Deluxe Reverb strangeness

[Randall]

Footswitch is not in play. Reverb is disconnected, and tremolo tube is out, because one tridode was shorted when it came in, so I pulled it. Customer has a replacement at home he wants to use.

I have buttoned this back up, I have other work I need to get to, and I have spent over a week with this on my bench. I fixed the complaints it came in with a long time ago. It only does this issue about once in 10 - 15 times of powering up cold, and is corrected by flipping the power switch a time or two. The customer only plays at home, so I guess in this case it is what it is. This is a very weird issue that I must accept I don't know how to solve, and move on.

[g1]

I guess no difference with footswitch connected or not?
Have you eyed up the layout drawing for any wiring dressed out of whack?

[Randall]

I have jiggled and jostled and pounded, it's not mechanical. Touching several HT places makes it reset, not just tube pins. As mentioned even probing B+ feed to OT makes it reset.

[Greg_L]

At first I thought it was always following the oscillation, but I have since stated that a few times I have seen it do the opposite. The latest failure mode did indeed have a collapsed wave, but I have seen the opposite.

Have you tried simply cleaning and maybe re-tensioning the tube sockets? You keep mentioning that touching a probe to tube pins makes things change. Maybe you're jiggling something into being bad and good.

[Randall]

At first I thought it was always following the oscillation, but I have since stated that a few times I have seen it do the opposite. The latest failure mode did indeed have a collapsed wave, but I have seen the opposite.

[g1]

I thought it had yet to be proven that the audio fault always corresponded to the oscillation seen on the scope.
I think that assumption may be preventing or at least slowing down solving the problem.

[Enzo]

Intermittents are always the hardest to solve, by their nature you can't sit there and watch it in failure mode.

[Randall]

This would be a lot easier to find out if it works if the issue weren't so fleeting, as in sometimes days between. Because as has been said maybe the issue doesn't happen because this mitigated it, or maybe just because of the random unknown nature of the failure.

[Chuck H]

We already know the problem mode is off and on - no pun intended. So we just use the amp as you have been, leaving the scope connected. If it never screws up, I am not sure what to conclude, but if it still screws up, at least we get the chance to see what is going on on the scope.

From post #32:

Just a thought... If touching a certain point in the circuit with your meter probe always fixes the problem you could just solder a 10M resistor parallel to a 10p cap from one of those points to ground and then see if you can't re instigate the problem. If you can't then the amp is fixed. Well, sort of. It's a band aid for an oscillation issue, but it might be fine and it would get the amp off your bench.

Hmmmm?

[Enzo]

We already know the problem mode is off and on - no pun intended. So we just use the amp as you have been, leaving the scope connected. If it never screws up, I am not sure what to conclude, but if it still screws up, at least we get the chance to see what is going on on the scope.

[Randall]

Yeah meter or scope. And I have it powered down with both grids with scope probes on them right now. But given how random this is, how will I know what effect this will have unless it goes into fault this way. If it doesn't, what have I learned?

[Enzo]

OK, you turn it on, it acts up, THEN you touch a probe to it, and it gives up. Does it do this with a meter instead of a scope?

In either case, connect the scope when the amp is off, NOW turn it on and let it go. Will it go into fault then? That eliminates the pop signal that seems to trigger things.

[Randall]

So I power it up in fault mode after sitting for a couple of hours, first time in two days. I touch my scope probe to the first grid on V2, and with a pop! it's back to normal.

Seriously, how am I going to ever fix this?

[Randall]

Thanks for sticking with on this one.

The caps I put in (F+T) were not that old. Several months maybe. ESR of the old caps I took out are 1 - 1.5 ohms. I didn't measure the new ones. I brought them up slowly on a variac the first time, as I usually do. Grounds were measured from the cap negs to chassis, all look good. Today I hit every solder joint on the amp, and verified the five chassis grounds coming off the board are solid. I pulled the jacks and cleaned those grounds as well. One thing I did find is the Normal jacks had three toothed washers each, and the Vibrato channel had three and four washers. I put them all down to two each, The internal speaker jack had four washers as well, I put both speaker jacks down to two each and gave everything a nice firm twist after a bit of pencil eraser and a squirt of DeOxit. So one speaker jack sticks out farther than the other, so what? I didn't like those stacked washers.

After that, if it doesn't go into fault again by the end of the day, I am sending it out. If it comes back, it comes back.

[Helmholtz]

I was primarily referring to the "touch sensitivity" of the amp when probing high voltage points. This is something I have seen with self-oscillating amps. It is likely that the circuit point being contacted carries high frequency. That should be verified if possible.
The meter leads mainly cause capacitive effects.

Old style ecaps typically had higher series inductance (ESL) than the better modern construction types which use multiple foil contacting. ESL does not change with age.

While the ESL + wiring inductance may add only a few ohms of impedance at high frequency, things change when the inductance resonates with a capacitive partner. Impedance at the resonant frequency can be quite high.
Connecting meter leads is likely to change the resonant frequency.

Self-oscillation often means that resonance is involved.

THat said I have no clue what might have caused the amp to become unstable.


In any case I would clean and verify all mechanical contacts. If the amp uses a multisection cap, good can to chassis contact is essential.