Help with voltage supply!

[bjm522]

I'm using a large beta electric power supply, rated at 30mA, but I'm only able to push it to about 10kV before the overload trips (about 32mA). I'm running air @ 1 micron with a turbopump, so I can get to any pressure required. as expected, lower pressures cause lower current, but below 1 micron there is little plasma formation. I have attached a picture showing the fusor in star mode at 8kV, as well as my overall setup.

What can I do to lower the current to ranges that other people are posting (~<10mA)?

Yes, that's a beer keg for a vacuum chamber.

[Tyler Christensen]

A chamber sealed off being pumped by a turbo should be well below the striking point of the plasma. It sounds like you have gross leaks in your system. I would do some leak checking, preferably before you run the turbo much more, you might be making the pump pretty sad if there are large leaks on either side of the turbo.

I also notice that you appear to have a plastic top on the chamber. It could just be that if you strike plasma at a higher pressure then attempt to lower it, the pressure is kept up by out gassing. This would be since hitting plastic with plasma will cause vigorous out gassing for most materials.

[bjm522]

The turbopump gets it down to .01 microns within 10 minutes of running, and I've done leak checks. This is not the issue- in fact, I actually use a controlled leak valve (bottom left of setup picture) to keep the pressure at 1 micron.

The issue is that I'm drawing far more current than other people have reported, and this is causing my power supply to trip at 1/3 of its rated voltage output. The only thing I can think of is that the grid is not transparent enough, and too many ions are hitting it.

Any other thoughts?

[Carl Willis]

I think it's likely that outgassing is the root of your problem. Watch the pressure gauge during a run, and if you see a surge of pressure as you add current, you can confirm this is the culprit.

Some outgassing is normal at the beginning of any run, but it can also be due to materials present in the chamber. I know nothing about beer keg construction...are they lined with anything? Tyler pointed out the acrylic lid, and with hundreds of watts delivered to the system, it's certainly going to be too hot on the inside surface.

Another possibility is that your leak valve isn't allowing you tight enough control of flow rate to dial in the low-current breakdown characteristic you are after. This takes practice as well as suitable equipment. Usually it's easiest to set the gas delivery valve to a fixed position and accomplish fine control by throttling the high vac line.

Can you post an Intro in the appropriate forum describing your effort more fully? That should be your first destination joining the forum, per the rules.

Thanks,
Carl

[Steven Sesselmann]

Brandon,

Nice job, lucky you have a current overload fuse on your supply, otherwise like most other beginners here, you would have blown up your power supply. Working with plasma is tricky, a vacuum can go from being a perfect insulator to a perfect conductor in no time at all.

It sounds as if your vacuum gauge is out by a factor of 10. usually a good plasma is sustained at around 10 micron, and it is a real balancing act to keep the pressure to voltage ratio just right.

You need a throttle on the turbo pump, such as a gate valve, to adjust the pressure, and a leak valve or a mass flow controller on the gas input.

After plenty of adjusting you will know where to find the sweet spot, and don't forget that the pressure changes with temperature, so as the fusor heats up, the pressure increases. This could be a real problem with your plastic window.

Tough having to drink all that beer in the name of science

Steven

[bjm522]

As I turn on the power supply and strike the plasma, the pressure does go up. I'm using an ion gauge good to nanotorr pressures, so I assumed that the fact that the gas was ionized was causing erroneous readings. What you said about the outgassing with ions striking surfaces makes complete sense in causing a pressure rise, and I will try to source a glass lid strong enough to hold 15psi.

The rays from the star are not directly striking the acrylic lid, but there are UV photons hitting it, causing a "blacklight" discharge in the plane of the rays. I never even thought about the fact that I had 300 watts currently going into the system, and I'm going to have to think about cooling or short operation times.

Beer kegs are just stainless steel, and I've sanded and acid washed the inside to remove all previous beer residue. This keg in particular was Lost Coast Brewery's Great White, so luckily it was a light beer and easy to wash off

The leak valve I'm using allows me to dial in the pressure very precisely, and it holds steady at any pressure I want. However, I'll switch to a throttling gate since I already have one in between the turbopump and the chamber.

