an emitter follower output device. Q11 is a current
source which acts as a load for the emitter follower Q10.
(Yes, I know by convention this is a current sink.) The
green LED acts as a voltage reference of about 2.1 volts.
The 2.2K ohm resistor supplies about 5 mA to turn it on.
The reference voltage is applied to the base of the current
source transistor. The emitter resistor of the current source
is 47 ohms, which sets the current at about 35 mA. The
exact value is not critical.
The current from the positive supply through Q10 is
limited by the 100 ohm resistor in the collector. The
power dissipation in the output transistors is about 0.35
watts in each with no signal. The output transistors could
operate without the heatsinks and be well within their
ratings, but I like to play it safe. This is a Class A output
The current source active load is better than a resistor
because it can supply full current to the output even when
the signal voltage nears the negative supply voltage. The
100 ohm resistor in series with the positive supply protects
the amplifier output from accidental shortcircuits of a
short duration, i.e., a second or two. In the case of a long
term short, this resistor will become a fuse!
In order to keep the circuit from becoming an
oscillator, there is a compensating capacitor from the
collector of the VAS to the base of the input follower
transistor. Ordinary ceramic capacitors are notorious for
changing value as the applied voltage changes. If you use
a ceramic here, it must be of the NPO type (indicating
that the voltage coefficient of capacitance is around zero,
but may be slightly negative or positive). A mica capacitor
is a little more expensive but guaranteed to work well.
The various transistors in the signal path each have an
upper frequency limit. That is, they act like a low pass
filter. Each additional transistor in the signal path
contributes phase shift at some high frequency. This phase
shift is cumulative and if it reaches the point where the
negative feedback becomes positive, the result will be a
peak in the high frequency response. A little more phase
shift produces an oscillator at some high frequency.
The purpose of the compensating capacitor is to
provide a low pass filter much lower than any of the
others, so that the frequency response is reduced as
frequency increases. As long as the "open loop" gain of
the amplifier reaches unity as frequency increases before
the phase shift exceeds 90 degrees by "too much," the
amplifier will be stable when negative feedback is applied.
"Too much" is rather ambiguous. That is a long story
beyond the scope of this article. (If you are interested in
pursuing this topic, look up "phase margin" in any of the
books listed in the sidebar.)
Notice that a 10K resistor is connected to the output
that goes back to the negative input of the first stage.
This is the negative feedback path. This resistor and the
1,500 ohm resistor to the 220 µF capacitor form a voltage
divider that sets the amount of negative feedback, and
thus the AC gain of the amplifier — about 7. 7. The 220 µF
capacitor allows full DC voltage feedback, so the output
DC voltage can be on the order of a few millivolts.
(The voltage on the 220 µF capacitor will be positive a
few tenths of a volt in normal operation, so it need not be
the non-polar type). The amplifier's voltage gain at DC is
unity. Note that the signal gain of the amplifier without
feedback is VERY high. The amplifier won't work without
the feedback connected because it is the feedback that
puts the operating point of the voltage amplifier stage
such that the output is near zero with no signal. (Whew!)
What we have built here is a discrete component
46 October 2013
High-Power Audio Amplifier Construction Manual
by G. Randy Slone, McGraw Hill 1999
This one is useful as a project book. Many of the designs are
suitable for building. It is much more than a construction manual,
however. Slone discusses the most prevalent designs in audio
amplifiers. Not a great deal has changed in the years since the
writing of this book, though some of the transistors used in these
designs are either hard to find or expensive due to their scarcity.
They are still available if you look hard enough. Slone discusses
The Audiophile's Project Sourcebook
by G. Randy Slone. TAB division of McGraw Hill 2002
This book has a number of audio projects of various
complexity including stepped attenuators, various preamplifiers
and tone control circuits, electronic crossover networks, power
amplifiers, an electronic reverberation circuit, and many more.
This one is a cookbook.
Audio Power Amplifier Design Handbook
by Doug Self, Focal Press 5th Edition 2009
This is the most technical of the books, containing lots of data
interesting to an engineer. It covers a number of design possibilities.
Self enumerates the causes of distortion in an amplifier and
systematically shows how to minimize each of them, resulting in
what he calls the "blameless" amplifier. It is not "perfect," but all the
known sources of distortion have been considered and adjusted to
minimize the distortion. This book has the best discussion of
subjectivity. An older version (i.e., the fourth edition) may still be
available. There is a LOT of new material in the fifth edition.
Designing Audio Power Amplifiers
by Bob Cordell, McGraw Hill 2011
Bob has the ability to describe much of the technical detail in a
way that is more intuitive than most authors. It is not a cookbook, but
it is a good discussion of the subject. The circuits in this book use
transistors that are newer and more available (and less expensive
and a bit less capable) than those in Slone's book.
All of the above books discuss the (now almost) standard three-stage Lin architecture and numerous variations for each stage. Self
describes some of these improvements in detail and concludes that
though the "numbers" are better, the improvement is not audible.
Perfection is a worthy goal, however, never to be achieved in an
audio power amplifier, but some designs come very close. I
purchased all four of these books at Amazon.com. I'm not sure all
are still available there.
What is subjectivity and the subjectivist movement? Follow this
link for some interesting reading. You don't have to understand all of
it to appreciate what Doug Self has to say at www.douglas-self.com/ampins/pseudo/subjectv.htm. The discussion is almost word
for word from a section of his book mentioned above.