White Box Troubleshooting 101

Every PC you encounter will seem to have its own personality—and an enigmatic, stubborn one at that. Sometimes you can pinpoint a problem quickly and correct it in a snap. But other times, even though you know you've done everything possible to overcome a problem, the darned box simply refuses to cooperate.

Don't fret. In this TechBuilder Recipe, I'll take you through the three major areas involved with analyzing and fixing any PC you're likely to come across: hardware, software, and security.

Whether you're just getting started as a professional system builder, or you're a serious whitebox hobbyist, this Recipe is for you. Let's get started.

Five Steps for Troubleshooting System Hardware

First, let's look under the hood of a typical system and identify five of the most common hardware problems a system builder will encounter. Some of these steps will serve as reminders to take care of the "no-brainer" stuff, such as checking a PC's connections. Others--such as problems with an optical drive subsystem--will take more careful consideration. We'll start with the simplest, easiest-to-fix problems, and then move along to the more complicated ones.

• Power supply or voltage selection switch: The most basic element of any electronic device is electrical power, of course, and that's certainly true for a PC. Yet there'll be times when you think you've done everything a technician is supposed to in order to get a PC up and running--and yet, for some strange reason, there's no juice getting to the system's components.
• It's easy to jump to the conclusion that you've got a bad power supply. But odds are the power supply is fine. Instead, check all the connections. Sounds simple, but it's often an easy step to overlook. A typical power supply is depicted here:
• If, after checking the connections, you still can't get the system to boot, check the system's voltage selection switch and make sure it is set correctly for your country. Since most power supplies originate out of the Far East, it's not uncommon to find the voltage selection switch has been set to 220 volts. If you're in the U.S. or other country that runs at 110 volts, you'll need to reset the switch.
• Power-on configuration: Another gremlin you'll occasionally come across occurs with the “power-on configuration" of the motherboard. Modern system boards control the power-on functionality of any system. That's most often accomplished by connecting a jumper wire from the case power switch to a connector on the system board, as shown here:
• Also, make sure the system board and all components on the system are properly grounded by attaching all mounting screws and mounts.
• Video, RAM or CPU: If power is not the problem, then the next likely candidates to check are video, RAM, and CPU. You'll know you're having problems with any of these three areas, as ailments are indicated by a series of audible beeps.
• Most motherboards have an integrated piezo electric speaker that beeps a certain number of times (it's different for each machine) when there is a problem of some kind. That beep count can be your best clue to drilling down on a hardware problem.
• Simply refer to the motherboard's documentation to locate exactly which problem corresponds to the beep count you hear. But note, in some instances you may need to physically connect an external speaker to the motherboard (via a jumper) to actually hear the telltale beeps. Let's look at each function separately:
  • Video : For troubleshooting video troubles, first swap out the video card and replace it with another. If the swap fixes the problem, you've found your culprit. However, there are other elements that could make video unworkable. First, make sure the video card is seated properly—that's another often-overlooked no-brainer. Next, be aware that certain video cards require additional power; so make sure the power supply is connected properly, and also make sure the power supply is the one required for the system. Here's a look a one common video card:
  • RAM : Some systems have "on board" video, in which a video controller is integrated directly onto the motherboard. In these cases, a lack of video can be attributed to a jumper or switch that is not set properly to enable the onboard video. This is usually a very simple fix. But when RAM is the root of the problem, often the modules have not been inserted properly. The first step is to reset the modules and check for any bent edge connectors. Next, check the system documentation to ensure you're using the correct RAM. Finally, be aware that for some systems, modules need to be paired together to make the system bootable. Here's a typical set of RAM modules:
  • CPU : CPU-related problems can be more complex to solve. If you're dealing with a faulty CPU, you usually will need to trash the chip. Be careful when installing the CPU, because many processor problems are due to plain old incorrect installation. For example, be careful not to damage the pins. A common chip (with inherent pins) is shown here:
  • Also, because modern CPUs are expected to power a myriad of parts on a system, adequate cooling is critical. Running a CPU without a heat sink or a fan can damage the processor in as little as a few seconds! Here's a look at a fan that has been correctly installed:
• Hard drive cables: If you encounter the dreaded “operating system not found" or “inaccessible boot device" error messages, then the problem is likely related to either the hard drive or the optical drive subsystem. The biggest culprit here is often (believe it or not) a reversed cable! Some cables or drives are not “keyed" properly, or they are incorrectly notched, meaning they're simply installed backwards. Happily, by merely by correcting the cable, this problem will be solved. A typical cable—notched correctly—is shown here:
• Checking the BIOS: Another system “gotcha" concerns the BIOS settings. In other words, the default bios settings are set wrong! Some manufactures default to "IDE drives disabled" or other such settings that can prevent an attached drive from being recognized. The simplest solution is to check the BIOS and make sure it's set correctly. But for new installs, you should also check the “boot order," as usually you'll need to boot off the optical drive (CD/DVD) to install an operating system. That should be set to boot first, as shown here:

Four Steps for Troubleshooting System Software

But what if you've checked a system's hardware, and it still won't work? It's time to check on the four most common software-related issues when it comes to troubleshooting a PC: the operating system (OS), refresh, boot sector code, and backup. Let's look at each separately:

• Reinstall the OS: On a new system, the easiest action to take is to simply reinstall the operating system. Most systems come with recovery disks that recreate the “as delivered" state of the system's software. System builders should take note here and consider creating that style of rescue disk required for their customer's system.
• Refresh: If you're working with a used system on which additional software has already been installed, then a "refresh" to the factory software setup may be what's called for. To refresh (a loose term for rejuvenate) means several things, such as deploying Disk Cleanup to reduce clutter on a Windows machine, running Disk Defragmenter to repair broken files and folders, and running System Restore to roll back a Win PC to an earlier, uninfected date. But refresh can also restore a system to its original factory-setup mode. Most tier-one computer vendors provide a system-restoration CD with their computers. This CD will reformat the hard drive, reinstall Windows, and reinstall any of the programs that came with the computer. But this CD cannot back up any data files or programs that were installed independently on the hard drive. These may be items your client wants to save, so be sure you save those data files onto blank CDs before restoring. Also, have available any CDs for application software that were installed independently.
• Reinstall the boot sector code: If the system still won't boot, before moving on to deploying recovery disks, you may want to try to reinstall the boot sector code. Most operating systems allow a technician to boot direct from the OS CD. In many cases, that step alone will revive a system enough to back-up any important data. It may even completely solve the problem.
• Backup: The last line of recovery is to perform a complete backup. The fastest way to do this is with a USB hard drive and an imaging backup program such as True Image from Acronis or Norton Ghost from Symantec. Using such programs can capture a complete system drive to a data file for later recovery.
  1. Three Steps for Troubleshooting Security Problems
  2. Sometimes, the most difficult problems a system builder will encounter are those caused by malware, or malicious code such as viruses and worms. Malware problems can be reduced, if not completely avoided, by following these three simple steps:
     1. Scan for viruses and worms: If the system boots into the OS and is usable, scan for viruses and worms. The easiest way is to run an online virus scan. A terrific free online virus scan is Housecall from Trend Micro. This online scan will completely examine the system you're working on for any known viruses or worms. It conducts the scan using the latest anti-virus information in Trend Micro's database. After the scan is completed, you have the option of hiring Trend Micro to repair the problem. But at the very least, you'll get general information on how to recover the system from a virus or worm attack.
     2. Get rid of spyware: Another major problem on today's Internet-connected systems comes from the pesky spyware programs a user often (unwittingly) downloads. Alternatively, a system can become so overrun by pop-up ads, also known as "adware," that it becomes practically unusable. In either case, the system slows to grind. This slow-down is most evident when you boot the system, launch applications, or receive Internet access.
     3. Solutions to combat these problems are readily available. On the commercial side, I recommend going to the major players, such as Symantec or McAfee. Both offer comprehensive security packages that can solve many, if not all, of the problems caused by spyware and adware. Be aware, however, that these companies charge two fees: a one-time license to get started, plus a monthly fee to keep the protection up-to-date.
     4. Freeware security tools: If you're looking to cut costs, there are plenty of freeware utilities to consider. One of the most notable is Microsoft's own AntiSpyware. This product can be applied to a system suspected of being infested by spyware and adware troubles. Another well-regarded, free spyware solution is Spybot's Search and Destroy, which can be deployed to disinfect a troubled system. Another good spyware and adware application is Ad-Aware from Lava Software, which can work wonders on a balky system.
     5. Now you know how to solve the most basic hardware, software, and security system problems. Remember to start trouble-shooting with the simplest, most obvious problems first. Only after you have eliminated them as possible causes of the system's problem should you move on to the more complicated problems. In other words: Keep it simple.
     6. CAROL ANN MUFF is the operations manager for Magnum Computer Consulting, a New York-based systems integrator specializing in custom accounting solutions for the small-business market.