I'm still not clear as to why more electrical current is flowing, though. Yes, outgassing can lead to contamination of my D2 atmosphere as well as pressure fluctuations, although, I'm not seeing those.

Thanks a ton for your input, and I'll post an introduction in the forum shortly. I overlooked that since this is a pressing issue in my mind.

[bjm522]

Yes, I'm very glad I have access to this power supply. It's pretty beefy.

The gauge is calibrated and has been tested in a known working environment. I do get my best plasma at ~10 micron, but at that pressure, I can only put 5kV on the electrode before the system overloads. I'm lowering the pressure only because I am trying to increase the voltage I apply before cut-off. I know very well of the balancing act you speak of; If I overload, I have to bring the pressure back up, strike the plasma, then drop it down again while increasing the voltage.

I have a gate valve on the turbopump, but I wasn't using it. Instead, I was using the controlled leak valve to control the pressure.

I'm starting to see problems with the plastic window, and I've limited my runtimes to compensate.

I still don't see exactly where the extra current is coming from.

Thanks for your input!

[Steven Sesselmann]

Brandon,

The extra current is most likely thermionic emission from your grid. As your voltage climbs so does the ion collision with the grid, thereby heating the wires. A hot negatively charged wire become an emitter of cathode rays, which are simply electrons running away to the shell.

These cathode rays are actually necessary in order to maintain a lit plasma, if you had a cold grid, you would need a separate ion source to keep the thing going.

Steven

[bjm522]

This makes a lot of sense. Is this a standard occurrence in other people's fusors? If so, how do you manage to run them at ~10mA currents at higher voltages? The only major difference that I can see is the thickness of the inner grid. Most pictures that I have seen seem to have a thin wire grid, where mine is 1/8" thick.

Also, will the conductivity of the plasma change with gas composition? i.e. will the current drop if I get a D2 atmosphere?

I guess the bottom line is that I need to figure out a way to increase the potential in the system. Any suggestions that anyone has will be tried; equipment is not lacking, and I can come up with anything.

[Steven Sesselmann]

Brandon,

What is the purpose of your second grid?

Steven

[bjm522]

originally it was the positive grid, with the keg grounded. It has been removed, since the new power supply is one high voltage output, and one case ground.

-Brandon

[Steven Sesselmann]

Brandon,

Right, so you have removed it, well that's fine, as it really won't add much to a beer keg.

I would try running the grid nice and warm <10 kV plasma and then slowly reducing the vacuum and watching the voltage climb through 10 kV and beyond.

Steven

[bjm522]

Thanks, by "reducing the vacuum", do you mean increase the pressure?

right now, at 10 microns, I can only apply ~5kV at max current. If I reduce the pressure to ~1 micron, the current drops and I can increase the voltage to ~10kV at max current.

Are you saying that as the grid heats up, I will get less current and therefore be able to increase the voltage? Wouldn't the current increase with a hotter grid, due to thermionic emission?

I'm not exactly following here. I think that if you clarify these things, I will have something to try and report back to you after the weekend.

Thanks for all your assistance so far!

-Brandon

[Steven Sesselmann]

Brandon,

Correction, I meant decrease the pressure ie. increase the vacuum gradually.

The ion current should fall and the voltage should increase as the vacuum increases.

Best of luck..

Steven

[Chris Bradley]

Brandon, I can't be sure but it sounds like you didn't fully appreciate Tyler's point:

If you can get down to the 0.01 microns your gauges are indicating, then you should be able to power up your power supply all the way to many 10's kV and see nominally zero current through the plasma.

If you are getting significant current drawn for, say, 25kV at 0.01 microns and plasma showing, then your gauge is likely wrong.

If you are getting significant current at 25kV at 0.01 microns and there is no plasma showing, then there is likely a short circuit/leakage path outside of your chamber.

