Triode gain stage switching scheme

[tmknight]

Hi, I'm hoping to get some guidance on switching in/out a triode gain stage. When switched out of the circuit, is it desirable to retain the plate and cathode voltage and merely "cut" the signal or is there some benefit to tube life if I dump the voltage to ground? Attached are snippets of three different options I'm pondering showing the voltage out of circuit.

Cathode to ground
Signal to ground
Cut the signal

I'd be interested in opinions.

Cheers

[pdf64]

I suggest to just route the signal around and away from the stage.
And don’t mess with the DC conditions, or there may be a thump when it’s switched in / out.
The grid needs a permanent grid leak, ie a DC path back to its cathode, otherwise it’ll float off and mess up the DC conditions.

[tmknight]

Thanks, pdf64, for the input. I've arrived at this, which loosely follows what happens in the Hi/Low input part of the circuit that also "drops" a triode: Grid to ground (via resistor), "output" to ground (after coupling). DC is retained and I think it should be silent switching. I'll admit the 2M resistor is rather arbitrary. Thoughts?

[pdf64]

That looks fine, though I don’t understand why the 2M grid leak doesn’t connect to the 0V common rail in the first scheme?
And why SW1 isn’t wired as a bypass? I thought the idea was to engage or bypass the gain stage.

[tmknight]

My thinking on the 2M grid resistor is that I didn't want the grid to have a path to ground until I wanted it out of the circuit - though, since the resistor is so large, I probably could leave it connected with negligible effect on tone.

As for the switch, that's not something (that I've found) easy to draw up in LTSpice, so I'm "faking" it by changing the net values to replicate the switch state. Possibly the attached layout snippet makes some sense out of my plan.

[sluckey]

Look at Hoffman's hot switch mod. Simple and works well...

https://el34world.com/Hoffman/files/Hoffman_Plexi50.pdf

https://el34world.com/Hoffman/files/Hoffman_Stout.pdf

[tmknight]

Thanks; yep, that's pretty much what I've come up with (slightly different connections, same result). My adding a resistor from grid to ground, I guess, is not necessary.

Thanks, pdf64/sluckey