Sil-pad/mica alternatives?

[Deactivated-1]

I need to mount a diode to a steel chassis, I have the diode on a large copper sheet that contacts the steel chassis in 2 places
The contact area to the chassis is 2 spots, 16x32mm with an M6 bolt in the middle. The steel chassis is 2mm thick mild CRS, the copper plate is 1mm thick, surface is not very level.
the diode will only output around 8 watts of heat but I want the diodes to be close in temperature so the bridge stays more-or-less balanced.

The voltage between the copper plate and the chassis is <20VAC, 0VDC.
I was thinking maybe plain paper, maybe soaking it in BLO so it doesn't absorb a ton of water, waiting for it to cure and then installing it.
maybe cooking paper, iirc it's like silicone impregnated paper so maybe that has higher than average thermal transfer
or many layers of saran wrap?

[tooki]

Why would you not use mica or silpad? It’s what they’re literally made for…

[Siwastaja]

Kapton tape is OK-ish. You want to make the surfaces and their edges 100% burr-free so that you don't accidentally cut through the tape, shorting the thing out.

If you don't have Kapton tape, then I guess all you can do is use any random packaging / office tape etc. Using thin tape helps with thermal transfer, at the cost that you need to be even more careful with deburring the edges and cleaning the surfaces to avoid punch-through.

Plastics usually are ten times less thermally conductive (so around 0.3 W/(mK)) than properly engineered silpad materials, so 0.1mm thick tape would correspond to a 1mm thick silpad.

Kapton seems to be around 0.8 W/(mK) so surprisingly good for a plastic tape.

[pqass]

See https://sound-au.com/heatsinks.htm#s7 for alternatives and their performance.

[mikerj]

Why would you not use mica or silpad? It’s what they’re literally made for…

Mica isn't made for insulating thermal pads.

Not quite sure what you are trying to say, but mica has been used to provide electrical insulation between power semiconductors and heatsinks for a very long time.

[jonpaul]

silpad is good, but we just eliminate the problem with floated HS.

j

[Andree Henkel]

What a stupid argument. What are actual Problems you face using Mica? Except that you don´t want to buy it?

modified citation:
"rubber (from tree) is a natural material, it was not actually made for making tires"

[Andree Henkel]

silpad is good, but we just eliminate the problem with floated HS.

j

well in my former company we had a rule: basically never let a metall part attach to a pcb be floating, you can run into a lot of trouble otherwise.

[tooki]

Why would you not use mica or silpad? It’s what they’re literally made for…

Mica isn't made for insulating thermal pads.

mica is a natural material, it wasn't "literally meant for cooling and insulating components from heatsinks"

No, but thermal insulators made of mica are intended for it.

(Also, some mica products are not straight mica sheets, they're composites made from ground mica. An absolutely manmade product.)

Because 1. they're clearly not that good for the job.

[Citation needed]

and 2. because I don't have any, and I'm not gonna order a big sheet of mica or expensive sheet of sil-pad just because I need to insulate 20 volts and barely need any thermal transfer.

So... you claim mica is "clearly not that good", yet you need barely any thermal transfer. So... what's wrong with mica again?

What type of plastic, or paper has the best thermal conductivity?

Paper? A thermal insulator? 


Just why do you think mica got used so much? Because it's not expensive, and works well. Many alternatives can get expensive real quick.

[Siwastaja]

mica is a natural material, it wasn't "literally meant for cooling and insulating components from heatsinks"

I personally only use supernatural materials, especially those from the ninth dimension.

[Conrad Hoffman]

Dumbest thread in a while, but what the heck, I'll play. Electrical insulators are, with few exceptions, terrible thermal conductors compared to metals. Mica, polyimide, Sil-Pads, paper, packing tape- all terrible! "Thermal" grease- terrible, even the fancy filled stuff. The successful ones are successful because they're thin and have good dielectric strength so they don't arc or punch through, not because they're good thermal conductors. Don't ever rely on grease for insulation. I shouldn't even have to say that, but some people have. Thin goes hand in hand with having decent flat surfaces. The minute you say the surfaces aren't that good, all predictions go out the window. There are some modern thick pads that are specifically designed for non-flat surfaces and they do way better than the old stuff. Are they good thermal conductors compared to metals? Nope, terrible. Keep your interfaces thin.

Now, if I were incredibly frugal (I am) I'd consider finding some discarded electronic item, say an old piece of audio equipment or appliance at the dump or on the curb, and seeing if it had some mica or sil pads inside. Odds are pretty good. If dumpster diving isn't your thing, I'd probably go with Mylar shipping tape because Mylar (polyester) is a known thing. I'd probably clean off the adhesive with solvent and use some thermal grease on both sides, unless the application is really non-critical. Then I'd just stick it on. Not obscure enough for you or not sufficiently labor intensive? Kodak made a lot of film using Mylar as the base. Find an old negative and soak it in chlorine bleach to remove the gelatin emulsion. Voila, clear Mylar. 

