Page 1 of 2

Non-Corrosive and/or Dielectric Liquid Cooling?

Posted: Sat Oct 12, 2019 5:28 am
by ian_krase
If you're like me, then in the course of vacuum work you come across the fact that glow discharges and particle beams easily heat things up... too hot. Hot enough to demagnetise, to outgas or oxidize, to melt, to enter thermionic runaway, or to destroy plastics or rubber seals.

The solution of course is liquid cooling. But this has some problems. As every liquid-cooled gaming PC afficionado knows, mixing copper and aluminum in the same system, even if they are not electrically connected, is just asking for a mess of galvanic corrosion that will pit and pinhole the aluminum pretty rapidly. And this applies more generally.

Also, if you're using water with additives for coolant, you can't have something you're cooling at other than ground potential. The coolant will conduct.


The solution to this is to use dielectric coolant, and that has... problems.

The best-known dielectric coolants seem to be "Fluorinert" and various other fluorocarbon or PFPE-based liquids. These are expensive and probably not great for the enviroment or health. Most ordinary hydrocarbons are far too flammable, as well as being polluters if they evaporate and tending to destroy plastics.


What are my options?

Re: Non-Corrosive and/or Dielectric Liquid Cooling?

Posted: Sat Oct 12, 2019 6:51 pm
by Andrew Seltzman
Fluorinert is best. It has no health issues at all, unless you burn it (don't).

https://www.ebay.com/itm/3M-Fluorinert- ... SwpOVdhGs6
https://www.ebay.com/itm/3M-Fluorinert- ... SwsR9c73pt

Re: Non-Corrosive and/or Dielectric Liquid Cooling?

Posted: Sat Oct 12, 2019 7:56 pm
by Richard Hull
The idea of cooling in the fusor goes way back on this site. From cooling a grid made of hypo needle tubing with flowing coolant, to cooling the fusor body when Carl Willis brazed1/4" copper piping to his hemispheres and pumped ice water through them.

Cooling the grid seems the most difficult. Cooling the chamber walls might just assist in allowing more D2 wall loading. General plasma streamlines require steering gear or potential shielding screens over sensitive regions where such beams might hit sensitive items which might out-gas or be damaged within the vacuum chamber, itself. As noted, cooling within a vacuum and among HV components within can be daunting.

Fluorinert has been mentioned many times here.

Richard Hull

Re: Non-Corrosive and/or Dielectric Liquid Cooling?

Posted: Sun Oct 13, 2019 12:18 am
by ian_krase
Interesting.

Flouorinert has a fairly low vapor pressure. Do you need to make a completely closed system?

Re: Non-Corrosive and/or Dielectric Liquid Cooling?

Posted: Sun Oct 13, 2019 1:53 am
by Bob Reite
My vacuum chamber does have a water jacket (or liquid jacket if you prefer) I never got around to hooking it up yet. I believe there are some archived posts here with someone cooling the grid with Flouorinert. Because of the expense, the Flouroinert loop goes to a water cooled heat exchanger close to the grid feedthrough.

Re: Non-Corrosive and/or Dielectric Liquid Cooling?

Posted: Sun Oct 13, 2019 4:21 am
by John Futter
Professional systems like accellerator mass spectrometers use deodorised kerosine to cool the ion source
I use ordinary kerosine to cool my ion sources at work ie 500 - 1500 watts of heat input tends to destroy the ion source magnets if the cooling stops.
Kero runs closed loop into a heat exchanger that is water cooled. Ion source has a copper backing plate that the kero flows through
and I use a petrol pump as used at service stations to circulate but you could use a high flow rate car type fuel pump


Cheap as chips

Re: Non-Corrosive and/or Dielectric Liquid Cooling?

Posted: Sun Oct 13, 2019 5:28 am
by ian_krase
By "completely closed system" I mean one that's all sealed up, not just a closed loop.

Re: Non-Corrosive and/or Dielectric Liquid Cooling?

Posted: Sun Oct 13, 2019 5:43 am
by Andrew Seltzman
Here is the documentation on the fluorinert cooling of my grid:
http://www.rtftechnologies.org/physics/ ... -index.htm

Fluorinert does evaporate slowly (much slower than ethanol), so a mostly closed system is recommended. A simple plastic cap or over pressure vent valve on the reservoir will do. It doesn't have to be a hermetic seal, but it shouldn't be an open reservoir either.

Re: Non-Corrosive and/or Dielectric Liquid Cooling?

Posted: Sun Oct 13, 2019 11:24 am
by Dan Knapp
Behlke (Behlke.com) uses a sealed closed loop fluorinated fluid system to cool and insulate their HV pulse switches. They use a pump which is sold for cooling computer CPU’s. There are a large number of different pumps sold for this purpose largely to the high performance gaming community. This would be a source of reasonably priced cooling components.

Re: Non-Corrosive and/or Dielectric Liquid Cooling?

Posted: Sun Oct 13, 2019 7:14 pm
by Rich Feldman
Two angles here.
1) Water is used as a HV dielectric in some pulse forming networks, where its high dielectric constant is favorable for electrode geometry reasons.
But even if deionized to double-digit resistivity (in megohm-cm), which is hard to keep that way, the self-discharge time of a water capacitor is on the order of a millisecond (IIRC).
Edit: RC for 1 meter cube is (8.85 X 80.1 = 709) pF X 0.18 megohm = 128 us (at 20 degrees C). The bottom line value is the same for any geometry.

2) How about long, thin tubing to make the electrical resistance tolerable in a water cooling connection? Trading off water resistivity, pressure drop, and temperature rise. Hot water is less viscous and less electrically resistive than cold. I think the electrical leakage current will galvanically eat at HV-positive pipes and not at those which are HV-negative.

I bet the main advantage of Fluorinert over refined hydrocarbons is its non-flammability.
As with PCB's in capacitors and transformers, there's an obvious & immediate benefit because it reduces the risk of fires.
As with all liquids, you need to pay attention to compatibility of wetted materials like elastomeric seals.

p.s. Regarding toxicity, ozone depletion, and global warming potential, the numbers for propane and butane are many times lower than even the latest nonflammable refrigerants.
I bet the same applies to the higher-boiling-point nonaqueous coolants under discussion in this thread.