Different approach to Hi-Lo power output modes. Idea for active switching using a few mosfets.

[SoulFetish]

I think this should work? So, here's the thing. I'm really liking how my 18W EL84 amp has come along. It's so responsive and I'm really enjoying the sound when I can get to push the output tubes. The problem is it is so Eff-ing loud it's ridiculous. Of course, there are the standard solutions –
Master Volume: not great in this circuit;
Half Power mode: Don't have the option of dropping out two tubes, and I loath the sound of switching them into faux triode mode. Totally useless IMO;
Attenuator: This is my preferred solution, and I've had good results in this amp using a purely resistive bridged-t to give me -6dB to -9dB

I'll try and make this quick because I have to go to work soon. But I think I found a way of actually incorporating a true low wattage output amp in the existing design using a pair of 3W 6945 subminiature Beam Pentodes. The beauty of these particular tubes, is that many of operating conditions are nearly identical to EL84s. The major challenge is they have significantly lower voltage limitations.
My idea is to use mosfets in series with the screen supplies for each of the 4 tubes, and use PIV gate drivers to switch between the two pairs of tubes by opening the screen supply to ensure that the two different types of tubes aren't conducting at the same time.

To limit the amount of things that need to be switched in and out to use the subminiature tubes safely, I'm using Mosfets in the cathode circuit in a kind of cascode configuration. Only I'm trying to maintain a stable voltage at the gates in order to lift the 6945 cathodes to around +85V. This is where I'm inexperienced and could use some advice on optimizing this circuit (or let me know if I'm wrong and it won't actually work). Anyway, here is the simplified schematic I drew up quick to show the basic idea:

[Helmholtz]

Regarding the 85V cathode supply:

Firstly I would recommend to use a common supply for both tubes. This makes sure that the zener current doesn't cut out during negative going signal cycles.

Not sure how well the Mosfet circuit would work, but a BJT version seems more straightforward:
https://makingcircuits.com/blog/how-...-zener-diodes/

Also a single 10W Zener with adequate heatsinking might do.

[SoulFetish]

Regarding the 85V cathode supply:

Firstly I would recommend to use a common supply for both tubes. This makes sure that the zener current doesn't cut out during negative going signal cycles.

Not sure how well the Mosfet circuit would work, but a BJT version seems more straightforward:
https://makingcircuits.com/blog/how-...-zener-diodes/

Also a single 10W Zener with adequate heatsinking might do.

That's a good point about the zener when the tube is in cutoff. However, wouldn't an adequately large capacitor from the source keep the Provide the Zener with a stable enough voltage during cutoff periods, and solve this problem?

[SoulFetish]

Wait.. I just realized that there doesn't even need to a mosfet for each tube to bias them at +85V. In fact, by tying both tubes a single mosfet regulator, it solves the problem of cutoff altogether, and the mosfet/zener will source current at all times when the screens are activated.

edit: In fact, I can actually cut the entire number of mosfets need by half, now that I think about it. I don't know why I didn't see it's simplicity before, but I wouldn't have realized it as quickly if it weren't for you bringing up the point about the zeners. Here is the revised and simplified design which functions exactly the same:

[Helmholtz]

In fact, by tying both tubes a single mosfet regulator, it solves the problem of cutoff altogether, and the mosfet/zener will source current at all times when the screens are activated.

That's what I meant.

[Helmholtz]

The cap across the Mosfet introduces a time delay, meaning the tubes will see full B+ at turn-on.

[Greg Robinson]

If I were you I'd skip the cathode fet, it may cause problems with Vhk leaving the cathodes floating anyway. Rather, just add a resistor from B+ to the amplified zener to keep it conducting a couple of milliamps when the screens are shut off.

EDIT: please ignore my statement about Vhk, I didn't interpret the schematic correctly.

[Helmholtz]

If I were you I'd skip the cathode fet, it may cause problems with Vhk leaving the cathodes floating anyway. Rather, just add a resistor from B+ to the amplified zener to keep it conducting a couple of milliamps when the screens are shut off.

I like your proposal of keeping the zener conducting all the time - for the reason I posted above.

But I don't see how the cathode voltage could get higher than the zener voltage (+Vgs). What is the Vhk limit?

[Greg Robinson]

I like your proposal of keeping the zener conducting all the time - for the reason I posted above.

But I don't see how the cathode voltage could get higher than the zener voltage (+Vgs). What is the Vhk limit?

Ah of course, I looked it over too quickly and completely mis-interpreted the amplified zener as another switch like the screen nodes, big forehead slap moment. SoulFetish, please ignore my nonsense comment about Vhk above.

[SoulFetish]

Ah of course, I looked it over too quickly and completely mis-interpreted the amplified zener as another switch like the screen nodes, big forehead slap moment. SoulFetish, please ignore my nonsense comment about Vhk above.

Yeah, no problem. I actually appreciate this kind of dialog, even when mistakes are made. I'm an experimentalist by nature and this kind of feedback helps me learn. Plus I almost always come away with either a better design, or a better understanding determine if it ends up being a sensible design at all.

[vintagekiki]

Variable Voltage G2-Control

https://www.tube-town.net/cms/?DIY/Amp-Tools/VoCo_Variable_Voltage_Control-EN