Distortion on SFSR power amp section

[Buxton_Flux]

Hello all and happy holidays.

I'm trying to troubleshoot some bad sounding distortion in my silverface Super Reverb if anyone could chime in between their holiday glass of wine and avoiding the in-laws.
 
The distortion comes on by about 2.2 on the volume, and is really saturated by about 3.

I have used a signal tracer to record the guitar and determined that it's not in the preamp section, the distortion seems to happen at the power amp.

Power tubes are so pricey now that I'm trying to diagnose by testing before just buying to see if it solves the problem. I have an emission tube tester and the two tubes both test good, but one will always bias much lower, and it's not the tube, because it does that if I swap them around. I read somewhere that oscillation can cause "one tube to have to work harder than the other." Biasing up to 50% caused one tube to redplate, so it's biased lower than that, but this ain't right. Hotter bias did not affect the distortion anyway.

Voltages are normal.

Filter capacitors are good.

The choke shows 98 ohms between leads, although I didn't unsolder it from circuit.

OT shows 93 ohms from primary to primary lead, and 45-48 from one primary to CT. Hard to get a read on secondary with my basic meter but it wavers from 0.2-0.5 ohms. I had read that a bad OT sounds like blown speakers, but I not sure this reading is off.

Here's a recording of what it sounds like at the preamp, phase inverter, and power amp stages:
https://on.soundcloud.com/aTyyh

Please note the three separate tracks. Each recording is with the amp at about 2.2 first, and then at 3. It does distort some at the PI by 3, but not as much as I'm hearing from the power amp, through the speakers or as recorded at a line out off the speaker jack, in the last track.

Am I overthinking it and it's just bad tubes? I have only one other 6L6, which tests weak but works, and yields the same distortion from the amp at the same point. This led me to think it's not the tubes.

What does this sound like?

[sluckey]

Tubes have to be high on the suspect list. Need to be eliminated as suspects.

[Latole]

I would not eliminate the preamps section, not knowing if the tests were well done.

So to check; all the tubes and the output transformer with an audio signal.

Most tubes testers are no reliable to test outputs tubes. It's like testing a car in a idle on a parking lot, not on a highway.

If the output tubes have already been replaced in the past without bias, which is very frequent, a too hot bias, besides damaging the tubes, can damage the output transformer.

[acheld]

I'm hearing your distortion in the PI recording, and it's amplified at the power tubes (of course). 

Somethin's happening there.

[Buxton_Flux]

I'm hearing your distortion in the PI recording, and it's amplified at the power tubes (of course). 

Somethin's happening there.

Okay good, I will take a look in the PI again. Have tried various tubes there, so I will seek to figure out what else could cause that.

[sluckey]

I have used a signal tracer to record the guitar and determined that it's not in the preamp section, the distortion seems to happen at the power amp.

What did you use for a signal tracer?

[Buxton_Flux]

Now noticed that heater voltages are higher than they should be, at 3.35v, where they should be 2.1v or so. All other voltages seem normal. What would be the possible causes of this, and is this a possible cause of the distortion?

[sluckey]

Now noticed that heater voltages are higher than they should be, at 3.35v, where they should be 2.1v or so.

Heater should be 3.15V when measured from tube pin to chassis. 3.35V is fine. Why do you think they should be 2.1V???

[Buxton_Flux]

Now noticed that heater voltages are higher than they should be, at 3.35v, where they should be 2.1v or so.

Heater should be 3.15V when measured from tube pin to chassis. 3.35V is fine. Why do you think they should be 2.1V???

Misread my notes on prior voltage readings (cathode 2.1v). Disregard prior post.

[Buxton_Flux]

Using an oscilloscope plugin in Reaper helps find where the distortion starts;

It is clean going through the channel mix resistor.

Then at the right grid leak resistor (adjoining .001 cap), by 3 on the volume it takes this shape (2nd image)

However, at the left grid leak resistor (adjoining AC shunt cap), it clips hard starting right at around 2.2 on the volume. The signal looks like this also on the tail resistor and AC shunt cap.

Have been unable to trace the fault in the PI circuit. Have tried replacing the 0.001 coupling cap, the AC shunt cap, and both coupling caps at the power amp grid leak. Tried replacing PI plate load resistors. All other resistors are new and check out. Have tried obviously reflowing solder points on circuit and tube socket. Tried lifting NFB from circuit, to no effect.

B+ 3 = 460v
Bias = -53.4v
V6 cathode = 103.6v
V6 plates = 249/254

Can this be caused by something further downstream, including power tubes, choke, or OT? Tried switching position of tubes and the more severe distortion remains on the left grid leak side.

[AlNewman]

Silly question..  Does your amp have the output tube balance pot on the back? 

Possibly a bad component or connection in your negative feedback loop?

[Buxton_Flux]

Silly question..  Does your amp have the output tube balance pot on the back? 

Possibly a bad component or connection in your negative feedback loop?

This one does not have a bias balance pot (if that what you mean), but I know the two power tubes are not exactly balanced. I only have three power tubes on hand; one good, one biases much lower, another biases close (4ma off) but tests weak in tube tester.

