Thermaltake Intros TG-30 and TG-50 Thermal Compounds with Honeycomb Pattern Stencil Applicators

[Tomorrow]

Pea drop in the center....

'nuff said

If you can't accomplish this, then you have absolutely NO business building a computer in the 1st place, IMHO

Depends on the CPU. This logic might have been good years ago but today in addition to AM4 where chiplet might be in the corner instead of middle and LGA1151 (soon LGA1200) we also have Threadripper etc where using this method would yield bad results. Gamers Nexus tested this years ago. Even their modmat has patterns for best spread printed on it.

 

[kiriakost]

actually the honeycomb applicator will be good for beginners to avoid putting too much tim; if this tim has silver particles even is not conductive may be capacitive and can cause damage if close electric paths

And also for CPU cooler reviewers, about more accurate comparisons.
The Thermal grease does not seem high performance, this could be bad idea for air-cooling.

 

[Chrispy_]

Depends on the CPU. This logic might have been good years ago but today in addition to AM4 where chiplet might be in the corner instead of middle and LGA1151 (soon LGA1200) we also have Threadripper etc where using this method would yield bad results. Gamers Nexus tested this years ago. Even their modmat has patterns for best spread printed on it.

Whilst that's true if you do A-B testing of various spread patterns the same day you apply the paste, I can guarantee you that as long as you put enough paste in there and the clamping pressure is sufficient, all methods result in a thin, even spread after a couple of weeks.

The only thing you have to do is avoid making an air-trap, like drawing a circle of paste, for example. This is why pea, rice grain, single-blob - call it whatever you want - is best. It will guarantee no air bubbles and it'll get squeezed into a very thin, even layer with the excess being pushed out the sides. As I mentioned a few posts back, I also put peas in the location of each individual dies in a multi-chip package, but I have a colleague that built some of our threadripper farm with the "big blob in the middle" method and having to RMA it for a failed board, the single dollop had made full-IHS coverage even without the one-blob per die method.

If that doesn't happen then you probably have poor clamping force which is a much more serious problem that worrying about the thickness of your thermal paste layer.

But either way, the most important thing is that it simply doesn't matter. People love to argue this until they're blue in the face but if there's enough thermal paste then pretty much any shape or pattern that doesn't trap air is going to be fine and they'll all perform identically after two weeks, if not sooner.

 

[kiriakost]

Whilst that's true if you do A-B testing of various spread patterns the same day you apply the paste, I can guarantee you that as long as you put enough paste in there and the clamping pressure is sufficient, all methods result in a thin, even spread after a couple of weeks.

I think that we have graduate from the same CPU-Cooling university.
I am following the INTEL train for many years and TPU record that too https://www.techpowerup.com/forums/threads/p4c800-e-and-xp120-thermalright.84829/ , from metal clips to plastic hook clips, from plastic hook clips to plastic anchor pins, and only after market coolers are using the force of screws.
A couple of weeks of patience before making a worth noted comparison this is a golden rule.
But no one he should expect anything better than 1.5 Degree as difference.

 

[Chrispy_]

I've always enjoyed this article

Thermal transfer compound comparison