Thermal paste job went way different than expected

Built a new rig for a client last week. i7-14700K with a Noctua NH-D15. Used the pea method like always. Idle temps dropped from 42C to 29C after I reapplied it with the spread method instead. Anyone else see a 10+ degree difference just from changing how you apply?

5 comments

5 Comments

mila_brown1023d ago

Gotta disagree with you there. The application method itself probably didn't drop temps by 10 degrees. More likely the first pea application was just bad - maybe too little paste or an uneven mount. Spreading it just fixed the original mistake. I've done the pea method on dozens of builds and never saw more than 3-4 degrees difference between methods.

the_wren23d ago

Fair enough, but you're making a big assumption that the first application was bad. I've done the pea method on a lot of builds too and I've seen bad spreads just as often as with the spread method. The whole point of the pea method is that the mounting pressure does the work for you, but if your chip is slightly uneven or the cooler base isn't perfect, you get a gap. Spreading guarantees full coverage across the whole heat spreader. That 10 degree drop tells me the first attempt was fine, just not covering a hot spot. The spread method fixed a gap the pea couldn't, plain and simple. Your 3-4 degrees is what you see, but I've seen tests online showing 5-8 degree swings on certain chips with weird die layouts.

kim.zara17d ago

Try a cross pattern instead, covers better on uneven coolers without leaving gaps.

oscarc5317d ago

Kim's cross pattern is actually what I ended up switching to after a similar issue. I had a cooler that was slightly convex in the middle and the pea method left a dry spot no matter how careful I was. Once I started doing a thin cross, my temps evened out and I haven't looked back since.

robertb4717d ago

People get way too wrapped up in the perfect application method. If it works, it works. I've used the pea method for years and never had a problem. The spread method works too, but it's a lot messier and wastes more paste. The real reason for that 10 degree drop is probably the pressure was slightly off or the first mount was crooked. The method itself isn't magic. You fixed a bad mount, you didn't discover a secret paste application technique. It's not that deep.