Floating on High Voltage

[vrulg]

Hi,

If I wanted to float a voltage supply on top of a very high voltage (~100kv) what type of power supply should I use? The floating supply would be used to power a heating element. The high voltage will be generated using a cockcroft walton generator.

 

[thylacine1975]

I smell ozone ... or an X-ray/particle accelerator project?

Hmm. Something with good dielectric isolation! I've been pretty surprised at the isolation achieved by a simple DC/DC flyback converter on a large magnetic core (with loads of plastic insulation between the windings and the core) immersed in motor oil. 100 kV is a factor of a few higher than I've ever attempted this way though

I used to work on an accelerator with a 2 MV Van De Graaf generator at it's heart. The electronics at the top potential was powered by a motor (at ground potential) driving a (long) plastic rod, coupled to a generator housed in the Faraday shield at the hot end. Control signals were transmitted across the potential difference via fibre optics.

Unfortunately the Cockroft-Walton multiplier output voltage droops badly under load, so make sure you minimise your leakage/corona losses whichever approach you choose.

Have fun!

 

[RetroTechie]

(Rechargeable) battery-powered. Seriously. If not (mains powered): how the hell do you think you'll keep that ~100kV from going through the power supply parts, and still be able to transfer enough power? Other than hand-winding a custom transformer, shining light on a solar panel or crazy stuff like that.

The floating supply would be used to power a heating element. The high voltage will be generated using a cockcroft walton generator.

Hey I built one of those once! (didn't know it was called that) :razz: 9V battery to ~600V DC, few mA's at most (CMOS oscillator + driver + a long row of 1N4148's + small capacitors). Going into a Geiger counter tube that I picked up at a dumpstore.

I'd try & find a way to keep whatever needs heating at a near-ground potential, and whatever goes high-voltage (negative or positive!) not need another power supply of any kind (apart from the high voltage you'd put on it).

Another option might be to replace one of the diode/capacitor sections with your load, as long as you don't connect ends between beginning & end of the chain, it might get a voltage across it in same order as AC that goes in. But I'm not quite sure if this would work, how exactly to wire it & how to calculate the numbers. Maybe some simulation a la Spice would be in order.

What exactly is it that would need heating, and what order of magnitude in heating power are we talking about?

 

[BradtheRad]

You can attach an ordinary diode bridge supply anywhere in a C.W. doubler, as though it is just another stage. And that's what it is. You can make it carry high current similar to an ordinary bridge supply.

However the C.W. stages remain low current.

I have simulated this successfully. You can attach one bridge supply, but no more than one. If you try two or more, then short circuits occur.

The rationale is this. The center tap (at transformer) does not carry current (despite a connection to the center tap being on every schematic you've seen of the CW topology). You can disconnect the center tap. Then you connect the HV load across the capacitor stack.

You can attach the bridge supply at the lower terminal, or the high terminal.

I'd post a schematic, however drawing it would be time and effort.

 

[vrulg]

Thanks all, for the responses.

It's for an electron accelerator. Unfortunately, keeping what needs to be heated at ground isn't an option, because what needs to be heated is the cathode(where the electrons come out of), and that can't be at ground(well...it could I suppose, but it would actually make the rest of the design much complicated).

I did have the idea of sticking a battery in there, but I thought there must be some sort of problem with that that I can think of

 

[RetroTechie]

It's for an electron accelerator. Unfortunately, keeping what needs to be heated at ground isn't an option, because what needs to be heated is the cathode(where the electrons come out of), and that can't be at ground(well...it could I suppose, but it would actually make the rest of the design much complicated).

See ordinary CRT tube: elektron guns are heated but using relatively low voltage (few hundred volts max) with respect to outside world / connected equipment. The high voltage for accelerating those electrons (few dozen kV) is applied to parts further down in the tube, but these don't require heating. So (at least in that setup) there's no need for a power supply that sits on top of the high voltage.

---------- Post added at 07:07 ---------- Previous post was at 06:37 ----------

You can attach an ordinary diode bridge supply anywhere in a C.W. doubler, as though it is just another stage. And that's what it is. You can make it carry high current similar to an ordinary bridge supply.

However the C.W. stages remain low current.

I have simulated this successfully.

You may have misread topic starter. Suppose you use a simple AC transformer to feed those voltage doubling stages. Primary side of transformer is hooked up to mains AC (=ground potential or within a few hundreds Volts of that). And there's a few kV of electrical isolation between primary and secondary transformer winding.

Off it goes into all those stages, and at the end you have, say +100 kV with respect to potential of the things we can touch. Now you include "an ordinary diode bridge supply" somewhere near that +100 kV point, and you've just created a return path that only has a few kV's isolation so that 100 kV will flash through. There goes your supply (and if you're lucky no-one gets electrocuted in the process!).

