Software to help troubleshoot Windows crashes?

September 10, 2007 6:38 PM

I have a hand-me-down Intel server running Windows XP Pro... most of the time. Once or twice a day it locks up completely and without warning. Connected peripherals go dark but the screen stays frozen. I'm not a hardware guy; is there some free/cheap software to monitor the hardware & OS and tell me what's causing this?
posted by Tubes to computers & internet (8 comments total)

#1 cause for lockups like that is heat. Open it up and look for dust. Try running it with the case open to see if the problem goes away.
posted by Argyle at 6:40 PM on September 10, 2007

Assuming you've throughly examined your machine for viruses and malware, and patched it to current levels, Step One is to examine the Events Log for clues. Step Two is to check for thermal problems, as Argyle suggests, but also listen/look for power supply fan, processor fan and case fan functions, and to make sure that nothing external has blocked air flow through the case (I've seen people stack papers on a computer case, such that the main power supply fan got blocked by hang over, and they suddenly had problems). Step Three might be to download and install SpeedFan or a similar app, to monitor your hardware. Step Four is to consider that bad RAM is the #2 reason after heat for unexplained Windows XP lock ups. Substituting/replacing RAM DIMMS, or even moving them from one slot to another, is required to troubleshoot some problems, but you can begin diagnosing RAM problem with MemTest86.
posted by paulsc at 7:05 PM on September 10, 2007

I second overheating as a likely cause. Check the fans, the case, etc. There are a bevy of free temperature monitoring apps for windows systems, but I can't recommend one in particular, since I'm a linux guy.

Good luck.
posted by chrisamiller at 7:13 PM on September 10, 2007

Check for heat -- BIOS may have a hardware monitor that lets you see the temp your CPU is running at. Generally anything over 45C is suspect, but depends on the CPU.

If that doesn't help, make sure you have the latest version of all your device drivers. Video card and hardware I/O (such as disk drive controllers, etc) are the usual suspects.

You can also try flashing your BIOS to the latest version available for your board (but this can also hose your system completely, so follow the instructions and make backups).

If you still can't get it working, try taking pieces out one at a time and replacing them with other pieces. Swap out the video first (or switch from add-in to on-board video if you have that option).

Also I second checking the event log for clues. If it's a blue screen, you can turn on full debug dumps and use the Microsoft Debugging Tools for Windows (download from microsoft.com, use google to find) to look at the dump and see if you can see what device driver it dies in...

USB has some issues too. Try turning off USB in BIOS if you can.
posted by jeffamaphone at 7:27 PM on September 10, 2007

From most to least likely, for an older machine: Overheating due to dust buildup in the CPU heatsink fins and/or failed fans, failing power supply, malware, bad RAM, Capacitor Plague.
posted by flabdablet at 8:20 PM on September 10, 2007

Dust is excellent insulation; it traps heat in electronic components. Get a can of compressed air from the camera store. Open up the computer case. Vacuum hose in one hand, can of air in the other, spraying in very short bursts (if you spray too long, you'll get propellant liquid on stuff, but fear not). Do this every 3 months or so.
posted by neuron at 12:50 PM on September 11, 2007

Dust-as-insulation is less troublesome than dust-as-total-airway-clogger, in my experience, especially for old machines. A finned heatsink simply can't work if the air from the fan never makes it as far as the fins because a plug of dust is in the way.

Quickest and easiest way to fix this is take the cover off the case, take the machine outdoors, hold it up above your head, tilt your head back, shut your eyes, and blow as hard as you can into the CPU fan area, then take a couple of steps upwind before opening your eyes or breathing in again.

You get absolutely covered in crap doing this. Impresses the customers no end when you come back inside :-)
posted by flabdablet at 5:02 PM on September 11, 2007

Thanks all. I've discovered that my Supermicro motherboard came with "Supero Doctor III" software, which doesn't log anything but at least tells me in real time that my system seems to be running a bit warm (Dual Xeon 2.8... CPU1:48C, CPU2:42C, System:51C).

So, upon the next crash I will unplug everything, pull the box, and give it a good de-dusting before moving on to BIOS and driver updates.

I appreciate the help; it's given me some paths to follow.
posted by Tubes at 10:04 PM on September 11, 2007