[Patrick Turner]

While surfing around the web for
any decently measured THD or IMD measurements
carried out on pentode tubes, I came up wuth very little of interest
that at least I didn't know a lot about already.

There is more mention of triode THD and some about IMD.

Some of what i found was up to the standard of the 1950s
RDH4 examinations of THD and IMD, and basically nobody
has provided the well set out analysis
of a typical 6AU6 including the schematic,
and with levels of THD products from 2H to 10H
as they rise in level for the progressive rise of fundemental F.

So we really only know that THD and IMD exist in devices, and that
it all rises with output levels and load values.

But while looking around Google under 'pentode distortion measurements'
I came across a thesis drafted up by one Mr Daniel H Cheever, posted at

http://web.mit.edu/cheever/www/cheever_thesis.pdf

Here he talks about the increase in harmonic trash within a signal
caused by NFB.
You all thought NFB reduced the trash, but he has other ideas supported
by
his calculations and observations fairly well
presented.

But his conclusions must be considered in the context of his ideas
presented about the Total Aural Disconsonance figure of merit, or TAD.

Basically, he seems to be saying the sound we hear is distorted by the
ear which is a very non
linear microphone, but the brain filters the harmonics out, but where
an amp produces harmonics, ( or speakers, microphone, or other device
)then
unless the relationship of the harmonics relate
to what a brain does with harmonics, it affects the sound far more than
we think using accepted methods,
and an SET amp can thus sound a heck of a lot better than a high NFB amp
using bjts.

He has a lot of mathematics to proove his point, or tend to proove it at
least.

He cites samples of new tube amplifiers costing many thousands made by
CJ and others
where the use of loop NFB has been reduced drastically.

So what do people think about My Cheever's thoughts?

Patrick Turner.