PC Troubleshooting and Repair Guide
This material is © of http://www.pcguide.com/ts/gen/index.htm. There may be updated versions of this document on that site; the document is also broken down into smaller chunks on that site.
Introduction:
There is nothing more upsetting for a PC user than when there is a problem with their machine. This upset can turn quickly to frustration when the problem seems to be impossible to solve, or even to understand. Every PC user has experienced these feelings, but it is in most cases possible to both diagnose and correct most problems with the typical PC. And with some help, you can usually do it yourself. The most important resource you can have at your disposal when you are trying to troubleshoot a problem with your PC is: experience. Those who have done a lot of work diagnosing and correcting problems with a wide variety of PCs develop a knack for recognizing problem situations that they have seen before. They also learn (and invent) techniques that make it possible for them to get to the root of a problem quickly. There's no substitute for experience, but I'm hoping that this Guide will be the next best thing. I have accumulated here the experience of myself and many other knowledgeable PC users, upgraders and maintainers, to help you learn how to detect and correct many common problems that plague PC users. This includes both general rules of thumb regarding how to troubleshoot your PC, as well as information on dealing with vendor warranties, and repairing your machine. The heart of this Guide is The Troubleshooting Expert, an expert system database of questions and answers that will help you to quickly find a solution to the most common problems with PCs. Over time I will continue to add to and expand this Expert so that it becomes an increasingly more and more useful tool for tackling the woes that plague the computer user. Tip: You may be able to get assistance with many PC problems by using The PC Guide Discussion Forums.
General Troubleshooting Techniques
"Give a man a fish, and you feed him for a day. Teach a man to fish, and you feed him for a lifetime." -- Chinese proverb
Many people will probably come to this Troubleshooting Guide looking for the answer to a specific problem that they are having with their PC. The Troubleshooting Expert can often help with this. However it is worthwhile if you are having a problem to read first the general troubleshooting information in this chapter. Using it you may be able to not only solve your current problem, but also develop your own skill as a troubleshooter so you can correct future problems more easily and quickly.
This chapter contains descriptions of most of the techniques and tricks that I use when I have a difficult problem that I need to diagnose and correct. They are based on my own experiences and my study of problem-solving methods, as well as knowledge of various problem situations that I have accumulated.
Warning: Please read the site warnings before you do any work on your PC.
Troubleshooting and Your Mental State
Very few people seem to understand how incredibly important the mental state is of the person who is attempting to troubleshoot a system. I have personally seen many times (and often with myself as the subject) the difference between a person in the right mental state and one who is not, in the ability to identify and correct system problems. It can easily be the difference between a problem that is fixed in minutes and one that languishes for hours or days.
I would urge all troubleshooters to read this section and keep these rules of thumb in mind as you try to work out your PC problems. If you get stuck, come back and refer to them again. You may be surprised how much just thinking about the problem the right way can make it much easier to solve.
Next: Don't Panic or Overestimate the Magnitude of the Problem
Don't Panic or Overestimate the Magnitude of the Problem
Probably the most important single thing to keep in mind when you find a problem with your PC is not to panic. Being in "panic mode" makes it extremely difficult for you to work on solving the problem, and in many cases the panic will turn out to be unwarranted anyway. Remember that most PC problems do not have to be solved with a clock ticking away in the background; this isn't a time bomb you are dealing with and if it is easier to turn off the PC and deal with the problem later, do so.
In fact, there are many problems with PCs that appear to be very serious but in fact are not. It isn't always possible to tell at the beginning how serious a problem is just by its outward symptoms. For example, there are some problems that can manifest themselves with your hard disk appearing to be crashed and all of its data lost. Sometimes real crashes do occur, but there are other problems that can cause a disk to appear crashed when really the problem is simple and can be fixed in a matter of a few minutes.
Panicking can also lead you to jump to a solution to the perceived problem before you really understand it, which can make matters worse.
Don't Jump to the Solution Before You Understand the Problem
I have noticed in my studies of problem-solving techniques that there seem to be differences in the way that many Westerners approach problems and the way that many of those of the East seem to. In Japan, for example, when a problem arises, the focus is generally kept on analyzing the problem for a lengthy period of time, and using specific techniques. In the U.S. and many other Western countries, when a problem arises the focus seems to shift very quickly from the problem to implementing the solution.
The difficulty with jumping quickly to the solution to a problem is that if you don't take enough time to really understand what the problem is, you may end up applying the wrong fix to the situation and making things worse. You will also make it more difficult to figure out what caused the problem so that you can prevent it from recurring. Even if you resolve the situation this time, you may find a similar issue cropping up again in the near future.
