Random crashes and shutdowns.

Hi, I have what seems to be a hardware problem, and I can't decide which part is at fault (I have ram and psu from corsair).

 

I've had this system for soon two years (corsair ram and a third harddisk were added a few months ago) and its been running fine till about a month ago when two problems started appearing intermittently:

 

1. "Memory problem": I get a BSOD with some random error, and if I run memtest upon reboot, memory errors show up within a few seconds, and keep appearing in hundreds and more if I let it continue running. Though this only happens if the computer is restarted without powering down after the blue screen. If it is powered down even for 5 seconds, it will run memtest error-free, even for many hours (longest I've tried was about 20 hours), and windows etc will work normally for a period of time ranging from 30 minutes to a couple days. Then the whole thing repeats itself. It also doesn't seem to be stress related, I could play games for hours, run various torture test apps such as prime95, and then it would crash while idling, like when I'm afk or just reading forums.

 

2. "Shutdown problem": Another problem is that sometimes the pc "partially" shuts itself down, which means that the signal to the monitor is cut, the sound disappears and it doesnt react to any keypresses, however it doesn't power down, all the fans keep spinning, external devices receive power etc. The only way to get control over it is reboot (by resetting or powering down and up again). Unlike the "memory problem" it does not produce memtest errors, if ran immediately after reset, so windows can boot right after restart, and so far it only happened during gaming, never when pc was idle. However it hasn't happened during stress testing either.

 

Both of these problems started appearing in the same period (about a month ago).

 

 

Things I've done so far (in no particular order):

 

- Reflash bios to latest version, interestingly this made the memory problem go away for about a week, however the shutdown problem appeared more often then. Or maybe it was because the pc could stay on long enough for it to appear? The memory problem came back again about a week ago though. So the net effect is - no change

 

- Remove all overclocking, set bios settings to default, except for ram settings which are set as specified (5-5-5-15, 2.10v, 1066MHz) - no change

 

- Relax memory timings, put frequency to 800MHz, also underclock cpu slightly - no change

 

- Run each ram stick by itself, and in different slots - no change

 

- Run ram from different manufacturer (2x1gb ******** ballistics), but theres no guarantee it not broken too - no change

 

- Remove all usb devices, disconnect one hdd and cdrom, in case psu was dying and couldnt give enough power or something - no change

 

- Check temps - they are low (cpu cores at 30C when idle, 45-50 with prime95 running, everything else also well within spec). Since the mb doesn't report NB temp, I just felt it with my hand and it only feels warm, there also a small fan on the NB heatsink. Also my case has extremely good ventilation, so heat is unlikely to be the problem here, unless it some random IC thats for some reason overheating when it shouldn't.

 

- Check psu voltages with voltmeter:

5v - 4.94

12v - 12.04 on 8981 connector, 11.96 on atx and pci-e., is the difference normal?

-12v - 12.36

5vsb - 5.00

3.3 - 3.36

 

- Run prime85, furmark 1.6.5, everest stability test set to stress hard drives, and also burn a cd - all at once, mostly for fun but also to pull a lot of power from the psu to see if it was at fault. Also plugged in all the usb devices. A couple of hours of this hasn't crashed to pc though. Too bad I don't have a usb cup warmer or something.

 

- Most of this was done in WinXP 32bit, I also have 64bit Vista in dual boot, but it behaved the same there, so its obviously isn't software related.

 

 

In short, I have a hardware problem, and I suspect either the mobo or psu, however since I can't even reliably reproduce the crashes, I can't pin down the problem. Any insight or similar experiences with solutions are welcome.

How likely is it that the graphics card is causing memtest errors? :roll:

 

Also, I suspect it could be the psu (I have a corsair psu), what is your take on it?

maybe take a look at your event logs, see if there is a background application that causes the crash in windows.

check your battery on the MB see if it still gives out 3.3v outside of the case.

you can also try reseating all the PSU connections.

theres no extra 4 or 8 pin connector on the MB for the supplemental power input that may have been overlooked?

maybe take a look at your event logs, see if there is a background application that causes the crash in windows.

Event viewer has no errors, windows minidumps always show some random stuff. Not so relevant though, because whatever is causing the crash in windows cannot produce errors in memtest ran from boot-cd.

check your battery on the MB see if it still gives out 3.3v outside of the case.

Checked that, battery is fine.

you can also try reseating all the PSU connections.

theres no extra 4 or 8 pin connector on the MB for the supplemental power input that may have been overlooked?

Done that, and no, tbh I doubt the pc would even post if that was the case.

 

I also spoke with asus support and they kinda blame corsair because the ram I use doesn't appear in their QVL. However, it is listed as compatible on corsair website.

 

Moreover, I just plugged in a stick of VS1GB667D2, which I kept since a while ago for just in case, and it actually seems to be stable now, neither of the problems I described happened in the last 3 days.

 

So, what do you think, ram problem or something else?

I also doubt it is the PSU, so I'm not going to send it anywhere just yet. I ran a single stick of VS1GB667D2 and its been stable for a week, now I removed that and put the dominators back in, but atm I also run them at 667mhz, and it is stable. I need some time to be sure the problems don't happen with these settings, so I'll get back to you in a week or so.

Well it seems it didn't take long until the computer had enough of the dominator sticks, and started crashing again. Now this happened with them running at only 667mhz, all other settings as specified (5-5-5-15, 2.10v). It seems that only the single corsair value select 667mhz stick can run stable. So lets sum it up again:

 

1. When I run it with either both or one dominator sticks, at pretty much any reasonable settings, it fails. The problems are intermittent though, I have no clue how to reproduce them in controlled conditions, or say with certainty that they have disappeared. All I can do is wait till it happens. It usually does so about once every day.

2. Same happens with competitor sticks.

3. A single corsair VS1GB667D2 runs fine though.

 

Is it screwed up mobo, or I'm just being unlucky so that different ram modules from different manufacturers died twice? Or maybe my mobo has turned ebil and is killing the ram, and its only the matter of time before the VS1GB667D2 dies too?