Getting the most out of your heatsink?

[mfo]

Hello everyone, I had a quick question in regards to getting the most out of your heatsink. How do you place your module inside of your heatsink? Do you just press it right in? Do you add some sort of thermal filler to bond it to the heat sink? I remember Larry saying that he used epoxy mixed with aluminum oxide powder. How effective would that be? What if one were to use something else like epoxy with copper powder since copper conducts heat better than aluminum? I was thinking of using silver grease since silver conducts heat better than both aluminum and copper. What are your thoughts on this? Thanks.

 

[gotwake424]

The lasers I have the builders used thermal paste between the host and heat sink.
But thermal paste can "dry out" and chip away (thanks for the correction HIMNL9)
I would go with silver grease.

 

[mfo]

Well i know that the lasers i have the builders used thermal paste between the host and heat sink. But thermal paste can burn-up so, i would go with silver grease.

hope that helped

That's what I was thinking. I mean they use silver grease for CPU's, so why not a laser...

 

[StridAst]

as an interesting note, there are many non metals you could dope a paste you are trying to use around a heat sink with, that conduct heat better then silver. The problem is abrasiveness, silicon carbide (most dark colored sandpapers are made from this) approaches the thermal conductivity of diamond. In fact, a synthetic diamond called moissanite is made from pure silicon carbide, and will pass a thermal conductivity test. (meaning cubic zirconia testers read them as diamond) catch is, it will scratch anything short of diamond. including sapphire or ruby. And last I checked optics are allergic to scratches...... However. If anyone is looking for a practical method for gap filling around a heatsink. You might want to try some silver foil http://www.pmc123.com/proddetail.asp?prod=Fine_Silver_Foil

Fine silver refers to 99.9% pure silver, which is commercially pure. It's also very malleable and soft. Would probably work quite well if you have too much space around your heat sink. best part is, filling with solid silver has got to be more conductive then any paste. As for how well it works in practice, not theory, I couldn't tell you, as I haven't tried it. yet.....

StridAst

 

[mfo]

as an interesting note, there are many non metals you could dope a paste you are trying to use around a heat sink with, that conduct heat better then silver. The problem is abrasiveness, silicon carbide (most dark colored sandpapers are made from this) approaches the thermal conductivity of diamond. In fact, a synthetic diamond called moissanite is made from pure silicon carbide, and will pass a thermal conductivity test. (meaning cubic zirconia testers read them as diamond) catch is, it will scratch anything short of diamond. including sapphire or ruby. And last I checked optics are allergic to scratches...... However. If anyone is looking for a practical method for gap filling around a heatsink. You might want to try some silver foil http://www.pmc123.com/proddetail.asp?prod=Fine_Silver_Foil

Fine silver refers to 99.9% pure silver, which is commercially pure. It's also very malleable and soft. Would probably work quite well if you have too much space around your heat sink. best part is, filling with solid silver has got to be more conductive then any paste. As for how well it works in practice, not theory, I couldn't tell you, as I haven't tried it. yet.....

StridAst

Thanks for the input. I do think however that I'd be better off using silver grease, do to the fact that the space is already tight in there. The silver foil looks cool, but it looks like it would be too much of a pain to improvise.

 

[HIMNL9]

Uh .....

Silicone based thermal paste don't burn (but can become "dry" with the time, and chip away)

Silver based grease is partially conductive, and if you use it on the back of your module and it go on diode pins .....

I suggest you to use thermal glue, something like arctic silver alumina if you want to made the assembly permanent, and something like fujik thermal silicone glue if you want to keep the possibility to dismantle it.

Other than this, glues don't spill around (after hardened) as sometime greases do, especially the first times they get hot

 

[ZRTMWA]

The difference between copper ans silver in terms of heatsinking is very minimal. I doubt it would have any affect on the heatsinking of a diode if you used copper or silver. But copper is cheaper.

See here: Thermal conductivity - Wikipedia, the free encyclopedia

StridAst: Are any of those compounds available for sale? I've never heard of silicon carbide or moissanite but they sound like they would be expensive and hard to get (unless you could remove the slicon carbide from the sand paper).

