links to theory on screen grid resistor purpose

[mikepukmel]

Ive searched the blog looking for more details on what a screen grid resistor is supposed to do, more than the obvious "drop the screen grid voltage with respect to the plate".

Reason for looking for more info: Tinkering with franken-amps, how do I know if I did something bad, screen voltage/current draw is "too high", etc.

Thanks.
Tinkerer.

[mikepukmel]

Another related question: Many amps, many vintage Marshalls, Fenders, etc have a choke between the connection to the plates (through the OT primary) and screen grid. As I understand it, the chokes don't drop voltage much, just smooth out ripples. So, in those amps, my guess was screen grid resistors are needed for REASON X whatever that is, because chokes don't drop voltage much?

But a few, e.g. Supro and Trainwrecks, don't use a choke, they have a resistor between the plate and grid nodes in the power supply, which do drop voltage compared to chokes. Is the reason that they didn't put screen grid resistors in e.g. Supro 6420 Thunderbolt, because it didn't have a choke, but a resistor instead, so the screen grid was at lower potential compared to the plate?

[Helmholtz]

Another related question: Many amps, many vintage Marshalls, Fenders, etc have a choke between the connection to the plates (through the OT primary) and screen grid. As I understand it, the chokes don't drop voltage much, just smooth out ripples. So, in those amps, my guess was screen grid resistors are needed for REASON X whatever that is, because chokes don't drop voltage much?

But a few, e.g. Supro and Trainwrecks, don't use a choke, they have a resistor between the plate and grid nodes in the power supply, which do drop voltage compared to chokes. Is the reason that they didn't put screen grid resistors in e.g. Supro 6420 Thunderbolt, because it didn't have a choke, but a resistor instead, so the screen grid was at lower potential compared to the plate?

While there may be a tendency, I don't think there is a general answer to this question as too many variables are involved and its a dynamic problem.

You may find some useful info about measuring/evaluating screen dissipating in the discussions contained here:
https://music-electronics-forum.com/...l=1#post518963

[g1]

This was posted here recently.

[galaxiex]

All this tech stuff is waaaay over my pay grade... but I do know...

KOC recommends 1K to 2.2K screen resistors in his TUT book series.

That's all I got...

[nickb]

I'd say the most important reason for screen grid resistors in the typical guitar amp is to avoid damage to the screen grid as a consequence of dissipating too much power. To small a value offers no protection, too large a value reduces power output and introduces non-linearities.

[Helmholtz]

BTW, screen resistors can also reduce the risk of oscillation.

[dmeek]

Back in the early 80s a local guitarist brought me his Marshall amp for repair. Unrelated to the repair was a modification he had had done called a "Claret Power Control" (patented).
It was basically a variable screen grid supply and he loved it, raved about it and I had to admit it sounded good. Of course, being a guitar player, within a year he got
tired of it and asked me to return his amp to original. Oops, am I being too cynical?

I traced the schematic, though forgot to note the transistor number. Might have been TIP50 Here's the schematic:

https://music-electronics-forum.com/...1&d=1560802637

Just looked for the patent today and waddayaknow:

https://music-electronics-forum.com/...1&d=1560802637

[Gingertube]

The Screen Grid (g2) exhibits gain when a signal is applied to it with respect to the cathode - just like the Control Grid (g1).
The 1st function of the screen resistor is therefore as a grid stop to prevent parasitic oscillation. An absolute minimum value of around 47 Ohms should always be used.
This is why when triode strapping a pentode you will see a small value resistor between screen and anode rather than just a direct connection (usually 100 Ohms to 1KOhms).

The second function leads from the fact that tube current is FAR more dependent on screen to cathode voltage than on anode to cathode voltage.

The HiFi guys will tell you (correctly) that for minimum distortion you should regulate the screen grid supply, well except for those who like Ultralinear connection where feedback voltage is applied to the screen grid.

As anode current peaks then so does screen current.

You can see why this might be so. As anode current goes up the anode voltage dips as more voltage is dropped across the load. The screen grid becomes more positive with respect to cathode than the anode so more of teh electron stream arrives at the screen rather than the anode which is the same as saying screen grid current increases.

A series screen resistor drops screen to cathode voltage on screen current peaks and so imparts some compression. As this is limiting peak screen current it acts to protect the screen from melt down (over dissipating).
KOC also claims that this compression adds "Touch Responsivity" to the amp.

In a push pull pair then individual screen resitors will compress both sides of the waveform.

You will see one other "trick".
In push pull as signal goes positive, one side delivers more current and the other side delivers less, and then these swap for the negative signal swing. The screen grid currents follow this. The TOTAL screen grid current, however, is fairly constant. So some early HiFi designers used a common (470 Ohms for EL34) screen grid resistor which held the screen grids voltage at near REGULATED supply performance.
Dave Reeve copied this in the HIWATT. You don't get any compression from this common screen grid resistor.

I've copied it in a Trainwreck clone with cathode biased EL34 output and I liked it - but at B+ of +380V for 32 Watts out I did'nt need screen protection for the EL34s.

Only ever use as high as 2K2 individual screen grid resistors on one amp. Push pull 6GW8/ECL86 but used individual 1K on many INCLUDING Ultralinear connected EL34 quad output. In this last case (with Plitron super dooper output trannies) the amp oscillated with 150 Ohms screen grid resistors and I had to go to 1K to ensure no parasitic oscillation. Counter intuitively, the better the output transformer (In Ultralinear anyway) the more you needed the grid stop function of screen resistors.

Every Fender which comes across my bench has the 470 Ohm 2W carbon screen resistors replaced with 1K 5W wirewound.

Because at least part of the screen grid resistor's function is as a grid stop then it is important to fit these resistors with the resistor body close as possible to the screen grid pin on the tube socket. Just as you do for Grid 1 gridstops.

So my "pot shots at the barn door". Hope there is something useful in there.

Cheers,
Ian

[g1]

You will see one other "trick".
In push pull as signal goes positive, one side delivers more current and the other side delivers less, and then these swap for the negative signal swing. The screen grid currents follow this. The TOTAL screen grid current, however, is fairly constant. So some early HiFi designers used a common (470 Ohms for EL34) screen grid resistor which held the screen grids voltage at near REGULATED supply performance.
Dave Reeve coppied this in the HIWATT.

Pete Traynor also liked this trick. And his YBA-3C is the only amp I'm aware of that called for a particular voltage across that shared screen resistor to set the bias. So biasing according to screen current, rather than plate current.

[mikepukmel]

Wow, wow, thanks everyone, reading, reading . . .

[Helmholtz]

Every Fender which comes across my bench has the 470 Ohm 2W carbon screen resistors replaced with 1K 5W wirewound.

Because at least part of the screen grid resistor's function is as a grid stop then it is important to fit these resistors with the resistor body close as possible to the screen grid pin on the tube socket. Just as you do for Grid 1 gridstops.

I think one reason to mount grid stoppers close to the grid pin is minimizing the influence of parasitic wiring inductance and coupling in order to avoid self-oscillation (Tube literature seems to be vague on good reasoning.)

As wirewound resistors have some inductance I am not sure if they are the best choice for screen resistors. I would prefer metal oxide types.

[catalin gramada]

http://www.oestex.com/tubes/screens.htm

[Gingertube]

Yes - minimise stray inductance is the goal
AND
Yes - metal oxide resistors would be better.
Cheers,
Ian

[nsubulysses]

http://www.oestex.com/tubes/screens.htm

Great link