600V bench variable supply to test transmitting valves

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Tired of my HT (+600V) variable bench power supply to suffer collateral damage when accidentally shorted whilst testing transmitting valves for output stage (i.e. FQP3n80c MOSFET passive regulator blowing out), I decided to make a simple but effective valve stabiliser. As nothing comes for free, here are my design constrain factors:
1. Input raw supply is +620V @ 100mA
2. Filament secondary winding is 15V @1.5A
3. No additional secondary winding is available for a floating screen supply (e.g. pass valve is pentode)
4. Output voltage ideally should be 0-600V

So with the restriction of not using a pentode as pass valve, I looked out for candidates to match my requirements and instantly thought about GU-50 in triode-strapped mode. Yes, I know that UG2 limit is 250V, not 1,000V as anode max voltage. But, in triode strapped specs are not shown.

I recently checked this with the 814 triode strapped, and seems to be ok UG2=Ua in triode mode. 7N7 also said this was ok and Morgan Jones previously tested this as well with similar valves.

So, question here is: can the GU-50 withstand 600V in triode mode or do I need to look out for options?

What about adding a zener string across G2 and cathode with a resistor between G2 and anode? Not ideal, but perhaps I can get a compromise solution that will allow me to get 0-600V swing here.

Thoughts?
Cheers,
Ale
 

Attachments

  • voltage stabiliser.jpg
    voltage stabiliser.jpg
    70.4 KB · Views: 858
Member
Joined 2009
Paid Member
I would recommend using an actual error amplifier rather than just a cathode follower for the supply. Unless the 620 V is regulated, you won't have much ripple rejection and the output voltage of your circuit will vary depending on the incoming line voltage.

Obviously, this doesn't answer your question, but I thought I'd point it out.

~Tom
 
Hey Tom,
Thanks, yes I know the rejection of this stage is worse than a simple RC stage as Merlin Blencowe clearly pointed out in his book.
Raw supply is well regulated. I'm trying to reuse what I have here, otherwise I should be looking for a bigger chassis and a different power transformer. I know that the floating screen supply is the best choice in this topology, but that will require an additional transformer and I don;t have the space to fit it in here...

What about an LND150 cascoded CCS feeding a zener string to ensure G2 stays below 250V? This should only kick in when output voltage is below 350V..
I may need to build this and try out whether the sturdy GU-50 have the stones to survive the 600V in triode mode!
 
Hi Ale,
Why go through the frustration of using a regulated HV supply only to have it damaged by a gassy, arcing or shorted tube? I have been testing high power tubes for many years and learned that regulation just isn't needed. The plate supply in the rig pictured below is a very simple LC combination plus a bleeder. Nothing is regulated, but everything is metered. This setup is good for 1500 watts input class A, the limit of my mains. A 1KHz 1v sine wave is amplified for Gm measurment.

In order to have a tube regulated supply supply adjustable between 0 to 600 volts, you will need a variac in the primary of the plate transformer to limit voltage drop and dissapation across the pass tube(s). Usually this is about 150 to 200 volts. The small voltage adjust pot can be mechanically coupled to the variac to rotate with it. This is the way Kepco, Lambda, HP and others did it.
 

Attachments

  • tester.JPG
    tester.JPG
    116.2 KB · Views: 798
Very interesting and wasn't aware of this!
Not sure if I understood your suggestion. From what I read, you are suggesting a variac in the primary, classic LC filtering stage in secondary and the voltage setting pot mechanically coupled with the variac?
Thanks
Ale

Thats how I do it. Have a 1250V PSU connected to a high power variac, so I get from 0 to 1250V 500mA, easy and rugged.
 
Administrator
Joined 2004
Paid Member
I think you also need a supply of negative grid bias if you want to be able to cut off the tube fully. Possibly you could use a small second transformer for the required filament voltage and use a tripler on the 15V to get the required grid bias. The variac idea on the primary side of a plate transformer is an excellent idea to limit power dissipation in the pass tube(s) but would require a dedicated transformer providing only plate voltage.
 
Administrator
Joined 2004
Paid Member
For that I have another psu (an old Farnell PSU 0-350V with a EL34)

I meant in the B+ supply the OP designed, I also used a separate supply for controlling the grid bias when I tested high voltage transmitting tubes. I used an Isco electro-phoresis supply to provide B+ of up to 1.5kV. Recently gave the set up to a knowledgeable friend as it took up too much space..
 
What about solid state regulation

something like this might work (correct me if im wrong)

An externally hosted image should be here but it was not working when we last tested it.


opamp goes High. transistor turns on. cuts the tube of.
opamp goes low tube conducts fully.

The reference depends on divider ratio but could be powered from a RC Filtered PWM signal omitting a Dac

For a current limiter.
You could use a low side current shunt and compare that voltage against the reference. (could also be a dac) and Pull the inverting input of the first opamp Low. this will make the output high turning the transistor fully On . and cutting Of the pass tube.

Heck if it works you could use this to make LF Sawtooth shapes for a curve tracer. (whit current limiting)


V4lve.
 
I didn't consider the fact that you can run out of voltage headroom for low voltages:confused:

You might be able to fully cut off whit a negative voltage of say, 100 volts. But it will complicate matters as you will have to find a way to Turn on a transistor that is below ground

and you will need to find a transistor that can operate at High voltages i remember there are some TV darlingtons that could be up to the task..

v4lve
 
All you guys have lost track of the original post and problem. When testing transmitting and large power tubes, especially of unknown quality, they can and do flash or arc sometimes for various reasons. Transistors, opamps or any solid state componets will not tolerate this. This was mogliaa's original problem. Tubes are very forgiving, transistors are not. They will not serve well in this application. One may think they are somewhat isolated by the series pass tube itself, but that is a bad assumption. In this particular instance sand should be avoided.

