RAM or HDD Issue?

[chywork]

I'm at my wits end. Here's the lowdown.

I built a new matx machine. It has an 8800gt, 2 gigs of Ballisitix (DDR800) RAM, a maxtor 320 gig drive, a G0 Q6600 quad2core, corsair 620W power supply on a gigabyte board. I had some blue screens of death and eventually the computer won't boot at all saying that a config file was missing or corrupt.

I reformatted the drive (not quick method) and installed windows again (XP Pro SP3), installed just the basic drivers, Firefox, and Nvidia drivers. I kept the computer on for over 24 hours. Then I received this message saying that files in system32 are corrupt use chkdsk. I rebooted and used memtest+ and I got a bunch (hundreds) of errors. I turned off the machine waited a couple of hours and did memtest+ again, no errors this time.

I'm baffled. Is this a HDD or RAM issue? Can anyone explain?

 

[imported_BadBlock]

My first suggestion would be Memtest86+ but it sounds like you have already tried that. Depending on how long you let it run you may want to let it do several passes, like run it when you are going out for a while or going to sleep. I have seen a few instances where memory would be ok on the first pass but then error on the second or third pass if there were inappropriate timings set in the bios.

If you want to test the hard drive try using seatools (you want the dos boot cd, not the windows version). Do the short test, if it passes that then do the long test which could take an hour or two depending on the drive.

If both of those come back negative I would start looking at what you are installing immediately after the os install, if anything.

 

[RebateMonger]

Large changes in the results of Memtest86+ are pretty strange. I've never seen it or heard of it. One possibility in your case could be overheating of the memory modules. If you are overclocking, then set everything back to the default values and try again. From what you describe, it appears to be a memory problem rather than a disk problem.

 

[imported_BadBlock]

I don't think it's common, I've memtested several dozen problem computers over the years and only seen it happen a couple times, and it very well could have been temperature related. A marginal problem that was only brought out when it got above a certain temp.

Main reason I wanted to encourage chywork to run memtest longer is like you said, it sounds a lot like a memory issue.

 

[chywork]

Overheating is one possibility that I haven't really investigated too much. I should have mentioned that I've using a silverstone SG01 case and bought the passive CPU cooler that is suppose to fit inside. The system is cooled by drawing the hot air through the power supply fan. Maybe there isn't enough air circulation. Part of issue might also reside that this RAM needs 2.2v instead of the normal 1.8v

 

[unholy414]

Originally posted by: chywork
Overheating is one possibility that I haven't really investigated too much. I should have mentioned that I've using a silverstone SG01 case and bought the passive CPU cooler that is suppose to fit inside. The system is cooled by drawing the hot air through the power supply fan. Maybe there isn't enough air circulation. Part of issue might also reside that this RAM needs 2.2v instead of the normal 1.8v

I take it your BIOS doesn't go up to 2.2? If it doesn't, get it as close as you can, then retest.

 

[RebateMonger]

Originally posted by: chywork
Part of issue might also reside that this RAM needs 2.2v instead of the normal 1.8v

Power dissipation (heat) with the higher voltage will be quite a bit more than at 1.8v.

 

[chywork]

Thanks for all the help. I'll let people know if it works. (The board does support 2.2v)

 

[chywork]

Hey, I'm pretty confident now that heat dissipation was the reason for the memory errors. I purchased a 80mm fan and put it in my sff case and had it blowing over the ram. Not a single issue since - 4 straight days of running at 2.2v. The only issue now is that I have to buy a quieter fan. Thanks for everyone's the help.