Fender Bassman output waveform oscillations

[Chevy]

Hi, thanks in advance for your input !
Got a silverface Bassman with the AA371 circuit, but have changed the output section to look like an AA864 ( changed feedback loop configuration to 820/100 ohms type, removed 2000pf caps from output tubes, changed bias to be variable bias voltage, and reversed the output xformer leads). The amp appears to run great, puts out more than 50 watts into 4 ohm resistive load, right from around 70 HZ to somewhere around 6 KHZ.... all great except for one detail: when overdriven, I can see the output waveform "wiggling" just where the signal is clipping, on the top and bottom of the waveform. In other words, the squared-off portion of the wavefrom. This to me looks like oscillation..., plus it may get even worse when into a real speaker rather than the 4 ohm load.

I suppose putting the 2000pf caps back in there would tame the oscillations.... but is there something else I can do to get rid of them ?

[g1]

Sounds like maybe you are seeing the ripple of the filter caps at clipping? Is it 120Hz "wiggle"?
What is the purpose of reversing the OT leads?

[Enzo]

If the OT primary leads are reversed, what was negative feedback becomes positive, and the amp can oscillate. This is not uncommon when someone replaces an OT, they get the wires backwards.

It sounds like you have parasitics. You have made major changes in the power amp and removed things, and now it is unstable. Other than the brute force of those caps, certainly work with lead dress.

It MAY get worse into a speaker? It MAY not. DOn;t confuse yourself with maybes, just connect a speaker and find out.

[pdf64]

'What is the purpose of reversing the OT leads?'
The AA371 power amp / negative feedback is configured as an inverting amplifier, see http://www.webphix.com/schematic%20h...a371_schem.pdf
The AA864 as a non-inverting amplifier, as per the normal arrangement.
So, the polarity of the feedback has to be flipped when changing from one scheme to the other.

As per Enzo's suggestions of course, also you could try tacking the 2000pF caps back in place place, then check if the symptom you describe improves.
It may just be one of those things and it did the same before your works.
Pete

[jimboyogi]

Sounds like maybe you are seeing the ripple of the filter caps at clipping? Is it 120Hz "wiggle"?

I've seen this on all of my builds so far, where the squared-off top and bottom of the waveform carries a ripple (in my case 100Hz), when viewed on an o'scope.
Is this not normal, and is it a problem?

[pdf64]

It depends what effect the OP is describing - if possible a photo of the scope waveform may help.
But yes, it may just be B+ ripple modulating the signal (which is normal).
Pete

[Chevy]

It depends what effect the OP is describing - if possible a photo of the scope waveform may help.
But yes, it may just be B+ ripple modulating the signal (which is normal).
Pete

OK, never thought of that... ripple on the power supply showing up there. These caps are 37 years old, after all... Just assumed it was oscillations... the last Fender amps I tested, had "perfectly" square/flat tops and bottoms of the overdriven waveform, no wiggle on them at all. Mind you, on my main Fender amp I put in 220uF caps in the B+ power supply, so that may have eliminated ripple altogether. Don't work on amps much anymore, so a little rusty here. Will check this idea out... but don't know if I can capture a photo or not.

Come to think of it now, I recall repairing an old Hiwatt amp years ago, and it went into actual oscillation at the clipping peaks... you could see the test waveform break out (at the neg and pos peaks) into an oscillating frequency that wasn't the same as the test frequency. Now... this seems to make sense.... I think you're right.

[pdf64]

If you slowly sweep the test signal frequency, the wobbling speed may be seen to 'beat' - slow down, stop then speed up again, as the test signal goes through octaves of the line frequency, if it is B+ ripple modulation.
Do the sweep by hand, it doesn't need to be automated.

Also try measuring the Vac ripple on the B+, at the rectifier output.
May be 2-3V static, ~10V at onset of clipping, even ~15V when overdriven.
Some (cheaper) meters may not be able to make that measurement though.
Pete

[loudthud]

If your scope has a "Line Sync" triggering source option, switch to that and slow the sweep down to 5mS per div. The output signal won't be stable, but you will clearly see the power supply ripple on the top and bottom of the clipped output wave.