Signal leaking through with volumes off

[Greg_L]

I'd start grounding grids from the input on to see if you can determine at what gain stage it's starting.

Ground the grids...like jumper wire from grid pin to ground?

[Chuck H]

+1 (and two and three) to those suggesting directly grounding the grids. If you want to change this situation the first thing to do is isolate the problem and directly grounding grids is the first step. That said...

CTS has been brought up recently on a couple of posts for sloppy pot making. If you started with CTS and replaced with CTS and you bought those pots at the same time from the same place (so perhaps the same "batch") there's a possibility that none of the pots you've installed are reaching zero.*** If you have a good meter an ohm measurement of the pulled pots may reveal something.

***In truth no pot or grounding scheme is ever actually at zero ohms. In guitar amps, which are high gain circuits (even the vintage designs) you don't need much of an error on early gain stages to hear the result at the output.

[Greg_L]

+1 (and two and three) to those suggesting directly grounding the grids. If you want to change this situation the first thing to do is isolate the problem and directly grounding grids is the first step. That said...

CTS has been brought up recently on a couple of posts for sloppy pot making. If you started with CTS and replaced with CTS and you bought those pots at the same time from the same place (so perhaps the same "batch") there's a possibility that none of the pots you've installed are reaching zero.*** If you have a good meter an ohm measurement of the pulled pots may reveal something.

***In truth no pot or grounding scheme is ever actually at zero ohms. In guitar amps, which are high gain circuits (even the vintage designs) you don't need much of an error on early gain stages to hear the result at the output.

The first pots were just no-name pots. They were very stiff, hard to turn, and didn't measure out well. I thought surely they were the problem. Nope.

The new pots are CTS and I did test them before going in. They looked good so I'm hesitant to blame them at this point.


So back to grounding the grids....grounding the first grid should kill everything, right? It's tied directly to the input. If there's nothing at that grid, there's nothing to amplify.
Grounding the second grid will stop everything before that point, right? That would include the input stage and tone stack.
And then the next stop is the phase inverter. Can I just ground PI grids?

[J M Fahey]

Grounding PI grids is fine, and I´d start there, of course, but it´s too wide a funnel.

After that go backwards (towards input) until you find one which does NOT kill crosstalk.

Once you pinpoint the area, look around, think grounding, wire layout, etc.

You might find a cause you can correct, you might not.

As of potentiometers:believe it or not, I have *actually* manufactured potentiometers.

Basic procedure is different from what customers imagine:

1) you coat a long (1 or 2 meters) Pertinax (phenolic paper) strip with conductive paint (phenolic varnish + finely ground graphite)

2) after cooking in a controlled oven, you punch the Omega shaped tracks.

3) here´s the key to your problem: you attach terminals with rivets

4) wiper must not touch the rivets because it will get damaged, so it stops a fraction of a mm away from it.
So you have a very short resistive path between wiper and terminal, which by definition is NOT zero.
perfectly fine for tone controls and 1000 other uses, but volume will never be absolute zero, a tiny amount of sound will still pass.

5) solution? yes, there is one, but it may be overkill and cause the opposite effect: manufacturer drops over the area where the terminal rivet will be applied a drop of silver conductive paint, practically zero resistance.
Now wiper *can* reach the silvered area undamaged and reach zero.

Problem?: silver paint extends beyond expected area, now 0 is 0 put often 1 is still 0 and if very sloppy even 2 us still 0 ,so it´s not an undiluted joy.
And you can coat 2 meters or 100 pot worth of tracks in 2 seconds, but silver dots must be applied one by one, carefully, by a semi skilled operator.

[Greg_L]

Grounding PI grids is fine, and I´d start there, of course, but it´s too wide a funnel.

After that go backwards (towards input) until you find one which does NOT kill crosstalk.

Once you pinpoint the area, look around, think grounding, wire layout, etc.

You might find a cause you can correct, you might not.

As of potentiometers:believe it or not, I have *actually* manufactured potentiometers.

Basic procedure is different from what customers imagine:

1) you coat a long (1 or 2 meters) Pertinax (phenolic paper) strip with conductive paint (phenolic varnish + finely ground graphite)

2) after cooking in a controlled oven, you punch the Omega shaped tracks.

3) here´s the key to your problem: you attach terminals with rivets

4) wiper must not touch the rivets because it will get damaged, so it stops a fraction of a mm away from it.
So you have a very short resistive path between wiper and terminal, which by definition is NOT zero.
perfectly fine for tone controls and 1000 other uses, but volume will never be absolute zero, a tiny amount of sound will still pass.

