic curves that the degree of any `wasted' remanence as it is when d.c. biasing is used. In addition
residual magnetism is not a linear function of the maximum to this, it will be seen from Fig. 9 that both linear sections of the
magnetising field. Because of this non-linearity, any signal applied to transfer characteristic can be used. It follows then that the limitations
the microphone would, on recording, suffer distortion. All signals imposed by d.c. biasing on the strength of the audio signal can be
(for example, the `unbiased' sine wave shown in Fig. 8) would suffer relaxed considerably without increasing the danger of approaching
`bottom-bend' distortion resulting in even harmonics of the the saturation level of the tape. A.C. bias, therefore, reduces `bottom
fundamental appearing in the recording. Large signals would be bend' distortion and, compared with d.c. biasing, gives an improved
further distorted by the magnetic saturation depicted by the flattening signal-to-noise ratio.
Initially, an increase in the a.c. biasing current for a given signal accordance with the recommendations of manufacturers of tape heads
strength lessens the distortion of the recorded signal and increases the and oscillator coils.
level of recording. But too much bias will cause `top-bend' distortion
and will also cause a lessening in the output, especially for treble FREQUENCY RESPONSE
signals. (The bias acts very much in the manner of the erase signal -
see below.) A typical variation of output with bias current is shown in
Fig. 10. The degree of magnetisation of the tape which results from recording
with a current through the head which is constant at all frequencies is
of the form shown by the curve ABC of Fig. 11. The reduced
intensity in the treble region is attributable in the main to the
self-demagnetisation of the tape.

To outline the recording process briefly - the tape is drawn steadily

past the recording head, and the field produced by the signal current
in the head induces some degree of residual magnetism in the section
of the tape nearest to the head. The amplitude of the signal governs
the amount of magnetism, and the sense (whether positive or
negative) controls the direction of the magnetism in each section.

On the basis of the molecular theory of magnetism, the `molecular'

magnets in the tape are aligned by the magnetising field: the strength
of the field governs how many of these magnets are brought into line;
the direction of the field dictates whether the alignment is with north
or south poles leading.
ERASING

It was stated above that the field associated with the a.c. biasing
signal did not cause any residual magnetism in a tape passing through
the field. If the tape is unmagnetised when the signal is applied, it
will be unmagnetised when it has passed through the field. If the
amplitude of the a.c. signal is sufficient to saturate the tape, then even
if the tape is not unmagnetised originally, it will be so when it has
travelled through the field.

This fact is used in the erase heads of most magnetic recording

equipment. The tape is drawn through the strong local magnetic field
arising from a high frequency alternating current in the erase head.
As each point of the tape travels into the field, it is saturated, and as
each point recedes, it is subjected to a field strength which diminishes
to zero some distance from the head, By this process it is cleaned of
all previous magnetisation.

OSCILLATOR COIL
The field set up by a sinusoidal signal current in the head would
The bias oscillator should generate a sine wave with negligible change direction regularly, so that the pattern of magnetism included
harmonic distortion to keep noise to a minimum. Symmetry in the in the tape would have adjacent sections of the tape with north and
waveform is essential. An asymmetrical sine wave has, in effect, a south poles leading alternately (see Fig. 12). Thus, the poles of each
d.c. component made up of the difference between the positive and section of the tape would be adjacent to the like poles of the
negative amplitudes of the sine wave. This will introduce noise in the neighbouring sections.
recording and impart some 'wasted' remanence to the tape.
The length of these sectional magnets in the tape depends on the
It is usual for the oscillator valve to provide both bias and erase speed at which the tape passes the head, and on the frequency at
currents. In most of the commercial tape decks, the combined which the magnetising field changes direction. For a constant
record-playback head is usually of high impedance, while the transport speed, the length varies inversely as the frequency of the
impedance of the erase head may be high or low. To accommodate signal. For low-frequency signals, the magnets are comparatively
low-impedance erase heads, a separate oscillator output, in the form long, but for high frequencies they are short.
of a secondary winding to the oscillator coil, is required for matching
purposes.

Because the characteristics of tape-recorder heads are not

standardised, it is impossible to design an oscillator coil that will suit
all types of head. Some degree of flexibility is possible if a tapped
secondary winding is used, but the oscillator stage of a tape amplifier
circuit may have to be adapted to a greater or lesser extent in
 frequencies. Also, the physical dimensions of the gap in the playback
head cause losses when the size of the sectional magnets and the gap
are commensurate. (This is usually referred to as `gap effect'.) These
combine to produce a more accentuated drop in the response curve in
the treble region.

Unfortunately, equalisation is not simply a matter of compensating

for bass and treble deviations. Several performance requirements are
in conflict, and a compromise is the best that can be achieved. A wide
frequency response, low distortion or a high signal-to-noise ratio can
each be obtained only at the expense of the others. Over a restricted
range, an increase in the bias current, for example, lessens distortion,
but causes treble attenuation and tends to lessen the signal-to-noise
ratio. Furthermore, equalisation is a function of the physical
properties of the tape used. The equalisation provided with any
combination of tape deck and amplifier may, for example, provide a
level response over a wide frequency range with one brand of tape,
but may give rise to a large treble peak with another brand.

The ability of the tape to resist demagnetisation depends, in part, on EQUALISATION

the length of these sectional magnets. If they are short, the opposing
fields set up by the neighbouring like poles will cause misalignment A tape amplifier can be represented essentially by the block
more readily than if they are long. Hence the demagnetising effect is schematic diagram of Fig. 13. Normally, the design of both
greater at the higher frequencies than at the lower ones, and amplifying units will follow general a.f. practice, although attention
consequently there will be less residual magnetism at the higher
frequencies even if the signal amplitudes are constant throughout the
frequency range.

For the playback process, the magnetised tape is drawn past the
playback head at a constant speed, and the magnetic fields associated
with the sections of the tape move and cut the coils of the head,
thereby setting up voltages in the coils. The rate of change of the flux
cutting the coils governs the voltage generated. Thus, the rapidly
changing fields set up by the sections of the tape on which
high-frequency sounds have been recorded must of necessity give rise
to larger voltages than the sections containing low-frequency
impressions. This is so, even if the impressions at both high and low
frequencies have the same intensity of magnetisation. The actual must be paid to certain requirements of the equipment. Precautions
speed of the tape has no bearing on the relative levels of the output at will be required to minimise hum pick-up, and the overall
the various frequencies. An increase in the transport speed simply signal-to-noise ratio should be at least 40dB.
multiplies all levels by the same amount.
The attenuation of the bass response in a tape recorder takes place
The rate at which the playback voltage rises with frequency is mainly in the playback head. The treble response suffers its greatest
equivalent to 6dB per octave. In Fig. 11, the line DEF indicates the losses in the recording process. Thus, in order to load the tape more
frequency response of the playback head only. (A tape which has or less evenly, it is best to provide treble boost in the recording
been ideally magnetised is assumed for this response curve.) The amplifier and to apply bass boost in the playback amplifier. Some
response curve for the recording head alone will depend on various treble accentuation may also be required in the playback amplifier,
practical considerations. For example, a tape made of a magnetic but this has to be limited in order to avoid much amplification of
material of high coercivity is difficult to demagnetise, and if one is background noise.
used, then the demagnetisation losses at high frequencies will be less
than if a low-coercivity tape is used. The use of a high coercivity
tape, of course makes intentional erasing more difficult.

The response for the complete equipment, (that is, the response
combining both recording and playback deviations) is indicated by
the curve DEGH of Fig. 11. To obtain an equalised output from this
response curve it is necessary to introduce boost at both ends of the
frequency range. For ideal equalisation, the compensating curve must
be a mirror image of the curve DEGH, so that the equalising response
curve needs to be the curve IGJ.

There will be, in fact, further high-frequency losses encountered in

tape recording. The tendency of the a.c. bias signal to act as an
erasing signal during the recording process is more pronounced at
high frequencies because the remanence is less deeply seated at these
 CHAPTER 3

High-quality Amplification

The principal features of good amplifying equipment can be otherwise good reproduction. Thus the amplifier should not give rise
summarised as follows: to noise or hum at an audible level.

1. The distortion produced by the amplifier should be negligible up to PERFORMANCE OF AMPLIFYING EQUIPMENT
the maximum output level. Distortion is the presence in the output of
frequency components which were not present in the input signal. The performance of amplifying equipment is normally stipulated with
These consist, in the first place, of harmonics of the original reference to some or all of the points enumerated above, and the
frequencies and, in the second, of sum and difference tones caused by interpretations to be given to the values quoted in subsequent
intermodulation between different frequencies. chapters of this book are discussed in the following paragraphs. In the
chapters on stereophonic equipment, the performance of each channel
2. The response of the amplifier should be uniform throughout the is quoted. Combined values for the two channels can be misleading.
audible frequency range. The average ear will respond to frequencies For example, a reserve of output power (see below) of, say, 5W per
in the range of 30c/s to 15kc/s. The upper limit of this range may channel does not mean that an output of 10W will be delivered before
extend to 20kc/s, particularly with young people, or be well below overloading occurs. Overloading will occur in the left-hand channel if
15kc/s with older people. The upper partials of some musical that channel has to deliver more than 5W, and similarly in the
instruments extend to about 10kc/s and the sound spectrum associated right-hand channel For sound sources placed symmetrically between
with them to about the upper limit of the audible range. To make the right- and left-hand microphones, the total power reserve will be
realistic reproduction possible, therefore, the amplifier should handle 10W, but for sounds originating in either extreme left- or right-hand
a range of frequencies at least as wide as that which can be heard. positions, the total reserve is little more than the reserve of the left- or
right-hand channel
3. The response of the amplifier to signals of a transient nature should
be good. Many sounds, particularly those deriving from musical Power Reserve
instruments, rise very rapidly to a high intensity and decay relatively
slowly. Such sounds are spoken of as `transients', and examples of The audio power available at the output of an amplifier is defined as
these are the sounds resulting from the clashing of cymbals or the (Vload)2 / R1oad, where Vload is the voltage developed across a load
plucking of strings. resistance Rload connected to the output terminals of the amplifier. The
rated output power, or the power reserve, of the amplifier is the
The steeply rising wave fronts of transient sounds can be shown to maximum audio power which can be obtained without exceeding
consist of a wide range of component frequencies. The ability of an either the limiting values of the valves or the level of distortion
amplifier to reproduce these transients faithfully will therefore permitted in the specification for the equipment.
depend on two things: first, the frequency response of the system
must be wide and, second, the phase shift over the whole frequency Distortion
range must be small. Variations in the relative phasing of the
component frequencies of a transient would result in a change in its The principal form of distortion occurring in the output from
aural character. amplifying equipment is non-linear distortion which is normally
divided into harmonic, intermodulation and beat-note distortion. Each
4. An adequate reserve of output power should be available. For subdivision contributes some power to the amplifier output at
faithful reproduction, the sound level should be comparable with that frequencies which differ from those occurring in the input signal.
of `live' conditions. The amplifier should thus be capable of handling
peak powers considerably above the average level to allow peak Harmonic Distortion
sounds to be reproduced without overloading and audible distortion.
Power which occurs in the output at second, third, fourth, and so on,
5. The output resistance of the amplifier should be low. This will harmonics of the fundamental signal frequency comprises harmonic
improve the performance of the loudspeaker and ensure crisp and distortion. It is expressed as a percentage ratio of the power
clean reproduction, particularly of transients. Air-loading of the associated with the particular harmonic to the total output power of
loudspeaker limits, to a large extent, the low-frequency resonance of the amplifier. Total harmonic distortion is the ratio of the power
the cone and suspension. The electromagnetic damping afforded by a associated with all the harmonics to the total output power. The total
low output resistance in the amplifier is, however, effective in harmonic distortion Dtot is given in terms of the individual distortions
ensuring adequate control of the cone movement over the whole D2, D3, D4 , etc. as:
frequency range.
Dtot = √ (D22 + D32 + D42 + ...)
The output resistance should preferably be much less than the
impedance of the loudspeaker voice coil, the ratio of the two being
termed the `damping factor'. 1n practice, a damping factor of above Intermodulation Distortion
10 is desirable.
If an input signal contains several different frequencies, any
6. The level of hum and noise should be low: the presence of these non-linearity in the amplifier will give rise to modulated waveforms
in the loudspeaker output will detract from the enjoyment of having frequencies which are the sums and differences of the
interacting waveforms. The extent of this distortion is assessed by phase) thus reducing sensitivity. It is usual to refer to the amount of
measuring the degree of interaction between two pairs of test signals. feedback in terms of the ratio of voltage gain of the amplifier without
The interaction between signals of very different frequencies is feedback to the gain with feedback. Thus feedback of, say, 26dB
quoted as intermodulation distortion and that between signals of would mean that the gain without feedback is 20 times the gain with
nearly equal frequencies is given as beat-note distortion. feedback.

Intermodulation distortion is measured between test signals having The gain of an amplifier without feedback must therefore be great
frequencies of 40c/s and 10kc/s. The ratio of the peak-to-peak enough to allow for the reduction in gain caused when the feedback is
amplitudes of the l.f. and h.f. signals is 4 : 1. The output obtained applied. This drawback is outweighed by the following advantages to
with the two signals is taken to be equivalent to the output obtained be obtained from using feedback: (1) reduced distortion; (2) wider
with a single sine-wave signal the peak-to-peak amplitude of which is and flatter frequency response; (3) reduced output impedance; (4)
equal to the peak-to-peak amplitude of the combined waveform. The reduced phase shift; (5) less dependence on small changes in supply
distortion is quoted as the r.m.s. value of the amplitudes of the sum voltages, etc.
and the difference waveforms, expressed as a percentage of the
amplitude of the h.f. signal. Circuit Design

Beat-note distortion is measured with two test signals having Although the power-handling capacity of an audio amplifier is not the
frequencies of 14 and 15kc/s. The amplitudes of these signals are property which is most important to the listener - a low level of
equal, and the resulting difference waveform, expressed as a distortion is usually considered to be so - it is nevertheless the prime
percentage of the test signal amplitudes, gives the beat-note concern of the circuit designer. The reserve of output power required
distortion. The equivalent output power of the combined waveform is for realistic reproduction of orchestral music depends mainly on the
defined as for intermodulation distortion. size and acoustical nature of the room and, to a lesser extent, on the
taste of the listener. In the home, it is generally considered that a peak
The test frequencies given above are British Standard test output power of 7 to 10W will be adequate (assuming a loudspeaker
frequencies. American Standard test frequencies are 70c/s and 7kc/s efficiency of 5 %). If simplicity and economy are the governing
for intermodulation distortion and 9 and 10kc/s for beat-note considerations, a reserve of about 3W can, if the amplifier is carefully
distortion. Distortion figures would be lower when referred to the designed, give a generally acceptable standard of performance. In
American Standard test signals than when referred to the British large rooms or small assembly halls, however, the conditions will
Standard signals, but the actual distortion is obviously the same in the probably merit a maximum output power of at least 20W. The type of
same amplifier. Examples of this are shown in Fig. 12 on page 51 for output stage used will depend on the maximum audio power
the 10W amplifier. demanded of the amplifier. Consequently, the design of amplifying
equipment will normally proceed from output to input, the
Hum and Noise requirements of the output stage dictating to a large extent the design
of preceding stages.
Contributions to the output from various stray signals picked up at
points in the amplifier are normally assessed together, and are often OUTPUT STAGES
measured as the voltage developed at the output when the input is
short-circuited to earth. This voltage is expressed in decibels as a For output-power reserves of up to about 5W, single valve output
fraction of the rated output voltage measures across the load stages in which the valve operates under class A* conditions, are
resistance, so that: considered satisfactory. Triode valves used as power amplifiers yield
a relatively distortion-free output, but their power-handling capacity
Hum and Noise (in dB) is generally inadequate. The power-handling capacity of pentodes,
however, is sufficient to permit their use in such stages and although
the inherent distortion is high, negative feedback can be used to
= 20 log10 { voltage with input closed

rated output voltage

} reduce it to an acceptable level.

There exists a choice of two basic forms of output stage from which a
A level of hum and noise of, say, 60dB, means that the rated output
good-quality audio output of more than 5W can be delivered to the
voltage is 1000 times the voltage developed when the input is
voice-coil of a loud-speaker. The two well-known forms are:
short-circuited to earth.
1. A class AB† push-pull pentode (or tetrode) output stages;
Cross-talk Interference
2. A class A or AB push-pull triode output stage.
Cross-talk interference can arise in stereophonic equipment as a result
of the signal in one channel `breaking through' into the other. Such
The choice between these is largely a balance between economy and
interference is measured as the voltage developed at the output of one
performance.
channel when the input of that channel is opencircuited and the signal
applied to the other channel is sufficient to give the rated output from * Class A operation is that in which the values of bias and signal voltages applied to the
this other channel. This voltage is expressed in decibels as a fraction control grid of the valve ensure that anode current always flows.
of the rated output voltage obtainable from one channel. †Class AB operation is that in which the values of bias and signal voltages applied to the
grids of the valves cause anode current to flow in each valve for appreciably more than
half, but less than a whole, cycle of the signal voltage.
Negative Feedback

Negative voltage feedback is used in high-quality amplifiers to

improve the performance. Part of the signal is taken back and injected
in an earlier stage of the amplifier in opposite phase (180° out of
Pentode Push-pull Stages proved to be a very satisfactory arrangement. A secondary feature of
the use of these operating conditions is that the 12W output valves
The use of pentodes of the 12W anode-dissipation type operating in a each run at a mean dissipation of only 7-5W. The corresponding
conventional class AB push-pull stage enables an effective output of fixed bias conditions in this case are represented by Curve 3 in Fig. 1.
12 to 13W to be obtained easily assuming an output transformer
efficiency of about 80% (which is a fairly conservative estimate for The low-loading form of operation is, however, suitable only for use
the efficiency). The appropriate supply voltage, limited by valve in the reproduction of speech or music and cannot be used with
ratings, is about 300 to 320V. The power efficiency (that is, the ratio sine-wave input without excessive distortion resulting. For this reason
between the audio output power and the d.c. input power) of such a it is difficult to measure directly the distortion levels which hold
stage is fairly high, being 50%, for example, in a typical stage using under practical conditions.
two Mullard output pentodes, type EL84. Harmonic distortion,
however, is of the order of 3 to 4 % at full output and, thus, a large A second method of improving the performance of pentode push-pull
amount of negative feedback is necessary to reduce the distortion to output stages is the use of `distributed-load' conditions. This is
an acceptably low level (say 0.5 %) at maximum output. discussed in detail later in the chapter. Depending on the precise
loading used, the variation in anode and screen-grid currents can be
The conditions for class AB operation normally recommended and reduced to such a level that almost identical performance is obtained
published by valve manufacturers are based on measurements made with cathode and fixed bias.
with continuous sine-wave drive. The bias under zero-drive
conditions and the anode-to-anode load resistance are so chosen that
the optimum performance is achieved when the working points of the
valves are displaced under driven conditions. This displacement is
caused by the effect of increased anode and screen-grid currents on
the cathode-bias circuit. For a typical output stage working from a
310V h.t. supply and using EL84 valves, the rise in cathode current -
and thus in cathode bias voltage - with a sinusoidal signal voltage is
about 40 % at full drive.

When such a stage is used in the reproduction of speech or music,

however, the operating conditions are different. The mean amplitude
of the input signal is now very small compared with the peak values
which occur from time to time, and thus the mean variation in
cathode current is also very small. Because of the relatively long time
constant of the bias network, even under peak signal conditions, the
displacement of the working point is small enough for the stage to be
considered as working with fixed bias. When a nominal,
cathode-biased, class AB stage is examined under the corresponding
fixed-bias conditions with a sine-wave input, it is found that, at high Triode Push-pull Stages
output levels, distortion is greater than when cathode bias is used.
These two conditions are shown for two EL84 by Curves 1 and 2 in The level of distortion is inherently low in a triode push-pull output
Fig. 1. The quiescent bias is the same in both cases, Curve 2 showing stage in which the valves operate more or less under class A
normal operation with cathode biasing and Curve 1 showing conditions. It is found that if 25W pentodes (or tetrodes) are
operation with fixed bias. The results indicate that, in practice, a connected as triodes, an output power of 12 to 15W with a level of
cathode-biased class AB stage designed on the basis of a sinusoidal harmonic distortion below 1 % can be obtained using a supply
drive will produce increased distortion when peak passages of speech voltage of 430 to 450V.
or music are being reproduced.
Maximum output power and the corresponding distortion vary
One practical method of improving performance is to adjust the appreciably with the value of load impedance. The curves of Fig. 2
quiescent operating conditions in the output stage so that they are illustrate the performance of two Mullard high-slope output pentodes,
nearly optimum for fixed-bias working, although cathode bias is still type EL34, connected as triodes in a push-pull output stage, operating
being used. This entails a smaller standing current Ia(o) and a lower slightly below their anode-dissipation rating of 25W. For
anode-to-anode load impedance Ra-a. These changes result in larger anode-to-anode load impedances below 7kΩ, either a shared,
variations in the instantaneous anode and screen-grid currents when bypassed cathode resistor or separate bypassed cathode resistors can
the stage is driven, but the effect of these is lessened because the time be used. Above 7kΩ however, improved operation is obtained if a
constant of the cathode-bias network has also been increased. The shared resistor with no bypass capacitor is used. Operating conditions
excursion of the working bias is still kept very small under driven approach class A as the anode-to-anode impedance is raised, and the
conditions. optimum performance for high-quality output stages is obtained with
a load impedance of about l0kΩ. An output of 14W is then delivered
It is found that good short-term regulation of the h.t. voltage is by the valves with total harmonic distortion well below 1 %.
ensured by the use of large (50µF) electrolytic capacitors for the
anode and screen-grid supplies. Peak currents corresponding to near- This type of output stage found favour for a number of years in
overload conditions are effectively provided by the capacitors with a high-quality amplifiers giving about 12W effective output. Because
reduction in the line voltage of well under 0.5%, and the of the low inherent distortion of these triode-connected stages, less
instantaneous power-handling capacity of the stage is not impaired negative feedback is required to give acceptable linearity than has to
be used in pentode or tetrode stages giving similar output powers.
This modified design is described more fully in Chapter 6 as an Furthermore, in a three-or four-stage design in which feedback is
alternative `low-loading' version of the 10W amplifier, and it has applied over the whole amplifier, including the output transformer, it
is thus possible to obtain greater margins of stability for a given level under cathode- and fixed-bias conditions since, with the closer
of distortion. approach to class A triode operation, variations in anode and

Distributed Loading

The conditions of distributed loading are achieved by applying

negative feedback in the output stage itself. In the simplest form of
distributed-load operation, the screen grids are fed from suitably
positioned taps on the primary winding of the output transformer, and
the stage can be considered as one in which negative feedback is
applied in a non-linear manner via the screen grids.

