Computer occasionally boots into a grey screen.

[Posmo]

The thread title pretty much expains the issue. Computer boots up okay until the Windows XP loading bar thing comes up, after which sometimes instead of giving the "welcome" screen, it just turns gray. Occasionally the gray screen has some colour "dots" (sort of like artifacts with an over-overclocked Gpu) on it. Up to this point I have resolved the issue by just rebooting to make it work, but it'd be nice to know what's causing it, and how I might go about fixing it.

Specs:
Xp sp.3
Q6600 @ 2880 Mhz (Overclocking not the cause, Ive tried with it and without it)
Abit Ab9 pro mobo.
2x1gb Buffalo 800mhz ram 15-5-5-5
Gigabyte 8800gt 512mb @ stock
40gb IDE Hd
300gb Sata hd.

 

[cvrefugee]

Video card drivers maybe?

 

[Posmo]

Video card drivers maybe?

Not the issue, I've swapped them out maybe a dozen times and currently run a fresh install(driver removal tool thingy) of the latest ones, and yet the problem prevails.

 

[awdrifter]

Does the computer crashes when the grey screen happens or can you still hear sounds or see the cursor moving.

It could be either the PSU is not providing enough power, or the video card is dying.

 

[Dogbert]

Tried reinstalling XP?

 

[Posmo]

Does the computer crashes when the grey screen happens or can you still hear sounds or see the cursor moving.

It could be either the PSU is not providing enough power, or the video card is dying.

I can hear the xp welcoming sound, and I can see the Cpu light flickering so I think it's safe to assume that something is working "behind" the gray screen. I have an LC power (cheapo, I believe) psu, I think it's 500 or 550 watts, can't be bothered to crack the case open to check. Everything works fine if I get through the boot ok. Is there any definite way to check whether the Gpu or Psu is the culprit? Unfortunately I have no other Pci-e gpu's or sufficiently powerful psus to test with, so I would need a program to see if they work ok.

Tried reinstalling Xp, no effect, so it's probably a hardware issue. If it's the gpu, im willing to swap it out (since the failure could break other parts, and Psu's are not that expensive) If it's the Gpu, im going to wait until it dies. Wouldn't want to replace a 8800gt, since it's quite sufficient for my needs atm..

 

[awdrifter]

Not sure if there's a program to check the PSU, but for video card, you can use ATI Tool to stress test it. If not artifact is detected within 30 mins, your video card is most likely fine.

 

[MadCow809]

sounds like the GPU is dying.

 

[Posmo]

Not sure if there's a program to check the PSU, but for video card, you can use ATI Tool to stress test it. If not artifact is detected within 30 mins, your video card is most likely fine.

I've stress tested the Gpu with atitool for about 2 hours now. Everything works fine, it's running a bit hot, peaking at 82C. But I doubt the problem is related to heat because it always happens at startup., when the computer is the coolest.. The card is a pre-overclocked one, so that explains the heat a bit, and no artifacts were detected.

I'd like a way to check if it's the gpu because I really don't want to go out and buy a new one just to find out that the old one was alive and well. Thanks for all the help.

 

[awdrifter]

If the GPU is 2hr stable it's most likely fine. When did this started to happen? Did you add or changed a part on the computer recently?

 

[Posmo]

If the GPU is 2hr stable it's most likely fine. When did this started to happen? Did you add or changed a part on the computer recently?

No parts have been changed on the computer, and it started happening about.. 2 months ago, for absolutely no reason I can think of.. Could it be possible that the RAM is the culprit? If I run the ram at even 801mhz (800 stock) I get errors in memtest86 test 4.. I'm thinking it might be faulty. DO you know of anyone who has had similar problems due to borked ram?

 

[PacketCollision]

B0rked RAM can cause all sorts of problems, but if memtest doesn't show any errors after a round or two of testing, it's probably not the problem. I'd place my bet on the PSU, mostly because cheap PSUs are missing most of the power-conditioning circuitry and can output 20+ times the allowed noise and ripple, which can cause intermittant issues even if they are able to power all the components (off-brand power supplies can often only supply about 1/2 of their rated wattage). Having a good-quality power supply is vital to a stable and long-lasting computer, so even if it's not the problem, I'd consider it a good investment.

 

[awdrifter]

No parts have been changed on the computer, and it started happening about.. 2 months ago, for absolutely no reason I can think of.. Could it be possible that the RAM is the culprit? If I run the ram at even 801mhz (800 stock) I get errors in memtest86 test 4.. I'm thinking it might be faulty. DO you know of anyone who has had similar problems due to borked ram?

Try giving your ram a little bit more voltage. Maybe 0.05v-0.10v over what the specification says, see if it'll fix the problem.

 

[Posmo]

Try giving your ram a little bit more voltage. Maybe 0.05v-0.10v over what the specification says, see if it'll fix the problem.

I've upped the ram voltage from the mobo stock 1.8v to 19.5v volts in the hope of fixing the problem already, but since I'm not sure about the specs on my ram, how far can I go safely on the voltage?

 

[awdrifter]

It depends on the memory modules they use on the ram. The JEDEC DDR2 spec is 1.8v, but some rams like my Patriot Extreme spec says it can run at 2.2v.

 

[MadCow809]

I've upped the ram voltage from the mobo stock 1.8v to 19.5v volts in the hope of fixing the problem already, but since I'm not sure about the specs on my ram, how far can I go safely on the voltage?

2.1V for most CAS 4-4-4 timing ram. or 1.9V for most CAS 5-5-5 timing ram.

You can turn it all the way up to 2.1 without even worrying about frying your ram, in fact, from personal experience, anything less than 2.4v won't cook your ram.