Traynor YGM, sparks flying!

[g1]

Can you post some pics of the area where the mods were done/reverted to stock.
Have you verified that it otherwise matches the schematic version you are using?

[hylaphone]

The switch reads 1 ohm to ground or the CT reads 1 ohm to ground ? Easy enough to unsolder the switch and see if it's shorted to case or CT shorted to case?

True, please disregard. With the switch closed there is ~ 1 ohm resistance between poles, no shorts to case or anything

[hylaphone]

Can you post some pics of the area where the mods were done/reverted to stock.
Have you verified that it otherwise matches the schematic version you are using?

Sure here are some. Aside from some value changes, adjustable bias, etc. it's functionally the same. As mentioned the amp was extremely noisy when I received it. Reworking the grounds, distributing the filter caps, and adding another filter node for V1 solved the noise, and the amp worked great until I started frigging with the B+...

[g1]

Did you also add in the individual screen resistors? Not sure if I've seen high value like 1K5 for EL84 before. And R40 change from 470R to 1K? No idea if that could contribute, but I'm thinking about what Helmholtz said about meter probe induced oscillation. Any chance there were any other changes introduced at the same time as the zener mods? If so, they could possibly be contributors. If not, you need to be certain you had probed the same spots when all work was done except zener mod, and that the sparks did not occur then. Even change of bias could contribute if it's probe induced, and phase reversal of output might also do it. May be worth breaking the NFB loop to check.
It may not seem like it to you, but there are a lot of changes from the stock unit here. Again, if the issue is probe induced, it could even be ground scheme changes, filter node changes, or simple lead dress.

[hylaphone]

Yes I changed those resistors, and will switch them back for further testing. I don't think anything else was changed.
Any work on the actual board was my attempt to clean up some sloppy stuff from a previous tech, particularly the bias section.

Admittedly I've been pretty frantic since the 1st meter died, but I would like to clarify one point - the 2nd meter was not used on the anodes, since Helmholtz mentioned that. I did check the screen voltage, but it was the reservoir cap that sparked and killed it.
I do have an old Fluke that's good for 1200VDC, but I'll hold off until a safe path forward is revealed...

Any thoughts on swapping the OT or is that ruled out at this point?

[Helmholtz]

Excessive voltage at the reservoir cap could mean that its ground connection is bad/missing.

[hylaphone]

Excessive voltage at the reservoir cap could mean that its ground connection is bad/missing.

Hmm I'll reflow and triple check.

Incidentally, why would C28 rated for 250V?

[Helmholtz]

Use your Ohmmeter to check from C30 minus to chassis.

[g1]

Does that lug connecting C30 (and stdby switch) to chassis ground have a star washer on it? It is carrying all the current from main reservoir and rectifiers.
C28 is a bias cap that should only have around -30V so 250V rating should be more than enough.
As far as OT question, that was related to my comment about NFB in post #19.

[hylaphone]

Use your Ohmmeter to check from C30 minus to chassis.

Yes already checked that, that ground is good.

[hylaphone]

Does that lug connecting C30 (and stdby switch) to chassis ground have a star washer on it? It is carrying all the current from main reservoir and rectifiers..

No, that's a good point. I'll try that first then bust out the big iron to solder it chassis.

[g1]

All those nyloc bolts holding ground lugs to chassis should have star washers under the lugs.
AC cord ground, 1R cathode sense resistors ground, main filter cap ground, other 2 filter caps mounted on strip ground.

[Helmholtz]

C28 is connected to a high voltage winding. If R39 fails C28 might explode if not rated high enough.

[hylaphone]

OK third time's the charm, I secured all of the grounds and am safely reading B+ throughout the amp. It also sounds lovely.

Thank you for bearing with me through this.

Now back to my original dilemma, 430V B+ seems insane for EL84. Would like to return this to my friend as a reliable amp for gigging, would you leave as is or try to lower it somewhere below 400V?
A new PT is too costly for now, whats the next best/most foolproof method in your experience? Now that I'm staying well away from those plates, I could try the amplified zener thing again, or VVR? Bucking transformer?

[g1]

The 430V seems insane but it's close to the schematic's 400V and it worked for them (at factory bias level). And no more notorious for eating power tubes than many other EL84 amps out there.
If the heater voltage is on the high side a bucking transformer would be a good solution, otherwise I'd be interested in hearing how the amplified zener works out.