In article , Steve Swift
wrote:
A better description is that it compares the input with the output and
adjusts the system to reduce discrepencies. Thus it changes the effect
of any nonlinearity within the loop. It doesn't do this by using
"another" form of distortion as a comparison for "offset".

I was thinking of the Texan amplifier. It uses really cheap 747 op-amps
and a hefty feedback mechanism so that the output waveform is all but
identical to the input waveform (except in amplitude).

[snip]

It is not unusual for a system to include more than one source of
nonlinearity. However the feeback can then act upon the entire system,
and reduce the nonlinearity of the system's response. So far as the
feedback is concerned, it can normally treat all the sources in terms
of their composite effect. It need not employ some 'extra' source
which would otherwise have been absent in order to 'offset' the effect.


One fascinating aspect of the design was that the volume and tone
controls affected only the feedback loop. In effect, the wrong volume,
or a lack of treble was just another sort of error that had to be
corrected.

Well, it would be unusual to do that as part of the power amp stages. But
it was common for feedback to be used for the tone control mechanism.

For example, the widely used 'Baxandall' tone controls used the feedback to
adjust the response in most cases I know of. Indeed, when I pointed out
that it wasn't correct that what you'd said was "all" feedback did I was
thinking of its use to control frequency and phase response (i.e. linear
properties, not nonlinear distortions).

I suspect we should be in alt.semantics :-)


That may well be so for various of the conversations on this group. :-)

However I don't think it is the case here. The problem is that the
comment you made about negative feedback presented it as working
on the basis of requiring "one form of distortion to offset another".
i.e. requiring *two* sources or forms of distortion. My point was
that this is simple incorrect. No 'other' form or source of distortion
is required.

So my point was that the engineering isn't as you described. I don't
regard it as 'semantics' to find that a description is incorrect. I
regard it simply as a mistake. Although I appreciate that the wording
did not say what you meant, and expect that you do understand feedback
perfectly well. But my concern is that someone who does not might be
mislead.

Your description would be more appropriate for a 'nulling' or cancellation
method where two sources are employed. But need not involve any feeback.
In the context of this thread, this distinction is relevant as the
'dynagroove' system was a method which tried to cancel or null one
distortion with another - but this involved *no* feedback mechanism
to compare the two in reality. The LP cutting system had no way to
sense what the LP replay system would actually do when playing the
LP.


Jim

--
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Audio Misc http://www.st-and.demon.co.uk/AudioMisc/index.html
Armstrong Audio http://www.st-and.demon.co.uk/Audio/armstrong.html
Barbirolli Soc. http://www.st-and.demon.co.uk/JBSoc/JBSoc.html