Strange oscillation?

[Chuck H]

Ahem... I know I often have weird difficulties. I like to think that's because I have little trouble diagnosing normal problems So here goes...

I'm working on a Champ-ish right now. All was fine except there is a crackly distortion on the low notes. While testing I noticed that as I was adjusting the output of the signal generator at low frequencies I could hear a little crackle in the OT (where you would normally hear a sing at higher frequencies). There's that thin, hairy hash line over the scope trace that is typical of oscillation, but it only happened when ADJUSTING the level up or down with low frequencies. I traced and found it didn't happen before the plate of the last preamp stage. I disconnected the power tube grid and grounded it and retested the last preamp plate and there it was still. It doesn't happen on the grid or cathode of that stage nor any part of the rest of the preamp. . I checked all components and found nothing wrong. I tried a different tube. Then I went ahead and rebuilt the circuit with a new plate, cathode and load resistors and new decoupling and cathode bypass cap. The problem is still there. There is no anomalous behavior with a steady sine wave. It only happens with changes to the preamp input level at LF. Any ideas?

TIA

EDIT: Here's a crappy Fender schem I modified to show changes from stock. Sorry for the illegible original. I couldn't find a better one. Y'all know the circuit anyway.

[nickb]

Try using a square wave input say 400Hz and set the tone controls to get a flattish response i.e not too much tilt or peaking on the waveform as best you can. The amount of ringing indicates the degree of stability. With that scope shot as a reference you can try things to reduce it.

Ideas to try:-Layout issue - lead dress; Add 10k grid stopper; Add 10p cap grid to plate.

Edit: While you're scoping the plate, of course

[g1]

Wrong phase OT issues can be manifested in different ways, not always a horrible howling.
I'd guess you have already tried flipping it (OT phase), but thought I'd mention it anyway.

[Rob's Radio-Active]

You may need to up capacitance of the cathode capacitor on the last preamp stage to 10uF. It may be charging and discharging to a different DC level during adjustments and affecting the bias cyclically.
Rob

[Chuck H]

Thanks guys. I had hoped it wasn't an oscillation, but just looked like one. It's definitely an oscillation though. Earlier I said that the tone controls had little effect. Well, that was true enough when I was just jostling them between 3-7 watching the scope. But trying Nick's ideas I found that if I turn the treble all the way down I can almost eliminate it and if I turn the treble all the way up it gets worse. Still, shifting the volume control doesn't cause the problem. Only changing the amplitude of low frequency at the input. Hmmm.?. So, for some reason low frequency amplitude shifts at the input are causing a visible oscillation from the plate of the final preamp stage BUT NOT the first (and only other) gain stage or the grid or cathode of the final gain stage. And turning the treble full up or full down DOES affect the issue. So there is a loop.

g1, I've isolated the problem to the preamp by eliminating the power amp from the signal chain. I did consider the OT phase but had already traced the problem to the power tube grid/last preamp stage area. So lifted and grounded the power tube grid and the anomaly was still present on the final preamp tube plate.

Nick, The amount of classic oscillation mitigation with bleeder circuits that is necessary has proven too great to be a plausible solution. Lead dress isn't that bad but I fussed with it a bit anyway with no change at all.

Interestingly, if I turn the treble full up and adjust my scope to close up on the wave form I can see a 180* image of the hairy wave form popping in and out of the shot.?. So, assuming this MUST be a loop, even though it's only visible on the last preamp plate circuit, my next move is to alter the stock input circuit to include shielded cables and improve lead dress there and for the second stage grid lead. If that doesn't make a difference all bets are off and I'm going to try adding a phase shift circuit somewhere inside where I think the loop is. I may also idealize some grounding.

It's just the weirdest manifestation of an oscillation I've seen. And I've flown much closer to the sun with other, higher gain amps with no such problem.

The only other possibility (I almost don't dare say it) is the stock, black fiber board. If that turns out to be the problem it would make five out of six vintage black fiber boards I've dealt with being responsible for issues. I can read high millivolts off the board now, but nothing unusual appears on any eyelets. We'll see.

[Chuck H]

You may need to up capacitance of the cathode capacitor on the last preamp stage to 10uF. It may be charging and discharging to a different DC level during adjustments and affecting the bias cyclically.
Rob

Haha! That sounds so clever coming from someone else! Because I thought I was reaching for straws when I considered it and temporarily changed that bypass cap to 22uf. No joy

[nickb]

The volume doesn't effect it (all very low) due to the 'bright cap'.

Short V1a pin 2 to it's local ground and see if that kills it. If it does, then the screening is you suggest is the next thing. Often the 68K are mounted on the jack. This lowers their effectiveness as a grid stopper so you try adding a stopper right at the pin . if that fails then I'd start looking at the input jack grounds.

[Chuck H]

The volume doesn't effect it (all very low) due to the 'bright cap'.

Short V1a pin 2 to it's local ground and see if that kills it. If it does, then the screening is you suggest is the next thing. Often the 68K are mounted on the jack. This lowers their effectiveness as a grid stopper so you try adding a stopper right at the pin . if that fails then I'd start looking at the input jack grounds.

Nick?!? Of course shorting the input grid will kill it, because then no LF can reach the next stage. The amp isn't squealing, motor boating, whistling, grunting or losing power. It's crackling on low notes when played through and on the scope it demonstrates a hairy oscillation trace overlying the normal trace when I modulate LF into the input, but the trace looks fine when the LF isn't being modulated. I've HEARD amps do this sort of thing before. I've even tried to troubleshoot for it. But that was in my earlier years before I had enough chops to really tackle it so I figured, like many players I'm sure, that some amps just crackle a little on low notes. But I need to turn this amp over to a friend and I'm just not willing to half ass it.

