Lapping or TIM dropped temps 18 degrees

[ochadd]

Bought a used Q9550 that the previous owner had lapped. Installed it last night and core one was pushing 82c on my tuniq tower in OCCT @ 3.85 but was otherwise stable. Other cores maxing out around 72c. My previous e7400 never had problems and went to 4 ghz with the same cooler and only maxed out at 55c.

I wanted to make sure it wasn't an air bubble in the arctic silver or something so I figured I might as well try lapping my cooler with some 400 and 600 grit wet sandpaper while it was apart. Results are all cores maxing out at 64c which is an 18 degree drop on core one and eight degrees on the rest. Idle temps are down about five degrees across the board.

Wish I would have tried it years ago as it only took about 15 minutes and cost under $3. It's not a mirror finish as 600 grit was as fine of paperfoudn locally but it seems to have worked wonders. Not sure I'd have the balls to try it on a CPU quite yet but I'm impressed.

 

[DrMrLordX]

Lapping the CPU + cooler is a proven way to improve temps . . . and you may find that 600-800 grit paper will give you a better surface for thermal transfer than a "mirror finish" anyway.

Some coolers ship with a nice, smooth surface already but many of them could use a good lapping.

 

[frostedflakes]

You're sure you didn't have the cooler installed incorrectly before? Maybe it wasn't making good contact or something, because I seriously doubt a 15 minute lapping job would make that big of a difference.

 

[novasatori]

CPU heat spreader and heat sinks can be convex or concave by the way they were formed, and if you only lap one it can exaggerate the problems.

Like if your CPU was convex and the HS&F was concave, it would have lower temps than if you lapped the CPU and kept the HS&F untouched.

 

[lenjack]

Could not make that much of a difference. Had to be mounted wrong.

 

[SanDiegoPC]

Could not make that much of a difference. Had to be mounted wrong.

I am inclined to agree with this post, but on the other hand this is one field where I am always seeing new things and learning.

 

[Zap]

CPU heat spreader and heat sinks can be convex or concave by the way they were formed, and if you only lap one it can exaggerate the problems.

Like if your CPU was convex and the HS&F was concave, it would have lower temps than if you lapped the CPU and kept the HS&F untouched.

This is true. Wasn't there a big bruhaha about TRUEs because their base wasn't flat and their explanation was that it matched the curvature of CPUs, except people buying high end coolers like that seemed more likely to lap their CPUs causing problems with contact?

 

[ochadd]

You're sure you didn't have the cooler installed incorrectly before? Maybe it wasn't making good contact or something, because I seriously doubt a 15 minute lapping job would make that big of a difference.

It was installed in the correct direction and tightened down just as tight. I've installed and removed this heatsink probably half a dozen times or maybe more on different CPUs without issue. First one I had a problem with but it's also the first quad core it's been used on.

The only thing that changed was reapplying the TIM and a lapped cooler base.

 

[novasatori]

This is true. Wasn't there a big bruhaha about TRUEs because their base wasn't flat and their explanation was that it matched the curvature of CPUs, except people buying high end coolers like that seemed more likely to lap their CPUs causing problems with contact?

Yes I think there were a couple heatsinks that were spec'd and made with a concave spreader, I don't have time to look for the posts though.

Its funny people dismiss it saying it couldn't impact it, but there's been numerous posts that I've seen over the years of people only lapping one and having adverse temps.