Re: OT-Computer Reboots at Random (LONG)
JohnKJordan in East TN
>Bummer. Check the Event Viewer under System Tools to see if there are any hints there just before the crash. (I don't use XP home, but the Event Viewer should be in there somewhere.)
It's hard to guess without having the machine in hand, but I agree that it is very likely a hardware problem. If so, if you go to all the trouble of reformatting and manage somehow to reinstall the OS you will still have the same problem. That you are not getting a blue screen error message is typical for some problems - remember, to display an error message means the system has to be stable enough to catch and pass an error condition to a low level software routine to display the error text on the screen.
A typical problem is with bad memory. On high-end graphics machines, we often ran into hardware stability issues with mismatched memory - someone would install simms from two different sources and they sometimes would not play well together.
I'm not familiar with the Microsoft memory test you mentioned - we always used and recommended running Memtest86. You have to create a boot disk to run this one and it is doesn't miss much. It's a free/public domain program but I see one web site selling a bootable CD (http://www.memtest86.com/). I also found a site still providing free downloads (http://www.memtest.org/). The binary version can be copied to a bootable floppy or CD; the ISO version can be dumped to a CD to create the bootable CD. If the Microsoft memory diagnostic works the same way, I'll bet a nickle they just packaged the open source memtest86 with the MS name!
If you have the resources, another common memory test is to temporarily replace the memory with known good modules from another machine. Another old memory trick is to remove modules one at a time - if a problem is with one the system may run with it missing. This is not possible on some motherboards (they need to work in pairs), but occasionally swapping the modules will be revealing if the problem goes away or the symptoms change.
As also mentioned, a faulty power supply is a common cause and difficult to catch in the act even with a good scope. Fortunately, you can replace the power supply inexpensively and see if that fixes things. Power supplies are pretty good these days, but if the power is iffy in your area (voltage transients) a UPS unit can do wonders.
A display card (or other card) can also cause reboot problems. The diagnostic routine here is to simplify the system as much as possible (remove all unnecessary cards), then temporarily substitute a different display card. Unfortunately, this is easier said than done unless you have a stack of hardware on the shelf or other machines to rob.
One simple thing I've done which has fixed hardware problems like this - reseat all connectors, cards, and chips in the system. Connectors can fail over time from oxidation/corrosion on contacts. Reseating simply involves removing and reinstalling each component. This action tends to scrape away any oxidation. For example, pull out the power supply connector and push it back in. Remove each card and reinsert it. Unclip each memory module (carefully) and reinstall. If there are any chips in sockets, push them down with your finger or thumb (they can work loose with heat cycles) - if they don't push down a bit, I pry them up slightly and push them back. I can't tell you how many computers I've fixed over the years with my finger!
Unfortunately, there are other motherboard problems that are harder to diagnose and typically repaired by replacing the motherboard. These days it is often cheaper to replace the computer. If you have to go that route, since your HD directory system appears fine, you can likely remove and temporarily connect your old hard disk to a new computer and copy much of your existing data.
And as for backup, what used to be so difficult in the past (with piles of floppies, CDs, or ack, tapes, is easy now. Buy an inexpensive USB or firewire external HD and copy everything you don't want to lose. I was so paranoid when I used these computers to make a living that I rotated between seven such drives. After I retired, I cut back to three! :)
BTW, since starting with computers in the mid 70s, I've developed some of my own theories about how to manage computers. One of my hardware rules is to never shut off a machine. Each time it heats up and cools down, things can be physically stressed on cards and even within components. Along with that, I make sure I have sufficient air flow - I'd add fans as necessary moving air across drives and components (my tower case 3D graphics/video machines might have every slot full, fully loaded with memory, and up to eight hot drives). And as mentioned, always power from a UPS.
JKJ