thermal compound gets "squished" out

I have a Thermaltake Silent Boost (a1889) and I'm using arctic silver 3 compound. It works great for about a week or 2 and then the processor starts running about 10-15 degrees higher even at idle. For some reason the arctic silver is getting "squished" out to the sides of the cpu die and the cooler is almost sitting directly on the cpu. This cooler does seem to latch on quite a bit tighter than my previous (a volcano 9)... is the extra pressure squeezeing the compound out? Should I try some sort of thicker compound then? I'm not really sure what to do and I figured someone here would be able to help out. Thanks in advance.

The more it squeezes out the better. The less you have there the more the core will touch directly to the sink. Thats not whats causing your heat problems.

Maybe too much compound? How much is actually squeezing out?

You should only be using a paper thin layer...so hardly any should "squish out". Thermal grease is used to fill microscopic pores to increase heat conductivity between the CPU die and heatsink surface.

I put the absolute thinnest, most even layer possible. And Like I said it's fine for a week or maby 2 and then it's starts heating up. If I reapply then it goes back to normal.

There must be another problem. As the other folks have already stated, the thinner the AS the better. Look for other problems like the cooler slipping off center, or loosening of the mounting clamps/screws. In addition, visually check your cooler to make sure that it is correctly contacting the die before remounting it.

And not to cause undue concern (although a _little_ paranoia is always a good thing) make sure that you're not a victim of counterfeit AS. The AS website shows how to identify the fakes.

Huckleberry said:

make sure that you're not a victim of counterfeit AS. The AS website shows how to identify the fakes.

Click to expand...

Where is this at? I'd like to check it out.

DoubleJ8 said:

Where is this at? I'd like to check it out.

Click to expand...

Here.

There was some suggestion in the past that AS3 could pump out under heavy vibration. Dunno whether that was true or false though.

actually artic silver III comming off the sides shorted out my cpu and made me get 50 C idle.

*EDIT* that was the first time i ever used it and the first time i had used compound ever ive learned since then.

Then you used way too much. You should only have enough so that it *starts* to come off the core, but never touches the wafer.

Nat what CPU are you using?

Nat what temps are you experiencing?

Nat how much did you applied?

My guess is this is normal if the temps are not high. My AS5 was quite low at beginning of burn-in, then it raised drastically and then level-off, however, it did not take 1-2 weeks to do this. It took less than a hour or 2. See my sig below for temps.

I have a 2100+ and I did use just enough to cover the die. I did wonder about vibration maybe being a culprit, because as far as I can tell nothing else is wrong with it. The hottest I've seen it so far running at idle is 115 fahrenheit. That's using soyo's h/w monitor so that would be the external temp.

if you going off a hardware monitor remember those aren't 100% correct.

A 46C or 115F temperature is not too bad at all, depending on the settings of your processor. That doesn't sound like an unreasonable result with that heatsink to me.

But if you still think there is a problem...

Remount your heatsink, clean the base and CPU, reapply thermal paste as per the directions on www.arcticsilver.com.

First thing you do, use prime95 and take load temperature readings. After allowing ~200 hours of settling time for the paste, do this again taking load readings with prime95.

Continue to take load readings every couple days, or weekly (whatever you feel like) and report if you still observe temperature degradation.

If you have already done something similar to this, then can you explain the method you used to observe the temperature difference?

RoadWarrior said:

There was some suggestion in the past that AS3 could pump out under heavy vibration. Dunno whether that was true or false though.

Click to expand...

The suggestion mainly came from RonnieG who basically portrayed some kind of inexplicable vendetta against AS, or atleast thats how it appeared to me, and that seems to be how AS interpreted his statements.

Supposedly he said Shinetsu paste is the only one that won't pump out under thermo-cycling because it's thicker... But with the number of people using the paste around here, if pump out were a problem then it would be complained about a lot more often.

Also, we have people using virtually every combination of heatsink and paste around here and if people were experienceing problems 2 weeks into using their paste because of pump out, we would hear more about it.