5) solution? yes, there is one, but it may be overkill and cause the opposite effect: manufacturer drops over the area where the terminal rivet will be applied a drop of silver conductive paint, practically zero resistance.
Now wiper *can* reach the silvered area undamaged and reach zero.

Problem?: silver paint extends beyond expected area, now 0 is 0 put often 1 is still 0 and if very sloppy even 2 us still 0 ,so it´s not an undiluted joy.
And you can coat 2 meters or 100 pot worth of tracks in 2 seconds, but silver dots must be applied one by one, carefully, by a semi skilled operator.

Interesting! Thank you for the info.

So, for example, say I move from the PI backwards and ground the grid of V1b and it doesn't kill the leakage getting to the output. The problem would lie between that grid and the next stage, correct?

[Enzo]

You cannot ground a plate, but you can add a cap to your grounding wire, and use THAt to "ground" a plate.

Remember, your problem has two parts. Something is picking up the signal, and something is transmitting it. I forget, did we actually eliminate the non-zero aspect of the pots? Does grounding the pot wipers kill the signal?

[Greg_L]

You cannot ground a plate, but you can add a cap to your grounding wire, and use THAt to "ground" a plate.

Remember, your problem has two parts. Something is picking up the signal, and something is transmitting it. I forget, did we actually eliminate the non-zero aspect of the pots? Does grounding the pot wipers kill the signal?

Not sure but I will definitely test that tomorrow.

[J M Fahey]

Just one detail: when answering do not quote the full meassage above, it´s already there
At most DO quote it BUT edit so just the first line (or the relevant line) remains, so we know who you are answering to but no more.
Thanks.

[Greg_L]

Just one detail: .

Okay like this?


Test results:
The leaked signal at the output is 190mv pk-pk.

Grounding the PI grids kills the signal.
Grounding at PI coupling cap kills the signal.
Grounding V1a grid kills the signal.
Grounding V1b grid DOES NOT kill the signal.
Grounding wiper of volume pot DOES NOT kill the signal.
Grounding V1 cathodes kills the signal.

So what is this telling us?

[Helmholtz]

You should not directly ground circuit points with a DC potential as this will obstruct circuit/tube operation. This includes PI grids, all cathodes and plates.
It is ok to use a 0.1µ cap to ground, which means AC/signal grounding.

[Greg_L]

You should not directly ground circuit points with a DC potential

I didn't ground any plates. I only briefly grounded the grids and cathodes (which are grounded anyway) to see the waveform.

[Helmholtz]

As said, you must not DC ground PI grids or any preamp cathode as these point (need to) have DCV. It just renders the tube more or less inoperative and such test makes no sense. You only want to ground AC signal.

[Greg_L]

As said, you must not DC ground PI grids or any preamp cathode as these point (need to) have DCV. It just renders the tube inoperative and such test makes no sense. You only want to ground AC signal.

Noted. I'm merely performing the tests suggested by others. I'm new to this and trying to learn.

In any case, I did note that grounding V1b grid did not kill the leakage. It's getting through that stage somehow.

[Helmholtz]

In any case, I did note that grounding V1b grid did not kill the leakage. It's getting through that stage somehow.

Yes, that's interesting. Means that leakage signal the doesn't enter V1b via the grid. And it outrules a bad vol pot. Rather the strong signal at the V1a plate somehow leaks to the V1b plate.
I could think of 3 possibilities:
- bad grounding (check with Ohmmeter)
- close parallel V1a and V1b plate wires
- unsufficient node D filtering: Check filter cap for C and ESR, scope voltage at node D for residual signal.

[Greg_L]

Yes, that's interesting. Means that leakage signal the doesn't enter V1b via the grid. And it outrules a bad vol pot. Rather the strong signal at the V1a plate somehow leaks to the V1b plate.
I could think of 3 possibilities:
- bad grounding (check with Ohmmeter)
- close parallel V1a and V1b plate wires
- unsufficient node D filtering: Check filter cap for C and ESR, scope voltage at node D for residual signal.

Yes, thank you.

Grounding - I've verified good grounds with an ohmmeter. But I have thought of moving some grounding points just to see what happens.
Plate wires - They're not physically close and I have moved them around with a chopstick. No change.
Filtering - This is one area I have not explored too seriously. I have looked for the signal in the B+ rail but didn't see it. I have not actually dug into the individual filter caps though. They're new, but that doesn't mean they aren't bad.