white paste

The white paste is heatsink compound, which is silicon grease loaded with magnesium oxide. In the UK I find the easiest brand to get hold of is servisol from Maplin, any electronic component distributor worth dealing with should stock an equivalent product.

"I.F." wrote in news:E9R2g.57197$ snipped-for-privacy@newsfe4-win.ntli.net:

Mmmm, goop. :)

Once you get it, remove the old with a lightly parafin-soaked rag. (Or white spirit), taking care not to rub against plastics which could build a dangerous static charge of you did so.

Then apply as little of the goop as you need, Don't be tempted to over do this, it's hard to squeeze out the surplus, and you'll get some VERY overheated transistors in a guitar amp if you make that film too thick. If the heatsink surfaces are very flat, a thin semitranslucent smear is all, if they're standard extrusions, make it slightly thicker, definitely a milk-white smear, but no more.

I second the very thin layer. It is not a very good conductor of heat, just beter than air. The object is to fill in all the tiny spaces where you do not actually have metal to metal contact. If too thick , then it will act as a heat insulator and the transistor will actually run hotter.

As the paste is expensive, I don't like wasting it either! But if it doesn't squeeze a little out round the edges as the fixing screws are tightened - then there are probably voids between the mounting surfaces, and air bubbles - no matter how highly compressed, make very poor heat conductors!

That is the biggest load of crap I've ever heard! - The manufacturers would be prosecuted under trade description legislation if their HEAT SINK COMPOUND was only slightly better heat conductor than air! Shit is a slightly better heat conductor than air - which would put it on a par with your brain!

The only other possible explanation is that you are spreading mis-information to make mischief!

"I.F." wrote in news:3FR2g.3030$ snipped-for-privacy@newsfe5-win.ntli.net:

True. Still, you'll not that likely get bubbles if you make it thin, far more likely if it's thicker. :) (Because it would be more uneven).

Ideally, if the surface is big, make the smear of even thickness as I described before, then use a fingertip to sweep inwards from all angles across the plane, so it is translucent at the edges, and slightly and slowly rising in thickness to the centre. Sounds tedious but it's easy to practise with few quick light strokes, and the result is excellent. Maybe fit it by making one long edge take full contact then hinging it into place, then (with just a few degrees of rotation) twist one plane against the other a few times, under gentle pressure, till the goopy resistance to that motion is consistent and strong. You won't have any bubbles after that.

What I've just described is the standard method for mounting Peltier coolers for high performance, so if it's good for those it should be good foe pretty much anything.

Thanks, everyone!

Since there is already a thin layer there, I'll just leave it as is instead of adding more.

If it wasn't for you guys, I would have made the mistake of putting a thicker coating on.

Alan Whittier, California

"I.F." wrote in news:7TR2g.3035$ snipped-for-privacy@newsfe5-win.ntli.net:

Steady, it's not that bad..

It's true, if you have a thickish layer it is a poor conductor of heat. Same would be true for metal loaded epoxy. This stuff works on the assumption that the thickness is very small compared to its area, and if you have too much you have too high a thermal resistance. It's designed to take high compression, and if there's too much there, the cushioning make that impossible.

I've seen the burned out transistors in power amps where too much was used, and repaired them by using a properly applied thin film after replacing parts. Once you see that, you'll know it's true that it's a poor thermal conductor if improperly used.

The only problems with putting too much on are the cost of the paste and what squeezes out when you re tighten the heatsink is messy! In the years when I made my living servicing PC monitors & PSU boxes, about 4 or 5% of PSU blowups were due to insufficient heatsink compound. In every case, replacing the blown components and using sufficient paste provided a permanent repair.

"I.F." wrote in news:_5S2g.2240$ snipped-for-privacy@newsfe3-win.ntli.net:

I've made money from repairs too. I've seen the damage that happens when a power amp got too much paste to allow proper compression and conduction of heat. Fixed it too, with proper method. I said this twice now, because I don't want this point to get lost. If thick film was the way, there'd have been loads there already, and the Mysterieux said there wasn't.

Maestro Mysterieux, I've often left the original there like you're suggesting. Check that it's ucontaminated by even a single hair though. And smear inwards like I said in that other post. This will give it a bubble- free profile and also remove surface contaminants to a large extent. But if it's got furry while you were working, replace it for sure.

