I never really messed with screen resistors much back in the old days nothing ever had them hardly or very small values... 470R is still pretty common and what I'm use to....but I went up and tried 1.5k like WOW a lot more compression/tighter/focused sound it does reduce power maybe why in old days they were just maximizing power? that and better tube quality....I like it I understand thats more ideal for EL34s because of excessive screen current? hard on the 470Rs? 470R is more of 6v6/6l6 area? because of less current? I don't know just trying different things...
most British amps with EL34 usually had 1K screen resistor and Fender had 470 ohm(6L6/6V6/5881) I may have to try 1.5K to see how it sounds...
Regards, Elwood
Regards, Elwood
Higher values are likely to give more distortion and less power. In hi-fi the screen stopper resistor is best regarded as a necessary evil.
Its alway best that the screen Voltage ( G2 ) is always lower than your plate voltage when the amp is operational.
to calculate the resistor you have to know the plate Ma's and the output transformers dc resistance and the screens ma
normaly a 1 k resistor wil do in pentode connection.
in triode connection a 100 Ohms resistor can be used.
Last edited:
Pentode and UL mode allows the screen voltage to go below the anode voltage on signal peaks. In fact it insists on it, unless the pentode has an unusually low screen voltage.
The g2 resistor used in UL and triode mode is rarely calculated. It is essentially a screen stopper. In pentode mode you do need to calculate, but then the OPT DC resistance is not an issue.
The g2 resistor used in UL and triode mode is rarely calculated. It is essentially a screen stopper. In pentode mode you do need to calculate, but then the OPT DC resistance is not an issue.
I have wondered about this. A screen resistor introduces current feedback, much like an anbypassed cathode resistor. Should that not improve distortion (albeit by a possibly negligible amount) ?
Ive heard this before, and I cant quite figure it out. If youve got 300v on the anode, and 250v on the screen, then you can only swing your output voltage 50v before your anode goes below your screen. The datasheets certainly dont show this, with the anode voltage going all the way down to 80v and below. Have I misunderstood?
Hi Merlin, I guess it would increase distortion, simply because you would be dropping your screen voltage at max output, therefore reducing headroom. If you increased the HT supplying the screens, this this would bring the headroom back up.
Last edited:
Hi Merlin,
I could be completely mistaken but I would explain the (measurable) effect of rising distortion by rising Rg2 values in typical power stage operation conditions as follows:
Ia:Ig2 ratio indeed is relatively constant over a considerably range of Ea swing for typical small signal pentode operation (not comming anyway near the Ia->Ig2 takeover area).
Power stage operation is a different thing. On very wide Ea swings, reaching into the takeover area (very low Ea) as often reached when squeezing the utmost from a power pentode, the ratio is not linear anymore, actually the screen can draw considerable more current than the anode.
This nonlinear behaviour will lead to nonlinear voltage drop across a mere resistor in front of the screen grid, introducing distortion. Even worse, the effect only occurs on one side of the anode voltage swing (towards low Ea and high Ia) but not when the anode swings in the opposite direction (high Ea, low Ia).
Regards Tom
I could be completely mistaken but I would explain the (measurable) effect of rising distortion by rising Rg2 values in typical power stage operation conditions as follows:
Ia:Ig2 ratio indeed is relatively constant over a considerably range of Ea swing for typical small signal pentode operation (not comming anyway near the Ia->Ig2 takeover area).
Power stage operation is a different thing. On very wide Ea swings, reaching into the takeover area (very low Ea) as often reached when squeezing the utmost from a power pentode, the ratio is not linear anymore, actually the screen can draw considerable more current than the anode.
This nonlinear behaviour will lead to nonlinear voltage drop across a mere resistor in front of the screen grid, introducing distortion. Even worse, the effect only occurs on one side of the anode voltage swing (towards low Ea and high Ia) but not when the anode swings in the opposite direction (high Ea, low Ia).
Regards Tom
Tests done on RF linear amplifiers, which are a reasonable match for audio output stages, show that very low g2 source impedance is needed for low third-order distortion. Resistance in the screen circuit generates re-entrant distortion, as g2 is both an output and an input. You would need a much higher resistor value to get enough useful negative feedback to suppress the distortion, but that would reduce output power significantly too. So the general rule is that UL and triode mode should have as small a g2 resistor as possible to maintain stability. Pentode mode is different, as there the screen is decoupled by a big capacitor.
Well just with the power tube quality these days....probably not a bad idea to have the screen resistors obviously... higher values might not be a bad idea either....
I have a pair of Menno's VDV70/100 boards with the recommended Plitron toroidal power and output transformers.
