voltage divider for screen grids

[Bob Simpson]

Thank you, Andy.

I guess I learn more from diagrams...

Bob

[roberto]

Nice idea! So we can do a pyramid like with two 1k resistors (talking about EL34) between pin 4 and 6, and a 220k between midpoint of two resistors and pin 8: voltage divider plus "screen stopper".
With 450V B+ we'll have around 2mA due to the voltage divider, up to 10mA due to voltage divider plus screen current. 1k is ok around 5W, 220k sould be 2*[(2E-3)^2]*2.2E5 = 2W

Just about different voltages on power tubes, does anyone connect pin 1 (g3) to raw bias voltage (around -60V), à la old Traynor?



EDIT
Does one (or more) 1N4007 just before the 1k screen resistor can achieve the same results?

[Andy Le Blanc]

yes you can "de-couple" with diodes, but the bleeder arrangement gives you control, and that few volts does seem to have a sonic impact.

set up a + 0 - see and listen to the dynamic circuit paying close attention to the difference measured between the plate and screen, and correlate it to bias.


I was playing around with a 12ax7 dressed as a diode, before the screens, when I stumbled over the difference measurement, can't get a good screen voltage reading over a screen grid resistor... right

[roberto]

What do you mean with "gives you control"?

Simple screen resistor:
- with no signal we have few mV voltage drop.
- with full signal we have around 12V drop (10mA on 1k)

Diode drop:
- with no signal we have around 0.7V drop each diode added. Suppose we have three of them, so 2.1V drop.
- with full signal we have around 12V drop (10mA on 1k plus 2.1V)

Voltage divider drop without "screen stopper":
- with no signal we have around 2V drop (it's 1/221 on 450V)
- with full signal we have around 12V drop (10mA on 1k plus 2V)

Voltage divider drop plus "screen stopper":
- with no signal we have around 2V drop (it's 1/221 on 450V)
- with full signal we have around 22V drop (10mA on 2k plus 2V)

[John_P_WI]

Using a big ass rheostat instead of the fixed voltage divider as previously shown could be used as a crude power control / volume control.

[roberto]

Larry "Novosibir" used something similar in its early amps as an output control:

[img:800:533]http://www.larry-amplification.de/briti ... pur091.jpg[/img]


But with a variable voltage to the screens, you need to change the bias voltage too.

[John_P_WI]

But with a variable voltage to the screens, you need to change the bias voltage too.

Yes agreed, that is why I said "crude" power control....

Damn, Larry builds some nice looking stuff.....

[Structo]

OK, please excuse me because I am a visual person and I like pictures.

And although I hate beating something to death, is this what you guys mean?

[roberto]

Instead of diodes we can use leds as clipping indicators, or even better some vactrol used as compressors (more current flowing to screens, more light into vactrol leds, less resistance on the vactrol, less voltage on the output of a voltage divider pilotet by the vactrol, less current flowing to screens).

[ChrisM]

Hey Chris, haven't seen you around these parts in a while.

Hope everything is going well for you.

Thanks for the welcome back!
Been very busy, both in life and also building a ton of amps

OK, please excuse me because I am a visual person and I like pictures.

And although I hate beating something to death, is this what you guys mean?

Yep that's it.
Just putting a voltage divider before the screen resistors to lower the voltage a bit.
You'll probably want to use a 5W resistor.

Another way you can do this (the way I would) would be to place a resistor from each output tubes screen grid (4) to ground. So for a 50W dual 6L6GC amp you'd have two resistors instead of the proposed one. I think this would be better as the resistor is then dissipating less power (heat) and has less strain on it. Also each tube will get a unique screen voltage that is proportional to it's screen current and bias.


Edit: you don't need that 1K 5W resistor you drew in. I believe Andy was just showing that for example. In the ODS pictured the screen supply is derived from the output of the choke.

[roberto]

you don't need that 1K 5W resistor you drew in. I believe Andy was just showing that for example.

And wrote that the amp sounds better with that resistor:

I've even placed 220k bleeders right from the screen grid terminal on the tube socket going to ground,
does the same thing but so far I've like the result of the voltage divider before the screen grid resistors

[ChrisM]

You can do the same thing by altering the value of the screen to ground resistor.

[roberto]

Not only, you have to change the screen resistor too.
Look at my post on the previous page.

[Andy Le Blanc]

It used to be said... "mind the screen grids"

its really is just making sure that screes really are at less than the plate V

you implement a design choice to address and control the screen voltage

I like the simple fix, ends up reducing the voltage by just enough to resolve the issue without jumping the shark, but a fully variable control should be fun

[roberto]

Hi Andy,

What do you mean with "gives you control"?