Different HS "paste"

16th May 2009, 23:23

This might not be as crazy as it first sounds.

"Pure" gold is very very soft.
Couple mil thick gold foil, 1" square should cost less than $10. Conductivity should be about 75% of pure silver, but higher than most silver filled pastes. About the same as Arctic Silver 5.

Purer grades should be soft enough to conform to surface imperfections.

Silver foil would have higher thermal conductivity but may suffer from increase in thermal impedance at silver-heat shink and silver-cpu interfaces because of oxidation.

Is this too crazy even to think about trying a foild instead of a paste?

Any heat sink guru out there feel like trying it?

jmke · 17th May 2009, 00:08

the possible increase in conductivity is not worth the cost; even now the difference between cheapo paste and expensive after market is only a few degrees at best. if you can manage to squeeze another °C out of your HSF with gold paste it would be a lot

leeghoofd · 17th May 2009, 08:13

I agree with Jmke, the cost is not worth the expected small gain...

You also have to verify if your IHS and Cooling block are 100% flat. If there's already a small deviation (concave, convex) not even "gold" will do it

17th May 2009, 14:19

Good point on the flatness.
Good point on how big the gain is...I was thinking in terms of reducing the interface thermal impedance by maybe up to a factor of 2. This sounds good, thus my post.

But if you do the math even a very hot OC CPU with proper thickness on the best thermal paste has temp drop over the paste of only 2 deg max. So twice as good is not much. Its the diminishing returns scenario.

Thanks for setting me straight.

16th May 2009, 23:23	#1
photonblaster Posts: n/a	Different HS "paste" This might not be as crazy as it first sounds. "Pure" gold is very very soft. Couple mil thick gold foil, 1" square should cost less than $10. Conductivity should be about 75% of pure silver, but higher than most silver filled pastes. About the same as Arctic Silver 5. Purer grades should be soft enough to conform to surface imperfections. Silver foil would have higher thermal conductivity but may suffer from increase in thermal impedance at silver-heat shink and silver-cpu interfaces because of oxidation. Is this too crazy even to think about trying a foild instead of a paste? Any heat sink guru out there feel like trying it?

17th May 2009, 00:08	#2
jmke Madshrimp Join Date: May 2002 Location: 7090/Belgium Posts: 79,021	the possible increase in conductivity is not worth the cost; even now the difference between cheapo paste and expensive after market is only a few degrees at best. if you can manage to squeeze another °C out of your HSF with gold paste it would be a lot __________________

17th May 2009, 08:13	#3
leeghoofd [M] Reviewer Join Date: Dec 2008 Posts: 3,209	I agree with Jmke, the cost is not worth the expected small gain... You also have to verify if your IHS and Cooling block are 100% flat. If there's already a small deviation (concave, convex) not even "gold" will do it

17th May 2009, 14:19	#4
photonblaster Posts: n/a	Good point on the flatness. Good point on how big the gain is...I was thinking in terms of reducing the interface thermal impedance by maybe up to a factor of 2. This sounds good, thus my post. But if you do the math even a very hot OC CPU with proper thickness on the best thermal paste has temp drop over the paste of only 2 deg max. So twice as good is not much. Its the diminishing returns scenario. Thanks for setting me straight.