The output tubes are a quad of EL34. A screen resistor (on each EL34) of 150 Ohms was fine for Pentode mode and Triode Mode but the things oscillated in Ultralinear Mode (on the Oscilloscope it was observed as a burst of RF just after the turnover of a sinewave peak, quite possibly as one of the push pull sides reached or came out of cutoff - BUT that is a WAG). Taking the screen resistors up to 1K fixed that. So YES, in UL Mode it is the Screen Stop function which is important rather than the screen ptotection function. I did'nt try other values because the 1K is what is RECOMMENDED on the Mullard Datasheet (see Distributed Load Conditions)
http://www.mif.pg.gda.pl/homepages/frank/sheets/129/e/EL34.pdf
Counter-intuitively it seems that the better the OT the more the screen stops are required or at least the larger the value required.
It is also significant to note that the Telefunken EL34 Datasheet recommends a common 470R Screen resistor for push Pull pentode mode for operating voltages up to 375V and a common 1K for 400V and above.
ASIDE: The common screen resistor is a cheap but reasonably effective way to stabilize the g2 voltage (in an amp which is not overdriven). The average screen current from the 2 tubes is reasonably constant (as one side pulls more screen current on one "half cycle" the other pulls less and then they swap for the signal excursion in the other direction).
The Guitar Amp gurus all say 1K minimum for EL34 and many also recommend pulling the "traditional" 470R 2W resistor on the 6L6 in Fenders and replaceing them with 1K.
This IS a tube protection thing - when overdriving the tube the screen current goes up significantly.
Larger INDIVIDUAL screen resistors impart some compression which in a Git. Amp is not necessarily a bad thing but I have found that for HiFi smaller screen resistors impart a tighter sound.
My absolutely favourite Guitar Amp is a DIY own design and build thing - It runs EL34 (Actually Philips 6CA7) in Pentode Mode, Cathode biased (individual resistors and bypass caps for each tube) with a COMMON 470R Screen Resistor and a B+ of 380V. It puts out 32 glorious watts instead of the 50 to 55W that most Git. Amp manufacturers normally try to squeeze out of a pair of EL34.
My 6V6G U/L HIFI Amp runs 150R screen resistors which sounded better (tighter) than the 270R I started with.
My pot shots at the barn door - FWIW
Cheers,
Ian
The output tubes are a quad of EL34. A screen resistor (on each EL34) of 150 Ohms was fine for Pentode mode and Triode Mode but the things oscillated in Ultralinear Mode (on the Oscilloscope it was observed as a burst of RF just after the turnover of a sinewave peak, quite possibly as one of the push pull sides reached or came out of cutoff - BUT that is a WAG). Taking the screen resistors up to 1K fixed that. So YES, in UL Mode it is the Screen Stop function which is important rather than the screen ptotection function. I did'nt try other values because the 1K is what is RECOMMENDED on the Mullard Datasheet (see Distributed Load Conditions)
http://www.mif.pg.gda.pl/homepages/frank/sheets/129/e/EL34.pdf
Counter-intuitively it seems that the better the OT the more the screen stops are required or at least the larger the value required.
It is also significant to note that the Telefunken EL34 Datasheet recommends a common 470R Screen resistor for push Pull pentode mode for operating voltages up to 375V and a common 1K for 400V and above.
ASIDE: The common screen resistor is a cheap but reasonably effective way to stabilize the g2 voltage (in an amp which is not overdriven). The average screen current from the 2 tubes is reasonably constant (as one side pulls more screen current on one "half cycle" the other pulls less and then they swap for the signal excursion in the other direction).
The Guitar Amp gurus all say 1K minimum for EL34 and many also recommend pulling the "traditional" 470R 2W resistor on the 6L6 in Fenders and replaceing them with 1K.
This IS a tube protection thing - when overdriving the tube the screen current goes up significantly.
Larger INDIVIDUAL screen resistors impart some compression which in a Git. Amp is not necessarily a bad thing but I have found that for HiFi smaller screen resistors impart a tighter sound.
My absolutely favourite Guitar Amp is a DIY own design and build thing - It runs EL34 (Actually Philips 6CA7) in Pentode Mode, Cathode biased (individual resistors and bypass caps for each tube) with a COMMON 470R Screen Resistor and a B+ of 380V. It puts out 32 glorious watts instead of the 50 to 55W that most Git. Amp manufacturers normally try to squeeze out of a pair of EL34.
My 6V6G U/L HIFI Amp runs 150R screen resistors which sounded better (tighter) than the 270R I started with.
My pot shots at the barn door - FWIW
Cheers,
Ian
Last edited:
If you are tube rolling or switching Triode/Pentode/UL
Then 1K is a must fit...Just my thoughts.
Flash over is a possible maybe should you get grid current.
Just my thoughts..😕😀
Regards the Git.amp ideas..these amps were given a real battering and hard use. Some of the ideas for protecting tubes are worth implementing.
Regards
M. Gregg
Then 1K is a must fit...Just my thoughts.
Flash over is a possible maybe should you get grid current.
Just my thoughts..😕😀
Regards the Git.amp ideas..these amps were given a real battering and hard use. Some of the ideas for protecting tubes are worth implementing.
Regards
M. Gregg
- Status
- Not open for further replies.