System slightly unstable. Video card?

[Scottee]

Just put together a new system. The only non-new part is a Ti4200 that I got used here on FT. There first time I powered up the system, I noticed the fan on the video card sputter and then stop. A couple of posts later, it seemed to loosen up and run continuously ok. Later while installing drivers and basic software and stuff, the system randomly rebooted a couple of times. It wasn't happening often, maybe 3 or 4 times over the course of like 3 hours. I touched the heatsink on the video card once while it was running and it seemed pretty hot to the touch. I've also gotten a few flickers on the screen upon start up of Windows, and this started after I downloaded the latest drivers from Nvidia. But the reboot happened before the new drivers and also after. Is it possible the fan's not sufficiently cooling the video card and that's causing the reboots?

P4 2.53
ECS PT800CE-A
2x256mb pc2700 ddram (512 total)
160gb WD 7200rpm hard drive
used ti4200
Samsung 16x dvd burner
generic 300W power supply that came with Apex case


I'm also most probably gonna change out the power supply as soon as I find a quality one on sale, possibly at Fry's.

 

[Kogan]

It's been a while for me, but when my geforce cards were overheating, they would just have weird video glitches instead of rebooting the system. Check some other components after you've ruled out the video card

Run some video card intensive stuff like loops and benchmarks in 3d games. if it messes up a lot there, has video problems, freezes, etc, then there's a good chance it is the video card.

Also try to improve the cooling on the video card or underclock it temporarily to see if it makes a difference.

Oh, and an even easier test - put in a different video card if you have one available

 

[Scottee]

So I'm now pretty sure it's the video card, cuz I starting turning down the hardware acceleration on the video card. I got it down all the way to the second lowest setting, and it still had the screen glitches. Then I completely turned off the acceleration on the video card and there seems to be no more screen glitches, and I haven't had any random reboots.

So now the question is, can I fix this problem in any easy way? I'll go try some old nvidia drivers and see if that helps. Any other advice on this acceleration problem? I assume the card won't run things the way it should like this? I haven't installed direct x yet either. Would that make things better or worse?

 

[Scottee]

So I went back and installed some pretty old nvidia drivers (44.03) and the video glitches are gone. The video card's running at full acceleration now just fine.

Then I found out XP defaults to auto resetting on errors, and I found out how to turn that off. (I'm new to XP)

Everything was then going pretty well, and I installed a bunch of new stuff. Then I decide to load and try running Warcraft III, cuz that's the most computer intensive game I have laying around. It loads fine. The intro video runs fine. Everything seems fine. Then while playing around with the settings, I get the error and the game closes.

FATAL ERROR!

Program: C:\Program Files\Warcraft III\war.exe
Exception: 0xC0000005 (ACCESS_VIOLATION) at 001B:0043A184

The intruction at '0x0043A184' referenced memory at '0x40060064'.
The memory could not be 'written'.



Does this have to do with the ram? Anyone have any ideas?

 

[fishmonger12]

why don't you try kogan's idea and try and ISOLATE the problem. torture tests are the way to do this. if you think it's a ram problem now, try prime95. if you think it's video, try 3dmark or something. but don't just go uninstalling stuff and screwing with your hardware. you're going to create more problems than you solve. old drivers can help, but they can also cause incompatibilities.

 

[Scottee]

I ran Prime05 overnight. The ram seems to be fine.

So I suspected the power supply and downloaded motherboard monitor, as someone suggested in another thread. I noticed the +12.0V line is normally running between 11.75-11.81V. is this normal?

Then I run Warcraft III to try to get the system to crash. Sure enough it does. I come back and check the monitor log, and it shows a drop on the +12.0V line down to 11.56V right before the crash. And it may have even dipped lower, as it's only reading every 10 sec.

So is the +12.0V line the culprit? Also, isn't it the +12.0V line that runs the video card? Which would explain some of the video glitches upon Windows startup? New power supply needed?

 

[fishmonger12]

worth a shot. a decent antec powersupply doesn't cost too much.

 

[Paperlantern]

yeah, massive, massive, massive importance of a good power supply... its like buying all performance parts for your car, and still using cheap crap gas in it, that gas isnt gonna run thos parts like they are designed, and may damage them

 

[Scottee]

After all that investigating of the voltage and the power supply today, I decided to check out the ram again, even tho it passed Prime95. So I swap one stick of ram out, cuz if it's the ram, then the system should still run on one stick. So I take one out, and go back and run 3dmark2001, which had been crashing earlier. It runs fine for an hour. Then I'm thinking that either the other ram stick is bad, or the other dimm slot is bad. So I start swapping around the ram in different configurations, and eventually, everything's working fine with both sticks in. Ran Warcraft III to make sure, and played a whole game just fine, whereas before it was crashing just going thru the menus.

Conclusion: one of the sticks of ram just needed reseating.

Now I feel kinda dumb for not trying that earlier in my tinkering. Will run the system overnight just to make sure.