Reverb Wet/Dry Mix Problem

[dchang0]

Okay, here's the problem described in two simple diagrams:


Here's what's wrong right now:



Here's what I'm aiming for:



And here's the schematic in question:


As you can see, the 500KA reverb pot isn't functioning the way I'd like, and since I cannot remove the pot easily, I'd like some help and suggestions on how I can add or change resistors to improve the mix. I have easy access to the 100K balance resistor and to the legs of the pot (to add resistors in parallel or series) but cannot remove the pot itself from the chassis.

The goal is to bring the entire wet signal up dramatically so that it reaches Fender surf levels between 8 and 10.

Correct me if I'm wrong--here's what I'm currently thinking:

1) Increase the 100K balance resistor to force more dry signal into the wet circuit

OR

2) Add a 100K resistor to the top leg of the 500KA reverb pot so that it acts like the pot cuts off at 8 when it's turned to 10. This of course doesn't bring the wet signal up in the mix, so I'd rather not do this except as a last resort.

3) Some other idea...

Any help is much appreciated!

[dchang0]

Update:

Here's something interesting: the reverb recovery circuit is actually working properly. If I smack the reverb unit by hand, it gets progressively louder as the reverb pot gets turned up all the way up to 10.

So the problem MUST be in the reverb driver circuit. I'm either not driving the tank with enough current OR I'm not feeding enough signal into the reverb send circuit. What's weird (to me) is that the reverb pot actually affects the INPUT into the reverb driver!

[Enzo]

How would the drive circuit be involved?


And I could be wrong, but I doubt your issue has anything to do with pot taper.

Your mix level drops like a stone, as does the dry signal. Your circuit is not a balance control, it is a straight reverb level control mixed into the signal path. The dry signal should remain untouched, yet it plummets as well.

personally I'd have used a larger mix resistor, you have 100k, Fender used 3.3MEG. But I don;t get the feeling that is the issue, but maybe.

You think the drive is involved? I doubt it, but disconnect the reverb pan drive cable, and connect it to another amp. Now you can hear if the sound goes away. And if it does, it will be because the input to the stage is disappearing, not because it lacks drive.

My first reaction was that the reverb control is mis-wired. What if the wiper is wired out to the mix resistor, and the cap from the plate of the tube is going to the top end of the pot? If so then as you turn the pot down, the wiper goes closer and closer to ground, and pinches off the output at the mix resistor. However, you say the reverb continues to come out, so that kinda precludes that theory.

The thing to do is monitor the signal at the left end of the mix resistor while you turn that pot up and down. It should have little effect there.

Here is another possibility, what if that 100k mix resistor is really a 100 ohm or a 1k, or a 10k?


And yet another, due to layout issues, at higher reverb levels maybe the circuit goes into oscillation at RF and squashes the signal.

A scope would help.

[dchang0]

Hi, Enzo--

Thanks for helping me diagnose this. Bit by bit, I'm ruling things out.

I just increased the mix resistor (which does measure at 100K) to 470K, and the reverb effect is much nicer. I get the Fender-y surf tone between 8 and 10 on the reverb pot. HOWEVER, the wet/dry mix problem has not improved--the dry mix drops out at 7.5 (a little earlier than 8, which was where it cut out with the 100K mix resistor), and the reverb signal still has that weird curve where it drops suddenly after 8. I did go back and look at the Ampeg Reverberocket 2's schem, which uses a 1.5M mix resistor. The RR2 is the closest match to the circuit I've got here.

Dang, I don't have an oscilloscope, and I won't have access to one in the foreseeable future. I doubt that it's oscillation--like I said--if I smack the reverb tank, the sound of the springs comes out loud and clear (and is distorted between 8 and 10).

That idea to hook the drive up to another amp is a good one--I will do that right away.

As for the reverb pot--I had it wired with the wet signal to the top leg, bottom leg to ground, and wiper to the grid of the 12AX7, but three different people pointed out that it would cause problems that way. The way they suggested (as it is currently wired) seems to work properly, at least in that the reverb pot does increase the volume of the reverb return circuit progressively as it should be.

The recovery stage is still very powerful. It's looking like a must to tone it down, because it goes into overdrive on its own (even with the dry signal totally clean). Perhaps the recovery stage is sucking too much current and causing sag in the dry circuit? Just guessing.

Bit by bit, we'll figure it out... Thanks!

[dchang0]

Well, using the second amp was a bust. I hooked the reverb driver RCA plug to the guitar input, and while I got a signal, I have no way of knowing whether that signal is strong or weak or normal. It did start feeding back like crazy at even as low as 1 on the 2nd amp's volume knob (there is no master volume), so maybe it's actually strong.

I'll try the 1M mix resistor next. That's the highest value I have on hand.

[dchang0]

Okay, I upped the mix resistor to 1M. Here's what happens:

1) Predictably, the dry mix is weaker, with more clean headroom on the main volume pot (used to overdrive at 5/12, now at around 9/12).

