Bonding and potting materials with good thermal properties

You obviously know absolutely nothing about Kapton.

Look at an old-style (or what we old folk think of as a normal style) electric stove element. Spiral and planar (other than the ends that connect.)

I have no idea what the OP is up to, but it's a perfectly understandable description of a coil shape.

--
Cats, coffee, chocolate...vices to live by

They saw you coming then. Even CONAP, the ONLY medium NASA approves for space HV encapsulation doesn't cost anywhere near that much.

The clear crap has no thermal qualities either. You have to use fillers. Practically no matter what you get. Fillers improve thermal properties. Clear "rubber", no fillers = poor thermal properties.

I'd break the structure up into smaller parts.

Spiral planar turns is a confusing description to me. Is there a magnetic media?

RL

OP might be better off using water cooled copper pipe than trying to glue a coil onto a heatsink plate that's going to take energy from the coil?

Someone here on s.e.d does that for induction heater "wiring" :)

Grant.

But not a suitable one for a simple power filter component, and one for which it could be very difficult to confine the magnetic and capacitive fields being generated.

If it were necessary to physically couple it to a heatsink, things would just get worse.

At 120KHz and 100A and 3KV, even a 30uH minimum energy storage element is difficult to fabricate with air as the medium. With a suitable magnetic structure, it might be possible to reduce turns sufficiently to reduce copper loss, while restricting flux induced current in heat-transfer, mechanical support or other hardware.

The use of electrically isolated and sectioned ferrite slabs, to short-circuit the pertinent magnetic paths ( to raise Q), while somehow avoiding an unacceptable increase in capacitive current, should be examined. I recon you'd want to avoid the turns actually carrying 3KV. Ferrite is a pretty good thermal conductor to couple into, and can have fairly regular surfaces to couple out from, even if the whole thing ends up potted.

If not cold-plate cooling, the volume assigned to the heatsink might more profitably be applied within the part dimensions itself. Isn't skin depth something like 0.2mm at 100KHz? 10AWG is more than 10x that, so bundled lower gauges might cut down on heat generated. This doesn't sound like the Harry I used to know...... Perhaps less attention to UL graded wire and a quick call to Nelco or VIP is in order? They can wrap wires and bundles with suitable Q of film and degree of overlap.

I won't suggest cooling the winding (tube conductor) directly with a cold-plate's intended fluid, or oil immersion, but there does seem to be an certain disregard for material cost indicated.

3KV ( AC? Peak? RMS? ) is a problem, not just as physical isolation, but as a corona inducement. Advice elsewhere re vacuum impregnation (before potting with higher viscosity thermally conductive media) is apt, as air paths will eventually work on any insulation or surface. Some forms of gap-pad have fairly high thermal conductivity and might be easier to use in an assembly, but they can't prevent air pockets.

Anyways, intriguing.

RL

h

It

r

he

on

Hey Seph, No I did not know, so I just measured and the inductance did drop by a factor of five when mounted on the Al plate. So that eliminates Al HS. I would like to keep the wire temp below 125C for practical reasons. This is a "one of" to be used in my lab test equiptment for testing Class E power amps. I would like to use easily obtainable, low cost parts but must meet electrical requirements. A ruggard unit is needed and your alumina sound fragel (sp) and expensive at 16" on a side. The thermal epoxy looks good if available. Seph, stop picking on JL, I think his quardrule sketches are priceless! I have a few on the wall. Cheers, Harry

Hey Rob, I really need the losses in this air core inductor. #10 AWG solid wire yields about 0.025R and the required 250W. Thanks, Harry

Originals? I think of JL as sort of the David Hockney of the schematic art world, but what I'd really like is an orginal from the Dutch Master Panteltje, complete with the original flux deposits.

Oh, Speff and I get along. We just rag one another now and then. He stayed at our house one night, passing through SF, and he's really a decent human being. For a Canadian.

John