Sil-pad, substitutes

Kapton tape qualifies. I'll let someone else fight over whether it's better or worse than product X, but you can go look at the properties if you have the name. Perhaps Kapton MT is a variant with better thermal properties than "plain" Kapton, which is all I ever used in that line (may have been all there was at the time - 20th century and all)

Yes, Kapton tape does work well and it's easy to apply :)

Jamie

The challenge will be to keep the board in good flat contact with the heat sink. A sil-pad with adhesive on both sides would be great in that respect. But sil-pads are thick and don't really conduct heat very well. Some thinner double-stick tape would be about as good.

An anodized heat sink stuck to the board with thermal epoxy would be great. But that's messy.

Hard anodize is tough. It's a good heat conductor mostly because it can be very thin, 0.5 mil for modest voltages. A sil-pad will be 10 or

Do you really need the heat sink on the bottom? Do you really need to insulate? Will the board be solder masked?

Is there air flow? Can you use the things Lasse suggested, maybe top and bottom?

You'll get roughly 300W through that footprint with most materials, give or take a factor of two. If the surface is uneven, you'll need a thicker filler pad at the expense of heat transfer; materials Kapton and mica tend to do better despite their poor transverse thermal conductivity, just because they can be made so thin.

Tim

-- Deep Friar: a very philosophical monk. Website:

A bit O/T but recently we were given an electric treadmill ("walker") which wouldn't operate - kept tripping the earth leakage. In fact it had NEVER worked. Finally tracked the fault to a TO-200 with a live tab mounted to an earthed un-anodised heatsink, and separated by silpad. After cleaning, the silpad showed a small (~1mm dia) hole, perfectly circular, while the tab and the heatsink both had significant etching/oxidisation at the (silpad) hole site.

Really a question of chicken and egg. Was the hole the result of a dag which pierced the silpad at assembly time? (The QC tag about 2cm away on the heatsink was meaningless, this was a yumcha machine.) If so, the dag could presumably be eaten away by the fault current each time the machine was tried (up until the ELCB tripped - the donor had tried it repeatedly over a month or more).

Or was the hole in the silpad to begin with? If so, there was only an air gap between the tab and the sink. Maybe it did pass some QC test and escape from the factory in a working state.

Apart from that one example, I haven't had any bad experiences with silpad. But like any other approach (e,g, mica washers) the end result will be dependent on the mating surface finish QC.

There are a wide array of silpads with varying degrees of thermal conductivity. Some are much better than others.

The only thing to be careful of is over torquing the screws and pinching the metal tab through.

True. But a sil-pad has more than enough thermal conductivity for this particular application.

It's going into an existing application.

I wouldn't trust solder mask to insulate over time and scuffing. I need to investigate the thermal insulating properties of solder mask before I decide, but if it's going to mess me up thermally I won't use it.

I didn't see Lasse's post. Is it in this thread?

The customer would rather not have air flow, I'd rather not count on it.

Well, do the math. I generally divide their specified thermal conductivity by 2 or so to approximate reality. More, if the contact pressure isn't extreme.

The new tacky stuff seems to be better than the more rigid, rubbery stuff of old. Although these were probably at peak thermal conductivity before they failed, as you can tell from the contact pattern:

Downside to the tacky stuff is, if it's too tacky, it's like putty, and squeezes out of the joint and you don't know how much dielectric strength you have left. If it's too rubbery, it's not very conductive. I've never tried the self-adhesive stuff but it sounds dubious at best. It would still be better than nothing when you've got just a couple excess watts to suck out of a board. Good way to save on a fan or bulky heatsinks.

It's too bad that metal core PCB stuff isn't more readily available. Bergquist has a flier on their version, claiming all sorts of fancy plating and laminating possibilities, but I haven't heard anything about it beyond that. Sounds very elaborate and expensive for more than one layer and any kind of vias. That said, the only thing better would be a water pipe soldered to the backside!

Tim

never

Some of the fairly tacky stuff has fine fibre-glass mesh embedded in it. This limits the amount of compression and tends to hold it together.

John