voltage divider for screen grids

[Andy Le Blanc]

you can measure the difference between the plate and screen...
so... you can use a simple divider, or other means, to set the voltage difference
with a bleeder, the value of the resistor just prior can set the difference.

so from the tap on the power rail that first resistor could be the "set value",

or if say you only needed to shave a few ma. difference off the screens at the bias point you've arrived at or favor for the tube type,
a larger value of bleeder right off the screen lug on the tube socket would do it.

once you observe what the relationship between the plate and screen you have the means to control it.

as opposed to guessing about the value of screen grid resistor, you can set up a series known differences,
any where from the screen being a few ma. positive to how ever many volts negative relative to the plate
and see where your preferences lay as you tone out and play the amp.... control

[pula58]

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)

In the last case, "Voltage divider drop plus screen stopper", I think maybe the thing Andy is finding is because now there is 2K in series with screen back to the choke. So when the screen draws current it is dropping (in voltage) quite a bit (twice) lower than with just the normal 1K screen resistor and a 220k bleeder to ground.

[roberto]

It's easy to check: just use a 2k 220k voltage divider.

[Colossal]

I had a chance to install and spend some time with this mod. With a (high gain modded) Marshall preamp into a 4xEL84 cathode biased power amp with NFB, presence, resonance, PPIMV. I tried both 100R and the 1k screen resistors, each with a 220k divider to ground. 100R gives more clarity but changed the overall quality of the distortion (sounds more like an amp without NFB, more Voxy), 1k darkens the amp considerably but gives significant compression, more like the squish and metallic shimmering grind of a larger pentode. In both cases, the added divider with 220k to ground did seem to enhance the overall clarity and separation while smoothing the distortion slightly (for the better to my ear...for what I am after tonally).

I prefer the sonic qualities of the larger screen resistor but I may try a 1k trimmer in series with 100R to find the best compromise between clarity/dynamics and shimmer/compression. 220k to ground is a keeper though. Sounds good!

[roberto]

Nice to read some hi-gain impressions on EL84 based amps.
Did you noticed some compression too, expecially with the 1k-220k divider?

[Colossal]

Nice to read some hi-gain impressions on EL84 based amps.

I do love high gain, tight grind tone!

Did you noticed some compression too, expecially with the 1k-220k divider?

Very much so. 1k gives quite a dark sound with an almost hall-like depth to it. There is more sustain and the tone sounds further back in the cabinet. 100R was much more focused, up front, and a bit harder and more glassy sounding, if that makes any sense. More like that Voxy/Rocket midrange and a little less bassy. There is a lot of fatness with the higher screen resistor. I prefer the higher value and the added compression. It still sounds very warm and organic, but just fatter perhaps at the expense of some clarity.

[eddie25]

Nice to read some hi-gain impressions on EL84 based amps.

I do love high gain, tight grind tone!

Did you noticed some compression too, expecially with the 1k-220k divider?

Very much so. 1k gives quite a dark sound with an almost hall-like depth to it. There is more sustain and the tone sounds further back in the cabinet. 100R was much more focused, up front, and a bit harder and more glassy sounding, if that makes any sense. More like that Voxy/Rocket midrange and a little less bassy. There is a lot of fatness with the higher screen resistor. I prefer the higher value and the added compression. It still sounds very warm and organic, but just fatter perhaps at the expense of some clarity.

Did you use the extra resistor in series to to the screens, or just the 220k bleeder and the standard screen resistor?

[Colossal]

Did you use the extra resistor in series to to the screens, or just the 220k bleeder and the standard screen resistor?

I did not use the extra 1k, just a screen resistor with 220k to ground off each one. I am taking the screen voltage right off the second node, no additional dropping resistor in series with B+ before supplying the screens.

[roberto]

So a 220k 1W from pin9 to pin3 of the EL84's tube socket, isn't it?

[eddie25]

Did you use the extra resistor in series to to the screens, or just the 220k bleeder and the standard screen resistor?

I did not use the extra 1k, just a screen resistor with 220k to ground off each one. I am taking the screen voltage right off the second node, no additional dropping resistor in series with B+ before supplying the screens.

Thanks. I am curious to know what you heard from just the 220k comparatively, besides what you heard from the different screen resistor values...

[Colossal]

Thanks. I am curious to know what you heard from just the 220k comparatively, besides what you heard from the different screen resistor values...

