Control pots issue, help please

[catalin gramada]

In this conditions seems to be a problem related more to 12ax7 preamp triode. Can be by grounding, by the fact is shared the same double cap with 12au7 stage, maybe high voltage or layout position, wiring or shielding as was already stated.
How is better to start to investigate , please ? Thanks

[Helmholtz]

If you don't want to do the variable grid load test I suggested, I would give the CF a try.

I already buit an amp exactly same which not perform any issues and work flawless...who knows why...?

Do both amps have exactly the same layout and OT? Both influence stability.

[J M Fahey]

Not to nitpick, just for reference, you are not showing 12dB NFB but 17.35 dB
Voltage ratio: 5000mV/680mV=7.35 X so dB ratio=17.33 dB
Straight from very useful: http://www.sengpielaudio.com/calcula...velDecibel.htm

That said, your power amp waveforms look perfect, so I´d move on and distrust/check preamp, mainly focusing on grounding and layout, with emphasis on the first.

Just curious about indicated speaker (amp output) voltage as shown:

If RMS, 31.6V into 8 ohm means 125W RMS ; more than I expected from 4 x 6L6 .
If peak, 31.6V into 8 ohm means 62.5W RMS ; less than I expected from 4 x 6L6 .

Side comment: when I started doing this thing, "everything" oscillated/buzzed/hummed ,even resorting to desperate measures such as shielding everything, using a separate grounding busbar, even shielding filament wires ; a couple years later most anything worked without big problems, and after 10 years could build visually sloppy amps wired with zipcord (not kidding) and no problems whatsoever.

Physics Laws didn´t change, so only option is that I "learnt" to do it right.
Problem is, besides gross errors, it´s hard to explain in words
Grounding "here" is a mess, and grounding "there", 1 puny inch away, solves a problem ... but how to first detect and then explain/suggest that?

Promise nothing but will reread all posts, look at your chassis pictures and maybe can suggest something.

In principle, on a problematic build, I´ll make the last stage stable and then progress towards the input, step by step, and often "hang" a stage ground to the following one (already stabilized)

EDIT: only see 1 chassis picture, and it´s incomplete, can fully see only 3 octal sockets, the first on the left is half shown.
Power tube grounds go straight to each socket mounting screw so that´s set in stone.
Would love to see the PI and earlier stages, all showing respective grounds; otherwise this is just wild guessing or posting generalities, which may or may not apply here.

[Helmholtz]

Amp output and input of 12au7 with signal from generator

[ATTACH=CONFIG]54404[/ATTACH]

and clipping

[ATTACH=CONFIG]54405[/ATTACH]

12db nfb

From the first scope picture I calculate the closed loop gain of the power amp as 31.6/0.68 = 46.5. This is very high. Typical guitar power amps have closed loop gains < 10. Together with the strong feedback divider attenuation of 50/4750 = 1/95 (from your schematic) this means extremely high power amp open loop gain.
High open loop gain and a lot of feedback is always critical regarding oscillation.

How did you calculate the feedback ratio?

[catalin gramada]

The pictures did not shows the same output level. For feedback loop I just followed the amount of voltage at 12au7 input when output shows it START to distort. In both case with/without nfb that is happen at 32Vp over my 4 ohm power resistor. Just using the same calculator as JMF said with values of input voltage with or without nfb. Using what you see in my pic is not relevant cause the image of clipped signal is away over the moment when it START to clip, so the input voltage is bigger than should be taken in consideration to calculate the loop. I should put pics with exactly the moment when power stage start to clip but was not my intention. I just put the pic to see if are or not any signs of oscillations in those context.
So, taking in consideration the input voltage when power amp START to clip without nfb and input voltage when power amp START to clip with nfb ( meant we rich in both cases the same clipping point or it create the same output effect). In my posted pic the signal NOT started to clip, but is strongly clipped

[Helmholtz]

The pictures did not shows the same output level. For feedback koop I just followed the amount of voltage at 12au7 input when output shows it START to distort. In both case with/without nfb that is happen at 32Vpp over my 4 ohm power resistor. Just using the same calculator as JMF said with values of input voltage with or without nfb. Using what you see in my pic is not relevant cause the image of clipped signal is away over the moment when it START to clip, so the input voltage is bigger than should be taken in consideration to calculate the loop. I should put pics with exactly the moment when power stage start to clip but was not my intention. I just put the pic to see if are or not any signs of oscillations in those context.

O.K. then, what are your real closed loop and open loop gains?

https://en.wikipedia.org/wiki/Negati...back_amplifier

[catalin gramada]

These pictures are more appropriate and shows exactly the moment when power stage starts to clip ( as far I can see on my scope) . First without nfb second with nfb applied. In respect with the input voltage used to get the same effect 200mv vs 700mv the amount of feedback is 10.9db around in conformity with my calculator, and this should be no precise , but more close to reality

[catalin gramada]

The outputs are EL34 into 4 ohm resistor load. The voltages are peak voltages

[Helmholtz]

[ATTACH=CONFIG]54411[/ATTACH]

[ATTACH=CONFIG]54412[/ATTACH]

These pictures are more appropriate and shows exactly the moment when power stage starts to clip ( as far I can see on my scope) . First without nfb second with nfb applied. In respect with the input voltage used to get the same effect 200mv vs 700mv the amount of feedback is 10.9db around in conformity with my calculator, and this should be no precise , but more close to reality

From your pictures I get open loop gain = 162 (huge!) This is a problem. Typical guitar/bass power stages have open loop gains < 20.
Closed loop gain = 45.7 (about the same as I calculated above). Typical values are 5 to 10.
Gain ratio = 3.55 corresponding to 11dB.

You have incredibly high gain levels in the power amp.

[catalin gramada]

Thank You. There are 3 gain stages included in my power amp: 12au7 / 12at7 / EL34. The gain of 12au7 is pretty low, 12at7 give a half of its gain as it is used as common cathode PI.

[Helmholtz]

The gain of 12au7 is pretty low

Don't underestimate the gain. It is most certainly around 10 (easy to measure). And this makes the open loop gain of your amp 10 times higher than typical.

[catalin gramada]

Don't underestimate the gain. It is most certainly around 10 (easy to measure). And this makes the open loop gain of your amp 10 times higher than typical.

Actually 13 around, already measured. Yes, but this did not seems to do the power stage more unstable, as it is fed in from external source, no matter how hard is pushed. Those bloody 12ax7 from input did

[catalin gramada]

I physical separate the preamp ground point from the rest of the rail and also separate supply cap,instead a dual one. Now the preamp is completely decoupled from the rest and I.m hundred percent there are not conflicts. This help for sure as I see there are less interaction between preamp and power stage from an osscilation point of view. But still oscillate over certain level of treble control.



I will try to see if some wiring / screening job can fix it. Otherwise I feel it is time to tame the hights a little, have no ideea how will be better to do. Can You help, please? Thanks.

[catalin gramada]

And this is with signal generator at input of second gain stage (2nd half of 12ax7 preamp) before volume control, with all pots dimmed.



I still have to see what I can do with the first stage...

[catalin gramada]

Question please:
I have to look at a point where are created conditions for parasitic oscillations or at the point where this condition manifest the oscillation effect ? It is not clear for me where the intervention should be done. Need an advice please to get rid of this parasitic issue. Thanks.