Rockford Fosgate 4600X series one power supply

I’ve got a Rockford 4600 that blew an output FET. I pulled the two FETs on that channel, and poking around I discovered that the main rail + caps are at +18V which is probably about right as the caps are 25v rated. But the - rail is at -27V. Any ideas where to look? Can’t find any bad components. I even replaced the TL494 and LM339M chips, no change. No 'scope to check waveforms yet, but I have one coming...
 
Although it's unlikely, it's possible that the positive rail cap is defective and allowing a lot of ripple to be on the rail and the meter is reading that.

Do you have another capacitor that you could parallel across that capacitor. It doesn't have to be identical. Anything 100uF or greater will likely be OK.
 
One FET was blown, so I pulled that pair and the driver to test. Now I can't remember why I pilled the remaining two pairs. One pair is still in.
(this is a four channel amp, there is one pair per channel) I have replacements for the one channel, do you think I should re install them all?

I checked all of the empty pads with a multimeter, no shorts.

I have been powering it up with a current limited power supply set to 10V, and the amp is drawing 0.5A, nothing getting hot. I tried stepping it up to 12V, and the rails went up to +34V (definitely too high, as the rail caps are 25V) and -22V, current stayed at 0.5A. The power LED is coming on through all of this when I apply voltage to the remote terminal.
 
Does the channel that still has FETs produce audio. If you have to check for audio, you need to clamp all heatsink mounted components to the heatsink.

I don't know that anyone has tried to measure this voltage and would not have done it myself. It's possible that voltage is leaking from other components and this is normal.
 
OK, I re installed the driver transistor on channel 4 and that seemed to calm things down, voltage and amp draw are stable. Who knew pulling a component would create such trouble. Well I'm learning, and hopefully no harm done.
Turned the bias down in all channels. Channel 2 and 3 I am getting good audio. Channels 1 and 4 are working but badly distorted. I played with the bias a bit, it didn't seem to make any difference with the distortion. Channel 4 was the original one that had a bad FET.