Alright... I'm stumped.

[Wakarusa]

Have a mid 70's twin reverb on the bench that's blowing fuses.
With power tubes removed the fuse won't blow.
With power tubes installed, the amp runs flawlessly for around 30 minutes and then, with no warning, blows the fuse.
The transformers and choke are cool to the touch when this happens
The transformers do not have shorted windings (hot or cold)
Yes, of course I've tried different tubes :)
The amp was recapped previously, all caps test good
It does it sitting on the bench (so not as much heat buildup as you get in the cabinet and no vibration)
None of the HV leads are in contact with a grounded element, none of the insulation is damaged.
There don't appear to be any carbon traces on the power tube sockets inside or out.So I'm stumped. I'm about to either rebuild the entire power section, or pull the main board to see if there's something funny underneath, but the customer's on a budget, so pot luck isn't a good thing here. So save the guy a few bucks... what in the heck am I missing here?

[Tommy_G]

Maybe its something external to the power amp...like a mismatched transformer, faulty speaker jack or cable, intermittently faulty speaker, etc.

[VaughnC]

It seems that the current draw of the output tubes is causing something to break down. Current traveling through any conductor creates a corresponding increase in the conductors temperature which, in turn, causes the conductor to expand. My best guess would be that one of the output transformers windings is expanding just enough as it heats (via the current flowing through it) to cause intermittent primary to secondary or winding to case leakage or short.

[Tommy_G]

Sounds like a good theory to me.

[phsyconoodler]

Bias supply problems?Did you put a bias meter on the output and watch it til the fuse blew?
I've had funky diodes in bias supply's blow fuses before.They get warm and poof!
Especially if they are those little round one's fender liked to use in '70's amps.
What happens to the current draw when the bias supply goes more positive?

[brad347]

I had an amp that acted like that once and it turned out to be a rectifier that was on the fritz. This particular amp was a tube rectifier and I know the Twin has a silicon rectifier, but diodes are cheap... maybe it'd be worth a stab in the dark?

[Groovey Records]

:munch

[SatelliteAmps]

You could use a heat gun to carefully heat different areas to try and force the fuse to pop. That would tell you if it is something overheating, and should narrow down the area.

[Wakarusa]

First, thanks to all who took the time to take a stab at it.
Various responses:
The externals shouldn't be in question. Running it through my own speaker wire to my own 4x12 test cab. Supplied by a regulated AC power supply on the bench, so I shouldn't have to sweat wall transients etc.
I'm with you on heating in the OT, but if that's the case it'll be near impossible to prove. As I said, the transformer itself is cool to the touch. I've also put my transformer short test rig on it immediately after the fuse blows and it seems to read fine.
Some day when I'm feeling flush I'll upgrade my Fluke 87 to a couple of 187s or 189s (whichever does the recording trick). I've had bias supplies go flaky before and would normally expect the tubes to red plate before blowing the fuse, usually with some sort of audible accompaniment. Doesn't seem to be the case here, but I can look again. I will say that the bias supply filter cap has been replaced and the work is a bit suspect. Easy and cheap to replace, so we'll give that a go.
Been using a hot air pencil in lieu of a heat gun for thermal testing (I like that you can concentrate the heat on individual parts with the HAP). Checked most of the stuff around the power tubes (leads, resistors, etc.) and power supply, but don't think I hit the rectifier(s). Agreed that this is a trivial thing to replace and doesn't run the bill up, so I'll swap these out and see what we get.Short answer is that I agree with VaugnC -- if the indications aren't somehow misleading me, the problem has to be one of heat from the power tubes, the power tube sockets themselves, something that doesn't like the additional current draw of the power amp, or the power amp bias supply.

Anyway, I'll swap out some diodes and the suspect bias filter cap and report back.

[alltone]

Hi Todd..If your twin is series #AA769 or #270 you might check the 100 ohm filament resistors near the pilot light.With the addition of the output tubes in circuit the filament draw may be loading the P.T. down if one of the 100 ohmer's are bad??Are the tube sockets and bases clean..no signs of ANY carbon arching. Sounds like you know what you're doin'..Just sittin' here with my morning coffee..thinking outside the box or not awake yet ..or both! Cheers, Doug

[alltone]

http://s303.photobucket.com/albums/nn150/douglasamps/?action=view&current=ACSeriesLoadTester.jpg

Do you have one of these on your bench..:crazy

[Nolatone Ampworks]

Like guys have said, heat related seems most likely considering the time delay before fuse blows. Cool transformers seems to rule that out. Power AND output, right?

Are you passing a signal through the amp during the 30 minutes before fuse blows?

I'm voting rectifier.

[RedMan]

If it's doing it consistently after 30 minutes from cold it could be heat related, but 30 minutes is a long time to heat up. I would check to make sure that one of the lugs from the pilot lamp isn't too close to one of the HT wires. I've seen it happen more than once. It will take a while to blow the fuse too since it may not be arcing/shorting all the time.

[Wakarusa]

Filaments (and pilot lamp) leads are all well clear of HV.

Later SF, so has the hum balance pot instead of the pair of 100ohm resistors.

Don't need the load tester since I've got a fancy AC power rig anyhow. The transformer tester, however, is scratch built and uses the same idea ;)

Anyhow, rebuilt the bias supply, replaced the main rectifier diodes. On the bench now with power tubes out. I'll give that a half hour or so to make sure nothing pops then install the power tubes. FWIW, the bias supply rebuild was a good tip in any case. I got lazy/tired and didn't look that closely -- somewhere someone tried to blackface it and just got it wrong. It's correct now, so maybe was the original problem. We'll know in an hour or so.

Thanks again to all. I spend a fair amount of time here dishing out tips and advice (even correct on occasion!) so it's really great to get some love back. So let's all cross our fingers that I can get this stinkin' thing out of here today :)

[Wakarusa]

Update:

The good news is that it looks like replacing the rectifier and bias circuits has solved the fuse blowing (I suspect it was the bias circuit, but who knows for sure?). The bad news is that the reverb seems to have died somewhere along the way (no idea if it was working when it came in tho'), but that won't be any trouble to fix.

Thanks again for the suggestions. I was definitely at a dead end.

[brad347]

good work!

[RedMan]

Congrats! I hate it when I get stumped on something. Never would have guessed bias diode. I assumed the bias caps were good since a cap job was done. If it was a B+ diode I would think it would happen without the tubes loaded too, but could be what it was.

[TheAmpNerd]

Todd,

I've found it is always helpful when stumped to post something here.
When we are so close to something at times we are too close to it.