Amp runs for 30 seconds, then fades out

[pine]

I have a 5E3 clone that has worked well for years, but now, I power it up, it plays for 30 seconds or so, then it gets fainter and fainter until there is no sound. I thought it might be a loose connection that was interrupting the circuit when it got hot enough, so I re-flowed every joint I could find, but of course the problem is still there. Now I'm thinking it is a component failure. Has anyone seen these symptoms before, or know what the problem might be?

[markr14850]

Maybe a bad grid-cathode reference resistor? The tube will work until it builds up enough charge on the grid that it shuts that stage off.

[Structo]

Yeah, first rule out the tubes for sure.

Usually something like that is heat related, in other words, after a while when the amp heats up the symptom happens.

Even though you checked, many times it is a bad solder joint or a cracked joint or contact that spreads when it gets hot.

Have you tried chop sticking around when the amp fades out? (be careful here)

Does your amp use carbon comp resistors?
Sometimes those can go way up in resistance after getting hot (possibly cracked resistor).

You can make a listening probe hooked up to another amp and trace the signal through the amp. (be sure to use a blocking capacitor on the probe to prevent DC from going into probe).

Good luck

[tubeswell]

Electrolytic caps can fail with too much heat. When the sound fades out, what are the DC voltages on the power rail?

[pine]

Thanks, everyone, I'll get some voltages and report back. There are some carbon comp resistors in there as well.