470R vs 1k screens?

[iknowjohnny]

Just wondering...i never used anything lower than 1k for screens on my EL34 builds, tho i have tried the other direction as high as 5k and i always felt the larger i tried the less i liked it. Conversely i then wonder whether i'd feel any difference between my current 1k's and 470R. I noticed the 470's in that hi gain build in another thread and it got me to thinking maybe i should try 470R's. And thoughts on whether i'd noticed any difference at all and is so what? Seems most use 1k so i'm wondering why a few go with 470R.

[Andy Le Blanc]

Measure the difference, and use your ears, I found that the g2, +/0/-
can have a large impact with a small voltage change, for the positive...
I suspect that its the voltage difference between g2 and plate, and not the
value of the screen grid resistor that may be important.

[Super_Reverb]

Just wondering...i never used anything lower than 1k for screens on my EL34 builds, tho i have tried the other direction as high as 5k and i always felt the larger i tried the less i liked it.

If you look at JJ and Svetlana EL34 data, their max screen grid voltage is 450-500V with max power dissipation at 8W each, so...

For your EL34s, you can calculate max. screen grid current or measure your actual current by voltage drop across known resistor value and datasheet tells you max screen voltage. Data on your specific tube is better.

Other than that, the sound that you hear, power output, and tube longevity (more screen current stresses g2) should determine your component values. As Andy said, experiemnt and find the value that makes it sound the sweetest.


rob

[Gaz]

If you're driving EL34s hard at anything higher than 400vdc on the screens, you're looking for big trouble. No doubt, you're screens will glow, and melt down eventually. Try it for your self though, and make sure you really let 'em cook - they'll last for a little while . You should see a glowing with the signal along with the notes from inside the plate structure.

If you're operating at low voltages, or simply running the PA clean (with a pre-PI MV, for example), the you'd probably be okay.

This has been my experience from new production EL34s. 6L6s have less screen draw, and a different construction, and in my experience, take higher screen voltages with ease.

Anyway, you shouldn't listen to me, but rather measure the voltage drop across the screen grid resistor under max load, and solve for the screen current draw. IME, even new EL34s can dissipate 10W before glowing, but anything over that will burn 'em up!

I wish I could live life with no screen grid resistors, but the $$$ isn't worth the tone.

[iknowjohnny]

I'll pass then. I was just wondering, but the schematic was indeed showing 6L6's so i guess thats why 470 instead of 1k. That was mainly what i was curious about.

[David Root]

As Tony put it. "EL34s have wimpy screens" and I would bet current production is wimpier than NOS.

I am thinking about a #183 type build so this is in my mind. The standard old school screen resistor for EL34s is a 5K6 sandbox (cement) 5W. I would think this would protect the tube even at 500V screen voltage (recall the old 100W Marshalls with 560Vp and Vs).

At least 1K/3W, up to 5K6, again depending on the screen voltage. Tone and feel is better to most people at the lower end of that resistance range. If you are using a Music Man power supply with only 350Vs you might even get away with 470R but I wouldn't bet the farm on that.

An alternative would be 6CA7s (see especially Tony's JJ 6CA7 review thread), these do not have wimpy screens!

[Alexo]

Really it depends on your load.

With a small load, 4K or so, your screens won't get hit too hard when you crank it up, but if you're approaching a 6K6 load, or even the 5K range, setting your EL34 screens at 400+ will cause them to draw a LOT of current. I think the 1K 25W R, in conjunction with the 6K6 load, in the TW contributes a lot to the singing tone of that amp because of this ...not that I've built one, mind you.

I was tooling around with this sort of thing last night, and at 6K, 400 plate, 400 screens, the screens went up to 30 ma (12watts) apiece when the amp was pushed into square wave territory. Upping the screen resistors caused a larger voltage drop as current increased, shifting the knee of the grid curve down and reducing the amount of current they would draw. Basically as you traverse your load line, you reach a point where the plates approach 0V and the high voltage screens look mighty tasty to the electrons jumping off your cathodes. They flock to the screens, treating them like the plates, which is something they were not designed to handle. But if you fool around enough with the screen R value, you can find a value that will drop their voltage low enough when they start drawing current that they will keep your plate voltage from swinging super-low, while at the same time look less appealing to the elctron demons.

The downside is that if this value is too large, you will end up with a LOT of compression in the PA (not necessarily a bad thing, but you lose power and punch) but as the screen voltage drops, your bias effectively cools off as well, so with this form of compression comes increased crossover distortion. Again, not necessarily a bad thing, but not necessarily a good thing either.

...Just as an example, running 7591a's at 425 plates and screens into 5K gave me 36 watts clean, but the screens would light up the room when I pushed the amp too hard. Once I figured out a screen R value that fixed things to the point that my screens weren't dying, I found the amp only put out 25 clean watts. :l 6L6's in the same amp didn't draw as much screen current, so the screens didn't drop as much, and they yield about 30W clean. 7591's & EL34's are true pentodes, while 6L6's and 6CA7's are good ol' BPT's and hence draw less screen current.

[David Root]

Alexo, yes, that is an important parameter too. I was thinking only of "standard" EL34 primaries, 3K4 or 1K7 for 100/50 W. Which as you noted are low enough to have some benefit when you boost the screen voltage.

In your example using 7591As, if you boosted the plate voltage ONLY to say 500V from 400, you could get back most if not all of that lost power I think. But that may well entail a tonal change too, which might not be to one's taste. I have not used 7591As or EL34s but that is generally true with beam tetrodes, and so pentodes too I would imagine.

[Alexo]

In your example using 7591As, if you boosted the plate voltage ONLY to say 500V from 400, you could get back most if not all of that lost power I think. But that may well entail a tonal change too, which might not be to one's taste. I have not used 7591As or EL34s but that is generally true with beam tetrodes, and so pentodes too I would imagine.

Hey David, good point, I guess it really comes down to the relationship between plate and screen voltage.

...raising the plate voltage in this amp might constitute a tonal change, but it's already SS rectified, so it's really the power transformer change that I don't want to face. $ $