You go on to say;

> As I turn on the power supply and strike the plasma,

OK, so taking that literally, then change your protocol by setting the pressure to 0.01 microns and PSU to 25kV, then slowly increment the pressure. At some point the voltage will begin to drop and the current will increase, and you will then be walking a tight-rope of control to keep the volts high and the power input low (to prevent the 'thermal runaway' from outgassing increasing the plasma pressure, thus current, thus power).

You need to inch your way around the parameter-space [pressure/volts/current] very slowly, starting from an 'unlit' plasma, not just whack on voltage and see the result because that way you end up with high power inputs and the positive feedback from outgassing as has been described.

Richard has written a very detailed FAQ on a step-by-step method to explore the plasma parameter space of such a device.

Also, I think rather than finding one great big piece of glass for the top, you should get a piece of metal with a hole in the middle and put a small piece of glass over it. You'll end up with two seals instead of one, but should make no real difference if it is as good a seal as the much larger outer ring seal. Just an o-ring around a hole will do, providing the metal sheet is flat and free of deep marks, then flat glass on top of that will do you. It would also be advisable to position a second piece of glass (can be thin) on the inside of the glass that will be sacrificial if it takes an electron beam hit, or becomes sufficiently sputtered on that you throw it and replace with a clean one.

[bjm522]

Thanks for your reply.

at .01 microns, I can turn the voltage all the way up to 30kV without getting any current or plasma. I can only get a plasma started at about 1 micron.

I will try your suggestion of 25kV @ .01 microns, and slowly increasing the pressure. I think that I tried this before with an overload result the moment the plasma fired, but I can't be sure as I've tried a lot of things. Most efforts have started from a lit plasma, and that might be my problem.

I will also check out that FAQ. It's starting to seem like this is a common issue, so I don't feel so bad anymore.

-Brandon

[Chris Bradley]

You're almost certainly still out-gassing a stack of water vapour in the thing. Has it been pumped down for very long? What does the chamber pressure rise to when valved-off overnight?

After more pumping and more plasma, I expect you'll see things begin to stabilise. Just be very slow on the valve and as the pressure goes up the voltage will be pulled down* and the current will rise.

*I was just thinking now that if your power supply puts out a 'constant voltage' then it may start pumping in the power quicker than you'd like as the current gets drawn. 'Simpler' power supplies will naturally have their voltage outputs pulled down as current rises (I tend to feel that more like 'power' sources, rather than voltage or current, are better for this sort of work as the power source reacts in the 'right sense' to do the job).

You'll have to experiment as to what is the best voltage to start off with is, with your particular PSU and chamber. What sort of a ballast resistor are you using, with this much current running through? All in the FAQs....

[bjm522]

I was unable to find any specific information about ballast resistors in the FAQs. There is a ballast resistor within the power supply, but I have not put a meter on it to test its resistance.

I am sure there's plenty of water vapor coming off of it, as I have not done a full bake-out. The pressure does not quite get to the vapor pressure of water overnight, but it does rise quite a bit.

I'm heading over to the lab right now to try setting the voltage high and slowly raising the pressure. Will return with results.

Thanks again for your assistance.

-Brandon

[Jeroen Vriesman]

You need a high-power resistor between the supply and the grid.
If you look at the paschen curve you wil see that it's very steep at low pressures.
Without a resistor your supply is a constant voltage, en the higher you need it to be, the smaller the possible presure range is.
With a resistor your power source behaves more like a current source.
Watch out for long cables between the grid and the resistor, cables can behave as a capacitor.

[Richard Hull]

It is all a matter of conditioning the chamber, getting all the water out and then things will stabilize. I have to run my fusor for about 2-3 hours after a long latent period before its settles down into a sweet spot. You have a lot of ion bombardment bakeout time ahead.

The pressure versus voltage and current is a very tricky business, but once the sweet spot is found, you get an inherent feel for it. It is as much art as science as you are creating a special environment in a very narrow zone of operational range.

It will all come with time.

I will add you to the Plasma Club.

Good work. Keep it up.

Richard Hull