Signal leaking through with volumes off

[Greg_L]

If my hypothesis of a grounding issue is correct, the signal leakage showing at the V1b cathode would depend on where you connect the ground lead of your probe. A good grounding point should be the negative lead/terminal of C20.

Ok thanks I'll try it again.

[Greg_L]

If my hypothesis of a grounding issue is correct, the signal leakage showing at the V1b cathode would depend on where you connect the ground lead of your probe. A good grounding point should be the negative lead/terminal of C20.

Okay back to this....clean signal at V1b cathode with probe grounded at C20

[Helmholtz]

Okay back to this....clean signal at V1b cathode with probe grounded at C20

I wonder if your scope is able to display/resolve a signal of only 1mV?

Anyway, if it's not a grounding issue (to further exclude a grounding issue check for zero resistance between ALL grounding points as indicated by triangles in the schematic) I could only think of direct plate-to plate leakage within V1. Try different brand tubes.

[Greg_L]

I wonder if your scope is able to display/resolve a signal of only 1mV?

Anyway, if it's not a grounding issue (to further exclude a grounding issue check for zero resistance between ALL grounding points as indicated by triangles in the schematic) I could only think of direct plate-to plate leakage within V1. Try different brand tubes.

I don't know how small the scope will go, but it will measure in the single mv range.

I will try some different tubes.

I will quadruple check my grounds, but every time I check them they're good.

The amp is grounded in a....multiple star ground scheme I guess?

The power cord has it's own chassis-bolted ground.
The power node A, PT center taps, and bias ground are all at one point near the PT mounting. I'm not using a mounting bolt, but it's in that area on it's own secure spot.
The other three power nodes, eyelet board grounds, and pots ground along a buss rod floating behind the pots. One end of this rod is bolted to chassis near the normal channel input jacks.
Input jacks ground via chassis connection.
OT grounds at the speaker jack.
Footswitch/reverb jacks ground via chassis connection.

I've built similar amps using this grounding method before and they work great. Very quiet, silent, no-hum, no noise, and no signal leakage operation.

Probing the output with no input signal shows a very clean stable empty flat line on the scope. If I scale way down to like 10mv/div it shows a little flickering hairline noise, but that seems to be nothing.

[Helmholtz]

If you used Fender type fiberboard you might have a conductive board.

[Greg_L]

If you used Fender type fiberboard you might have a conductive board.

It is a Fender-type fiberboard, and the tube sockets are fairly cheapo. I wonder if it could be either of those things.

Also, I checked my grounds again, even moved some around. Grounded my buss rod at different spots. No change. Still get leakage on both channels.

[Greg_L]

Before I just give up on this thing, one last bit of info I've noticed if anyone cares...

The leakage happens with the volume pots all the way down. That much we know. BUT.....as soon as you just start to crack the volume pots open, the leakage vanishes as normal signal starts to flow through the pot.

Probing the output, I see leakage with volumes on zero and as I soon as I start turning the pot it goes away and then normal signal comes through.

What does this suggest? Besides me being too deep into the weeds on this one?

[Helmholtz]

Means that normal signal is out-of-phase with leakage signal.

[Greg_L]

Means that normal signal is out-of-phase with leakage signal.

I thought that too. I need to check the phase again coming out of the second stage. The 2nd stage should be in phase with the input signal.

[Enzo]

Your amp is working correctly as far as signal goes, you have no phase issues. WHat you have is crosstalk, and that has no guarantee of phase relation. A resistor in some input stage can be radiating into a cap nearby, for example, or two wires run along side one another., 50/50 chance of phase alignment. The answer for you will not be found in the schematic.

[Helmholtz]

If e.g. a conductive board provides a galvanic leakage path between V1a and V1b plates, the leakage signal at the plate of V1b would be out-of-phase with the normal input signal.

To check if the board provides the leakage path you may temporarily lift the connection between the V1a plate wire and plate resistor from the board.

[Greg_L]

Your amp is working correctly as far as signal goes, you have no phase issues. WHat you have is crosstalk, and that has no guarantee of phase relation. A resistor in some input stage can be radiating into a cap nearby, for example, or two wires run along side one another., 50/50 chance of phase alignment. The answer for you will not be found in the schematic.

It would appear so. At this point I'm just trying to exercise my brain. It needs it. I'd love to figure this out but I'm satisfied with just letting it ride.

[Greg_L]

If e.g. a conductive board provides a galvanic leakage path between V1a and V1b plates, the leakage signal at the plate of V1b would be out-of-phase with the normal input signal.

To check if the board provides the leakage path you may temporarily lift the connection between the V1a plate wire and plate resistor from the board.

I'm glad you suggested that, I was thinking that myself. At least my troubleshooting process is somewhat getting on track. I'll try it and report back.

[Enzo]

This is the sort of "problem" that teaches you to troubleshoot. It isn't a circuit defect, it won't be on the schematic. You have a signal output that comes from somewhere. SO you must systematically isolate the problem. There are so many "tricks of the trade", some of which we have already seen - like grounding. You can apply a signal to the input and move back from the output to determine just where it is being picked up. Then a similar approach on the front end to determine just where along the signal path the spurious signal comes FROM. Then you have to find what links those two points together. AS mentioned, it could be adjacent wiring, conductive eyelet board, improper wire dress, lack of shielding, etc.

Here is a trick. Get a small piece of metal, maybe the size of a business card or even a playing card. COnnect a wire to it. Now cover it with tape so it can't touch anything live. Ground the wire to the chassis. et voila you have a moveable shield. You can stick it between parts or under wires or wherever looking for possible sensitive areas.

[Greg_L]

This is the sort of "problem" that teaches you to troubleshoot. It isn't a circuit defect, it won't be on the schematic. You have a signal output that comes from somewhere. SO you must systematically isolate the problem. There are so many "tricks of the trade", some of which we have already seen - like grounding. You can apply a signal to the input and move back from the output to determine just where it is being picked up. Then a similar approach on the front end to determine just where along the signal path the spurious signal comes FROM. Then you have to find what links those two points together. AS mentioned, it could be adjacent wiring, conductive eyelet board, improper wire dress, lack of shielding, etc.

Here is a trick. Get a small piece of metal, maybe the size of a business card or even a playing card. COnnect a wire to it. Now cover it with tape so it can't touch anything live. Ground the wire to the chassis. et voila you have a moveable shield. You can stick it between parts or under wires or wherever looking for possible sensitive areas.

Ahhhh cool idea. Thanks. I'll make one of those.