The characteristics of the output stage under these conditions of

operation lie between those for pentode and triode operation. They
approach the triode characteristics as the percentage of the primary
winding common to the anode and screen-grid circuits increases.
Under optimum conditions, about two-thirds of the power-handling
capacity of the corresponding pentode stage can be realised with a
much lower level of distortion while, at power levels corresponding
to triode operation, the distortion also is of the order corresponding to
triode operation. At the same time, the output impedance is reduced
to a level approaching that obtained if a conventional triode push-pull
stage is used.

An output stage with distributed loading can thus be used with

pentodes of the 25W type in high-quality amplifiers designed for
output powers well in excess of 15W, the power efficiency being
appreciably greater than with triode operation. Alternatively, the
performance of 12W pentodes can be improved considerably and,
although the power handling capacity is reduced somewhat, effective
output powers of 10 to 12W can still be obtained.

A comparison is given in Table 1 of triode, pentode and

distributed-load operation for the Mullard output pentodes, types
EL34 and EL84. For the EL34, the comparison between
distributed-load operation and triode operation is of most interest. It
will be seen from this that distributed-load operation using a
tapped-primary output transformer enables the power-handling
capacity to be more than double that possible with triode operation
whilst, at the same time, distortion in the stage can be kept very low.

Although with a common-winding ratio of 0.2 (that is, with 20 % of

the primary winding common to the anode and screen-grid circuits)
the distortion level is comparable with triode conditions, it has been
found that appreciable improvement is obtained at higher output
powers if the ratio is further increased. Progressive improvement has
been obtained as the percentage of common primary winding is
increased up to 40 or 45 %. The power-handling capacity of the stage
is reduced further as the ratio is increased, but at least 35W can be
obtained with a level of distortion at the onset of grid current of about
2.5 %.

Typical performance curves of the EL34 when used with an output

transformer having the primary winding tapped at 43 % of the turns
are shown in Fig. 3. The output powers quoted are those delivered to screen-grid currents are reduced when the stage is driven. Second, as
the load in the secondary circuit. with normal triode operation, output power and distortion are less
dependent on the value of load impedances: with a common-winding
With valves of the 12W dissipation class, comparison with normal ratio of 0.4, little change in performance is produced by altering the
pentode operation is more significant. Appreciable reduction in anode-to-anode load from 6 to 9kΩ. However, Table 1 shows that
odd-harmonic distortion is again obtained under distributed-load there is little benefit to be achieved in respect of distortion by
conditions, and an output of approximately 15W is delivered by the increasing the common-winding ratio beyond 0.2 and, because of the
valves if the common winding ratio is 0.2. greater power-handling capacities possible, the circuits described in
this book which use distributed loading are designed for output
There are two advantages not apparent in Table 1 in using a transformers having 20 % of the primary winding common to the
common-winding ratio of about 0.4 when a high output power is anode and screen-grid circuits.
available. First, almost identical performance figures are obtained
INTERMEDIATE STAGES cathode input, and R2 ensures correct d.c. conditions in the second
section.
To ensure stability in an amplifier when negative feedback is used to
reduce distortion, the phase shift must be small. Thus, the number of The input grid of the circuit is directly coupled to the anode of the
stages should be a minimum. For low-output equipment such as the preceding valve so that the grid is at a direct potential of between 70
3W amplifier described in Chapter 7, it is usually possible to obtain and 80V. The bias is obtained by the voltage developed across the
the desired output with only a voltage-amplifying input stage and a shared cathode resistor R5, and the direct potential of the cathode is
power-amplifying output stage. However, for equipment using a therefore high. The cathode currents of both sections flow through
push-pull output stage, the penultimate stage must be capable of R5, the d.c. components being additive and the a.c. components in
providing a well-balanced push-pull drive voltage of adequate opposition.
amplitude. Consequently, extra stages will be required to give the
push-pull drive voltage and to provide sufficient amplification of this
voltage.

With push-pull stages using 25W pentodes such as the EL34, the
maximum drive voltage required is approximately 2 X 25Vr.m.s. For
stages using 12W pentodes (the EL84, for example) the requirement
is about 2 x10Vr.m.s.( These input-voltage requirements are similar for
triode, pentode or distributed-load operation.) If a high-impedance
double triode such as the Mullard ECC83 is connected as a cathode-
coupled phase splitter, it is possible to produce the above drive
voltages and also to perform the necessary phase splitting in a single
intermediate stage. .

A cathode-coupled phase splitter using an ECC83 is shown in Fig. 4.

A negative pulse at the grid of the first triode section of the ECC83
produces a positive pulse at the first anode, so that Vout is
positive-going. The cathodes are strapped together, and therefore go
negative at the same time. The negative pulse on the second cathode
is equivalent to a positive pulse on the second grid so that V’out is
negative-going. Hence V’out is 180° out of phase with Vout. The grid For there to be some a.c. signal remaining on the cathode to inject
of the second section of the valve is capacitively earthed by C3 for into the second section, the two sections cannot be operated with
exactly equal anode loads nor with equal grid resistors in the
following stages. The anode load R4 of the earthed triode section is negative-going, so that Vout and V’out are 180° out of phase. The
should be slightly higher than R3 to give perfect balance. Because of values of R1 and R2 are chosen so that the gain of the
the high amplification factor of the ECC83 (µ is 100) exactly equal phase-reversing stage is unity (that is, V'out/Vin = 1). The ratio R1/R2
resistors cannot cause more than 3 % lack of balance and nominally is equal to the ratio (A - 1)/(A + 1) where A is the gain of the phase-
equal resistors, matched to within 5 %, result in a lack of balance reversing valve measured from grid to anode.
which cannot exceed 2 %.
The circuit is less widely used than the phase splitter described above
The grid resistors R6 and R7 for the output valves must be of close because a lack of balance between the two outputs can result from the
tolerance (±5 %) as they form part of the load to the phase splitter. coupling capacitors. However the circuit has two advantages:
The frequency at which low-frequency unbalance occurs depends on
the time constant R2C3 and this can be made long enough to maintain 1. It does not require a voltage difference between heater and
adequate balance down to low audio frequencies.* High-frequency cathode, so that it is suitable for directly-heated valves or
balance is largely determined by the layout of the wiring which can for double-valves in which the second sections are needed
make shunt capacitances unequal. The inter-electrode capacitances of for other purposes.
the ECC83 are sufficiently small and equal between the two sections 2. If a voltage does exist between heater and cathode as in the
for their shunting effect on the circuit to give negligible unbalance at heater chain of a d.c./a.c. amplifier for instance - the large
high frequencies. amount of negative feedback from anode to grid can be
used to counteract excessive hum pick-up.

* R2 will be of the order of 1 MΩ so that unbalance is caused if any leakage current flows INPUT STAGES
through C3. This is a common cause of distortion.

A low level of hum and noise is essential in the input stage of an

The effective voltage gain (Vout/Vin or V’out/Vin) is about half that of amplifier because high gain is required in this stage to provide a good
one section used as a normal voltage amplifier. Nevertheless, there is signal-to-noise ratio and to compensate for the reduction in gain
sufficient gain in this stage because of the high amplification factor of resulting from the use of negative feedback. The recommended valve
the ECC83. The absence of the input coupling capacitor reduces for input stages is the Mullard low-noise pentode, type EF86 (or, in
low-frequency phase shift; and makes for good stability when d.c. equipment with series heater chains, type UF86). A conventional
feedback is applied. EF86 voltage-amplifying stage is drawn in Fig. 6.

Because of the direct coupling to the preceding stage, the anode

voltage of the preceding valve determines the operating conditions of
the ECC83. If this anode voltage is too high, the negative bias on the
phase splitter is too low, and overdrive on peak signals will result in
grid-current distortion. If the anode voltage is too low, the bias of the
phase splitter will be too high and the valve will operate away from
the linear part of its dynamic characteristic, and distortion will be
greater.

An alternative type of phase splitter - or more correctly, phase

reverser - is illustrated in Fig. 5. This circuit has various names such
as anode follower, see-saw and paraphase.

The anode load is shown shunted by a CR network which produces

an advance in phase and thus increases the stability of the amplifier at
high frequencies. Negative feedback ( from the secondary winding of
the output transformer ) is introduced across a high stability resistor
R5 in the cathode circuit. The feedback resistor R8 should be of the
high-stability type. The feedback network C8,R8 attenuates
One input for the push-pull output stage is the output Vout from the frequencies in the ultrasonic range and again gives increased stability.
stage preceding the phase reverser, and the other is the output V’out The stage is intended for direct coupling to the input grid of a
from the phase reverser. The input for the phase reverser is taken cathode-coupled phase splitter so that the phase shift at low
from the potential divider R1,R2 connected between the two outputs. frequencies is minimised and the low-frequency stability of the
When Vout is positive-going, the grid is positive-going, and the anode amplifier with feedback is improved.
 Microphony
The following sections describe the precautions needed to ensure that
the good properties of the EF86 are realised in the typical circuit of The EF86 has a rigid electrode structure. When correctly sited, the
Fig. 6. action of sound waves (acoustic feedback) and vibration on the
electrode structure will not produce any audible howl. When the
Noise valve is sited in an obviously doubtful or unsuitable position as, for
example, in tape recorders, a flexible mounting for the valveholder or
High stability, cracked-carbon resistors should be used for the anode a separate weighted sub-chassis is advisable. There are no
load R6, the screen-grid resistor R7, and the cathode-circuit resistors appreciable resonances in the EF86 below lkc/s. At higher
R4 and R5. If normal carbon resistors are used, the low noise level of frequencies, the effect of vibration is generally negligible because of
the EF86 may be masked by additional non-thermal noise generated the damping provided by the chassis and valveholder.
by the flow of current in the resistors.
Starvation Operation
The low-frequency noise generated by the EF86 itself is most
conveniently specified as an equivalent voltage at the control grid. In a single-valve output stage, pentodes are essential if adequate
For an h.t. line voltage of 250V and an anode load of 100kΩ, the output power is to be obtained. Inherent distortion with pentodes,
equivalent noise voltage is 2µV for a bandwidth of 25c/s to 10kc/s. however, is high, and a large amount of negative feedback is needed
The bandwidth has to be specified because the random noise of the to reduce the distortion to an acceptable level. Consequently, a very
valve covers all possible frequencies, and also because the noise high gain is essential in the input stage to compensate for this,
voltage is more or less proportional to the bandwidth. For particularly if only one stage of voltage amplification is used.
comparison, a resistance of 10MΩ in the grid circuit produces a
thermal noise voltage of about 4µV for the same frequency range. It is well known that the gain of a voltage amplifier can never reach
the theoretical maximum represented by the amplification factor µ of
Hum the valve. The gain, in fact, is given by the expression µRa/(Ra + ra),
where ra and Ra are the internal slope resistance of the valve and
The cathode resistor R4 is bypassed at signal and hum frequencies by external load resistance respectively.
the capacitor C2. Heater emission or the choice of an unsuitable
valveholder leading to capacitive coupling or leakage between the Practical considerations in conventional circuits restrict the extent to
pins, may give rise to a heater-to-cathode current. The following list which the anode load can be increased. For example, the maximum
indicates points which should be considered when constructing an will depend on the effect of shunt capacitances and the need to
amplifier: maintain the frequency response at the upper end of the audio
spectrum. The load will normally be less than 500kΩ and this value,
1. A low-loss valveholder must be used. A nylon-loaded used with an EF86 operating with a line voltage between 250 and
phenolic holder, preferably skirted, is generally adequate. 350V, will give a stage gain of about 250. In these conditions, the
anode and screen-grid voltages will be 60 and 70V respectively.
2. The external leads to the heater should be run as a twisted
pair to neutralise their external magnetic fields. (The heater If a more severe limitation of frequency response is acceptable, much
of the EF86 is itself made in the form of a spiral.) The higher values of anode load may be used. These higher values
heater wiring should be kept as close to the chassis as produce `starvation' operating conditions: valve currents and voltages
possible. are much smaller than those of conventional stages, but the stage gain
is much greater. The frequency response is restricted under starvation
3. Earth returns should be made to a busbar. Connections can conditions, bunt can be extended considerably by the use of
be made to the central spigot of each valveholder in the first corrective feedback.
instance. The busbar should only be connected to the
chassis at one point, preferably near to the input terminal. Because of the high value of anode load for starvation conditions, the
Hum picked up from currents circulating in the chassis is input impedance of the following stage will need to be high.
then kept to a minimum. Wherever possible, an external Preferably, it should not be less than 10MΩ. Direct coupling to the
earth connection should be made. grid of the following stage is possible, and the screen grid of the
voltage amplifier can be fed from the cathode circuit of the following
4. Where a recommended layout for the components is given, stage. Because of the low values of current and voltage and the
there will be no danger of hum being generated by the favourable effect of these on negative grid current, starvation
magnetic field of the mains transformer acting directly on conditions are useful in high input-impedance voltage amplifiers.
the electrode supports. If some other layout is tried, the
possibility of this form of hum occurring should be PRE-AMPLIFIER STAGES
considered very carefully.
Pre-amplification may be required with power amplifiers to provide
When used as a normal voltage amplifier with an h.t. line voltage of extra voltage gain when certain kinds of input source are used. Often,
250 V, an anode load of 100kΩ and a grid resistor of 470kΩ, the it is convenient to include the treble and bass tone controls and the
maximum hum voltage of the EF86 itself when one side of the heater volume control in the pre-amplifier. Also, the preamplifier can be
is earthed, is about 5µV at the grid, and the average value is about designed to give compensation for the bass attenuation and treble
3µV. The maximum level with a centre-tapped heater is only 1.5µV. boost which have been applied by the manufacturers during recording
Additional external screening of the valve is not normally required. to obtain more favourable ratios of signal to motor rumble and signal
to surface noise. Pre-amplifiers for use with tape-recording
equipment will also provide equalisation for the inherently non-linear
nature of magnetic recording.
Pre-amplifiers for stereophonic equipment must necessarily be more degree of control available in either sense should be the same, so that
complex than monaural units because a stereophonic pre-amplifier is a symmetrical, ‘centre-zero’ arrangement is needed.
more or less two monaural circuits in one. Furthermore, coupling
must exist between corresponding controls in each section of the Another facility normally required in a stereophonic pre-amplifier
stereophonic circuit so that comparable adjustments can be made to which is not necessary in a monaural circuit is a switch to allow the
both channels. If the principle of design adopted is that both channels transference of the input signals from one channel to the other. Also,
must be identical, then rigid ganging of the controls can be used. this switch usually serves to combine both channels, so that the
However, if it is intended to cater for non-identical channels,* equipment can be used with monaural recordings.
concentrically-operated controls which will permit individual and
coupled adjustment of the channels, will be needed. All pre-amplifiers, because they provide the inputs to the power
amplifiers, must be designed to send as little hum and noise as
Even if nominally identical channels are used, the acoustical outputs possible into the main amplifier. Hum and noise in the early stages
from the two loudspeakers will not be exactly the same unless count as signal voltages and are not reduced by feedback. There is
precautions are taken. Differences can occur because of (i) the therefore no advantage in including a pre-amplifier in the feedback
difference in the outputs from the halves of stereophonic pick-up loop. In fact, the switching required for compensating and equalising
heads, (ii) the unequal sensitivities of the loudspeakers and (iii) the circuits and the frequency-selective networks in a high-gain
very small differences in gain of the two amplifying channels. If the preamplifier are likely to produce large phase shifts and,
volume control consists of a dual-concentric potentiometer, consequently, to increase the risk of instability if included in the
individual adjustment to each channel will rectify any lack of feedback loop.
balance. If a dual-ganged potentiometer is used, a special balance
control is required. It should be possible with this control to increase *There is strong evidence suggesting that the two channels of a stereophonic arrangement
need not be identical. Excellent results are obtainable, for example, if one channel uses a
the gain in one channel while simultaneously decreasing it in the 20W power amplifier and the other, a control-less 10W amplifier. Of course, the settings
other. Also, it should be possible to increase and decrease the gain of of the gain and tone controls in each channel of the pre-amplifier will need to be different
one channel with respect to the other, and it is desirable that the because of the different characteristics of the power amplifiers.
 CHAPTER 4

General Notes on Construction and Assembly

Details are given in subsequent chapters of the dimensions, layout
and wiring of each item of equipment. This information is extensive,
but no claim is made that it is exhaustive. The constructional sections
of each chapter should be regarded principally as giving some
guidance for assembling the equipment, but should not be considered
to be absolutely inviolate sets of rules and instructions. An intelligent
use of these sections, coupled with references to circuit diagrams,
should produce excellent results.

The suggested arrangements of components in the various chassis are

those adopted in the prototype equipment. The convenience of the
constructor has been borne in mind in evolving these layouts, and
economy of space has been effected whenever possible. However, the Components such as the germanium diodes (used in the tape-recorder
underlying principle has always been to obtain the best possible equipment of Chapters 11 and 12) which are particularly sensitive to
standard of performance for a given expenditure. A good margin of temperature, will also require long lead wires. These wires should
stability, a low level of hum, and so on, have never been sacrificed in also be held with a pair of pliers to act as a heat shunt when
an attempt to achieve attractive, but less fundamental, properties such soldering. The polarity of Mullard germanium diodes is decided by
as miniaturisation. the outer markings: usually the negative pole is the end on which a
band is marked.
There is no theoretical method of deducing the best layout for audio
equipment. It has to be found experimentally, and it follows that the EARTH CONNECTIONS
effect of any deviations from an established arrangement cannot be
predicted. Consequently, any alterations to layouts recommended in In audio equipment, it is advisable to have only one direct connection
this book should not be undertaken lightly. They should be dictated to the chassis, and this should be made close to the input socket. With
by strong reasons, and the subsequent variations in performance a number of connections, eddy currents can be set up in the chassis
should be analysed closely. and hum voltages can be induced in the sensitive sections of the
equipment by the magnetic fields associated with these currents.
Similarly, recommendations in this book for good-quality Normally, a common earth lead or busbar is recommended, and the
valveholders and for high-stability resistors, for example, are based centre spigots of the valveholders often serve as convenient points for
on experimental proof of their necessity. Any change from the securing this lead.
recommendations can alter the performance considerably. The
voltage ratings quoted for capacitors, however, should be regarded as To maintain a low level of hum when two or more units are used
minimum values and capacitors with higher ratings may be used if so together (see Chapter 1), the chassis of the units should be connected
desired. In these respects, the home constructor is advised to choose by one lead only. If the negative return line is conveyed from one unit
his components very carefully, and complete sets of components or to the other by way of the power cable, the connection between
constructional kits purporting to be used with specific Mullard chassis should not be duplicated in the a.f. coaxial lead, and one end
circuits should be examined closely. Mullard Limited only design and only of the outer conductor of this lead should be earthed.
publish the circuits: they have no control of material and components
sold for use with their circuits. Whenever possible, an external connection should be made from the
equipment to earth. This lessens hum pick-up and may also help in
SMALL COMPONENTS eliminating any undesirable r.f. signals that can occur in equipment
situated near radio or television transmitting stations.
The smaller components are normally mounted on tagboards, and it is
generally more convenient to solder them to the tags before the OUTPUT TRANSFORMER
boards are mounted in the chassis. The small carbon resistors should
be laid across two tags and the lead wires bent around the tags in the The quality of the output transformer will govern to a great extent the
manner shown in Fig. 1(a). If the tagboard wiring diagrams show that quality of the output from the loudspeaker. A poor component can
neighbouring tags should be connected together, the lead wires of the give rise to a high level of distortion and can cause instability. The
appropriate components should be cut to lengths sufficient to allow most important factors determining the performance of an output
this. Waxed capacitors, silvered-mica capacitors and high-stability transformer are:
resistors will require longer leads so that they will not be overheated
when soldered into position. The leads should be bent as shown in 1. Primary-winding inductance
Fig. 1(b), care being taken in doing so to avoid making sharp bends 2. Leakage inductance
very close to the ends of the components. 3. Core material and core size
4. Intrinsic capacitance of the coils (especially of the primary)
5. Capacitances between the various coils
6. Power efficiency.

In an actual design, a compromise has to be found between the

inductance and capacitance of the primary winding. The inductance
must be high enough to minimise the fall in response which occurs at
low frequencies. On the other hand, a high inductance necessarily
produces a high self-capacitance because of the large number of turns
required, and this can result in a large loss in voltage at high
frequencies.

