Graphics card/driver issues

[LeakySeals]

I donate parts and labor to people in need. In one a case a guy fried his graphics card on a Dell 8300 running XP pro connecting the monitor (bent pins shorted the card). I gave him an older card out of my stash, installed the driver for it. A Ti-500 Was fine for several months then blue screened complaining about the driver. I re-installed with a different driver version. Was fine for a few weeks, eventually blue screened again. I went over last night, re-installed the driver, it was fine. But I replaced the card anyways.

Its XP, so there haven't been any updates the would effect it. Is it sharing something with another device that causes incompatibility over time? or is it just a lousy card and scrap it?

 

[LeakySeals]

What makes me suspicious about compatibility over time are the observations when the machine boots. When its having the issue - as it boots up the screen flashes, distorts, etc. as the driver is initialized. I run ccleaner (using modded winapp.ini), reboot, the boot process is normal.

Strange, usually it works or it doesn't.

 

[uc50ic4more]

Originally Posted By: LeakySeals
Strange, usually it works or it doesn't.



... And when hardware is (unpredictably) somewhere between "works" and "doesn't work" it's usually a hardware problem.

 

[2004tdigls]

i bet if you take out the graphics card you will see damage on the motherboard, take a look at the mounting surface on the motherboard, you may see a crack

 

[jjcom]

I would wonder if when the previous card fried, it might have lightly toasted a few traces on the board. I've seen that happen when cards have fried and even a few fans that have shorted out. It'll toast the board just lightly and it'll still work most of the time.

 

[LeakySeals]

Probably so. The plastic connectors between the motherboard and the power harness are brown on a few connectors, so something caused some heat. But thats nothing new, I've seen that on many older machines. I did try using some electrical connection spray on it, plugging it back in.

But from experience it just seems like something goes on with the drivers/bios/os where it gets "out of sync" after a while. Like they share something, and one or another gets updated causing an issue. Something like that. Suppose the truth will come out if that known good card fails. Will update this thread if it does.

 

[OVERKILL]

I would wonder if potential lack of power is causing this. What was the original card? Dell tends to spec the minimum possible PSU that would work for their units in a given configuration. That'll cook a card over time.

 

[LeakySeals]

It was a factory NX6200 LE 256. You can see the caps are popped. Probably all thats wrong with it. Both are AGP slotted, likely the motherboard connector pins that show some heat. Maybe go with a PCI card, avoid the slot altogether?

 

[OVERKILL]

Originally Posted By: LeakySeals
It was a factory NX6200 LE 256. You can see the caps are popped. Probably all thats wrong with it. Both are AGP slotted, likely the motherboard connector pins that show some heat. Maybe go with a PCI card, avoid the slot altogether?


I would say that the Ti-500 would draw more juice than that 6200.

 

[LeakySeals]

I agree. The 6200 has no fan, the Ti-500 does. That may change things. I heard from the guy, he says its fast again. But he is unaware of TCPOptimizer I added.

The ti-500 is in my next giveaway to test. Looks like a dell, but she's got HP, emachines and gateway parts in her

I tried a different driver. Not new or old, a driver the kids claim gave them "points" whatever that means.

 

[PandaBear]

I think this card is probably at the end of life, it was a dated card from late 90s, and electronics are usually designed for 10 year of use.

I'd put the 6200 back in (may be similar in speed anyways) or spend $10 to buy a more recent card even if used.

 

[LeakySeals]

Its the fifth day of testing on the PC scheduled for giveaway. using it right now. The NVIDIA driver version I'm using seems to have stabilized the Ti-500. I think its something to do with the heat probe and the on-board card fan. More active than other driver versions. I can hear the little bugger, noisy.

The guys PC has been fine with the card swap, so I think the motherboard survived.