computerwriter.com pc_type.gif (30736 bytes)


Advanced

Home
Bio
Search
Contact
Troubleshooter
Articles
Links
Typetronics
Calendar
Books
Notes for PR Folk
About

Troubleshooter Checklist

Before you click on the link below, read this article first. You may be able to solve your own PC problem (I make no promises for Macs) by following the checklist.

Send a problem report

A new PC, running the software that was pre-installed on it, should not be producing "Illegal Operation" or "Fatal Exception" (aka the "Blue Screen of Death") errors. It should work, and if it doesn't, that's not your fault. It's a warranty issue you need to be discussing with whomever sold you the equipment - and you need to do that before the warranty runs out.

Once upon a time, many name-brand companies provided universal warranties that included not only the software they installed, but anything you did later. Now, you're lucky if the guarantee covers pre-installed software products beyond 90 days and, for anything you install, all bets are off. You're on your own to discuss it with whomever provided the products.

Those are just the rules of the game, like 'em or not. But what do you do if you can't figure out what's causing the problem in the first place? The biggest hassle most users have is diagnosing the nature of the problem. Never mind figuring out which product caused it; let's start with whether it's a software or a hardware issue. Then we can begin to track down the actual culprit.

Baselines

  • Was there a time when this phenomenon did not happen?
    • If there was, what has changed since then?
      • Did you install or re-install any hardware or software that might be causing the problem?
      • Was there an electrical storm? If you don't have at least a power conditioner, your system could have been hit by a power surge or spike that damaged one or more components
        • Hint: even if your electrical circuits are protected, is your phone line or broadband connection also protected? Spikes can take these routes, too.
      • Is it possible you opened the computer's case and touched the interior without grounding yourself against static?
        • You may not have touched hardware components per se, but if you grabbed a data cable leading to a hard drive without dumping static first, it may have affected the contents.
      • Did you change the wall plug into which the computer is connected (and/or added any heavy electrical appliance to the circuit it's on)?
        • Power surges (a little slower and not quite as damaging as a spike) can occur when the heavy draw cuts out and can wreak havoc in your system. But so can a mini-brownout caused by a heavy appliance coming on - and it may reduce the power to your system enough to corrupt data in memory just as it's being written to a drive.
     
  • If these errors have always occurred since the time you got the computer, or if none of the above applies, we can still begin to trace whether it's a hardware problem or a software problem as follows (but keep in mind what I said earlier about warranties):
Software or Hardware?
The answer to the broad question of whether your problem is caused by software or hardware hangs on repeatability.
  • If a problem is predictable (examples only, not references to specific programs): "...every time I start Word," or "...every time I try to print from Excel," or "...every time I connect to the Internet," and you can demonstrate it whenever you want, then the cause is most likely a software compatibility issue. The exception to this rule is, "...whenever the machine has been on for a while (see below)."
    • The solution is to try to figure out which software program is causing the problem, then to call for technical support and/or get an update (if such a thing exists) from the publisher's Web site -- or to have it sent to you if you cannot connect to the Internet.
    • You get the clues about which program is causing the problem by clicking the "Details" button in the Illegal Operation error message, or checking the details on the Blue Screen. In both cases, the offending program and the offended program are identified. See the resources listed at the end of this article for where to go and what to do next.
    • The most common cause of Illegal Operation errors is defined as a page fault. It is most likely to occur when one or a combination of conditions exist. If you're low on electronic memory resources, Windows doesn't stop you from loading programs. Instead, it uses blank space on your hard drive as a temporary parking area for code it can't fit into memory. Aside from slowing your system down (the hard drive is orders of magnitude slower than your electronic RAM), this hard drive space is usually limited - either because of a limit imposed by you or because you're also short of hard drive space. And the system works on a first in, first out basis. As new code is added to the cache, older code may be pushed out. When the program whose code is on the hard drive looks for it in memory, it first complains to Windows that it's not in memory. If the code is available, Windows swaps it back into memory and you never notice. If the code has been pushed out of the cache, however, you get the page fault error and the program goes to heaven.
    • You may also be a victim of what is affectionately known as "DLL Hell." It's a problem that crops up when you're either installing software that insists on replacing core Windows dynamic link libraries (DLL files) with its own versions, or when you upgrade a product with a slightly changed installation process (or one that isn't efficient).
      • It's possible that the newer DLL isn't well-written, or that Windows is using an older version stored in one of its default folders (\Windows, \Windows\System or \Windows\System32).
      • This is where Norton CleanSweep or a shareware program such as Duplicate File Finder or Space Hound 32 will come in handy to find duplicate DLL files.
        • Keep in mind that it isn't the dates of the files that are important here; it's the version numbers you're interested in.
        • And, here's another hint. Until you get things straightened out, don't delete the older files. Rename them instead (example: suspect.dll to suspect.d~l).
     
  • If a problem occurs completely at random, with no discernible pattern or predictability, and you cannot duplicate it at will, then the cause is most likely hardware. An exception to this rule occurs the problem identified is a "page fault" (see above).
    • The most likely cause is a flaky memory module, but it could, if the problem only presents when the system has been running for a while, indicate what is known as a "cold solder join" somewhere on the motherboard.
    • One of the first things to do is to invoke Device Manager (right click on My Computer, then select Properties), to see if any devices are flagged by a yellow circle with an exclamation mark in it. Double-clicking on the device listing (or right-clicking, then selecting properties) will bring up a dialogue that reports on the device. See the article reference below to find out what the error codes mean.
When problems collide
Problems can become confusing when they're caused by a hardware device, but not by the device itself. Modems, printers, graphics controllers, sound controllers, and other devices have small pieces of software called "drivers" that translate what Windows is sending into instructions they can understand. Sometimes these drivers have compatibility issues with other software (and some have specific issues with Win98SE, WinMe, NT, or 2000). Once again, the solution is to see if the manufacturer has an update at the company Web site, then to try to get it. If you cannot access the Internet because of the fault, you may have to enlist a friend with a connection to help.

If your problem is the Blue Screen, copy down the error number (0D, 0E, - that's zeroE or zeroD - etc.), then go to Microsoft's support site and search the online Knowledge Base for this error number. Essentially, you'll get the same advice as above, but some of the error messages are quite specific to certain products and conditions. You can ignore most of the numbers on the screen, but the information you need is in this example: "A Fatal Exception 0E Has Occurred at 0028:C02A0201..."

Other resources at the Microsoft Support site that may come in handy include:

Back to top

Send a problem report

Contacting me
Copyright © 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003  Myles White. All rights reserved.
Revised: December 20, 2002 .