High voltage DC PSU (4 kV 250 mA)

[pmartin]

Hi, I need to build a PSU for a sputtering deposition chamber. The requirements that I was given are:

4 kV DC 250 mA for 5 min, turned on and off manually, powered from mains

I have a microwave transformer (220 V - 2.2 kV, 1450 W) which would give about 3.1 kV DC after full wave rectifying and filtering (diode bridge+cap) and more than enough current. I was thinking about a few options:

1.- Making a sine wave PSU capable of delivering 400 V peak (sinusoidal oscillator at a low voltage, around 12 V peak for example, connected to a 220/6 V transformer, then to the HV one).

2.- Same as option 1, but using a square wave from a 555 instead of sinusoidal. I don't know if it's okay to apply a square to an iron cored transformer (after removing the DC component) because of the distortion introduced by its attenuation to high frequencies.

As a side question, if I needed to make it regulated, how would I be measuring the voltage if I wanted to do it in the HV side? Simply a voltage divider, a voltage follower op amp and an optocoupler? (I know that I can measure from the intermediate stages, ie the 400 V one or even 12 V one).

Thanks in advance for any help that you can give me.

PS: I only need a single unit.

[Arjan Emm]

Microwave transformers are already pushed to their very limits in normal use. You can't feed them with a higher voltage.

[T3sl4co1l]

Might be able to get some resonant action: put a capacitor across the secondary.  It will resonate against the leakage inductance, which is intentionally large on an MOT.  As a bonus, this has the effect of regulating voltage, because more voltage drives it harder into saturation, shifting it out of resonance.  (And because line frequency is pretty stable.)

I don't think this would actually help achieve more output voltage though, nor more VAs.  You could demonstrate this effect (regulation and voltage gain) at somewhat lower line voltage (perhaps 50-70% of nominal?), but not if it's already at the limit. :/

FWIW, Hammond makes a few transformers nearly in that range.  Perhaps a couple stuck together would do?

Tim

[oldway]

Voltage doubler + triac or variac primary voltage control/regulation

[tautech]

Bit more voltage and less current than your spec but what about a furnace ignitor transformer ?
Looking at one I have here: 2 x 4 KV @ 15 mA continuous.

[Richard Head]

You need more detailed specs than that. If they can't give you more specs you'll have to research it yourself.
ie Does it have to be short circuit proof? That makes a big difference to the design.
Is there a requirement for output voltage smoothing? If so what sort or ripple is required?
What degree of line regulation is required? Probably not an issue but you need to ask the question.
I sounds like a custom step-up mains transformer would work fine. You can probably use a 500VA transformer since the duty cycle so low. 

[mikeselectricstuff]

For simplicty and robustness it's hard to beat a motor driven variac, assuming you don't need super-fast transient response.

Maybe also look at saturable reactors if you need regulation over a fairly small range.
Microwave transformer with  a doubler is probably the simplest & cheapest, especially as you can get the diodes and caps from the microwave as well.

You might also want to investigate parts used in "inverter" microwaves, which use high-frequency magnetics - I can't immediately think of another mainstream source of suitably high power, high voltage, kHz frequency transformers, and this is cedrtainly something you don't  want to have to make yourself.

And of course don't forget to keep an eye on Ebay for something surplus

[Someone]

And of course don't forget to keep an eye on Ebay for something surplus

Yes, buying something designed for the task where available is almost always cheaper. For the surplus prices of these specialised power supplies you'd be hard pressed to purchase the parts and enclosure to build something similar.

[mikeselectricstuff]

Maybe also look at saturable reactors if you need regulation over a fairly small range.

I wonder if you could get some amount of adjustment for regulation by putting DC into the filament winding to partially saturate the core.
You'd probably also need some primary ballast to deal with the saturation current.
 

[james_s]

Maybe also look at saturable reactors if you need regulation over a fairly small range.

I wonder if you could get some amount of adjustment for regulation by putting DC into the filament winding to partially saturate the core.
You'd probably also need some primary ballast to deal with the saturation current.
 

I don't think that would work. IIRC saturable reactors use a pair of control windings on the outer poles of a 3 pole core, wired out of phase so the induced AC voltage cancels out. If you try to feed DC into a single winding on a transformer the AC will be fighting against the DC you're trying to put in.

[T3sl4co1l]

I don't think that would work. IIRC saturable reactors use a pair of control windings on the outer poles of a 3 pole core, wired out of phase so the induced AC voltage cancels out. If you try to feed DC into a single winding on a transformer the AC will be fighting against the DC you're trying to put in.

You get ripple on the control winding, so you need to use constant current.

It would take a tremendous amount of current to saturate an MOT on the filament, though -- you need, I think, a few hundred amp-turns, and there's maybe two or three turns there.

Tim

[CM800]

There's two things you could do:

Go ask a local transformer house to wind you a nice iron core unit (£200-£500 I guess, not too sure?)

Wind your own ferrite unit. It's quite common for hobbiests to make their own high voltage transformers for fun, check out 4hv.org, they are one of the bigger high voltage communities
I know more then handful of people have made their own units & drivers.

http://uzzors2k.4hv.org/index.php?page=hvxfrmrwinding
http://uzzors2k.4hv.org/index.php?page=slrconverter

[David Hess]

Ideally I would use a plate transformer designed for a legal limit HF amplifier.

Alternatively if it has to be done as cheaply as possible with microwave oven transformers, what about using two in series so they can be operated below their rated voltage and the output can be adjusted?

[mikeselectricstuff]

Ideally I would use a plate transformer designed for a legal limit HF amplifier.

Alternatively if it has to be done as cheaply as possible with microwave oven transformers, what about using two in series so they can be operated below their rated voltage and the output can be adjusted?

MOTs typically have one leg of the secondary connected to the frame, and minimal insulation if you decide to disconnect it, so this can risk insulation breakdown unless you can use them in a centre-grounded configuration.  With a capacitive voltage doubler you may be able to get round this issue.

[james_s]

I've wired two MOTs in series by disconnecting the inside from the frame and then tying both of those newly free leads together. Wire the primaries out of phase and this configuration maximizes the insulation between any two points of differing potential. It might help to vacuum pot the thing in epoxy although I haven't tried.

[pmartin]

Thanks for all the help and ideas. They ended up doing it themselves with a variac+MOT (with its diodes and capacitor) after speculating that plasma could be generated at a lower voltage (the threshold voltage turned out to be 340 V at the input of the MOT).