Best way to cure output osscilation

[Chuck H]

Are you experiencing crossover distortion prior to clipping? More important,.. You mentioned that gain increased with reduced NFB, so, are you experiencing crossover distortion at the previous maximum gain level? It's possible the increased gain is allowing the tubes to attempt drawing some grid current and the coupling caps are charging. You could try decreasing the coupling cap value.

[catalin gramada]

Are you experiencing crossover distortion prior to clipping?

Yes, that I talking about, there are not severe but still. Sending more feedback into loop cure it, otherwise running power tubes hotter did not help. Thanks

[Chuck H]

Yes, that I talking about, there are not severe but still. Sending more feedback into loop cure it, otherwise running power tubes hotter did not help. Thanks

Is the crossover distortion occurring above or below your previous gain level when you used more NFB?

[catalin gramada]

Trying to post a pic, found is much better today, but still can see. The mains voltage get lower with 10v around from 232v yesterday to 223v now.....
I measured everything slightly before clipping in respect with the bias voltage was set
Sorry for images was not pretty accurate, it is still pretty noisy, I have a spider on the table

[pdf64]

I don't think it's likely to be crossover distortion in the sense of 'tubes being too near cut off at zero crossing', rather it's probably 'gm doubling' distortion, as discussed https://music-electronics-forum.com/...ad.php?t=41210
Due to that, the transfer function is nowhere near straight, so waveforms will get distorted; try plugging the numbers into nickb's interactive valve data sheet page http://bmamps.com/ivds.html
I suggest not to worry about it, it's part and parcel of class AB tube amps, so don't lose sleep and tube life by attempting to run the tubes so hot that the transfer function straightens out.

[Chuck H]

Is the crossover distortion (or transfer function) occurring above or below your previous gain level when you used more NFB?

[catalin gramada]

I never get before clipping before, that was with 15 db nfb instead 8db now that means top of semialternances was squashed first then crossover distortion appears. Now get crossover dist before peaks to clip

[Chuck H]

I understand. But does the amp currently produce the same gain without crossover distortion as before when you were using more NFB. Or, let me ask this a different way...

What was the maximum gain before clipping when you were using more NFB?

And

What is maximum gain before crossover distortion now that you are using less NFB?

[catalin gramada]

I understand. But does the amp currently produce the same gain without crossover distortion as before when you were using more NFB. Or, let me ask this a different way...

What was the maximum gain before clipping when you were using more NFB?

And

What is maximum gain before crossover distortion now that you are using less NFB?

Understood too, I didn.t recorded, give me couple of minutes to measure, please. Thx

[catalin gramada]

Ok, let me show you some numbers:
With 470 ohm nfb resistor: 26vpp PI input -77vpp PI output -58vpp OT (meant no trace of clipping, it clips at 64vpp arround)
With 1.5k nfb resistor: 10vpp PI input - 56vpp PI output- 45vpp OT (the output is far away from clipping but at this voltage I can see first trace of crossover distortion on the scope)

[Chuck H]

Ok, let me show you some numbers:
With 470 ohm nfb resistor: 26vpp PI input -77vpp PI output -58vpp OT (meant no trace of clipping, it clips at 64vpp arround)
With 1.5k nfb resistor: 10vpp PI input - 56vpp PI output- 45vpp OT (the output is far away from clipping but at this voltage I can see first trace of crossover distortion on the scope)

Thank you. That answers my question. I don't actually have a suggestion at this time, but I thought the information could be beneficial to the thread. I had thought it might be that your maximum undistorted gain might not have changed, but it seems it did.

I would ask though, is this amplifier intended for sound reinforcement or as a guitar amplifier?

[Helmholtz]

I don't think it's likely to be crossover distortion in the sense of 'tubes being too near cut off at zero crossing', rather it's probably 'gm doubling' distortion, as discussed https://music-electronics-forum.com/...ad.php?t=41210
Due to that, the transfer function is nowhere near straight, so waveforms will get distorted; try plugging the numbers into nickb's interactive valve data sheet page http://bmamps.com/ivds.html
I suggest not to worry about it, it's part and parcel of class AB tube amps, so don't lose sleep and tube life by attempting to run the tubes so hot that the transfer function straightens out.

The signal waveshape shows (slightly) reduced slope near zero-crossing. That's just the opposite of what you would expect from "gm doubling" and is typical for dominating tube non-linearity near cutoff (reduced gm).

Of course "gm doubling" doesn't mean actual increase in gm. Rather voltage gain increases in class A region as the load current is shared between 2 tubes, making each tube see "doubled" load impedance.

But what's wrong with a little distortion in a guitar amp?

[pdf64]

The signal waveshape shows (slightly) reduced slope near zero-crossing. That's just the opposite of what you would expect from "gm doubling" and is typical for dominating tube non-linearity near cutoff (reduced gm)...

Yes, I agree, I posted before I'd seen the scope pics; I'd been expecting the kink to further along from the zero crossing.
As the issue is only reported to manifest at very high signal levels, I wonder if there's some degree of grid rectification going on, increasing the effective bias voltage.

[Chuck H]

Yes, I agree, I posted before I'd seen the scope pics; I'd been expecting the kink to further along from the zero crossing.
As the issue is only reported to manifest at very high signal levels, I wonder if there's some degree of grid rectification going on, increasing the effective bias voltage.

Maybe smaller coupling caps? You could probably go as low as .033uf without noticing.

[J M Fahey]

Found crossover distorsions appears as amplification goes up.this was not present with 15db nfb before. There is a way to get rid of it?

If itīs not present at lower volumes then it is NOT crossover distortion by definition, although waveshape looks *somewhat* similar.

It is bias shifting which appears at high power, a very diffeent thing, in my way is "natural", it was "always there since day one" , is an important part of tube sound (so much so that Transtube and others create it on purpose, go figure) and people startted worryoing about it *only* after they started scoping amplifiers , misinterpret it and, of course, "listen with their eyes" so itīs now "unbearable"
Snowflakes.
[/rant mode]