>"What the heck are 82kohm resistors R1 and R2 doing in the power amp circuit ?
Even if they are 1% precision types, they can only creates a DC offset that is not desirable."
If the open loop gain is high enough it doesn't matter. Plus they got R4 in there which should bias the thing.
Possibly it is a method to keep DC offset from happening with highly non-sinosodial waves with extreme second harmonic distortion (such as the human voice) so that when it loaded one side of the power supply the imbalance would cancel out.
Revox is a weird company like that. I got the print of an FM tuner I can't even follow. They got more transistors in there than, well, anything. Ceramic filters all over the place n shit. Seriously ridiculous. And I mean to the point where I consider it severely overdesigned. On FM they transmit up to 15 KHz, it is not worth ten grand to receive. But this thing by golly can receive, I am sure of that. I'll have to see if I can find the model number, the print will make you drink. I can imagine it on paper, each of a ton of pages would be bigger than almost any desk, and some rooms.
I am not saying their stuff is not good, I just think they go too far sometimes.
You need two supplies for an amp, a low and a high. If you want, have a boosted supply like Pioneer et alii did years ago to run the drivers and smooth out the linearity when you got to the rails. That's three. This amp does not do that.
The only other thing those 82Ks might do it stabilise it during power up and power down. Remember the discussion about the QSC circuit with no center tap on the transformer ? Well they have to manage that DC from power up till power down. But this amp is not like that.
Really, those 82Ks are working into an 18K, so just how much can they do ?
Weird company, engineers probably get better drugs than anyone.
My best guess at the moment.