Thermal Testing

[Evilsizer]

Looking at some of the other benchmarks of thermal paste, got me thinking. i see thermal pads that have as good or better W/mk ratings. yet actual temps say when used on GPU memory do not see to equal that of say Paste+copper shim+paste mounting. to me the paste+shim+paste again shouldn't give lower temps. my thinking is that one review i saw past was about 8 W/mk then copper is how high? sandwich that togther seems with the differences it should not be as low as a thermal pad rated at 12W/mk. which led me to think, is there even a standard to rate paste/thermal pads. looking for new pads for my 2060 super led me down looking at thermal pad testings. one video threw me for a loop when he did the paste+shim+paste and it was lower temps then a Gleid 15W/mk thermal pad.

 

[Zerileous]

I'm not to clear on your specific application, but copper's conductivity is like 400 w/mk. Generally thermal paste and thermal pads are used differently, thermal paste being only to displace air in an otherwise metal on die contact (or metal on IHS, metal on package, etc). Thermal pads on the other hand are compressible and occupy space. You wouldn't normally use a thermal pad in place of paste or vice versa. I'm guessing the shim takes up the space and then it's a question of whether the extra interface is better or worse than a pad. Sounds like it's better in this case. My concern would be that a copper shim would either not take up all the space, or not be compressible, leading to a less effective mount on the GPU die itself (in the case of a GPU cooler). Of course there are many different GPU cooler designs and some of these have discrete components for cooling the memory and power stages and the GPU die so in those cases that wouldn't be an issue.

 

[Evilsizer]

i may not be describing correctly with paste+copper shim+paste.