6BQ5 Screen Voltage Regulator

[Troncones]

Yes I'm still working on the SE 6BQ5 amp. I should have it running an a couple of days.

I thought this may be a good time to look into how to regulate the screen voltage on Pentodes. The power supply I modeled in PSUD II looks stable but I will know more once I run the amp.

Is there a simple (something a Newbie would understand) solid state way to regulate screen voltage?

Would it help the performance?

[mastodon]

A simple zener (stack) with a resistor should do fine for g2 regulation. Lots of info available, see e.g. here: Zener Diode Voltage Regulator - Electric Circuit

Or use the 6BQ5's in triode mode, but I assume you already have your OPT's, which might get you into trouble if you decide now to change the required load impedance

[toprepairman]

Why bother, it won't make a useful difference.

[SY]

A quick examination of tube characteristics shows that regulation makes a VERY big difference in pentode circuits.

I use a simple Maida reg, very inexpensive, excellent regulation, and more than quiet enough for this application. You can swipe the reg design right out of the Red Light District amplifier.

[DF96]

A pentode screen is both an input and an output. If there is significant impedance there, the output current will generate a voltage which will then be fed in again. This would simply be a source of harmless negative feedback if a pentode was linear. Pentodes, like all valves, are not linear but have a lot of second-order distortion. This second-order, if fed back into g2, mixes with the input signal to create third-order (etc.). In most cases this extra 3rd will be in phase with the intrinsic 3rd produced by the pentode anyway.

Class A SE might not suffer from this effect too much, as at least the pentode will be biased in a reasonably linear part of its characteristic. It gets serious for P-P Class AB/B. Similar issues arise for tetrodes, and tests on RF linears have shown that g2 needs to be fed from a very low impedance: a string of zeners is not good enough, although still popular.

[FLT]

Originally Posted by DF96

: a string of zeners is not good enough, although still popular.

Putting an Transistor on top of the zeners helps(Mosfet preferred).

[DF96]

How? The problem is the dynamic resistance of the zeners. The exact voltage at g2 does not matter too much, but it must not vary at audio or envelope frequencies.

[tubelab.com]

Putting an Transistor on top of the zeners helps(Mosfet preferred).

I use a mosfet source follower biased by a zener or gas tube string. A pot can be used to allow for voltage adjustment. One of my recent projects uses 8 6BQ5's. I use a single zener string feeding 8 seperate mosfets. There are 8 pots tied across the top zener, each feeding a seperate mosfet, which feeds its own screen grid. It is a simple but very effective circuit that allows for individual tube current tweaks since the cathode voltages are all fixed at 12.5 volts.

[DF96]

Yes, a zener feeding a follower solves the problem. A fat capacitor might do instead, as the main issue is dynamic impedance not exact bias voltage.

[toprepairman]

Tubelab, that's a very nice little trick for multiple tubes, I like it.
I suspect that the original request was for just a simple single SE, so in the interests of keeping it simple, a resistor and decoupling cap will be perfectly sufficient (or just tie the screen to supply).