Mogliaa,
I strongly recommend not bothering with a regulated HV supply to test a transmitting tube. It just isn't necessary if you monitor and adjust levels as you test. Designing a wide range DC regulated tube supply is not a simple task for most of us unless one is a seasoned engineer. I'm not, but can tell you that what you originally proposed will not serve you well. While you may get some voltage adjustment, you will not go to zero. Like Kevin said you will need a negative voltage source well below common to fully cut off the pass tube.

Also, to achieve regulation, you will need an error amplifier stage that is referenced to a stable voltage. Passive feedback componets will not provide the necessary control. A high gain feedback stage, either single ended or differential, can be used. I don't like zeners as a reference because they drift as they heat. I prefer gaseous diode reference tubes like the 5651, 0G3 or the more common 0D3s and etc.

You asked about connecting the voltage control to the variac in the transformer primary. Yes, the voltage adjust potentiometer should be ganged with the variac. You will need a dedicated plate transformer for this. This will allow you to use a common series pass tube like the 6336, 6C33C or a pair of 6080s to supply 100 mA and not exceed the max plate/cathode level of around 200 volts.

I use a simple LC filter in my plate supply. But I do this for a specific reason, and I'm not regulating. I didn't necessarily mean for you to use this. I think a good way for you to go if you must regulate is to copy an existing circuit from a commercial power supply. Perhaps from a piece of HP equipment or many others.

Victor
 
Last edited:
Administrator
Joined 2004
Paid Member
Any good surplus electro-phoresis supply should handle faulty tubes with aplomb.. I used an Isco 494 picked up cheaply on eBay for precisely this reason. (all solid state) Programmable voltage, current and power clamping - with extremely effective protection against load faults including gas discharge, and random arcing. Scary to watch but supply was unfazed by it. There may be equivalent surplus ones available cheaply in the EU.. Otherwise the variac based supply seems a good idea - no off the cuff design with solid state devices is likely to survive this level of abuse.
 
Yes, if you don't ask it to dissipate more than 40W constantly, or more than 100W during couple of seconds.

Thanks all for suggestions. As Victor pointed out, I've tried sand there and without a more elaborated circuit, there is no chance it will survive. Also, don't have the time (and patience) now to build many versions of Mosfets regulators and refine the circuit until it can survive a short at these voltage levels.
@Anatoliy: I tested 814 in this mode and survived operating in triode mode above max g2 voltage stated for tetrode mode. I was suspecting that the sturdy GU-50 should be able to cope with this, so great news!

A Variac will probably be the best option here. In the meantime, without adding any additional transformers I can get away with it with the GU-50 in triode mode. I may get an output swing between 85V and 510V @75mA to stay within the GU-50 Pd limits (i.e. 40W in continuous operation). I have everything to fit the valve here in and with a bunch of additional passive components I have this bench supply up and running. Yes, it won't go down 85V, but generally most of the tests I do in the bench are above 100V. It won't allow me to bring any circuit slowly from 0V, I know.
I will need to fit an additional transformer and rectifier to provide a negative 100V. In this case I will be able to provide 0 to 530V easily. I don't think I can use the filament winding for this supply as it has to be referenced to cathode.

Probably what I should do next is to buy a variac for this supply :)
thanks
Ale
 
Combine the opamp idea with a zero bias tube like an 811A. Handles 45 watts continuous and 65 intermittent.

awsome idea. and if you take +-18v rails you could Tie the transistor to the negative rail and pull the grid below common. A 811A will likely cut off pretty good. @600v But it might pull grid current at higher output voltages.

Ive heard chinese 811A's have trouble going into full cutoff could this be an issue ?

V4lve
 
Transistors, opamps or any solid state componets will not tolerate this.....In this particular instance sand should be avoided.

I believed this too......until I stumbled across these mosfets on the clearance page at Allied Electronics a couple of years ago. With a BIG (250 watt) 15 ohm resistor in the drain lead they will not blow into a direct short with 600 volts of supply. Of course they will be over dissipated in short order so a foldback circuit or a fuse is needed. For those who don't want to look through the data sheet, the important numbers are 600 volts, 600 WATTS, 43 AMPS of continuous drain current and 172 amps pulsed. They were clearance priced at $5 each when I got some. I don't know if they are still available. I haven't blown one yet.

I made a simple power supply using a large variac feeding a reverse wired 480VAC (2 X 240) to 120 VAC industrial control transformer rated at 1 KVA. There was a solid state FWB on each 240 VAC secondary making two 0-310 VDC variable supplies that could be wired in series or parallel. A smaller 120 VAC to 24 VAC industrial transformer with a small variac provided variable heater power. I used this for years, but there was no current limiting and several experiments ended in smoking or exploded parts.

I tried several post regulators that usually blew up to protect my experiment. That includes sweep tube pass devices. A huge current spike created by a momentary short takes out the metal link inside the tube that connects the cathode to the base pin. The Fuji fet survivied. A dead short across the output at maximum voltage would trip the bench breaker before blowing the fet (about 2 seconds). 60 amp peak currents will stil fry parts. That power supply was large and heavy, nearly 100 pounds. It has been dismantled.

I got a HP 0 to 600 volt 0 to 1.5 amp supply for big amp experiments. I am using that today, but it's current limiter is too slow to save parts. In fact this power supply has created some serious parts explosions.
 

Attachments

  • 2SK3681-01.pdf
    47.2 KB · Views: 192
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.