[Siwastaja]

Dumbest thread in a while, but what the heck, I'll play. Electrical insulators are, with few exceptions, terrible thermal conductors compared to metals. Mica, polyimide, Sil-Pads, paper, packing tape- all terrible!

That's true and good for first-order approximation (i.e., I agree one needs to learn what you said first). But there is a lot of variation from terrible to not-so-terrible.

For example, some random paper might be 0.1 W/mK, some random plastic 0.3 W/mK, a better plastic 0.8 W/mK, a typical cheap glassfiber reinforced Sil-Pad 3.0 W/mK, and a modern well-engineered non-reinforced (but it doesn't matter if you have the physical separation) Sil-Pad at 15 W/mK! While all of them are far from the aluminum heatsink itself (200 W/mK), the better SilPads are not that far. If the power density isn't ridiculously high (and you can always make it lower with a copper heatspreader), these can be useful even in large thicknesses like 2-3mm (sold under names like gap filler or gap pad). Typical example would be mounting a complete PCB assembly to a heatsink, with some tiny SMD passives in the way, requiring a conforming pad. In such case, let's say a DPAK which otherwise can dissipate only 1W can now dissipate 2-3W or so just fine, a big difference.

[Siwastaja]

Kodak made a lot of film using Mylar as the base.

Surprisingly, most of the color or B&W negative shooting stock was and still is, AFAIK, on celluloid triacetate base. Motion picture print films were polyester (trade name Mylar) for mechanical robustness (against wear and tear when projected gazillion of times).

Some negative stocks did use Mylar as you say because I remember people complaining how they curl more when drying but can't remember any specific examples now.

[Gyro]

The average Mica washer can be cleaved 2-3 times surprisingly easily with a razor blade on the edge. I can't remember where I first read it, but for low voltage applications (and with well de-burred surfaces), this can significantly reduce thermal resistance.

[ejeffrey]

Attempting to readibetween the lines because the OP has not said this explicitly it seems they have chosen a custom copper heat spreader that is not insulated from the diode package.  This means that off the shelf transistor insulation kits probably won't be the right size or have the mounting holes in the right place to mount the heat spreader.

So one question is why?  If your diode is a standard package there are probably insulating kits available for it that could attach it to the copper heat spreader.  Then you don't have to worry about insulating the heat spreader from the case.  If the power requirements are not very high, that's a pretty easy option.

If that isn't desirable for some reason, sil-pads are a lot easier to cut to size and add holes than mica, here is a $1.32 option that could be cut to size: https://www.digikey.com/en/products/detail/3m-(tc)/35MMX35MM-8810-SQUARES/3529589

Anything else will probably perform worse or cost a lot more than the sil pad, but kapton is probably the best of whats left considering cost, availability, thermal, mechanical, and electrical properties.  It just requires careful deburring and mounting to avoid piercing holes in it.

There is more than the interface material.  You need insulating bushings for the screws, and to chamfer and deburr the metal parts where you drill the holes in order to avoid shorting out at that point.  If you get a "TO220 insulating mounting kit" it will usually come with all the parts you need, but if you DIY it you will need to take care of that yourself.

[Conrad Hoffman]

Kodak made a lot of film using Mylar as the base.

Surprisingly, most of the color or B&W negative shooting stock was and still is, AFAIK, on celluloid triacetate base. Motion picture print films were polyester (trade name Mylar) for mechanical robustness (against wear and tear when projected gazillion of times).

Some negative stocks did use Mylar as you say because I remember people complaining how they curl more when drying but can't remember any specific examples now.

Many Kodak sheet films (4x5 stuff) were advertised as having "Estar thick base", which was nothing more than their name for Mylar.

[jonpaul]

Rebonjur,

We used many insulating tapes in magentics, Kapton is hard to cut or diecut and can be easily punctured.

The thin tpaes like 3M56 same issues.

Tiny burrs on the HS can eventually puncture the film.

The thermal R of the interface is easy to calc , a func to area, Delta T and thermal R.

See the many app notes and papers on thermal management

j

[TimFox]

Kodak made a lot of film using Mylar as the base.

Surprisingly, most of the color or B&W negative shooting stock was and still is, AFAIK, on celluloid triacetate base. Motion picture print films were polyester (trade name Mylar) for mechanical robustness (against wear and tear when projected gazillion of times).

Some negative stocks did use Mylar as you say because I remember people complaining how they curl more when drying but can't remember any specific examples now.

Many Kodak sheet films (4x5 stuff) were advertised as having "Estar thick base", which was nothing more than their name for Mylar.

Mylar^TM is a DuPont trademark, and Estar^TM is an Eastman Kodak trademark for the same thing:  polyethylene terephthalate (PET) polyester.

[Conrad Hoffman]

Kapton tape handles quite high temperatures so is generally preferred, but it isn't sitting on everybody's shelf, though IMO it should be. In this case almost anything thin would appear to work.

[coppercone2]

You can put ceramic pads, but they are even harder to engineer in, and you can get toxic dust from cutting them, and when you buy them, they always come in the wrong size for your heat sink

so when someone complains about silpads being hard, just show them ceramic

I use them to keep instruments consistent on restorations,.

