Mica and Goop

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Jump to Latest
D

dotneck335

Member
Joined 2012
I don't know which version of Kapton it is----write to Rod and ask him. He is not only one of the most knowledgeable audio gurus on the planet, but as straight a shooter as you will find. He usually answers emails in a timely manner. But since he is recommending it for his projects, it's gotta be all you need.
 
L

lhquam

Member
Joined 2010
dotneck335 said:
I don't know which version of Kapton it is----write to Rod and ask him. He is not only one of the most knowledgeable audio gurus on the planet, but as straight a shooter as you will find. He usually answers emails in a timely manner. But since he is recommending it for his projects, it's gotta be all you need.
Click to expand...

Farnell has Thermal Pads using Kapton MT shown here: POLYIMIDE FILM | Premier Farnell CPC UK. I wonder how their performance compares to Sil-Pads and some of the other commonly used insulators.
 
A

AndrewT

R.I.P.
Joined 2004
AndrewT said:
The 30m roll of 25mm wide koptan arrived from China.

Does the spelling mean it is not Kapton?

The plastic film plus adhesive has a total thickness of ~1.6thou (~1.6mil, 0.04mm) making it fairly good as a Thermal conductor.
I'll try to measure the cleaned film later.

I have noted in the past with pre-applied adhesive that it flows out of the interface when under pressure from clamping and with a bit of heat from the semi device.
I'll see if this version does similar.
Click to expand...
update:
As reported in another Thread this koptan does resist temperature.
ordinary packing tape shrivels and wrinkles when touched with a soldering iron @ ~<150°C
This koptan does not seem to suffer any change with a few seconds touching of an iron @~300°C.

I have taken apart a very high bias ClassAB which was dissipating 15W through each output device for many hours. Tc~60°C and Ts~50°C
I can see some migration of the adhesive to the edges of the devices. The temperature and pressure (of clamping) has forced the adhesive to flow towards the low pressure areas. I can see bulges where the adhesive has escaped from that pressure.
This must result in a thinner average thickness of adhesive under the device and that can only help in reducing the Rth c-s of the koptan insulator.

I will continue to use the misspelt version of the name for my insulating, thermally conductive tape, since I have no idea whether it really is "Kapton".

I am pleased with what I have seen in this short term test.
 
Last edited:
A

AndrewT

R.I.P.
Joined 2004
bwaslo said:
OK, I had some time I had to be down in the shop, so did the mica and the nuthin tests.

Same heatsink, same FETS, same power (25W per TO-220 device). Same heatsink temperature (or pretty close anyway).

I had to run the 'nuthin' test twice, as I was seeing one FET about 7 degrees higher than the other. Apparently some crud had gotten under the part when I took the alumina insulator out (?). I hadn't cleaned, just smoothed and added a little goop first time. Whatever, I cleaned off the FETs and the heatsink, put new goop on the FETs and did it again.

One of the interfaces done with the premium nuthin measured 0.16C/W, the other was 0.144C/W. Average 0.152C/W.

With new mica insulators, one was 0.53C/W, the other was 0.71C/W. average is 0.62C/W. The difference between the two seemed like a lot, so I checked under the parts and insulators, all looked clean. Then I wiped off the insulators and measured the thickness of each (using an ancient micrometer that was my grandfather's). Unfortunately I lost track of which was which, but one measured 2.0mil (51um), the other 2.5mil(63um). Not a huge difference, and I don't know whether the thicker one was on the FET which had the higher temp, though the ratios aren't too far apart.



The results are a lot more in favor of the aluminas than I would have thought. The tests were done on different days, though, tightening of screws might not have been the same (I don't have a torque wrench, and was going by when I saw deflection on the long allen wrench I use on the caphead screws). There might have been differences with the contact of the thermocouple, that was hard to get stable both days.

So, if taken at face value, what's it mean? Well, if your parts are getting close to their max (125C for most devices, 175C for some MOSFETS), then do whatever you can, alumina, nuthin, or BeO if you happen to have a stash put away ...and don't expect the amp to be passed down as an heirloom to your decendents. If you have margin, then expected semiconductor life is said to double with every 20C decrease in junction temperature. If the thin alumina insulators are indeed 0.43C/W better than mica, then you'd have to be dissipating 46.5W in the TO-220 device to double the part's life by switching to alumina. I'm dissipating about half that, 25W, so I gain 10.7C with alumina, or about 45% more expected life. This is all assuming failure was not caused by other things, and that the semiconductor heat/age is what will be making the amp eventually die. I'll be operating the junction 95C below its max allowed (84C below if with mica), so the FETS should live long and prosper . But since I got the thin alumina insulators, I'll use them of course. Anything worth doing is worth overdoing, right?

______________________
______________________
Click to expand...
I went back and read Bwaslo's post again.
For the test I just did with 15W passing across a To264 device I think I am seeing ~2Cdegrees difference between the heatsink temp and the device temp. This is not at all accurate since I cannot measure the Tc under the device, nor can I measure the Ts under the device.

But if any other insulator can halve the deltaT, then the increase in lifetime or increase in power handling would be difficult to prove given the potential for ~1Cdegree reduction in temperature.
 
Last edited:
P

plasnu

Member
Joined 2011
anatech

anatech

Administrator
Joined 2004
Paid Member
I'm still using the Wakefield type 120 goo. It has always worked well, and if its good enough for industry, it is more than good enough for our stereo equipment. Don't forget that the expected service life of anything industrial is far longer than audio systems or computers. Yet, we expect industrial life-times from our commodity audio equipment.

