Computer is borked, any advice appreciated

[Ralence]

  So I built a new PC about 3 months ago, and really had no problems at all until about 2 weeks ago.  I started playing games again, and I started getting random crashes.  It starts with some graphics glitchiness, sometimes it fixes itself, most often it just locks up completely.

  First thought was temperature, CPU is running at 39C, MB at 33C, so that's fine.  Next step was to reboot with memtest86 and run some diagnostics.  It made it appear that there was some bad ram (38 errors over 16hours on the same slot).  So I replaced the ram (Both sticks), and that didn't help either.  Ran memtest again with the new ram (What are the odds?) and no errors at all this time.  Plugged the old ram in with the new ram, ran memtest, and it showed no errors either?  So I'm not sure where the faulty memtest86 results came from.

  I'm running WindowsXP Pro, and it's showing me no errors at all in the event viewer, so I'm thinking it's not driver/software related.  (It crashes on me in Eve, and in Civ4, so it's not game specific)

  I don't have an extra PCI-E card to throw in to test, or else I would have done that already.

  Is it possible the bad ram I had previously has corrupted some files or drivers?  Or am I going to be forced to dig down and start replacing the last 3 components that could possibly be faulty (mb, pcu, vid card)?

  Any advice/flames/suggestions would be appreciated.

[Murgos]

Off hand I would say power supply.  There is usually a mom and pop geek computer place around that will have the equipment (if not always the experience or knowledge, ask them when their best troubleshooter will be in and go then.  One near my house the owner has an BSEE and lots of experience) around to troubleshoot these types of problems on the off chance you will buy the replacement parts from them.

The number of subsytems in a computer are for all intents and purposes infinite, even though it appears monolithic from the outside, you could swap parts back and forth for ever and not find the culprit because it COULD just be a bad driver.  If you can't isolate the parts and test for faults then you are really just guessing.

[Shrike]

Power supply.

I had a VERY similar experience when I replaced my CPU and vid card for more gaming goodness. Everything was fine for about 3 months, then I tried EQ2 (yes, I know...). Then the real fun began. Crashes, hardlocks, eventual windows corruption. The final act was physical mobo damage and complete replacement of same.

Anyway, I"d be looking at the PSU first. An inadequate or failing one can cause of lot of weird stuff like you're seeing.

[Engels]

PSU or GPU. One or the other. Considering the random occurence, I'd venture a gues its the PSU, since vid cards tend to crash hard when they do crash. Good news is a good PSU isn't that expensive; bad news is it may actually be your home's power supply. The older the home the 'dirtier' the power coming through the walls. You may also want to invest in a UPS, which provides stable, squeaky clean power, even if you have an outage or a serious surge. Surge protectors do not protect against minor fluxuations, only big ones.

[Lantyssa]

It could be the PSU, but from the description it sounds possibly like your GPU overheating.  Is the fan on it still working?  Are you overclocking the GPU at all?

[Flood]

1. GPU overheating
2. PSU issues
3. Way outside shot - IDE Drivers.  If you have an Nvidia chipset mobo, the IDE drivers included in Forcewares do weird things sometimes.  Some (I hate to say most so) manufacturers advise not to install them at all.  Extremely unlikely IMHO but just throwing it out there.

[Engels]

This is true, particularly with the old NForce2? chipsets. I had similar problems and I just told the store to give me a different MoBo altogether, this time with something solid yet dumb like a ViA chipset. Nforce4 I hear is stable.

[Trippy]

nForce 3/4 has issues too, especially their ActiveArmor Firewall -- do not install those drivers unless you enjoy pain and suffering. There's also data corruption and compatibility issues with certain combos of nForce 3/4 motherboards and Maxtor SATA drives.

[Ralence]

It could be the PSU, but from the description it sounds possibly like your GPU overheating.  Is the fan on it still working?  Are you overclocking the GPU at all?

  GPU is still good, it just crashed again on me about 4 seconds ago with GPU temp at 43C while running nothing except Firefox.  I should have mentioned that was also a part of my "overheating" original theory of the crime.

