Plexi Clone Build - Cannot figure out why harsh distortion

I've posted many times about trying to debug this plexi clone amp I've built (see also https://www.tdpri.com/threads/how-t...-down-weird-overtones-in-plexi-build.1025094/ , https://www.tdpri.com/threads/not-sure-if-my-plexi-build-has-problems.984480/ and https://www.tdpri.com/threads/marsh...riations-get-breakup-at-lower-volume.1041560/). I've been trying to figure out over the past 11 months why the distortion sounds harsh and buzzy and I've made close to no progress. I'm losing my mind.

At first I built it with a bunch of mods, but I've rolled it incrementally completely back to stock. Here are all of the things I've tried to figure this out. Throughout every change, the buzzy distortion persisted, and I heard no difference (in the harsh buzzyness).

1. Take circuit completely back to stock, including removing tube rectifier
2. Shorten leads as much as possible, make sure tube pin leads are not too close to each other
3. Check all voltages in the amp
4. Swap out all tubes for a brand new set
5. Use my audio probe to listen to the signal at various points in the circuit. Built an attenuation circuit into the probe capable of attenuating the post power-tube signal to a listenable level, so I can listen to the signal post power tube and pre output transformer.
6. Chop sticking
7. Try using the 8 ohm impedance selector instead of 16
8. Completely disconnect NFB from circuit
9. Change the NFB wires to use shielded wire
10. Change the input wires to use shielded wire
11. Change the effects loop wires to use shielded wire
12. Try upping the value of the post-phase inverter "stability" cap
13. Try swapping in a different power transformer
14. Try swapping out each filter capacitor for a larger value
15. Try taking the filter capacitor for the preamp tubes completely out of the circuit, and connect it via alligator clips, in case it was too close to the signal caps.
16. Take out most of the signal caps and figure out which end is the foil-end, mark it, and reorient appropriately
17. Switch V1 from shared cathode to split cathode
18. Take the effects loop out of the circuit and instead connect the treble pot directly to the PI section
19. Try varying levels of NFB, from lots to completely disconnected.
20. Buy an oscilloscope and poke around with it
21. Take amp to a professional tech to debug

I really don't know what else to try.. The amp sounds fine right up until it starts to break up. Then it gets this horrible buzzy noise, which I think might be parasitic oscillation. But I've tried everything I can think of to alleviate it, and I cannot figure it out.


I've tried probing the circuit, and to be honest, I think I may be hearing the start of the oscillation way back in like the first or second tube stage. It's only once we get to the post-phase inverter stage that the oscillation starts to break up, which makes the buzzy sound really noticeable.

I really built this amp so that I could play it overdriven, not clean, so I want to fix this so badly. Does anyone have any clue what else I could try to figure this out?

 

Shot of amp:


Close up of board:


Close up of power section:


Closeup of input section:


Closeup of V1


Closeup of V2 and V3


Closeup of output jacks:


Closeup of power tube sockets:

 

Closeup of left power tube:


Closeup of right power tube


Closeup of filtering section


Closeup of tone pots:


Closeup of more tone pots:


Closeup of more tone pots and input section


closeup of input section:

 

Back of chassis:


More back of chassis:


dent on PT (i originally thought this might be the problem, but swapping out the PT didn't seem to have any effect)

 

Here are some short clips of the sound, probed at various points throughout the amp. To me, it sounds like the parasitic oscillation begins somewhere before the PI section, but is very faint. Then maybe it gets much louder afterward when it starts to distort and generate its own overtones. But I don't know if I'm reading into it and hearing what isn't there.

This is my test loop, playing through a different amp:


This is my test loop playing through the plexi amp, everything on 10, through attenuator and into a greenback speaker:


This is my test loop playing through the plexi, probed at the output of the V1 stage


This is my test loop playing through the plexi, probed at the output of the V2 stage, but prior to tone stack:


This is my test loop playing through the plexi, probed at the output of the tone stack:

 

This is my test loop playing through the plexi, probed at the left PI section:


This is my test loop playing through the plexi, probed at the right PI section:


This is my test loop playing through the plexi, probed at the right tube output:


This is my test loop playing through the plexi, probed at the left tube output:

 

Thanks! What should I do when probing? Play a sine wave through the input? What frequency should I use?

I have a digital oscilloscope, so the signal always looks a little noisy. Should I use the "averaging" function, so it displays the average of the last few trigger points, to make it display more like an analog scope?

 

Thanks, but yeah, it's the same either channel, even un-jumpered

 

Do you have a recommendation for multi-stranded wires (preferably not crazy expensive) that actually stay put where you put them? Posing the wires has proven to be an issue, because I prefer multi core to single core.

If I'm going to have to start moving around things like the OT wires, I'll have to splice them. They're out of slack cause I cut them short, trying to reduce the length of wire runs. And FWIW, the bad distortion sound preceded many of the board-splices.

Which wires are you referring to running across V1? Do you mean the heater filament wires? Is that bad to do that?
You may not be able to tell, due to the lack of 3D in the pictures I took, but I did try to generally keep wires apart from each other in the 3rd dimension, especially signal wires coming and going from the tubes, and power wires, except for wires that I though I'm supposed to keep twisted, like the post PI signal wires. Ground wires, I give myself some leeway on. And FWIW, chop-sticking all the wires around V1 and the power tube did not make any sonic difference.

But, yeah, I could rebuild it from scratch... just not sure how I'll feel about having to buy basically the whole thing new again, and then what if it still doesn't work?

 

Are you certain that I should not twist the OT primaries? The Metroamp instructions @dan40 sent indicate that the OT primaries should be twisted.

Upon playing through my amp again, I actually think that while the bad sound is definitely still there, it's at least POSSIBLE that the recent changes I made caused it to be a little less obvious than it used to be. Maybe the reason I haven't been able to track it down is that the PO is not caused by one single smoking gun, but that it creeps in lots of locations all across the amp. Maybe some of my recent cap reorintations, etc made it slightly better.

I just ordered a bunch of stuff from the Valvestorm site. Maybe I'll try to redo most of it and do everything like in those instructions you sent, in the hopes that it really is just lead dress, etc, that is causing the problems.