Reverb electrical feedback

[Michael Allen]

Working on a single ended amp that I've running a one tube reverb in. driving a MOD 3 spring, long delay tank. 190 ohm input.

I am having a terrible feedback problem with the reverb tank that I can't seem to solve. If I have the reverb and treble at noon or higher (so if one is at 12 and the other is at 1) feedback slowly builds up at around 370Hz I think. I can't get it to go away! It's not acoustical, if I have the master all the way down and turn the reverb or treble up, the system starts feeding back and then I can turn the master up and hear it ringing.

I've tried a cap to ground on the grid of the driving tube. Higher caps to ground on reverb return. Plate snubber on recovery stage. The only way I was able to max out both controls without ringing is by adding a .03uF cap from the reverb wiper to ground. But that pretty much kills all reverb since I'm passing so much signal to ground.

The bummer is since it's a frequency that's not really high, I can't get away with simple filtering. I think it may be the tank. It happens with preamp vol down and master down, so it's not input signal related or acoustical feedback. I've tried putting tape on the tank back, wrapping in bubble wrap, putting tape on the transducers/spring connections. I think maybe the springs are just so sloppy that it doesn't take much to get them moving. Any ideas or thoughts on what to try?

Also, not on the schematic is a tremolo circuit with a 12ax7/IRF20 follower. Pot values are 1M vol, 1M Treb, 1M Bass, 1M MV.



Yeah i know, its a push pull tranny in single ended operation with no current sink or anything on the other side of the primary. But it's working and sounds great, and the OT doesn't get hot at all...

[tboy]

You could try reversing the polarity of the wet signal, like by swapping the drive transformer's output leads. It might help.

[Michael Allen]

Thanks, tried that, no difference.

[Enzo]

OK, first things first. Pull the reverb pan out of the cab, and get it as far from the speaker and the rest of the circuit as possible on its cables. Does distance help at all? Whether we think it is acoustic or not, we still must determine if proximity to anything matters

Your attachment doesn;t come up.

[tubeswell]

I am having a terrible feedback problem with the reverb tank that I can't seem to solve. If I have the reverb and treble at noon or higher (so if one is at 12 and the other is at 1) feedback slowly builds up at around 370Hz I think. I can't get it to go away! It's not acoustical, .

Are you sure? Did you put your finger on the back of the pan when it was feeding back?

[Michael Allen]

Enzo and tubeswell,

Right now the chassis is just on my desk, 1x12 cab on the ground, reverb pan a few feet away on the carpet. Starting with the amp off, I can turn preamp and master to 0, reverb and treble to 10, turn the amp on and after the tubes have warmed up I can turn the master up and the circuit will be feeding back. It doesn't seem to require any interaction with the speaker or volume level at all.

I"ll try again with the schematic. I've used this reverb before as many others have and it works great. Never used this tank before though. I'll try upping the 100k mix resistor and see if it helps without losing too much reverb.

[tubeswell]

So with everything else up, but the reverb level to 0 - no feedback?

[pdf64]

I think you may be getting a feedback loop back through the 470k - 100k mix resistors. A virtual earth mix would fix it, but lose a lot of gain.
As an after thought, could put a FET buffer before the 100k dry mix resistor to greatly reduce the feedback.
How about the standard AB763 value mix resistors? 3M3 dry, 220k to ground, 470k from reverb?

[loudthud]

I agree with pdf64. The feedback path goes backwards through the 100K and feeds back to the drive tube depending on the position of the treble pot. Increase the 100K or pick the drive signal off before the tone stack.

[Michael Allen]

Ultimately there is always a feedback loop through the mix resistor. I've used this before and it's fine, so i'm a little boggled why it's not working here, but I'll try and change the mix resistor and/or move the reverb circuit tonight. Thanks guys,

[pdf64]

Previously was the 100k dry mix resistor being supplied direct from a common cathode stage, or via a tone stack?
The tone stack will have a higher source impedance, which will result in there being more positive feedback coming back around.

[Michael Allen]

hmm, interesting. I'm pretty sure i never used it following a tone stack, just between stages. I think you're probably right on, i'll check this out in the next few days. Thanks!

[pdf64]

My thinking is that the source impedance feeding the dry mix will form a potential divider (dry mix - source impedance) that will attenuate the positive feedback coming back around.
The source impedance may be <40k for a 12AX7 common cathode stage, but possibly >500k from the tone stack, depending on the tone settings, and you've identified that the treble control setting affects the oscillation.
Given that, loudthud's idea of feeding the reverb from before the tone stack may be the best solution.
Pete.

[Michael Allen]

Yep, putting the reverb before the tone stack worked. Still used a 100k mix resistor. Now it's 2nd stage plate to reverb send and 100k reverb mix into tone stack

[pdf64]

Great, thanks for posting the result.