Edit: This article helped a little:
http://en.wikipedia.org/wiki/Silicon_carbide

 

[gotwake424]

Uh .....

Silicone based thermal paste don't burn (but can become "dry" with the time, and chip away)

Silver based grease is partially conductive, and if you use it on the back of your module and it go on diode pins .....

I suggest you to use thermal glue, something like arctic silver alumina if you want to made the assembly permanent, and something like fujik thermal silicone glue if you want to keep the possibility to dismantle it.

Other than this, glues don't spill around (after hardened) as sometime greases do, especially the first times they get hot

Thank for the correction

 

[Krutz]

well, better than any glue/epoxy/paste would be directly combining the metal parts. thats why i only use copper for heatsinks (whenever possible), and solder both the aixiz into the heatsink, as well as soldering the heatsink into the flashlight-host.
..i posted pics here:
http://laserpointerforums.com/f51/presenting-mokume-heatsink-43427.html

edit: wont work with greenie-modules, of course! well, maybe with super-low-melting solder.. some indium (130°) or gallium (30°!), or better some nice alloy from both, 60° melting point shouldnt hurt the module too much?

manuel

 

[StridAst]

The difference between copper ans silver in terms of heatsinking is very minimal. I doubt it would have any affect on the heatsinking of a diode if you used copper or silver. But copper is cheaper.

See here: Thermal conductivity - Wikipedia, the free encyclopedia

StridAst: Are any of those compounds available for sale? I've never heard of silicon carbide or moissanite but they sound like they would be expensive and hard to get (unless you could remove the slicon carbide from the sand paper).

Edit: This article helped a little:
Silicon carbide - Wikipedia, the free encyclopedia

Silicon carbide is sold free of sandpaper, and fairly inexpensive per weight. again, *very* high thermal conductivity, and *very* high hardness which means it WILL scratch ANY optics it touches. Not might, will. however if you want to try using it. remember to embed it in some sort of glue first. there are several places that sell silicon carbide grits. and of course the finer the grit, the smaller spaces it will fit in, the larger the grit the larger caps would fill nice with it. Most lapidary stores sell it, and virtually any place that sells equipment for rock tumblers will sell it. here is an example
Silicon Carbide - Grit - Kingsley North

It can be found in brick form from some places, however the only way to cut it is with diamond abrasives. (diamond is the only thing harder than it that is commercially available. there ARE synthetic substances that would work, but they are made in labs, and not in large quantities.)

Again, use at your own risk.

I would HIGHLY suggest using the silver foil I linked to above if you have much of a gap to fill around your heatsink, the ideal method for heatsinking would be soldering a heatsink to your module of course as stated, however fine silver, esp in a foil form, is MUCH softer then you would expect. it should be malleable enough to work if you need to make a module fit a heatsink that's loose.

StridAst

 

[mfo]

Silicon carbide is sold free of sandpaper, and fairly inexpensive per weight. again, *very* high thermal conductivity, and *very* high hardness which means it WILL scratch ANY optics it touches. Not might, will. however if you want to try using it. remember to embed it in some sort of glue first. there are several places that sell silicon carbide grits. and of course the finer the grit, the smaller spaces it will fit in, the larger the grit the larger caps would fill nice with it. Most lapidary stores sell it, and virtually any place that sells equipment for rock tumblers will sell it. here is an example
Silicon Carbide - Grit - Kingsley North

It can be found in brick form from some places, however the only way to cut it is with diamond abrasives. (diamond is the only thing harder than it that is commercially available. there ARE synthetic substances that would work, but they are made in labs, and not in large quantities.)

Again, use at your own risk.

I would HIGHLY suggest using the silver foil I linked to above if you have much of a gap to fill around your heatsink, the ideal method for heatsinking would be soldering a heatsink to your module of course as stated, however fine silver, esp in a foil form, is MUCH softer then you would expect. it should be malleable enough to work if you need to make a module fit a heatsink that's loose.

StridAst

From my understanding of using foils (except lead foil) they all seem to tear very easily. If I were to use the foil, wouldn't I have to pull it very hard in order to wrap it very tightly around the module? Wouldn't it tear if I did that? When I load it into the host?

