Screen resistors - please weigh in

[gregarious]

I read recently read a comment made regarding screen resistors which piqued my interest at the mention of touch sensitivity.

"For both longer tube life & resistor life, and the added bonus of even better touch sensitivity from the amps, every amp I've ever built, repaired or modified, and running 6L6 tubes, has received 1000 ohm 5 Watt screen resistors.

One of the biggest design mistakes that EVERYONE was making in the '50s was that they didn't use any resistors on the screens of the power tubes. This caused constant premature power tube failures.Leo Fender was the first to recognize this and added, eventually, 470 Ohm 1 Watt resistors to the screens. Tube life increased dramatically. But in reality, 470 Ohms was still not enough for the longest tube life,and 1 Watt wasn't enough for long resistor life. I can't tell you how many Fender amps have come across my repair bench with the screen resistors burnt, and they invariably burn the 1K5 Ohm grid stop resistor mounted directly beneath them, a poor idea to place them there in the first place.

Due to the success and popularity of Leo's amps no one ever really questioned that screen resistor value, and the industry has stayed with Leo's values for the most part, with the exception of a handful of actual electronic engineers with college degrees in their hands, and who also are into tube audio."

For me this should read, touch sensitivity with the added bonus of longer tube life.

If seems that HAD was/is of a like view as the schematics / layouts attest (1k Ohm on a 6L6).

I thought this was an interesting exchange:
http://blueguitar.org/new/text/scr_res.pdf

Any insights as to the role / significance screen resistors make is appreciated.

[Ears]

When tubes are driven to the knee of the plate curves and beyond a virtual cathode is formed between the screen and Plate and screen current rises dramatically. They then tend to melt.

The resistor limits the current and also drops screen voltage which also has the effect of compressing the gain of the tube. [Edit: That's where your "touch sensitivity" comes in.]

Terman provides a good explanation, I posted some chapters of his stuff here months ago, in PDF format . I think it included this topic.

[Noel Grassy]

So, if I use 1K2 on my EL34 screens I'm not in danger of hurting anything
physically in my tubes? Reason being is I found the correct looking 5W
brown ceramic Ohmite R's for my current K-60 build but only in this value.
If the tone's turdish I'll scrap them regardless of esthetics. I'll put the called-out value of 1K back in. Just a wee bit of extra "squashment" or compression of the overdriven signal sounds like my cuppa tea.

Thanks for pulling our coat Ears. By becoming a virtual Cathode, has the action of electrons being focused on the Plate come to a halt?

That Terman cat has a street named for him around Stanford University in our little village. (Somewhere I've got that big dark green tome he wrote so many moons ago. I need to find a lotta my books since I moved. pityful)

[Firestorm]

This is a complex topic.

Early Fender amps actually had their output tube screens at higher potential than the plates (and even specified that in the transformer specs). It's pretty much a bad idea, but the reason they did it is sort of interesting. In a pentode or a tetrode, the screen grid is much closer to the cathode and input grid than the plate, so it is essentially the "anode" of the tube (in terms of setting the positive voltage). Because the screen is just a tiny spiral of wire, the electrons from the cathode wind up blowing past it to the plate. Relatively few wind up hitting the screen grid so everything is copacetic (at least with the tubes and voltages Fender used early on).

Some other early amps, like Vox, used common screen dropping resistors instead of individual screen resistors and worked more or less like the old Fenders. Also, not an especially good idea.

Some tubes are very sensitive to screen grid dissipation -- most of the "KTs" and the EL34 (if you take into account the huge difference in screen and plate voltage ratings). And new manufacture tubes seem to be pretty spotty in this regard, so so it seems smart to use something of at least 470R and up to 1K or more for screen resistors.

Ears is spot on when he says this is a source of touch sensitivity -- as more current flows in the tube with the amplitude of the input signal, the larger the screen resistor is, the more voltage it will drop. And since the screen is effectively the anode of the tube, it triggers the compression.

Hi-Fi amps using EL34s sometimes had a completely separate screen supply (derived from another PT tap) so you could run the EL34s with 800V on the plates and only 400V on the screens. I doubt that would sound good for guitar.

[Noel Grassy]

Firestorm, I think that separate screen tapped tranny was called an
'Ultra Linear'? I've got that name rattling around my brainbox but I could be wrong as I've never encountered one in a guitar amp. I think Torres may have spoke of them when I used to visit his shop.

Could you describe the process of "triggering compression"? I'm not savvy enough to grasp that in one soaking. Is this part of the voltage falling off our screen with respect to the current on the grid? Or much differenter?

Thanks you shall have, NG.

[Ears]

Firestorm, I think that separate screen tapped tranny was called an
'Ultra Linear'? I've got that name rattling around my brainbox but I could be wrong as I've never encountered one in a guitar amp. I think Torres may have spoke of them when I used to visit his shop.

