I've tried searching the forums for info on this, so I'm sorry if I overlooked anthing:
My computer completely shut down when playing the game "Dead Space" and when I turned it back on it would stall at the black screen after the windows logo. After crashing to a white screen in safe mode I thought that the CPU might be overheating and when I opened the case, the CPU heat sink was completely caked with dust. I cleared the dust from the sink and case and i was able to restart windows normally.
I was then able to use the Catalyst Control Center to determine that my Sapphire 3650HD AGP was idling at 80 celsius at stock settings! Perhaps against all good judgement, I then ran rthdribl to see if the card was indeed over heating at the computer again shut down within 30 seconds. The last temp I read was 108 c.
So, my question, if I have diagnosed the cause of the shutdowns correctly, is can there have been damage done to the GPU if I have not overclocked or are these shutdowns the result of a safety measure to prevent the card from burning up?
Yes the card getting that hot would cause the PC to shut down. It would become unstable at the temperatures you are talking about.
There could have been damage but that depends on how often its been doing it. Just the once or twice you may get away with but if its been doing it for a while you may have cooked it.
I would suggest that the GPU is probably caked with dust internally like your CPU cooler was.
You could try getting a can of compressed air and carefully blowing out the plastic casing (if its that kind of card). Or you could take the plqastic off so you can see whats going on and clean it manually.
There are online guides for doing this kind of thing and also for removing the heatsinks and replacing the paste if that is needed.
Basically if teh card is clean but the temps are up then that indicates that the paste that connects the heatsinks to the chips needs replacing.
So I got into the card today. Here's what the fan and the chip looked like:
Think it's clogged?
I was surprised to see what looked like arctic 3 on the chip, but as you can see it was sparse and perhaps the over heating caused it to burn somewhat? I replaced the thermal with arctic 5.
As for the fan, even after "unclogging the sink" and airing out the fan motor it still wouldn't spin in the machine. I checked the header (2 pin) and it was in place. The fan seemed to only spin for a half-second every so often. After tapping it a few times it started up, but its making a slight grind noise (I checked to make sure i reassembled it correctly so there is no interference with the hardware) and I think the fan is still running really slow since I'm still idling at around 70 degrees.
It also seems to heat up very quickly. It hit 90 in about a minute using rthdrib, but still no artifacts. Could quick heating be a sign of damage or is it more likely the fan still being slow from dirt?
I'd like to try and just get a replacement brushless because I don't have a lot of cash to spend on aftermarket cooling. I'm having trouble finding one that match the specs.
It's a DOCENG FS125010
Unless I measured wrong, it seems to be exactly 2 inches and not 50mm. Maybe GPU fans aren't standardized like case fans though.
Please stop be now and ridicule my actions thus far.
Got this thing solved, but I really can't fathom how I was able to avoid damaging this card in light of what has transpired.
I was finally able to oil the fan, which was no easy task seeing how their was no bearing access under the sticker. I had to pop the fan right off the motor while avoiding damage to the fins; the spindle was completely gummed inside.
After cleaning and oiling the bearings I'm now idling at a cool 40 degrees and topping out at 67 with the ATI fur cube and still no artifacts. I guess I got lucky this time.
So what do you think? Should I start overclocking now?