(deleted)

[matt h]

(deleted)

[martin manning]

The grid could get its ground reference through the following stage's grid leak. In the Designing Preamps book the schematic accompanying this subject Rf has its own dc blocking cap, and it includes a grid leak resistor.

[HeeBGB]

You could ask Merlin but I think he has had enough of TAG. He hasn't posted since "Gerry's Follies".

I am sure someone who can splain will be along shortly. Your question is over my head otherwise I would try to answer it.

[matt h]

(deleted)

[xtian]

You could ask Merlin but I think he has had enough of TAG. He hasn't posted since "Gerry's Follies".

Naw, it's just that we haven't sufficiently tickled his brain since then.

[martin manning]

Yes that could become a problem. Honestly I think the following stage's grid leak R has just been left out. Blencowe borrows from Aiken here, and provides a (dead) link to Aiken's site. There, Aiken states: "Note that the tube gets its grid DC bias reference from Rf and Rl." Rl is the load resistance following the output coupling cap Co. In Aiken's circuit Rl is only 100k, and he also says: "If the sum of the feedback resistor and the load resistor is very large, a 1Meg resistor should be added from the grid of the tube to ground, to keep the grid bias stable."

[matt h]

(deleted)

[sepulchre]

Matt, I think it would be well worth it for you to acquire a copy of Merlin's book. I've got both the first and second editions (though they are embarrassingly dusty at this point). As someone pointed out in a different thread, "You can read his book 500 times and come away with something new each time." That might be a little exaggerated . . . but not much.

[matt h]

(deleted)

[sepulchre]

Ah. Gotcha. Great book.

[pdf64]

My initial thought to this was - 'as the grid is a (imperfect) ac virtual ground, what's the downside of adding a default grid leak resistor, eg 1M, 2M2?'

The AC50/4 MkII has a version of this in its paraphase circuit http://schems.com/manu/vox/ac50_1.pdf

For the more general case, I can see that as the closed loop gain increases so will the actual Vac at the grid (due to the lack of open loop gain), and so a grid leak resistor will compound the problem by further reducing the open loop gain (primarily by creating a potential divider with Rg).
In that case, the grid leak resistor may best be moved to the 'input' side of Rg.
Perhaps the best way forward is to add a design guideline for such circuits, such that the resistance between grid and ground should be limited to ~1M, to be achieved by adding ground reference resistors to the input / output as required?
Pete

[jazbo8]

Aiken has a tutorial for LNFB here. Glad to see his site working again.

[tubeswell]

In the book he has a grid leak (as Martin noted)

[martin manning]

My initial thought to this was - 'as the grid is a (imperfect) ac virtual ground, what's the downside of adding a default grid leak resistor, eg 1M, 2M2?'...

...For the more general case, I can see that as the closed loop gain increases so will the actual Vac at the grid (due to the lack of open loop gain), and so a grid leak resistor will compound the problem by further reducing the open loop gain (primarily by creating a potential divider with Rg).
In that case, the grid leak resistor may best be moved to the 'input' side of Rg.

I don't think the grid fits the definition of a virtual ground... In fact we want it to be quite the opposite, no?

Aiken discusses your second point, and proposes the same solution. See at the link posted by jazbo8, which is where I'm quoting from

[matt h]

Ah-ha! Thanks for the link. Suspicions confirmed all around. Now where did I park my car?