Does copper conduct heat better than water?

[Mac42]

I was just noticing something on the front page comparing copper to silver... the only thing is that if copper conducts heat better than water (ie- the water is the heat-transfer bottleneck) then havig a silver waterblock wouldn't do you any good. This might change with the addition of chemicals, or I might be wrong all together, but it just seems silver would be a waste of money.

This also reminds me of a comment I saw on Spode's site saying that eventually we would have to get rid of the metal all together and go to direct water/die cooling. To me it seems that if copper was the better conductor (again assuming I'm right) then it would be able to spread the heat to more water surface than direct water-cooling would.

Anyone have any thoughts on the subject?

[dimmreaper]

If memory serves, copper is a much better conductor than water.

However the advantage to direct die cooling is that you don't have a layer of thermal paste on the die. Water conducts heat better than thermal paste. Well at least thats how I understand it, soemone correct me if I'm wrong . . . .

[dimmreaper]

Originally posted by Would71
Is that your phrase of choice for the week? "Someone correct me if I'm wrong..." :D :D :D It's easier to say that now than to appologize for being wrong later if I am . . . .

[Richard]

It's difficult to say. I've direct die cooled a Pentium III coppermine processor, but haven't attempted it with a Tbird. My results with the coppermine were quite good. (Then again, the heatload that a P!!! gives off is pretty insignificant in comparison to a high end Athlon.)

My overall feeling is that with adequate volume of water and assuming you don't have an air bubble in the "die_chamber" you should be able to cool the die to just about water temperature. Of course, this is already nearly achievable with a good copper waterblock. (Which I shouldn't need to mention is umpteen times safer.)

My final answer is that while direct die cooling may provide marginally better cooling than a copper waterblock; it is not enough to justify the potential risk.

[Mac42]

I bet artic silver is pretty close if not better than water. The problem with direct-die cooling is that there's a very small amount of water transfering heat at any given moment. At least with copper you can spread the heat out to a larger amount of water. I've seen a few attemps at direct cooling and they weren't that great. I really don't know that you could get much better results than a copper block as far as water-cooling is concerned.

Also, if CPU's get much hotter then direct water cooling might be a bad idea. If the die runs enough heat to evaporate the tiniest amount of water on the die then it could create a small bubble and start a chain reaction. It might not fry the chip, but you never know.

[johns linux box]

why not put the none moving parts in transformer oil and then cool the oil
that way the only fans that you would need are to cool the drives and a radiator to cool the oil

[fuzzba11]

I believe the thing with copper is that it will absorb and dissipate the heat more easily. Water can absorb a lot of energy without increasing temperature very much, but copper increases it's temperature more, and depends on high airflow or water flow to dissipate that heat. Water flowing over your core? Hmmm...I think what would work really well would be a small water tank over the die, which flushes itself out for new water every few minutes or when it hits a certain temperature. I say that rather than water constantly flowing since I would imagine the water would have more time to absorb heat from the processor...

[Naeleros]

This also reminds me of a comment I saw on Spode's site saying that eventually we would have to get rid of the metal all together and go to direct water/die cooling. To me it seems that if copper was the better conductor (again assuming I'm right) then it would be able to spread the heat to more water surface than direct water-cooling would.

My opinion is that this will come to pass someday. Someday.. some major company will put together the right combination of waterblock/radiator/pump/etc to make watercooling cost effective and noise effective for consumer PCs. There's a lot to be overcome, though. Its almost hard to imagine where we are at right now. But.. I still think its coming. It might even get funded by a chip manufacturer... depending on their various challenges with making chips able to be air cooled.

I think it will come down to a noise driven solution mostly.

James