I tried lifting the wire to the speaker in the negative feedback loop, with no effect.

Does this waveform look familiar to anyone? It's weird that it "turns over" pretty suddenly at 2.2 on the volume. What can I safely disconnect to isolate and test?

[sluckey]

Those waveforms do look nasty. However, connecting a scope directly to the grid of that LTP PI is not recommended because even the high impedance of your probe can upset the bootstrapped bias of that PI. Probably be better to connect the scope to the other side of the PI input cap, or even just monitor the PI plates.

[AlNewman]

I really didn't have much faith in the nfb being the issue, but it ties in right around the same spot.  Unhooking it won't solve the problem, but if it isn't hooked up properly it might create the problem, as that's sort of what it's designed to do is stop disortion and oscillation.  As for the matching tube adjustment, it creates a voltage divider between the two output grids, so perhaps it could throw one tube out of whack.

https://www.aikenamps.com/index.php/what-is-blocking-distortion

I've been doing a lot of reading on this lately.  Your problems sound similar.  I was modding mine though, so if you've made no changes in a known good circuit, you're down to bad solder joints, a bad component in the phase inverter or output tube, or a bad tube.  From there it gets into bad transformer or bad speaker, which I doubt.  Have you tried switching the phase inverter and reverb tubes?  How about switching the power tubes, does the problem stay in the same place?  Try cleaning tube sockets, and moving wires and components with a chopstick.

[Buxton_Flux]

I really didn't have much faith in the nfb being the issue, but it ties in right around the same spot.  Unhooking it won't solve the problem, but if it isn't hooked up properly it might create the problem, as that's sort of what it's designed to do is stop disortion and oscillation.  As for the matching tube adjustment, it creates a voltage divider between the two output grids, so perhaps it could throw one tube out of whack.

https://www.aikenamps.com/index.php/what-is-blocking-distortion

I've been doing a lot of reading on this lately.  Your problems sound similar.  I was modding mine though, so if you've made no changes in a known good circuit, you're down to bad solder joints, a bad component in the phase inverter or output tube, or a bad tube.  From there it gets into bad transformer or bad speaker, which I doubt.  Have you tried switching the phase inverter and reverb tubes?  How about switching the power tubes, does the problem stay in the same place?  Try cleaning tube sockets, and moving wires and components with a chopstick.

Appreciate the input. Read your article with interest - still somewhat over my pay grade.

Mine is wired just like the layout on Robinette's page. Standard. I have ruled out the components in the negative feedback circuit here. Have long since tried all different tube and speaker combinations. I'm out out of ideas for how to test anything. Was hoping to learn definitively whether this was a bad tube or other component through testing, but I guess it's not that well documented. I did try switching the power tubes, and it does stay in the same place.

I'm going to have to resort to what the guys who fix amps do, and just buy parts and try them. I don't mind buying power tubes if that's the problem, but if it doesn't fix it, it will be a drag to have to also buy an output transformer. That's about the only thing that hasn't been swapped around so far. It's a pity because it sounds pretty good at 2.

[AlNewman]

I'm wondering... did the bias issue follow the tube when you swapped them?  You said one was red plating?

[Buxton_Flux]

I'm wondering... did the bias issue follow the tube when you swapped them?  You said one was red plating?

Yes that's a good question. I will have to get in there are try it again to clarify.

With the standard 27k resistor on the bias, the bias maxes out at around 35ma or so (plate voltage 465 or so). I wanted it biased somewhat cool anyway. But I tried up to 47k and down to 10k bias pot resistors. With the latter, I can get the bias higher - which doesn't solve the distortion - but one of the tubes was red plating at around 55% dissipation, and I think it was the tube, not the particular tube socket. But I will try that again.

The bias board components are not original, but I have not changed them.

[Buxton_Flux]

Tried this now with the power tubes out.
(Previously only went up to about 3 on the volume because I don't have a dummy load resistor and have to keep the speaker load plugged in).

The clipping is still there without the power tubes in, so I guess I can rule out the power tubes and output transformer. It is just clipping at the phase inverter.

However, the preamp section earlier will also clip at a higher volume setting: somewhere over 5-6. Different tubes will clip as expected; 12at7 higher and 12au7 not at all.

I tried raising and lowering the first (1k) dropping resistor to raise and lower voltage, but things come out about the same in the wash. Tried 're-silverfacing' the PI plate load resistors to 47k (from 82/100) but that didn't help.

I am puzzled why the up and down of the wave is asymmetrical (always clips on one side first), if not because of the difference between the 82k and 100k resistors there.

As a head check, I looked at the high-ish gain signal from another Fender, solid-state amp, and the signal is also a bit weird looking - not flat on top - probably because of the phase shifts in the Fender tone stack.

Although I can believe that v1 would clip at 6, I find it hard to believe this amp should clip at 2.2 in the PI, when it should have headroom. It's as though it's just clipping because there is more gain. Can't really find any documentation on the internet about where clipping levels typically are, so if anyone that knows wants to tell, that would be good to know. Right now, haven't advanced an inch on solving this.