What would cause high frequency oscillation with neg feedback loop connected?

[jasonguitar]

What would cause high frequency oscillation with neg feedback loop connected?

[Enzo]

Same thing that causes it anywhere else - the output is getting back to the input.

All that component brand and chemical compostion stuff is irrelevant, these circuits are not rocket science, not sensitive to tiny changes. However, if a wildly wriong value resistor was installed somewhere - like 100k instead of 100 ohm or something - odd ssymptoms could result. Hey, we've all done it.

What leaps out at me is that you changed a transformer. By any chance did you change the output transformer? Or dewire and rewire the old one?

The negative feedback wire is only negative because of the phase relation between where it comes from and where it is going. A common problem in home builds and OT replacements is guys getting the plate wires from the power tubes backwards. The NFB wire comes from the "hot" side of the OT secondary. The phase of the signal there is determined by how the primary of that OT is wired.

IN a nutshell, try reversing the two plate wires.

ANother possibility is in all the component replacing, perhaps the ground end of a filter cap is not connected to ground itself. We call all those larger e-caps on the B+ "filter" caps, but really, after the plpate and screen sup-plies, little filtering is needed. Those ones for the preamp are more "decoupling" caps. What they do is prevent the signal in one stage from affecting the B+ to another stage. If the B+ from say the phase inverter gets some of the signal from that stage on it, and the preamp uses that B+ also, then that signal can leak into the preamp. And if the phase relation is constructive, feedback can result.

And come to think of it, your amp could be oscillating hardcore at RF when running unimpeded, and the NFB drags it down enough to bring the oscillation into the audio range. In that scenario, your NFB doesn;t kill the oscillation when disconnected, it gets worse, you just can;t hear that high. What makes that plausible is that when amps are oscillating hard at RF, the oscillation itself can;t be heard, but the net result is the audio still passes, but it sounds weak and hummy.

SCope things and see if RF is going on.


There three unrelated but possible means to your symptom.

[jasonguitar]

Same thing that causes it anywhere else - the output is getting back to the input.

All that component brand and chemical compostion stuff is irrelevant, these circuits are not rocket science, not sensitive to tiny changes. However, if a wildly wriong value resistor was installed somewhere - like 100k instead of 100 ohm or something - odd ssymptoms could result. Hey, we've all done it.

What leaps out at me is that you changed a transformer. By any chance did you change the output transformer? Or dewire and rewire the old one?

The negative feedback wire is only negative because of the phase relation between where it comes from and where it is going. A common problem in home builds and OT replacements is guys getting the plate wires from the power tubes backwards. The NFB wire comes from the "hot" side of the OT secondary. The phase of the signal there is determined by how the primary of that OT is wired.

IN a nutshell, try reversing the two plate wires.

ANother possibility is in all the component replacing, perhaps the ground end of a filter cap is not connected to ground itself. We call all those larger e-caps on the B+ "filter" caps, but really, after the plpate and screen sup-plies, little filtering is needed. Those ones for the preamp are more "decoupling" caps. What they do is prevent the signal in one stage from affecting the B+ to another stage. If the B+ from say the phase inverter gets some of the signal from that stage on it, and the preamp uses that B+ also, then that signal can leak into the preamp. And if the phase relation is constructive, feedback can result.

And come to think of it, your amp could be oscillating hardcore at RF when running unimpeded, and the NFB drags it down enough to bring the oscillation into the audio range. In that scenario, your NFB doesn;t kill the oscillation when disconnected, it gets worse, you just can;t hear that high. What makes that plausible is that when amps are oscillating hard at RF, the oscillation itself can;t be heard, but the net result is the audio still passes, but it sounds weak and hummy.

SCope things and see if RF is going on.


There three unrelated but possible means to your symptom.

Thanks for the reply Enzo.
I deleted the whole story cause I thought maybe I was rambling on but I guess I'm glad you read it. I changed the power transformer years ago. Never touched the OT although I was surprised by the oscillation cause it is similar to what it would sound like if the plates were reversed. On a stupid whim, the perfectionist in me decided to improve an amp that was already great! Guess the powers that be decided I needed another lesson!

I only changed the carbon comps in the power supply cause they had drifted and when I replaced the filter caps a while back I neglected to change those too.

The amp actually sounds good if the NF is disconnected.

All the grounds seem good. I get continuity from all supply cap grounds to chassis. This thing is humming alot like its under-filtered but doesn't sound like it at all with a guitar plugged in. If I didn't know any better the thing cranks and sounds really good as long as NF is disconnected. Doesn't sound like it should but not a bad sound. All the output is still there but something is very wrong here.

[jasonguitar]

I'm at a total loss here. With all preamp tubes out it still oscillates with just the phase invertor, which makes since of course but just to be sure its not coming from an earlier gain stage. Everything "looks" right. Does this indicate anything: My b+ is 475v with the NF disconnected and when I connect it, it starts oscillating and the voltage there drops to 455v?

[MWJB]

The B+ drop is your amp eating up current trying to amplify the oscillation.

Did you reverse the OT plate wires?

Anything you might have changed with lead dress between the old PT and the new one? Anything that might have got changed with regard to the PI wiring/lead dress? Any change to NFB loop values?

Something is driving your negative feedback positive.

out of interest what's the amp?

[Enzo]

And we still have not eliminated my RF scenario.