1. GPU overheating
2. PSU issues
3. Way outside shot - IDE Drivers.  If you have an Nvidia chipset mobo, the IDE drivers included in Forcewares do weird things sometimes.  Some (I hate to say most so) manufacturers advise not to install them at all.  Extremely unlikely IMHO but just throwing it out there.

  You know, now that you mention it, one of the things I did a month or two ago was to "uninstall" the Nvidia drivers after reading about problems with them, I wonder if there's a residual driver floating around thats causing these problems, I definitely wouldn't be surprised.

  Since I really don't have that much installed on this PC yet, I'm going to try to do a format and clean install, you may be onto something with the driver thing, especially since yesterday I actually got a bluescreen with a DRIVER_THREAD error, the first time I've ever gotten any sort of error at all from this machine.

[raydeen]

Do the crashes occur when you're online only or when you play an offline game as well (with no other programs accessing the network)? I've read forum posts (CoH) about out of date network card drivers contributing to system crashes. Try upgrading your net card drivers and see if that fixes things.

[Ralence]

  Just wanted to say thanks for all the ideas/suggestions so far, it's really appreciated.  As far as progress goes, I pulled a good PSU out of my other comp (Antec Neo HE 550), and it didn't fix the errors, as for the other poster, I'm already using a UPS, so that can be eliminated as a problem.

  I did a full format/re-install of WindowsXP, to eliminate driver issues, installed only nvidia drivers for video, and the onboard lan/sound drivers, and it crashed on me then as well.  So I disabled the onboard lan/sound, and it crashed then as well.

  When it crashes, it puts a lot of about 1 inch horizontal lines on the screen, then starts to garble and eventually locks up.  The interesting thing is that it did it while I was playing with BIOS settings after a reboot, eliminating all Windows/Software from consideration, there was nothing loaded at all and it still started to create those lines on me.

 

nForce 3/4 has issues too, especially their ActiveArmor Firewall -- do not install those drivers unless you enjoy pain and suffering. There's also data corruption and compatibility issues with certain combos of nForce 3/4 motherboards and Maxtor SATA drives.

  This is going to be my next step, I do have an Nforce4 chipset, and I'm using a Maxtor SATA drive as my C:, so I'm going to pull that drive, and re-install onto an IDE drive that I have lying around.  I can't find any information concerning the Maxtor/Nforce problem and my particular firmware version, but it's worth checking out to be sure.  If the MB is "losing" the HD, as people have claimed from their problems with this combo, it would explain my crashes, since no C: drive means no OS.

   Again, I really appreciate the input, you all have definitely pointed out some things that I had forgotten about.  Worst case scenario is that tomorrow I take my box down to the local shop and let them test the MB/CPU/Vid Card, since I believe those are the only possible problems for me at this point, if it turns out to not be the nForce4/Maxtor problem.

[Trippy]

  When it crashes, it puts a lot of about 1 inch horizontal lines on the screen, then starts to garble and eventually locks up.  The interesting thing is that it did it while I was playing with BIOS settings after a reboot, eliminating all Windows/Software from consideration, there was nothing loaded at all and it still started to create those lines on me.

Sounds like memory (again) or you need an entirely new motherboard.

  This is going to be my next step, I do have an Nforce4 chipset, and I'm using a Maxtor SATA drive as my C:, so I'm going to pull that drive, and re-install onto an IDE drive that I have lying around.  I can't find any information concerning the Maxtor/Nforce problem and my particular firmware version, but it's worth checking out to be sure.  If the MB is "losing" the HD, as people have claimed from their problems with this combo, it would explain my crashes, since no C: drive means no OS.

Here's the official support reply from Maxtor. Note of course that they totally deny the data corruption problems are their fault.

[Flood]

  Just wanted to say thanks for all the ideas/suggestions so far, it's really appreciated.  As far as progress goes, I pulled a good PSU out of my other comp (Antec Neo HE 550), and it didn't fix the errors, as for the other poster, I'm already using a UPS, so that can be eliminated as a problem.

