screen resistors

[daz]

It's said higher values give more compression and i've experimented with that in the past and experienced that. But i'm wondering..is it the value of the resistor in and of itself, or is the screen voltage in relation to the plate v? In other words, tho i suppose it can't happen but hypothetical.....if screen v somehow was so much higher than the plate that u needed huge screen resistors to make the screen and plate v the the same, would the result be minimal compression? So in simple terms, is it the difference in screen vs plate voltage that makes for the compression or the resistor size? I have 470R in my 6v6 amp now and screen and plate are the same voltage so i suppose i should go back to 1k but i don't want more compression.

[Helmholtz]

Screen voltage is a kind of second bias. Lower screen voltage causes less plate current and less gain.
The larger the screen resistors the more screen voltage drops when screen current increases with signal.
Result is dynamic (signal dependent) screen compression.
A stiff screen voltage doesn't allow for screen compression.

It is no problem if screen voltage at idle is a few volts higher than plate voltage as long as both screen and plate voltage are not too high for the tube.

As the screen resistors protect the screens from overdissipating at large output, I wouldn't go below 470R.

[mhuss]

Some other food for thought:

Even 470 ohms must be dropping some voltage, unless Ohm's law has been repealed. It's usually small, though, mA across a small resistance. Try (carefully!) measuring the voltage across the screen resistor.

The screen resistor also serves to prevent oscillation and reduces 3rd harmonic distortion slightly as the value increases. Too small a value may cause the screen to overheat (e.g., An EL34 screen can only dissipate 2 watts.)

[Helmholtz]

Try (carefully!) measuring the voltage across the screen resistor.

Screen current at idle typically is rather low, so you won't see much voltage drop across the screen resistors.
But with signal screen current strongly increases because of the higher plate current and because screen voltage is higher than plate voltage most of the signal half-cycle.
That's where the screen resistor needs to drop screen voltage to protect the screens from overdissipating.

[daz]

Well, sounds like 1k would be best. I'll change em. There was a bit less compression with 470R but it wasn't much so better safe than sorry when it gets cranked with band.

[kg]

keep in mind current day tubes rarely have the QC of the golden age, where g2 wires would be precisely aligned to the g1 "shadow." obviously this greatly reduced g2 dissipation and increased longevity.

operating into a lower than expected plate load impedance will greatly reduce screen current, as max plate voltage swing is decreased. the converse is also true--higher load impedance will cause quite a bit more screen current. screen grid current curves skyrocket once they are the most positive element in the envelope while there is heavy cathode current.

feeding the screen from a higher b+ but with larger value series resistance will increase screen compression, not reduce it. the compression comes from the impedance of the supply voltage in a proportional relationship. a lower screen grid supply voltage along with lower series resistors will have less compression, and may also have less g2 dissipation.

hth
ken

[soundmasterg]

operating into a lower than expected plate load impedance will greatly reduce screen current, as max plate voltage swing is decreased. the converse is also true--higher load impedance will cause quite a bit more screen current. screen grid current curves skyrocket once they are the most positive element in the envelope while there is heavy cathode current.

hth
ken

I've experienced the screen compression thing with a Silvertone 1484 that wanted to see a 4 ohm load but was going into an 8 ohm load. I doubt if the screen voltage was ever above the plate voltage in that amp as it is normally a fair bit lower, but the impedance mismatch made that mismatch sound GREAT for leads, but rhythm tones were pretty dirty and midrangy and undefined. I'm actually surprised the output transformer held up for as long as we used to play that amp that way as it is soooo undersized. We always ran it full blast too.

Greg