Best solution for reducing Reverb tank feedback (howling/screeching)?

[nevetslab]

Something I haven't yet done but often consider it is attenuating the buffered output of the Fender reverb tank (4AB3C1B), which one rarely has turned up more than a few numbers on the Reverb control (surf bands not accounted for). At the moment, I have a Fender Vibrolux Reverb combo amp in the shop, which howls quite well if you turn the Reverb level up. I first thought maybe I'd just increase the cathode resistor of V4A, presently 1.5k with 100k plate resistor. Then, looking at the schematic of a Fender Twin Reverb, I see they broke up the mixing resistor following the wiper of the Reverb pot in two sections....82k (R74) & 390k (R30), with a shunt to ground following R30 of 220k (R31). The Vibrolux Reverb doesn't have that
shunt resistor, though one could tack-solder it in between the 82k and the 390k, as the circuit is slightly different in that mixing area.

I've already pulled the tank out of the vinyl pouch, added a foam damping pad between the open pan and the cardboard one usually finds covering the open pan, though that made no difference. Just too much return circuit gain to deal with if the pot is turned up.

Custom_Vibrolux_Reverb_schematic.pdf
65_Twin_Reverb_SvcMan.pdf

[Dave H]

Increasing the cathode resistor won't have much effect on gain because it's bypassed with 22u. It just shifts the bias point. You could remove the 22u bypass cap for a 6dB reduction in gain.

I think the extra shunt resistor to ground in the Twin (R31) is to compensate for the reduced attenuation of the wet signal by the dry side 3M3 resistor compared with 470k (R32) in the same position in the Vibrolux.

[nevetslab]

Increasing the cathode resistor won't have much effect on gain because it's bypassed with 22u. It just shifts the bias point. You could remove the 22u bypass cap for a 6dB reduction in gain.

I think the extra shunt resistor to ground in the Twin (R31) is to compensate for the reduced attenuation of the wet signal by the dry side 3M3 resistor compared with 470k (R32) in the same position in the Vibrolux.

I have the chassis removed now, so I think the plan will be to inject signal thru the RCA Reverb return cable, and add a shunt resistor to ground at the junction of the 82k (R65) and the 390 k (R31) mixing resistor and knock that level down 10dB to start with.

[Leo_Gnardo]

I've already pulled the tank out of the vinyl pouch, added a foam damping pad between the open pan and the cardboard one usually finds covering the open pan, though that made no difference. Just too much return circuit gain to deal with if the pot is turned up.

That's been a problem ever since 1963. Dropping the return gain is a good option, but in the reissue amps a big pain because they're on pc boards. Using a split resistor pair in the plate circuit could net you a workable solution. But let's look at something else.

What I do in amps with intractable reverb howl problems, is I go all Boris Karloff and mummify the reverb tank. Home Despot sells a handy weatherstrip tape that's intended to keep the wind from whistling thru the gap between camper tops and pickup truck sides. It's a bit over 1/8" thick, and about 1 1/4" wide. (That's 3mm x 30mm for those metrically inclined.) There's enough on a $9 roll to put a layer on a long tank's top, sides and ends, with enough left over to treat two more tanks. The thin steel on these tanks transmits quite a bit of energy into the tank's interior, generating annoying feedback. Damp those steel panels and abracadabra zut alors, your reverb racket is (usually) tamed.

[Helmholtz]

...and never completely tighten the 2 screws that secure the pouch.

[Helmholtz]

I have the chassis removed now, so I think the plan will be to inject signal thru the RCA Reverb return cable, and add a shunt resistor to ground at the junction of the 82k (R65) and the 390 k (R31) mixing resistor and knock that level down 10dB to start with.

So the -6dB from disconnecting the cathode bypass cap (as suggested by Dave H) is not sufficient?

[nevetslab]

So the -6dB from disconnecting the cathode bypass cap (as suggested by Dave H) is not sufficient?

I never did try lifting the cathode bypass cap. I left that circuit as it was, and dealt with it at the mixing network. The junction of the 82k and the 390k made a convenient place to tack in a resistor to ground. 33k gave me roughly 10dB attenuation. It might have been too much, from the amount of reverb I now hear thru the amps' input, but if I made it less, say 6dB loss, I'd be on the threshold of feedback again. It sounded like enough loss, still plenty of reverb (though probably not enough for surf guitarists....though they'd be using the stand-alone Reverb unit anyway).

I forgot I had a stash of adhesive-backed weather stripping, but, after screwing down the tank, which is now cushioned from the foam sheet I put in before, it sounds ok to my ears. And, didn't have to lift the main PCB up to make the mod's.

[Enzo]

I've never had a howling reverb pan I couldn't tame without modifying the circuit.

VArious thoughts. SOmetimes a pan works better upside down, regardless of which way it is "supposed" to be mounted. The broad flat top surface of a pan can resonate. A strip of self-adhesive foam tape, run down the center of that surface prevents this. I use the gray tape PV uses for the 12AX7 hatch on 5150s for example. The four little short corner springs ought to be damped - a small piece of foam rubber stuck inside the spring coils.

Did we decide just WHAT was feeding back? Electrical or acoustic? Moving the pan away from the speaker make a difference? Does replacing the pan output with a resistor still howls? In fact does it howl with nothing plugged into the return jack on the chassis? Does position of pan change freq of howl at all?

[nevetslab]

I've never had a howling reverb pan I couldn't tame without modifying the circuit.

