March 8, 2008 – 4:18 PM
There’s a famous scene in the classic movie Caddyshack in which Bill Murray (playing demented greenskeeper Carl Spackler) says that in order to defeat the gopher that’s ruining his golf course, he must think like a gopher. Troubleshooting computers can be a bit like that: To recognize and fix what’s ailing your system, you must understand how it works.
For instance, when my 10-month-old laptop started to hang occasionally during boot-up before Windows could launch, I figured the problem might be caused by a cold hard drive not quite ready to leap into action, or perhaps an overstrained power supply falling short at a critical moment. But when the problem evolved into crippling Windows crashes, called the Blue Screen of Death (or BSOD), I knew it was something more.
Was it software? No, couldn’t be, I told myself. I was sometimes getting those crashes before Windows was even running. Was it the hardware? Probably, but why wasn’t the problem always the same? Before I could go much further, I had to mind the four corners of PC troubleshooting: backup, document, Google, and isolate.
Backup and Document
First, I did the smart thing. After that initial BSOD, I backed up all my data. I keep an external USB 2.0/FireWire 250GB hard drive around for just such a contingency. I love how the quick drive can get files off my ailing PC fast. And because it’s external, it puts no additional load on the system that I’m trying to troubleshoot. A fast network connection to a second PC offers similar benefits.
Confident that my work and files were protected, I soldiered on, keeping a close eye out for patterns and trends in my mysterious crash investigation.
Next, I went about documenting the problem. PC veterans know that intermittent or delayed crashes are often caused by heat. Heat can cook CPUs, lock up graphics cards, and render disk drives inoperative. Because different applications can drive components at different levels, these crashes can be frustratingly unpredictable. A PC may crash within ten minutes, or it might run all day before giving up.
Facing just such a situation, I downloaded a heat-monitoring utility. I also used a watch and a case thermometer to time how long and how hot the system would run before crashing. When the case sensor pushed past about 40 degrees Celsius, I would often get the BSOD. Now, hitting 40 degrees in my power-mad laptop isn’t that unusual. Still, I thought, maybe a specific component wasn’t getting properly cooled and was passing a critical threshold. I was officially on to something.
I wrote down everything. I tracked the times and temperatures from my tests. I jotted down cryptic error messages and memory addresses from the BSOD screen dumps. I even took digital pictures of the Windows crash screens so I could send them to my vendor’s tech support folks.
Great Google-y Moogle-y
That’s when I worked the third trick of PC sleuths: the old Web search. I Googled all those error messages and codes. Then I did the same at Yahoo Search. I used the search engines to find helpful and obscure technical forums, where discussions relating to my problem raged.
I also went to the highly useful Google Groups service and performed searches in Usenet newsgroups, which are online bulletin boards where a lot of computer-savvy people post information and discuss topics. I soon learned that failing memory chips could be causing my BSOD error code.
Which brings me to the fourth step in a good troubleshooting program: isolate. My PC has two RAM modules. After studying the documentation and checking some user forums for detailed guidance, I opened my system and removed one of the 512MB RAM DIMMs. With 512MB left in my system, I booted up. Windows launched. Applications launched. The PC ran for one full day, then two. I ripped MP3 files and compressed video. My PC was cured, but that other RAM DIMM was most certainly toast.
Just to make sure I’d isolated the problem correctly, I swapped the two 512MB modules in the PC. I couldn’t get the poor beast to boot. Ten minutes later, I described my sleuthing to the tech support guy, and immediately arranged to receive two new RAM modules to replace mine (RAM needs to be closely matched, so you often need to replace both modules even if only one is bad). As it turns out, the specific model of memory in my system was known to be temperamental. The replacement stuff has been rock-solid ever since.
Different Strokes for Different Folks
The RAM odyssey was trying, but I was fortunate. I was able to get my PC to run for significant stretches, so I could notice patterns (like temperature levels and BSOD codes) that ultimately identified the problem. What if a PC just plain won’t start? At those moments, I urge folks to check the simple stuff. Is the power cord plugged in tightly at both ends? Is the outlet powered? Don’t laugh: I had a network problem drive me bonkers for 45 minutes before I realized a blown circuit had shut down the outlet powering my router.
After that, you have to think like a gopher and wonder about things like: Is the hard disk working? If you listen closely, you may be able to hear a drive in distress (clicking, grinding, or undue vibration can all be telltale signs of a hard disk fixing to bite the dust). Remove the case and put your eyes and ears to work. An inoperative cooling fan may cause components to cook like ants under a magnifying glass.
Don’t overlook simple explanations, either. A dead PC could have been damaged by being knocked over or moved. Look closely for possible physical damage. Also power off the system and make sure all the internal and external cables, as well as internal cards, fit properly and securely. Then restart again with the case off and your eyes and ears peeled. Look closely at the boot screen and see if any weird errors–like a drive not found error–come up.
Tools of the Trade
Of course, your eyes and ears aren’t always enough to pinpoint problems. Software programs can also help you get past pernicious PC problems. Here are a few to consider:
RegWorks: This low-cost Registry editor is tuned to guide you through routine tasks and help you solve Registry issues.
Task Manager: It’s built into Windows (just right-click the taskbar and click Task Manager). Use the Processes tab to find programs that might be hogging CPU time or refusing to close, and close them out by right-clicking the app and clicking End Process. Works every time.
System Restore: It’s also built into Windows XP (go to Start, All Programs, Accessories, System Tools, System Restore). It takes a snapshot of your Windows configuration so you can roll back to a previous, working setup. System Restore is great for backing out of hardware upgrades or software installations gone horribly wrong.
X-Setup Pro: It’s tweaker madness with this low-level utility that lets you change almost everything in your Windows setup. Great for disabling misbehaving Windows components and streamlining and securing your PC.
Hey, computers may never be 100 percent reliable, but maybe they’re becoming more predictable. The reason your system acts funky is rarely mysterious–except to you. You simply need to take the proper steps and tap the World Wide Web of troubleshooting knowledge to fix what ails you.
http://www.pcworld.com/resource/printable/article/0,aid,122156,00.asp
Posted in Hardware, Software | No Comments