Asking for a solution for screen grids supplying

[catalin gramada]

Hello. I have a new project p-p using el34 pentode mode. Plates 490V with 320V into the screens.
The screen grids are supplied by independent power supply and this makes the life very complicated from a safety point. I mean ramping the plate voltage always before screen grids. ( I had a fault happening with input signal applied when the mains dropped off and reconnected in seconds-the screen grids circuit blowed off ).
So I want to make the screen grids supply circuit dependent by the presence of the voltage at the plate.
Just for now I thinking at one simply solution but I don't see all aspects how reliable could be : I meant using a relay ( no more than 20-30mA ) controlled by plate voltage which comand dc contacts for screen circuit. I have no experiences with relays but I know the engaging voltage are related by an voltage ecart:- say using a 12V relay it have a minimum voltage to engage at 8.5V or so. For sure it exist a minimum voltage to disrupt: it is the same minimum voltage to engage ? How should I choose the relay for my application please ?
The relay solution is the fastest comes to my mind but I like to hear any practical solution I may use to avoid an accident using separate supplies for plate and screens please. Thanks.
The minimum operating voltage for relay is 70% by nominal, that corresponding to a 70 percent raport between plate and the screens which should ensure the screens will be engaged/disengaged at a point the plates voltage will be no less than the screens. Any cons please ? Thanks.

[Mars Amp Repair]

Hello Catalin,
Well, I'm not a designer, but am curious to begin with, why are you designing this with 2 separate supplies ro plate & screens? Perhaps we can better help with that information. Thanx, Glen

[catalin gramada]

Hey. That was my choice to test a circuit with regulated screen grids. I tested series and shunt regulators. Now is time to test with independent supplies. It works very well with 320V into the grids for 6.5W max dissipation on set clip. 175w/8ohm rms from a el34 sextet. 160va screens PT ensures a decent 5Vdc regulation for 100mA dc (30 iddle / 130mA max). I still have to play with the ballast resistor to see what is the minimum current to draw to place it in the best part of the current/voltage transfer for transformer I used. Is a project, but the problem have now is related by ramping voltage. I already toasted the circuit by a mains problem fault.
...well, I know, I should made a winding on existing iron...but was to convenient for me to use an extra PT instead to work on the big transformer I have...At the end I.m not convinced I keep it...but sound promising still...

[Helmholtz]

Just an idea: What about a clamping diode between the 2 supplies that ensures that the screen supply voltage can't get higher than the plate supply?
In normal operation the diode doesn't conduct and if a shorting tube takes out the plate fuse, the screen fuse will also blow.

[J M Fahey]

Hello. I have a new project p-p using el34 pentode mode. Plates 490V with 320V into the screens.
The screen grids are supplied by independent power supply and this makes the life very complicated from a safety point. I mean ramping the plate voltage always before screen grids. ( I had a fault happening with input signal applied when the mains dropped off and reconnected in seconds-the screen grids circuit blowed off ).
So I want to make the screen grids supply circuit dependent by the presence of the voltage at the plate.

In those cases (I feed plates scary high voltages but always pamper the screens) I use a single supply, the main one, and drive screen voltage from it, so screens will NOT have higher voltage than plates ever.

I use a voltage divider from +V to set reference voltage (say 47k>220k to get 400V out of 500V supply) and feed that to a MosFet gate (I use IRF820 simply because I have many, pick a higher voltage one for safety)
Mosfet source to ground 220k, just to provide a discharge path.

To get 320V out of 490V, your divider can be 120k>220k

Both plate and screen supplies will track each other at a fixed ratio, no surprises.

You can add a standard screen resistor from Mosfet Source to Screen grid.

Heatsink the TO220 Mosfet, just bolting it to chassis with proper grease and insulation is fine.

[g1]

Another idea would be to have the standby switch serving only to turn off the screen supply. And a simple yes/no detection to prevent screen supply from running when plate voltage is not present.

[catalin gramada]

Another idea would be to have the standby switch serving only to turn off the screen supply. And a simple yes/no detection to prevent screen supply from running when plate voltage is not present.

I had a stb switch on screens. The problem was it didn't prevent an fault when mains dropped off with signal on input. It blows instantly. Now, I will add protections for such "events" but really don't know if series fusible on screens are quite effective or not ?

[g1]

What are resistance values for S1 - S6 ?

[catalin gramada]

There are 100 ohm/1w. The screens draw 20mA onset clipping at 320V, meant 6.5w power dissipation. I installed a minimal value right to the socket, as preemptive measure against potential hf oscillation.(not sure if really need it)...

[catalin gramada]

...and triggering to work more precisely looks pretty simple task:

https://youtu.be/lfyGiwBMdQ4?si=-msm35927Ze0n9vl

[Helmholtz]

So you don't like my simple diode solution?

[catalin gramada]

There is a problem with actual fuses protections: due to very low impedance circuit, large caps there are way overated to resist at start up. There are mainly there to save the power transformers in case of fault and not to protect the circuits please. There are ineffective as it shows at the last fault when screens and cathode resistors blows without a fuse hint.

[Helmholtz]

There is a problem with actual fuses protections:...

I see, your reservoir caps are huge.
Why don't you wire the secondary fuses after the caps?
The primary fuses should protect the transformers.

[g1]

Using such low value screen resistors at each tube is part of the reason you are looking for protection in the first place. The screen R's normally limit the current and to a certain extent can act as fuses.

[catalin gramada]

Using larger values as current limiters will compromise the whole idea of regulated screens. I will provide series fusible protection for the next, should work but still I don't know if will act really fast as effective circuit protection for screens failure please ?7