Well I finally powered up my rev C to test it and neither channel is working. It's difficult to debug. There isn't a full schematic on the site and the PCB info doesn't match the rev of the board that I have.

The only documentation I have to go on is the pinouts for the two chips. Both boards show the same symptoms.

The output on the 3886 is close to the negative rail (about 7V above the -35V negative rail).
V+ and V- are about 70V apart (seems correct)
V+ and V- on the op-amp are about 25V apart (seems correct)

The 2 1k 3watt resistors each have about 24.5V across them (they get really hot, I don't leave the board powered up for long).

The ground pin on the 3886 doesn't seem to be connected to the output ground connector, should it be?

Also, I fried one of the 3886 chips by shorting the output to one of the rails with my voltmeter probe I pulled it out and probed the board without the chip and the output pin is centered between the two rails. So it's the chip itself forcing the output low. Strangely enough, the 3886 doesn't get hot at all.

Any advice on debugging this thing?

P.S. which of the input terminals is signal and which is ground?

IN_GND is the pad closest to the heat sink and the input pad is the pad closest to the input cap(C13) as indicated on the silk screen.

Do you have any pictures you can attach?

What sort of transformer are you using?

Also how is your transformer connected?

Cheers!
Russ

Well, looking at the schematic, the 1k resistors have V-(V+) on one side of them and -12V (12V) on the other side, so I guess it's not incorrect to have 24.5V across them. They do get mighty warm, though.

I'll look at it some more tonight, having the schematic will help a lot.

I poked at it some more tonight.

The relay is disconnecting the output ground.
The output of the 3886 is still close to the negative rail.
The input signal is making it through the lm318 and to the lm3886.

What's controlling that relay?

Could be, or possibly installed backward?

A pair of replacement zeners and a new LM318 (just in case) are on the way.

You should probably also (for kicks) check the resistance of C6 (and C11 for comparison) to be sure C6 is not shorting. I'll send you another one of those as well.

Yes

Well I pulled out ZD1 and tested it, it works fine.
Then I pulled out C11.
There's still a bunch of voltage across R1 (like 30V).

So it looks like -12V is shorted to ground somewhere.

Yep, pulled the lm318 and now there's no voltage across R1.

So the LM318 was pulling lots of current.

I wonder if that's because the chip itself is bad, or because the inputs are messed up and driving the output to one extreme.

Sorry, I'm mixing results from the two channels here. On this channel the -12V rail was measuring close to 0 and the 12V rail was measuring 4V.

So I pulled ZD2 and the -12V rail was still close to 0.
So I pulled the capacitor and the rail was still 0.
Then I pulled out the LM318 and it measured -35V as expected.

So the LM318 was sucking enough current to pull that rail down to 0, which takes 35 mA. Assuming something similar is going on with the +12V rail (forgot to measure it with the chip out), it's pulling 31mA off of that rail.