Debugging a nice 60/120 hum in a reverb circuit

[mikepukmel]

Hi All, I put a jumper on the grid of the recovery tube, but it picked up more noise. I think I need to make a really short cable and solder one end. I will do that later. I didn't think it reduced the hum but with the other stuff it was picking up, hard to tell. I chop stick moved what wires I could move, but didn't hear any improvement. Some photos:









Noticed that if I turn the volume on 10, I hear a similar hum, but much lower level than the reverb knob. I never had the volume on that high while practicing, so probably didn't notice. The hum in the reverb is worse. I will hook up the scope again and get a photo of the trace.

[mikepukmel]

I re-measured the voltages on the two reverb tubes. Didn't see anything odd, but .. probably would not know.

reverb recovery reverb driver
pin1 200 392

pin2 0 0

pin3 1.73 7.74

pin4 3.25ac

pin5

pin6 198 389

pin7 0 0

pin8 1.74 7.71



Voltage on either side of the added dropping resistor for the reverb driver.

426 391

[mikepukmel]

I soldered a very short jumper from pin2 on the reverb recovery tube to the ground lug on the nearest RCA plug, turned off all fluorescent lights, brought in a LED lamp from another room. Turning up the reverb knob is about the same: hum increases with increasing knob setting.

[mikepukmel]

Heck of a time getting a photo of the scope trace. scoped hooked up to speaker output inside the chassis. Kinda stinky but this is it:




Peaks are about 17ms across, so about 59hz (within the calibration state of the old scope). Reverb knob on 10, about 150mv. Reverb knob on 1, just some noise.

This was taken with the short jumper soldered from pin 2 of the reverb recovery socket to nearby ground.

Maybe I screwed up the cap board wiring?

Could this be a leaking filter cap but the reverb recovery just picked it up and maybe its there on the other channels at a lower level?

[mikepukmel]

Ah! Been at work all day and got home to see Tom asking the question I had. That is, what happens when you ground the grid of the recovery stage. Knowing that will narrow the possible sources further.

So...
The hum is 60Hz
Even with the tank cable unplugged from the amp the hum still manifests analogous to reverb knob setting.

That's actually a lot of info right there. But I'm stumped as to how 60Hz can get into that circuit isolated from the tank and transformer. I can't actually think of anything but the filament circuit. Since the rest of the amp isn't humming that seems odd. Might just be a bad tube.

Hi Chuck, maybe the amp is humming, but I didn't notice it. With the scope hooked up this time, after ther reverb knob test, I turned up the channel volume to 10 (never did that before, usually on 2 or 3). I did hear a similar hum, and the scope picked it up (didn't get an image, will do tonight.

[mikepukmel]

I searched old blog posts, and found one that sounded similar but it was a vintage repair. The owner said he found a leaking fiberboard. Any chance that something could be leaking across this board? Is there some way to measure that?

[Chuck H]

I searched old blog posts, and found one that sounded similar but it was a vintage repair. The owner said he found a leaking fiberboard. Any chance that something could be leaking across this board? Is there some way to measure that?

Did you use a crappy, black paper and resin circuit board? IMHE with older amps the problem isn't always moisture. Some here believe that old black boards that leak are ALWAYS fixable by drying them out, but in several cases with Fender amps and one Traynor there was no joy in attempting to dry out the boards. I've come to think that with repeated exposure to high heat and voltages that the carbon used to pigment those boards (most black pigments used are just "lamp black" = carbon) aligns to form conductive paths that can no longer be broken. Tell me you're not building on a black paper, carbon soaked board. With what we know today there's no reason other than the "authentic look" to do that. With cheaper practices for manufactured goods causing problems all the time I'd trust a newly made black paper board about as much as a politician. But...

I don't think it's the board. I've seen a lot of problems due to leaky boards but 60Hz hum hasn't been one of them. There's no 60Hz connections on the board! Even in the odd event that there might be it would be on the power supply end where it's as far from the preamp as possible. Further, the problems with conductive boards tend to be leaky B+ related. That is, high voltage DC. 60Hz could only be present at less than half the B+ so I would expect other leaky board symptoms to show first, or at least as well.

[mikepukmel]

Hi Chuck, Top of the morning to you! (Wow fi you're in the North West, you're up very early!) Nope, Didn't use the paper/cardboard/resin eyelet board. This version of the amp, Im using a fiberglass double turret board (at least what they say on the site) from Watts: https://tubeamplifierparts.com/turre...tegory=9635856

Im not so much interested in all authentic look, would be nice to get the amp working .

I ordered the board in 'black', should have gotten natural. The reason for asking about leaking board, is that this board, the surface is shiny, like as if it was painted. The other few fiberglass boards I have do not have a sheen.

The scope trace, to me, looks really weird. Like a superposition of a partially rectified sine, but with those sharp peaks. I done did something stupid.