The inductance and self-capacitance of the primary winding form a

parallel tuned circuit in the output stage, and the resonant frequency
of this must be as high as possible. It is usually between 50 and
100kc/s in a good transformer. Furthermore, the capacitance must be
low to prevent an increase in the phase shift when feedback is used.
A high capacitance can give rise to low-frequency instability in the
amplifier. The capacitance can be minimised by careful winding of
the primary coil.

If the coupling between the primary and secondary windings of the

transformer is not close enough (corresponding to increased leakage
inductance) excessive phase shift will be introduced at high
frequencies which may lead to parasitic oscillations. Close coupling
is often obtained by dividing the primary winding into sections and
winding the secondary turns between these sections.

It is advisable to obtain output transformers from manufacturers who

build the components for specific circuits. The commercial type
numbers listed with the circuits in subsequent chapters have been
supplied by the manufacturers, and the transformers should all prove The lines AGC and EGF are the characteristics of normally and
to be satisfactory in the appropriate equipment. inversely connected linear potentiometers respectively. If these are
used for the balance control, then the need for the centre-zero
NEGATIVE FEEDBACK arrangement will give an operating point at the intersection G of the
two characteristics. The resistance corresponding to IG will appear in
Negative feedback is usually taken from the output transformer to the series with each grid, and that represented by GH will be between the
input stage of a power amplifier to reduce the harmonic distortion in grid and earth. The signal attenuation caused in each channel by this
the output signal. The correct polarity of this feedback is essential. arrangement will thus be 50 %.
Connection to the wrong side of the secondary winding of the output
transformer will give positive feedback and will result in oscillation. The curves EDF and ADC are respectively the characteristics of a
logarithmic potentiometer connected in reverse and an
The correct connection can best be found by trial and error, and it is antilogarithmic potentiometer connected normally. The operating
advisable to use a low-grade speaker while doing this. With point with this combination would be the intersection D, and the
unsuitable phasing (positive feedback) the resultant violent oscillation signal attenuation in the zero position of the control would be only
might damage a delicate speaker-coil suspension. If a low-grade ID/IH. For potentiometers obeying a 10 % logarithmic law, the
speaker is not available, a momentary completion of the feedback attenuation would thus only be 10 %.
loop should be sufficient to determine the correct phasing.
POWER SUPPLIES
Should it be desired to experiment with the level of feedback -if r.f.
instability occurs, for example - this can be done by altering the value Allowance has been made in the circuits of subsequent chapters for
of the resistor in the feedback loop. If the resistance is increased, adequate h.t. power supplies. The requirements of each item of
there is a decrease in the level of feedback. (A corresponding value of equipment dictate the type of rectifier to be used and the specification
shunt capacitance in the feedback loop must be found empirically.) for the mains transformer.
Except in the unlikely event of r.f. instability, little advantage will be
gained from changing the values of components recommended in A minimum value Rlimmin. of series resistance is quoted on page 29
later chapters. Audible improvement is indiscernible when the level for the rectifiers used with the circuits in this book. If the
of feedback is increased above that indicated for each amplifier. specification of the transformer does not give a total winding
resistance equal to or greater than this minimum, then a resistor must
STEREOPHONIC BALANCE CONTROLS be included in each anode circuit. The amount of series resistance Rt
contributed at each anode by the transformer is:
The balance control in stereophonic equipment will often consist of
dual-ganged potentiometers connected in the grid circuits of Rt = Rs+n2Rp
corresponding valves in each channel. One of the potentiometers will
be connected normally with its minimum-resistance end earthed and where Rs = the resistance of half the secondary winding
the other will be connected in reverse, with its maximum-resistance Rp = the resistance of the total primary winding
end earthed. The ganged potentiometers could be either a pair of and n = the ratio of the number of turns on half the secondary
linear-law potentiometers or one logarithmic-law and one winding to the number of turns on the whole primary.
antilogarithmic-law potentiometer. The characteristics of these
components, connected normally and in reverse, are shown in Fig. 2.
The value R of the resistance which must be added to each anode
circuit to produce the necessary limiting value Rlimmin. is therefore
given by:

R = Rlimmin.-Rt.

A good mains transformer is essential for high-quality equipment. It

may be necessary to screen the transformer to reduce the transfer of
hum, and a mains filter may also be required if interference arising
from the mains supply is experienced.

The policy of using only one connection to the chassis should also be
applied to reservoir and smoothing capacitors. It is not advisable to
bolt the negative cans of these capacitors directly to the chassis. The
cans should be insulated from the chassis and the negative connection
made to the common earth lead.

Low-frequency instability can result from inadequate decoupling of

the power supply. If this occurs, an increase in the values of the
decoupling components or the addition of extra decoupling
components may prove beneficial. The extent of this increase must be
determined experimentally.
 CHAPTER 5

Twenty-watt Amplifier

The circuit to be described in this chapter is designed to give the

highest standard of sound reproduction when used in association with The main feedback loop includes the whole circuit, the feedback
a suitable pre-amplifier, a high-grade pick-up head and a voltage being derived from the secondary winding of the output
good-quality loudspeaker system. transformer and being injected in the cathode circuit of the EF86. The
amount of feedback applied around the circuit is 30dB, but in spite of
Two Mullard output pentodes, type EL34, rated at 25W anode this high level, the stability of the circuit is good and the sensitivity is
dissipation, form the output stage of the circuit. These are connected 220mV for the rated output power. The level of hum and noise is
in a push-pull arrangement with distributed loading, and give a 89dB below the rated 20W.
reserve of output power of 20W with a level of harmonic distortion
less than 0.05 %. The intermediate stage consists of a The rectifier used in the power-supply stage is the Mullard full-wave
cathode-coupled, phase-splitting amplifier using the Mullard double rectifier, type GZ34. This provides sufficient current for the amplifier
triode, type ECC83. This stage is preceded by a high-gain voltage (about 145mA) and also for the pre-amplifier and f.m. radio tuner
amplifier incorporating the Mullard low-noise pentode, type EF86. unit (about 35mA) being used with it.
Direct coupling is used between the voltage amplifier and phase
splitter to minimise low-frequency phase shifts.
 Separate cathode-biasing resistors are used in the output stage to limit
CIRCUIT DESCRIPTION the out-of-balance direct current in the primary winding of the output
Input Stage transformer. The use of other balancing arrangements has not been
thought necessary although it is likely that some improvement in
The EF86 input stage of the circuit of Fig. 1 provides high-gain performance, particularly at low frequencies, would result from the
voltage amplification, the stage gain being approximately 120 times. use of d.c. balancing. It is necessary in this type of output stage for
High-stability, cracked-carbon resistors are used in the anode, the cathodes to be bypassed to earth even if a shared cathode resistor
screen-grid and cathode circuits, and they give an appreciable is used. Consequently, a low-frequency time constant in the cathode
improvement in the measured level of background noise compared circuit cannot be eliminated when automatic biasing is used.
with ordinary carbon resistors.
Negative Feedback
The stage is coupled directly to the input of the phase splitter. The
purpose of this is to minimise low-frequency phase shift in the Negative feedback is taken from the secondary winding of the output
amplifier and to improve the low-frequency stability when negative transformer to the cathode circuit of the input stage. In spite of the
feedback is applied. A CR network (C1, R3) connected across the high level of feedback used (30dB), the circuit is completely stable
anode load produces an advance in phase and thus improves the under open-circuit conditions. At least 10dB more feedback
high-frequency stability of the amplifier. (obtainable by reducing the value of R13) would be required to cause
high-frequency instability. The most probable form of instability
Intermediate Stage would be oscillation with capacitive loads, but this is most unlikely to
occur even with very long loudspeaker leads.
The second stage of the circuit uses a Mullard double triode, type
ECC83,and fulfils the combined function of phase splitter and driver Power Supply
amplifier. The phase splitter is a cathode-coupled circuit which
enables a high degree of balance to be obtained in the push-pull drive The power supply is conventional and uses a Mullard
signal applied to the output stage. indirectly-heated, full-wave rectifier, type GZ34, in conjunction with
a capacitive input filter. The values of the limiting resistors R26 and
With the high line voltage available, the required drive voltage for an R27 will depend on the winding resistances of the mains transformer
output power of 20W is obtained with a low level (0.4 %) of used. Their purpose, when required, is normally one of voltage
distortion. The anode load resistors R11 and R12 should be matched control only. Where a transformer with a very low winding resistance
to within 5 %, R12 having the higher value for optimum operation. is used, a secondary voltage rated at 400-0-400V may be found
Optimum balance is obtained when the effective anode loads differ adequate. The rating of the mains transformer is such that about
by 3 %. The grid resistors R14 and R15 of the output stage should 40mA may be drawn from the h.t. supply to feed a pre-amplifier
also be close-tolerance components because they also form part of the circuit and f.m. tuner in addition to the normal current required (about
anode load of the driver stage. High-frequency balance will be 145mA) for the amplifier.
determined largely by the wiring layout because equality of shunt
capacitance is required. Low-frequency balance is controlled by the Extra decoupling will be needed for these ancillary supplies. The
time constant C6R10 in the grid circuits of the triode sections,. and the smoothing component R22 can only be chosen when the type of tuner
value chosen in Fig. 1 will give adequate balance down to very low to be used is known. A value of about 6kΩ would be suitable for
frequencies. typical current and voltage requirements of approximately 35mA and
200V. The components R16 and C8 depend on the pre-amplifier to be
A disadvantage of the cathode-coupled form of phase splitter is that used. The values given in Fig. 1 refer to the 2- and 3-valve
the effective voltage gain is about half that attainable with one section pre-amplifiers and the mixer unit described in Chapters 9 and 10.
of the valve used as a normal voltage amplifier. However, as the
amplification factor of the ECC83 is high (100), the effective gain of CONSTRUCTION AND ASSEMBLY
the cathode-coupled circuit is still about 25 times.
The chassis for the 20W amplifier are made from two separate pieces
Output Stage of 16 s.w.g. brass or steel sheet. The dimensions (in inches) of these
pieces are:
The main feature of interest in the output stage is the use of two EL34
with partial screen-grid (or distributed) loading, the screen .grids (a) Main chassis 21 X 1½
being fed from tappings on the primary winding of the output (b) Base 1529/32 X1013/32
transformer. As stated in Chapter 3, the best practical operating
conditions are achieved with this type of output stage when about 20 Each piece should be marked as shown in the drawings of Fig. 2, and
% of the primary winding is common to the anode and screen-grid the holes should be cut as indicated. Where bending is required, it is
circuit. important for the bends to be made accurately at the lines for the
pieces to fit together properly on assembly.
The anode-to-anode loading of the output stage is 7kΩ and, with a
line voltage of 440V at the centre-tap of the primary winding of the Most of the resistors and small capacitors are mounted on two
output transformer, the combined anode and screen-grid dissipation ten-way tagboards, and they should be connected as indicated in Figs.
of the output valves is 28W per valve. With the particular screen-grid 3 and 4. The larger components should be fitted to the chassis in the
to anode load ratio used, it has been found that improved linearity is positions indicated in the layout diagram of Fig. 5: this arrangement
obtained at power levels above 15W when resistors of the order of of the components will ensure good stability.
1kΩ are inserted in the screen-grid supply circuits. The slight
reduction in peak power-handling capacity which results is not The wiring between the components is also indicated in Fig. 5. A
significant in practice. busbar earth return is indicated, with only a single chassis connection
at the input socket. Of the valveholders fitted to the chassis, only the
holder for the EF86 needs to be skirted. This holder should also be At the full rated output, the distortion without feedback is well below
nylon-loaded. 1% and with feedback is below 0.05 %. The distortion rises to 0.1 %
for an output power of 27W. The loop-gain characteristic is such that
D.C. CONDITIONS a level of at least 20dB of feedback is maintained from 15c/s to 25
kc/s, and of at least 26dB down to 30c/s.
The d.c. voltages at points in the equipment should be tested with
reference to Table 1. The results shown in this table were obtained Measurements of intermodulation products were made in the
with an Avometer No. 8. prototype amplifier using a carrier frequency of 10kc/s and a
modulating frequency of 40c/s. The ratio of modulating amplitude to
PERFORMANCE carrier amplitude was 4 : 1. With the combined peak amplitudes of
Distortion the mixed output at a level equivalent to the peak sine-wave
amplitude at an r.m.s. power of 20W, the intermodulation products,
The total harmonic distortion of the prototype amplifier at 400c/s, expressed in r.m.s. terms, totalled 0.7 % of the carrier amplitude. At
measured without feedback and with a resistive load, is shown in Fig. an equivalent power of 27W, the products totalled 1% of the carrier
6. The distortion curve towards the overload point is also shown in amplitude. The beat-note distortion between equal-amplitude signals
Fig. 6 when feedback is applied. at frequencies of 14 and 15kc/s is 0.25 and 0.3 % at equivalent
powers of 20 and 27W respectively, and between frequencies of 9
and 10kc/s it is 0.2 and 0.25 % at the same equivalent powers. The
variations of intermodulation and beat-note distortion with equivalent
power are plotted in Fig. 7.

The output/input characteristics of the prototype amplifier are shown

in Fig. 6. Excellent linearity is indicated up to output voltages,
measured across a 15Ω load, of 20V, which corresponds to an output
power of 27W.

Frequency Response

The frequency- and power-response characteristics - that is, the

curves for output powers of 1 and 20W respectively - are shown in
Fig. 8. From this figure it will be seen that, at an output of 1W, the
characteristic is flat (±1dB) compared with the level at 1kc/s from
2kc/s to 100kc/s and the power-response characteristic is flat
(±0.5dB) from 30c/s to 20kc/s.

It is important that adequate power-handling capacity is available at

the low-frequency end of the audible range and this is determined
chiefly by the characteristics of the output transformer. The prototype
amplifier is capable of handling powers of at least 20W at frequencies
as low as 30c/s without excessive distortion, but for very low
frequencies it is desirable that the signal should be attenuated in the
associated preamplifier.

Sensitivity

The sensitivity of the amplifier measured at lkc/s is 6.5mV for an

output of 20W when no feedback is applied, and approximately
220mV with feedback, the loop gain being 30d B. The loop-gain
characteristic of the complete amplifier for the full frequency range is shown in Fig. 8. The sensitivity, with feedback, at the overload point
(27W) is approximately 260m V.
The level of background noise and hum in the prototype equipment is
89dB below 20W, measured with a source resistance of l0kΩ. This is
equivalent to a signal of about 5.5µV at the input terminals. It is
possible to increase the overall sensitivity of the amplifier by 6dB
while still maintaining a low background level, a high loop gain and a
good margin of stability, but various design requirements of
associated pre-amplifier circuits (the need for a high signal-to-noise
ratio, for example) render a higher sensitivity a doubtful advantage.

Phase Shift and Transient Response

Emphasis has been laid in the amplifier on a good margin of stability

and, consequently, the phase shift is held to a comparatively low
level. As shown in Fig. 11, the shift is only 20° at 20kc/s. Excellent
response to transient signals is obtained, the rise time of the amplifier
being of the order of 5µsec.

Output Impedance

The output stage has a low inherent output impedance, and this is
further lowered by the use of negative feedback. It is approximately
0352 with a 1552 load for an output of 20W at frequencies of 40c/s,
1kc/s and 20kc/s. This gives a damping factor of about 50.
 CHAPTER 6

Ten-watt Amplifier
An easily constructed three-stage amplifier with a rated power preceded by a high-gain voltage amplifier using the Mullard low-
reserve of 10W is described in this chapter. The circuit is drawn in noise pentode, type EF86. The intermediate and input stages are
Fig. 1. coupled directly to each other so that low-frequency phase shifts are
minimised.
Several versions of the amplifier are possible. The equipment can be
constructed to include volume and tone controls, when it is suitable The optional treble and bass tone controls are shown in Fig. 1. RV28
for direct use with high-level signal sources, or it can be made provides continuously variable treble control from +10 to -10dB at
without controls to be used with pre-amplifying equipment. In 10kc/s and RV30 provides bass control from +11 to - 5dB at 20c/s.
addition, three arrangements of the output stage are possible: normal
loading, low loading and distributed loading. The level of feedback taken from the output transformer to the
cathode circuit of the input stage is 26dB. The basic sensitivity of the
The push-pull output stage of the circuit uses two Mullard output circuit with this amount of feedback is 40mV for the rated output
pentodes, type EL84. These yield a rated output power of 10W with power, and the sensitivity when the tone controls are included is
a level of harmonic distortion, under normal loading conditions, of 600m V. The level of noise and hum is at least 75dB below 10W.
less than 0.3 %.
The rectifier used in the power-supply stage is the Mullard full-wave
The intermediate stage consists of a cathode-coupled phase-splitting rectifier type EZ81.
amplifier using the Mullard double triode, type ECC83, which is
CIRCUIT DESCRIPTION from overdrive on peak signals, will occur. If the anode voltage is too
low, the phase splitter bias will be too high, and the stage will operate
Input Stage away from the linear position of the valve characteristics with a
resulting increase in distortion.
The first stage of amplification, is provided by the EF86 in a circuit
giving a voltage gain of about 150 times. The negative-feedback Output Stage
voltage from the secondary winding of the output transformer is
introduced across the 100Ω resistor R3 in the cathode circuit. The output stage consists of two output pentodes, type EL84, in a
Because of the high gain in this stage, any excessive noise will lower cathode-biased, push-pull circuit The anode supply voltages are taken
the quality of the output from the speaker considerably and so from the reservoir capacitor C15, while the supply voltages for the
high-stability, cracked-carbon resistors are used for the anode and screen grids and for the rest of the amplifier are taken by way of R22
screen-grid circuits. The EF86 is coupled directly to the phase splitter and C14.
to reduce the phase shift at low frequencies. A CR network (C1, R1)
shunting the anode load produces an advance in phase which Compensation for the slightly different currents drawn by any two
increases the stability of the amplifier at high frequencies. EL84 is achieved by using the separate cathode resistors R18 and
R19. Matching of the output valves is therefore unnecessary.
Intermediate Stage
The stopper resistors R15 and R16 in the control-grid leads and R20
The double triode type ECC83 is operated as a cathode-coupled and R21 in the screen-grid circuits have been included as a normal
phase splitter to provide a push-pull drive voltage for the output measure to prevent parasitic oscillations. These resistors should not
stage. The use of the cathode-coupled arrangement gives low be bypassed, and a close connection should be made to the tags on
distortion and facilitates direct coupling with the input stage. The the valveholders.
voltage gain obtained with a cathode-coupled circuit is approximately
half that obtainable from each valve section operated as a normal A resistor R25 having a value of about lkΩ maybe placed across the
voltage amplifier. Nevertheless, it is sufficient because the output terminals to prevent instability from occurring when the
amplification factor of the ECC83 is 100. loudspeaker is disconnected.

It is necessary in a cathode-coupled phase splitter for the anode load Operating Conditions - Normal, Low and Distributed Loading in
of the earthed section to be slightly higher than that of the first Output Stage
section if perfect balance is to be obtained. Thus R10 should ideally
be slightly higher than R9, by a percentage which depends on the Three ways of operating the output stage are possible. These use the
amplification factor of the valve. Because of the high amplification conditions for normal loading, low loading and distributed loading.
factor of the ECC83, nominally equal resistors, matched to within 5
%, cannot give rise to more than 2 % lack of balance. Normal- or low-loading conditions are determined by the choice of
value for the equal cathode resistors R18 and R19 in Fig. 1. For
At low frequencies, the presence of C5 and R8 in the grid circuit of normal loading, the value of each resistor should be 270Ω, and for
the second triode section produces both phase and amplitude low loading it should be 437Ω (that is, 390Ω +47Ω). The conditions
unbalance. The frequency at which the lack of balance becomes for distributed loading are achieved by including part of the anode
significant depends on the time constant C5R8, which should be as load (that is, part of the primary winding of the output transformer) in
long as possible to provide maximum decoupling within the pass the screen-grid circuits of the valves.
band of the amplifier. Adequate balance can easily be maintained in
practise and the component values used in this circuit give it down to Normal loading is used when the amplifier is to be tested up to a full
low audible frequencies. output power with a sine-wave input. Low loading is used only with
speech and music signals when it gives reduced distortion.
The anode voltage of the EF86 determines the operating conditions of Distributed loading conditions are used to achieve a compromise
the phase splitter and, provided the values of the associated between the performance of the two EL84 when connected as triodes
components are within the stated tolerances, the stage will operate and as pentodes. A sufficient power reserve is maintained, yet
correctly. If the anode voltage of the EF86 is too high, the bias on the distortion is much less than with the usual pentode connection.
phase splitter will be too low and grid current distortion, resulting
Normal and Low Loading Larger peak currents are produced in the output stage under
low-loading conditions than with normal class AB operation. With
The normal loading conditions correspond very closely to the speech or music inputs, these peaks are of short duration. They are
recommendations for class AB operation of the EL84. The difference supplied by the reservoir capacitor C15 which is large enough to
is that, in Fig. 1, separate cathode resistors R18 and R19 (each equal provide satisfactory smoothing. When the amplifier is at the point of
to 270Ω) are used instead of a shared resistor of 130Ω. The overload on a peak signal, the momentary fall in line voltage from a
anode-to-anode load resistance for these conditions is 8kΩ and the nominal value of 320V should not be more than 2V.
anode current for zero input signal is 2 x 36mA.
As the current in the output stage increases, the bias voltages across
The low-loading adjustment is so called because, with the increased the cathode resistors also increase at a rate determined by the time
values of cathode resistance, the anode-to-anode load resistance is constants of the bias network. It is unlikely that this increase will
reduced to 6kΩ and the quiescent current to 2 x 24mA. Although the exceed 1V. The working conditions of the output stage are such that
output stage is connected for cathode-bias operation, operation with the output valves are driven back into a region where lower distortion
speech or music input signals approximates very closely to fixed-bias is obtained. As a result of this change in bias, a variation in gain will
conditions. A comparison has been made in Chapter 3 between the occur, but the distortion introduced by gain variation is held to a low
distortion in the output stage when speech or music signals are used level by the large amount of negative feedback. The total effect,
with normal and low loading conditions and this comparison is therefore, is to improve the performance at peaks of output power.
shown in Fig. 1 on page 19.*
It will be realised that, when the amplifier is being used for its
Curves 1 and 3 in that figure have been plotted for fixed bias primary purpose as a means of home entertainment, the low-loading
conditions. Curve 1 shows that distortion increases seriously above adjustment is to be preferred to normal loading. However, sine-wave
10W when normal loading is used. Curve 3 is for low loading and testing up to the full output power should not be undertaken with the
illustrates the lower level of distortion to be achieved with this low-loading arrangement. Such testing should only be undertaken
adjustment. Curve 2 has been plotted for cathode bias and therefore with normal loading. Measurements of frequency response may be
refers to a sine-wave input. The curve is for normal loading because made with low-level sinusoidal inputs provided the output power
this is required for sine-wave testing up to the maximum output. does not exceed 1 to 1.5W and square-wave testing may also be
undertaken if the signal is again restricted in the same way as the
With speech or music inputs, the maximum rated power of the sine-wave input. It is not felt that these restrictions will be of very
amplifier is only required in the output stage for a small part of the great significance because few home constructors will have the
time, the average requirement being comparatively low. However, a equipment needed for sine-wave and square-wave testing.
large power reserve must be available to cater for the widely differing
sound levels which can occur. Distributed Load .

