Modern HV EL34s?

[Gregg]

I've heard of this being done but curious how exactly to go about it. I would appreciate some guidance, maybe a quick schematic. I'm sure others would as well. Thanks in advance!

There's nothing particular to it. Just add a 5W zener to the screen as per the schematic below (Mesa) or you can use one for each pair of tubes in a 100W amp. It can handle the current of both.

[Chuck H]

With the zener arranged like that it can only reduce voltage by a fixed amount (where a resistance would be variable with current). It can't "regulate" the voltage to below a limit during spikes but it does reduce them by the amount of the zener voltage. I'm not being snarky. I'm just making an observation on what the circuit looks to be doing.

[Helmholtz]

With the zener arranged like that it can only reduce voltage by a fixed amount (where a resistance would be variable with current). It can't "regulate" the voltage to below a limit during spikes but it does reduce them by the amount of the zener voltage. I'm not being snarky. I'm just making an observation on what the circuit looks to be doing.

Not sure if I am getting your point. The zener reduces screen supply voltage by a stiff amount. But the screen resistor (as shown) is still there and allows for all the positive screen current depending effects.

[Chuck H]

Not sure if I am getting your point. The zener reduces screen supply voltage by a stiff amount. But the screen resistor (as shown) is still there and allows for all the positive screen current depending effects.

I got the impression it was implied that use of the zener in this manner would put an upper limit on the screen voltage. That's not what the zener does. That was my only point.

[Helmholtz]

I got the impression it was implied that use of the zener in this manner would put an upper limit on the screen voltage. That's not what the zener does. That was my only point.

The 50V zener reduces the screen supply voltage by around 10%. The effect on amp behaviour (sag induced screen compression, power, distortion) will be insignificant.
Producing the same voltage drop at idle (where screen current is low) via screen resistors would require high Ohm values. Such large value screen resistors strongly influence the amp's power, impedance and dynamics.

[Leo_Gnardo]

Producing the same voltage drop via screen resistors would require high Ohm values at idle. Such large value screen resistors strongly influence the amp's power, impedance and dynamics.

I've tried higher value screen resistors, and my results agree with your observations. Will be going on a zener buying expedition soon. Look out Mouser, here I come!

[Chuck H]

The 50V zener reduces the screen supply voltage by around 10%. The effect on amp behaviour (sag induced screen compression, power, distortion) will be insignificant.
Producing the same voltage drop via screen resistors would require high Ohm values at idle. Such large value screen resistors strongly influence the amp's power, impedance and dynamics.

Yep. Which is why I said:

With the zener arranged like that it can only reduce voltage by a fixed amount (where a resistance would be variable with current).

But that was short strokes. Your more complete explanation is better for the discussion.

[Reader]

There's nothing particular to it. Just add a 5W zener to the screen as per the schematic below (Mesa) or you can use one for each pair of tubes in a 100W amp. It can handle the current of both.

[ATTACH=CONFIG]50830[/ATTACH]

The data at https://www.mouser.com/datasheet/2/5...-55-596455.pdf list 5732 as 6.8 volt half watt zener.
I am trying to understand what am I missing here. Mistake in posted drawing or am I off base?

[Helmholtz]

...and increase screen circuit resistance as much as practical until those two approaches stop the problem.

Sorry, I might have misinterpreted you (surely happens sometimes due to my limited command of this foreign language - but I might not be the only one).
Are you saying that you prefer unusually high value screen resistors over zeners to get rid of ugly clipping artefacts, accepting other collateral performance influences? Would this be the discussion you mean?

[Helmholtz]

The data at https://www.mouser.com/datasheet/2/5...-55-596455.pdf list 5732 as 6.8 volt half watt zener.
I am trying to understand what am I missing here. Mistake in posted drawing or am I off base?

You're right, didn't check.
I was speaking of a 50V/5W zener or a string of zeners giving the same total rating - in the same place as shown in the Mesa schem.

[Chuck H]

Sorry, I might have misinterpreted you (surely happens sometimes due to my limited command of this foreign language - but I might not be the only one).
Are you saying that you prefer unusually high value screen resistors over zeners to get rid of ugly clipping artefacts, accepting other collateral performance influences? Would this be the discussion you mean?

Sometimes, yes. It depends on where the greater detriment to performance is. And I'm not going ridiculously high resistance with the screen grid circuit. I've only done it a couple of times. The last time was in an amp that used a resistor instead of a choke between the plates and screens on the HV rail. For the same voltage drop at the screens when clipping, the zener would still allow the 1k/5W individual screen grids to roast and increasing the rail resistor worked better. And I did end up needing to reduce grid drive on that amp too.

EDIT: Oh, and the change in resistance was to replace the 1k dropper in the rail with a 2.2k.

[Helmholtz]

And I did end up needing to reduce grid drive on that amp too.

Are you speaking of a linear drive reduction or using clipping diodes to prevent blocking?

...increasing the rail resistor worked better.

But the increased rail resistor will also drop preamp supply while individual screen zeners won't.

Did you verify if a rail choke also helps?

[Reader]

You're right, didn't check.
I was speaking of a 50V/5W zener or a string of zeners giving the same total rating - in the same place as shown in the Mesa schem.

OKay but please put up with me as I ask another question. My confusion is that the drawing shows a screen voltage on pin 4 of the V10 6L6 of 454 volts. It also shows a screen voltage on pin 4 of the V12 EL34 of 383 volts. Granted there is a 1K7 more screen resistance and a 6.8 V zener in the feed to the V12 EL34 but that seems not to explain the extra 71 V lower voltage at the V12 EL34 Pin 4. Can you tell me what I am missing? is the sample schematic just bogus?

[Chuck H]

Excellent questions!

Are you speaking of a linear drive reduction or using clipping diodes to prevent blocking?

Linear. These were different amps than the model where I use zeners to prevent grid loading.

But the increased rail resistor will also drop preamp supply while individual screen zeners won't.

Yes, but the following rail resistor was adjusted to give the same idle voltage. But I do know there would still be greater voltage drop in the preamp with more current.

Did you verify if a rail choke also helps?

This is the best question. I have suspected that it could if the right choke was selected for the reasons I mentioned on page 1. But since I'm not at actually at my bench much and I haven't had an amp with a choke around lately I haven't been able to test for this.

[Helmholtz]

OKay but please put up with me as I ask another question. My confusion is that the drawing shows a screen voltage on pin 4 of the V10 6L6 of 454 volts. It also shows a screen voltage on pin 4 of the V12 EL34 of 383 volts. Granted there is a 1K7 more screen resistance and a 6.8 V zener in the feed to the V12 EL34 but that seems not to explain the extra 71 V lower voltage at the V12 EL34 Pin 4. Can you tell me what I am missing? is the sample schematic just bogus?

You're right. The zener type number must be wrong. With a 6.8V zener the EL34 screen current would have to be ridiculously high to achieve 383V at the screens.