Classic 50 Neg V woes

[Smokebreak3]

What a nightmare dissasembly
Repaired one of these last week, came in with burnt R61, R62, which are (2) 22r is parallel in the neg voltage supply.
Also, U1 , the 4558 , was cracked, which is fed by the neg voltage. lucky it is socketed.
Repair seemed straightforward, I replaced the 4558 and 22r and the amp worked great for 10 hours, then came back to me.
The 22r resistors ( I replaced w a single 1W 10r) burnt up again, and the 4558 is swollen in the middle and toasted.
So I'll replace those again but wondering how far down the rabbit hole I should go with the repair. I figure I'll replace C42, C43, C49.
Customer says it happened right after he cranked it.
https://www.thetubestore.com/lib/thetub ... aAs-3VwND5

[R.G.]

Replace the diodes CR11, CR12 too. If one of these went shorted or very leaky. This could pump AC into the negative power supply, with similar results to what you're reporting. 1N1003, 4004, 4005, 4006, 4007 would all work there. They all differ only in their reverse voltage rating.

[Stevem]

Yup do the caps and look close for any signs of arcing on the output tube sockets.

[Smokebreak3]

Cool thanks guys

[nuke]

This amp has a couple of weirdnesses going on with the -27v supply.

Preamp tubes V1 and V2, have their filaments wired in series, and are powered by the -27V DC. If a tube

The 12AX7 V3 and the EL84 filaments are powered by the 6.3VAC power, that also powers the pilot lamp.

The schematic is a hot mess and hard to read, with jumps and labels that you have to hunt for.

The -27V runs the 4558 for reverb, the -14v is a mid-voltage reference for the opamp from the reverb, and it biases the effects loop buffer Q2, which is mis-drawn as that is actually a darlington, rated for 30V E-C.

Hunting around, "RAWV-" and "PREV-" are just alternate names for -27V.

If the -14v gets compromised somehow, then the opamp that drives the reverb is going to misbehave, and the only low-z DC thing there is the reverb drive coil, which maybe be zapping the 4558. The opamp is rated for +/-15v or 30v on a single supply as in this circuit. The opamp and the MPSA63 are both really close to their voltage ratings.

The other place -27v goes is the channel relay, and it might be worth checking the relay coil and the damping diode across the relay coil. If the diode has opened, then the relay coil's voltage kick when it is operated might be enough to pop the 4558 once in a while if it catches an unfortunate voltage spike.

And there's the problem of "fake" 4558's running around sometimes, depending on where you get your parts from. Could be a pin compatible, but lesser chip that isn't happy on -27v.

Could maybe disconnect the windings to the rectifiers and use a bench power supply to substitute the -27v and run it a while, see what the current is doing.

C50 on the effect loop board could short and take down the -27v. But it wouldn't pop the 4558.

Any other visible damage besides the resistor and the opamp?