The h.t. current consumption is smaller when the output stage is The arrangement of the output stage for operation under conditions of
adjusted for low loading. Consequently, the standing dissipation in distributed load is shown in Fig. 2. The screen-grid resistors R20 and
this stage is reduced from 11W at each anode with normal loading to R21 are disconnected from C14 in Fig. 1 and are connected to the
7.5W at each anode. The valves are thus being run well below the tappings provided on the primary winding of the output transformer.
maximum anode-dissipation rating of 12W. The h.t. current The centre tap of the primary winding is connected, as before, to the
requirement is reduced with low loading so that the mains reservoir capacitor C15. The dropper resistor R22 in the h.t. line no
transformer rating can be lower if the amplifier is to be used longer carries the screen-grid current and so is increased from 1.2kΩ
permanently with the low-loading adjustment. to 5.6kΩ to maintain the same d.c. conditions in the first two stages
of the amplifier.

*The distortion curves of Fig. 1, Chapter 3 are for the output stage only. The best results are obtained with each half of the primary winding
Distortion in the complete amplifier is discussed below. tapped so that about 20 % of the turns are common to the anode and
screen-grid circuits, when the anode-to-anode loading should be the voltage across C15 should then be near the specified 320V. The
6.6kΩ. voltage across C15 must not exceed 320V at either 80mA with
normal loading, (Ra-a = 8kΩ) or 60mA with low loading (Ra-a = 6 kΩ).

When an f.m. radio tuner unit is used with the amplifier, the
additional h.t. current drawn will not normally exceed 35 to 40mA
and the voltage across C15 will then drop to approximately 295V.
This additional current should betaken from C15 by way of suitable
dropping resistors R23 and R24 and should have adequate decoupling
(provided by C13 and C16). In these circumstances, the overload
point of the amplifier is reduced from 13 or 14W to approximately
12W.

The h.t. supply for pre-amplifiers should be taken from C14. The
values of the smoothing resistor R17 given in the table in Fig. 1 are
suitable for use with the pre-amplifiers of Chapters 9 and 10.

With distributed loading, total distortion is reduced very much

compared with normal loading. The rated power reserve of 10W is
maintained although overloading now occurs at 11W instead of 14W,
However, the rate at which distortion increases with output power
after overloading is less with distributed than with normal loading.
The frequency response of the amplifier is the same for either method
of loading, although the stability is improved with distributed loading
because the lower level of distortion is obtained with reduced loop
gain.

Negative Feedback

The level of negative voltage feedback taken from the secondary

winding of the output transformer to the cathode circuit of the input
valve is 26dB. The output impedance with feedback is 0.9Ω
measured at the 15Ω output terminals. This gives a satisfactory
damping factor of about 17 (that is, 15/0.9). CONSTRUCTION AND ASSEMBLY

Power Supply The chassis for the 10W amplifier in which tone and volume controls
are incorporated is made from three separate pieces of 16 s.w.g.
The power-supply stage shown in Fig. 1 uses the Mullard indirectly aluminium sheet. Two separate pieces of 16 s.w.g. aluminium sheet
heated, full-wave rectifier, type EZ81. Adequate smoothing is are required for the version without the controls. The dimensions (in
achieved using the resistor R22 and the large capacitor C15 (50µF) so inches) of these pieces are:
that the expense of smoothing chokes is avoided. The current
supplied by the EZ81 is sufficient for any arrangement of the output A - For version with controls
stage and also allows the use of a radio tuner unit. If the low-loading (a) Main chassis 20 x 13
version of the output stage is adopted and it is not intended to use the (b) Base 14⅞ x 7⅞
amplifier with an f.m. tuner unit, then the Mullard rectifier, type (c) Screen 8 x 3
EZ80, can be used in place of the EZ81, but it must be emphasised
that currents greater than 90mA must not be taken from the EZ80. B - For version without controls
(a) Main chassis 14 x 8⅝
Limiting resistors must be included in the anode circuits of the (b) Base 17½ x 4⅞
rectifier, and their values will depend on the type of mains
transformer used. The total resistance - made up of the added The pieces should be marked as shown in either Fig. 3 (with controls)
resistance and the effective resistance of the mains transformer - must or Fig. 4 (without controls) and the holes should be cut as indicated.
be at least 190Ω per anode for the EZ81 used in this particular circuit Where bending is required, it is important for the bends to be made
(215Ω for the EZ 80). For most 300-0-300V mains transformers, the accurately at the lines for the pieces to fit together properly on
effective resistance is about 130Ω so the value of R26 and R27 (in assembly.
the 20 % range of preferred values) must be at least 85Ω each (100Ω
for the EZ80). With the power supply giving an h.t. current of 80mA,
Two sets of assembly drawings are included in this chapter. One set
is for the version of the amplifier which includes the volume and tone
controls and the other set is for the control-less version of the
amplifier. The versions differ considerably in layout, but the two
arrangements will give good stability. The layout of the control-less
version has been dictated to some extent by the requirements of
stereophonic equipment. The narrow chassis with the sockets
arranged on one side will enable two amplifiers to be used
back-to-back with considerable economy of space. A symmetrical
arrangement of two amplifiers can be achieved if one of the units is
built as a mirror image of the other.

In the version of the amplifier having controls, the majority of

resistors and small capacitors are assembled on one ten-way tagboard
and one five-way board. These are drawn in Figs. 6 and 7
respectively. The position of .these boards and the other components
is shown in Fig. 5, and the wiring between the components is also
shown in this figure. A busbar earth return is used and only one
connection is made between this and the chassis. This connection is only one connection, at the input socket, between it and the chassis.
made, via the volume-control potentiometer, at the input socket. The Only the valveholder for the EF86 needs to be skirted and it should
only valveholder which needs to be skirted is that for the EF86, and it also be nylon-loaded.
should also be nylon-loaded.
D.C. CONDITIONS
Most of the components of the control-less version of the amplifier
are assembled on two ten-way tagboards and these are drawn in Figs. The d.c. voltages at points in the equipment should be tested with
8 and 9. The general layout of the components in the chassis is shown reference to Table 1. The results shown in this table were obtained
in Fig. 10. As before, a busbar earth return is used and again there is using an Avometer No. 8.

PERFORMANCE
Distortion Measurements of intermodulation distortion were made in the
prototype amplifier using carrier and modulating frequencies of (a)
The total harmonic distortion of the prototype amplifier with normal 10kc/s and 40c/s, and (b)7kc/s and 70c/s, the ratios of the amplitudes
loading and with feedback was measured at 400c/s. The variation being 1:4. The values are plotted in Fig. 12 against equivalent output
with output power is shown in Fig. 11. The relationship between power. For a stated equivalent output power, the peak value of the
input voltage and output power is also drawn in Fig. 11. At the full combined inputs is equal to the peak value of a single sine-wave
rated output power, the distortion is less than 0.3 % and at about 14W input which would produce that power. For frequencies of 10kc/s and
it rises to 1 % 40c/s, the intermodulation distortion (expressed in r.m.s. terms as the
percentage of intermodulation products relative to the amplitude of Hum and Noise
the higher frequency) is better than 1.0 % at 10W equivalent output
power. The level of background noise and hum in the prototype equipment is
at least 75dB below the rated output of 10W.
The magnitude of 'beat-note' distortion has been measured with
frequencies of (a) 9 and 10kc/s, and (b) 14 and 15kc/s, the amplitudes Phase Shift
of the input signals being equal. The difference component present in
the output, expressed as a percentage of either signal, is also plotted The phase shift in the amplifier is low. As shown in Fig. 13, the phase
against equivalent output power in Fig. 12. shift is 20° at 20c/s and at 10kc/s.

Frequency Response

The frequency-response characteristic for an output power of 1W and

the power-response characteristic for 10W output power under
normal loading conditions are shown in Fig. 13. It will be seen that
the frequency response is level (to within 1dB) from 5c/s to 20kc/s
and the power response is level from 20c/s to 15kc/s.

The loop gain characteristic is also plotted in Fig. 13. This curve
shows that 26d B of feedback are available at lkc/s.

Output Impedance

The output impedance of the amplifier is approximately 0.9Ω for a

15Ω load. This gives an adequate damping factor of about 17.

Performance with Normal and Distributed Loading

Performance figures are given in Table 2 comparing the amplifier

Sensitivity when a conventional pentode push-pull stage is used with the unit
when distributed loading is used.
The basic sensitivity of the amplifier (that is, the sensitivity at the
control grid of the EF86) is 2mV without feedback and 40mV with
feedback. The sensitivity at the input terminals in Fig. 1, thus
including the controls, is 600mV with feedback. The basic sensitivity
at the overload point (14W output) measured with feedback is 50mV.
 CHAPTER 7

Three-watt Amplifier
The circuit described in this chapter has been developed to meet the The comparatively high sensitivity of the amplifier (100mV for 3W)
demand for a simple amplifier of reasonably high quality. The permits the use of all types of crystal pick-up head and allows, if
amplifier, which is operated from a.c. mains, uses three Mullard required, the use of equaliser networks between the head and
valves: an EF86, an EL84 and an EZ80. The circuit given in Fig. 1 amplifier. The output terminations of the circuit are suitable for
includes three controls: volume (RV1), treble (RV2) and bass almost all kinds of loudspeaker, but, although the circuit is designed
(RV11). A modified version of this circuit, which allows the to make the most effective use of the single output valve, the best
amplifier to be used with a pre-amplifier or in stereophonic possible results will only be achieved if a suitably housed
equipment, is shown in Fig. 2. In this version, the three controls have high-quality speaker is used.
been omitted, and the feedback network has been simplified.

CIRCUIT DESCRIPTION
Because of the inherently high level of distortion with single-ended
output stages, appreciable negative feedback around the output stage The use of direct coupling between the stages necessitates a higher
is necessary to produce an output of acceptable quality. At the same cathode voltage in the output stage than is required with RC coupling.
time, an overall sensitivity of 100mV is required if the amplifier is to The value of R13 is thus greater than is usual for the cathode
be suitable for use with any type of crystal pick-up head. (The resistance. The screen-grid voltage for the EF86 is taken from the
attenuation resulting from pick-up equalisation with passive RC cathode of the EL84. In this way, negative d.c. feedback (which is
networks must also be borne in mind.) essential in a directly coupled circuit to stabilise the operating of both
stages) is applied to the voltage amplifier.
The basic sensitivity of the circuit without feedback should be about
10mV in order that the desirable level of feedback (about 20dB) can Negative a.c. feedback is applied from the secondary winding of the
be provided. From considerations of stability, this feedback should be output transformer to the cathode of the EF86. In Fig. 1, this feedback
taken around the minimum number of stages. loop incorporates the bass-boost control, the amount of feedback
being changed continuously at low frequencies as the resistance of
The EF86 in the voltage-amplifying stage is used under conditions the control potentiometer RV11 is varied. In the simplified version of
approaching those of starvation operation. With a high value of anode the circuit, the control RV11 is omitted, and the feedback loop
load resistance (R5 is 1MΩ) and reduced values of anode and consists of R6 and C5 only.
screen-grid voltage, the gain of the stage is raised two or three times
above that obtained under normal operating conditions. This increase Provision for volume and treble control is made at the input of the
is attributable mainly to the fact that, because the voltage at the anode amplifier. The potentiometers RV1 and RV2 constitute these controls
of the EF86 is very low, direct coupling can be used between this respectively. In the control-less version of the amplifier, the output is
anode and the control grid of the EL84 in the output stage. Thus the taken directly to the capacitor C1 in the control-grid circuit of the
shunt loading on the anode circuit of the EF86 is least at low and EF86.
medium frequencies.
The power supply uses the EZ80 in combination with a mains screen-grid is supplied through another filter network R12, C6. The
transformer meeting the specification given below. The resistor R15 h.t. supply for a pre-amplifier is taken from C9. Extra smoothing is
between the cathode of the EZ80 and the reservoir capacitor C9 is for provided by R16 and C10.
voltage control. The anode of the EL84 is supplied from C9, and the

CONSTRUCTION AND ASSEMBLY

Chassis details of the amplifier are given in Fig. 3. For the socket, and the feedback path will be completed by connecting the
control-less version, holes for the potentiometers will not be required. appropriate junction of R6 and C5 to the speaker terminal.
The chassis can be cut from one piece of 16 s.w.g. aluminium sheet,
12 in. long and 10 in. wide. A bottom cover plate to the amplifier is If the can of the double electrolytic capacitor is used as the negative
not necessary. side, then it should be isolated from the chassis. The earth connection
to the chassis should be made at the input socket only.
A suitable arrangement of the components in the amplifier with
controls is shown in Fig. 5. The position of the components on the The mains transformer should have an h.t. rating of 300-0-300V, 60m
A, and it is preferable, though not essential, that a separate l.t.
tagboard is shown in Fig. 4. For the control-less version, the winding (6.3V) be used for the EZ80 rectifier. This is indicated in the
potentiometers RV1, RV2 and RV11 and the capacitors, C2 and C7 circuit diagram, and also in the list of components.
are omitted. The capacitor C1 will be connected directly to the input
PERFORMANCE

Frequency Response

With the treble and bass controls in Fig. 1 in their minimum effective
positions, or with the control-less circuit of Fig. 2, the frequency
response is essentially flat from 35c/s to 30kc/s (Fig. 6). With
maximum application of the respective controls, a treble cut of 20dB
is available at 10kc/s, and a bass boost of 15d B is available at 70c/s.
The bass boost is obtained by reducing the main feedback at low
frequencies by means of RV11 and C7 (Fig. 1).

Distortion

The relationship between the total harmonic distortion and the output
power is shown in Fig. 7. It will be seen that, for a typical amplifier,
for outputs above about 3.5W, the distortion increases rapidly. This
indicates the point beyond which over loading of the amplifier
occurs.
 D.C. CONDITIONS
Output Impedance
The d.c. voltages at points in the equipment should be tested with
The output impedance of the amplifier for a loudspeaker load of 15Ω reference to Table 1. The results shown in this table were obtained
is less than 1.5Ω. This gives an adequate damping factor of more than using an Avometer No. 8.
10 (that is, > 15/1.5).
 CHAPTER 8

Seven-watt, D.C./A.C. Amplifier

The circuit described in, this chapter is intended primarily for those enough (200mV) to allow its use with most types of crystal pick-up
who require high-quality audio reproduction but who are limited to, head or with f.m. radio tuner units.
or are likely to encounter, d.c. mains supplies. The amplifier can be
arranged to accommodate any speaker impedance without the need Four Mullard B9A (noval) valves - types UF86, UCL82 (two) and
for rewiring if an output transformer having a tapped secondary UY85 - are used in the amplifier. The heaters of these valves (rated at
winding is fitted. The absence of a mains transformer means that the 100mA) are connected in a series chain. Because of this and the high
unit is much smaller than other amplifiers of comparable output a.c. voltages appearing at some of the heaters when the unit is
power. operating on a.c. mains, the standard of performance in respect of
hum pick-up is necessarily a little below that to be attained with
The equipment operates in the range of mains voltage (a.c. or d.c.) of equipment such as the 10 or 20W amplifier. But the performance is
from 200 to 250V, and the maximum output power available is still very good and, as satisfactory results have been obtained with a
between 7 and 8W at low distortion. Negative feedback of 21dB is variety of makes of crystal pick-up head, the d.c./a.c. amplifier should
used to improve the performance of the complete amplifier, but even satisfy most of the demands for equipment of this nature.
with this high value of feedback, the sensitivity of the circuit is high
CIRCUIT DESCRIPTION the low audio frequencies, so that the phase shift is limited. In
addition, the busbar is connected to the chassis by way of the
The circuit diagram for the d.c./a.c. amplifier is given in Fig. 1. capacitor C21.
Phase-splitting Stage
The input stage of the amplifier incorporates the Mullard high-gain,
low-noise pentode, type UF86, used as a voltage amplifier. In the The triode sections of the two UCL82 form an anode-follower
second stage of the equipment, the triode sections of two phase-splitting stage. Such an arrangement has been chosen because
triode-pentodes, type UCL82, form a phase-splitter. This provides the of the large amount of negative feedback that occurs between the
drive for the pentode sections of the two UCL82 which operate as a anode and grid of V3A. This can be used to counteract the large hum
self-biased, push-pull output stage with a nominal anode-to-cathode voltages transferred, with an a.c. mains supply, to the grid of V3A as
voltage of 200V. a result of running the heater of V3 some 62V above the neutral
voltage level.
Negative voltage feedback is used to improve the performance of the
complete amplifier. It decreases distortion, improves the frequency The balance between the output voltages from the triodes of this stage
response and reduces the output resistance. is controlled by the values chosen for the resistors R14, R16 and R17.
1f resistors which deviate from the nominal values by the extreme of
The power supply of the amplifier uses the Mullard half-wave the 5 % tolerance range are used for R16 and R17, the resulting lack
rectifier, type UY85. The rectifier output is RC smoothed, and the of balance should not be more than 4 %.
anodes of the push-pull pentodes are fed from the reservoir capacitor
C20 through the centre-tapped primary winding of the output Output Stage
transformer.
The pentode sections of the UCL82 are used in a push-pull output
Tone-control Circuit stage. Normal pentode loading is used so that maximum output power
is obtained with any h.t. voltage. The valves operate under the
The passive network comprising the tone-control circuit in this following class AB conditions:
equipment is identical with that of the 10W amplifier described in Anode-to-cathode voltage 200V
Chapter 6. It will provide all the control necessary for normal usage. Screen-to-cathode voltage 200V
Cathode circuit resistance (each pentode) 390Ω
Pick-up of hum voltages in the tone-control network is minimised by Anode-to-anode load 6kΩ
enclosing all the control components in a metal screening box. This Quiescent anode current 2x35mA
box is isolated from the chassis by a sheet of paxolin and is held at
the negative h.t. potential. The cases and spindles of the Rectifier Stage
potentiometers are connected to this screening box. The chassis
should not be connected to an earth point. The resistor R34 is included in the cathode circuit of the UY85 to
limit surge currents through the valve. The fuse FS, rated at 150mA,
Details of the tone control are given in the circuit diagram in Fig. 1. is also included in the h.t. circuit.
The treble and bass controls are RV1 and RV3 respectively. RV5 is
the volume control. The arrangement of the components within the Heater Circuit
metal screening box is given in Fig. 5.
The heaters of the four valves used in the amplifier are connected in
Input Stage series. The UF86 in the input stage (which has the highest gain) is
connected to the neutral side of the mains supply.
The UF86 in the first stage of the amplifier is connected to the neutral
end of the series heater chain. In this way, there is a minimum The nominal drop in voltage across the four heaters in series is 150V.
transfer of hum voltage from the heater to the control grid of the The difference between this and a mains voltage in the range 200 to
valve. To obtain a low level of noise in the amplifier, high-stability 250V is taken up by the following components connected in the
components should be used for the resistors R6, R7 and R9 in the heater circuit: the Varite thermistor R30 (type VA1010) used for
anode, cathode and screen-grid circuits of the UF86. surge suppression; the resistors R31 (470Ω), R32 (120Ω) and R33
(120Ω) used to provide tapping points for the ranges of mains
Instability in the amplifier can result from the large amount of voltage; the pilot lamp LP used as an indicator. The components R30,
negative feedback introduced into the cathode circuit of the UF86. It R31 and R32 are mounted under a ventilation hole cut in the chassis.
can also occur because of the impedance that exists between the This allows dissipation of the heat developed in these resistors.
chassis and the busbar which serves as the negative return line of the
amplifier. The following precautions are taken to prevent any Negative Feedback
possibility of instability.
The amount of feedback taken from the secondary winding of the
The simplest method is to reduce the loop gain of the amplifier. To transformer to the cathode circuit of the UF86, is 21dB. The output
do this at high audio frequencies, the capacitor C6 is connected to resistance with this feedback is 0.9Ω, measured at the 15Ω output
bypass the anode load resistor R6. To reduce the loop gain at bass terminals. This gives an adequate damping factor of about 17.
frequencies, the parallel combination of R10 and C10 is included in
the coupling circuit between the first and second stages of the
equipment. This method of coupling results in resistive attenuation at
CONSTRUCTION AND ASSEMBLY C21, between the chassis and the busbar return line of the amplifier.
(This connection is required for reasons of stability.)
Chassis details for the d.c./a.c. amplifier are given in Fig. 2. The two
metal pieces of the chassis should be cut from 16 s.w.g. aluminium The amplifier itself should be contained in a well-ventilated
sheet and the insulator from paxolin sheet, 1/16in. thick. The insulating box. If this is not possible then at least the ventilation hole
dimensions (in inches) of these pieces are: and the electrolytic capacitors on the top of the chassis must be
protected to prevent people touching them. The potentiometers
Main chassis 12 x 11 should be fitted with knobs which prevent any contact with the bare
Screening box 63/8 x 47/8 spindles and which have no protruding grub screws.
Insulator 55/8 x 17/8
To prevent hum pick-up, there are no isolating capacitors in the input
The arrangements of the components on the tag-boards is given in circuit. For this reason great care must be taken when connecting
Figs. 3 and 4 and a suggested layout of the components in the chassis auxiliary equipment to the amplifier. It is essential that no exposed
is shown in Fig. 5. metal parts are connected to either side of the input circuit of the
amplifier. Earth connections in associated equipment should be
In all d.c./a.c. equipment, certain precautions must be taken to dispensed with, as there is a risk that one side of the mains may
prevent constructors and users receiving electrical shocks. This become connected to them. They are also likely to cause hum trouble.
possibility of shocks arises because the amplifier can inadvertently be
connected wrongly to the mains supply. If the polarity of the mains It should also be borne in mind that the speaker leads are connected
connection is incorrect, several components will become `live'. to the busbar, and because of an incorrect mains connection this may
become `live'. Thus the metal parts of the loudspeaker should be
In this amplifier, the chassis is insulated electrically from the internal inaccessible.
components and wiring except for a single connection, by way of
PERFORMANCE
The maximum output for a mains voltage of 250V is 8W. When a
The performance data given in this chapter have been obtained from d.c. mains supply is used, the output power will be lower because less
measurements made on a prototype amplifier using a typical output h.t. voltage will be available.
transformer. Except for the relationship between output power and
mains voltage, all the data have been obtained with an a.c. mains Distortion
voltage of 240V. The measurements have been made with a
sine-wave input signal. The total harmonic distortion, measured at 400c/s, is about 0.5 % for
an output of 7W. For an output of 3W, the distortion is 0.2 %.
Output Power
Intermodulation distortion was measured at 40c/s and l0kc/s, the
The variation of obtainable output power with mains voltage is amplitude of the low-frequency signal being four times that of the
shown in Fig. 6. It can be seen that the rated output of 7W at an input high-frequency signal. At an equivalent output power of 7W, the
frequency of lkc/s is obtainable with an a.c. mains voltage of 240V. intermodulation distortion is 2 %. At an equivalent power of 3W, the
distortion is 1 %. The stipulated equivalent output power is the power Frequency Response and Stability
that would be obtained from a single sine-wave input of amplitude
equal to the peak value of the combined low- and high-frequency The frequency-response, loop-gain and phase-shift characteristics for
inputs. The distortion is expressed in r.m.s. terms as the percentage of the prototype amplifier are given in Fig. 7.
the intermodulation products relative to the higher frequency.
At the full rated output of 7W, the response curve is flat to within
The magnitude of the beat-note distortion was found to be negligible ±1dB (relative to the output level at lkc/s) for frequencies ranging
in the prototype amplifier. from 15c/s to 14kc/s.

