"Scott Dorsey" wrote in message
...

I think Mr. Stamler is exaggerating a little bit about a load that
is a factor of ten higher than the effective output Z being a problem.
Note that when the effective output Z is calculated, the gain of the
tube is taken into account... the more gain you have, the more the
feedback is helping you.

Except that into, say, a 6k load, the tube has a lot less gain to begin
with, so a lot less feedback when used as a cathode follower.

I did some tests a couple of decades ago to illustrate the point. The tubes
weren't completely open-loop, since they had unbypassed cathode resistors,
but still. For a 12AU7 with a 100k plate resistor and a 1.5k cathode
resistor, idling at about 2.2mA:

Load / Gain / THD at +15dBu

240k / 10.0 / 0.22
92k / 8.7 / 0.26
48k / 7.2 / 0.32
20k / 5.0 / 0.52
9.6k / 2.9 / 0.90

Note that whereas at a 240k load there's 10x gain, so the chance to reduce
distortion by that amount, at a 9.6k load, there's only 2.9x gain, so you
can only reduce distortion by that much when you turn the tube into a
cathode follower. Hence my suggestion that the 12AU7 be used with an A/D
converter of -10dBV nominal sensitivity.

For a 6SN7 with a 27k plate resistor and 475 ohm cathode resistor, running
at about 5.4mA idle current, the numbers a

240k / 13.0 / 0.072
92k / 12.0 / 0.082
48k / 10.0 / 0.10
20k / 8.0 / 0.18
9.6k / 5.6 / 0.27

Note that at 9.6k you have about twice the gain and less than 1/3 the THD,
so a 6SN7 should be about 6x cleaner as a cathode follower under those
conditions. It's a nice tube, although a power hog. But it's still happier
at higher load impedances; into 240k you have the possibility of around
..006% THD, and in fact that's what I've measured at that level into that
impedance.

By the way -- yes, I know I've written that I don't think THD is a good
measure of audio quality, and I stand by that. But I did these tests back in
the 1980s, when I was young and foolish -- and besides, the same pattern of
behavior will show up in IM tests.

Peace,
Paul