Ultralinear usually refers to "distributed load" taps of the OPT.

Could you describe the process of "triggering compression"? I'm not savvy enough to grasp that in one soaking. Is this part of the voltage falling off our screen with respect to the current on the grid? Or much differenter?

Thanks you shall have, NG.

Think of it this way. The electrons are pulled by the screen from the cathode. Most stream past it with enough momentum to carry them to the plate. But what happens when the plate potential falls to a critical point below the screen? The electrons can run out of momentum somewhere midstream between screen and plate, or at least the space charge field they are in dips the other way. If they are still closer to the screen than the plate they will then return to the screen and the screen current rises. At the same time the plate never receives these electrons and its current falls, in other words transconductance of the tube drops.

The total current from cathode is the same but the screen starts gobbling up more of the plate's share.

[Ears]

Continued from previous..

When the screen current rises its resistor causes the screen voltage to fall. Lower screen voltage pulls fewer electrons from the cathode. Its a feedback system. Lower screen voltage works against the condition that caused the screen current to rise. This is how the compression happens at peak levels.

[skyboltone]

Thanks Ears:
I've been advocating lower screen voltage/ higher screen resistors around here for quite some time but really only thinking about the plate and screen being at the same potential causing excess screen current; not looking at the feedback picture at all.

Thanks again for an easily understood explanation of a complex topic.

Dan

[Firestorm]

Firestorm, I think that separate screen tapped tranny was called an
'Ultra Linear'? I've got that name rattling around my brainbox but I could be wrong as I've never encountered one in a guitar amp. I think Torres may have spoke of them when I used to visit his shop.

Ultralinear taps are screen taps on the SAME winding used to supply the plates (usually at around 70% of the winding). You pick up a little power gain, a little better bass response and an OT that is better able to tolerate impedance mismatch.

When you run screens at a significantly lower voltage than your plates, you use a separate winding to supply the lower voltage. Yes, I said "tap." My bad.

All of the Fender "135" amps were ultralinear, and they do work fine for guitar (just not the way Fender wired them).

[Ears]

Thanks Ears:
I've been advocating lower screen voltage/ higher screen resistors around here for quite some time but really only thinking about the plate and screen being at the same potential causing excess screen current; not looking at the feedback picture at all.

Thanks again for an easily understood explanation of a complex topic.

Dan

[It's only my understanding from reading texts and writers in this forum, such as Firestorm. ] Dan, If you advocate lower screen voltages (say 50% to 80% of plate value, depending on the tube type), you will , as a general rule, extend tube life and reliability, but you probably won't observe so much of the compression effect and I suspect that's why so many guitar amp builders continue to run screen voltages close to plate values.

Beam tubes work differently to pentodes and have an effect Terman calls "bottoming", I think in essence the end result is the same as that I have described above but I will need to refresh my memory.

Edit: There is secondary emission from the plate as well. Some Electrons hitting the plate arrive with enough energy to cause electrons to dislodge from the plate, if the dislodged electrons have enough energy to travel far enough toward the higher potential screen they can be pulled toward it instead of falling back to the plate; all contributing to the screen current.

[Noel Grassy]

Thanks Firestorm & Ears for the valuable and clearly illuminated explanation.
As Skybolt said, "For a complex topic". I continue to learn so much
at this website. I appreciate everyone's willingness to share.

[funkmeblue]

mesa/boogies simul-class amps were ultralinear in the class a/b section, as well as the Z route 66

[Wayne]

Since it's "be kind to your screen-grids" week around here, what about the role of the suppressor grid in cases where it has an external connection? (guess that applies only to EL34 - can't think of another tube in common uses that has a G3 pin)

Could the voltage on the suppressor be manipulated somehow to provide better tone/enhanced tube reliability or both?

I have read somewhere (not sure where, I read a lot!) about tying G3 to the bias supply to lighten the load on G2. While this seems like it would work by reducing screen current due to secondary emission at the plate, it seems intuitively like it would put an end to any idea of peak compression.

I could theorize on and on, but it seems like a much better idea to ask this experienced group of builders/modders - has anyone ever played around with this idea, and what did you learn?

W

[mhuss]

That's an old Pete Traynor trick. If you make the suppressor voltage significantly less than the cathode, not much current will flow if the bias fails for any reason, saving your tubes and perhaps your OT. I have never gotten around to experimenting to see what it does (if anything) to tone. Of course, if you use the bias supply to bias the suppressor and the bias supply itself fails, you're still smoked.

I did a lot of experimenting with the screen resistors on a Hiwatt 'JP' clone, and we found the value makes a difference in tine as well as touch. Of course, it sounds best when the screens are glowing like a toaster.

--mark

[Wayne]

You've refreshed my memory! I did read that it was a Traynor trick, although I've never come across a Traynor that was wired that way including my own YBA-1.

I will say again that I never met a Traynor that sounded bad, provided it was working properly.

W