Current Thermal Compound with respect to Laptops

[DavronTheMighty]

I've done quite a bit of research recently to understand what testing has been done relative to laptop processors and thermal compounds. The majority of thermal compound testing that I've seen has been for desktop/workstation processor between the IHS and the copper/aluminum/nickel heatsink/waterblock. The most clearly applicable testing that I've found was done in 2012 and posted on the Anandtech forums. It involved a delidded 3770k and a lapped H100, which is obviously similar to Intel laptop processors and cooling system. Is anyone aware of another recent study of current thermal compounds specifically for laptops beyond the liquid metal? If there hasn't been, would anyone else place value in such a test?

In my mind the testing would probably include a sandy bridge quad core mobile processor, haswell quad core mobile processor and a skylake quad core mobile processor along with 5 to 10 thermal compounds.

If I were to do it, I wouldn't include the liquid metal compounds since they are a pain to apply and remove and are rather expensive. At the very least I'd expect some of the top performers to be in the list, such as grizzly kryonaut & hydronaut, gelid gc-extreme, IC Diamond 7, Prolimatech PK3, and Arctic MX-4.

Since laptops are thermally challenged with stress testing, I suspect that comparing average throttled processor speed would be the best measure rather than temperature which will be at the throttled point.

Does anyone have any thoughts or suggestions on the setup or reasonableness of the testing or perhaps are already performing such a test?

Thanks

 

[CountMike]

I wouldn't think there would be any appreciable difference from desktop application. Whatever there is loss of heat transfer by smaller contact surface in laptop CPUs, it would be offset by much lower TDP. I use same paste in both + GPUs too.

 

[DavronTheMighty]

The reason that I'm considering doing this is that there has been some past study of bare die silicon vs IHS and that has proven to have at least some effect on choosing the specific compound, refer to the pump-out effect that has been discussed which seems to have a strong correlation to the viscosity of the compound. Some of these effects have been large enough that many people have switched to only using ICD7 or the liquid metal solutions in bare silicon situations. Some of this is attributed to the force, which laptops tend to not have much ability to modulate. I doubt that I could get much long term pump-out type data with testing in any reasonable time period.

 

[CountMike]

Some laptops have thermal pads and most of them can't use TIM paste because of it.

 

[DavronTheMighty]

Yes, it is true that some laptops do have thermal pads. However, from my experience over the past 6 or so years of working on laptops with quad core processors (the ones that I have suggested as potential test rigs that I have available), I've not seen many or actually any of those with quad core processors that do. While they may not have amazing paste, from my experience those tend to have some form of paste. From what I can tell the thermal pads tend to have less conduction, which the quad core processors with 45W or better dissipation can not handle the handicap. I can see that you don't see any value in it, and I guess you represent the entirety of the Tom's Hardware community.