Advice on installing new cooler.

[Crow47]

I just bought a Zalman 9500 AM2 cpu cooler at Compusa because they're closing (wheee). I have some questions about the thermal paste. The cooler came with a small tube of it, and I'm wondering if I should use all of it. It is a fairly small tube, but the instruction booklet doesn't specify. Also, should I leave the paste in a dot after I put it on the CPU and then put the cooler on, or should I spread out the paste first over the CPU with like a credit card, then put the cooler on?

 

Does anybody have this cooler, or have they had this cooler? Did you like it?

 

Thanks.

[road-runner]

I always just put a dot of paste in the center on a dual core and a line across on quads. http://www.arcticsilver.com/pdf/appinstruc...d_dual_wcap.pdf

[j2t]

Yep, can't really go wrong with the instructions on arctic silvers site.

[werty316]

I just add enough TIM so it covers the entire CPU with thin layer.

[Crow47]

Well when I put this cpu on, I followed the instructions on that pdf. Well my temps are almost twice as hot as they were before (55C load compared to 30C) on the same cooler. Is there a reason you wouldn't want to spread the thermal paste over the whole chip?

[iKillSteal]

Well when I put this cpu on, I followed the instructions on that pdf. Well my temps are almost twice as hot as they were before (55C load compared to 30C) on the same cooler. Is there a reason you wouldn't want to spread the thermal paste over the whole chip?

Apparently spreading the paste catches tiny bubbles of air in it which then decrease the pastes ability to transfer heat to the heatsink. Just putting a big ol' dot in the middle of the IHS(or line for you quad users) supposedly eliminates that problem since the paste just gets smashed flat down with only the very edges, after the heatsink is positioned, being exposed to air instead of being slowly spread and aerated.

 

Somebody needs to do some tests and see if there is any actual difference in TIM performance using either method. I'd do it but I ran out of AS5 with my recent upgrade.

[razor]

Somebody needs to do some tests and see if there is any actual difference in TIM performance using either method. I'd do it but I ran out of AS5 with my recent upgrade.

 

 

I can provide un-verifiable proof that either way works just fine. when I first installed my Zalman 9500 using AS5, I did the "dot in the middle" method. a few weeks ago, I was taking apart and rebuilding my PC for fun, and did the "thin layer across the chip" method. both methods produced idles in the high 20's/low 30's and loads in the low-to-mid 50's.

 

when I get bored and dis/reassemble again, I'll take screenies and post them.