Stock, it would fluctuate just over 51Âº C with the Noctua NH-C12P. This is warmer than my SI-128, so I figure something must be wrong since it is supposedly better. I pulled it apart, and noticed there is a little more TIM than should usually be there. It looked evenly spread. I thoroughly clean the IHS and the NH-C12P with IPA. I checked that the NH-C12P is in fact flat with a straight edge. I reapply TIM and place the cooler on top of the IHS, and then I thought of something. When I bolted it down the first time, it did not really seem tight as compared to my SI-128. Even before I did the 20Â¢
, the SI-128 was tighter. (Lovingly ripped off from txtmstrjoe's 2Â¢ mod
So I got to thinking: what could I do to make it tighter? The change is not an option given the difference in the mounting bracket. Then I thought, what if the bracket were thicker? So I measure the screws (about 5 mm) and find out that they would fit well inside of an SAE #10 washer (5.5 mm or so). So I run to Home Depot and buy a bag (for $1.07) and put a washer between each spring screw and the bracket, thereby increasing spring pressure. I give it a wiggle; hm... still loose. So I put another one in on either side, and notice it is much harder to bolt down. Good; that is what I want.
So I boot into Windows and fire up S&M FPU. With Arctic Silver 5 that is literally a half-hour out, load temps are just about 49Âº C. Not bad. I am hoping that as the AS5 cures, it will drop a bit.
I realize now this same technique could have been used on my SI-128 as well. It also would have been much easier to implement than the 20Â¢ mod given the physics of asking the bracket to bend more than it should.
(Crossposted with my build thread.)