Oh well, not willing to go through the trouble to lap my cooling plate and the CPU and then test it without tim.
I don't think it would hurt anything to test the theory out at low loads though, I would still be curious to see how the temps were under those conditions. It would be easy enough to keep track of them during the test to prevent any damage.
From the same article:
"To increase the effective area of heat transfer (A), the voids created by the imperfect surfaces must be filled with suitable highly conductive thermal interface materials. Many different approaches have been adopted by the industry to fill in these voids. Thermal greases, soft metal films, soft metal plating, better machining, and surface finishing techniques are some of the commonly adopted approaches."
The portion "better machining, and surface finishing techniques" seems to me to indicate better lapping methods can be used as well as tim applications. After reading the article though, it seems to me that a lot of pressure would be required to perfectly mate the surfaces. It seems to me the level of machining the surfaces would make a great deal of difference. Personally I would still like to see a test on a CPU cooler specifically. I suppose it may be too much work, and a caliper would likely be needed to tell just how flat the surfaces are. I suppose this is the real reason for tim to fill the gaps, because the amount of work to get a perfectly flat mated surface between two devices would just not be economical.
Maybe when I upgrade my boards and CPU I'll try this idea out before installing the new hardware.Edited by Mergatroid - 3/25/11 at 7:17pm