Crazy temp spikes and dried out thermal paste

[firefiber]

Hi guys.

I recently built a PC (well, almost four months now). The thing is, after the first month of running smoothly, I started OC'ing it bit by bit. And when running benchmarks and stability tests, I occasionally hit 70°C and very rarely 80°C. I got my CPU (FX-8150) up to 4.13 stable. After this, I didn't really push any more, and didn't run benchmarks/stability tests, etc.

However, about a week ago, my CPU started giving me crazy high spikes in temperature. I'd be doing nothing, and then out of the blue, my fans would start screaming, and I'd get warnings saying my temps were at 70°C. About three days ago, my PC started freezing/crashing and BSODing. So I opened it up, and took off the heatsink. Well the thermal paste had all but dried up and vanished.

I wanted to know if this is actually possible? I mean, in this time frame? Isn't it usually supposed to last quite a while? It's only been four months since I started using it. Is this normal? And could there be any other reason for the spikes and BSODs?

I bought a TG-2 Thermaltake thermal paste stick yesterday anyway, so I'll clean out the stock paste and apply that.

 

[nekulturny]

This doesn't simply happen overnight. 70c is way too damn hot for that processor. If you were hitting 70c in the beginning during your stability tests, your overclock was not stable. That is again, too hot. What CPU cooler are you using? It may be heavily debatable but the brand of thermal paste isn't really that big of an issue. Arctic Silver 5 is just as good as the crap that comes with CoolerMaster HSF's.

You're frying your chip, no, with temperatures like that, I don't think theres another cause. What have you done to the voltage settings? You shouldn't need to touch the voltage at all to get an 8150 to 4.1GHZ.

For future reference, AMD states the max safe temperature for FX CPUs is 61c. If you're sustaining temps higher than that under load, your overclock is not "stable".

 

[abCasPeRR]

Agreed with nekulturny, to run that CPU that hot for that amount of time isnt good. A stable overclock means that it doesnt overheat and it doesnt shut itself down due to any problems. Im surprised that it didnt already shut itself down after having getting to that temp. As for CPU cooler what are you using ? Those FX processors get pretty hot so I wouldnt even recommend using the stock cooler on it.

 

[mocchan]

Just for reference, Antec's silicone thermal compound is pretty shotty. I've purchased it just for the heck of it, boy, I saw temperatures 5-8C higher than Antec's own Formula 7 compound :lol:

 

[firefiber]

This doesn't simply happen overnight. 70c is way too damn hot for that processor. If you were hitting 70c in the beginning during your stability tests, your overclock was not stable. That is again, too hot. What CPU cooler are you using? It may be heavily debatable but the brand of thermal paste isn't really that big of an issue. Arctic Silver 5 is just as good as the crap that comes with CoolerMaster HSF's.

You're frying your chip, no, with temperatures like that, I don't think theres another cause. What have you done to the voltage settings? You shouldn't need to touch the voltage at all to get an 8150 to 4.1GHZ.

For future reference, AMD states the max safe temperature for FX CPUs is 61c. If you're sustaining temps higher than that under load, your overclock is not "stable".

Nope, I didn't mess with the voltages. And yeah, you're right. I set the cut-off for the benchmark software at 80. I guess I set it too high. If it hit 80, it just stopped. But this was only when I was trying to get it above 4.2GHz. I found that 4.13GHz (with stock cooling and no voltage increments) was the sweet spot. At this speed, the benchmarks gave me 60°C to 65°C. But my point is that I did stress it a bit while trying to get it past 4.2GHz. I'm wondering if it's actually possible that all the thermal paste could have dried out so fast 'cause of this? It seems really unlikely, though.

 

[nekulturny]

It didn't dry out overnight no. This happened over time. Like I said, your CPU most likely was always running far too hot. My advice to you would be to remove your overclock completely and count your blessings that (and theres no guarantee you haven't) totally damaged your chip. Lest buying a new one is within your financial reach.

 

[abCasPeRR]

Nope, I didn't mess with the voltages. And yeah, you're right. I set the cut-off for the benchmark software at 80. I guess I set it too high. If it hit 80, it just stopped. But this was only when I was trying to get it above 4.2GHz. I found that 4.13GHz (with stock cooling and no voltage increments) was the sweet spot. At this speed, the benchmarks gave me 60°C to 65°C. But my point is that I did stress it a bit while trying to get it past 4.2GHz. I'm wondering if it's actually possible that all the thermal paste could have dried out so fast 'cause of this? It seems really unlikely, though.

Stock cooling was probably the issue, buying a after market heat sink would have saved you from frying the cheap. The hyper 212+ is a cooler ive used for the past year, and a lot of other people use because its so cheap yet performs amazing.