Simple Class A Amplifier by JL Linsley Hood Wireless World Dec 1970 Jlh1970

Wireless World, December 1970
Simple Class A Amplifier

A postscript to the design published last year
by J. L. Linsley Hood
The author has had the benefit of an extensive and frequently helpful correspondence with readers
following the publication of the circuit design. Attention has been drawn to some obscurities in the
original article and to certain possible improvements in the design. Details are given below.
Power supply
Although much interest was aroused among constructors by the good technical (and audible)
performance given by the amplifier, it is clear that the principal feature in the eyes of many users was the
relative simplicity of the circuit. This being so it must have seemed a pity that the power-supply unit was
somewhat complex. However, the purpose of this power-supply design was to avoid possible
degradation of the amplifier performance by h.t. ripple. The sawtooth ripple voltage across the reservoir
capacitor in a class-A amplifier drawing some 2-3A will be many times greater than in a class-B system,
particularly at the audibly important low-signal levels.
However, a number of measurements have been made since the publication of the original article on an
amplifier of this type, operated from a simple supply unit of the type shown in Fig. 1. There is little
difference in the performance above 100Hz either in total harmonic distortion or in intermodulation
distortion, although the shape of the output power/distortion curve at the onset of overload is modified,
as would be expected, by the ripple on the h.t. line. Below 100Hz the distortion curve rises more steeply
to about 0.2% at 20Hz.
A thermister is necessary, in this case, to slow down the rate of rise of the h.t. voltage. This will get hot in
use.
It now appears that the mains transformers used in the development of the prototype of this amplifier
were not as efficient in respect of apparent secondary circuit resistance or secondary leakage reactance
as some of those which have been supplied for this purpose since the publication of the article. The
reservoir voltage found with the 15 system may be above that given by the author. At switch-on this
can cause a transient overloading of the transistors specified for the series regulator circuit in the original
article. In view of this, it is suggested that these should be an MJ481 or 2N3055, used in conjunction with
a 2N1613 or, better still, a 2N699. These amendments are shown in Fig. 2.
Adjustment of amplifier output current and centre-line potential
The author had supposed, somewhat naively, that most constructors of the circuit would have
somewhere in their workshops a collection of odd-value resistors needed for trimming circuit parameters,
and it was mentioned in the original article that the desired quiescent levels could be set by adjustment
to R2 and R5 or R6. This sort of comment is unhelpful if one is writing away for a kit of parts. In view of
this it is suggested that R2 should be replaced by a resistor in series with a potentiometer, as shown in
Fig.1. The necessary value of resistors R5 and R6 to give an entirely adequate accuracy in the mid-point
voltage setting can be predicted, and the suggested amended values are shown.
Some obscurity arose, inadvertently, in the original diagram concerning the reason for the different
values of input decoupling resistor quoted for mono and stereo use. This was because it was intended
that the one decoupling circuit should serve both channels. Where an unsmoothed h.t. supply is used it
is recommended that the decoupling capacitor should be increased in value to 250F.
Page 1 of 4

