Control pots issue, help please

[catalin gramada]

It is not ripple noise, is a harsh loud noise like a noise from interference

[Helmholtz]

Please add the grid to ground resistors and post scope pictures of the noise at the plate of the third triode and amp output.

[catalin gramada]

I disconnected 12au7 from PI and scoped it output. The issue comes from output of 12au7, meant the third triode

[catalin gramada]

1 M grid resistors to ground did not difference. Have no ideea how to scope the noise, there are lots of components inside the envelope. Any thoughts?

[Helmholtz]

Is the 100k resistor soldered directly to the 12AU7 socket? (It's a very large value grid stopper and wiil steel some highs).
What happens if you directly ground the 12AU7 grid?
Did you try a different 12AU7?
Maybe a socket contact problem?

[catalin gramada]

Is the 100k resistor soldered directly to the 12AU7 socket? (It's a very large value grid stopper and wiil steel some highs).
What happens if you directly ground the 12AU7 grid?
Did you try a different 12AU7?
Maybe a socket contact problem?

100k is directly soldered to 12au7 socket. I don't know what is 12au7 input capacitance and how much treble will be rolled off with 100k grid stop. What I found is, if I put a metal sheet over the issue is gone. Not permanently. Sometimes it did , sometimes not so it is into a stability margin. I was very happy first time when I put a metal sheet over , the issue was gone and can put the treble pot at max without problem. But take it off and put it back and the issue was back. Sometimes it stops... Hmmm, is clear into a margin of stability I think.

[Helmholtz]

100k is directly soldered to 12au7 socket. I don't know what is 12au7 input capacitance and how much treble will be rolled off with 100k grid stop. What I found is, if I put a metal sheet over the issue is gone. Not permanently. Sometimes it did , sometimes not so it is into a stability margin. I was very happy first time when I put a metal sheet over , the issue was gone and can put the treble pot at max without problem. But take it off and put it back and the issue was back. Sometimes it stops... Hmmm, is clear into a margin of stability I think.

Or it is actually interference from an external source like a cell phone, a switched mode power supply or a dimmer. Make sure that the metal sheet is securely grounded. Amp chassis should always be covered/closed by a grounded plane (sheet metal or foil). Also use tube shields.

Amp oscillation/instability is typically a single frequency that should be easy to scope.

Grid-to-plate capacitance of the 12AU7 is specified in the datasheet. Multiply with stage gain the get Miller capacitance. Miller capacitance (+ grid-to-cathode capacitance) and series resistor (grid stopper) determine low pass corner frequency. (I just realized that the 12AU7 is very low µ=open loop voltage gain, so the roll-off frequency is probably above audio. Forget my comment about the high value grid stopper.)

[J M Fahey]

Strong suspect is instability/oscillation, caused by poor layout/grounding/shielding.
The thump/pop at certain pot positions shows the start/stop of oscillation, and "hiss" means it´s present.
We can not hear it directly, too high in frequency for our ears, but the indirect effect of it hitting the tubes, hard.

Have no ideea how to scope the noise, there are lots of components inside the envelope.

What components?
Are you playing the guitar while you test?
1) set controls so as to start/create the noise.
2) set all guitar controls to 0.
I assume you are going from guitar straight into input jack, no pedals or anything else involved.
Put guitar on a stand and scope output.
Take a screen picture and post it here.

[catalin gramada]

Thank You.
I tried to isolate the problem and disconnected first tube, including tone controls. Put the signal from generator to 12au7 input and it works without problem , even with huge 20db amount of nfb. With the preamp connected things goes wrong at the point I put in to max or reach the clipping power stage. It is a high freq problem I think but is weird as did not react to bleeding capacitors I put over the PI outputs or over 12au7 anode, it act the same. The layout is weird, it is SoundCity/ Hiwatt chassis I used, as I have 3 pieces, but is a usable layout. I built a similar one in same chassis gifted to one friend.My buddy didn't report any problem with it. The 12ax7 preamp alone also didn't show any issue on scope.If I connect it to 12au7 input power stage goes into instability , sensitive to treble tone control, volume control , nfb resistor and shows very weird waveforms when clipping.
As time hit instability condition I heard a noise like something arcing and waveform also flick on the scope. I cannot realise if those arcing noise comes from OT or from one of the power tube.
I tried an empirical approach and swapped randomly tubes, components and wiring, including OT. Didn't help I do constantly something wrong. I need a more tech investigation approach with my limited tools. How should I start , please ? Thanks.

[Helmholtz]

The 12AU7 is part of the global NFB loop and thus part of the power amp. Your are using heavy NFB and I think the power amp is close to the edge of instability.
It is possible that the stability depends on the impedance the 12AU7 grid "sees". The signal generator has low output impedance while with the preamp circuit the source impedance depends on controls' settings.

Is the amp stable if you lower or disconnent NFB?

[catalin gramada]

Thank You,
Without nfb the amp is stable, meant have not the issues mentioned before.

[Helmholtz]

Thank You,
Without nfb the amp is stable, meant have not the issues mentioned before.

This seems to confirm my theory above.
Now you may try to lower grid impedance by wiring a suitable resistor (or even a small capacitor of say 10pF) directly between 12AU7 grid and ground.

Or reduce NFB. 20dB is crazy high in a tube amp having lots of parasitic effects.

[catalin gramada]

I used 20db nfb just for test the power stage with signal generator to 12au7 input and as I said was not issues at all. Used in conjunction with preamp stage found 12db to be desirable for my bass aplication, but it osscilate as I said. The issues mentioned was with almost 12 db but with preamp connected. I did just a test with 20db nfb but just with power stage to see how much it counts. But as I said the power stage alone works perfectly even with 20 db feedback, but with signal generator on input of 12au7.

[Helmholtz]

But as I said the power stage alone works perfectly even with 20 dB feedback.

Apparently only if its input is terminated by a low enough impedance (signal generator).

[catalin gramada]

Thank You Helm
I still have half of 12au7 unused. I can made a cf to put in front of, between tone control and 12au7 gain stage. Do You think it solve my problem, please ? Thanks.