PC Random Crash - what else to check?

PhilipM · 20 May 2011 at 04:41

I recently upgraded my main rig from 2->8GB RAM and an extra Hybrid HDD. I memtested the memory overnight with no air con (so ambient up to 30C) and it was rock solid. Recently, I've been experiencing random lock up. It doesn't happen that frequently, but when it does, it's typically a few minutes after startup. Once it happens the first time, reboots tend to see the same error after a shorter interval. Leaving the PC for a while seems to make it go away.

My gut feel is the PSU , so I installed voltage monitoring software, which reported the 12v rail as low, but a multimeter check found no such problem.

Is there anything else I should check or try before I go out and buy a new PSU in desperation?

Thanks,
Philip

olivier renault · 20 May 2011 at 08:47

I'd try without the MomentusXT connected. It's possible the drive is not working correctly with your system.

If it's the drive that's locking up, maybe research on seagate diagnostic software and firmware for the drive. My feeling is that it's something to do with it.

What PSU you have with your i5 and 4870?

PhilipM · 20 May 2011 at 09:05

I hadn't thought of the drive causing an issue. Seagate have a pretty comprehensive self diagnostic tool. I'll download that tonight and give it a shot.

I'm using a Silverstone 500W MODULAR ST50F-230 PSU. The complete system it is driving is:
CPU : Intel Core i5 750 2.66Ghz (Lynnfield) (Socket LGA1156)
Graphics : XFX ATI Radeon HD 4870 512MB GDDR5 PCI-Express Graphics Card
Fans : Xilence XPF80R Red Wing 80mm Quiet Fan x2
Memory : 2x4GB DDR3 at standard speed
CPU Cooler : Silverstone NT06E (passive - uses PSU fan to draw air)
HDD : 1TB SAMSUNG HD103SJ SATA F3
Optical : SONY AD-7241S-0B 24XDVDRW OEM + 500GB Momemtus XT
WIFI : ABIT AIRPACE PCI-E WIFI CARD

Using http://extreme.outervision.com/PSUEngine and allowing for 25% capacator aging (as my PSU is about 18m old) this gives a minimum of 382W and a recommended of 432W. Actually, this is a bit more than I was expecting!

olivier renault · 20 May 2011 at 09:22

according to this, it's got 384 W peak.

It seems to be ok powering the rig, especially at idle, but at load it must be huffing and puffing. At idle and after system restart, it should cope no problems.

I know my system occasionally locks up for a few seconds as one of my hard drive spins up. Not a problem, but it's annoying. The Momentus is a strange beast, as it learns to cache files on the SSD portion. Could be something to do with the lock up after startup.

If the drive checks out and problems still occurs with the drive off, then could very well be the PSU.

Rockbottom · 20 May 2011 at 12:34

Honestly I suspect the PSU, in my experience all sorts of random/intermittent problems can be caused by the PSU especially after adding new hardware, do you have access to a different one you can try (a mates)?

PhilipM · 21 May 2011 at 08:33

More interesting results, but again nothing definitive!

I installed seatools last night. The PC started to crash, but I persevered and it passed all the basic tests (too unstable to run long tests).

I ran the windows memory checker for two full cycles, no crash and no errors.

I ran chkdsk at it's most detailed settings. It ran through, taking about an hour and a half with no errors.

Since the chkdsk, the system has been completely stable! I'm wondering if perhaps there was corruption on the cache (momentus XT is a hybrid drive). I know it's very aggressive with caching, so perhaps the chkdsk overwrote the bad data with new stuff.

Unfortunately I don't have a spare PSU to try. So, I'll wait for the next bout of instability and try chkdsk again to see if the result is repeatable.

olivier renault · 21 May 2011 at 10:12

I would talk to Seagate about that. It's possible they have a firmware for the drive to fix some issues. Or that the drive is faulty.

Could also be that the PSU struggling would make the drive malfunction, but I would go with a firmware check first. Maybe possible to RMA the thing.

PhilipM · 23 May 2011 at 02:46

I've now completed the long test using Seagate's own software, the drive is working correctly.

I have already upgraded to the latest firmware - that's worth doing out of the box as the drive spins down too quickly otherwise (to save power in laptops).

It's been stable the last few days!

PhilipM · 7 Jun 2011 at 09:43

It seems that if it's going to hang, it's within 20 minutes of startup typically. I've been running it continuously for 3-4 days at a time, including reboots. Once it's stable, it remains so.

I have also ruled out the new memory. I swapped back to the old memory and it experienced the same issue.

Given this odd timing quirk, I'm inclined to think it's something thermal. My suspicion is back to the PSU, perhaps with the standby power generating heat in one component which then overheats due to the power demands at startup.

Does that sound plausible? Any ideas on how to prove the theory (short of going out and buying a new PSU)?