Troubleshooting Methods Overview

To become a successful troubleshooter, you need to

• Learn as much as possible during the client interview.

• Evaluate the client's environment .

• Use testing and reporting software to gather information about the system.

• Form a hypothesis (a theory you will try to prove or disprove).

• Use the troubleshooting cycle to isolate and solve the problem.

tip

Remember, troubleshooting is the art and science of quickly and accurately determining what is wrong with a customer's system. Troubleshooting is an art because every technician will bring his or her own experience and personality to the task. Troubleshooting is also a science because you can apply a definite method that will bring you a great degree of success.

The First Step ”The Client Interview

The client interview is the all-important first step in solving any computer troubleshooting situation. During this interview, you need to determine the following facts:

• The software in use at the time of the problem

• The hardware in use at the time of the problem

• The task the customer was trying to perform at the time of the problem

• The environment in the office or work area at the time of the problem

• If new software or hardware has been added to the computer or network

• If other users are having the same or similar problems

The number-one question you're trying to answer is, "What changed since the last time it worked?" I learned this question years ago from the writings of longtime Byte magazine columnist Jerry Pournelle (check out his latest work at www.jerrypournelle.com), and it's been endlessly helpful over the years.

During the client interview, you need to ask questions to determine the following information:

• What hardware or software appears to have a problem? ” The user might have an opinion about this, but don't be unduly swayed by a statement such as "the printer's broken"; the device or software the user believes to be at fault might simply reflect a problem coming from another source.

• What other hardware or software was in use at the time of the problem? ” The user probably will answer these types of questions in terms of open applications, but you will also want to look at the taskbar and system tray in Windows for other programs or routines that are running. Pressing Ctrl+Alt+Del will bring up a task list in Windows that has the most complete information about programs and subroutines in memory. To determine the exact version of a Windows -based program in use, click Help, About. View the System properties sheet to determine the version of Windows in use.

• What task was the user trying to perform at the time of the problem? ” Ask the questions needed to find out the specific issues involved. For example, "Printing" isn't a sufficient answer. "Printing a five-page brochure from PageMaker to a laser printer" is better, but you'll probably want the user to re-create the situation in an attempt to get all the information you need. Don't forget to check the Event Viewer in Windows NT/2000/XP for details about the software running at the time of the error as described in Chapter 18.

• Is the hardware or software on the user's machine or accessed over the network? ” If the network was involved, check with the network administrator to see if the network is currently working properly. If the hardware and software are not networked, your scope for troubleshooting is simpler.

• What were the specific symptoms of the problem? ” Some users are very observant, but others might not be able to give you much help. Ask about the approximate time of the failure and about error messages, beeps, and unusual noises. If the user is running Windows 9x/Me, find out how long he has been using the computer since the last reboot. These versions of Windows can run out of resources when used with many open windows or over a long period without being rebooted.

• Can the problem be reproduced? ” Reproducible problems are easier to find than those that mysteriously "heal" themselves when you show up. Because power and environmental issues at the customer's site can cause computer problems, try to reproduce the problem at the customer's site before you move the computer to your test bench, where conditions are different.

• Does the problem repeat itself with a different combination of hardware and software, or does the problem go away when another combination of hardware and software is used? ” For example, if the user can print from Microsoft Word but not from PageMaker, this means that the printer is working, but there might be a problem with configuration or data types used by different applications. If the user can't print anything, there might be a general problem with the printer hardware or drivers.

Sometimes, the client interview alone will reveal the answer. More often, however, you'll need to go to the client's work area and evaluate the hardware and software that are involved.

How to Evaluate the Client's Environment

Depending on the clues you receive in the initial interview, you should go to the client's work area prepared to perform a variety of tests. You must look for three major issues when evaluating the customer's environment:

• Power issues

• Interference sources

• Symptoms and error codes ”this might require that you try to reproduce the error

You can select from the tests listed in Table 22.1 based on your evaluation of the most likely sources of problems. You might need to perform several tests to rule out certain problems.