I have to agree with others. Do not use Arctic Silver on anything but a CPU heat sink.

-Chris
 
wrenchone

wrenchone

Member
Joined 2004
I suspect that the expiration date on heat sink compound might have something to do with the constituents separating/settling out. I picked up a couple of largish tubes of Wakefield 120 from a local surplus house at a bargain price. When I opened one up, what came out was almost pure silicone oil with only a trace of filler. Kneading the tube for a while restored the stuff to expected color/consistency. Fortunately, the tube is plastic, and stood up to the kneading without splitting. It took a lot of kneading....

There was no expiration date on the tubes, but I suspect that expiration was the reason the stuff ended up at the surplus store.

BTW, telecoms won't use the usual silicone-based heat sink compound because of the tendency of the silicone constituent to migrate and end up in places where it shouldn't, like relay/switch contacts. This possibly accounts for the "drying" of old thermal joints as the silicone migrates out to greener pastures...
 
Last edited:
L

labjr

Member
Joined 2003
wrenchone said:
I suspect that the expiration date on heat sink compound might have something to do with the constituents separating/settling out. I picked up a couple of largish tubes of Wakefield 120 from a local surplus house at a bargain price. When I opened one up, what came out was almost pure silicone oil with only a trace of filler. Kneading the tube for a while restored the stuff to expected color/consistency. Fortunately, the tube is plastic, and stood up to the kneading without splitting. It took a lot of kneading....

There was no expiration date on the tubes, but I suspect that expiration was the reason the stuff ended up at the surplus store.

BTW, telecoms won't use the usual silicone-based heat sink compound because of the tendency of the silicone constituent to migrate and end up in places where it shouldn't, like relay/switch contacts. This possibly accounts for the "drying" of old thermal joints as the silicone migrates out to greener pastures...
Click to expand...

I bought a couple large tubes of Aavid Thermalcote at a flea market 25 years ago and I haven't noticed any separation. The guy had cases of them in cardboard boxes for $1. I wish I bought more.

I have noticed other brands like Radio Shack that were separated when new. Maybe it's just the way some of them are formulated.

If you use a lot there's an ebay seller that has 1LB tubs of Thermalcote for about $25 shipped. That's a lifetime supply for most and you can stir it up when it's in a tub.
 
wrenchone

wrenchone

Member
Joined 2004
After reading all the stuff about silicone creep in the literature, I may try transitioning to a no-silicone formulation that stays put, if there's no penalty in thermal resistance.

I have some tubes of the Thermalloy goop that have weathered well (It'd be nice if I could actually find them at this late date), but the Wakefield 120 definitely separated.
 
wrenchone

wrenchone

Member
Joined 2004
I've done some cursory nosing around, and Wakefield 126 seems to be the major non-silicone thermal paste stocked by the "big two" (Mouser and Digi-Key). It's about 6% lower in thermal conductivity (6% worse) than their silicone-based 120 paste, but will not creep, run, contaminate solder baths, gum up conformal coats, or sleaze its way into switches and relays. The Aavid non-silicone paste is actually better rated than their standard silicone formula (which is equal in thermal conductivity to Wakefield 120), but for some reason, the big distributors don't stock it. Maybe they know something I don't... I'll check Newark/Farnell to see what they have.

There are also some advanced greases loaded with higher conductivity solids (boron nitride?) that can be used if you need a little bit of extra umph. I have the MG Chemicals version of the hi-test compound floating around, but I haven't had sufficient incentive to break it out yet.

Edit -
Newark has Arctic Alumina in 14g syringes. this may be worth a look, as it is a non-silicone compound. I find it a bit troubling, though, that no figures on thermal conductivity are given in the data sheet.
 
Last edited:
anatech

anatech

Administrator
Joined 2004
Paid Member
Hi wrenchone,
I still use Wakefield 120 and buy it in the tubs from the big two. I just keep it mixed and it's fine. I find lot's of telco equipment where normal silicone grease is used, so I wouldn't say they don't use it! Most of the industry is going to sil-pads of varying effectiveness because of the cost savings in labour. It has zero to do with supposed problems with silicone based grease or any performance advantage claimed for some of those (expensive) pads. They look nice and neat, I'll give them that. I have used high performance pads (single use), and they are a pain to clean off if you need to service the device. You cannot reuse the pad for another part either. Truly evil things. Cyrus uses them in consumer products.

On the other fillers, I would stay completely away from anything conductive!

-Chris
 
wrenchone

wrenchone

Member
Joined 2004
I am intrigued by a product that may resist caking up due to silicone creep.
And yes, a lot of the things I design are Class A amps with high thermal loading on the devices, which is something I would consider "critical".

At any rate, I see no cogent argument for not using a non-silicone compound. I intend to try it. The Aavid stuff is available for real cheap on E-pray. This may be an opportunity or a warning, considering that none of the big three (Dig-key, Mouser, and Newark/Farnell) are stocking the Aavid thermal grease (listed, but zero stock level). Is it being dumped due to problems? Wakefield 126 is about $21 for a big tube. I've made worse investments involving more money.

One small caveat, though. The non-silicone paste cleans up with isopropanol, so if you aggressively wash a board with IPA after using the non-silicone paste, you may wash some of it out of the joint, and deposit it elsewhere on the board.
The usual silicone goo cleans up with board wash, which is usually formulated with crude hexane (several isomers). The same caveat applies for aggressive washing here, too.

BTW, One of the reasons I specifically mentioned Arctic Alumina is that it uses a non-conductive filler, unlike Arctic Silver...
 
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Top Bottom