You have it about 95%. right! I also put in a 20K 1W decoupling resistor between the PA and VA B+, along with the 22 uF/450V cap (see next). If you want possibly more channel separation, use a 39K resistor and a 10-16 uF / 350V (or greater) cap for each channel.
I feed both VAs off the 20K resistor and the 22 uf Cap. This resistor is probably not
absolutely necessary, but I like to follow the "old school" line of design thought, and just put it in. With both VA stages, the current consumption is about 1.2 mA, and the voltage drop about 24 VDC.
Each PA stage draws about 35-36 mA, so the total B+ (231 VDC) current consumption is 71-73 mA, which is 6 to 8 mA above the HV winding's published max ratings, but then I am also drawing about an amp less than the filament winding's max rating, so I figure that the load on the primary still stays well within the ratings. The 269EX trannie gets warm after a few hours' operation, but not overly hot to the touch - a PA774 in an ST35 gets significantly hotter.
I used 4.7K instead of 47K for the pentode grid stopper R7. I also used some different component values for C1 (22 uF / 450V), C2 (47 uF / 10V) C4 (330 uF / 50V MUSE Cap), just because these were what I had on hand. None of these actual values changes (except possibly R7) will have any noticeable effect on the amp's overall operation.
The real "secret" to the improved sound quality is the plate-to-plate NFB effected by adding the 1 Meg resistor R10. the other NFB loop has little or no effect, and could be left out, just eliminate the 150 ohm resistor, and the 4.7K & .001 uF NFB path. I just left it in, because trying to remove it would be very difficult due to the very close spacing and compact structure.
I measured the TF103's primary inductance last year, and I found them to be very close to 10H, from P to B (red to blue) leads. That inductance value definitely helps the LF response. For such puny sized OPTs they sould pretty darn good
I'm sure that full sized Edcors would be a major inprovement, as would any "fancy" OPTs.
In place of the 100K input resistor, a 100K stereo pot can be used, if desired. I just left out the VC because I didn't have enough room under the chassis plate (the wood frame consumes about 3/4 inch from each side, so that left me with approximately 6X5 inch under the chassis plate.
For the next version, I plan to go with PSE, and maybe use some 3K OPTs that I have, these have no UL taps, so I'll be limited to either triode or pentode modes with these. Or I'll order some iron from Edcor and wait until hell freezes over, errrrr, for the 8 weeks for them to be delivered...
I won't bother with any more PP 6BM8 designs, as I've already built up a couple of the K502 11BM8 / 16A8 designs, and I've done all the reasonable mods to get them sounding better. IMHO, the SE sounds nicer than the PP does, though at lower power levels.
Here are the actual voltage measurements (rounded off) from my amp:
B+ = 231 VDC
6BM8:
pin 8: 1.42 VDC
pin 9: 102 VDC
pin 2 16.8 VDC
pin 6: 219 VDC
pin 7: 227 VDC
Pins 1 & 3: 0.0 VDC
Pins 4 & 5: 6.5VAC, 3.25VAC each to ground
PSU:
UF4007 cathodes: 247VDC
100 OHM WW: about 8VDC drop
+ of 40 uF cap: 239 VDC
choke: about 8VDC drop
+ of 180 uF cap: 231 VDC (B+)
filament: 6.5 VAC
/ed B