Are you kidding me? Copper's heat conductivity is at least 30% more than aluminium.

But really, if anyone wants to be extreme, use a silver plate. Silver is the best heat and electrical conductor for any single metal at room temperature. At least 10% better than copper.

Well I did say metal; diamond isn't a metal. But yes, diamond is the best solid heat conductor known-a damned expensive one at that. There might be better ones, but only under very specific conditions. In any case, a typical heatpipe actually has better heat conductivity than any pure metal, but the heatpipe orientation and temperature range matters a lot in that aspect.

---

Added on May 29, 2007, 11:11 pmAnd yes, thanks for the table. So to set the figure exactly, copper conducts heat 87% better than aluminium, and silver conducts heat 5% better than copper and 98% better than aluminium

This post has been edited by lohwenli: May 29 2007, 11:11 PM

Not really. There is no such thing as heat attractivity. And the specific heat capacity has little effect on losing heat, only effects the time it takes for thermal conditions to equalise (aka hysteresis).

The proper answer is-density. Aluminium is much, much lighter than copper for the same volume, and you can easily put a hell lot of thin fins and increase the surface area without a significant increase in weight. And with that much surface area, the lower heat conductivity of aluminium can be minimised. Using a similar design in copper will significantly increase the weight of the heatsink, those who own all copper heatsinks should know this quite well. Dropping on your foot one of those all-copper heatsinks which are huge enough to perform equally to their aluminium fin counterparts will really hurt, and I'm not talking about the damage to the heatsink

Heatpipes on the other hand, are made of copper because it is absolutely critical for the heat be conducted to the fins, or else why even put fins at all? The non-heatpiped heatsinks suffer this problem-heat is not well distributed to most of the fins, and even on the same fin its hard to conduct heat from one end to another due to the low thickness of the fins.

All copper will win even for the same airflow. But it will be only by a slight margin, which will not be worthwhile considering the weight. You can see this clearly in XP90 vs XP-90c. The winner isn't terribly clear, because the high conductivity of the heatpipes already distibutes heat very well to the fins.

I'm not too sure about this. I know heatsink manufacturers have a tendency to use a design that takes up less space for copper heatsinks due to weight, which usually results in fins close together (which gives damn good cooling with high cfm, but hopelessly bad cooling at low cfm). Has anyone tried out the XP-90 and XP-90c with a variety of different fans?

The flatness probably does affect the heatsink's effectiveness. But contrary to most people's perception, thermal paste actually has very bad thermal conductivity compared to metal-so filling the gaps with thermal paste would probably have not much effect.

Filling it with solder however, might be a good idea as it would allow better heat transfer from the IHS to the heatpipes (more area). BUT, there's a huge risk...
http://forum.lowyat.net/index.php?showtopic=270141&hl=

Off topic,

I think sticking a small heatsink on the flat base of larget heatsinks would have minimal effect. It would be hard to get good thermal contact unless you lapped both surfaces to match, and the bigger heatsink would already be doing a good job of transferring heat away, a conventional heatsink would be insignificant by comparision. Even the best non-heatpiped heatsinks struggle to deal with 80W and above, noisy, high power fans would be required.

This post has been edited by lohwenli: Jun 1 2007, 07:45 PM

Corrected the link..sorry, had too many tabs open, forgot what I last copied.

Yeah, that would be the case, alternating cool and hot strips of area on the IHS. But anyway heat production on the processor die is by no means uniform to start with, and the processor die is pretty small even when compared to the IHS. So I guess there would be no real effect of uneven cooling.

Trust me on this-when soldering something that can rapidly dissipate heat, you need a TON of heat even to keep the solder workable. Regardless of whether you use a stove or a very high power soldering iron, you're setting yourself up for trouble.