Built a reverb tank drive/recovery circuit - dry sound leakage problem


2016-01-04 10:50 pm

I recently built this reverb tank drive and recovery circuit:


The only modification I would like to make to it is a switch to make the dry signal open circuit for a 100% wet sound, as this can't be achieved with just the mix knob. This is so the reverb unit can be used as a stand alone effects unit for music production as well as for live instruments.

When I make the direct dry signal open circuit (around R14) there still seems to be some minor dry sound leakage coming through (mostly high frequency from what I can tell).

Any ideas how to rectify this?

Thanks :)
Last edited:
If removing R14 does not kill the dry signal, then you have crosstalk from somewhere. Disconnect the reverb pan, and run a signal through. Do you now still hear dry signal?

R12 is the power supply feed, not sure how it got involved. If you were messing with it instead of R14, that may be the issue.

R15 becomes superfluous.

If you do have crosstalk, it might be the recovery side picking up the drive side, or it might be poor grounding layout or even a lack of decoupling. I might have had the ground return for the two drive transistors run separately to the supply ground from the returns of the two recovery side. A sort of two pronged star ground.

Make sure C6 is healthy, in fact I might add a second R12 and C6 just for the drive side, and thus separate the drive and recovery power feeds for decoupling better.


2016-01-04 10:50 pm
Hi Enzo,

You're right, I meant to say R14, not R12. I actually omitted R12 completely as I'm using a 30V supply, not a higher voltage amplifier supply rail as the circuit is intended to be used with. I will check with the reverb tank disconnected tomorrow when I turn my living room into a workshop again (everything is put away at the moment).

I should've mentioned in my first post, my knowledge of electronics is pretty basic, although this project alone is teaching me a lot/refreshing information I learnt as a teen studying electronics.

What is the function of C6 here?

I'll report back again tomorrow.
Again, still guessing. If loading the drive circuit with the pan input coil does it, I am led to believe either the wiring is radiating, or we have a grounding issue or power decoupling issue.Try a different cable from the drive, and move the pan away from the recovery circuits as much as possible. In fact, does moving the pan around affect the amount of crosstalk?

I earlier suggested dividing the power supply into two branches. That should be easy to tack together and see if it helps. Likewise separating the ground returns from the two halves. We need to know if this is coming from ground/supply or from signal radiating.

Oh, another thought on the radiating side. Reverb pans have the two jacks. On some pans, both jacks are grounded to the pan chassis. On others, one or both jacks are insulated from the pan chassis. Probably most common is the drive jack is insulated while the output jack is not. Depending on the pan you are using, is the input jack isolated And if so, try grounding that jack to the pan wall. That would make the pan a shield.
Try increasing the size of your filter cap C6. When you connect the input to the reverb it alls the transistor to pull or supply current to the input coil.

As Enzo said add 2 R12's into the circuit, and another C6. One for the drive side and one for the output side. If it still does it, then at least you know it can't be getting coupled through the power supply rail.

Can you take a picture of your circuit layout? I doubt it could be coupling at audio frequencies, but anything is possible if things are close enough.


2016-01-04 10:50 pm
Both the input and output of the tank are grounded (checked for continuity to pan).

As suggested, I separated the drive and recovery grounds and sent them both directly to supply ground; no difference.

Then separated power supply to drive and recovery, adding another 100uF cap to drive supply (existing cam remained on recovery side supply); also no difference :(

I don't have a larger sized capacitor to try in place of C6 but I tried replacing C6, then tried the two capacitors in parallel to increase capacitance - no difference.

I also tried a different cable to the tank (no difference). Moving the tank closer and further from the circuit also made no change to the leaked sound.


  • WP_20160210_19_08_12_Pro.jpg
    76.9 KB · Views: 63
Last edited:


2016-01-04 10:50 pm

I'm not sure I understand what you mean by having the drive cable plugged in without the pan. I have the input and output cables soldered directly to the circuit board, with an RCA plug on the end of each cable. I tried no cables connected to the tank, just the input connected and just the output connected.

*update*, while the leaked sound can be heard noticeably (although faintly) with the tank input connected, it can also be heard extremely faintly with nothing connected to the tank, with the volume up high and being close to the speaker. This level of leakage would be acceptable if it didn't get louder once the input cable is connected.

I have another tank here, different model and impedance but the leakage still occurs when the input cable is connected to it.

