I just installed a new i5 Sandy Bridge on an Asus P67 Sabertooth board. When I first turned it on, I received a Bios error stating there was a CPU overheating issue. I was able to navigate to the Bios settings, and the temp said 100ºC. That's obviously wrong because the computer was only on for less than a minute.

So, I ignored it and tried installing windows, etc.

I got nervous though as I read more with how bad or nonexistent thermal paste and burn out a CPU. I didn't apply any thermal paste because the heatsink came with stuff already on it. But I got nervous it wasn't properly applied, so I removed the heatsink. I saw that it was applied for the most part, and saw no issue, so immediately put the heatsink back on the CPU.

Now -- my question:

Do I need to scape off the stock thermal paste and apply my own since I removed the heatsink for a few seconds, or will I be fine?

Some data points:

  • using RealTemp, my CPU was clocking around 33ºC consistently.
  • the stock thermal paste spread over about 2/3rds of the CPU.
  • Unless things have changed, stock heat sinks came with thermal pads, which are a lot more forgiving than thermal compound – Journeyman Geek Dec 22 '11 at 6:34

Unless the paste looked like a dog's breakfast after separating them, there will probably be no need to clean it off and reapply. Of course, if it looks too clean then it may be that the protective plastic was never removed from the layer of paste in the first place...


It does not sound like a thermal paste problem... If you have removed the heat-sink multiple times it is advised you clean and re-apply thermal paste especially if you did not twist the heat-sink as you pulled it off discharging any vacuum that may have formed, because what will happen is the thermal paste will form bubbles in that already thin layer and then will act negatively as to the job it was there to complete in the first place...

Now for what I think your real problem is, your heat-sink should have some sort of 2, 3, or 4 pin connection to plug into your motherboard and it should be plugged in to the CPU_FAN slot or your computer will fail to realize that you have a heat-sink on your CPU and think there is a fire in your case often showing temps between 80C and 100C. So make sure it is plugged into the correct spot and if you don't have a plugin then you will need to disable it in your BIOS.


It really depends on how much I like the computer. If it's an older one that I don't care too much for, I would take Ignacio's suggestion. However, if it is my main workstation then I would reapply the thermal paste. Laptops, I always replace the thermal paste. Clean off the old ones with some firm cotton swabs and Denatured Alcohol.


[...] the temp said 100ºC. That's obviously wrong because the computer was only on for less than a minute.

I wouldn't be too quick to jump to conclusions. According to Arctic Silver's thermal paste application instructions:

A modern high-performance CPU can be permanently damaged in less than 10 seconds without proper cooling.

I doubt that's just scare-mongering; the heat that is generated in a CPU these days is enormous, that's why heatsinks are so huge these days and why modern cases have multiple fans.

Thermal paste isn't that expensive and you usually get it in quantities that will allow several applications (a very thin layer will do), but it needs to be there between your CPU and heatsink/fan, so I wouldn't mess around and hope the board's temperature sensors are wrong, just apply it correctly and you should be fine. I've found using an old credit card or similar works well to firstly remove left overs of old paste (if any) and secondly to spread the new paste evenly and to make sure it gets into all those microscopic little irregularities.

Another tip, but of course that may not be everyone's thing: if your machine tends to overheat after running Windows for a while, try installing Linux, it's much gentler on CPU demand in my experience. I've saved a few machines this way that I would otherwise have had to dismantle or throw away.

  • modern systems have thermal cutoffs. If a system overheats chances are it'll merely shut down. – Journeyman Geek Dec 22 '11 at 6:38
  • @JourneymanGeek: yeah, normally. My point was that this is not the sort of thing you want to use "chances are" phrases with ;-) – Amos M. Carpenter Dec 22 '11 at 11:18
  • and by modern, i mean in the last ten years. And if it dosen't just shut down, you get wierdness, but i rather doubt any modern processor after the infamous burning AMD would have lasting issues from a short period of overheating – Journeyman Geek Dec 22 '11 at 14:39
  • a) Are you seriously recommending, then, @JourneymanGeek, that people be incautious with this sort of thing and not take high temperature readings seriously...? Even if your CPU is not immediately fried, you'll reduce its life each time it overheats. b) Spontaneous iPhone Combustion and recalls of several types of Dell machines due to overheating would disagree with your "any modern processor" statement. c) Why are we even arguing about this...? If you have an alternate answer, feel free to post it. I'll continue to recommend caution. – Amos M. Carpenter Dec 23 '11 at 6:26
  • all of which are battery related. Overheating would have other symptoms - slowdowns, stuttering and shutdowns.It would shorten life, but i don't see any other symptoms noted. – Journeyman Geek Dec 23 '11 at 14:42

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Not the answer you're looking for? Browse other questions tagged or ask your own question.