2) The reverb effect kicks in much, much earlier, at 3/10 on the reverb pot. It gets very "surf" earlier too, around 5/10.

3) The overall dry/wet mix & volume problem persists. The dry volume cuts out at around 5-6 out of 10 and again, the wet signal drops suddenly between 8-10. Same as before, except the dry signal cuts out a tad bit sooner.

That said, using the 1M mix resistor did prove that the reverb drive is working fine. There is definitely enough reverb signal to get slapback off the springs.

[dchang0]

Problem solved. I stuck a 470K resistor in series with the top leg of the 500KA pot, and it effectively cuts off the "crap range" between 8-10. I can now crank the pot all the way up and just barely get into surf territory. I did try a 100K resistor there--it didn't do much but hinted that this might be the solution.

I'll try a 220K to see how that works (get more surf!), but the 470K works great.

Thanks, Enzo, for your help. I'll post up the updated schematic in the other thread where I was designing the circuit.

EDIT: The 220K didn't work. It only cut off the top part of the signals' drops. 470K it is!

[Chuck H]

I don't think Enzo was saying that your pot is wired wrong in the schematic. I think he was saying that it may be wired wrong in the amp itself. Is the reverb level pot wiper wired to the grid of the 6bm8 triode or the following 12ax7 triode? And/or have you double checked the pot as it's wired in place now? Also, as the circuit is now you don't have a resistor separating the 6bm8 triode decoupling cap from the dry signal path with the reverb on 10. Perhaps try a 100k resistor from the send off the pot and a 470k mix resistor (the one that was originally 100k but is now 1M). And triple check your pot wiring.

[dchang0]

Man, you guys are great! Expert help even in the middle of the night!

To answer your questions: The reverb pot wiper is wired to the plate of the 6BM8 triode (recovery) through the 0.022uF coupling cap. The top leg of the 500KA pot now goes through a 470K resistor to the grid of the 2nd half of the 12AX7.

If I understand you correctly, what you have suggested is to put a 100K off the top leg of the reverb pot (where the 470K is now) and lower the 1M mix resistor to a 470K. I can certainly try that right now.

One other thing: I just noticed that the whole chassis is microphonic. Yes, the 6BM8 is microphonic, but should the whole chassis be? Perhaps this is the real culprit?

[dchang0]

Hi, Chuck--

Just tried the 470K mix resistor with the 100K resistor off the top leg of the reverb pot. It works okay, but the wet/dry signals start dropping off at around 7.5/10. At least it doesn't get into the really sharp drop range between 8-10 out of 10.

The 1M mix with 470K off the top leg works better, though the dry signal suffers somewhat.

Either way, I am still facing the microphonics. I wish I had another 6BM8 on hand to test out. Still, I may simply be driving one or both halves of the 6BM8 too hard...

Thanks for your help!

[Chuck H]

I doubt your driving the recovery stage too hard. The signal coming off the tank is about 10mV. Unless the bias condition of that triode is grossly bad you are not over driving it.

If your still getting dry signal drop out with a resistor on the send leg of the reverb pot then something is wired cattywhompus or you have parasitic oscillations causing the amp to cut out.

[dchang0]

Thanks, Chuck. Most likely oscillations, then. I have triple-checked the wiring. But my layout is very, very, very tight (all this stuff in one tiny 5F1 chassis).

Ah well, I'm happy with using the 1M mix resistor with 470K send leg resistor. I'll source some 6BM8's elsewhere--was very surprised to find this brand new EH one so outrageously microphonic. Are 6BM8's naturally microphonic, or is this just a bad tube?

[tubeswell]

In regard to your chassis being 'microphonic', you may need to shield some signal wires. IME this condition can be caused where long lengths of unshielded grid wires run alongside the chassis. (At least when it happened to me on one build - shielding the signal grid wire(s) to the output tube(s) cured the problem). Another possible source of 'microphonic' noise (apart from a tube), is the reverb pan feeding back, which you would find out about soon enough by dampening the hard side of the pan. JM2CW

[dchang0]

Thanks, tubeswell. It's not the reverb pan (removed it from the amp and kept it several feet away). I suspect that it is, as you suggest, some combination of poor shielding on grid wires, crappy/cramped layout, the obviously microphonic tube, and possibly a bad ground. I've double-checked the grounds and so far they've turned out okay. Would having a ground loop introduce microphonics, or is that not possible?

Geez, it's too bad the reverb sounds so good between 1 and 5 out of 10. Otherwise, I could just rip it all out, LOL.

[pdf64]

The plate impedance of the 6bm8 is shorting out the whole signal path at high reverb settings, I think.
Try a resistor (start with ~100k, then increase) between the reverb wiper and the coupling cap. That will act to reduce this effect, but the 500k reverb pot is too high a value. Maybe fit a ~100k across the track ends also. The whole signal level will be reduced, but that's the downside of passive mixers.
Also the reverb footswitch is normally on the recovery side, rather than the send side - any reason for doing it that way?