VArious thoughts. SOmetimes a pan works better upside down, regardless of which way it is "supposed" to be mounted. The broad flat top surface of a pan can resonate. A strip of self-adhesive foam tape, run down the center of that surface prevents this. I use the gray tape PV uses for the 12AX7 hatch on 5150s for example. The four little short corner springs ought to be damped - a small piece of foam rubber stuck inside the spring coils.

Did we decide just WHAT was feeding back? Electrical or acoustic? Moving the pan away from the speaker make a difference? Does replacing the pan output with a resistor still howls? In fact does it howl with nothing plugged into the return jack on the chassis? Does position of pan change freq of howl at all?

It was acoustical. I did play with position of the tank installed in the pouch without any real change. I didn't think to try inverting the tank. I did at least add some foam padding between that cardboard bottom and the tank, while it still being housed in the thin vinyl pouch. Now, I also didn't give a listen to it with the tank moved out off the floor of the cabinet. On the previous amp...a Fender Twin, which was sent back for the reverb distorting, what I heard was it having LF feedback, always on the threshold of feeding back if the Reverb level was turned up. That turned out to be a Ruby tank, which had one of it's two springs broken out. Changed it to the Accutronics 4AB3C1B and all was forgiven.

I thought also about changing V4 to a 12AY7, but, that would have reduced the signal gain of the amp, so I didn't bother.

[Enzo]

I realize it is not a solution to have the pan two feet away outside, but it might illuminate just what part of the thing is sensitive.

By the way, I see the cardboard as forming the last wall of a (relatively) sealed spring compartment. I tape it right tight to the pan. I then put any added padding - which I sometimes use - outside the cardboard.

yeah, for sure a broken spring will cause grief.

[Mick Bailey]

Sometimes it helps to put a dab of hot melt on the transducers if they're a loose fit. Also, a couple of drops of water-thin superglue on the edge of the winding wrap so that it gets pulled in and pots the winding. I always do this after rewinding the coils.

[Leo_Gnardo]

On the previous amp...a Fender Twin, which was sent back for the reverb distorting, what I heard was it having LF feedback, always on the threshold of feeding back if the Reverb level was turned up. That turned out to be a Ruby tank, which had one of it's two springs broken out. Changed it to the Accutronics 4AB3C1B and all was forgiven.

I thought also about changing V4 to a 12AY7, but, that would have reduced the signal gain of the amp, so I didn't bother.

Having only one spring will make a reverb tank sound awful, no matter what brand. The reverb sound needs more randomization, a single spring sounds cheap & horrible. I've been satisfied just fine with Ruby & CE brands which look to be the same exact thing, with the vendor's name printed by the real (made in China) manufacturer. Just think, for a couple thousand buck order, plus a custom decal on each one, you could have your own line of Nevets boutique brand tanks! Dope 'em up with a swatch of foam or hot-glue or RTV in the right spot, maybe a stripe of weatherstrip, charge $75 apiece and you'll be a ... well a hundredaire ... eventually.

I like the 12AY7 solution, it's at least worth a try. Golly, might have to turn up the volume knob a pinch, big deal. Gonna have to give it a go.

Sometimes it helps to put a dab of hot melt on the transducers if they're a loose fit. Also, a couple of drops of water-thin superglue on the edge of the winding wrap so that it gets pulled in and pots the winding. I always do this after rewinding the coils.

I wonder if RTV "bathtub caulk" would work instead of hot melt. FWIW I have no hot melt glue around the shop, never really needed it. Also, you rewind bad reverb tank coils ?!? I'll give you the prize for dexterity & patience. I tried taking a coil/laminations assembly apart a couple times, came to the conclusion that I'm all thumbs, not cut out for this kind of operation. I can see, if spare tanks are way expensive to the point of being nonobtainium, this kind of thing might be worthwhile. We had a go-round about this with Juan Fahey a couple years ago. He was trying to make his own tanks because it would cost upwards of US$100 to import spares in Argentina. Said the process made his head spin, and that's from someone who's a pretty good mechanical engineer in his own right.

[Chuck H]

It would help to know more about the problem. Does it start with feedback in the LF (then bringing in all the springs frequencies) or is it higher pitched to start with? And, as already asked, does it seem to be acoustic or electronic feedback?

I tend to shoot for too much LF in the reverb for the amps I build and it's caused me no end of hardship. Another tank might not respond the same too. But you can go through a few tanks looking for one that behaves and that's not cost effective. But then, neither is taking half a day trying to figure out the right formula for mummifying and mounting a tank. It doesn't make it any better that, although this has always been an occasional problem, newer tanks seem even worse about it.

If it's acoustic feedback and it starts in the lower frequencies you can try reducing the value of the the coupling cap and/or the cathode bypass cap from the recovery stage. Most of what I've read indicates that you're not not supposed to push reverb frequencies below 500Hz. I think that's hogwash but that's what "they" say (including Accutronics while they were still their own entity). Too much LF is often a cause of creeping in feedback for reverb circuits. I've wrapped up and insulated tanks such that I can't imagine anyone's ever done more. Sometimes you just have to play to the limitations of the hardware. And those limitations are greater now than they use to be IMHE.

JM2C

[g1]

I've seen amps with tank mechanical feedback where the last +/- half inch from the speaker is a deal breaker. And ones that couldn't be cured because someone had replaced the speaker with a much larger magnet version.

[Chuck H]

Yepper g1. On my last reverb amp I found that changing one of the speaker types in a 2x12 combo squelched the resonant actuator for the feedback. I did go back to the original configuration with two like speakers in the amp because the customer liked the tone better. That made the hardship of stabilization worse. I managed, but I wasn't trying to do it for money on a schedule.