I have used both tanks with a little sealed amp I have and a separate preamp as recovery, and both produce a pure reverberated sound so I'm sure the issue lies within the circuit.

Thanks for all your help so far :)
I am used to amps that not only have the jacks on the pan, but also have them on the amp. So the cables are completely removable. SO yours are soldered to the board, thus they are permanently "plugged in" so to speak.

One last thought, tack a short across the return cable right at the board, not out at the plug end. Does THAT kill the crosstalk?

All my suggestions so far have been to find WHERE the signal is sneaking in. Once we know that, we can cure it.
OK, that means the recovery circuit is picking up the drive side signal.

Unshort it back to original, and now unsolder the return cable from the board. Now does it crosstalk or not. We are testing the cable as possible pickup agent. If it no longer picks it up with cable removed, then the cable was involved. If it still picks up without the cable, then the return circuit is sensitive somehow.


2016-01-04 10:50 pm
OK, with the return cable removed completely, the crosstalk is still present.

I have looked over the underside of the board very carefully and can't see any physical reason to cause it (stray solder etc.).

I also divided the input into two - one directly to the reverb and one only to the dry signal. I then left the dry disconnected/open at the very input of the board in case that signal line is where the interference was being picked up from.

I have still not solved it :confused:
It won't be anything like a shorted trace, your return input is sensitive enough to pick up the signal from the drive side. I think your solution will involve preventing the driver from radiating much.

Just a thought. Cut a piece of steel tin can and flatten it. CLip wire it to circuit ground. Now cover it with tape or something so it doesn't short circuit anything, and lay it on top of the recovery circuit. Does this shield help at all?

I suppose you could move it over and try shielding the drive circuit to see if that appraoch helps.


2016-01-04 10:50 pm

I have here a painted steel square, with a threaded hole in the corner for a screw (seems ideal for the test you suggested). I clamped a wire to the steel with the screw (continuity to the steel checked out) and attached it to supply ground and then to circuit ground, tried the steel plate in different positions on the top side and under side (worth a shot, I suppose) of the circuit and it seemed to have no effect.

I think I'm going to physically chop the board in half and make a few modifications (addition C6 and power terminal) and create more space between them.

On a side note, as you are knowledgeable of reverb maybe you can give me your thoughts; I am keen on adding a feedback knob for some CONTROLLED ringing of the tank (I'm aware this might not be desirable to most, but for experimental music it could be interesting). I know the Doepfer A-199 spring reverb has this feature and seems to work well.

I tried a couple of methods with scary results. First I took the output from the recovery fed it into the tank input via a 250K log pot and a kill switch. This gave me a kind of distorted feedback beyond a certain point, then at one point it went into a loud screech which I could only stop by unplugging the power.

Then I tried the same thing but with the feedback going back into the drive input. The result was frightening so I killed the power immediately.

The pot was wired like the 1M pot in this circuit:


An idea of what I'd like to achieve with the feedback control can be heard here:

Thanks again!
Last edited:
If you don't want to listen to the process, you can mount it all in an insulated cab, but instead of hard wiring feedback, I suspect you're more likely to like acoustic feedback. Have the recovery end driving a speaker as well as going off to wherever you want the signal. Have that speaker aimed at the springs so they help the springs vibrate. Reverb feedback is a real problem in many amps, and we usually do a number of standard things to fight it, but you could go the other direction and try to enhance it.

Your existing circuit of course cannot drive a speaker, so you'd add some sort of small power amp for that.

Without it sitting in front of me, I am running out of crosstalk ideas. (Not an invitation, I am retired)

In most guitar amps, if you had a little dry bleeding through, it would never be heard.


2016-01-04 10:50 pm

I went the butcher route and chopped the thing in half (with some redesigning or the back of both sections of course). Now there is no crosstalk. It occurs again when I put the two halves together again, but only as the boards are literally about to touch. So if I built this again using the same design I interpreted from the diagram, making both sides a little more compact and having them a quarter of an inch or so further apart would probably be fine.

The reason I need zero crosstalk is because I'll be using this as a standalone reverb effects unit where a 100% wet mix is needed. It will not be installed in a guitar amp.

I'll start a new thread about the feedback knob. I'd like to do it just by adding a pot if possible, if I can achieve anything like that Doepfer video I posted.

Thanks for your help throughout, much appreciated :)