Hot cpu after upgrade?

[imported_GeekWannabe]

Hey folks,
I replaced the stock HSF on my E6400 today after OC'ing to ~2.6ghz had pushed temps to 61c last weekend(prime95 for 2 hours+).

Removed the mobo, lapped the cpu and new tuniq 120 to 2000 grit, and carefully reinstalled everything to discover that temps now during a stress test are......<drumroll please>........60c?!?

I realize there's a likelyhood that it will drop as much as 5 degrees more after a few dozen thermal cycles, but I honestly expected much better from this highly-rated (and massive) unit. As a test I cranked the fan pot on the Tuniq up for about 15 minutes (1950rpm) and saw no noticable drop in temps, same with it dialed down fully (1050rpm). This seems to indicate poor heat transfer at the interface, correct?

The only possible shortfall I can think of was the use of the supplied TIM, but surely it's not THAT bad......
In any case, do these temps warrant an inpection of the TIM before I dial up the OC another notch or am I just going to have to live with temps in the 60's and the shortened lifespan of the proc to achieve 2.8Ghz+ speeds?
Footnote: I used Artic Silvers' primer/cleaner on both surfaces prior to assembly and applied the TIM according to AS's directions for a C2D proc.
thanks for your input,
Hugh

 

[Zap]

You didn't test it out before lapping? Three possibilities...

1) Bad lap job
2) Faulty heat pipe
3) Not mounted properly

 

[imported_GeekWannabe]

I saw no reason to assemble it "as-is" since an inspection of the CPU showed a large depression in the center (I was able to slide 3 thicknesses of paper under a new razor blade placed across the heatspreader), and the HSF had visible tooling marks on it.

I can posts pics of the lap job if you think it may be at fault, but as someone with above-average mechanical ability who took 4 hours to do the job, checking for flatness with the aforementioned razor blade, it would surprise me if that was the case........but I'm not ruling anything out.

I have no idea how to test for a faulty heat pipe...(run prime95 and feel the tops and/or bottoms of the individual pipes? or I also have a quality thermometer with temperature probe that may suffice?)

The instructions that came with the Tuniq unit can only be described as useless, however after laying all the parts out it's obvious how its assembled and I didn't have to modify or force anything to get it on.

I suppose I'll have to remove it just to eliminate the possibility of poor lapping/insufficient thermal paste/improper mounting..........it's not that hard but the upper left-hand thumbscrew is a bit of a chore to get to, so I'll probably machine a screwdriver slot in the top of it while I have it off.

Thanks for the list of possibilities Zap, I'll let you know how it looks when I pop it off, I think that's going to tell me alot about how to proceed, especially since I'm seeing differences of 2-3c between the cores at idle.

 

[yh125d]

With such a heavy tower like the Tuniq, sometimes its weight can bow it out away from the cpu. Is one core cooler than the other? Try stress testing with your case lying flat on the desk. If doing this lowers your temps noticably you may want to use zup ties to help support the tower so it doesn't get bowed out so bad.

As for seeing if a heatpipe is faulty, I don't know how one would go by doing that. when I'm stress testing my CPU i can touch my heatpipes and they're cool to the touch

 

[imported_GeekWannabe]

I removed the tuniq unit and it's obvious from the TIM pattern that the HSF was "drooping" and causing contact at the upper portion to be lost. Some of this is almost certainly caused by the orientation I chose which makes the mounting H-bar that crosses the HSF run horizontally, not conducive to good support when the case is upright imho. This orientation also results in the heatpipes running perpendicular to the direction that the 2 cores are placed on the die (I honestly don't know if this would have any effect since the heatpipes run through a solid copper baseplate....?)


The reason I did this is to allow the HSF to blow up and directly out of the huge top fan on my antec 900, but after looking closely at results I think I'll turn it 90 degrees and let it blow out through the rear fan.

My case temps are always very close to ambient and gpu cooling hasnt been an issue so I don't think it will cause any problems. I'll also support the upper end of the HSF unit after it's mounted and before I stand the case back up, that in itself may be the key to success!

We'll know something in about an hour, keep your fingers crossed for me guys, and thanks again.8D

 

[imported_GeekWannabe]

After further inspection, I couldn't rotate the HSF 90 degrees or the fan itself would be in contact with the NB heatsink on the d975xbx2......not a good thing.

Instead, I re-cleaned both surfaces and carefully reinstalled it the same way I initially had it, this time tightening the thumbscrews incrementally in an "x" pattern, and made sure that all 4 were FULLY tightened up to the shoulder on the screws.....something I did not do the first time around for fear I'd strip the threads and be up a creek.

It's currently running prime95 and seems to have leveled off at 58c, with both cores mirroring temps very closely through the loading/unloading cycles, so at least some of the problem was caused by less-than-perfect installation, and supporting the HSF with a zip-tie appears to have evened out the core temp differences for now.

I'm not thrilled with the results, but I'll reserve judgement until the break-in is complete and decide then if I want to try again with some super-premium paste.

Until then feel free to add comments/suggestions as you see fit, and thanks for the input from everyone!
peace,Hugh