The phase shift in the feedback loop is only 117° at a frequency of

22kc/s when the loop gain is unity. Consequently, the amplifier has a
stability margin at high frequencies of 63°. The phase-shift
characteristic below 100c/s is not given in Fig. 7. Nevertheless,
adequate stability at low frequencies has been achieved in the
amplifier.

Tone Controls

The treble tone-control potentiometer RV1 gives continuously

variable control from +10dB to -10dB at 10kc/s, and the bass
potentiometer RV3 gives such control from +8dB to -5dB at 100c/s.

Sensitivity

The input voltage required for the amplifier (including the

tone-control network) to give the rated output of 7W is 200mV. This
is enough to allow its use with most types of crystal pick-up head or
f.m. tuner unit.
Hum and Noise

In several copies of the prototype used with a.c. mains supplies, the
hum and noise level, with the volume control set to zero, was always
better than 60dB below 7W. This is not as good as the level normally
achieved with present-day high-quality audio equipment. But the
level is still very satisfactory, and can be considered good if the
limitations set by the series heater chain and the associated high
alternating potentials with a.c. mains operation are borne in mind.

D.C. CONDITIONS

The d.c. voltages at points in the equipment should be tested with

reference to Table 1. The results shown in this table were obtained
using an Avometer No. 8.
 CHAPTER 9

Two- and Three-valve Pre-amplifiers

The circuits described in this chapter have been designed principally short-circuits the unused sockets to earth, an arrangement which
for use with amplifiers built to the 20, 10, and 3W circuits discussed considerably reduces the amount of `break-through' between
in Chapters 5, 6 and 7. They can, however, be used with any power channels. The positions of the switch, from left to right, are:
amplifier which does not require an input signal greater than 250mV Auxiliary, Radio, Tape, Microphone (in the 2-valve circuit only)
for full output. Input facilities are provided for magnetic and crystal Microgroove and 78 r.p.m.
pick-up heads, tape-recorder playback heads, radio tuner units and (in
the 2-valve circuit) a microphone. An auxiliary input position is The equalisation for disc recordings conforms to the present R.I.A.A.
provided for ancillary equipment such as the tape pre-amplifier characteristics which have been adopted by most of the major
circuit described in Chapter 12. recording companies. The tape-playback characteristic is intended for
use with high-impedance heads when replaying prerecorded tapes at a
An additional output position, which is independent of the tone speed of 71/2 inches per second.
controls, is also provided with both circuits, enabling programmes to
be taken to a tape amplifier while they are being fed from the normal The tone controls used in both circuits cover a wide range of
output position to a power amplifier. Both the auxiliary input and the frequencies and provide boost and cut for high and low frequencies.
additional output positions have jack sockets which are situated at the Switched high- and low-pass filter networks are included in the
front of the chassis. 3-valve circuit so that unwanted signals such as rumble and record
scratch can be eliminated.
All the input sockets in each circuit are connected to one switch
which selects one input at a time. In both circuits this switch also

CIRCUIT DESCRIPTIONS anode and grid of the first EF86. This arrangement has been chosen
because the grid-circuit impedance of the first stage should be low. A
Two Mullard high-gain pentodes, type EF86, are used in the 2-valve low impedance at this grid lessens hum pick-up and reduces the effect
circuit of Fig. 2. In the 3-valve circuit shown in Fig. 3, two EF86 and of plugging-in external low-impedance circuits. Furthermore, the
a Mullard double triode, type ECC83, are used. arrangement also results in low gain in the first stage. Hence, Miller
effect between the anode and grid of the first EF86, which can be
First Stage of 2- and 3-Valve Circuits troublesome when high values of resistance are used in series with
the grid, is reduced.
All the equalisation takes place in the first stage of both circuits, and
is achieved by means of frequency-selective feedback between the
Series resistors are used in the input circuits so that the sensitivity and
impedance of any channel can be adjusted accurately. The component
values given in Figs. 2 and 3 are intended for sources encountered
most frequently, but the sensitivity and impedance* of each channel
can be altered by changing the value of the appropriate series resistor.

No feedback is used in the second stage of the 2-valve pre-amplifier.

The full output from the anode of the second EF86 is taken to the
passive tone-control network and an auxiliary output is taken from
the junction of R20 and R21 to the programme-recording jack socket.

*The impedance of the input channels includes the grid impedance of the EF86
modified by the feedback components as well as the impedance of the input
network.

Second Stage of 3-valve Circuit

The second stage of the 3-valve pre-amplifier circuit of Fig. 3 has a

linear frequency-response characteristic and its gain is reduced by a
small amount of negative feedback applied at the control grid by way
of the resistor R23. The output of this stage is taken from the anode to
the tone-control network, and also to the programme-recording output Output Attenuating Networks
socket. Because of the negative feedback, the tone controls have very
little effect on the frequency response at this anode. The sensitivity and impedance of each channel have been chosen for
sources which are likely to be encountered most frequently. The level
Third Stage of 3-valve Circuit of output signal from the pre-amplifier will have to be adjusted if it is
to be used with power amplifiers having different sensitivities from
The output stage of the 3-valve pre-amplifier is made up of the filter the 20W circuit. The signal should be attenuated by using a simple
circuits. The low-pass filter is situated between the two sections of potential divider as shown in Fig. 1. The values of resistance in Fig. 1
the ECC83, and is arranged on the switch SB. It incorporates a (a) are for an attenuator suitable for the control-less version of the
Mullard wound pot-core inductor, type WF1428, in a π-type network. 10W amplifier (Chapter 6) and those in Fig. 1 (b) are suitable for the
The high-pass filter is arranged on the switch SC and consists of two control-less 3W amplifier (Chapter 7).
RC networks in a feedback loop around the second triode section of
the valve.
H.T. Supplies

The h.t. currents taken by the 2-valve and 3-valve pre-amplifiers are 2-valve 3-valve
respectively 3mA at 230V and 6mA at 250V. The values of the pre-amplifier pre-amplifier
smoothing components for these supplies depend on the combination 1 1
(a) Front panel 9x3 12 /2 x 3 /2
of preamplifier and power amplifier. Appropriate values are given in 1 1 1
(b) Rear panel 16 x 2 /2 20 /2 x 2 /2
Tables 1 and 2. The points from which these supplies are taken are 7 5
(c) Mounting strip 8 /8 x 1 12 /8 x 1
shown in the complete circuit diagrams of Chapters 5,6 and 7 for the 1 1 5 1
(d) Cover panel (two) 9 /8 x 3 /2 12 /8 x 4 /8
20W, 10W and 3W amplifiers.
The pieces of the chassis should be marked and cut as indicated in
The l.t. current taken by the 2-valve pre-amplifier is 0.4A at 6.3V and
Fig. 4 (2-valve circuit) and Fig. 5 (3-valve circuit).
that required by the 3-valve circuit is 0.7A at 6.3V.
For ease of assembly, components should be mounted on the
CONSTRUCTIONAL DETAILS
tagboards, and these should then be bolted to the mounting strip
before the strip is attached to the rear plate. Diagrams showing the
The chassis design and layout of both pre-amplifiers are intended
positions of the components on the tagboards are given in Figs. 7 and
specifically for the home constructor. Conventional box-type chassis
8 for the 2-valve pre-amplifier, and in Figs. 10, 11, 12 and 13 for the
are not used. Instead the chassis are made up on the unit system, the
3-valve circuit.
separate parts being joined together during the assembly of the
equipment. Both chassis are made up of five separate pieces of 16
When the mounting strip and tagboards have been attached to the rear
s.w.g. aluminium sheet, the dimensions (in inches) of which are as
plate, connections should be made between the valveholders and
follows:
components. The valveholders should be mounted so that the valves
will be on the outside of the completed chassis and the valves should
be fitted with screening cans.
 It should be noted that the input-selecting switch for both circuits has
The potentiometers and switches should be mounted on the front a third set of controls not indicated in the circuit diagrams, which
panel of the chassis and the other components which make up the short-circuits the unwanted inputs to earth. Details of these switches
tone-control network should be connected to them. The front panel are shown in Figs. 14 and 15.
should then be bolted to the back panel, and the remaining
components should be connected in position as indicated in the Details of the low-pass filter switch SB used in the 3-valve
general layout diagrams of Fig. 6 (2-valve pre-amplifier) and Fig. 9 pre-amplifier are drawn in Fig. 16. The construction of the high-pass
(3-valve pre-amplifier). The series input resistors should be filter switch SC is such that the capacitors of the filter are
connected directly between the input sockets and the selector switch. progressively connected in parallel. The wafer arrangement is drawn
in Fig. 17. This. arrangement reduces switch clicking and also
obviates the need for large capacitors in the network.
The connection to the programme-recording output jack on the front valve circuit may be connected to a coaxial output socket on the back
panel of the 3-valve pre-amplifier should. be made with coaxial cable of the pre-amplifier if it is needed for a fixed recorder.
from the junction of R27 and C13. This jack socket and that of the 2-

PERFORMANCE

The values for hum and noise in the pre-amplifiers which are quoted
below for each input channel have been measured with each of the
pre-amplifiers connected to a 20W power amplifier. The
measurements were made at the output socket of the power amplifier
when the input terminals of the pre-amplifier were open-circuited.
The frequency-response curves were also obtained with these
combinations of pre-amplifier and power amplifier.

The sensitivity figures given below provide full outputs from both
pre-amplifiers. Total harmonic distortion in the two-valve
pre-amplifier is less than 0.15 % at the rated output level and is only
0.24% at ten times this output. Total harmonic distortion in the
3-valve pre-amplifier is less than 0.1% at the rated output and only
0.65% at ten times the rated output. A rapid increase in distortion
does not occur until the pre-amplifiers are considerably overloaded.
 Magnetic Pick-up Position

PICK-UP INPUT CHANNELS 2-valve 3-valve

pre-amplifier pre-amplifier
Equalisation curves for the magnetic and the crystal pick-up input Input Impedance 100kΩ 100kΩ
channels in both pre-amplifiers are drawn in Fig. 18. The difference Sensitivity at 1 kc/s
in sensitivities (indicated in this figure) between the positions for (a) microgroove 4.8mV 7mV
microgroove and 78 r.p.m. records is achieved with the different (b) 78 r.p.m. 13mV 12mV
basic levels of feedback provided at the respective positions of the Hum and Noise (below 20W)
selector switch. The change is necessary because standard discs are (a) microgroove 55dB 53dB
recorded at a higher level than microgroove discs. (b) 78 r.p.m. 57dB 58dB
This channel is most suitable for pick-up heads of the
variable-reluctance type, but moving-coil types which have higher
outputs can be used if a larger value of series resistance is included.

The tape equalisation characteristic used in each pre-amplifier is

shown in Fig. 19. Each channel is intended for replaying pre-recorded
tapes using high-impedance heads, and the characteristics adopted
result in good performance with these heads. The frequency-response
curve for playback with EMI test tape TBT1 is drawn in Fig. 20. If a
greater sensitivity is required, the value of the series input resistor can
be decreased until the desired sensitivity is obtained.

Crystal Pick-up Position

2-valve 3-valve
pre-amplifier pre-amplifier
Input Impedance 100kΩ 100kΩ

Sensitivity at 1 kc/s
(a) Microgroove 70mV 150mV
(b) 78 r.p.m. 210mV 270mV
Hum and Noise (below 20W)
(a) microgroove 55dB 53dB MICROPHONE INPUT CHANNEL
(b) 78 r.p.m. 57dB 58dB (two-valve circuit only)
Input Impedance 1MΩ
Low- and medium-output crystal pick-up heads can be used for this Sensitivity 7.5mV
channel. The input is loaded with the 100kΩ resistor (causing bass Hum and Noise (below 20W) 44dB
loss) in order that its characteristic shall approximate to that of a
magnetic cartridge, and to allow the same feedback network to be The frequency-response characteristic for this channel is given in Fig.
used. This produces the best compromise with most types of pick-up 21. The microphone input channel is intended for use with
head. high-impedance systems such as crystal microphones or magnetic
microphones with transformers.
However, if the head is not suitable for this form of loading, or if its
output is too high, then it can be connected to the auxiliary input RADIO INPUT CHANNEL
socket, which is discussed fully below. With this channel, the pick-up
output is fed into a 1MΩ load which compensates automatically for 2-valve 3-valve
the recording characteristic.
pre-amplifier pre-amplifier
TAPE PLAYBACK INPUT CHANNEL Input Impedance 1MΩ 1MΩ
Sensitivity 330mV 250mV
2-valve 3-valve Hum and Noise (below 20W) 63dB 63dB
pre-amplifier pre-amplifier
Input Impedance 80kΩ(approx) 80kΩ(approx)
Sensitivity at 5kc/s 4mV 2.5mV
Hum and Noise
(below 20W) 52dB 47dB
The frequency-response characteristics of both preamplifiers are
given in Fig. 21. With the values of impedance and sensitivity quoted
above, this channel should meet most requirements. Other values can
easily be obtained, however, by altering the feedback resistor and the
series input resistor. If the input impedance of the channel is too high,
it can be reduced by connecting a resistor of the appropriate value
between the input end of the series resistor and the chassis.

AUXILIARY OUTPUT POSITION

The output from the second EF86 of both circuits is available at this
auxiliary position enabling a record of programme material to be
made with tape equipment. This additional output is taken to a jack
socket at the front of the chassis. Excellent recordings can be made
even when the input signal is derived from a low-output magnetic
AUXILIARY INPUT CHANNEL pick-up.

It can be seen from the circuits that the auxiliary channel is identical An output of about 250mV is available at this socket, and the
with the radio input channel in both pre-amplifiers. The channels can impedance is low. Recording equipment plugged into this socket
therefore be used for high-output crystal pick-ups, for example, or for should not have an input impedance less than 500kΩ. The tone
a tape pre-amplifier such as that described in Chapter 12. If it is controls and the filter networks are inoperative when this output is
desired to use a microphone with the 3-valve circuit, a separate used.
microphone preamplifier or an input-mixing pre-amplifier can be
connected to the auxiliary socket. The auxiliary input is taken to a
jack socket at the front of the chassis. This makes it easier to connect
a portable tape recorder. The jack-socket termination of the auxiliary
input position is such that insertion of the jack disconnects the coaxial
socket on the rear panel.

TONE CONTROLS

The treble and bass tone-control characteristics of the pre-amplifiers

are shown in Figs. 22 and 23. These indicate that an adequate
measure of control is provided in both units for most applications.

Low-impedance controls have been adopted in the 2-valve circuit so

that any capacitance resulting from the use of long coaxial leads
between the preamplifier and main amplifier will have a minimum
effect on the output impedance of the pre-amplifier.
FILTER NETWORKS (three-valve circuit only)

D.C. CONDITIONS The characteristic of the low-pass filter has a slope of approximately
20dB per octave. The components of the network are arranged for
The d.c. voltages at points in the equipment should be tested with operation at 5, 7 and 9kc/s in positions (a), (b) and (c) respectively of
reference to Table 3 for the 2-valve circuit and to Table 4 for the the switch SB in Fig. 3. Position (d) of the switch gives a flat
3-valve circuit. The results shown in these tables were obtained using characteristic.
an Avometer No. 8
The characteristic of the high-pass filter has a slope of approximately
12dB per octave. Operation is at 160, 80 and 40c/s respectively at
positions (a), (b) and (c) of the switch SC. Position (d) of the switch
cuts at 20c/s: this is considered preferable to allowing the response to
continue to lower frequencies because of the possibility of the
amplifier or loudspeaker being overloaded.

.
 CHAPTER 10

Four-channel Input-mixing Pre-amplifier

It may often be desirable to use several signal sources with an microphone should be connected to a high-impedance input stage. If
amplifying unit which has provision for only one input signal. The a low value of resistance of 1.5MΩ is chosen for R1, for example, the
circuit described in this chapter is capable of handling four input combination of the series capacitive elements of the microphone, the
signals and of supplying a mixed output voltage suitable for driving a grid-circuit capacitance C1 and the grid resistance R1 will result in a
single-input amplifier. Two of the input channels of the mixer are loss of about one-third of the output voltage from the microphone at a
suitable for microphone signals, a third for a radio or equalised-tape frequency of 100c/s. Consequently, a value of 10MΩ has been
input and a fourth for crystal pick-up signals. selected for R1 to provide the high-impedance input for the crystal
microphone and to prevent the loss of bass output voltage.
The circuit as it stands provides an output voltage of 40mV and was
intended to be used with the version of the 10W amplifier in which If it is required to use a low-impedance microphone such as a
the tone-control network is disconnected, described in Chapter 6. moving-coil or ribbon type, the mixer can be made suitable by using
With a simple modification to the output stage, the mixer will provide a step-up transformer in the grid circuit of either EF86. A switched
an output voltage of up to 800mV. arrangement for low- and high-impedance microphones is shown in
Fig. 2. The connections marked A and B in Fig. 2 should replace
CIRCUIT DESCRIPTION those similarly marked in Fig. I. The output leads of the microphone
transformer should be made as short as possible to avoid hum pick-up
Both microphone input stages of the input mixing circuit are and loss in treble response.
identical. Each is equipped with the Mullard low-noise pentode, type
EF86, operating with grid-current bias obtained by means of the The output from each microphone input stage is RC coupled to the
high-value grid resistor R1. grid of one half of the Mullard high-µ double triode, type ECC83.
The radio and pick-up input stages are also connected to this grid by
The internal impedance of a crystal microphone is predominantly way of the resistors R11 and R12, RV13. This half of the ECC83 is
capacitive, and the capacitance is of the order of 2000pF. Therefore, arranged as a voltage-amplifying stage.
to avoid loss in terminal voltage at low audio frequencies, the

The potentiometers RV4, RV9 and RV13 serve for the they reduce any interaction between the channels considerably. These
adjustment of signal level and the mixing of the microphone fixed resistors also ensure that the grid of the ECC83 will not be
and pick-up input channels. Adjustment for the fourth channel short-circuited when any one of the potentiometers is set at a
will be achieved by means of the gain control incorporated in minimum.
the radio unit used at the radio input terminals. (If it is The output stage of the mixer unit is suitable as it is shown in Fig. 1
required, control of the radio input can be achieved by for use with the control-less version of the IOW amplifier. The circuit
connecting a potentiometer to the radio input socket in the way has been arranged so that the sensitivity at the microphone inputs is
that RV13 is joined to the pick-up socket.) 3mV for an output voltage of 40mV, and this is sufficient for crystal
microphones. The sensitivities of the other input stages, for the same
The values of the resistors R5, R10, R11 and R12 have been chosen output voltage, are 230mV and 250mV for the radio and pick-up
so that, in combination with the potentiometers RV4, RV9 and RV13, terminals respectively.
Feedback is taken from the anode to the grid of the first half of The output voltage is obtained from the second half of the ECC83
the ECC83 by way of the components R14 and C9. The which has been connected as a cathode follower. This type of
purpose of this is to provide a low -impedance at the grid and connection provides a low output impedance which, in Fig. 1, has the
hence minimise the loss in response at treble frequencies approximate value of 600Ω. Because of this low impedance, little, if
any, attenuation of high notes will occur as a result of cable
caused by Miller effect between the anode and the grid of the capacitance if a long cable is required between the mixer unit and the
triode. main amplifier. The input impedance of the main amplifier must be
greater than 100kΩ to ensure that the bass frequencies will not be
attenuated by the coupling capacitor C11.

