Testing MethodsThe stock cooler above will be the only heatsink used in this round-up. However there are problems with Intel's stock heatsink foremost being its push-pin mounting system. Made of plastic they have finite lifespan and repeated installations can eventually cause damage. Intel wanted to make things easy for PC-Hobbyists and push-pins were the solution. Given the number of times I would be removing the cooler to re-install thermal pastes, I decided upon a modification.
The mounting system pictured is similar to what I use for certain water-blocks which have problematic mounting hardware themselves. The testing process involved repeated installations to find the best "mount." For each paste I chose the best of three installs, which became a time consuming process. Cleaning the heatsink and IHS between installations used almost all my Arctic Silver
ArcticClean.
The only installation which could not be repeated was the Intel's factory installed thermal material. Comparing the effects (if any) of different application methods such as spreading or the
dollop method then doubled the overall number of installations. The dollop method involves placing a drop in the center or thin line of paste across the IHS and then simply mounting the heatsink ensuring the pressure is evenly distributed. The theory behind the dollop application is that "compressing" rather then spreading the paste forces air out and thermal material into the gaps along both surfaces.
As seen above Intel uses a combination installing their factory paste. We can see that the paste below is already spread and when the heatsink is installed TIM will be compressed filling the gaps and ideally displacing the air from those surface gaps. Examining differences between different application methods involved taking photos of both the heatsink base and IHS before and after.
Intel factory installed paste distributed evenly under the mounting pressure. A certain benefit of Socket 775 is the clamping mechanism which holds down the CPU. This prevents the CPU from potential damage as in Ziff Sockets when the CPU is pulled out attached to the heatsink. Even with the Ziff locked I've had this occur so many times I began unlocking the small lever, at least as far as I could. Note the large ridge of paste at the right of the photo.
Ultimately temps will provide the best proof for the specific application method ->
Have you tested the difference between the line and the blob method?