 

[StridAst]

tearing is possible, however it usually stretches and deforms more easily then it tears. one of the characteristics of fine silver is it's ability to deform easily without fract however it deforms much easier then it tears. In short, you can bend, draw, mash, or stretch silver MUCH easier then say aluminum foil. Different metals, different working properties. Ideally the way I would try if someone wished to, would be to push the module into the center of a small sheet of silver foil, kind of like poking your finger into the center, so it covers the end and wraps around the sides, then poke it into the heatsink, forcing the module into place, then cut any silver off the front with an exact o knife. The silver SHOULD deform enough to fill any gap, with minor, if any, tearing, although it would depend on the gap in question. if the gap is less then 30% of the thickness of the silver foil, then it probably will tear. this is a rough estimate, not a promise, as there are different things that can impact. most notably: any sharp surface, including dust, can and will cut the foil, starting a tear, and if the foil is already work hardened, and was not annealed. While copper does conduct heat well, it is also MUCH less ductile and malleable than silver. and aluminum is quite brittle by comparison.

If you have worked with lead foil, and felt how soft it is and easily deformed, silver is MORE ductile AND malleable then lead. i.e. softer, and easier to make work in this regard. It has been a while since I have worked with silver foil, and the company I last bought some from has discontinued carrying it. so I have not bought from the company I linked to, I simply googled it and posted the link. I would advise before buying it, calling or e-mailing them and asking if the silver foil is annealed. A common term for the annealed silver is "dead soft" simply verify that it is before purchasing it.

Ductility - Wikipedia, the free encyclopedia

StridAst

 

[mfo]

tearing is possible, however it usually stretches and deforms more easily then it tears. one of the characteristics of fine silver is it's ability to deform easily without fract however it deforms much easier then it tears. In short, you can bend, draw, mash, or stretch silver MUCH easier then say aluminum foil. Different metals, different working properties. Ideally the way I would try if someone wished to, would be to push the module into the center of a small sheet of silver foil, kind of like poking your finger into the center, so it covers the end and wraps around the sides, then poke it into the heatsink, forcing the module into place, then cut any silver off the front with an exact o knife. The silver SHOULD deform enough to fill any gap, with minor, if any, tearing, although it would depend on the gap in question. if the gap is less then 30% of the thickness of the silver foil, then it probably will tear. this is a rough estimate, not a promise, as there are different things that can impact. most notably: any sharp surface, including dust, can and will cut the foil, starting a tear, and if the foil is already work hardened, and was not annealed. While copper does conduct heat well, it is also MUCH less ductile and malleable than silver. and aluminum is quite brittle by comparison.

If you have worked with lead foil, and felt how soft it is and easily deformed, silver is MORE ductile AND malleable then lead. i.e. softer, and easier to make work in this regard. It has been a while since I have worked with silver foil, and the company I last bought some from has discontinued carrying it. so I have not bought from the company I linked to, I simply googled it and posted the link. I would advise before buying it, calling or e-mailing them and asking if the silver foil is annealed. A common term for the annealed silver is "dead soft" simply verify that it is before purchasing it.

Ductility - Wikipedia, the free encyclopedia

StridAst

Thanks for this information, very informative and helpful indeed. I will reconsider this as an option now. + rep to you.

 

[HIMNL9]

@ StridAst: uhm, i have at home carborundum, if you mean this

This is 400, have also 800, somewhere .....

And yes, it grind well ..... sometimes also too much ..... never tought to use it as thermal compound in glue making, anyway ..... must test also this, the first time i have a moment for that (but suppose that for better results, i need to get 4000 or 6000 mesh )

 

[StridAst]

yes, that would be silicon carbide. (different name, same thing) however yes, you are correct. That is a fairly course abrasive shown there. simply too large of grain size. the finer you get the better. main reason it's not heavily used in thermal glue is probably the sheer abrasiveness.

StridAst

Again, silver foil would probably be the FIRST pick for thermal conducting, a high heat conductive substance like silicon carbide would be a distant second

 

[ZRTMWA]

Couldn't silicon carbide be sintered though?