oh crap!

SockToy · Post by **SockToy** » Sat Mar 10, 2007 10:38 am

My Ultra 120 somehow managed to dislodge itself (I guess... it seemed pretty secure to me), and I didnt notice till I glanced at MBM after hours of Supreme Commander and realised my CPU was hitting 80 (!) degrees ... and I guess it had probably been pretty sustained around that level, maybe even higher given I saw it when I flipped out of SC.

I've cleaned off the thermalright goop with ArtiClean 1+2 and applied some Arctic silver 5 instead, and made sure the heatsink is well secured (again). Temps are back to 33 degrees idle and 54 max.

Am I likely to have kefrazzled the CPU?

Post by **NeilBlanchard** » Sat Mar 10, 2007 10:52 am

Hello,

In theory, it would have throttled itself to keep from overheating (too much), so it may be just fine. If it is up an running again, then that is the best indicator you have, I think.

bonestonne · Post by **bonestonne** » Sat Mar 10, 2007 11:58 am

heres a test that i created when a similar event happened to my P4 tower...turn it off, remove the heatsink, pop out the processor, and check the edges of the silicon. if any edges look like they're starting to seperate, you've got yourself a problem.

if all the edges look fine and crisp [you just need a visual] then theres nothing to worry about, however, if a processor reaches that temp, no matter what goes on, you're gonna want to check it, 80 is blazing hot, my old PII Xeons ran that hot once, ever since i chill them as much as i can.

Arvo · Post by **Arvo** » Sat Mar 10, 2007 12:08 pm

Don't worry.

About year ago I helped my friend - his PC started suddenly reboot. Looked at CPU utility - CPU reached 94C(!). Looked into PC case - even fan holder melted onto the heatsink

After replacing CPU cooler all went normal and system is running well.

SockToy · Post by **SockToy** » Sat Mar 10, 2007 12:23 pm

I cant really look at the silicon can I? Its under a heat spreader on the core2duos...

disphenoidal · Post by **disphenoidal** » Sat Mar 10, 2007 12:45 pm

bonestonne wrote:heres a test that i created when a similar event happened to my P4 tower...turn it off, remove the heatsink, pop out the processor, and check the edges of the silicon. if any edges look like they're starting to seperate, you've got yourself a problem.

How did you look at the die on a P4? I thought all of them had a metal heatspreader can on top of the die.

bonestonne · Post by **bonestonne** » Sat Mar 10, 2007 2:54 pm

not the metal heatspreader, i looked at the actual silicon. you have the metal on top, but you have the silicon where the pins are. the heatspreader can be banged up to a point without causing a problem, but if the silicon is damaged, it affects the processor the most.

Saribro · Post by **Saribro** » Sat Mar 10, 2007 4:06 pm

The green stuff where the pins are is not the silicon. The silicon is invisible and can only be inspected after prying of the metal heatspreader. Removing the heatspreader is risky, and more or less makes your CPU incompatible with your heatsink.

Chris Chan · Post by **Chris Chan** » Sun Mar 11, 2007 10:14 am

Yeah, the green stuff is plastic, and it's called the substrate. You're right though except in terminology, bonestonne, because if the substrate starts delaminating, that is a sure indicator of serious overheating problem.

disphenoidal · Post by **disphenoidal** » Sun Mar 11, 2007 10:21 am

http://balusc.xs4all.nl/srv/har-cpu-int-pd.php

Here are some pictures of CPUs with the heatspreaders removed, so you can see the actual silicon dice underneath.

bonestonne · Post by **bonestonne** » Sun Mar 11, 2007 11:21 am

ah..ok, i stand corrected in that part.

SockToy · Post by **SockToy** » Sun Mar 11, 2007 12:38 pm

Well... It seems happy with a reapplication of AS5. Bit worried it might all fall off again though ;/ Ah well.