I hate it when that happens. It’s one of the most annoying things a computer user can face. You’re working along and all of a sudden *poof* – everything’s gone. Be it a random reboot, or the infamous “blue screen of death”, it’s very,
The problem is that this can be caused by so many things, it’s difficult to nail it down quickly.
Let’s run through some of the possibilities, and what I’d do.
You’ve touched on one thing I always look at first: viruses and spyware. Do make sure that your scanners are running, and are being regularly updated with the latest databases of spyware and virus definitions. For anti-spyware tools, that should happen daily – anti-spyware tools typically do so less often.
If you’re running Windows 9x (meaning Windows 95, 98 or Me), there’s about a 50/50 chance that the problem is a software problem. Because of how their design evolved, those versions of Windows were all somewhat more vulnerable to crashing bugs in the software. A poorly written device driver, or even an application bug, could in the worst case scenario cause a system reboot or blue screen.
If you’re running Windows XP (or Windows NT, 2000 or 2003), the system is designed more robustly – meaning that it’s more difficult for these types of problems to result in a random reboot or crash. Not impossible, just much less likely.
If you’ve made a recent change to your system, perhaps installed a new software package, or a new piece of hardware, and these problems started happening thereafter, that’s a likely clue. Depending on the software or hardware, my first reaction would be to look to the vendor for reported problems and possible updates. In particular, device drivers are the most likely to cause crashes and reboots without warning.
As I said, software related reboots and blue screens have become more rare under Windows XP. However you should still make sure that your system is as up-to-date as possible, particularly including drivers for recently installed hardware.
I know that many folks have become suspicious of Windows Automatic Update for various reasons, but I’ve not heard of any crashes resulting its use. I still recommend it as the best way to keep Windows up to date.
I will say that if your computer is still under warranty, you need to talk to the manufacturer first. You should exhaust all your options in getting them to repair a failing machine before you head out on your own and possibly void your warranty.
Naturally, if you’ve recently installed new hardware, that’s a possible clue. You might consider removing it temporarily to see if the problem abates. If it does, it points to either the device itself being the cause, or perhaps the system power supply, as I’ll discuss in just a minute.
If your computer has been running fine for some time, and you haven’t installed any new hardware or software recently, then my tendency would be to start suspecting various hardware components.
Perhaps the most common are failing fans. The fans that move air through your machine to keep it cool are critical to its operation, and are often the first to fail – either due to accumulated dust and dirt, or simple age. When the fan stops working, the machine overheats, and when the machine overheats – it crashes. Randomly.
Next most common is a failing power supply. Power supplies can fail slowly – meaning that they can become ‘marginal’ before they fail completely. And the symptoms of a marginal power supply are – you guessed it – random crashes. This is one of those cases where replacing the power supply (or having someone replace it for you) is often an inexpensive test. Particularly if you’ve added more hardware to your system over time, you may simply be demanding more of the power supply than it was designed to provide, so an upgrade might well be in order as well.
I have this gut feeling that memory is failing just a little more frequently than in the past. I won’t speculate as to why, or even if my observation is accurate. The good news is that there are tools specifically designed for testing memory. Memtest86 is one such tool that performs an exhaustive test of your computer’s memory. Microsoft also provides a Windows memory Diagnostic. Both tools run from a bootable floppy or CD (Windows cannot be running for the tests to exercise all memory), and both tools are free.
Naturally, it’s also possible that the fault lies elsewhere. Your motherboard, an add-in card, even your disk drives or video card. Which leads to the ultimate predicament … it could be anything. Quite literally.
One of the more common repair techniques is to make an educated guess at what might be wrong, replace that component, and keep repeating until all the components of your computer have been replaced, or the problem goes away.
Unfortunately, doing that is beyond the resources or desire of most computer owners.
So here’s what I would do, when faced with a randomly rebooting computer:
- Yes, I would consider upgrading to Windows XP, if the machine is capable of it.
- Make sure that anti-virus and anti-spyware utilities are running and up-to-date.
- Make sure that the operating system and all device drivers are as up-to-date as possible.
- Run a memory diagnostic such as Memtest86 or Windows memory Diagnostic.
- Run a hard disk diagnostic such as SpinRite.
- Run a motherboard temperature monitoring tool such as Motherboard Monitor – it’s a free tool that will report your CPU’s temperature among other things, and will let you see if the machine is overheating for some reason.
At this point I’ve done pretty much everything I can that doesn’t involve opening the computer. If the problem isn’t evident or resolved, we need to get a little more serious. This might also be the time for some to simply take their computer in to a technician for diagnosis.
Next, I’d open up the computer and:
- Carefully vacuum all the dust out of the machine.
- Make sure that the fans which are accessible are running properly. If not, I replace them. If the machine doesn’t crash as quickly with the cover off, that’s often a sign of overheating.
- Remove as many optional hardware components as possible that would still allow the machine to run. If the problem disappeared, I would re-install components until it reappeared, and then remove other components to make sure that the problem was associated with only a specific component.
- Re-seat all remaining and accessible connectors and expansion cards – sometimes problems are as simple as a loose connection.
At this point we’ve done pretty much everything we can with what we have on hand. Next up, we start spending money (or, perhaps if you’re a geek, pulling from your spare parts bin), and go down the “replace parts until it works” path. This is another jump off point for many – it’s definitely easier to simply take the computer in to a technician for diagnosis.
- I’d replace the power supply first. Unless there’s other data that says the problem is likely to be elsewhere, I’m just playing the odds here. If I went this far, and I planned to keep the computer for some time, I’d also consider upgrading to a higher wattage supply at the same time. Replacing a power supply is only moderately difficult.
- Next up, would be the motherboard. This is a bit of work, as it often involves tearing the entire computer apart.
- Lastly, I’d consider replacing the computer.
In reality, unless you’re really interested in playing with the hardware and trying the “replace it ’til it works” approach, I’d recommend skipping this last set of items completely and taking it into a repair shop to let them figure it out.
And, naturally, before you do so, it might also be time for a cost/benefit analysis: will it be cheaper or more effective to simply replace the computer than to fix it? I’m not at all saying that it will be – it depends on the availability and going rates for computer repair in your area, and the potential cost of fixing whatever is broken. But this is the time to at least do the math and compare.