Just my .02$ - I was never a fan of spreading. Too easy to get small air bubbles captured. I usually just put one dab about the size of a grain of rice and clamp the heatsink. On most boards you can see that it flowed to the edge of the cpu die after the heatsink was clamped. I do remember using the "X" method on larger dies (AMD?). Still, no more than a couple grains of rice. Thermal compound flows pretty well especially after getting hot. The idea is less, not more (which is why I never liked the thermal pads). ...and in earlier days not all compounds were non-conductive like they are today so you had to be very vigilant about overflows....
[SOLVED] How's my CPU thermal paste application?
- Thread starter Appletax
- Start date
Sky-Knight
Well-Known Member
- Reaction score
- 5,418
- Location
- Arizona
And you didn't scroll down. Note, I did watch the video. I just disagree with magic processes that cannot be observed. I don't have transparent tools to play with. And I've also never had a mess on a CPU after mounting.Lol... pictures 1 & 2 show it's already overflowing...
Spreading the thermal paste as thin as you can over all the cpu * will * result in overflowing once the pressure is applied.
You didn't even look at the video you're referring to (hxxps://support.arctic.de/en/mx-4)
Note, where I live it's going to be 112F outside today, that's 44.44C. Indoors it's uncommon to see air conditioned temperatures below 76F, 24.44C, and the office I'm typing this in is currently 82F, or 27.77C.
If I get my cooling wrong, crap DIES. I used to test network hardware appliance cooling systems in an oven, outside, during the Summer! My methods may be a bit messy, but they are well proven.
Last edited:
Markverhyden
Well-Known Member
- Reaction score
- 10,850
- Location
- Raleigh, NC
The OP pictures clearly show the paste layer is to thick. Personally I'll use the X method but only .5" long for each leg if I can extrude a small diameter amount. If not I'll do a small dollop in the middle. Then seat the sink, but not latch, and then take is off to make sure the coverage looks good. After that clamp it down.
I'm pretty sure you will create air bubbles doing so...
Sky-Knight
Well-Known Member
- Reaction score
- 5,418
- Location
- Arizona
Yep... once the compound is spread when you separate the parts you have to reapply.
Appletax
Well-Known Member
- Reaction score
- 388
- Location
- Northern Michigan
Maybe it would be cool if there were a way for us to apply thermal grease as a bunch of little dots like on this AMD Wraith Prism cooler that I recently installed.
britechguy
Well-Known Member
- Reaction score
- 4,460
- Location
- Staunton, VA
I missed this when you initially posted it, and it pretty much confirms my preconceived notions.
Other than not using so much as to make things messy (which far too many do), once you have "enough" or "very nearly enough" the functional results are virtually identical.
Over the decades the conniptions I've seen thrown about "running hot" have been many, and virtually all stupid beyond belief. There are people who scream about temperatures 10 degrees or more beneath T[max]. Running at or near T[max] in certain conditions is perfectly normal if the workload is forcing "the hamster to run as fast as it can on the wheel" and can also be influenced by ambient temperature.
If you're consistently running between T[min] and T[max] sans throttling or any other misbehavior, then everything is fine. Anything between those two temperatures is fully within normal limits (WNL).
