Hi folks,
Tonight I worked on a circuit for the SCA35 that eliminated the need for gNFB.
The SCA is such a simple amp to begin with, so I decided to keep it simple. Working on a suggestion by DeathRex here: http://www.diytube.com/phpBB2/viewtopic.php?t=4760 , I took the FB off the output tube cathode resistors.
An interesting effect that I did not expect was the constant F3, which was 11Hz to 29KHz clean across the output power... 100mW to full out.
Here's the circuit:
Bias on the output tubes were best with 560R, but performance the best with 470. Hence the compromise in the output cathodes. Another advantage of this bias method is matched tubes aren't required.
The driver is a single 12AX7 in paraphase. The trimmer sets the symmetry for a wide variety of tubes of various health I tested. With the tweak on symmetry, you can gain an extra couple of watts before clipping over a single 10K or so resistor there (which may be fine for some, but not other 12AX7's I had in the bin).
Next step is to build a "speaker impedance simulator" to look for instability at resonances. I used a pretty inductive 8R2 load and had no oscillations or upper spectrum resonances.
Cheers!