  I did a full format/re-install of WindowsXP, to eliminate driver issues, installed only nvidia drivers for video, and the onboard lan/sound drivers, and it crashed on me then as well.  So I disabled the onboard lan/sound, and it crashed then as well.

  When it crashes, it puts a lot of about 1 inch horizontal lines on the screen, then starts to garble and eventually locks up.  The interesting thing is that it did it while I was playing with BIOS settings after a reboot, eliminating all Windows/Software from consideration, there was nothing loaded at all and it still started to create those lines on me.

 

nForce 3/4 has issues too, especially their ActiveArmor Firewall -- do not install those drivers unless you enjoy pain and suffering. There's also data corruption and compatibility issues with certain combos of nForce 3/4 motherboards and Maxtor SATA drives.

  This is going to be my next step, I do have an Nforce4 chipset, and I'm using a Maxtor SATA drive as my C:, so I'm going to pull that drive, and re-install onto an IDE drive that I have lying around.  I can't find any information concerning the Maxtor/Nforce problem and my particular firmware version, but it's worth checking out to be sure.  If the MB is "losing" the HD, as people have claimed from their problems with this combo, it would explain my crashes, since no C: drive means no OS.

   Again, I really appreciate the input, you all have definitely pointed out some things that I had forgotten about.  Worst case scenario is that tomorrow I take my box down to the local shop and let them test the MB/CPU/Vid Card, since I believe those are the only possible problems for me at this point, if it turns out to not be the nForce4/Maxtor problem.

I'd follow up on the Maxtor info provided, and other than that... eh I'm sort of stumped.  The artifacts on your screen immediately make me think GPU again, but you'd think it would be more reproducible if that was it...  Sorry    Main thing is do NOT install Forceware IDE drivers, or the Nvidia Firewall stuff, it can really make your rig go funky.


Just throwing out a convienience thing:  Make yourself an ISO of your WinXP install, and splipstream SP2 in there.  Makes a nice handy, clean install disc.  You can Google to find the steps for your specific setup, or go look at DFI-street.com in the stickies for a good little walk-thru.  I made a WinXP/SP2 disc, and a disc that has all the recent BIOS versions for my mobo, plus Memtest and a couple other system utils. (Okay I just followed the directions the smart guys on DFI board gave, but hey.)  Makes recovering from issues or installing/reformatting a snap.


Good luck - keep us posted.

[Righ]

Did you try different data cables yet? Its very unlikely based on the symptoms, but its cheap to test, so it might be an idea before you throw big bucks around.

[Ralence]

  Finally figured out what the problem is.  Apparently there is a bad batch of eVGA 7800GT cards running around, and bad ram on the video card is the culprit.  I actually found a BIOS upgrade for the card (Since when can you upgrade the BIOS on a video card?), but that didn't help me at all.

  Spent all of about 45 seconds on the phone with a tech from eVGA (With less than a 2 minute wait, might I add), and he said "yep, bad memory, RMA it and we'll cross ship you a new one", he was describing the problem as I was explaining what was happening, he obviously has been running into this problem commonly.  In fact, their policy is to not cross ship unless you purchased directly from them, which I did not, but they are making an exception for the 7800GT issues.


  So again, thanks a ton for all the assistance here, if not for eliminating all the possibilities, I'd probably have taken forever to narrow it down, I've honestly never had this show up before, and there's no way to actually test for it as far as I know, especially without a second PCI-E card to throw in.

Phil

[Miasma]

I think that's the same problem WayAbvPar had.

[WayAbvPar]

It sounds similar. I never had any graphics issues, however- just total crashes. After replacing the PSU and case fan, along with updating the GPU drivers with some ha><><ored version that someone linked to in my thread, I haven't had any problems. Card runs between 58 and 63 degrees Celsius after a hard night of gaming. I played EVE all day long yesterday (while alt-tabbing out to watch streaming video of horse racing!) and didn't have any trouble.

[Lantyssa]

I am glad it looks like you solved it.  Let us know how the new card turns out.