The h.t. current taken by the mixer circuit is 5mA at 190V. The value
of the smoothing resistor depends on the power amplifier used, and
appropriate values are given in Table 1. The l.t. current drawn by the
unit is 0.7A at 6.3V.
CONSTRUCTION AND ASSEMBLY and R16 were each 47kΩ, the output voltage of the unit would be
about 400mV.
Chassis details of the input-mixing pre-amplifier are given in Fig. 3.
Space has been left in the chassis for the transformer (Fig. 2) If the required output voltage from Fig. l had been obtained by
necessary for converting the high-impedance microphone input into attenuating the output voltage from the cathode load in the final
one suitable for low-impedance microphones. Space has also been stage, the low output impedance resulting from the cathode-follower
provided for the optional volume control for the radio input channel. action would have been lost.

The chassis is made up of five separate pieces of 16 s.w.g. aluminium Frequency Response
sheet. The dimensions (in inches) of these pieces are as follows:
The response curve of the mixer unit, measured between the
(a) Front panel 111/2 X 3 microphone socket and the output terminals is shown in Fig. 8. The
(b) Rear panel 181/2 x 21/2 curve is flat to within +3dB, relative to the response level at lkc/s,
(c) Mounting strip 1111/16 x 25/8 from 20c/s to 20kc/s. The response curve measured between either
(d) Cover plate (two) 111/2 x 311/16 the radio or the pick-up socket and the output terminals is shown also
in Fig. 8. Because the EF86 stages are not included for this second
A suggested layout for the components of the four-channel position of measurement, the bass response is slightly extended and
pre-amplifier is given in Fig. 7. The arrangement of the components the curve is flat to within ±2dB, compared with the 1 kc/s level, from
on the tagboard is shown in Figs. 4, 5 and 6. The first stage of the 15c/s to 20kc/s.
pre-amplifier has been modified in this layout to incorporate the
facilities for low-impedance microphones. The transformer is Hum and Noise
included, and a switching arrangement allows the first stage to be
used for either high- or low-impedance microphones. Also, the Measurements of hum and noise were made with 100kΩ resistors
optional volume control for the radio input channel is shown in Fig. connected across the microphone and pickup terminals and with the
7. potentiometers fully advanced. These arrangements simulate a
reasonable practical condition. The mixer was connected to a 10W
PERFORMANCE amplifier and measurements were made across a 15Ω load resistor.
The voltage measured in this way was 38mV.
Output and Sensitivity
The full output of the amplifier is 10W which corresponds to a
The maximum output voltage of the mixer unit as it is drawn in Fig. l voltage. of 12.3V across the 15Ω load resistor. Consequently the hum
is 40mV. This output is obtained with an input signal voltage of and noise level is 50dB below 10W. The background level of the
3mV at either microphone socket, 230mV at the radio terminals or power amplifier alone is better than 70dB below 10W, so the figure
250mV at the pick-up terminals. of 50dB must be attributed to hum and noise in the mixer unit.

If greater outputs from the unit are required to drive, for example, an D.C. CONDITIONS
amplifier incorporating a tone-control network, these can be achieved
simply by adjusting the coupling between the anode of the first triode The d.c. voltages at points in the equipment should be tested with
of the ECC83 and the grid of the second. If the capacitor C10 is reference to Table 2. The results shown in this table were obtained
joined directly to the anode of the first triode, an output of 800mV with an Avometer No. 8.
will be available. Intermediate values of output voltage can be
achieved by altering the values of R15 and R16. If, for example, R15
 CHAPTER 11

Three-watt Tape Amplifier

The functions associated with both the recording and playback In addition to the basic equalisation, some measure of tone control
processes are, in most tape recording equipment, fulfilled by a single may be desired in a complete recorder. Equalisation is arranged to
amplifier with suitable switching facilities. It is only with a few of the provide a level overall frequency characteristic. The tone control
more elaborate and costly items of equipment that separate recording provided allows for modifications to be made to the playback
and playback amplifiers are used. characteristic so that the treble response can be attenuated to suit the
demands of the individual listener. The control gives 12dB of cut at
The circuit to be described in this chapter combines the requirements 10kc/s.
of both processes, and it can be used with many of the tape decks that
are fitted with a combined record-playback head and a separate erase The quality of the performance of the complete equipment is
head. With such a deck, the amplifier forms a self-contained necessarily limited by the output stage of the playback system. A
recording and reproducing unit which is capable of an excellent higher standard of reproduction can be achieved if a good-quality
performance, whether it is reproducing recordings made on the same amplifier and pre-amplifier are fed from the low-level output of the
unit, or whether pre-recorded tapes are used. playback amplifier.

Any tape amplifier must, if an acceptable performance is to be Excluding the limitations mentioned in the previous paragraph, and
achieved, provide compensation for the unequal response over the the distortion introduced by the tape itself, the major source of
audio-frequency range that is inherent in the process of magnetic distortion in this unit occurs in the recording stage. The operating
recording. In keeping with general practice, the treble equalisation is conditions for the second section of the ECC83 are chosen as a
associated with the recording channel and the bass equalisation is compromise between the requirements of gain and output voltage and
incorporated during playback. the desirability of maintaining a high a.c./d.c. load ratio. The voltage
required from this stage to provide a recording current of 200µA
It has been pointed out in Chapter 2 that the extent of treble depends on the series resistance between the anode and the recording
attenuation is dependent on the tape speed. head. A low value for this resistance, whilst reducing the
output-voltage requirements, also increases the distortion of this
It is thus desirable to provide separate high-frequency correction for stage. The total harmonic distortion in the recording stage at the peak
the different tape speeds. In this amplifier, equalisation is provided level of 20V which has been adopted in this design is not more than
for speeds of 33/4, 71/2 and 15 inches per second. 2% at lkc/s.
CIRCUIT DESCRIPTION appropriate to the particular tape speed. The degree of boost is
controlled by the resistor R7 and the damping resistors R8 and R9
The circuit diagram of the combined record-playback amplifier is connected in parallel at SB2 with the capacitors C7 and C8. The steep
given in Fig. 1. fall in boost that occurs below the resonant frequency is modified by
damping the inductor and by partially shunting R7 at SB1 at the
Three stages of the circuit are common to both recording and appropriate frequencies.
playback processes. A fourth stage acts as an r.f. oscillator for the
biasing and erasing signals when recording, and is used as an The treble boost obtained will be correct for many combinations of
audio-output stage in the playback process. A subsidiary stage of the tape and head, but it may be too great for others. If this is so, the
recording amplifier, which is excluded from the playback circuit, is damping on L1 should be increased by connecting a resistor in
the recording-level indicator stage. parallel with C9 and reducing the values of R8 and R9. The optimum
values should be determined by listening tests.
Controls
During playback, the correct feedback for bass equalisation is
Four controls are provided in the circuit provided by the resistors R11, R12 and R13 arranged on SB3.

(1) The switch SA sets the amplifier for either the recording or A low-level output of 250mV, having a source impedance of 50kΩ,
playback condition. can be taken from the anode load of this stage of the amplifier and
(2) The switch SB gives the appropriate equalisation for a tape can be used either during recording for monitoring purposes, or
speed of either 15, 71/2 or 33/4 inches per second. The switch must be during playback for feeding an external pre-amplifier and power
in the correct position during playback as well as when recording. amplifier.
(3) The tone control RV15, which is operative only during
playback, is situated in the anode circuit of the first section of the Recording Output Stage
ECC83, and gives variable treble cut.
(4) The gain control RV20 operates during both recording and The output from the anode of the equaliser stage is taken to the grid
playback processes. It does not influence the low-level output which of the second section of the double triode by way of the gain control
is available at the anode of the second stage of the amplifier. RV20. Further high-frequency boost is added to the recording signal
by the capacitor C15 in combination with the resistor R23.
Valve Complement
The recording signal from the anode of the second section of the
The amplifier uses four Mullard valves and one Mullard germanium ECC83 is taken by way of a parallel-T network to the recording head.
diode. These are The network presents its highest impedance at the biasing frequency.
Bias is fed to the recording head immediately after the T-network.
(a) Type EF86, low-noise pentode, used in the input stage for both This arrangement produces a substantially constant current drive to
recording and playback functions. the recording head and provides efficient rejection of the bias voltage
(b) Type ECC83, double triode. The first section of the valve is at the anode of the output valve.
used in the equaliser stage for recording and playback. The second
section is used as the output stage when recording and as a voltage- H.F. Oscillator (Record) or Power Output Stage (Playback)
amplifying stage during playback.
(c) Type EL84, output pentode. When recording, this is used as the The output pentode, type EL84, acts as an audio output stage during
oscillator valve, and in the playback process, it is used in a playback. In the recording process, the EL84 is used to provide the
single-ended audio-output stage to drive the loudspeaker. h.f. oscillations for the biasing and erasing signals.
(d) Type EM 81, tuning indicator, used in the recording-level stage.
(e) Type OA81, germanium diode, used as the indicator-circuit The bias signal is introduced into the recording head through the
rectifier. capacitor C24, the value of which determines the bias current flowing
in the head. The bias voltage is obtained from the anode of the EL84.
Input Stage
The oscillator coil and the choice of component values for the
The pentode, type EF86, acts as a voltage amplifier for both oscillator circuit will depend on the types of record-playback and
recording and playback processes. It is possible to record from either erase heads used. The details given with Fig. 1 are those suitable for
microphone or radio sources. Both inputs are fed to the grid of the record-playback heads having an impedance between 15 and 30kΩ
valve, the radio being attenuated to the level of the microphone input. and erase heads with an impedance between 200 and 300Ω. Details
The switching is achieved by inserting the jacks so that only one for heads with other values of impedance should be obtained from the
input may be used at a time. manufacturer of the tape deck used.

Equaliser Stage The bias oscillator coil and the primary winding of the output
transformer are arranged in series. The latter is bypassed by the
One section of the double triode, type ECC83, is used in the second switch SA8 when in the recording position.
or equaliser stage of both processes. The tone control which is
operative only during playback, is also located in this stage. The presence of the capacitor C25 prevents an abrupt cessation of the
oscillations when the amplifier is switched from the recording to the
During the recording process, a resonant circuit, containing a playback condition, and thus prevents magnetisation of the record-
Ferroxcube pot-core inductor L1 is used to provide treble playback head. The erase head is earthed for the playback process by
equalisation. The value of tuning capacitance in the resonant circuit is the switch SA9.
selected by the equalisation switch SB to give maximum treble boost
On playback, approximately 10dB of negative feedback is used, the CONSTRUCTION AND ASSEMBLY
feedback voltage being taken from the secondary winding of the
output transformer to the cathode of the second section of the double The chassis is made up of five separate pieces of 16 s.w.g. aluminium
triode. The harmonic distortion in the output stage is not more than 3 sheet. The dimensions (in inches) of these are
% at 1kc/s for an output level of 3W. The output power from the (a) Main chassis 11 x 11
playback amplifier is fed by way of the transformer T1 to either a (b) Base 153/4 x 65/8
3.75, 7.5 or 15Ω speaker. (c) Internal screen 9 x 23/8
(d) Internal screen 7 x 23/8
Recording-level Indicator (e) Small bracket 11/4 x 1/2

Each piece should be marked as shown in the chassis drawings of

Fig. 3, and the holes should be cut as indicated. It is important that,
when bending the sheet, the scribed lines should be exactly along the
angles. This ensures that the pieces will fit together properly when
assembled.

Before assembling the record-playback switch SA around the screens,

it will probably be found convenient to fix the following components
to the chassis:

(1) The erase and playback sockets which have to be fitted to the
chassis beneath wafer 3 of switch SA.
(2) The nylon-loaded valveholder for the ECC83, complete with
the skirt for the screening can. Pins 1 and 9 should face towards the
coaxial sockets.
(3) A five-way tagboard. This should be bolted to the internal
screen (c) in the position marked `Tagboard No. 6' in the layout
diagram of Fig. 11.
(4) The small fixing bracket (f), which should be bolted to the
screen (c) as indicated in Fig. 3(a).

The construction is best continued by assembling wafers 1 and 2 of

switch SA around the internal screen (d). The wafers should be
arranged so that positions 6 and 7 are nearest the chassis and the face
A tuning indicator, type EM81, is fed from the anode of the second of each wafer described as the `rear' in the switch diagram (Fig. 4) is
section of the ECC83 through a detector circuit using a germanium farthest from the switch plate. The internal screen (c) should be added
diode, type OA81. to the assembly, both screens should be bolted together, and wafer 3
should be fitted in position, again with its `rear' face (Fig. 4) farthest
The value of the resistor R31 in the target-anode circuit governs the from the switch plate. The general arrangement of the switch wafers
sensitivity of the indicator, and has been chosen to give a sufficiently and internal screens is shown in Fig. 11. Details of spacers required
high sensitivity to allow a large series resistance R26 to be used for the assembly are shown in Fig. 4. The use of a shock-proof
between the diode and the second anode of the ECC83. This large washer is recommended between the switch plate and the front panel
resistance minimises the loading effect on the recording output stage. when the switch and screens are fitted to the main chassis. A suitable
layout of the other components is shown in Fig. 11. The arrangements
The operating conditions of the EM81 are normally chosen so that the of the small components on tagboards are shown in Figs. 5 to 10.
target shadow `closes' for a recording current of 200µA. They can Details of the equaliser switch are given in Fig. 12.
however be chosen so that the shadow `closes' at lower peak
recording levels if reduced peak distortion is desired at the expense of Frequency Response
the signal-to-noise ratio.
Treble boost is incorporated during recording and bass boost during
In the playback position, switch SA6 disconnects the recording-level playback. Separate equalisation is provided for tape speeds of 15, 71/2
indicator stage from the h.t. supply, and this gives a positive or 33/4 inches per second to give the following attainable overall
indication of the position of the record-playback switch SA. response (relative to the level at lkc/s):

If a meter indication of the recording level is required, the EM81 15 in./sec: ±3dB from 35c/s to 17kc/s
stage of Fig. 1 can be replaced by the circuit given in Fig. 2. The 71/2 in./sec: ±3dB from 50c/s to 13kc/s
valve used is an ECC83. The signal is taken from a point in the anode 33/4 in./sec: ±3dB from 50c/s to 6kc/s
load R22 of the recording output stage (V2B in Fig. l), and the
milliammeter (1mA, full-scale deflection) is included in the cathode The overall response at the higher frequencies depends on the type of
circuit of the second half of the ECC83 of Fig. 2. The variable head used and on the magnitude of the bias current. The response
resistor RV53 should be chosen to suit the resistance of the figures given above and the curves drawn in Fig. 13 will normally be
milliammeter. obtained with a bias current of 0.5 to 1.0mA through heads of
medium impedance.
The playback characteristic of the amplifier at a tape speed of 71/2
inches per second is designed to the C.C.I.R. specification, thus
permitting excellent reproduction of pre-recorded tapes. The
recording characteristic is arranged to give a flat frequency response
in conjunction with this replay characteristic.

60kc/s, and the erase-head voltage was about 25V again at a

frequency of 60kc/s.

Test I - D.C. Voltages

The d.c. voltages at points in the equipment should be tested with

reference to Table 1. The results shown in this table were obtained
using an Avometer No. 8.

Test II - Amplifier on Playback

Three pieces of equipment are required for this test:

(1) A signal generator covering a frequency range from 20c/s to
20kc/s;
(2) A valve voltmeter covering a frequency range from 20c/s to
20kc/s;
(3) A load resistor of 15Ω with a 6W rating.

The 15Ω resistor should be connected to the speaker sockets. The

Sensitivity record-playback switch SA should be in the playback position and the
tone control RV15 should be set for the flat response.
The sensitivity of each amplifier is measured with the control RV20
set for maximum gain. This, of course, does not apply to the A signal from the generator, having a frequency of 5kc/s, should be
low-level output measurements: the control is not effective until after applied to the record-playback socket (which normally
this point of the circuit. accommodates the connection plug from the record-playback head).
The consequent output signals should be measured on the voltmeter,
Recording Sensitivity both at the low-level output socket and across the load resistor.
(measured at 1kc/s, with recording-head audio current of 200µA)
(a)Microphone input: 2.5mV for peak The input voltage should be adjusted to give an output voltage across
(impedance = 2MΩ) recording level the load resistor of 67V for each tape speed, and the input required
(b)Radio input: 250mV for peak for this output should be noted. The voltage readings that should be
(impedance = 680kΩ) recording level obtained are given in Table 2.

Playback Sensitivity With the switch SB set to 15 inches per second, and with the
(measured at 5kc/s for all tape speeds for 3W audio conditions given in the previous paragraph, the gain control should be
output or 250mV low-level output) varied until the output voltage across the load resistor drops to 50mV.
(a) 15 in./sec: 5.2mV The frequency of the signal should then be reduced to 100c/s and the
(b) 71/2 in./sec: 2.8 mV values of boost given in Table 3 should be observed at the 15Ω load
(c) 33/4 in./sec: 1.1 mV output. The switch SB should be changed to 71/2 and 33/4 inches per
second, and the boost measurement should again be made.
TEST PROCEDURE
The bass-boost characteristics for the three tape speeds are shown in
The four tests outlined below are intended as simple, yet quite Fig. 14.
effective, checks for the combined record-playback amplifier.
The treble cut introduced by full application of the tone control RV15
The values given in the various tables and figures were obtained from should be determined for all tape speeds, and the curve obtained
the prototype amplifier, using Brenell record-playback and erase should correspond to that shown in Fig. 15.
heads. The bias current used throughout was 1.0mA at a frequency of
Test III - Amplifier on Record

The instruments required for this test are:

(1) A signal generator covering a frequency range from 20c/s to
20kc/s;
(2) A valve voltmeter* covering a frequency range from 20c/s to
20kc/s.

The record-playback and erase heads should be connected to the

appropriate sockets in the amplifier, and the equipment should be
switched to the recording condition.

For a tape speed of 15 inches per second, a signal at lkc/s should be

applied from the generator to the radio input socket. The magnitude
of this signal should be such that an output of 30m V is obtained at
the low-level output socket.

When the signal frequency is switched to 15kc/s, the boost indicated

in Table 4 should be observed.

For accurate results two separate pieces of p.v.c. covered wire are
recommended for the connection to the valve voltmeter. A coaxial
With tape speeds of 71/2 and 33/4 inches per second, the input signal at
cable may result in considerable errors in the measurements because
lkc/s should again be adjusted to give a low-level output of 30m V.
of the parallel capacitance which is introduced.
The change to input frequencies of 10 and 5kc/s should give the
treble boost indicated in Table 4.

The treble-boost characteristics for the three tape speeds are shown in
Fig. 16. These characteristics will be suitable for many combinations
of tape and head, but the peaks may occur at too low a frequency and
may indicate too much boost for other combinations. If this is so,
modification of the equalisation network (that is the components
connected to switch SB2) may prove beneficial (see page 89). Any
alteration to the treble-boost characteristics will, of course, cause changes
in the overall frequency-response characteristics of Fig. 13.

Values for the recording sensitivity for an output voltage measured at

the second anode of the ECC83 are given in Table 5. A test of the
recording-level indicator should show that the EM81 `closes' with
approximately 20V at this anode.
An alternative method of checking the recording amplifier is
possible. For any of the tape speeds, the voltage developed across a
50Ω resistor connected in series with the recording head can be
observed for the full range of signal frequencies. The response curve
so obtained should agree with the appropriate curve for the prototype
amplifier, plotted in Fig. 13. For these observations, it will be
necessary to disconnect one end of the resistor R34, otherwise only
the bias signal will be measured.
Test IV - Bias Level

For this test, two pieces of equipment are required:

(1) A valve voltmeter which will indicate accurately at frequencies of
up to 70kc/s;
(2) A resistor of 50Ω.

The resistor should be soldered in series with the earthy end of the
record-playback head, and the voltage developed across the resistor,
with no input signal, should be measured with the voltmeter.

The voltage developed across the resistor should be 50mV, which

corresponds to a bias current of 1.0mA flowing in the 50Ω-resistor.
 Power-supply Unit

CIRCUIT DESCRIPTION

The circuit diagram for a power supply suitable for use with the 3W The values of the dropper resistors R100 and R101 shown in Fig. 17
tape amplifier and the tape preamplifier described in the following should be chosen to give respectively a potential of less than 350V
chapter is given in Fig. 17. The value of R101 will depend on across the reservoir capacitor C100 and a potential of 300V across
whether the unit is to be used with the amplifier or the preamplifier. C101. It may be found that R100 is not needed.
The requirements of the unit are that it should provide (i) a direct
voltage of 300V at a current of 50mA for the amplifier and 25mA for CONSTRUCTIONAL DETAILS
the pre-amplifier and (ii) an alternating voltage of 6.3V at a current of
2A.

