Fusor Forums

Target for ICP needs cooling. I am thinking about using tap water to cut the costs but with three phase and about 130Kv connected to target only possible way of using tap water I can think of is using a closed recirculating water loop using a battery powered DC circulating motor and pump with the whole system floating in good isolation (about 50 cm above the ground). Any idea or advice with this?

You might want to look into Fluorinert. It's not cheap, but it's a nonconductive fluid that has good thermal cooling properties. Running a charged cooling loop is very difficult and might be asking for disaster unless done extremely well.

Use anything other than water. Mineral oil is a lot cheaper than fluoinert, and doesn't have the heat capacity of water (few things do), so just pump it faster. Which is now easy as the pump and so on don't have to float off ground. Hydraulic fluid (cheap and pretty thin) would probably work too. Most refrigerants would be good if you can stand the hassle of keeping them contained at some pressure.

Water is actually a good insulator but only if it is *very* pure, and it will only stay that way if you purify it more or less continuously, as it leaches impurities out of just about everything it comes into contact with -- they call it the universal solvent for good reasons. Not worth it on our smaller scales IMO.

Hi Steve,

From experience I can attest to the use of deionized water conveyed in rolls of PE tubing to cool kilowatt loads at ~50 kV. Extension to 100 kV would not be difficult in principle, but much thought about the layout is necessary.

Obviously, having mobile, pressurized water at high potential creates a special brand of electrical hazard, one that needs to be contained.

Fluorinert is a good idea also. It evaporates easily, so should be kept in a sealed system. The circulation pump and other components must be chemically compatible. PE tubing is adequate.

Moving insulating liquids can transport charge a la Van de Graaff generators, and water can conduct if your DI cartridge dies or the water has not had time to clean up. Attention should be taken to ground all metallic components in the flow loops that are not at the high potential by design.

-Carl

Fluoinert comes in many flavors, some of which do *not* evaporate easily at all, but it's still way too expensive. I use a type here for vapor phase reflow soldering that doesn't boil until a couple hundred C plus -- solder melting point or a little above. It stays in the coffee can for years, no loss, no problem, so it matter to say which part number/boiling point you are talking about within this family of chemical products.

I bet you could push some refrigerant in gas form through most things fast enough to get decen cooling, but haven't tried that. Or just use a refrigerator compressor and some other parts stolen from a junk fridge, like the condenser and so on. You'd just have to find some high pressure tubing that was also a good insulator -- maybe hard to do?

I hadn't considered the VanDeGraff effect, but no kidding, the laws of physics don't change just becasue you forget one or two of them....Ow! But assuming you're going to have a grounded reservoir and pump, it should only cause some parasitic current draw, not other troubles, I think.

We use
Low odour Base solvent to cool the cathode on our Negative ion source.
This stuff is like keroscene with out the smell
small gear pump drives it around to the ion source 85kV away from earth with water cooled heat exchanger at earth end.
tubes are 3/8 bore polythene

John Futter wrote:
> We use
> Low odour Base solvent to cool the cathode on our Negative ion source.
> This stuff is like keroscene with out the smell
> small gear pump drives it around to the ion source 85kV away from earth with water cooled heat exchanger at earth end.
> tubes are 3/8 bore polythene

sounds like automotive parts for diesel and diesel would work with a domestic boiler heat exchanger

I once used Activated Alumina to de-odorize Kerosene. Also removed all color, too. But I never measured it resistivity. I did it because we were trying A.Alumina on everything, and this stuff really smelled. Afterwards, it was great.

Ordinary mineral oil or silicone oil, works quite well with a gear pump and low pressures. AS Carl notes, you should have a piece of grounded metal tubing somewhere in the loop, down near the pump, obviously to handle leakage and injected charge.

Dave Cooper

What about HV transformer oil? I have seen they are used in 110 KV power transformers?

Just for the record pure water isn't an insulator. It's a lousy conductor. Low odour mineral spirit seems as good an answer as any, but remember its still combustible.

All I know is that we buy this stuff from NEC

transformer oil should be more than ok

Well, by that definition quartz is also a conductor, and so is high vacuum -- put a big field on either and current will flow. It's always a matter of degree, order of magnitude. Look at ohms per sq of really pure water and it looks decent -- problem is keeping it that way. The big Z pinch machine is water insulated. So....obviously it can work as an insulator as long as you keep ionic stuff out of it. I was kind of surprised how good it is when I went to use cataphoresis to put Y2O3 on some filaments to make them emit electrons better, using plain old store distilled water -- even in a not that clean beaker with whatever contamination the Y2O3 had in it,it took hundreds of volts to get it to draw current. But let that sit on the bench a couple of days and get some dust in it, it becomes a much better conductor.

The other problem with water is lubricity -- it stinks as a lube (great as a cutting fluid) so pumps that have rubbing parts need not apply. No gears etc, have to go centrifugal and even then if the pump is not designed for it, the seals and bearing fail quickly. Best I'd seen for that was a magnetic coupled pump with graphite bearing on alumina for that part. Lived forever.

Yup - Transformer oil IS the coolant in.... um.... transformers.

Can be a bit ornery, if your outside temps are -40F or something, but if it's warm, it flows real nicely.


Dave Cooper

If you are still considering water, look at 12 volt bilge pumps, they are cheap and available in lots of different sizes (litres per hour). I think RULE is the most common make. They are also mostly plastic.

And, if you can find one the right size, they really do rule ;~} I've used them here on the farm with a large tractor towable tank to get lots of water out of streams etc. They are very power efficient. They may not make enough head for your application, but these were done by guys who really understand the hydrodynamics of water, unlike nearly all other water pumps. They are certainly one of the best for pre pressurizing another non-self-starting pump from a sump of some kind Good thought, Lyn.

Seems like they may also handle some of the other fluids mentioned above, too. As an aside, I've had these frozen solid in the bottom of a tank, no damage once things thawed. They are better than they look.

They are a lot more powerful than they look, and they will maintain a head. The reason I suggested them is if you need to float a 12 volt system in a HV circuit they will take some beating.

They are very efficient, true, but only meant to get water from a bilge to over the side, or at least the ones I've seen, up to the 1000gph range. At 10' head (about 5psi) the flow drops to zero. I used to use them daily to get our drinking water out of a rather remote spring, and had to dig a road to get the tractor closer because of that. Takes more than 5 psi pressure drop to run *most* cooling systems but great for priming a more powerful pump you can't put into the water like you can one of these.

Think one of those garden fountain pumps, on steroids, and you're about there. Very low power drain for the amount of water moved.

I'm not going to press the point, but the larger ones do create some pressure. depends how much flow you require. Well worth some experimentation. I respect your comments Doug.

Fluorocarbon fluid period is the best I think , either fluorinert, or better the newer Galden line of fluids, with just about any reasonable temperature tailored product being available. They insulate like crazy, and they are totally stable, and non-toxic. Many commercial accelerators use Galden fluids to cool the terminal ion source, so it will isolate into the MEV range. The semiconductor industry is also a heavy user.

I have some Galden 110 which is like "fluorocarbon water" it boils at 110C, but the viscosity is very low so it pumps fast, and it is compatible with everything except, get this: Fluorocarbon polymers which it makes swell. On the other hand cheap rubber or poly tubing is just fine. Check out the Galden web site:

http://www.solvaysolexis.com/products/b ... 2-0,00.htm

You can often find it for a reasonable price on ebay. There was some HT200 the last time I looked with "best offer"