To my ear, I notice a bit more smoothness and clarity under heavy distortion with just the addition of the divider. I tested this first with the stock 1k screens so as not to bias my ear before making any other changes. I then changed the 1k to 100R just to hear the effect. The amp is already very warm and is set up for high gain, wide/fat distortion. But I do notice perhaps a bit more clarity with the added 220k resistor to ground. I prefer 1k to 100R screens (on this amp...not all), however 1k may be a bit much and I may experiment with finding a happy medium between clarity and compression.

[Andy Le Blanc]

glad youre getting around to playing with it. I was surprised at the tone difference its made for a range of tube types.

Whats the measured difference between the plate and screen with the set up you find you like the most?

[Liquids]

Sorry to ressurect this.

I've been messing with the concept of using a voltage divider for the screen voltage in my amp (Ampeg V4b). The V+ is 520-540+ volts depending upon where I plug it in...and in general, I'd like to reduce the screen voltage.

I tried utilizing the center tap of the standard voltage divider/filter cap arrangement, but for 6L6GC tubes and the cold bias, it was too low for the screens.

I started experimenting with an independent voltage divider that was higher in voltage for the screen, and then tapping off of the 'center' of that for the screen arrangement (cap bypassing the resistor from B+ to 'center' and another bypassing the resistor from 'center' to the amps 'floating' ground). Note that the amp utilizes no choke - just an independent tap off the B+ through a 470/500 ohm resistor and a 20/25uF screen capacitor, to each individual screen resistor.

I was able to swap back and forth between 'standard' configuration and the voltage divider that I seemed to like - the arrangement I tried and liked the most, so far, was a voltage of about 80% of the B+ voltage.

I currently achieved this with a 47k resistor bypassed with a 40uF cap to 'center,' and 2x100k resistors in series from 'center' to ground, bypassed by another 40uF cap. Is the 47k resistance effectively 'not seen' by the screen supply because it has a capacitor across it? And/or, have I effectively created a voltage divider but more so a point where the screens 'see' 2 40uF caps as if in parallel and hence 80uF more filtering?


I don't seem to get sag like I would expect from adding 47k in series to the screens - in fact, things seem to be nice and stiff (suggests the added capacitance), but in my head, the voltage divider must be creating some kind of series resistance for the screens, even though the divider utilized bypass capacitors, no?

Or could the voltage divider scheme effectively be increasing the stiffness of the screen supply due to the added capacitance, and what I am theoretically preferring, audibly is the sound of a stiffer screen supply? It's hard to know as my understanding of this from a technical perspective drops off in this area.

So with that said - as much as I prefer the idea and sound of the '80% voltage' for the screens, I'm not sure I've isolated the variables in comparing it to the full screen voltage. My other option is doing further experiments once more parts arrive - like simply comparing stiffening up the 'stock' screen supply by adding/removing capacitance to what is there when additional parts arrive, or paralleling resistance to reducing the 47k resistor to ~22k and trying out the sound of '90% screen voltage vs 80%, etc.

Any insights would be appreciated - in parallel.

[Firestorm]

Not to try to out-Andy Andy, but the secondary reason screen grids were added to triodes in the first place was to limit the effect of plate voltage on plate current. There is still an effect, but the higher the screen voltage is, the lower the plate voltage can be pulled before the plates stop conducting.

Addendum: I didn't really address the question, did I?

When the screens are tied to the plate, you have a pseudo-triode where plate voltage directly affects plate current; when they're tied to fixed V+, you have a textbook tetrode or pentode (where plate voltage and current are (semi) independent); and if you tie the screens to a tap on the OT... well, you know the drill.

But in most of the amps we seem to like to talk about (the ones without a choke between plate and screen- actually some others if they drive the power supply hard), the screens are run from a very marginally reduced B+2 and let the screens overdissipate for part of the cycle.

These tubes were designed for radio-use at frequencies we are not interested in so if we misbehave with them at AF then we become Leo Fender (God bless) and might find that next thing.

[tubeswell]

The thing I was wondering about using a voltage divider to set screen voltages in an output stage is:

Allowing for the required amount screen current, wouldn't you need to set up the resistances in the divider so that they pass a relatively high amount of current through the divider - such that any changes in screen current do not hugely affect the voltage at the knee of the divider so much? In which case wouldn't the resistors in the divider need to be lowish resistance (and high-power) resistors, and in which case would they not use up a whole heap of B+ current?