The choice of rectifier will depend on the tape deck used. Normally,
the Mullard full-wave rectifier, type EZ80, will be suitable. However,
with tape decks that use electrical braking for the tape transport The chassis consists of one piece of 16 s.w.g. aluminium sheet, 83/4
system, it is essential that the Mullard type EZ81 be used, so that the in. long, and 7 in. wide. It should be marked as shown in the chassis
current which is required for the short braking periods can be drawing of Fig. 18 and the holes should be cut as indicated. Mounting
supplied. holes for the mains switch, the fused voltage-selector and the pilot
lamp are shown in the figure, but if it is so desired, these components
If the EZ80 is used, the series resistance in each anode circuit of the may be mounted elsewhere when, of course, there will be no need to
rectifier must be at least 21552; if the EZ81 is used, the minimum cut the particular holes in this chassis.
series resistance is 200Ω for each anode. Very few transformers
meeting the specification given below will have a total winding The chassis drawing shown in Fig. 18 is for a mains transformer of
resistance less than these minimum requirements, but should it be the inverted-mounting type. If a different type is used, it will be
lower, a series resistance large enough to make up the minimum necessary to drill grommet holes to enable the leads to be taken
should be added to each anode circuit. The required value of this through the chassis.
series resistance is derived in the way shown on page 28.

The assembly of the power unit should be accomplished quite easily

by referring to the layout diagram of Fig. 19. This figure again caters
for a transformer of the inverted-mounting type.
To avoid unnecessary expense. input and output plugs have not been
used: securing clips suffice to anchor the mains and h.t. leads. The
pilot lamp is also an optional component.

It is very important when wiring the electrolytic capacitor to ensure

that the correct section is used as the reservoir capacitor C100. The
section which is identified as the `outer', or else marked with a red
spot, should be used for this component.
 CHAPTER 12

Tape Pre-amplifier

The tape-recording pre-amplifier described in this chapter is intended CIRCUIT DESCRIPTION

to be used in conjunction with a high-quality pre-amplifier and power
amplifier. The unit combines the function of both recording and The circuit diagram of the combined record-playback pre-amplifier is
playback amplification, although, for the playback operation, it acts given in Fig. 2. The record-playback switch SA is shown in the
only as an equalising stage, giving sufficient output voltage to drive position for recording and the equaliser switch SB is set for
the high-quality amplifying system. equalisation at a tape speed of 71/2 inches per second. The playback
output is taken from the anode of the second EF86, and the remaining
The general principle of the design has been to preserve simplicity as stages are used only for the recording operation.
far as is compatible with high-quality performance. The distortion
introduced in the recording channel has been reduced to such an Controls
extent that it is probable that the quality of performance will be
limited only by the magnetic tape itself, provided, of course, that the Two switch banks are used in the pre-amplifier. Switch SA provides
tape deck used is of satisfactory performance. The level of total the change from the recording to the playback process, and switch SB
harmonic distortion in the recording process should not be greater provides the equalisation appropriate to one of the tape speeds: 33/4,
than 0.5 % with a recording current of 150µA through the head. 71/2, and 15 inches per second.

Equalisation to correct for head and tape characteristics is provided The gain control RV13 is the only other control in the pre-amplifier.
for each of the tape speeds: 33/4, 71/2, and 15 inches per second. The
high-frequency equalisation is applied during the recording process Valve Complement
and the low-frequency correction during playback.
The complete pre-amplifier uses five Mullard valves and one Mullard
Treble equalisation is achieved by using a wound Ferroxcube germanium diode. These are:
pot-core inductor, Mullard type WF816, in a resonant circuit between
the first and second stages of the amplifier. The frequency at which (a) Type EF86, low-noise pentode, used in the input stage.
maximum treble boost occurs is determined by the tuning capacitance (b) Type EF86, used in the second stage.
which is adjusted by the switch SB3. The value of resistance in the (c) Type EF86, used in the output stage for recording.
feedback network to give bass equalisation during playback is (d) Type ECC82, double triode, used as a push-pull oscillator for the
selected for each tape speed by the switch SB1. bias and erase signals.
(e) Type EM 81, tuning indicator, used in the recording-level stage.
There is no provision for tone control in this preamplifier. It is (f) Type OA 81, germanium diode, used as the indicator-circuit
anticipated that such control will be available with the associated rectifier.
amplifying equipment.
Input Stage Recording Output Stage

The pentode, type EF86, acts as a voltage amplifier for both The third stage of the unit, operative only during the recording
recording and playback processes. It is possible to record from either process, uses another EF86, the grid of which is fed from the volume
microphone or radio sources, the radio input also being convenient control RV13. The stage is designed to give low harmonic distortion
for recording from crystal pick-ups. Both inputs are fed to the grid of at peak levels of recording current, and the distortion should not
the valve, the radio input being attenuated to the level of the exceed 0.5 % for a recording current of 150µA.
microphone input. Switching is achieved by inserting the jacks, so
that only one input may be used at a time. The recording current is fed to the recording head by way of a
parallel-T network, which acts primarily as a bias-voltage rejector
Equaliser Stage circuit. The series resistance of the network is needed in this stage to
ensure a constant drive to the head, and its inclusion is also desirable
As was stated above, no tone control is incorporated in this stage. The to preserve a satisfactory a.c./d.c. load ratio for the EF86 of the third
output is taken during playback from across part (R10) of the anode stage.
load of the second EF86. The output supplied is 250mV at a source
impedance of 15kΩ. A rearrangement of the anode load resistance H.F. Oscillator Stage
(that is, R10 and R11) can be made, if required, to give an output of,
for instance, IV at a source impedance of 60kΩ. The halves of the double triode are arranged as a push-pull oscillator.
The stage is designed so that the ECC82 draws approximately the
The output of the second stage of the amplifier is fed during the same h.t. current during playback, when it is inoperative, as it does
recording process from the anode of the EF86 by way of the gain under its oscillatory conditions during recording. In this way, the total
control RV13 to the grid of the following EF86. current drain for either the recording or the playback process does not
alter greatly, and the design of the power supply is consequently
A resonant circuit containing a wound Ferroxcube pot-core inductor simplified.
L3 (Type WF816) is used to provide treble equalisation. The value of
tuning capacitance in the resonant circuit is selected by the switch The oscillator coil consists of a wound Ferroxcube pot-core, type
SB3 to give the maximum treble boost at frequencies appropriate to WF1388, in which the secondary windings are tapped so that the coil
the tape speed used. The extent of treble boost is controlled by the can be used with most of the commercially available tape decks. The
resistor R30 and the damping resistors R34 and R35 connected in value of the coupling capacitor C18 will vary with the type of
parallel with the capacitors C26 and C25. The steep rise in boost record-playback head used, and a list of appropriate values is given in
which occurs below the resonant frequency is modified by damping Table 1.
the inductance, and by partially shunting the resistor R30 at the
appropriate frequency. Power Supply

The treble boost obtained will be correct for many combinations of The unit described in Chapter 11 is suitable for use with the
tape and head, but it may be too great for others. If this is so, the pre-amplifier circuit. The heater-current requirements of the two tape
damping on L3 should be increased by connecting a resistor in circuits are similar; allowance for the difference in h.t. currents is
parallel with C27 and reducing the values of R34 and R35. The made by a suitable choice of value for R101.
optimum values should be determined by listening tests.
CONSTRUCTION AND ASSEMBLY

The chassis for the tape pre-amplifier is made up of seven separate

pieces of 16 s.w.g. aluminium sheet. The dimensions (in inches) of
these are:

(a) Main chassis 11 x 11

(b) Base 153/4 x 65/8
(c) Internal screen 8 x 23/8
(d) Internal screen 7 x 23/8
(e) EM81 mounting bracket 13/4 x 13/4
(f) Small bracket (two) 11/4 x 1/2

Each piece should be marked as shown in the chassis drawings of

Fig. 3, and the holes should be cut as indicated. It is important that,
when bending the sheet, the scribed lines should be exactly along the
angles. This ensures that the pieces will fit together properly when
assembled.

Recording Level Indicator

The tuning indicator, type EM81, used in this stage is fed from the
The values of the resistors R31, R32 and R33 arranged on switch SB1 anode of the recording stage. The large series resistance R18 is used
have been chosen to give appropriate feedback for bass equalisation to minimise the loading of the recording stage. If a meter indication
during playback. To avoid capacitive coupling, this section of the of the recording level is required, the circuit given in Fig. 1 can be
equaliser switch is arranged on the front of the switch wafer.
 correct position for this to be so. The valveholder can be fixed in the
main chassis, as described in Chapter 11, if it is required to mount the
EM81 at the end of extension leads. An extra hole with a diameter of
1 in. will have to be drilled for the valveholder, in place of the
grommet hole C shown in Fig. 3(a).

Identification of the coil winding suitable for the erase heads of

various decks can be made by referring to Table 1. The appropriate
value of C18 is also given in this table. The bias capacitor is
connected to the grey lead from the secondary winding for all the
makes of deck indicated. The connections in the diagram for the
primary winding of the coil and for the blue (earth) lead from the
secondary winding are also the same for the decks listed.

PERFORMANCE
Frequency Response
used in place of the EM81. This alternative arrangement is similar to
that suggested on page 92 for the 3W tape amplifier. The overall response of the recorder depends on the type of head
used, the magnitude of the bias current, and, to some extent, on the
Details of the record-playback switch SA are given in Fig. 4. Before tape employed. The low-frequency response depends on the amplifier
assembling this switch around the screens, it will probably be found used rather than on the type of head, and in this design, it will not be
convenient to fix the following components to the chassis: more than 3dB down at 50c/s. The high-frequency response depends
on the tape speed and the gap width of the head. With heads having a
(1) The erase and record-playback coaxial sockets which have to be gap width of 0.0005in., the following performance figures (relative to
fitted to the chassis beneath wafer 3 of the switch. the level at lkc/s) can be obtained:
(2) All the valveholders. Only the three EF86 valves should be
skirted, and the holders for these valves should be nylon-loaded. The 15 in./sec: ±3dB from 35c/s to 17kc/s
holder for the input valve V1 (EF86) should be of an 71/2 in./sec: ±3dB from 50c/s to 13kc/s
anti-microphonic type - that is, having a flexible mounting. 33/4 in./sec: ±3dB from 50c/s to 6kc/s
(3) The two small brackets (f) which should be bolted to the internal
Flexibility has been achieved in the treble-boost circuits of the
screens (c) and (d) as shown in the layout diagram in Fig. 10.
pre-amplifier. Consequently good control of the complete response of
the pre-amplifier is possible; and equalisation to reasonably high
The construction is best continued after the above components have
frequencies is practicable. However, the degrees of control and
been fitted by assembling wafers 1 and 2 around the internal screens.
equalisation depend on the individual adjustments of component
The wafers should be arranged so that positions 6 and 7 are nearest
values to suit the head and tape being used. The component values
the chassis and the face of each wafer described as `rear' in the switch
given in this chapter apply only to the pre-amplifier when it is used
diagram (Fig. 4) is farthest from the switch plate. The internal screen
with Scotch recording tape and a Collaro tape deck.
(c) should be added to the assembly, both screens should be bolted
together, and wafer 3 should be fitted in position, again with its `rear'
Ferroxcube pot-core inductors are adequately screened to prevent
face farthest from the switch plate. The general arrangement of the
excessive hum or stray bias being picked up, and they do not appear
switch wafers and internal screens is shown in Fig. 10. The use of a
to cause more circuit-ringing than RC networks which produce the
shake-proof washer is recommended between the switch plate and the
same treble boost.
front panel when the switch and screens are fitted to the main chassis.
The arrangement of the smaller components on tagboards is shown in
The playback characteristic of the pre-amplifier conforms to the
Figs. 6 to 9, and a suitable layout of these boards and the other
C.C.I.R. specification, thus permitting excellent reproduction of
components in the chassis is shown in Fig. 10. In fitting the EM81
pre-recorded tapes.
mounting bracket, the valveholder should be fastened to the bracket
so that the solder tags are on the same side as the flange on the
The recording characteristic is arranged to give a flat frequency
bracket. The gap between pins 1 and 9 on the holder should face the
response in conjunction with this playback characteristic. Additional
flange to ensure that the tuning eye of the EM81 faces forward. The
head losses occurring during playback will normally be capable of
indicator should appear in the centre of the front panel of the
correction by the tone controls in the associated amplifying systems.
equipment, and the bracket should be bolted to the chassis in the
 TEST PROCEDURE

The four tests outlined below are intended as simple, yet quite
effective, checks for the pre-amplifier.

The values given in the various tables and figures were obtained from
the prototype pre-amplifier, using Collaro record-playback and erase
heads. The bias current used throughout was 1.0mA at a frequency of
60kc/s, and the erase-head voltage was about 25V, again at a
frequency of 60kc/s.

Test I - D.C. Voltages

The d.c. voltages at points in the equipment should be tested with

reference to Table 2. The results shown in this table were obtained
using an Avometer No. 8.

Test II - Amplifier on Playback

Sensitivity
Two pieces of equipment are required for this test:
The sensitivity of the recording process is measured with the control
RV13 set for maximum gain. This does not apply to the playback (1) A signal generator covering a frequency range from 200/s to
sensitivity because the gain control is not operative at the point from 20kc/s;
which the output is taken for the associated equipment. (2) A valve voltmeter covering a frequency range from 200/s to
20kc/s.
Recording Sensitivity
(measured at lkc/s, with recording-head audio current The record-playback switch SA should be in the playback position. A
of 150µA) signal from the generator, having a frequency of 5kc/s, should be
(a) Microphone input: 0.5mV for peak applied to the record-playback socket (which normally
(impedance = 2MΩ) recording level
(b)Radio input: 250mV for peak
(impedance = 1.2MΩ) recording level accommodates the connecting plug from the record-playback head).
The consequent output signal should be. measured on the voltmeter at
Playback Sensitivity the output socket.
(measured at 5kc/s for each tape speed for output of
250mV) The input voltage should be adjusted to give an output voltage at the
(a) 15 in./sec: 5.5mV output socket of 250mV for each tape speed, and the input required
(b) 71/2 in./sec: 2.4mV for this output should be noted. The voltage readings that should be
(c) 33/4 in./sec: 1.0mV obtained are given in Table 3.
 Test III - Amplifier on Record

The instruments required for this test are:

(1) A signal generator covering a frequency range from 20c/s to

20kc/s;
(2) A valve voltmeter1 covering a frequency range from 20c/s to
20kc/s.

The record-playback and erase heads should be connected to the

appropriate sockets in the preamplifier, and the equipment should be
switched to the recording condition.

A signal at lkc/s should be applied from the generator to the radio

input socket. The magnitude of this signal should be such that an
output of 15m V is obtained at the output socket.

The boost indicated in Table 5 should be obtained at the appropriate

tape speed when the signal frequency is altered to the value shown in
the table.

The treble-boost characteristics for the three tape speeds are shown in
Fig. 12. These characteristics will be suitable for many combinations
of tape and head, but the peaks may occur at too low a frequency and
may indicate too much boost for other combinations. If this is so,
modification of the equalisation network (that is the component
connected to switch SB3) may prove beneficial (see page 103). Any
alteration to the treble-boost characteristics will, of course, cause
changes in the overall frequency-response characteristics of the
pre-ampler.

'For accurate results, two separate pieces of p.v.c. covered wire are recommended for the
connections to the valve voltmeter. A coaxial cable may result in considerable errors in
the measurements because of the parallel capacitance which is introduced.

Values for the recording sensitivity for an output voltage measured at

the anode of V3 (EF86) are given in Table 6. A test of the
recording-level indicator should show that the EM81 `closes' for each
tape speed with approximately 15V at the anode of the
recording-output valve.

An alternative method of checking the recording amplifier is

possible: for each tape speed, the voltage developed across a 50Ω
resistor connected in series with the recording head can be observed
for the full range of signal frequencies. The response figures so
obtained should agree with the values obtained with the prototype
pre-amplifier and listed on page 110. For these observations, it will
be necessary to disconnect one end of the resistor R26, otherwise
only the bias signal will be measured.

Test IV - Bias Level

For operation at such high sensitivities, great care should be taken to
ensure that the signal measured is, not composed mostly of hum. It is For this test, two pieces of equipment are required:
advisable, therefore (a) to use screening cans on the three EF86, (b)
to screw on firmly the base of the amplifier, and (c) to use coaxial (1) A valve voltmeter which will indicate accurately at frequencies of
cables for the connections to the measuring equipment. up to 70kc/s;
(2) A resistor of 50Ω.
The input voltage at 5kc/s should be varied until the output voltage
drops to 50mV. The frequency of the signal should then be reduced to The resistor should be soldered in series with the earthy end of the
40c/s and the values of boost listed in Table 4 should be observed at record-playback head, and the voltage developed across the resistor,
the output socket. with no input signal, should be measured with the voltmeter.

The bass-boost characteristics for the three tape speeds are shown in The voltage developed across the resistor should be 50mV, which
Fig. 11. corresponds to a bias current of 1.0mA flowing in the 50Ω resistor.
 CHAPTER 13

Seven-watt Stereophonic Amplifier

The circuit diagram drawn in Fig. 1 is for a high-quality, stage and the pentode sections are arranged as a push-pull output
dual-channel amplifier designed principally for stereophonic stage with distributed loading. The EZ81 forms a conventional power
reproduction but also providing facilities for monaural applications. supply with resistance-capacitance smoothing, providing the h.t. for
Only one channel of the amplifier is drawn. The circuitry appearing both channels.
between the dotted lines is for the left-hand channel: it should be
duplicated for the right-hand channel. The circuitry drawn outside the The rated output-power reserve of each channel is 7W, at which level
dotted lines (the power supply, for example) is common to both the total harmonic distortion is always better than 1.0 %. The low
channels. level of distortion is achieved by using 21dB of negative feedback,
the feedback voltage being taken from the secondary winding of the
The total complement of valves used in the amplifier consists of one output transformer in each channel to the cathode circuit of the
double triode, type ECC83, four triode pentodes, type ECL82, and corresponding input stage. The sensitivity of the circuit, even with
one full-wave rectifier, type EZ81. The double triode is shared this high value of feedback, is 120mV, which is ample for use with
between the two channels, one section of the valve being used in each existing stereophonic crystal pick-up heads.
channel for voltage amplification. Two of the ECL82 are used in each
channel. The triode sections of these valves form a phase-splitting

CIRCUIT DESCRIPTION being connected to position b of SA1, single-channel reproduction is

possible.
Resistors and capacitors appearing in the left-hand channel of the
amplifier are numbered 1, 2, 3, etc.; the corresponding components in (c) Dual-channel monaural reproduction from an f.m. tuner unit.
the right-hand channel are numbered 101, 102, 103. etc. The input socket in Fig. 1 is connected for monaural applications. If
position c of SA3 is connected to the right-hand input terminal
Input Selector Switch instead of the left-hand terminal, the circuit will be suitable for
reproducing stereophonic transmissions. If position c of SA3 is
The input stages of both channels are connected to the 3-way selector earthed instead of being connected to the input socket, the system
switch SA. The switch positions indicated in Fig. 1 provide the gives single-channel monaural reproduction from an f. m. tuner unit.
following facilities:
Input Stage
(a) Stereophonic reproduction from stereophonic crystal pick-up
heads. The triode sections of the ECC83 are used for voltage amplification,
one section being used in each channel. A simple treble tone-control
(b) Dual-channel monaural reproduction from a monaural pick-up network consisting of the components RV2 and C1 (RV102 and
head. In this position, the left-hand pick-up input terminal is `live', C101) is incorporated in the grid circuit giving continuously variable
and both channels are connected in parallel at the input. The input treble cut. A volume control RV1 (RV101) is also included in the
terminal for the right-hand channel of a stereophonic pick-up head is grid circuit. Dual-ganged potentiometers are used in the prototype for
earthed at position b of SA2. If position b of SA3 is earthed instead of the tone-control components RV2 and RV102 and also for the
volume control components RV1 and RV101, so that equal
adjustments can be made simultaneously to both channels. It may be Phase-splitting Stage
found more convenient to use dual-concentric potentiometers for
these controls because adjustments can be made separately to each The output from the anode of V2A is taken to the grid of V3A by
channel with them. way of the resistors R15 and R12 (R115 and R112). These
components also form the grid-leak resistance for the pentode section
Instability can result from the large amount of negative feedback V2B, and R16 and R12 (R116 and R112) comprise the grid-leak
introduced to the cathode circuit of each input stage, and the simplest resistance for V3B. Balance between the output voltages from the
method of ensuring stability is to reduce the loop gain of each triodes of the phase-splitter is governed by the values of these three
channel. To achieve this reduction at high audio frequencies, the resistors.
capacitor C2 (C102) is connected to bypass the anode load resistor
R3 (R103). To reduce the loop gain at bass frequencies, the parallel Output Stage
combination R6 and C6 (R106 and C106) is included in the coupling
circuit between the first and second stages of each channel. This The pentode sections of the ECL82 are used in a push-pull output
manner of coupling results in resistive attenuation at low audio stage. Distributed loading is used, and the primary winding of the
frequencies, so that the phase shift is limited. output transformer is tapped so that 20 % of the winding in each
anode circuit also appears in the corresponding screen-grid circuit.
Balance Control (The choice of tapping points is discussed in Chapter 3.)