Lostgallifreyan wrote in news:Xns97AEE30026B83lostgallifreyangmail@140.99.99.130:

Small followup: Failures due to too little goop usually happen on small surfaces. Failures due to too much goop usually happen on big ones.

PA and guitar amps usually have bigger heatsinks than the surfaces involved in the large transistors used in PSU's because they mount several devices on one plate them mount that to an even bigger one, or to the chassis.

It's easy to squezze a little excess from small parts, but very hard to do it from between two plates of metal.

The claim that provoked my rebuttal was that bad! The suggestion that heatsink compound has only slightly better thermal conductivity than air is ridiculous, and persuading the OP to damage their equipment by using insufficient heatsink compound is outrageous!

Or maybe they spread the heatsink compound that thinly themselves, and feel that the heatsink is running cool (But the transistor isn't!) - Oh! that's working well then.

I never suggested using excessive amounts - I was rebutting someone who was suggesting scraping it so thin it has no chance of filling any voids, and from their comments on the efficiency (or lack of) of the compound - they even had me wondering if cooking oil would work better!

Some of the cheaper manufacturers still use plain silicon grease which is a lot less efficient than the white paste - but still orders of magnitude better than air.

"I.F." wrote in news:_iS2g.3042$ snipped-for-privacy@newsfe5-win.ntli.net:

Point taken about the comparison to air, but the rest of the post makes sense.

[quote] I second the very thin layer. It is not a very good conductor of heat, just beter than air. The object is to fill in all the tiny spaces where you do not actually have metal to metal contact. If too thick , then it will act as a heat insulator and the transistor will actually run hotter. [/quote]

I've said plenty now about how to optimise for large surfaces such as in power amps, so I won't add to that, but this guy did make the main point, that metal to metal contact is all, and the goop is only to fill gaps. Anthing that prevents that is badm be it too much or too little, and context (as in size of surface and amount of goop) is all.

"I.F." wrote in news:hvS2g.3080$ snipped-for-privacy@newsfe5-win.ntli.net:

You'd said this:

[quote] The only problems with putting too much on are the cost of the paste and what squeezes out when you re tighten the heatsink is messy! [/quote]

That could be just as dangerous with a large surface. I agree with you for small ones though, I do it just as you say for those. But guitar amps built as the first poster describes are definitely an exception. Better to go with the advice as for fitting Peltiers if you want to get those right.

It's also rather subjective, I have a few old stock tubes of paste that have dried out so the paste is a bit stiff, if this was put under a large plate heatsink the heatsink would visibly bend as the screws were done up - I can get away with using this on smaller items like TO220 cases so long as I spread it evenly. More recently purchased tubes of the product I have found rather runny, in fact it is so runny it could be painted on with a brush - I could chuck the whole tube on a large heatsink and the excess would squeeze out with the screws not much more than finger tight!

"I.F." wrote in news:ROS2g.48440$ snipped-for-privacy@newsfe6-win.ntli.net:

Hell, mine's not even consistent within a single tube. >:) True about what you said, about the bending if you apply pressure on a thick smear under a large plate. That's why I do that smear-to-the-centre thing, if it's going to bend slightly anyway, best that you work with that fact to minimise any bad effects it might have. That'a also the reason for the slight rotation under pressure before final clamping, it tends to shift the stuff outwards. The smearing takes care of inconsistency, as does the rotation afterwards. The really runny stuff I don't like, when the tupe does that I soak up the whey to leave the curd. If the tube was so dried the stuff was chalklike and wouldn't squeeze without straining my hand, I'd bin it. I've still got the same tube after many years though, they last a LONG time if you don't use a lot at once, and it hasn't dried out. (Is in a syringe, 10 ml, I think I've bought two in ten years or more. If I'd worked full-time I'd have got through more, but it will always last a lot longer than a large reel of solder.

On TO220 I usually put a bead in the middle and squash it toward the edges a bit with the end of the nozzle before fitting, like a small-scale crude version of the big-plate treatment.

I've sometimes wondered if toothpaste would do, at a pinch.. :)

Give it a try and let me know.......................................

The stuff you buy for babies, "desican", not sure of spelling is same as in heatsink compound except for the grease is not silicon. I saw this talked about in a ham group. It will smell a litle at first but works the same. But take taht with a grain of salt as I have never tried it myself. JTT