Sorry for the long winded post, but here it is..
Since the discussion is regarding distortion and frequency
response, I would like to share some experience I have had with
improving the fidelity of otherwise mundane audio power
amplifers. Although some of the technicques presented may require
some ingenuity to apply to a very large plate modulator, I feel
that they have worthwhile and audible merit from the standpoint
of linearity, 'tilt', and transient response. In the second case
presented below, A/B tests with various CD recorded music
including violin, Cuban percussion, and industrial styles, the
improvements were very noticeable. Surely there is then merit for
the improvement of the fidelity of the voice waveform.
I would also like to suggest that a triangular wave is an
excellent way to check the linearity of an audio amplifier. The
proper circuit configuration around the phase inverter can make
the sides of the wave obey an extremely straight 45 degree angle,
to the point where the input and output signal shapes are nearly
indistinguishable. 'Tuning' is done by superimposing the input
and output signals on a scope and matching them.
(In an amplifier with poor linearity, the sides of the triangle
wave can 'bellow' out, or 'hourglass' in, just like the trapezoid
RF waveform in a poorly designed or operated linear amplifier (go
figure).)
The first two circuits embody this method in the phase inverter
circuit. I realize that many people may prefer a balanced circuit
throughout, but it may not always be practical.
The first two circuits also embody a cathode feedback scheme
surrounding only the output tubes and the audio output
transformer. This application will 'straighen up' and otherwise
poor transformer quite well, although there are requirements for
plenty of grid drive voltage and ample B+ voltage along with a
large value filter capacitor on the plate supply. If the
transformer is too small, the low frequency power output fidelity
will be less improved since the transformer cannot handle the
extra current demanded when correction takes place. A solution is
to demand less power.
This article describes the conversion of a fairly poor performing
(actually bad sounding) 70W theater amplifier which suffers from
lack of iron. The result is a very clear 40 watt amplifier. A
method of in-circuit measurement of tube matching is shown.
http://www.montagar.com/~patj/mi12188a.htm
This article describes the improvements made to a decent-sounding
50 watt public address amplifier to the point where it is as
faithful as a Quad or McIntosh as far as most people's ears are
concerned. About 50 watts was still obtained (large iron).
http://www.montagar.com/~patj/aph1050.htm
This article demonstrates a method of coupling the output signal
back to the grids around a transformer coupled high power (50
watt) AB2 audio stage.
http://www.montagar.com/~patj/tubeart.htm
For those interested in a QRP experience, this article describes
an extremely linear series modulator for an 807 RF stage.
http://www.montagar.com/~patj/amxmtr_dc_classamod.html
Patrick