If I short the first triodes grid I'd need to inject a (roughly) 5V to 11V (modulated) signal below 100Hz into the tone stack to simulate what the first triode was doing. I don't think my sig genny will do that I might be able to get 4.5V out of it below 100Hz. It'll do about 5.5V over 1kHz. I'll check it and see. I'll run that test if I can.

[km6xz]

Can you capture it in a photo and post? The wave form when it is happening will tell a lot. Does it crackle only with bass control change or is it volume change? If you monitor the grid of that tube do you see and dc shift with pot movement?. What is the amplitude on the grid when it does it, and what is the level on the grid when using 1khz test tone at for the same output level? These are general questions that involve FB, bias exceeding level for Class A, or DC leakage. If you disconnect the FB entirely and readjust the signal generator to the same level out, does the crackle still viewed on the plate of the second stage? At the output? If on the latter but not the former, I would suspect transformer.

[nickb]

Nick?!? Of course shorting the input grid will kill it, because then no LF can reach the next stage. The amp isn't squealing, motor boating, whistling, grunting or losing power. It's crackling on low notes when played through and on the scope it demonstrates a hairy oscillation trace overlying the normal trace when I modulate LF into the input, but the trace looks fine when the LF isn't being modulated. I've HEARD amps do this sort of thing before. I've even tried to troubleshoot for it. But that was in my earlier years before I had enough chops to really tackle it so I figured, like many players I'm sure, that some amps just crackle a little on low notes. But I need to turn this amp over to a friend and I'm just not willing to half ass it.

If I short the first triodes grid I'd need to inject a (roughly) 5V to 11V (modulated) signal below 100Hz into the tone stack to simulate what the first triode was doing. I don't think my sig genny will do that I might be able to get 4.5V out of it below 100Hz. It'll do about 5.5V over 1kHz. I'll check it and see. I'll run that test if I can.

No, no! I meant an HF short using a 100pf cap or so, not a bit of wire.

And do post a scope shot with the square wave so we can see if these is any ringing.

[eschertron]

I know you already tried grounding the grid of the power tube, with no change in results. Even though logic says that it is not involved, I still wonder what might change if you upped the power tube 'grid bypass' cap from 47p to closer to the 330p that Fender shipped with certain champs. Presumably to prevent oscillation.

Have you tried taking out the bright cap to see? That's also a bit of a deviation from the original design. Just thinking of things to eliminate as possibilities...

[Chuck H]

Argh! I hate to be THAT guy, but...

Stan, I did disconnect the feedback loop before grounded the power tube grid as part of this process. The problem only happens when LF signal is altered at the amps input. Adjusting the bass control has little affect on the problem and doesn't instigate it. An accurate measure of the voltage at the input is hard to gauge since I'm turning the sig genny output knob by hand. But I am visibly keeping the scope wave unclipped. I can guess that I'm modulating the sig genny output between 50mV and 300mV at 82Hz. Not at all unlike what a guitar might do. The transformer is new and unsullied. I would have suspected it anyway if I hadn't eliminated the entire output circuit and still witnessed the anomaly on the preamp. The crackle I heard in the transformer would seem to be a HF artifact of the problem being accurately recreated by the OT similar to the way any OT "rings" when a higher frequency test signal is used. There is no specific input level that induces the problem. It is only changes in LF amplitude at the amps input that induce the problem. I don't know how to use my scope to check for DC modulation. Obviously my digital meter can't do it. I can dig out a needle meter and see about it though. Advice on the scope procedure perhaps?

I apologize for the lack of scope shots. It would be frightfully hard to capture while wanking on the sig genny volume I might try, but I don't have a working camera right now.

Nick, Ok, I actually do have a 4.7pf shunt between the grid and plate on the input triode (yes, I tried disconnecting it too) I'll try an additional shunt cap straight to ground and across the plate load. But not tonight. Tough day job day. Just having a beer and browsing the forum right now

eschertron, unfortunately the problem existed BEFORE the bright cap was installed. I figured to solve for the crackle at some point, but I'm also voicing the amp. I don't doubt it makes the problem worse though I don't really know because I didn't test the same before that cap was installed. Also, because of the new components (OT and speaker) the tone of the amp would be unacceptable without a fairly bright preamp. Further, attempts to use "bleeder" circuits demonstrated that cap values in the range of .01 or greater are needed to effect a solution. Obviously unacceptable. I still need to idealize a couple of things that the stock circuit ignored (like shielded input leads and ground scheme).

[nickb]

The point of the shunt caps ( wherever you put them) is simply to alter the signal path to determine which nodes are part of the feedback loop.

You can twiddle the the sig gen at 82Hz! That's impressive. Years of practice on your guitar has clearly paid off

To see the DC on the scope, set the channel coupling to DC and slow the timebase down so you can see 2-4 cycles of your 82Hz-ish modulating frequency and look for asymmetry - it would nee to be pretty substantial to see it I would think.

BT, the ghost you saw is probably just a triggering issue - the waveform is not stable so neither is the triggering.

PS: Have a beer for me too.

[pdf64]

...You can twiddle the the sig gen at 82Hz! That's impressive. Years of practice on your guitar has clearly paid off ...

I think Chuck meant that he was twiddling the sig gen 'output level' to mimic a guitar's 'attack,decay sustain, release' signal envelope

[nickb]

I think Chuck meant that he was twiddling the sig gen 'output level' to mimic a guitar's 'attack,decay sustain, release' signal envelope

Yes, that is what I understood too.