To compensate for any differences in acoustical output resulting front Negative Feedback
unequal outputs from the stereophonic pick-up head or unequal
sensitivities of the loudspeakers, a balance control consisting of The amount of negative feedback which is used between the
RV23 and RV123 is incorporated between the volume and tone secondary winding of the output transformer in each channel to the
controls. This is made up of a logarithmic potentiometer connected in cathode circuit of the input stage of that channel is 21dB. The output
reverse in one channel and an antilogarithmic potentiometer resistance of each channel with this feedback is 0.54Ω measured at
connected normally in the other. The characteristics of this control the 15Ω output terminal. This gives an adequate damping factor of
are discussed on page 28 in Chapter 4. If dual-concentric approximately 28.
potentiometers are used for the volume control RV1, RV101, balance
can be achieved with them, and the potentiometers RV23, RV123 can Power Supply
be omitted.
A conventional power supply using the Mullard full-wave rectifier;
type EZ81, with resistance-capacitance smoothing, provides the h.t.
for both channels. Both supply connections are made to the common
reservoir capacitor C15, but smoothing is achieved separately in each CONSTRUCTION AND ASSEMBLY
channel with the components R19, C12 and 8119, C112. The chassis for the 7W stereophonic amplifier is made from three
separate pieces of 16 s.w.g. aluminium sheet, The dimensions (in
The resistor R22 in the cathode circuit of the EZ81 should, with the inches) of these pieces are:
transformer resistance, provide the minimum limiting resistance (a) Main chassis r 16 x 101/2
quoted for the valve. The choice of value for limiting resistors is (b) Tagboard mounting strip 127/8 x l
discussed on page 28. A total h.t. current of 140mA at 278V is (c) Cover plate 13 x 71/2
required for the amplifier, and the total heater current needed is
4.57A at 6.3V. Each piece should be marked as shown in the drawings of Fig. 2, and
the holes should be cut as indicated. Where bending is required, it is
A reduction in the number of components is achieved if an h.t. line important that the bends should be made accurately at the lines for
which is common to both channels is used. The necessary the pieces to fit together properly on assembly.
modifications (with reference to Fig. 1) are as follows: R8, R11 and
R19 should be decreased to 15, 5.6 and 5.6kΩ respectively; C5 and Most of the smaller components should be mounted o n tagboards,
C8 should be increased to 16µF each; R108, R111, R119 and C105, the wiring of which is shown in Figs. 4 to 7, and the tagboards should
C108, C112 are not required; R103, C102 should be connected to C5; be fixed to the mounting strip. The larger components such as
R110 should be connected to C8; R114 should be connected to C12. switches, valveholders and transformers should be mounted in the
chassis in the positions indicated in the layout diagram of Fig. 3. The
recommended wiring between these components is also indicated in Hum and Noise
Fig. 3. It will be seen that the two channels of the amplifier are
separate, one channel being in each half of the chassis. The level of hum and noise in each channel of the prototype amplifier
is better than 60dB below the rated output.
PERFORMANCE
Tone Controls
Distortion
A continuously variable treble-cut tone control is included in each
The total harmonic distortion was measured in the prototype channel of the amplifier. The characteristic of this control is shown in
amplifier with a continuous sine-wave input signal at 400c/s. With Fig. 8, from which it can be seen that full application of the control
21dB of negative feedback, the distortion for the rated output of 7W gives about 23d B of cut at 10kc/s.
per channel is always better than 1.0 %. The variation of distortion
with output power is shown in Fig. 9. D.C. CONDITIONS

Intermodulation distortion, measured with carrier and modulating The d.c. voltages in each channel should be tested with reference to
frequencies of 10kc/s and 40c/s respectively, is better than 1.5 °/. Table 1. The results shown in this table were obtained with an
Avometer No. 8.
Cross-talk interference is better than 40dB below 7W at 100c/s, lkc/s
and 10kc/s.

Frequency Response

The frequency-response characteristic of each channel of the

prototype amplifier, measured for an output of l W is shown in Fig. 8.
At 20kc/s. the response is 3dB below the level at lkc/s and at 15c/s it
is 1.5dB down. The response at the rated output power of 7W is 3dB
down at 15c/s.

Sensitivity ..

The sensitivity of each channel of the circuit (including the controls)

with 21dB of negative feedback is 115mV at balance for the rated
output of 7W.
 CHAPTER 14

Three-valve Stereophonic Amplifier

The circuit drawn in Fig. 1 has been designed to provide dual-channel (approximately 6dB) of feedback in each channel. The distortion
stereophonic amplification of a reasonably good quality at a fairly occurring when a continuous sine-wave signal is applied to the input
low cost. Only one channel of the amplifier is shown in Fig. 1. The is about 5 %, but under the normal input conditions of speech or
circuitry appearing between the dotted lines is for the left-hand music signals, the level of distortion will be somewhat lower. The
channel only: it should be duplicated for the right-hand channel. The amplifier, although not comparable in specification with high-quality
circuitry drawn outside the dotted lines (the power supply, for equipment, has been found to give most pleasing stereophonic results
example) is common to both channels. Each channel uses one during listening tests in normal-sized living rooms. The feedback
Mullard triode pentode, type ECL82, the triode section of which is voltage is taken from the secondary winding of the output
used for voltage amplification and the pentode section for power transformer of each channel and applied to the cathode circuit of the
amplification. The h.t. supply for both channels is obtained from one triode section of the ECL82 in that channel.
Mullard full-wave rectifier, type EZ80.
CIRCUIT DESCRIPTION
The sensitivity of each channel for an output of 2W is 300mV at
balance, which is adequate for use with the majority of stereophonic Resistors and capacitors appearing in the left-hand channel of the
crystal pick-up heads. This sensitivity has been achieved in spite of amplifier are numbered 1, 2, 3, etc.; the corresponding components in
the small complement of valves by using only a small amount the right-hand channel are numbered 101, 102, 103, etc.
Input Selector Switch 10kc/s. Dual-ganged potentiometers are used in the prototype for the
volume control components RV1 and RV101 and also for the
The input stages of both channels are connected to the 3-way selector tone-control components RV2 and RV102 so that equal adjustments
switch SA. The switch positions indicated in Fig. 1 provide the can be made simultaneously to both channels. It may be found more
following facilities: convenient to use dual-concentric potentiometers for the volume and
tone controls because these also allow separate adjustments to be
(a) Stereophonic reproduction from stereophonic crystal pick-up made to the channels.
heads.
Output Stage
(b) Dual-channel monaural reproduction from a monaural pick-up
head. In this position, the left-hand pick-up input terminal is `live', The output stage of each channel consists of the pentode section of an
and both channels are connected in parallel at the input. The input ECL82. Unless precautions are taken, the acoustical outputs from the
terminal for the right-hand channel of a stereophonic pick-up head is loudspeakers of the two channels will not be exactly the same. In
earthed at position b of SA2. If position b of SA3 is earthed instead of order that allowance can be made for any differences, a dual-ganged
being connected to position b of SA1, single-channel reproduction is balance control (RV7 and RV107) is included in the grid circuits of
possible. the pentode sections of the ECL82. This control is made up of a
logarithmic-law potentiometer connected in reverse in one channel
(c) Dual-channel monaural reproduction from an f.m. tuner unit. and an antilogarithmic-law potentiometer connected normally in the
The input socket in Fig. 1 is connected for monaural applications. If other channel. The characteristics of this type of control are discussed
position c of SA3 is connected to the right-hand input terminal on page 28. If dual-concentric potentiometers are used for the volume
instead of the left-hand terminal, the circuit will be suitable for control RV1, RV101, balance between the channels can be achieved
reproducing stereophonic transmissions. If position c of SA3 is with them, and the potentiometers RV7, RV107 can be replaced by
earthed instead of being connected to the input socket, the system fixed resistors (470kΩ).
gives single-channel monaural reproduction from an f.m. tuner unit.
Power Supply
Input Stage
A conventional power supply using the Mullard full-wave rectifier,
The triode section of one ECL82 is used in the first stage of each type EZ80, with resistance-capacitance smoothing provides the h.t.
channel for voltage amplification, providing a stage gain of about 50 for both channels. The supplies are taken from the junction of the
times. A volume control RV1 (RV101) is included in the grid circuit smoothing components C7 and R11, these components and the
of the triode section. A simple tone-control network, consisting of the reservoir capacitor C8 being common to both channels. The total h.t.
components RV2 and C1 (RV102 and C101) is also incorporated in current drain of the amplifier is 80mA at 240V and the total heater
this grid circuit. The control provides continuously variable treble cut current is 2.31 A at 6.3 V.
and, with full application of RV2 (RV102), gives at least 20dB cut at
CONSTRUCTION AND ASSEMBLY PERFORMANCE
Distortion
The chassis for the 2W stereophonic amplifier is made from one
piece of l6s.w.g. aluminium sheet 14 in. long and 11 in. wide. This The total harmonic distortion was measured in the prototype
piece should be marked as shown in the drawings of Fig. 2, and the amplifier with a continuous sine-wave input signal at 400c/s. With
holes should be cut as indicated. 6dB of negative feedback the distortion for the rated output of 2W
per channel is approximately 5 %. The level of distortion for speech
Most of the smaller components are mounted on two tagboards. The or music input signals is somewhat lower than this. The variation of
components for the right-hand channel are arranged on Tagboard No. distortion with output power is plotted in Fig. 6.
1, and those for the left-hand channel are mounted on Tagboard No.
2. Details of these boards are given in Figs. 4 and 5 respectively. It Cross-talk interference is better than 40dB below 2W at 100c/s, lkc/s
will be observed that the arrangement of the components on one and 10kc/s.
board is the `mirror-image' of the arrangement on the other.
Frequency Response
The larger components such as switches, valve holders and
transformers should be mounted on the chassis in the position The frequency-response characteristic of each channel of the
indicated in the layout diagram of Fig. 3. The recommended wiring prototype amplifier is shown in Fig. 7. This response is 3dB below
between these components is also shown in Fig. 3. The wires the response level at lkc/s at frequencies of 40c/s and 40kc/s.
connecting the input sockets to the selector switch should be taken
along the edge of the chassis beneath Tagboard No. 2. This Sensitivity
arrangement follows the practice normally adopted in audio
equipment of separating by as much as possible the sensitive input The sensitivity of each channel o£ the circuit with 6dB of negative
circuit from the rest of the amplifier. feedback is 300mV at balance for the rated output of 2W.
Hum and Noise

The level of hum and noise in each channel of the prototype amplifier
is at least 62dB below the rated output.

Tone Control

The tone-control characteristic is given in Fig. 8. From this it can be

seen that 23dB of treble cut are available at 10kc/s with full
application of the control

D.C. CONDITIONS

The d.c. voltages in each channel should be tested with reference to

Table 1. The results shown in this table were obtained with an
Avometer No. 8.
 CHAPTER 15

Stereophonic Pre-amplifier

The circuit diagram for a high-quality dual-channel stereophonic The halves of the input sockets are connected to the switch SB which
pre-amplifier is drawn in Fig. 1. Only one channel is shown in the selects one input at a time. This switch also short-circuits the unused
figure. The circuitry which appears between the dotted lines is for the sockets to earth, an arrangement which reduces considerably the
left hand channel and it should be duplicated for the right-hand amount of `break-through' between inputs. The positions of the
channel. The switches and sockets which are drawn outside the dotted switch, from left to right, are: Auxiliary, Radio, Tape, Microgroove
lines are common to both channels. and 78 r.p.m.

The circuit is fundamentally a combination of two two-valve The equalisation for disc recordings conforms to the present R.I.A.A.
pre-amplifiers (Chapter 9), so that the total complement of valves is characteristics which have been adopted by most major recording
four low-noise pentodes, type EF86. Each channel is arranged to give companies. The tape playback characteristic is intended for use with
an output voltage high enough to drive an amplifier built to the 20W high impedance heads when replaying pre-recorded tapes at a speed
design. A simple potential divider (Fig. 2) can be used to attenuate of 7½ inches per second. The tone controls used in each channel
this output to the level required for amplifiers built to either the 10W cover a wide range of frequency and provide boost and cut for high
circuit or the 3W circuit. A n auxiliary output taken from a tapping and low frequencies. The switch SA permits either or both channels
point on the anode load of the second EF86 is available in each of the pre-amplifier to be used. In its central position, both channels
channel for programme monitoring. are available. In position a, the switch short-circuits the left-hand
channel to earth while in position c, the right-hand channel is
Facilities have been provided in this equipment for magnetic and inoperative. The switch SC between the volume control RV28
crystal pick-up heads, tape-recorder playback heads and radio tuner (RV128) and the output to the power amplifiers connects the two
units. An auxiliary socket for any input source convenient to the user channels for normal or reversed* stereophonic reproduction or for
is also provided in the equipment. dual-channel monaural reproduction.

*Reversed stereophony is achieved when the left-and right-hand input signals are
conveyed to the right- and left-hand loudspeakers respectively.

Prototype of
Stereophonic
Pre-amplifier

CIRCUIT DESCRIPTION

Each channel of the pre-amplifier is made up of two stages, and each Series resistors are used in the input circuit so that the sensitivity and
stage uses a Mullard high-gain pentode, type EF86. All the impedance of any input can be adjusted accurately. The component
equalisation takes place in the first stage, and is achieved by means of values given in Fig. 1 are intended for sources encountered most
frequency-selective feedback between the anode and the grid of the frequently, but the sensitivity and impedance* for each input can be
first EF86. There is no feedback in the second stage, and the output altered by changing the value of the appropriate series resistor.
from the anode of the second EF86 is taken by way of the capacitor
C12 to a passive tone-control network. Maximum output is obtainable with the arrangement shown in Fig. 1,
and the output voltage developed is sufficient to drive the 20W circuit
This arrangement has been chosen because the grid-circuit impedance described in Chapter 5. This output voltage can be attenuated to the
of the first stage should be low. A low impedance at this grid lessens level required to drive the 10W or 3W circuits of Chapters 6 or 7 by
hum pick-up and reduces the effect of plugging-in external low- introducing the appropriate network shown in Fig. 2 between the
impedance circuits. Furthermore, the arrangement also results in low volume control RV28 and the switch SC in Fig. 1. An auxiliary
gain in the first stage. Hence, Miller effect between the anode and output signal, which is suitable for monitoring programmes, is taken
grid of the first EF86, which can be troublesome when high values of from a point on the anode load of the second EF86.
resistance are used in series with the grid, is reduced.
To compensate for any differences in acoustical output that may ganged to an antilogarithmic potentiometer connected normally in the
occur in the two amplifying channels, a balance control consisting of other channel. The characteristics of such a control are discussed on
RV16 and RV116 is included between the two stages. This consists page 28.
of a logarithmic potentiometer connected in reverse in one channel

Dual-ganged potentiometers are used in the prototype for the tone Each piece should be marked as shown in Fig. 3 and the holes should
and volume controls. It may be found more convenient to use be cut as indicated. It is important that, when bending the pieces, the
dual-concentric potentiometers which allow separate adjustments to scribed lines should lie exactly along the angles. This ensures that the
be made to the channels. If this type is used for the volume control pieces will fit together properly when assembled.
RV28, RV128, there will be no need for the balance control RV16,
RV116, which can thus be replaced by fixed resistors 1MΩ. For ease of assembly, components should be mounted on tagboards,
and these should then be bolted to the mounting strips before the
The d.c. supply of the power amplifiers can be used to provide the h.t. strips are attached to the rear plate. Diagrams showing the positions
supply for the channels of the pre-amplifier. The smoothing of the components on the tagboards are given in Figs. 4, 5, 6 and 8, 9,
components should normally be mounted in the chassis of the power 10.
amplifier used with each channel. The values of these components
will vary with the type of power amplifier used, and suitable values When the mounting strips and tagboards have been attached to the
are indicated in Table 1. The h.t. current drawn by each channel of rear plate, connections should be made between the valveholders and
the pre-amplifier is 3mA at 230V, and the l.t. current is 0.4A at a the components on the tagboards. (The valveholders should be
heater voltage of 6.3V. mounted so that the valves will be on the outside of the completed
chassis, and screening cans should be fitted.)
'The impedance at each input includes the grid impedance of the EF86 modified by the
feedback components as well as by the impedance of the input network.
The ganged potentiometers RV22-RV122, RV26RV126 and
RV28-RVI28 should be mounted on the front panel of the chassis,
and the other components which make up the tone-control network
CONSTRUCTIONAL DETAILS should be connected to them. The balance control RV16-RV116 and
The chassis and layout of the pre-amplifier have been designed
the switches SA, SB, SC should also be connected to the front panel,
specifically for the home constructor. A conventional box-type
which should then be bolted to the back panel. The remaining
chassis is not used. Instead, the chassis is made on the unit system,
components should be connected in position. as indicated in the
the separate parts being joined together during the assembly of the
general layout diagrams of Figs. 7 and 11. Details of the selector
equipment.
switch are given in Fig. 12.
The chassis is made up of six separate pieces of 16 s.w.g. aluminium
PERFORMANCE
sheet, the dimensions (in inches) of which are as follows:
The values for hum and noise in the pre-amplifier which are quoted
for each input position have been measured with each channel
(a) Rear panel 19 X 4
connected to a 20W power amplifier. The measurements were made
(b) Front panel 10 x 4½
at the output socket of the power amplifier when the input terminals
(c) Mounting strip (two) 97/8 x 1 of the pre-amplifier were open-circuited. The frequency-response
(d) Cover plate (two) 10¼ x 45/8 curves were also obtained with this combination of pre-amplifier and
power amplifier.
The sensitivity figures given below provide an output of 250mV value of the resistor R3 can be decreased until the desired sensitivity
when the full anode load of the second EF86 is used. All is obtained.
measurements were made with the balance control set for balance.
RADIO INPUT POSITION
PICK-UP INPUT POSITIONS
Equalisation curves for magnetic and the crystal pickup positions are Input Impedance 1MΩ
drawn in Fig. 13. The difference in sensitivities between the positions Sensitivity 330mV
for microgroove and 78 r.p.m. records is achieved by the different Cross-talk at 1kc/s 49dB below 20W
amounts of feedback provided at positions (d) and (e) of the switch Hum and Noise 58dB below 20W
SB1 (SB101).
The frequency-response characteristic for this position is given in
Magnetic Pick-up Position Fig. 17. With the values of impedance and sensitivity quoted above,
this channel should meet most requirements. Other values can easily
Input impedance 100kΩ (approx.) be obtained, however, by altering the feedback resistor R7 and the
Sensitivity at 1 kc/s series resistor R2. If the input impedance of the channel is too high, it
(a) Microgroove 5mV can be reduced by connecting a resistor of the appropriate value
(b) 78 r.p.m. 15mV between the input end of R2 and the chassis.
Cross-talk at 1kc/s
(a) microgroove 50dB below 20W AUXILIARY INPUT POSITION
(b) 78 r.p.m. 60dB below 20W
Hum and noise It can be seen from the circuit of Fig. 1 that the auxiliary position is
(a) microgroove 58dB below 20W identical with the radio input channel The input with the component
(b) 78 r.p.m. 59dB below 20W values shown in Fig. 1 can therefore be used for high-output crystal
pick-ups, for example, or for tape pre-amplifiers such as the circuit
This arrangement is most suitable for pick-up heads of the described in Chapter 12. The auxiliary input is taken to a jack socket
variable-reluctance type, but moving-coil types which have higher on the front panel of the equipment. The insertion of the jack
outputs can be used if a larger value of series resistance R4 is disconnects the coaxial input socket on the back of the chassis.
included.

Crystal Pick-up Position TONE CONTROLS

Input Impedance 100kΩ
Sensitivity at 1 kc/s The treble and bass tone-control characteristics of the pre-amplifier
(a) microgroove 70mV are shown in Fig. 16. These indicate that an adequate measure of
(b) 78 r.p.m. 210mV control is provided in the unit for most applications.
Cross-talk at 1kc/s
(a) microgroove 48dB below 20W Low-impedance controls have been adopted so that any capacitance
(b) 78 r.p.m. 59dB below 20W resulting from the use of long coaxial leads between the pre-amplifier
Hum and Noise and main amplifier will have a minimum effect on the output
(a) microgroove 58dB below 20W impedance of the pre-amplifier.
(b) 78 r.p.m. 59dB below 20 W
HARMONIC DISTORTION
Low- and medium-output crystal pick-up heads can be used with this The total harmonic distortion of each channel of the pre-amplifier is
position. The input is loaded with the l00kΩ resistor R6 in order that less than 0.15 % at normal output levels. At outputs of ten times this
its characteristic shall approximate to that of a magnetic cartridge, level, the harmonic distortion is only 0.24 %.
and to allow the same feedback network to be used. This produces the
best compromise with most types of pick-up head. However, if the AUXILIARY OUTPUT POSITION
head is not suitable for this form of loading, or if its output is too
high, then it can be connected to the auxiliary input socket, the An additional output from the second EF86 of each channel is
function of which is discussed below. available at this position, enabling a record of programme material to
be made with tape equipment. This additional output is taken to a
TAPE PLAYBACK POSITION jack socket at the front of the chassis. The voltage at the auxiliary
output socket of each channel is about 250mV and this output is
Input Impedance 80kΩ (approx.) available at a low impedance. Recording equipment plugged into this
Sensitivity at 5 kc/s 4.0mV socket should not have an impedance less than 500kΩ. The tone
Cross-talk at 1kc/s 49dB below 20W controls are inoperative when this output is used.
Hum and Noise 54dB below 20W
D.C. CONDITIONS
The equalisation characteristic used for this position is shown in Fig.
14. For frequencies above 100c/s, the curve follows the C.C.I.R.
The d.c. voltages in each channel should be tested with reference to
characteristic, but below this frequency, slightly less boost is used.
Table 2. The results shown in this table were obtained using an
The channel is intended for replaying pre-recorded tapes using
Avometer No. 8.
high-impedance heads, and the characteristic adopted results in good
performance with these heads. If a greater sensitivity is required, the