Table 22.1. Troubleshooting Tests and Requirements

Test	Requires
Power	Multimeter, circuit tester
BIOS beep and error codes	List of BIOS codes, POST card
Printer self-test	Printer and paper
Windows bootlog	Start Windows with Bootlog option enabled
I/O port	Connect Loopback plugs and run third- Loopback tests party diagnostics
Video tests	Run third-party diagnostics
Hardware resources	Windows Device Manager
Device drivers	Windows Device Manager, Autoexec.bat, Config.sys (Windows 9x)

For more information about the requirements listed in Table 22.1, see "Useful Hardware and Software Tools" later in this chapter.

Testing Power

Systems that won't start or that have lockups or shutdowns with no error messages could be the victims of power problems. To determine whether power problems are located inside the computer or are coming from outside the system, use the tests and tools described in Chapter 5, "Power Supplies and Circuit Testing." If a system malfunctions at a customer site but works properly at your test bench, power problems due to improper wiring might be to blame.

Looking for Sources of Interference

Power problems also can be caused by interference from other devices, such as copiers, vacuum cleaners, elevators, and alarm systems. If a system performs properly when moved away from its normal work area, but malfunctions when it is returned to its normal location, or if it works during the business day but not after hours (when an alarm system is activated), interference might be to blame.

See "Power Conditioning Devices," p. 175 , for suggestions on dealing with sources of interference.

If the problem is network- related , it might be necessary to reroute UTP ( unshielded twisted-pair) cabling away from interference sources or connect the cable to a different port on the hub or switch.

Recording Symptoms and Error Codes

If your tests rule out power and interference, you must proceed to tests that focus on the hardware or software that appears the most likely cause of the problem.

Which test or diagnostic routine is the best one to start with? Before you perform any specific tests, review the clues you gathered from the client. Here's an example: Microsoft Word for Windows would print, but PageMaker would not. If these were old MS-DOS “based programs, you might suspect a problem with the PageMaker printer driver because MS-DOS programs control the printer directly.

However, all Windows-based programs use the same Windows printer driver. Printer hardware or driver failures would prevent all software programs from printing; however, in this case, printing works from some programs but not others when the same printer and printer drivers are in use. Before you can solve this problem, you need more information about the printer. It's time to use the printer's self-test, (a technique listed in Table 22.1), for more information about the printer.

A laser printer's self-test usually indicates the amount of RAM on board, the emulation (HP or PostScript), and firmware revisions. The amount of RAM on board is critical, because as you learned in Chapter 10, "Printers," laser printers are page printers: The whole page must fit into the laser printer's RAM to be printed.

Thus, there are two variables to this printing problem : the size of the RAM in the printer and the size of the documents the user is trying to print. The self-test reveals the printer has only the standard amount of RAM (2MB) on board. This amount of RAM is adequate for text, but an elaborate page can overload it. A look at the PageMaker document reveals that it has a large amount of graphic content, whereas the Microsoft Word document is standard- sized text only with a minimal use of bold and italic formatting.

Your theory is to add RAM to the printer, and it can print the brochure. If you don't have a suitable RAM module, how can you prove it?

Because Microsoft Word printed a text-only document flawlessly, you might be able to convince your client from that fact alone that the printer isn't "broken" but needs a RAM upgrade ”or a workaround.

Devising a workaround that will help the printer work is good for client satisfaction and will prove that your theory is correct. Have the client adjust the graphics resolution of the printer from its default setting to a lower amount, such as from 1,200dpi to 600dpi or from 600dpi to 300dpi, and print the brochure again. If a lack of printer memory is the cause of the problem, reducing the brochure's dots per inch for graphics objects will enable the brochure to print. The client will look at the lower print quality and at that point you can recommend the RAM upgrade. Point out the provision for RAM upgrades in the printer manual if necessary. Remember, you're not selling anything, but solving problems.

If the printer will not print at all, other tests from Table 22.1 are appropriate, such as the I/O port loopback test or hardware resources check.