And is it possible when changing parts that a couple swapped places? and now when you put the older parts back in, the mis-placement is continued? I once had a fellow send me a kit he had built and just couldn;t get to work. I found that on his eyelet board, he had a cap and resistor taking each other's place, creating a very odd circuit. The board looked great, but by going down the row of parts and verifying each connection I discovered his error.

Any chance the injection point of the NFB is wrong? Wire connecting to wrong end of resistor or something?

[jmaf]

Did you reverse the OT plate wires?

Putting my money on this.

jasonguitar, have you tried reversing either the pin 5 leads or pin 3 leads on the output tubes? You have positive feedback as Enzo pointed out, even the hum seems to indicate you're adding hum feedback instead of cancelling hum.

[jasonguitar]

The B+ drop is your amp eating up current trying to amplify the oscillation.

Did you reverse the OT plate wires?

Anything you might have changed with lead dress between the old PT and the new one? Anything that might have got changed with regard to the PI wiring/lead dress? Any change to NFB loop values?

Something is driving your negative feedback positive.

out of interest what's the amp?

'71 Super Reverb. Blackfaced. The thing has been working well for a long time. Never had oscillation probs before. I didn't change the lead dress. I changed the NF feedback to 1k and installed new tight tolerence caps and resistors in the phase inverter. I went back to the old parts one by one after the hum/oscillation started but the problem persists.

[MWJB]

OK, the most likely thing is that it is related to the work that you have done at the PI, or set off by it at least.

I would be checking for voltage creep accross the board material in the region of the PI, possible conductivity to the insulator board, which may be conductive too. Check under the circuit board for any large blobs/pools of excess solder (gather very easily when the boards warp).

Triple check those PI values, upping the NFB ratio to 11:1 shouldn't push the amp into instability, but a 10K could easily.

Whilst we're here, what are voltages & plate current like with the new PT?

[jasonguitar]

I haven't scoped it yet. I just recently got a working scope. Its an antique though, not ideal. A mint Heathkit IO-12 which could use a cap job. I've only used it a couple times to check output waveform while adjusting bias and it worked but was a tad unstable. I just don't understand why RF would be a problem all of a sudden. I know 70's Supers are notorious for oscillation problems due to lead dress but this one never had the issue and frankly I have yet to come across one that did oscillate when snubber caps were removed. It sure seems like I have the wiring wrong but I am very, very familiar with this circuit and have gone over the schematic and layout and even sat it side by side with a perfectly functioning blackface bassman which has the same PI values and voltages and everything looks right. 1m resistors on each side of a 470 ohm feeding a 22k at the junction of the 100ohm to ground and 820 ohm to NF wire to tip of output jack. Between the junction of the 820 ohm and 100 ohm resistors and the left 1m resistor is a .1 400v cap and between the 220k channel mix resistors and the right side 1m is a .001 600v cap. Seems like the problem might have to do with the b+ to the phase inverter filtering?

[jasonguitar]

OK, the most likely thing is that it is related to the work that you have done at the PI, or set off by it at least.

I would be checking for voltage creep accross the board material in the region of the PI, possible conductivity to the insulator board, which may be conductive too. Check under the circuit board for any large blobs/pools of excess solder (gather very easily when the boards warp).

Triple check those PI values, upping the NFB ratio to 11:1 shouldn't push the amp into instability, but a 10K could easily.

Whilst we're here, what are voltages & plate current like with the new PT?

I actually loosened the boards, pulled out the insulator and sanded all the flux off. I then removed all excess solder blobs from the eyelet board prior to posting this. Did a pretty thorough job. Didn't clean any flux of the eyelet board itself and of course I can't get underneath it without removing it. Could flux on the eyelet board be causing this?

I had a 1.5k feedback resistor in this same amp in the past with no probs. It was a little too bright but I figured I might like the slight edge of a 1k opposed to the stock 820 ohm. I guess being a 2ohm output, upping feedback resistor values might be more dramatic in these amps?

Won't be at the bench for a little while but I do recall with near 100% certainty that the B+ is 475v, plate and screen voltage were 470v and plate current measured 34/35ma per tube with shunt method.

[jasonguitar]

Putting my money on this.

jasonguitar, have you tried reversing either the pin 5 leads or pin 3 leads on the output tubes? You have positive feedback as Enzo pointed out, even the hum seems to indicate you're adding hum feedback instead of cancelling hum.

It would definitely make sense but I know for sure that I haven't reversed the plate leads. I did freshen up the connections as they were looking a bit frayed but I did them one at a time with brown wire on the left and blue wire on the right. I am pretty sure I double checked the grid wire polarity but I'll check again when I get to the bench.

[jasonguitar]

SURE ENOUGH! GRID WIRES WERE REVERSED!!! They are both black wires and even at first glance they looked right but turns out they were switched at the 220ks when I replaced those resistors. Thanks for mentioning that JMAF! You live and learn! It still seems to have a bit more hum than than usual though.

[jmaf]

SURE ENOUGH! GRID WIRES WERE REVERSED!!! They are both black wires and even at first glance they looked right but turns out they were switched at the 220ks when I replaced those resistors. Thanks for mentioning that JMAF! You live and learn! It still seems to have a bit more hum than than usual though.

Glad it worked for you, but it was actually MWJB and Enzo who solved it. Have yourself some good sounds!

[jasonguitar]

You guys all rock. Very grateful for this forum. Thanks gentlemen!