(On another note: owe you all an apology for too-many-edits. I peeked at my post history and eek, like 25 edits on 3 posts, which of course generated one email for each edit! Sorry)

[Chuck H]

Shiny happens. I wouldn't expect it to be a "finish" per se, but rather a thin layer of the same epoxy the glass is bonded with. And if it's a clear finish there wouldn't likely be anything in it that's conductive. Unless the manufacturer did something really stupid. An easy tell for a leaky board would be to put your black probe on the chassis and firmly press the tip of the red probe into the board near a HV turret.

[mikepukmel]

Thanks Chuck, will try that leaky board test this afternoon. (Have to run out and paint some pieces of siding on the saw horses in the back yard before it rains. Have a roofing company coming soon, they'll need to replace a few pieces around the roof line.) Does the scope trace look like anything you'd expect?

[Chuck H]

Ok...

The hum is coming from the reverb recovery.

The hum is still present with the grid grounded. (this eliminates the grid and anything before)

The hum is dependent on reverb knob setting. (this eliminates the ground for the recovery as the source)

The normal channel is fine. (this eliminates the D HV supply node as the source of the hum)

A tube swap with a tube that was hum free in the normal channel did not stop the hum. (this eliminates the tube)

So that only leaves the 100k recovery stage plate resistor and the .003 coupling cap. Nothing about either of those components or their proper, or even improper implementation should cause 60Hz hum.

There is nowhere left to go except to assume that there is a mistake in the wiring, most likely having something to do with the filament circuit on the recovery triode OR there have been errors in your reporting.

Please measure AC voltage on top of the reverb pot. Not the grounded lug or wiper. If you find AC you should look for a wiring error on the socket or possibly a stray wire strand coupling the filament supply where it doesn't belong.

[mikepukmel]

Chuck, just to double check, when you wrote "Please measure AC voltage on top of the reverb pot. Not the grounded lug or wiper." you mean the metal case on the back of the pot? I tried all 3 lugs, and the back of the pot, didn't see any AC. I get about 0.4vac with one probe grounded on the chassis, and one probe in the air.

The only place I can find AC, is about 2.1vac at the turret where the wire comes from the rectifier tube, and either side of the choke, and either side of the standby switch.

[mikepukmel]

Double checked, no AC on any part of the reverb pot.

I did also turn up the volume for the channel to 10 (never do that) and I do get a hum, louder than the reverb hum. It gets loud at 8 or so. Nothing plugged into the input jacks. Something big wrong.

[Chuck H]

Chuck, just to double check, when you wrote "Please measure AC voltage on top of the reverb pot. Not the grounded lug or wiper." you mean the metal case on the back of the pot? I tried all 3 lugs, and the back of the pot, didn't see any AC. I get about 0.4vac with one probe grounded on the chassis, and one probe in the air.

The only place I can find AC, is about 2.1vac at the turret where the wire comes from the rectifier tube, and either side of the choke, and either side of the standby switch.

Well that eliminates any stray strand to or from the filament supply. There is literally nothing else I can think of to check. Unless the lead dress and wire routing is so incredibly bad, like intentionally bad, like if you bundled the lead between the .003 cap to the reverb pot with the untwisted filament leads for most of it's length bad.

I actually have a problem with no readable AC. For starters you wouldn't use the open air meter reading as a null. The null would be with the probes touching each other. And since you have excessive audible hum with the reverb pot turned up, there HAS to be AC on it. No AC = no hum.

This would be my next move...

Ground the recovery stage grid and turn the reverb pot up full. Confirm that in this state the amp is, indeed, humming. Now turn the reverb pot all the way down and confirm that the amp is, indeed, not humming. Set meter for AC (appropriate range if applicable) and touch the probes together and scribble down any null other than zero. Turn the reverb pot back up full and NOW measure AC across the two outer lugs of the reverb pot. What happened?

EDIT: Even a few mV of AC will be greatly amplified by the mixing stage because there is very little padding between the reverb recovery stage and the mixer. So something like 4mVAC ends up being something like 150mVAC going into the PI grid. This is why a very controlled reading is critical here.

[mikepukmel]

OK ran through your instructions carefully. I found the meter has a mv knob setting. Shorted, reads 0. Shorted the recovery input with a small piece of bus wire. Looking at the back of the pot, left lug is ground, center wire goes to a 470k resistor, right lug wire goes to a 0.003 uf cap.

Pot at 1, no hum. Left lug ground reads 0, center lug reads 0, right lug reads about 20mv (fluctuates 15mv to 20mv ac). Turn the pot on max (10). Left lug reads 0, center lug reads about 15mv right lug reads 15mv.

So, its that yellow wire! Dang I knew I shouldn't have used yellow. Bad tube socket maybe? I have a couple more tubes, will try them.

I removed the extra 16uf cap and resistor and put the reverb driver wiring back to the factory AB763 routing, through the grommet up to the cap board to point marked B on the schematic. Hum seems a little louder.