Problems? See this thread at archive.org.)
My arctic thermal expoxy has gone bad in the tube, years old and i just need to stick a tiny copper heatsink to a small power regulator on an pcb board, now efficiency/thermal conductivity is not of major importance in this application, according to the link below Superglue Loctite 382 is good up to 130C
Read the link at the bottom of the post for a discussion on how well it works.
This totally your choice and i will bear no responsibility for any damage.
http://www.candlepowerforums.com/vb/showthread.php?44209-super-glue-as-thermal-expoxy
John.
-
-
No reply, I thought at least a few disparaging remarks might have been made, don't force yourself to make there though
John -
-
having been in a situation myself where proper TIM wasnt available, i just used mayonnaise. toothpaste or dental cream is also a viable alternative for short term situations
Sent from my Nexus 5 using TapatalkTomJGX likes this. -
-
About 9 times lower, that will reduce the heat transfer by basically a factor of 9 as it is directly proportional to the thermal conductivity of a material. 0.9 is very low as well. Here are some ballpark values for you:
Copper: 380-400 W/mK
Aluminium: ~250 W/mK
AISI 1010 steel: ~60 W/mK
Banana: 0.481 W/mK
Cake batter: 0.223 W/mK
Air: 0.0263 W/mK
Moral of the story, use sticky cake batter while you wait for that thermal tape
The resistance to heat transfer in your case can be written as t/(kA) t is the thickness, k the thermal conductivity and A the surface area. So either you have a very thin layer of glue (good luck with that) or your heat transfer will drop dramatically.Peon, King of Interns, jaybee83 and 2 others like this. -
I... think the assumption that the glue would actually transfer much heat at all is the most risky part
But I guess if you could somehow predict that the processor surface will hit 105 degrees, you have a two-component glue type that will melt at exactly that temp, achieve a perfect coating, and then run the cpu at that same temp for as long as you're going to use it.. Since things start happening with the glue when it hardens again. Then you might get a goop coat that has slightly better results than a good seating without any goop
It does work in practice, though. With glue, I mean. But for a different reason than why ceramics will work as a heat-conductor at very high temperatures. Because you get some physical contact with the metal, or have a very thin layer under the actual contact points. But when the glue hardens, the actual glue doesn't transfer heat at all, and then the point where it does transfer heat tends to become higher. As... some of us... bitterly learned after frying our 3dfx ram chips ..that happened to be possible to fry if the chip was partially covered in insulating material, even if the heatsink still transfered heat off the chip. And it actually did work until the glue hardened.
So it isn't necessarily a horrible idea to use a dot of glue to get some heatsink stuck on a chipset chip, now that the chips are more robust, and don't burn up or crash if the heat is 10-15 degrees higher on a small sector of the chip. I mean, it's better than tape. Or a rubber-band across the graphics card... Probably.
In the meantime - not going to recommend improvised home-cast ceramic/plastic compounds for cpu-goop, no matter how brilliant that idea is on paper.
-
As a bonus, you also get to a yummy slice of cake to eat afterwards
Supergule can be used as an basic thermal epoxy in a pinch
Discussion in 'Hardware Components and Aftermarket Upgrades' started by Tinderbox (UK), Oct 9, 2015.
