Export thread

Halforums, I come to you troubled and anxious to my very core. Troubled, because this is the third major problem this computer has had in its short life, and anxious, because for the next month, I have 2 ESEA TF2 matches per week.

Problem: My computer seems to hate vidya games. I can't play TF2 for more than 5 or so minutes before my system completely crashes. If I surf the interwebs for 30 or so minutes beforehand, it takes less than 2 minutes to crash when I launch TF2.

I have some reasons to suspect that this is a heating problem, because:
A) the reasons above
B) this computer is roughly 3 years old
C) she also crashes when I play less taxing games, but it takes anywhere from 5-20 minutes, depending on the game
D) rebooting immediately after a crash inevitably leads to another as soon as I reach my desktop.

But my problem with this theory is that the computer (she's a laptop) doesn't feel hot at all on bottom, which it always has if I crash from overheating. Also, my fans are all perfectly clear.

So please Halforums, do my team and I a favor, and figure out just what the fuck is wrong with this thing.

It may very well be a hard drive issue or something could be damaged from the times it's overheated.

These kinds of things are very difficult to diagnose over the web.

Laptops are tough to fix if hardware is the issue. If you think it's over heating you could try putting a cooler underneath it. http://www.newegg.com/Product/Product.aspx?Item=N82E16834998921. It could be that after 3 years your heatsinks aren't making good contact with the CPU and GPU, so it's not transferring heat like it should.

Also make sure your video card drivers are up to date.

if you've previously been crashing your computer from overheating, you're doing progressive damage to it. It will eventually begin crashing even at lower temperatures. heat is enemy #1 for chips.

Thanks for the replies guys.

This thing has probably crashed approx. 5 times due to overheating, and I hope to god that's not the problem.
Tin, when you say that it's progressive, do you think that would apply to all cases? I ask because probably in the past 6 months or so I haven't accidentally crashed it due to an overheat, and it has seemed just as resistant to going down as when I originally bought it. Would you say this sudden decline is possibly due to progressive damage?

Shakey, I've tried to put a cooler underneath it, it didn't seem to make her last any longer.


Updating drivers now.

What type of laptop do you have? There is a known issue with the Nvidia cards overheating in some Dell models. Mine didn't feel hot at all when it was overheating.

One other thought is RAM, but everything seems to lend itself to the overheating.

Try installing some software to monitor your temperatures. Some of the graphics adapter makers make sofware to monitor their cards.

Basically, can you give more information your hardware?

Running a scandisk to make sure that you're hard drive is working on all sectors would be my first step towards resolving the problem.

It sounds overheating related, but I'd try to block out the simple issues first.

Run a check disk on your hard drive.
Run tests on your RAM. Some versions of Windows have a RAM test in the advance boot menu, and Ubuntu has a RAM test on it's live disc.
Uninstall and clean out all you know how of your old drivers.
Run a program like Speedfan to check temperatures of your CPU and mobo chipset. I'm not sure what to run to check graphics card temps.
Run a CPU intesive task (like Prime95) to see if the CPU is the weak link, or if it's likely your graphics card.

Methuselah said:

Thanks for the replies guys.

This thing has probably crashed approx. 5 times due to overheating, and I hope to god that's not the problem.
Tin, when you say that it's progressive, do you think that would apply to all cases? I ask because probably in the past 6 months or so I haven't accidentally crashed it due to an overheat, and it has seemed just as resistant to going down as when I originally bought it. Would you say this sudden decline is possibly due to progressive damage?

Click to expand...

It's possible, but as others have said, it's hard to diagnose over the internet.

All cases of overheating do damage to your chips, and it's progressive damage..one day the computer might seem fine, and the next, it's too late.

Speedfan is putting my GPU around 70 and my cores are hanging around 50.
I rann Prime95 for around 2 minutes, it brought my core's temps to 80 and they seemed to be still climbing when I stopped the test.

Yeah, hitting 80 is hot enough to fry your CPU.

So, what processor and GPU do you have? Acceptable temps vary a little depending on the architecture.

Processor: Intel® Core™2 Duo Processor P86001 (2.4GHz, 1066MHz FSB, 3MB L2 Cache)
GeForce® 9800M GTS Graphics with 1GB of GDDR3 Discrete Video Memory and Intel® PM45 Chipset9

Bleh, you guys post too fast

80C isn't too bad, your CPU can take up to 105. GPU's tend to be able to go even higher then that.

Do a memtest, hdcheck with forced repair. If those both turn out fine, I'd say try a reformat and reinstall.

