Overheating GT430 in htpc machine

[lpallard]

Hi all!

I am experiencing a strange overheating problem with a Nvidia GT430 videocard in my htpc and cant figure out why...

Basically, the machine has been sitting unused in a storage room for about 5 months until I put it back in action tonight. Ive checked for dust inside the casing and inside the GPU/CPU fans, no dust (except maybe a slight film, nothing more).

Last time I used the machine (back in january 2012), it was working perfectly. Same hardware back then, nothing changed...

Tonight I started the machine and ran XBMC for about 10 minutes before the machine randomly rebooted.. Then back on the desktop, I tried to figure out what happened, to no avail.. Then I launched XBMC once more and sure enough, about 10 to 15 minutes later, Boom! it rebooted..

This time I suspected that the videocard was overheating, sure enough, launching the Nvidia control panel showed idle temperatures of 45-50C with fan speed at 65%, and temperatures steadily increasing up to 110C when XBMC is running.. Just running the glxgears (I am in Linux) makes the GPU go up to 60-65C...

This machine never overheated before. Like I said, no hardware change. The ambient temperature in the room is around 27C. The machine is sitting on the very same entertainment unit as it was before, therefore there are no air flow obstructions caused by lets say, books, a wall, etc...

Anybody could comment on this issue??

The only thing I am thinking is the thermal paste... It may have become bad but why?

 

[drums101]

yea the thermal paste could have gone bad with age...you should take the gpu apart and clean it out and replace the paste...you should also make sure that the gpu fan still works after sitting around not being used...there is a possibility that it isnt spinning....you can also try installing some new drivers to try and fix it

 

[Deemo13]

Make sure the fan is actually working. The thermal paste may have also dried up or something, you might want to replace the goop sometimes gets hard over time.

Also make sure there's no dust in the fan itself.

 

[lpallard]

Thanks guys for suggestions, unfortunately except replacing the thermal paste I had already tried everything else ... ive made sure the heatsink was clear from dust, sure was, and of course the first thing I checked was if the fan was turning which is the case..

Im VERY surprised about the thermal paste because this computer is about 18 months old and never overheated before .. or at least not to the point of rebooting.

I did a test after I posted here yesterday : I took the case cover off and no more overheating. This would not explain why I ran this machine in a less ventilated and more cramped area and never overheated before ....

Very strange. All I can think is that the thermal paste got bad suddenly ... I however noticed when i took the case cover off that there was a tremendous amount of heat radiating from the video card as I literally burned my hand touching it and could feel the heat about 10 inches away... not to mention the smell of overheating electronics

Im clueless. If the paste wasn't doing its job transferring the heat to the heatsink, it wouldn't be that hot right???

 

[Deemo13]

What are you doing in terms of monitoring? Is it just a stock computer case?

 

[lpallard]

No monitoring per se.. just launch the nvidia control panel and I go in the thermal settings... there I see a thermometer with the temperature.

Yes a stock case.. large horizontal lian li case with nothing special..

Same machine, same software, same furniture, same everything but now it overheat.

Ill do more testing. Normally I'd suspect either the fan doesn't run or its packed with dust but neither are true.

 

[Sunius]

That's very strange indeed. The card should not be able to heat the heatsink so the point of burning your hand. No way.

 

[Deemo13]

Maybe try to get like MSI Afterburner to turn up the fanspeed.

 

[Sunius]

The card has 49 watt TDP. That's really little, too little to heat it that much... For some reason I think the card may be defective in a way that it uses much more voltage than it has to. Is it still under warranty?

 

[Fongleitus]

Hi all!

I am experiencing a strange overheating problem with a Nvidia GT430 videocard in my htpc and cant figure out why...

Basically, the machine has been sitting unused in a storage room for about 5 months until I put it back in action tonight. Ive checked for dust inside the casing and inside the GPU/CPU fans, no dust (except maybe a slight film, nothing more).

Last time I used the machine (back in january 2012), it was working perfectly. Same hardware back then, nothing changed...

Tonight I started the machine and ran XBMC for about 10 minutes before the machine randomly rebooted.. Then back on the desktop, I tried to figure out what happened, to no avail.. Then I launched XBMC once more and sure enough, about 10 to 15 minutes later, Boom! it rebooted..

This time I suspected that the videocard was overheating, sure enough, launching the Nvidia control panel showed idle temperatures of 45-50C with fan speed at 65%, and temperatures steadily increasing up to 110C when XBMC is running.. Just running the glxgears (I am in Linux) makes the GPU go up to 60-65C...

This machine never overheated before. Like I said, no hardware change. The ambient temperature in the room is around 27C. The machine is sitting on the very same entertainment unit as it was before, therefore there are no air flow obstructions caused by lets say, books, a wall, etc...

Anybody could comment on this issue??

The only thing I am thinking is the thermal paste... It may have become bad but why?

I just bought a second-hand GT 430, installed it in a Dell Inspiron and began to play Second Life when the card overheated and quit, the computer screen went black, I felt the heat sink on the card, it was so hot I couldn't touch it it with my hand, I gave it a few minutes to cool down and fired it up again and found the fan was spinning very slow, although GPU-z reported it to be be spinning in the "Normal" RPM range. Having experience with fan bearings running out of lube and the fan either running slow or not at all, I got to work on removing the fan from the heat sink and peeled the labels of the fan looking for a rubber plug, customarily those fans have one serving as a bearing cover, there wasn't one, so I was able to get a couple of shots of WD-40 in between the fan and the housing, gave it a few preliminary spins manually, using my fingers, and it freed right up. Cleaned any excess WD-40 of the blades and housing and re-installed it in the heat sink and put it back in the machine. It's been running Second Life 24/7 for about 3 days now at between 60 and 70 degrees celcius. I used NVidia inspector to crank the fan up to 100%, but it sits at the 96% mark. Good enough for Me. From what I gather these fans seize up from sitting for any length of time.