Reverb oscillation in SF Pro Reverb (Master Volume)

[waspclothes]

Hey guys,
I have a SF Pro Reverb (the really complicated one with the master volume and pull-boost) that's making an oscillating sound as the reverb is turned up past 4.

Here's the schematic since it was a bit hard to find:



Here's my trouble-shooting so far:

Reverb driver tube (V3) pulled = still oscillates
Tank disconnected = still oscillates
Reverb recovery tube (V4) pulled = oscillation stops
Reverb recovery tube (V4) replaced with new tube = still oscillates

Reverb recovery tube (V4) grid (pin 2) grounded = still oscillates
Reverb recovery tube (V4) plate (pin 3) grounded (After .003 coupling cap) = oscillation stops
Reverb recovery tube (V4) cathode (pin 1) grounded = oscillation stops

V4 voltages:
pin 1 = 2.0V (2.3 expected)
pin 3 = 232V (225 expected)


So I guess where I'm confused is.. If the oscillation is coming out of V4, why didn't grounding the grid on it stop the oscillation?

I gotta say, this is a pretty ugly amp.. lots of wires, badly dressed. But I poked around quite a bit and adjusting any of the wiring around V4, and the wiring leading to the reverb knob itself didn't yield any reduction in the oscillation.

So what I mean to say is.. I'm totally stumped. Has anyone seen this before, or have any suggestions?

[waspclothes]

I should also add that the amp appears all original and the electrolytic capacitors are dated 1978. The amp also has no rectifier tube, just a bridge rectifier inside instead. But other than that, the schematic looks the same as the one posted.

[Jecarroll7]

What do your supply voltages look like? Are you seeing any AC ripple on your filter caps?

[Chuck H]

Coincidence? Two microphonic tubes? The one that was in it and the one you replaced it with. Other than that my only suggestion is that your "filter" caps aren't decoupling. You can check this by tacking a good 400+ volt cap of 10uf or more from the top of any preamp plate load to ground (NOT THE REVERB DRIVER OR PI). If that stops the oscillation, replace all the filters.

[waspclothes]

Success! It ended up being the capacitor at the cathode of V4. Whowouldathunk?

[Jecarroll7]

The bypass cap? Not the filter (off the plate?) Interesting.

[Chuck H]

The bypass cap? Not the filter (off the plate?) Interesting.

"Interesting"... For me in this case that's an alternative to WTF?!?! If it were shorted (cathode grounded) the tube would be hot, but I don't see how that should result in oscillation above a setting of four on the reverb pot. If it were open (no bypass) the recovery and mixer stages could interact on the cathode and overall gain would be reduced by about 3dB. The interaction shouldn't oscillate because the reverb signal is sort of phase ambiguous. Or not... I've noticed that NEW reverb pans seem to have more representation of the original signal at their output than the vintage pans did. I have to wonder if that amp has a replacement reverb pan then. But even then it wouldn't be the first (or second, or third) thing I would have suspected.

Kudos to waspclothes for finding the problem. Tenacity is all we have left when reason fails. And to give up is an admission of defeat!

[Enzo]

The interaction shouldn't oscillate because the reverb signal is sort of phase ambiguous.

The reverb signal is ambiguous, yes, but this is not the reverb signal, it is the reverb recovery amp feeding back. Remember, this happens with the pan unplugged - no reverb signal.

The plate signal of the reverb triode is fed to the grid of the other triode, which shares the cathode of the first triode. Without the bypass cap, I could see a sort of phase shift oscillator occurring. At least with only brief analysis.

[waspclothes]

The reverb signal is ambiguous, yes, but this is not the reverb signal, it is the reverb recovery amp feeding back. Remember, this happens with the pan unplugged - no reverb signal.

The plate signal of the reverb triode is fed to the grid of the other triode, which shares the cathode of the first triode. Without the bypass cap, I could see a sort of phase shift oscillator occurring. At least with only brief analysis.

I had to get the scope out to figure out this one. Coming off of the reverb recovery stage was the input signal, but maybe 10Vp-p (if I remember) and close to 180 degrees out of phase with the input signal.

When I put another 25uF across the existing cathode bypass cap on a whim, that signal dropped big time.

[g1]

Not that it matters now, but you said it had the SS rectifier, perhaps it is the 70W ultra-linear version? (screens will be supplied by taps on OT).

[waspclothes]

Not that it matters now, but you said it had the SS rectifier, perhaps it is the 70W ultra-linear version? (screens will be supplied by taps on OT).

Cheers, this is the proper schematic, thank you.

[Chuck H]

I had to get the scope out to figure out this one. Coming off of the reverb recovery stage was the input signal, but maybe 10Vp-p (if I remember) and close to 180 degrees out of phase with the input signal.

When I put another 25uF across the existing cathode bypass cap on a whim, that signal dropped big time.

That doesn't explain why the oscillation was still present with the grid of that stage grounded though.?. I think Enzo has it. With the cathode coupled there is a sort of oscillator circuit formed. The cathode bypass cap is necessary for decoupling as well as the added gain it affects on the recovery stage.

[TimmyP1955]

I had the same issue on an amp. Removed the cathode cap on V4 and it got much worse. Installed a new cathode cap and all was well.