Help with signal bleed on a 1964 Gibson GA-30 RVT

[8bitBagel]

I have a 1964-ish Gibson GA30 RVT amp (9 tube version). I bought it a couple months ago with a known audio popping problem, which turned out to be a failing output transformer. The transformer died a couple weeks ago, taking out a couple diodes with it. I replaced that with a model suggested by Hammond, and it works almost perfectly.

The problem i'm having is that it bleeds signal with all of the pots set to 0 on both channels, as well as on the reverb/tremolo effects. It happens when the input cable is plugged into either channel, and interestingly adjusting the trebble pot does impact the loudness of the signal.

I verified that there's no signal coming through the loudness pot, and even with the output from that removed the signal comes through.
I did some spelunking through the wiring with an osciloscope and it seems that it's crosstalk from a preamp tube, V2B in this case (input 2). If I disconnect the grid from the tube (pin it stops, but only if I remove it on the tube side. Unsoldering it on the peg side and leaving a wire dangling still causes the test signal to come through. I tried chopsticking that wire a bit, but it didn't seem to help. It does run very close to the plate wires, but all of the wiring seems original and doesn't seem to have been rerouted. 

I tried a bunch of other things without anything conclusive, like checking the filter caps (replaced by a previous owner), isolating the channels, unhooking the effects input. The final filter that runs the 250V preamp bus does have a little bit of the test signal on it, but mostly removing it with an additional 40uf cap didn't change the signal bleed at all. There's also a ton of RF noise in that area, i can clearly see the test signal if the probe is anywhere near the components in the preamp stage without actually touching them. 

Sorry for the long spiel, but I'm out of ideas. My googling has been inconclusive, it seems like it's not an uncommon problem without any great fixes. I've seen people suggest trying different tubes, but swapping them around made no difference.

Thanks for any suggestions/info! 

TL;DR: amp still amplifies input at 0 everything, seems to appear out of thin air right before the second preamp tube grid pin.

[johnnyreece]

When that happens to me, it's usually a bad ground. Have you checked your pots to be sure they're properly grounded?

[8bitBagel]

I checked the ground on the pots with a multimeter, it looked like everything was grounded properly. I'll triple-check it though.

[Phil_S]

How did you check there is no problem with the loudness pot?

If you haven't done it yet, I suggest trying a jumper from the wiper of the loudness pot to V2 pin 8. Disconnect the grid wire at R45. This will bypass the whole tone stack but leave the suspect in play. What happens? By leaving the existing wire on pin 8, you will know if it is functioning as an antenna. If there is no noise, you know the problem is in the tone stack. If there is noise, you can be pretty sure it is the wire (actually, lead dress.)

If the problem really is the actual bit of wire, you might try some shielded wire from R45 to V2 pin 8, grounding the shield only at one end. Make a chassis ground at whichever end is convenient to a ground point. This is NOT a circuit ground. Think of the shield as part of the chassis, not the circuit. Alternatively, you might replace that wire and find a different route to pin 8. The Gibson amps from this era are a rats nest and can be difficult to deal with.

Post of photo of V2 wiring if no progress.

[8bitBagel]

Thanks for taking a look at the schematics!
I have actually tried a shielded wire to pin 8, didn't help. I had it grounded to a grounding point on the board, IIRC it made it more staticky.
I'll try your suggestion of jumping from wiper to the grid, we'll see what happens.
I did try removing a bunch of connections leading to the V2B grid and to see where the signal originates, and IIRC the earliest I can see it is before R45.
I'm attaching a picture of the entire board as well as a couple pictures of V1/2. The dark green wires are from the wipers of the loudness pots.

FWIW IIRC I tried grounding the output from the loudness pot, and even that didn't even remove the noise. It seems like it's somehow coming through from the first tube, even though at that point there's no direct signal connection going through.

20221120_114416.jpg

20221120_122950.jpg

20221120_122943.jpg

[Stevem]

Since it sounds like this issue was not taking place with the old Ot then I would reverse the two plate leads on the OT.

[Phil_S]

It seems like it's somehow coming through from the first tube...

In that case, jumper from the input (C15) to V2 pin 8 might be a useful diagnostic.

I have 2 Gibson amps from the mid 1960's. My GA17-RVT is more similar to yours than the GA20-RVT. I spent countless hours on the GA20-RVT and finally got it "right." Once you get beyond the immediate problem, you might look at bypassing the icepick notch filters (C19/R44/R45 and C5/R10/R11 to improve the tone.

BTW, be really careful with the opto thing that runs the tremolo. These are unobtainium at any price. For a time, there were some old style Vactrols that might have worked, but the modern ones are LED, so you'd need to modify the circuit. If you wreck it, you'll need to build your own with black heat shrink, an incandescent, and an LDR. Finding the right parts is a trial and error undertaking.

[8bitBagel]

Thanks for the suggestions. I'll take the amp apart and update.

[wpaulvogel]

I notice that there’s no coupling caps before the 330k and 100k mixing resistors originally. Are you 100% certain that the signal bleed you talk about is not a normal condition of the amp?

[8bitBagel]

I notice that there’s no coupling caps before the 330k and 100k mixing resistors originally. Are you 100% certain that the signal bleed you talk about is not a normal condition of the amp?

Well, interesting you mention that. I tried to find out and the response is mixed: Someone on reddit said that their '65 GA-30 RVT does this as well. I also asked a couple people on reverb who are selling these amps and they said that theirs are quiet on 0.
Either way though, I thought that if the loudness pot was at 0 then no signal should be getting through downstream since it was all being sent to ground.

[martin manning]

There are several paths for signal to bleed through, including components or leads in close proximity, and conductive circuit board material. You might be able to isolate the entry point by poking around with a large cap connected to ground, basically shunting the bled-through signal to ground.

[8bitBagel]

How did you check there is no problem with the loudness pot?

I think i forgot to respond to this part - i verified it by going so far as to unsolder that wire from the board, which didn't fix the problem. Grounding it didn't help either, and i didn't see anything there on my oscilloscope.

[sluckey]

What is the amplitude of your test signal? If connected to the input jack you should set the amplitude to about 200mV peak to peak.

Do you have bleed-through when playing a guitar through the amp?

[8bitBagel]

What is the amplitude of your test signal? If connected to the input jack you should set the amplitude to about 200mV peak to peak.

Do you have bleed-through when playing a guitar through the amp?

It starts being audible at around 20-30mV? With a guitar plugged in and guitar pots set to 10 it's at maybe speech volume? I think volume-wise there's no difference between loudness set to 0 and to 1-ish on the amp.

[Phil_S]

The pots are old. I wonder if 0 and 1 are the same. IOW, 0 used to really be off, but isn't any more. Maybe you can unsolder the pot and see what it meters when on the 0 position. Maybe the pot needs replacement? There are a variety of possibilities. If you want to solve the problem, I suggest you hunker down for the long haul. Eventually, you'll find it.