Stability of output current setting
Some criticism has been voiced because there is no specific control over the output current value in the
simplest form of this circuit, other than that due to the stability of the current gain of Tr2, whose
performance determines this parameter. In order to meet this point (in anticipation) a circuit was
described in the original article which allowed precise control over the operating quiescent current
without detriment to the performance of the amplifier.
However, measurements made on an amplifier without this addition have shown no significant change in
operating current in somewhat over two years use, and there is also little measurable difference in
current from a minute or so after switch-on to the end of a six-hour period of continuous use. In practice
therefore, in temperate climates at least, the simplest form of the circuit is adequate in this respect. If any
user cares to experiment with an alternative and somewhat more elegant form of quiescent-current
control another regulation circuit is shown in Fig. 3. The transistor used as Tr5 requires to be somewhat
more massive than that used for Tr3 since the mean collector current is twice that of Tr3 and the
maximum voltage and current occur simultaneously. The 2N2905A is just about adequate with a good
heatsink, but a larger power device such as the 2N4919 is preferable.
Alternative transistor types
The amplifier has been built successfully with a wide variety of transistors, including fully complementary
versions to operate from an existing negative h.t. line, and in one case two identical amplifiers have been
made for use with the inputs in paraphase, in order to double the available output voltage swing. One
constructor has, indeed, made a stereo 30W system using two such pairs of amplifiers plus input phase
splitter, as shown in Fig. 4.
However, one transistor change which is recommended is the use of a 2N1711 as Tr3. This has a high
voltage capability equal to that of the 2N1613, and a current gain which is double that of either the
2N1613 or the 2N697. The use of the 2N1711 instead of the former types suggested for Tr3 increases
the feedback factor and approximately halves the typical distortion factor of the system (0.025% at 9W or
0.05% at full power) without detriment in other respects.
Also, a 2N1711 as Tr3 allows the use of 2N3055 devices as Tr1 and Tr2, with a final performance which
is equal to that of the original specification below 100kHz. (The typical current gain of the 2N3055s is
only half that normally found with the MJ480/1 output transistors, and their use was not originally
recommended for this reason.)
Gain/frequency and power/frequency characteristics
These are, in fact, better than the curves published in April 1969. As mentioned in a letter to the editor
published in October 1969, the h.f. fall-off shown was mainly due to an error in the measurement
instrument. Although the performance at h.f. depends to some extent on the layout employed, the small
signal voltage gain, with the component arrangement shown, is flat (within 1dB) to beyond 2MHz. This
may be a snag in some cases because even a small feedback capacitance between output and input (as
may happen, for example, if the output heatsinks are not earthed) may cause the amplifier to oscillate. A
suitable circuit change to reduce the amplifier h.f. response to more normal levels was described in the
letter above. This is not an essential modification the authors own units are still exactly as described in
April 1969.
The output power response of the unmodified amplifier is flat within 1dB to 200kHz.
The l.f. response shown in the original gain/frequency and power/frequency graphs was that determined
for an earlier prototype of the amplifier. During the development of the circuit the values of some of the
capacitors were increased to improve the l.f. performance, and by an oversight the graphs
accompanying the article were not amended. In fact the gain and power graphs can be shown as flat
from 10Hz-200kHz. In this respect, and that of transient response, the class-A design is probably better
than any circuit so far published. The i.m. distortion, at 10W output, (70Hz and 7kHZ, 4:1) is less than
0.1%.
Page 2 of 4

Miscellaneous
Surprise and even alarm has been caused to some constructors by the fact that the output
transistors get hot. However, with adequate heatsinks, which should be black painted, the dissipations in
the transistors are only a small fraction of the makers permitted level, and provided that some care is
taken in the layout to make sure that sensitive components, such as electrolytic capacitors, remain cool,
no reduction in the working life of such a system, in comparison with an equivalent class-B unit for
example, is to be expected.
Some difficulty has apparently been encountered by some constructors because the power supply
regulation system is inoperative when the supply is operated without a load. If an equivalent resistive
dummy load is connected for bench-testing, all should be found to be well.
Finally, it is prudent to wire a small resistor of about 2k across the loudspeaker terminals to make sure
that the output capacitor charges even with the speaker disconnected. Charging of the capacitor by an
accidental short-circuit could cause damage. This addition is shown in Fig. 1. No damage is caused by
operating the amplifier on an o / c output.
Figures
Fig. 1. Suggested amended circuit for 8 or 15 use employing a simplified power supply.
The dotted components reduce the h.f. response and should be used with capacitive loads.
Page 3 of 4

Fig. 2. Amended circuit of power supply for 15
systems.
Fig. 3. Alternative method of quiescent-current control. R1, R2 and C1 in the

original have been deleted. Tr5 is 2N4919 on heatsink or alternative type.
Fig. 4. Scheme for 30W class-A system. The two identical class-A amplifiers, each designed to
give 15W into 8 (34V at 1.5A), are driven by a phase-splitter. The output capacitors have been
removed.
Page 4 of 4

Simple Class A Amplifier by JL Linsley Hood Wireless World Dec 1970 Jlh1970

Uploaded by

Copyright:

Available Formats

Simple Class A Amplifier by JL Linsley Hood Wireless World Dec 1970 Jlh1970

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Simple Class A Amplifier by JL Linsley Hood Wireless World Dec 1970 Jlh1970

Uploaded by

Copyright:

Available Formats

Wireless World, December 1970