EL34 suppressor grid

[SoundCity85]

Have an unusual Super City DEA 200 4x EL34 amp. It needs troubleshooting (another subject), but I noticed a 100K resistor between suppressor grid and cathode pins. The HT fuses are connected per pair between cathodes and ground.

What can one possible gain from a 100K resistor between g3 and cathode?

[pdf64]

Perhaps it provides a bit of local nfb, similar to screen grid resistors?
It may impede the secondary emission suppression action a bit, adding a bit of a kink to the anode characteristics, non linearity to the transfer function, and so more harmonics added to the signal waveform?

Whatever, as far as I’m aware, it’s an operating condition outside anything suggested by the manufacturers.
Ideally, someone with a curve tracer would investigate this. Martin Manning has a long thread on his utracer, over at TAG 

Failing that, I’ll try adding them to my Sound City 50+ beater, which makes a good test bed, and see what happens 
That may take a week or 2.

[SoundCity85]

That would be really cool thing to try. Right now the amp has a few failed components in the power section, and my idea was to fix these issues and get the amp up and running again. But the next step would be to convert to power amp to more typical Marshall specs. There's no schematic available so it's quite some work to figure everything out. It has two diode bridges with HT for the anodes probably in the 700-800V range, where the screen supply only comes from one of the rectifiers. I think a bit similar to the Dynacord Gigant. But I don't want an amp running on these kind of plate voltages in any case

[PRR]

We normally wire power pentode G3 to ground or cathode. Cathode is the more common.

G3 current is really very small in normal operation. 100k works a lot like a jumper.

7027 has a G2 where most other power Octals have G3. If 7027 is put in a fully-jumpered EL34 socket, BOOM!

With 100k, the 100k just gets hot. If well-sized, nothing smokes. At 400V a 100k has to be well over 1.6 Watts, so a physically large part. Back when 7027 was readily available, this may have been a slick trick. Today 7027 are very rare.

If you truly have 800V in there, it gets silly. 6.4W heat, spec a 13++W part, a 20W resistor is large. Per socket! Four 20W would hide the amp. Anyway almost no other common tube has a hope of living with 800V.

If you find another explanation, I'd like to hear it.

[SoundCity85]

Thanks for your input. I'll start with fixing it, then check voltages, and then redo the entire thing

[SoundCity85]

It looks like the circuit was already modified, so the split rail 2x diode bridge is not longer interconnected. The reservoir cap for the plate supply is grounded now instead of having the minus connected to the screen supply HT. Still haven't powered up but it's clear the 700+V are not longer present. So I guess I'll just wire this thing up properly like a Marshall power section