ATI Radeon HD 4650 Overheating

Howdy,

Recently (within the past 2 weeks), my ATI Radeon HD 4650 RV730 started to overheat. I've scoured the internet and tried several times to address the issue, but so far I've had absolutely no luck.

My computer is a little less than 2 years old and only very recently did this overheating issue surface, where the GPU's heat rises inexplicably and forces the computer to completely shut off in the middle of video games. For example, today the video card started at around 60 C idle and right now it's at 94 C idle.

I've tried to address the issue in several ways. I used compressed air to clean all fans as well as the onboard fan on the GPU. I've underclocked the GPU and made sure the fan output was at 100%. I've removed the GPU, inspected it, cleaned it again with compressed air, replaced the thermal compound, and verified that the fan is indeed active. I've tried moving wires around to help with air flow to the GPU. The drivers are completely up to date.

Basically, I've done everything that I can think of short of attaching a new fan so that it blows directly on the video card/replacing the video card itself.

I've had absolutely no problems until very recently. I have Kaspersky Internet Security 2012 installed on my computer, so I have no viruses. The highest I've seen my GPU rise to while in game was 108 C. My CPU idles at about 40 C. The mobo idles at about 75 C.

I've already ordered a replacement card, so by now I just want to know what could possibly be causing this overheating issue and what I can do in the future to fix it/stop this from happening.

Here are my computer's specs:

Windows 7 Home Premium 64-bit
AMD Athlon II 620 x4
4 GB RAM
ATI Radeon HD 4650 1GB
750 GB Hard Drive

Thanks,

Hero

 

First off, 108C??? YIKES! :eek

Second, you have covered pretty much everything except broken micro circuitry. You might have a broken thermal sensor, of course, but when they break they tend to stop working entirely, and that leads to one of two things: 1, Card claims to be running at 400C or something equally ridiculous, or 2, Card thinks it's over-heating to the point of melting, and the computer shuts off in self-defense.

If I had to guess, I'd say the GPU itself is broken and over-heating as a result. And the easiest fix for that is a new card, either via a warranty exchange or an upgrade.

 

I think the only way you can tell for sure is to get an infa red thermometer or something similar and try fire it at the card heatsink to get a reading, you'll see if its the sensor then??

Obviously unless you can borrow or rent one its going to cost a bit but just an idea

 

A low tech approach to checking the heat is to put one's hand in the air exhaust but one needs a standard to compare it to. I've seen folks tie wrapping fans to a vid card to cool it. As mentioned, everything was done by you except changing the fan.

I had an issue with my Intel Q6600 CPU getting hot, I also underclocked it to the minimum(1.8GHz) even though the HS/F had the proper reported normal RPM.:confused A HS/F replacement did the trick. Since you're getting a new card, this is a moot point for you, though it may help others. Good luck!

 

I' going to go for the less technical, Is it not possible that the card is simply had it's life or maybe been put under more pressure than it likes, more times than it should?
My boys 4850 is basically dead now, wont run anything other than desktop Aero without setting off the neighbours fire alarm lol
It's only 2-2.5 years old but has been overused and is simply dying, overheated a few times before which have weakend the card. It's perfectly possible the same thing has happend to your card.

While we all want to get 20 years out of our PC's, sometimes that just doesn't happen.

 

I must also go with "It's reached the end of it's useful life.".

:cry

 

Thanks for the replies.

Thanks for the input. I don't believe the sensor is broken because it does smell like something's cooking from time to time. Also, the sensor responds to stress I put on the card, which indicates that it's reading the temperature of the card accurately.

But yeah, I am replacing the card with the same one off ebay. Hopefully this will fix the issue.

That would be a bit expensive and I don't know anyone with one of those, unfortunately. But I don't believe the sensor is broken based on the reactions I am getting from it. I could be wrong, of course.

Thanks for the tip. I do wish I had another fan I could mount next to the card or something. I suppose I could rip one out of an old Compaq that died, but it would seem as though there's something wrong with the card itself, so it probably wouldn't h elp that much.

Perhaps water cooling might keep the card cool? I wish I knew/had the money to test these things out. Thanks, though. I appreciate the input.

Indeed it is possible. I am just searching for the reason a video card's life might be less than 2 years, though. For example, I'm just wondering what causes the GPU to out like this. I saw, for instance, someone's onboard fan had gone out so they were able to dissect it and replace the ball bearings, thus repairing the fans. I am going through the same thought process right now.

And also, what would cause that kind of stress on the card? Did I perhaps attempt to adjust my video settings too high?

Thanks for the input!

 

If the sensor reacts to changes in work load, then it's fine. But GPUs usually don't age very well, and the smell of something cooking is definitely not a good sign.

A good video card shuts itself down before it gets to the point where the stress causes damage. And by good I mean good quality components. Unfortunately, computer component brands are not manufactured equal, and cheap cards tend to be cheap for a good reason.