by EWBrown » Fri Jan 22, 2010 6:33 pm
Two "quick and dirty" tests to try:
Since the LTPI plates are clipping, and forming those "teats" on the peaks of the sine wave, the tube is getting "starved" for a proper plate voltage at higher drive levels. This is being caused by the 40V approx voltage drop in the PSU, at higher power levels. 191V on the plates seems somewhat low, and 150V on them at full drive is far too low.
Increase the value of the PSU's input cap (before the choke) it isn't a "sin" if you somewhat exceed 500VDC, try kicking it up to (no more than) 525VDC at idle (if your electrolytic caps can handle the voltage) The drop should then be less at full power vs idle conditions, Before you crank up the B+, turn the bias down on the output tubes, as the current will also rise as, the B+ rises.
The plate dissipation is the plate to cathode voltage (V P-K) multiplied by the cathode cathode current (IK), aim for about 80-85% of the max safe PD ratings.
The other Q&D test would be to parallel the 6SN7's 68K plate resistors with 220K, this would bring the effective plate load down to around 51.9K
This will increase the 6SN7's plate voltages by another 45VDC or so, and should not affect the stage's gain significantly. The CCS will keep the plate currents the same as they were with 68K. This alone may eliminate the clipping / distortion on the LTPI outputs.
I'm going strictly on memory here, from when we were working on this design last year., so I could be "off" on some of my "presumed" existing resistor values
/ed
Real Radios Glow in the Dark