Not all bad
Ian Bell
using the well loved 6SN7 and it B9A equivalent the 6CG7/6FQ7. Just for
fun, I decided to try out some other tubes in the same circuit to see
what happened. I made no changes to the component values so there's a 1K
in each cathode and a 10K between the two triodes all fed from an
unregulated PSU of about 320V.
The results were surprising. As a reference I first tried some new Ei
6FQ7 tubes. These consistently achieved a 2H of -50dB at 2Khz for a 20V
rms output. 2H was below -80dB and higher ones were immeasurably small.
Then I tried four 6N1P Russian tubes (a sort of ECC88 lookalike) with a
mu of around 35. These generally achieved 2H around -46dB, 3H around
-70dB and higher ones immeasurable. That's not much worse than the
Matsushita 6FQ7s I have tested apart from the much higher 3H. Stage gain
measured around 35 as expected
Then I tried some 6N2P tubes - a 12AX7 lookalike but with a single 6.3V
heater. These achieved slightly better than -46dB 2H on average with 3H
around -72dB with higher ones immeasurable. Stage gain measured
consistently around 80 so either mu is lower than expected or the mu
follower action does not give a high enough effective load (more
likely). However, with a little local NFB these could perform better
than a 6CG7 although they would lack in output drive capability.
Lastly I tried some ECC82 tubes by JJ (12AU7), the types much reviled by
Morgan Jones who measured average 2H distortion at this output level of
just -38dB. All 3 that I tested were much better than Jones' -38dB with
the average 2H being -43dB, but 3H was a disappointing -65dB on average
and 4H was better than -80dB.
In conclusion, the 12AU7 appears to be not as bad as reputed but still
not good enough. The 6N1P is quite good but does not have enough stage
gain to make local NFB worthwhile. The 6N3P performs better than the
6N1P and has an 80 times stage gain which means with local NFB it could
be quite useful.
Cheers
Ian