A Dual Regulated HV DC Power Supply

[KK4YY]

Needing a dual rail HV regulated power supply, and not wanting to buy p-channel mosfets, I put this together and it worked. As can be seen, there are two identical voltage doublers/filters/regulators connected in series and tapped at their junction for the common leg.

A dual secondary transformer is required and I had a Triad N-68X isolation transformer on hand. Wiring it "backwards" did the trick.

I was worried that the negative rail might oscillate, but it didn't. Maybe I'm just lucky.

Most of the components used were items I had in the junk box so the values shown are not specific to the design. The drawing is "as built".

I haven't seen this implementation used anywhere else, so I'm posting it here.


Don

[Opcom]

Very nice! I'm glad to see worked-out linear type MOSFET regulators of one polarity for these voltages. Some things like this were done with tubes back in the old CRT scope days. I appreciate this for tube op-amps. 

[MikeKE0ZUinkcmo]

FETS make these types of projects SO easy.   

I made a few HV and bias, adjustable, regulated supplies so I could run most any tube rig.

[KK4YY]

I built this to power a mosfet audio driver that I've been planning for my DX-100.

I could have taken excising power from the transmitter by bridging the LV power supply, but then I would need a p-channel fet to regulate it because there's no way to work this out with a center-tapped transformer. So, I ventured into the unknown (at least to me) and went with this circuit.

My thought was to let this supply run constantly and bias the audio driver fets to cut-off during stand-by. This would remove the load from the supply and prevent audio drive to the tube grids during receive. I haven't yet worked-out the details of how I'm going to do that, but it will be part of the T/R switching one way or the other.


Don

[WBear2GCR]

Hey Don!!

Maybe you could share the supplier of those "0 ufd/400v caps??   

[WBear2GCR]

FETS make these types of projects SO easy.   

I made a few HV and bias, adjustable, regulated supplies so I could run most any tube rig.

Mike, very nice job!
I got worn out just looking at the work that went into it!
Had to sit down, rest.

Did you use a decal for the meter face, or was the paint to keep from the old meter face
from bleeding through?

No need to sand it off, just flip it over??

                           _-_-

I had no idea that Daystrom was ever in Newark NJ!!

[KK4YY]

Bear,

Thanks for noticing the zero cap — I can see that you're paying attention. It's on the drawing because I intend to use a cap there but, being true to "as built", I zeroed it out for this posting. I can tell that I won't be able to get much past your keen eye.


Don

[KK4YY]

Admittedly, this design is a work-around for not using a p-channel fet. I wondered if I could find a way of doing it. It looked good on paper, so...

About the only time this circuit may be necessary would be where the desired voltage is high enough that p-channel fets aren't available, perhaps above 600V.


Don

[Opcom]

The regulation quality and simplicity combination is exceeding good compared to similar-function bipolar or vacuum tube circuits. I tried LTspice on it, very nice performance.

Modulator!! Hmm.. a conventional MOSFET audio amp running on a nice regulated +/- 250V supply could destroy many speakers! haha >3KW into 8 Ohms.

[MikeKE0ZUinkcmo]

...No need to sand it off, just flip it over!!...

Randy, I'm not that smart, I had to do it the hard way!!

Actually just printed it on regular paper with the Lazer, using TONNE software.

[KK4YY]

Patrick,

My measurements show...

At turn-on the voltage quickly reaches about 98% of the stable voltage. The last 2% takes a few minutes before settling down (the rails track each other well though). This would cause a very small, proportional change of the modulator grid bias of perhaps 0.2%. I could live with that. Better still, letting the supply run constantly eliminates that warm-up characteristic.

When I apply a 10k load (24mA) I get an initial 1% voltage drop that slowly settles to about 2%.

Considering that there's no thermal compensation or feedback circuitry, I'm well pleased with the result. It should be more than adequate for my application.

Thanks for running the sim.


Don