I am designing a 250W inductor of spiral planar turns on a 16 inch square AL heat sink with fan, operating at 100A and 120KHz for testing. It contains voltages of 3KV and I am using #10 AWG rated 600V for the proper resistance needed. I have good separation between turns so only the separation of turns to heat sink is important. I would like to elevate the spiral winding above the heat sink and flood the area with a bonding material with good thermal properties. The better the electrical isolation the closer I can place the winding to the heat sink. What are good materials for this task? Cost is a third order driver. Cheers, Harry
--------------------------------------- Posted through
I guess you know this, but just in case--- the aluminum heat sink will significantly lower the inductance of your pancake coil if it's close.
What maximum temperature are you expecting to see at the wire? I'm thinking something like laser-cut alumina insulator with thermally conductive (but electrically insulating) epoxy. Maybe something like this:-
The current induced into the aluminum may heat the aluminum more than the copper coils!
Are you using solid #10 wire? That may have a lot of eddy-current losses and get hotter than necessary.
It might be better to use bare copper wire in a conventional solenoid pattern and blow a lot of air through it.
One thing I do is epoxy current shunts to heat sinks, at 100+ amps but not at kilivolts. A major issue for me is the inductive ciupling between the shunt and the heatsink, and the resulting eddy currents in the aluminum.
If you must use your suggested geometry, consider an insulating layer of BeO or AlN. You could also use flat or square wire to maximize the thermal transfer surface area.
When you say the wire is rated for 600 volts, that suggests you are using plastic insulated wire, like regular household wiring stuff. That will be terrible thermally. It will have awful eddy current losses, too.
Well, first off, lose the 600V insulation - it's a thermal nightmare and useless for 3000 volts.
Dupont Kapton can be had that will hold your full voltage at ~1 mil thickness. You might want to go a bit thicker for safety factor. Bond a sheet to your heatsink and you can press your coil right down on it. Or lay a sheet on each of two heatsinks and sandwich the coil between them along with some thermal conductive epoxy (though it might work fairly well without the epoxy, if the heat sinks are rigid and flat.) You can also get Kapton tape and wrap the coil.
Then again, a coat or two of corona varnish (or some precoated magnet wire) might be adequate at this fairly low "high voltage". Regardless, the general concept is to take care of the basic electrical insulation in as little thickness as possible, then you can concentrate on getting the best possible thermal contact, rather than be faced with a problem of having to maintain a certain thickness of thermal epoxy to maintain electrical insulation.
Mind you, with 250W spread over 256 square inches (if you mean 16 inch square, not 16 square inch), I suspect you could also get there with a ceramic heat sink (likely a 16" floor tile would work fine) and kiss your dielectric issues goodbye. It's not a particularly high power density. Contrariwise, suspending the coil and blowing air over or through it might work just as well.
If you mean what you wrote (16 First off, lose the 600V insulation - it's a thermal nightmare and does nothing for 3000V.
Corona varnish or Dupont Kapton sheets or tape will insulate 3000V without being nearly so thick. Basic idea if using conductive heat sink is to take the electrical insulation part out of the picture in as thin a layer as possible, then you can concentrate on getting the best thermal contact without having to maintain a layer of thermal epoxy of a particular thickness.
Given the low power density, a simple ceramic heat sink (ie, a 16 inch ceramic floor tile - no need to go all crazy expensive on the fancy ceramics) should work fine, and remove the insulation to the heat sink part of the problem. Plop coil on floor tile, wrap masking tape around the edges for a form, pour in epoxy to cover coil.
Another option would be suspending the coil in space (on some ceramic or phenolic supports) and blowing air through or across it directly.
Air wouldn't be the ideal heat transfer medium. Water has a much higher heat capacity, and with forced convection can shift a lot more heat. Pure - distilled or at least de-ionised - water is a pretty good insulator. One x-ray generator that I ran into had a water-cooled anode, 30kV above ground. Transformer oil doesn't have the same heat capacity - though orders of magnitude more than air - and can be a very good insulator.
Kapton really sucks thermally though. I was using a 1 mil piece of kapton as electrical isolation on a little TO-220 transistor. The transistor got a bit too warm at the maximum current. (~65C IIRC) I switched to a 10 mil Sil pad and the maximum temperature went down about 1/2 ~45C.
We've wound small-production custom flyback inductors using MIL-spec PTFE wire. It's not so bad- pretty thin insulation and rated for 200°C & 600V. No UL rating, but that wasn't a factor.
Any heat sink compound? Was it 1 mil film or tape with 1.5 mils of adhesive?
Except I don't think it does. There is maybe 5-6:1 difference between silicones (plain silicone should be similar to plain polyimide in thermal conductivity), but a 10:1 difference in thickness cancels that out, and then some.
Yeah Kapton tape. I don't think there is 1.5 mils of adhesive. But perhaps that was a contributing factor.
And I didn't use any heat sink compound. So maybe the sil-pad did a better job of filling the gaps. This was just one measurment done several years ago. Still I was surprised at the time that the thick sil-pad did better than the kapton tape.
A TO-220 can sit not-flat over an insulator and have air gaps. Grease probably would have helped. Sil-pads are thick and don't conduct heat all that well, but they are compliant.
Dow-Corning has amazing stuff - call your local rep, and he'll probably fix you right up! I once worked on a HV power supply about the size of a brick, that was potted in some clear stuff that dried rubbery, but it was about $300.00/ounce.
For a transformer, you want a soft compound as hard compounds can cause magnetostriction issues.
To gain thermal advantage, you add silica fibers (fiberglass) to it or silica beads.
It sounds like you do NOT perform a vacuum impregnation of transformer varnish. That can also cause problems as the potting compounds you want to use do not have the same degree of adhesion that the varnish has. I would impregnate it first, then start thinking about a thermal conduction medium.
That is also the standard rating for PTFE insulated wire.
He should make a Litz configuration using about 7 16Ga wires either with the insulation or using mag wire. Six around one nests nicely and will still yield the Litz effect as long as they remain insulated from each other (except at the ends).
It will likely handle a higher amperage and will assuredly perform better than the single solid conductor.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.