copper disks

If you are doing a purely conduction cooled design, then your enclosure needs to be a precise fixturing element for the PCB assembly. It is also a good idea to go to an .090 form factor. That's where those thermal bridge devices would/could come in on the heat generating device side of the board, coupling off to large pad segments, full of also thermally conducting vias.

Then a mere assembly time only added pad pressure will not do anything.

Once compressed, there is no ongoing force either. The stuff is compliant to the point that only the pressure generated at assembly occurs and that passes after assembly, *unless* you packed the can so tight that you cause a volumetric mismatch issue.

Good point. The Amazon disks are not likely to be very pure copper.

I can measure their electrical conductivity, which correlates to thermal. And I should allow my mechanical design to accommodate thicker disks; they are available in various gages.

Eventually, I have to verify the actual chip temperatures.

I have several product possibilities for chip-scale parts that will need heroic cooling, so some experimenting is worth it.

Thanks for that link. Page 6 is scary.

There is a little bit of variation in isotopic composition of copper from place to place but nothing to make any difference in practice. Given a sample and its trace impurities there is a sporting chance of saying whereabouts in the world it has come from. The purity does affect the properties a little bit so you get about 1% improvement by going "oxygen free" - important in a hard vacuum but otherwise not.

More of concern is that at least over here some of the stuff sold as copper for jewellry making is actual bright copper plated aluminium. The latter does not react well to being coated and hot fused in a furnace.

Sure you do. The specs and practices have been in place for a very long time, and as I said before, it really doesn't matter except during assembly.

It is like that quicksand stuff they make that people run across a pool of. If one stops... one sinks.

The pad material poses minor pressure during assembly and then fully abates within seconds.

It is like taking a fresh biscuit out of the *POP* container right after you hit it on the counter edge. Take one and squeeze your fingers together pinching the center. The 'back pressure' is only there as you move inward. Once you stop, so does the pressure. Then, there is actually an intimate 'vacuum' attachment created. It is like soft superglue, but not quite so aggressively attached. All air gaps are displaced (hopefully), and you have a state of the art conduction cooled brick.

Oh, the enclosure needs to be something like fully hobbed out Al or a mil spec plated magnesium or zinc type thing.

Hard Anodized Al is nice and non-conductive. You could laser ablate it after to expose places to tie grounds and mesh shield seals, etc. to. Hard Al is very hard done right. Even bead blasting has a hard time ablating it.

Those numbers are fine, as long as it is not too aggressive a heat source.

Then again, you should likely not be attempting to couple away too much to start with. Hopefully you are merely cooling an MCU or FPGA or such. The overall homogenous settled in temp is easy to manage that way. If there is an overly aggressive heat gereation source, you may have to find a way to case-couple it more directly.

We have a quick turn enclosure proto guy who does small parts too.

I had him cut me a brass washer with a groove in it for an o-ring.

That, I solder on to a standard PCB mount SMA connector long enough to get though my enclosure wall, and add the o-ring to make an IP67 level sealed component out of it.

A slightly parametrically modified version does the same thing for my SMC panel pass connectors.

He did all the little rounds and chamfers and everything.

Not even then.

Electrical might move enough to bother someone, but thermal cannot move enough to pose any problem. Conduction cooled designs are supposed to be a device which homogenizes all the internal heat sources into one overall (raised) operating temp. The conductivity of the copper or an alloy thereof is only going to affect the settling times, and even then you may need a microscope to see the change in the needle position on the meter face or stopwatch. It really is trivial.

Copper and Aluminum work. That is why we do not see many stainless enclosures when some form of conduction cooling is needed because thermals move through stainless so slowly that hot spots would result.

It would be nice to see a photo of the actual product and compare it to the computer-generated image.

There is some reason, physics or something, why electrical and thermal conductivity track. Most metals are about 140,000 K/W per ohm.

Some common copper alloys (see Chris' link) have 1/3 the thermal conductivity of pure copper. Some are 5%.

I did the CAD rendering as well, and it (the actual part(s)) look(s) exactly the same.

In my case, mushy is fine. It wouldn't be expensive to get some OFHC copper sheet, or an alloy that's not too badly polluted, and have someone punch up a bucket of disks.

It's all about the free electrons, they carry all the electrical and most of the heat current.

George H.

Yep, and when they get so hot that they radiate, we make diodes (valve) out of 'em.

Well if there are to be no machined holes or such... just sticking it on top, then mussy is OK. OFHC is probably spendy, where alloy 110 (or whatever the 'pure copper' number is.) might be fine. I always find McMaster-Carr to have pretty good info.

George H.

Brass seems to have about twice the thermal resistance per ohm. But my sources may be wrong; "brass" is poorly defined.

I have 60lbs of pre 64 pennies in a box. Let me know how many you need, I'll be thinking about a price plus shipping cost while I'm waiting. :-) I guess I'll will 30lbs each to my kids. I probably should split them before I'm to old to lift them. Mikek

Pure copper is the champ. I generally use thin (1/2 or 1/3-oz) FFC cables to connect to cold plates. It's way better thermally than a zillion tiny wires, and way easier besides.

Cheers

Phil Hobbs

snip

I bought a hundred 6" heat tubes for a project that uses a peltier pad to make a battery charging current.

These guys are expensive mil level stuff...

BTW folks. Heat pipes are a spin-off of NASA technology. NASA and Los Alamos.