Strange reboots / data losses

[velis]

I have an Asus P5K-E WiFi (specs here)
RAM is 2x1GB Kingston PC2-6400 + 2x2GB Patriot PC2-6400
CPU C2D E8400
Primary disk == Intel X25-M 80GB G2

Running on Win7 HP 64bit

The issues are all very hard to repeat, but frequent enough to cause me serious pain:
1. The SATA controller will suddenly "lose" all the attached drives in the middle of operation. The POST after the inevitable BSOD stops at drive recognition (can't detect any drive). A hard reset fixes this either in the first go or the second.

2. The primary disk (SSD) will have its data corrupted. I just reinstalled Windows and on 8th boot my account got totally borked. Had to create a new account. This is just an example, the data corruption goes many ways, but it progresses steadily towards me having to reinstall in 6 - 12 months time. The corruption is very limited though - a sector here, a sector there, the above example is the most serious problem resulting from this issue so far.

3. For some reason, the cold boot (from computer being turned off overnight) has >50% chance of BSODing and then the second go boots just fine.

The 3rd problem got so annoying that I decided to reinstall a few days ago. I thought the problem was due to problem #2, but today's boot proved me wrong (I got BSOD on the first boot).

The problems are so sporadic that I can't really even test them by removing the second pair of RAM, using a standard disk, etc., because I could be running just fine for weeks and then got a problem.
BTW: Running in IDE mode because AHCI mode would hang POST at SSD detection. This continues even afted updating SSD and MB to latest firmwares.

Any ideas what may be wrong with my setup?

Edit: There is also this issue that when I leave the computer out of power for a few days, it will invariably "forget" BIOS settings (as if I had reset the BIOS)

 

[velis]

I figured my problem was rather rare, but seriously, nobody has any thoughts on this?

 

[Blitz KriegeR]

If the board "forgets" the BIOS setting it could be power related? Is the date/time correct when this happens? If not, then the MB battery needs to be changed. If it does have the correct date/time then that idea is null/void.

BTW I also have a P5K Wifi in one of my other systems. It doesn't use an SSD, but I've never seen this type of behavior with it.

The issuses that are occuring prior to OS boot tells me it is most likely a dying board. One thing you could do is test to see if the crashes are load related. Try running something like Linpack, Prime95, or Futuremark when the system is up and see if it enduces a crash. If so, one other simple test I'd run is connect another PSU to the system. Just the basics, 24pin ATX, 4/8pin EPS, GPU, and the boot drive; nothing else. Then repeat the load test. If it now passes then it is the PSU not providing reliable power (unlikely, but possible). If it still fails I'd say go ahead and RMA the board.

Post back with any results/more info and I'll try to help some more.

Good luck mate!

 

[DaveSimmons]

Are you overclocking or running memory with custom timing?

Have you checked for loose cables?

Is the CPU temp OK? (Loose HSF)

> Edit: There is also this issue that when I leave the computer out of power for a few days, it will invariably "forget" BIOS settings (as if I had reset the BIOS)

That does sound like the battery or motherboard could be dying or dead, but the battery should only affect settings on POST, not cause devices to disappear while Winodws is running.

 

[velis]

I have now bought a new battery. Will see what effect it has on the entire setup.
I really want to fix the problem because I want to upgrade this winter when the 7xxx series gfx cards come out. I already decided I want to move to a small case like TT Armor 30 and fix that to a nice WC loop. The current comp will be sold, but I need it to work in order to do that...

 

[velis]

Edit: Turns out the statement in this post was wrong...

 

[Iron Woode]

check the caps on the board. If you see swelling or leaking, the board is toast.

 

[mindless1]

For future reference: turns out I suck at overclocking. Seems I left a voltage too low (don't know which), but now that I run at stock, all of my issues disappeared...

Had it passed running a Prime95 Torture Test for a few hours? That's always the first thing I do after changing clock settings, voltage, etc.

 

[velis]

Yes, it passed prime, prime + furmark, memtest, all for 8+ hours testing.
The proc will go to 4.6GHz, also 8+ h prime stable - i just didn't use it since it required too high voltage and tortured my poor Zalman Reserator 2.

But this was not a bad overclock regarding the processor. Seems I messed with the wrong south bridge voltage since I had issues with SATA controller.

 

[velis]

No luck. Seems I'm not that bad at overclocking after all.
The issues continue.

The problem is only that I don't know which component is the troublemaker.
I think I'll just swap the MB + CPU + RAM for an Athlon II which will be free now that I bought a new set for my HTPC.