links to theory on screen grid resistor purpose

[tubeswell]

When you're done reading Ian's very informative post, just for laughs, get your head around this Sulzer PI (at pages 10-11).

https://www.americanradiohistory.com...ns-1948-08.pdf

(note: some fun was had a few years ago with this on the Hoffman forum: https://el34world.com/Forum/index.ph...4041#msg154041,

but I believe PRR cracked it: http://el34world.com/Forum/index.php...8348#msg268348)

[Helmholtz]

get your head around this Sulzer PI (at pages 10-11).

https://www.americanradiohistory.com...ns-1948-08.pdf

My qualitative explanation of the Sulzer PI (fig. 5) - assuming it works:

As long as idle screen current is not higher than idle plate current and as the plate resistor is larger than the screen resistor, plate potential can be assumed to be lower than screen potential, which makes the screen more attractive to the electrons than the plate.

A positive going control grid signal will produce a decrease of screen potential by increasing screen current. Decreasing screen potential in turn decreases plate current and thus increases plate potential.
Consequently the screen becomes the inverting and the plate the non-inverting output.

I am sure balancing of this circuit will be tricky.

Edit: Just saw fig. 6. It indicates that at idle screen and plate potentials are similar and screen current is actually about twice plate current in the design example. So the first part of my explanation above doesn't apply.

[tubeswell]

It relies on having g3 biased negatively enough to repel electrons back to the screen apparently.
(and it does work, but is tricky to set up, and ‘works’ with high gm tubes better)

[Helmholtz]

It relies on having g3 biased negatively enough to repel electrons back to the screen apparently.
(and it does work, but is tricky to set up, and ‘works’ with high gm tubes better)

Yes, the elevated cathode potential with grounded G3 seems to be essential to make screen current higher than plate current.