Basically what you suggested is fine from a simulation or theoretical point of few, if you have "an ordinary diode bridge supply" that has 100+ kV electrical isolation between mains AC and secondary side. The "ordinary diode bridge supplies" that I have are double isolated, which means a little over 4 kV isolation is guaranteed. Not 100 kV...

Theory is one thing, but don't forget the practical issues when trying something like this. That ~4 kV safety isolation that we use to prevent AC mains from electrocuting us, means nothing when you put 100 kV across it. 8-O

 

[BradtheRad]

Yes 100kV is much different from 1kV.

My oscilloscope has the electron beam (1000 volts) riding a 6.3VAC transformer winding. It is in proximity with mains supply (both in the original transformer and in the new one I installed to fix an arcing problem).

Since the OP stated he was using a CW multiplier, I jumped to the assumption that he had determined it is the best design to suit his purposes. He didn't say what his supply voltage was, nor how many stages he was making.

I get the idea that the CW design is popular because it can use components that have the same rating. There's no need to increase capacitor V ratings with increased number of stages. No one section has to bear the entire output voltage (with the exception of the load if attached).

=================================

To show how my idea plays out...

Here is the CW layout which worked for me. As you can see, the highest voltages appear at the rightmost section of the schematic. And there is no center tap (zero ground) connection.

Here is the layout which has a bridge supply at the top of the capacitor stack. I see that it is unworkable. The caps at the supply rails will need an unreasonably high V rating.

This is another multiplier known as a series-parallel type. It is more suitable for higher current needs. It has a disadvantage since it needs increasingly higher capacitor ratings as more stages are added.

I'm posting these schematics to show my concept can work if there are few stages. I realize (now that you point it out) that they won't work for the OP.

 

[vrulg]

See ordinary CRT tube: elektron guns are heated but using relatively low voltage (few hundred volts max) with respect to outside world / connected equipment. The high voltage for accelerating those electrons (few dozen kV) is applied to parts further down in the tube, but these don't require heating. So (at least in that setup) there's no need for a power supply that sits on top of the high voltage.

Maybe I'm missing something, but I don't see how that can work. If the electron source isn't held at the cathode potential, it will be essentially at ground. Which means the electrons should experience a force directed from the cathode to the electron source. Essentially the electron source would act as another anode, but on the wrong side of the accelerator. I don't see how very many electrons would make it past the cathode to be accelerated.

 

[RetroTechie]

Remember voltage potentials are relative... I suppose what you're thinking of is (heated) electron source at -100 kV, with target (that electrons are drawn to) at near ground potential.

CRT type just moves the 0V reference point: (heated) electron source at near ground potential, target (near the front of the screen) at some +24 kV.

In both situations you have a high voltage difference, in both situations you have (heated) electron source at - side, and electrons accelerating towards + side. Only difference is in which side is kept near ground level.

 

[vrulg]

Remember voltage potentials are relative... I suppose what you're thinking of is (heated) electron source at -100 kV, with target (that electrons are drawn to) at near ground potential.

CRT type just moves the 0V reference point: (heated) electron source at near ground potential, target (near the front of the screen) at some +24 kV.

In both situations you have a high voltage difference, in both situations you have (heated) electron source at - side, and electrons accelerating towards + side. Only difference is in which side is kept near ground level.

Ah, I see you meant. Keeping the cathode at ground (as opposed to keeping the anode at ground) complicates the design of the accelerator. This is because the walls of accelerator, beyond the anode, need to be at the same potential as the anode. Functionally this means that most of the accelerator setup needs to be at the same potential as the anode, and keeping it all at +100kv could get annoying with safety considerations.

 

[BradtheRad]

Maybe this idea was already covered, but you can make a negative voltage multiplier by reversing diode directions.

Then you have a single wire with -100 kV coming from it. I believe that's what an electron gun is.

Keeping one point at -100 kV beats keeping a lot of things at +100kV.

 

[RetroTechie]

Keeping the cathode at ground (as opposed to keeping the anode at ground) complicates the design of the accelerator. This is because the walls of accelerator, beyond the anode, need to be at the same potential as the anode.

Okay I see... this does present a question: if walls of the accelerator are at same potential as the target, what's preventing the electrons from taking a shortcut & go for that wall instead? :?: What do you intend to use as target(s) & building materials?

---------- Post added at 20:58 ---------- Previous post was at 20:51 ----------

Maybe this idea was already covered, but you can make a negative voltage multiplier by reversing diode directions.

Then you have a single wire with -100 kV coming from it. I believe that's what an electron gun is.

Keeping one point at -100 kV beats keeping a lot of things at +100kV.

In principle it shouldn't matter much what side you have at high voltage. In practice that can matter a lot depending on function & construction. Problem is that electrons leave a wire more easily when that wire is (thermally) hot. Hence the need to heat - side (and why poster is having this problem).

 

[dfullmer]

Out of box.

Heat your element with a laser beam delivered via fiber optic. Glass is a good insulator. Sounds like a fun project.

Good luck