Thermal Paste on MB + CPU Overheating until shooting down

[Isaac Queiroz]

Hi! First, thanks for looking into my problem. (specs in the end of the topic)

Short History: My CPU stuck to my heatsink, I managed to take it out after trying a lot and I know this may have damaged CPU. I've got some new thermal paste on the processor but the temperature on idle goes from 30º C to 85º C+ in a few minutes. Underclocking doesn't work and the heat is spreading to the heatsink (it's an Akasa Nero 2) as it is close to 80º C when the computer shuts down. To make things worse, I've found some paste inside the socket (covering an entire pin). I've cleaned both the MB and the CPU with Isopropyl but the temperature problem is still there. The question is: The CPU is busted therefore the temperature problem or my MB may be dirty and therefore the overheating? I'm afraid of buying a new CPU and burning it as well.

Long story:

Ok. So... I've been "building" computers my whole life, for me and for others, always for free (a lot of karma points earned but the worse problems always affect my hardware.

I don't recall the last time I changed my thermal paste and I may have used some epoxy glue by mistake because it was nearly impossible to separate the CPU from the heatsink. I tried a little heat (using the hairdryer), not enough to burn anything and a lot of isopropyl, but it was impossible. Eventually, I put it inside a bowl full of isopropyl and left for a few hours and nothing. After giving up and trying to rescue the heatsink, i pried the CPU from it using several tools and on the first try the thermal paste broke, freeing my components. There is no visible damage to the CPU but I know better, it probably broke something.

After all this drama, I bought a better Thermal Paste (from Zalman, nothing fancy). I was tired, used the "rice grain" technique and the heating problem was there: On idle, the CPU would reach 85º C just after Windows starting up. I tried disassembling, applying paste again and nothing.

Eventually I found out there was a small grain of the old thermal paste inside one of the pins inside the socket. It probably melted down with the heat and it was inside the socket all the way. I've done my cleaning, it's not visible at all but the temperature problem is still there.

The CPU is damaged, right? If so, is there a way to be safe and know if the motherboard is ok too? Is it safe (or at least, "ok") to buy a new CPU and put on the socket?

Sorry for the wall and thanks for all and any help!

Specs:

CPU: AMD FX 6100 @3.3GHz about 3 or 4 years old
Normal Temperatures: 65º C max when benchmarking, before the incidents described in this post
Cooler: Akasa Nero 2
Thermal Paste: ZM-STG2 (Zalman)
Motherboard: M5A78L-M LX/BR (a small variation from the LX version)
GPU: EVGA GTX 650 TI
PSU: Brand new Seasonic 620W

Update 1: Tried 3 more times to reinstall the heatsink with the proper amount of thermal compound and all, just in case I was missing anything and, as stated before, I can confirm the 70ºC on the cooper's pipes (won't let it reach 80ºC again). It boots into Windows and works like a charm until it gets to 80º+ and shutdown. I'd say the CPU is fine, except it's reaching its maximum temperature way too fast in a relatively nice vented system. Has anyone seen this behavior before?

Update 2: Finally was able to test the CPU in another motherboard and it seems to be ok, the thermometer is crazy (jumping up and down, with difference between the cores) but the overall temperature is stable bellow 40ºC on benchmark. I'm looking into my heatsink: changed the cooler but it doesn't seems to help. I wonder how it stopped being efficient all of the sudden?

Final Conclusion: The heatsink was busted. I found several fissures around the copper pipes.
As you can see on the link below:
http://www.tomshardware.com/forum/269575-29-modifying-heatsink-copper-heat-pipes
The solution is to buy another heat sink! Thanks everyone for the help!

 

[ffg7]

is the cpu heatsink fan assembly firmly seated as sounds like it is not with those temperatures rising quickly?

 

[danny2000]

1. You submerged the whole processor and heat sink in bowl of alcohol?
2. I doubt that you used any kind of glue; CPU(S) sometimes stick , especially with some of the older Pentiums (9 years ago?); I forget the socket number, but it was one of the one's where you could not get the processor out of the motherboard until the arm was lifted; If the processor was stuck, you could not lift the release arm; A catch 22; I tried to explain that to my boss one time....but she did not understand/believe.

 

[ffg7]

socket 479 for intel. I've pulled a cpu with a stuck heatsink out of the socket without releasing the arm & cpu still worked. alcohol will evaporate quickly.

 

[Isaac Queiroz]

@ffg7 Thanks for the answer. The heatsink is assembled as firm as it gets, it won't move in anyway. The temperature with the computer before the thermal paste incident was normal, 65ºC max when benchmarking the whole system. I should have said, it's not a "omg my pc is heating" situation, the CPU is probably lost, I just want to check if anyone heard of this behavior before and if the Motherboard is faulty as well.

@danny2000 Thanks for the answer. I may have used a thermal adhesive (http://www.arcticsilver.com/arctic_silver_thermal_adhesive.htm), I think I borrowed from a friend and never bothered reading the label as I also thought there "is no glue". And yes, the processor was in a solution of laboratory grade Isopropyl Alcohol as last resort for cleaning it (I would not recommend another kind of alcohol for it).

 

[ffg7]

we used that alcohol to clean circuit boards & cpu's when I was working at Celestica.

 

[danny2000]

"submerged" is the scary part of the sentence; that is not advisable by anyone, regardless of what type of alcohol that you choose to use; It is more of a last-ditch effort before you throw it away.

When you said glue, I thought that you meant some type of glue, and not something designed for CPU(S). Update: I see what you meant here. "perm" stuff.