ATI Card Overheat Auto-shutdown Problem

[djriley]

Hey this is my first post here, I joined just now so I could ask this question.

I like to play games a fair bit on my computer. Not overly performance/graphics-intensive ones, mainly Team Fortress 2 and Portal 2. Anyway, all the games I have been playing have been playing have been crashing after playing for a while (usually after 15-30 minutes). No error message or anything, the game process just abruptly ends and I'm left staring at the desktop.

I think the cause of the crashing is some sort of safety feature in the graphics card where it will "shutdown" or something when it gets to a certain temperature to prevent overheating. I have CPUID Hardware Monitor and whenever the game shuts down it's only at around 50ºC, with the "Max" value at 61ºC. I've been told that the card will perform comfortably at 80ºC or even 90ºC so I'm not convinced that it really is overheating.

My specs are:
OS: Windows 7 32-bit
Motherboard: ASUS M4A88TD-M
PSU: Thermaltake Litepower 700W
CPU: AMD Phenom II X6 1090T
RAM: 2 x Kingston 4GB 1333MHz DDR3
Graphics: Sapphire Radeon HD6870 1GB

Also, my entire system will crash to a blue screen every once in a while, around once a week. I'm not sure if that's related. Also, there's a little red light on the motherboard which says "MemOK" which turns on for a few seconds when I turn on the computer. The manual says it means that the RAM isn't installed correctly but I have taken both sticks out and tested them separately in each slot and the light still comes on. I have tested them in every slot using the MemOK button and every time it tells me no errors, but the light still comes on.

Any help for any of these problems would be much appreciated.

 

[GreenDutchAlien]

Well, it shouldn't be a heat problem, 50-60 is perfectly ok.
The little red light is only on for a few seconds i believe ? I think it will always light up at the start and when it stays on it means there is a problem, not sure though.
Did you try a clean install of the newest drivers ? Or used driversweeper to remove old drivers ?

 

[djriley]

Well, it shouldn't be a heat problem, 50-60 is perfectly ok.
The little red light is only on for a few seconds i believe ? I think it will always light up at the start and when it stays on it means there is a problem, not sure though.
Did you try a clean install of the newest drivers ? Or used driversweeper to remove old drivers ?

I've reinstalled the drivers but I haven't tried driversweeper. I'll do that now.

 

[GGTY]

You could check the Event Viewer (Run > eventvwr.msc) to see if there are any error logs related to those two games you mentioned.

 

[djriley]

Well, it shouldn't be a heat problem, 50-60 is perfectly ok.
The little red light is only on for a few seconds i believe ? I think it will always light up at the start and when it stays on it means there is a problem, not sure though.
Did you try a clean install of the newest drivers ? Or used driversweeper to remove old drivers ?

Did as you said but still no luck.

You could check the Event Viewer (Run > eventvwr.msc) to see if there are any error logs related to those two games you mentioned.

I couldn't find anything in the event viewer but to be honest I'm not familiar with it so I don't really know what to look for in there. I don't think it's any particular game causing the problem, as it happens with all of my games, not just those two.

 

[sayantan]

First of all run all kind of stress test for your gpu (furmark) ,cpu (prime95) separately.Then run both of them together.If there is any fault it should show up.

 

[djriley]

Well it seems it may be a heat problem after all. I ran furmark (as sayantan suggested) a few times in the preset 720p mode and it would crash once it got to around 85ºC every time. I then took the side panel off of my case and tried it again. No crashing, and the max temp reached during the test was a little over 70º every time. I then ran the test with the panel on and then removed it halfway through the test and the temperature would almost immediately start to drop. Looks like I'll need to get get some better airflow inside the computer.

 

[sayantan]

Add two 120mm high RPM fans on the side of your case.Its the most effective way of cooling your gpu.Make sure they blow inside the case directly over the motherboard.
Hope you have provision for two fans on your side panel.

 

[djriley]

Add two 120mm high RPM fans on the side of your case.Its the most effective way of cooling your gpu.Make sure they blow inside the case directly over the motherboard.
Hope you have provision for two fans on your side panel.

Nope, the only fan mounts I have are at the rear just below the PSU and at the front down the bottom (directly behind the power button), and of course the CPU fan. There's only two fan connectors on my motherboard: the CPU fan and one chassis fan (I have a microATX board), so all I have at the moment is a fan just below the PSU blowing air into the case. I wanted to add another one at the front down the bottom blowing air inwards and out the back of the case but yeah, only one connector. I suppose I could mod the side panel a little and add some fan mounts. I'll buy a Y-connector or two and add a few more fans, but in the meantime I have a little USB desktop fan blowing into the case (not the best method of cooling but it works).

 

[sayantan]

good luck.

 

[djriley]

Nope, games are still crashing. GPU temperature isn't even reaching 60ºC and they're still crashing. Anyone got any more suggestions?

 

[sayantan]

I could't find any review of your psu or any tech spec on web.So I've to ask this:
What is the Amps rating on your 12v rails??

 

[GGTY]

In your original post, you said the games were crashing to desktop.

During the stress tests, is your entire system crashing or is it just furmark?

Can you clarify?

 

[djriley]

In your original post, you said the games were crashing to desktop.

During the stress tests, is your entire system crashing or is it just furmark?

Can you clarify?

Only furmark is crashing during the stress test, not the entire system. It's like every time a program puts too much strain on the card it just crashes to desktop.

 

[sayantan]

I'm mostly sure there some problem with your gpu.Your mobo has an integrated gpu.You may try running these test on the IGP to make sure your motherboard(north bridge controller) is OK.

 

[leon2006]

Well it seems it may be a heat problem after all. I ran furmark (as sayantan suggested) a few times in the preset 720p mode and it would crash once it got to around 85ºC every time. I then took the side panel off of my case and tried it again. No crashing, and the max temp reached during the test was a little over 70º every time. I then ran the test with the panel on and then removed it halfway through the test and the temperature would almost immediately start to drop. Looks like I'll need to get get some better airflow inside the computer.

What case do you have? It seems you need more cool air to be circulated in your PC. More cool air in , then hot air out.

 

[GGTY]

Your GPU temps are all over the place (You mentioned 50, 61, 70, and 85C) when your programs crash to desktop.

85C should be doable for the HD6870... I owned a HD6850 that I flashed to 6870 and it could go up to 95C without locking up.

If it was your GPU overheating, your entire system would lock up and you'd need to cold reboot. This makes me think it's not a bad card or poor airflow, but something else in your system that's causing the program crashes.

The symptoms you've described so far don't add up.

When you said you tested your memory, does that mean a few passes of memtest?

What is the blue screen error you get? Can you recall the message?