Seraphyn said:

80C isn't too bad, your CPU can take up to 105. GPU's tend to be able to go even higher then that.

Click to expand...

I was about to say "wut?" because desktop chips don't like to operate at 80C, but apparently the mobile parts can.

Regular operating temps for an Intel core duo 2 is between 50 and 70 degrees. 80 is getting hot, and he said it was still rising when he ran the test.

105 is the MAX operating temperature, meaning that if it exceedes that, it will fry the processor. The chips are designed to shut down before the heat reaches that threshold.

Yeah, I wouldn't want it to go above 90. Just because you can get to a certain temp doesn't mean it isn't going to do some damage in the long term.

If 105 is the max, 80 isn't a disastrous temp, certainly if it's under full load. It's not the temp at which it'll shut down, that'll probably be at the 95 mark.

If you don't have warranty anymore, I'd say open the thing up and give it a good cleaning with some clean compressed air. That's never a bad thing to do and it might help with the temperatures. Then run prime again see if it'll finish.

Bowielee said:

105 is the MAX operating temperature, meaning that if it exceedes that, it will fry the processor. The chips are designed to shut down before the heat reaches that threshold.

Click to expand...

You're right that the thermal specification is the maximum safe temperature, but compare it to that of a desktop chip. My E7200 has a limit of 74C, an E6700 is listed at 60C. I doesn't seem to me that 80C is in the danger zone for a mobile part.

Regular operating temps for an Intel core duo 2 is between 50 and 70 degrees. 80 is getting hot, and he said it was still rising when he ran the test.

Click to expand...

Are you talking mobile or desktop? 70 is over the limit for many desktop processors, but 80 doesn't even stretch the limit for mobile parts.

Well considering I was running the CPU at 100% as it climbed to 80, I'm not sure TF2 would crash it in such short a time. Once I finish some work I'll check my temp and run TF2 till a crash.

Thanks for all the help so far guys <3

So I just played a bit of TF2 while tracking my dual core and GPU temps. At the crash, which happened about 2 minutes into launching, my cores were both at about 63, and didn't seem to be rising significantly. My GPU was at exactly 100 upon crash. Have we found our problem?

I'm not sure if something like ntune or rivatuner works on mobile chipsets, but you could try underclocking it or forcing the fans to go all out. Use it at your own risk though, you can do even more damage if you change the wrong stuff. Problem is, with a laptop, you can't add a better cooler. You can clean it out and hope it works better, but that's it.

Have you done everything else that was suggested though?

Aight so I just ran an nvidia stability test on all system components, and it crashed me in record time! (20 seconds). I think I'm going to run it on individual components, and see if I can isolate the problem. All while tracking my temps, of course.


I ran the CPU test for about 3 minutes. The temp climbed slowly, and was at 78 at its peak.

annndddd ran memory and GPU stability tests. The memory test raised the temp of my cores slightly and gradually, and I let it run for 5 or so minutes. The GPU test, however, seriously spiked my card's temp, and I cut it off once I hit ~96 degrees. The thing is, after I ended the test, it continued to climb until it peaked at 99, when it slowly started to drop off. This kinda screams graphics card fan problem to me.

Finally weighing in.

The symptoms you describe perfectly match a GPU overheat. The testing you do later seems to confirm it. You say the fans are clean...but are they actually spinning? It's quite possible the bearings might be going in the fans (they still spin, but never hit their former RPMs) or something else going on with them. One thing's for sure, though. If you keep running your chips that hot, they will fail sooner (and at lower and lower temperatures, until you won't be able to run them at all). It's also possible the 'sinks may be clogged, the air path through your machine might be otherwise obstructed, or even just be that your GPU is 'weak' and unable to handle these temps.

--Patrick

Yeah I feel pretty confident I've found the problem, not that it's encouraging in the slightest. I don't even think I can open this thing up.

Any warranty recourse?

--Patrick

Pretty sure it's expired.

Does speed fan or the nvidia utilies show that the fan is working?

doh! nvidia is putting my fan speed at 193 RPM.

Well there you go. Probably bad bearings, then.

--Patrick

Is that when it is being stressed or idle? It seems slow either way, but it does slow it down if it isn't needed.

Most internal fans (the ones smaller than 100mm) idle at around 1000-1500, ramping up to 5000-6000 (or even 7000+! Loud!) when under full load. If his can't even hit 250, I'd say it's time for a new fan.

--Patrick

Yeah, but it doesn't hurt to check.