Just an idea on better cooling

[jahjahbinks]

As far as I know, all the current cooling scheme (air, water, phase) involves putting the cooling solution on one side of the motherboard. I am thinking that by putting smaller heatpipes on the underside of the motherboard directly beneath the CPU area and have a heatsink or waterblock to cool it will help dissipate heat faster. Of course a larger case would be needed. What do you guys think?

[eddieate]

well you can't really get a good contact on the other side of the mobo without risking shorting something out.
its a good idea maybe just get some airflow there? like i got on my gfx card i noticed the top above the core was getting warm so its got a 120mm fan on it now seemd to sort it out all right :)
experiment around maybe you could somehow mount a heatsink there with artic ceramique thats meant to have 0 conductance & capacitance you might end up just trapping the heat there though...
/Ed

[jahjahbinks]

Well, not with existing motherboard, it's very likely to short things out. Maybe with a newly design motherboard with more layers so that there are less interconnects on the back side.

[Serra]

I think I once saw some tests run on that idea, where the testers took a board and put mounted it in the middle of a large plexi container by its edges with a fan blowing on the back. As I recall it didn't do a lot for temps, something in the 1-2 degree area. The synopsis though (iirc) was that that temperature drop was more likely to have been caused by the increased air circulation in general rather than the actual cooling. I wouldn't be surprised if more active solutions, like subzero cooling or phase change for example, did help... but I'm pretty sure that the difference for air was effectively 0.

Prolly worth nothing that back when those tests were run, it would have been in the day of the 86w Pentiums or earlier... so it could make more difference now with Intels rating at ~135w(+/- depending on exact model).

~ Serra

[Dragoon42]

I wouldn't think that it would generate a lot of heat back there, but I could be wrong. Could test it by having a huge fan blowing in the back of it?