[TimFox]

Your country flag isn't set.
If you are in the US, see  https://www.mcmaster.com/products/washers/material~mica/  or  https://www.mcmaster.com/products/mica/

[TimFox]

I try to give useful information.
Even though you reject mica for not-well-specified reasons, I thought you might like to see the specifications for normal mica.
Note that I did not disagree with you, nor demand that you use mica, I just gave you some factual information.
Perhaps someone else looking at this thread for information can make use of that site.

[sparkydog]

If I make a thread that focuses entirely on my question, it's hardly helpful for others, instead if I ask a question like "mica alternatives", it can be helpful for many.
otherwise the forum will turn into just "do my homework plz" instead of a place for helpful information.

There's a Stack Exchange for that. I see forums like this as rather a place for people to get together to discuss specific projects, as a way of gaining access to people with different knowledge that might not otherwise be available. A way, in other words, to learn by tinkering... and, yes, as a way to ask other people to check your "homework" because it didn't come out of a book that has a "right" answer and because you don't have a handy teacher to tell you if you got it right.

That's not to say that general questions aren't (or shouldn't be) allowed, but please don't try to turn this into another ivory tower where applied problems are off limits.

...my 2¢.

[Siwastaja]

Your question was formed OK, maybe you could have elaborated a bit why mica is not an option. "I have to do it with what I have available because I'm on a deserted island and can't mail order any mica or silpad" is understandable.

However, remember this is a discussion forum, not a question - answer site. We are not here to answer your specific question, but rather using it as a starting point. Which is great because then we will see more variance in the replies and the replies could help others to learn. You can also get inspiration from some ideas you did not think about.

8W through 2*16*32mm=1024mm^2 is not a problem. For example, if you use random plastic film 0.3W/(mK), with thickness of 0.2mm, you will develop 8W / ((0.001024 m^2 / 0.0002 m  ) * 0.3 W/mK) = 5.2 degC temperature difference over the thermal interface material. Hardly a problem.

Although, if the surface is, as you say, "not very level", if you effectively only use the thermal conductivity of 1/4 of that 1024mm^2 - then the temperature difference would be 20degC already, eating quite a bit into the margins.

Which brings us to the purpose of thermal interface material - it's needed to fill the gaps, providing contact over the whole area. Plastic sheets, mica, etc. are just insulators and need additional thermal interface material like thermally conductive grease. Sil-Pad on the other hand is a complete solution.

You might want to experiment with paper soaked in mineral oil (the oil would fill the gaps), but no one can give you guarantees how it works. Will the oil escape? Or get dirty and conductive?

I have had excellent results using bog standard double-sided tape. The thinnest type you can get, not the thick foamy stuff! When the surfaces are relatively smooth and level, the glue of the tape itself fills any microscopical voids, and the tape carrier being just some 100µm thick, it's OK even if not engineered to high thermal conductivity. And, when the surfaces are smooth and clean, the tape is also so strong it suffices for mechanical mounting. But no one knows how well it ages.

I'm a bit worried about the heatsinking capability of that steel chassis. I tried to google CRS steel as this was a new term to me, and , it can mean two totally different thing: cold rolled steel and corrosion resistance steel. This is why I hate acronyms. If you mean stainless, the thermal conductivity sucks greatly, some 15 W/mK. Normal carbon steel would be closer to 40-50 W/mK, still crap compared to aluminum, but not that bad.

It's a more complex modelling question, but for a rough idea, calculate the temperature difference when the heat tries to go through the 4 sides of your 16x32mm window. The total surface area the heat has to go through is 2*2mm*16mm+2*2mm*32mm = 192mm^2. Given carbon steel, this area and 4W of power, now each centimeter of depth means temperature drop of 4W / ((0.000192 m^2 / 0.01 m  ) * 45 W/mK) = 4.6 degC. The heat clearly can't go very far. Most of the dissipation has to happen within your coupling window area 16x32mm plus maybe 20mm in every direction, i.e., something like 50 x 70 mm. Now the silver lining is, 4W per such area is not a lot of power.

[Conrad Hoffman]

I once mounted a TO-220 device to a thin steel chassis wall and was amazed at how poorly the heat spread and  how how it got directly under the device. We added a spreader and it was sufficient, but just barely so.

There's a crazy expensive product, Aavid Ultrastick, that's a filled paraffin. Fancy term for wax. https://media.digikey.com/pdf/Data%20Sheets/Aavid%20Thermal%20Technologies%20PDFs/ultrastick-100300F0000G.pdf The stuff melts when the device heats up, giving you the thinnest possible layer and a liquid interface, and thus excellent heat transfer. It works very well, but it used to be way more affordable. Several other companies have similar stuff. One caution- don't use it in optical systems, especially vacuum, as it outgasses and will deposit on the optics. I suspect one could DIY something that would work as well- candle wax or a block of paraffin, plus a filler. The Ultrastick MSDS says it's 10-25% zinc oxide. That or aluminum oxide (alumina) powder would work. You could also melt some paraffin into a sheet of paper or card to good effect.