proper voltage at PI input ?

[Alan0354]

Too many suggestions to try guys ! Thanks, but i think whats going on is that marshall and others who use the 470R do so because they want the drive which i don't, and they relieve the oscillation with that 50 or 100pf cap across the PI plates. That gets rid of it, but i don't like the effect it has so i'm just leaving it at 1.2k on the cathode, which i just tried and that works.

On the split load PI....i thought about that, but for now i'm going to leave the mods to the master. they seem to work well, but if i decide not i may try the split load. The reason i haven't so far is that the plate resistors and caps there are well in place the way i installed them and are a pain to remove and change.

Regarding to the thick trace on the scope, scope the input of the PI stage from the tone stack, if you see the same thick trace, you definite don't have oscillation with the power amp. My suspicion is you don't have oscillation.

[daz]

Regarding to the thick trace on the scope, scope the input of the PI stage from the tone stack, if you see the same thick trace, you definite don't have oscillation with the power amp. My suspicion is you don't have oscillation.

No, i already did that and the signal is great all the way up to the PI. But there are some questionable things and possible hints...

1-it Happens in the master's middle range, not at very low settings or very high.
2-it only shows up at the output, not even at the EL34 grids.
3-only happens with NFB resistance of 50k or less.
4-either a small cap(50-100pf) across the PI plates or a higher PI cathode eliminates it.

[Alan0354]

No, i already did that and the signal is great all the way up to the PI. But there are some questionable things and possible hints...

1-it Happens in the master's middle range, not at very low settings or very high.
2-it only shows up at the output, not even at the EL34 grids.
Does not make sense on this one.


3-only happens with NFB resistance of 50k or less.
4-either a small cap(50-100pf) across the PI plates or a higher PI cathode eliminates it.

It is oscillation then. a 100pF is not going to change the sound. It is a good fix. I would have done it this way. This is called dominant pole compensation that you introduce a pole in the forward path ( open loop gain from PI to power tubes and output transformer).

If you want to learn about close loop feedback stability, pick up an opamp text book. The power amp is an opamp with +ve and -ve input. The NFB is exactly the same in opamp. This is called the non inverting opamp configuration. Everything concern the stability is same as opamp stability consideration. You will find the meaning of Forward gain, feedback gain, Bode plot and dominant pole compensation.

[daz]

#2: may not make sense, but it's the truth. It only shows up when i scope the speaker out. Then i can take the probe and put it anywhere before that and normal wave. Even tho it's the PI where i can eliminate it. I'm just the messenger.

[guitician]

What your seeing could be related to the triggering of the scope. Is the ground on the OT common?
Another thing I noticed in your schematic was the 30K combination of 10K PS resistors to V1. I recently found that higher V1 voltages 300V+ on the node, seem to distort sooner, or that they just have a less linear transfer characteristic. Which is opposite to what I thought would happen, so I usually designed then to run on 250V. Just a thought, I know you like where you are now with it, tonewise.

[daz]

My V1 plates are only at about 200v each. The PI is at around 290 with the 1.2k cathode. with the 470R it was about 230 at each PI plate.

On the OT thing....do you mean is the scope grounded to the OT ground? If thats what u mean, no, the scope is just grounded to the amp thru the AC receptacle.

EDIT: Wait....you mean 300V at the node, not the plates of that tube, right? Then i DO have 300 on V1 and 371 at the PI. I think your observation would not be agreed with by others, but what the heck, i may try that. Gotta wonder tho.....what would the node voltage matter....wouldn't it be the plate voltages that would be all that matter? I mean granted, the more node V the more plate. But what if the plates were at XXX voltage with 300V at the node then you drop the node to under 300 but bias the tube so the voltage stays the same as it was before. I think that might be telling as to whether it's the node voltage or plate. Like i said, it would seem the plate voltage would be what matters, not the node V. But i;m an idiot so theres likely something i'm missing.

[Alan0354]

It is very hard to talk by drawing, a scope picture speaks a 1000 words.

Your original NFB is 100K feedback resistor and 4.7K gain setting resistor. If you change from 100K to 50K on the feedback resistor without reducing 4.7K to half, you reduce the close loop gain by half. That by itself can sometimes cause instability. That's the reason some opamp specified to have minimal closed loop gain. I bet if you half the 4.7K so you keep the same closed loop gain, you won't have problem.

But again, until I see the picture, I cannot say for sure, I just listen to what you said.

[daz]

It is very hard to talk by drawing, a scope picture speaks a 1000 words.

Your original NFB is 100K feedback resistor and 4.7K gain setting resistor. If you change from 100K to 50K on the feedback resistor without reducing 4.7K to half, you reduce the close loop gain by half. That by itself can sometimes cause instability. That's the reason some opamp specified to have minimal closed loop gain. I bet if you half the 4.7K so you keep the same closed loop gain, you won't have problem.

But again, until I see the picture, I cannot say for sure, I just listen to what you said.

My camera is missing so unless i find it thats not gonna happen. I think it may have fell out of my jacket pocket onto the street.

The resistor is at 33k, and stock marshalls use from 47 to 100k and the 4.7k is 4.7k in every marshall i've seen. So i don't think that should be changed. The 100k is nothing more than a size I chose myself. I could have chosen a 22k or anything. In other words, it's just a random choice and the 4.7k is there because it's a marshall design that works with thier FB resistor, which as i said varies depending on the model and tap used. I also have a 250k variable NFB pot, so theres no point in that anyways unless i decide to do away with that, and thats not likely for good reason.

[teemuk]

In case we're still wondering what's on the scope screen here's one visual example of how oscillation can look like on the scope screen. In this case it's spurios oscillation triggered by something related to high enough voltage swings. Scope frequency is set to capture that 40 Hz wave so the resolution is too low to display very high frequencies so the "waveness" simply shows up as thickening of the trace. The thickness is dictated by oscillation's amplitude and scope's speed to capture it on display.



I do have to say though that the drawn example of that output wave is a bit confusing. It displays both halfwaves right, because how it's drawn it's not really clear? Is that the output with a sine wave input to power amp, or is the input signal passed through the preamp as well. You get most reliable indication of what the power amp section is doing by hooking the signal generator straight to the power amp input, bypassing the processing introduced by the preamp completely.

[J M Fahey]

FWIW what you show is oscillation.

Applying NFB is difficult when you have a transformer within the loop ..... what happens in tube amps , so you can only apply so much NFB before it becomes unstable.
That´s why you don´t have problems with little or no NFB , which appear as you increase it (by decreasing resistor value).

It´s a tribute to Fender´s *excellent* transformers that they can use so much NFB (8:1 from 4 ohms tap) and stay clean and stable.

PS: adding a small cap in parallel with the NFB resistor may make things worse, by *increasing* NFB at higher frequencies.

If anything, lower HF gain locally, not end to end.

To that end, you can add a 47pF or 100pF ceramic or mica (x 1000V) cap from plate to plate at the PI.

Of course, that´s what they wanted: loud and clean.

EDIT: "you" means DAZ
By sheer chance Teemu´s post appeared while I was answering.
And yes, his scope image shows oscillation on wave peaks.
Daz one shows more oscillation, not surprising because his is WAY overdriven, so *all* of the bottom side is "peak"
Simulposting.

[daz]

Daz one shows more oscillation, not surprising because his is WAY overdriven, so *all* of the bottom side is "peak"
Simulposting.

Do you mean my preamp OD is high? Because some have said that, but i keep saying that's not so. I don't think you can be so sure just looking at my schematic because thats just not the case. That wav i drew and the one that was a actual pic i posted were done with the gain somewhere around 12:00 to 2:00. At that point theres about as much gain as a tube screamer. The OD it produces maxed is literally the same or less than a peavey classic 30's disto channel or maybe 20% more than a blues Jr's preamp maxed with the fat switch on. If you ever played one of those you know thats not a lot. Also, it will do that even with lower preamp drive, quite low. Also note it will NOT display that oscillation at the master's input, so how can the preamp be generating it?

[guitician]

On the OT thing....do you mean is the scope grounded to the OT ground? If thats what u mean, no, the scope is just grounded to the amp thru the AC receptacle.

The probe has it's own ground that should be used only at the nearest ground point to the point being measured. At least that's how my scope works. check your owners manual.

[guitician]

EDIT: Wait....you mean 300V at the node, not the plates of that tube, right? Then i DO have 300 on V1 and 371 at the PI. I think your observation would not be agreed with by others, but what the heck, i may try that. Gotta wonder tho.....what would the node voltage matter....wouldn't it be the plate voltages that would be all that matter? I mean granted, the more node V the more plate. But what if the plates were at XXX voltage with 300V at the node then you drop the node to under 300 but bias the tube so the voltage stays the same as it was before. I think that might be telling as to whether it's the node voltage or plate. Like i said, it would seem the plate voltage would be what matters, not the node V. But i;m an idiot so theres likely something i'm missing.

Yeah. I meant the 300v on top of the load resistor(Ebb). 200v on the plate is about right. I thought it would be lower then that with the 10K resistors. The load line is formed from the node voltage and will change the slope of the transfer characteristic. A chart on the Tung-Sol data sheet for the 12ax7A shows the most gain(69) is had with a Ebb voltage of 300v and 3k6 Rk and a 510k load resistor.

[daz]

By the way people, SORRY, but the schematic i uploaded shows the outputs as 6L6 when they are actually EL34 ! I was using 6L6 for a while and thought i'd end up sticking with them. So i will change that schematic i posted, but to those who already looked at it, they are EL34.

EDIT: seems you can edit a post past a certain time, so the post where i included the schematic i cannot edit to put the revised one in. So here it is with all changes i made since then reflected except the PI cathode resistor which is no longer 2k, but I'm not sure yet what i will do with that. I may go back to 470R and use a cap to kill the oscillation. But i changed 3 or 4 things since and those changes are reflected in this schematic...

[guitician]

Well an EL34 is about three times as sensitive to input voltage as a 6L6GC, which may be the cause of problems.