The key is to exercise patience. If possible, you want to take enough time to make sure that you understand what you are seeing and what the root cause is. (Although with some sorts of intermittent problems you may not be able to explore the root cause easily). Only by addressing the root cause of the problem will you be able to take the most effective corrective steps and ensure that the problem will go away permanently.
Don't Be Stubborn
I'm a very stubborn person, and I know many others are as well, but believe me when I tell you that nobody is going to out-stubborn a computer. If you are having a persistent problem and everything that you have tried isn't working, try to think of something else you haven't tried. Change tracks. Clear your mind for a while and then try to think of a totally different possible explanation for what you are seeing. You may discover what the problem is this way.
Keep an Open Mind
Very often the real cause of the problem you are seeing is the last thing that you can possibly think would be it; problems simply are not always cause-and-effect. Keep in mind the famous quote of Sherlock Holmes (Sir Arthur Conan Doyle): "When you have eliminated the impossible, whatever remains, however improbable, must be the truth". It is very common for a symptom to be the result of a problem in a completely different part of the PC, which is why you need to make sure that you fully explore the problem before looking for a solution. In addition, bear in mind that changes to a system usually are the cause of problems that arise shortly thereafter, even if they are in unrelated areas.
Give Yourself Time
Starting a troubleshooting session on a difficult problem one half-hour before you have to go away for the weekend is not a great idea. Since you haven't allowed yourself nearly enough time to do a proper job of diagnosing the situation, you are going to feel rushed. It's likely that you will misdiagnose the problem, have to stop half-way before you are finished, and that you will generally feel uncomfortable during the entire process (even more than usual).
Give yourself a fighting chance to resolve the problem properly by allowing yourself enough time to do the job right.
Be Wary of Fatigue
Problem-solving, which is what troubleshooting is at its core, requires concentration. Concentration is hard to hold when you are feeling tired. If you start to feel fatigued, you may be much better off to call it a night and try again the next day, unless there is a rush to get the problem addressed for some reason. (While you are taking this rest, as well, your subconscious mind may even help to solve the problem for you.)
A fresh look after a good night's rest has been enough to solve a problem that seemed "unresolvable" the night before more than once for me.
If You're Stuck, Try Using Your "Background Processing Capabilities"
An amazing capability of the human brain is that under certain circumstances you can put your own subconscious to work for you, to help you solve a problem without expending any energy at all. I am no expert in the field of psychology, but I have used this technique myself with great success on many occasions.
Have you ever tried to remember something, some detail of a past conversation, or a phone number of an old friend, and felt that you knew it but just couldn't remember the exact words or numbers you were looking for? And then a few hours or a day or two later the answer will "pop into your mind" apparently for no reason, while you are doing something else? This is what I am talking about, and sometimes it can help you to resolve problems with your machine.
Some day, you may have a particularly difficult problem that you have been trying to solve for quite a while. It may seem that the answer is obvious, yet you can't quite grasp it. If this is the case, try just going on and doing something else temporarily. Take your conscious mind off the task at hand and try to do something enjoyable, relaxing or distracting for a few hours or the rest of the night. Let your subconscious work on the problem "in the background".
Then come back the next day and look at the problem again. You may find it much easier to resolve the problem now--I've even had the answer come to me in a matter of minutes after spending the previous entire day trying to figure it out. You may even find the answer popping into your head while doing something completely unrelated.
Two Heads are Better Than One
If you have at your disposal someone who knows about PCs, say a peer in your office or a family member, have them take a look at a difficult problem. Even if you are great at troubleshooting, there will be the occasional problem that for one reason or another, you just won't be able to crack. Sometimes another person who can look at a problem from a fresh perspective (who hasn't been staring at the same thing for days trying to resolve it) will be able to set you on the right path to fixing the problem.
Don't be afraid of asking for help. There is no shame in needing assistance, and it can be an excellent way to learn.
Tip: You may be able to get assistance from others with your PC problem by using The PC Guide Discussion Forums.
It Isn't Personal...
OK, this section hits me very close to home. Before you take umbrage at any of the comments that I make here, I will start by saying that I am one of the worst people I know in doing all of the things that I say in this section one should not do. :^) I have a pretty bad temper and tend to take things too seriously in general, so it's not surprising that I tend to lose it sometimes when I am stuck with a particularly enigmatic situation with a PC.
What I am saying here is that despite the fact that it may seem otherwise when you are feeling frustrating about some sort of difficulty, the PC is not doing whatever it is doing intentionally to annoy you. It is not alive, it has no feelings, no motivations, no agenda. There is a logical reason for everything that a PC does--it is just hard sometimes to understand what it is. It can seem like the PC is "out to get you", and then to start feeling very frustrated and angry. Usually when this happens your ability to actually solve the problem will drop pretty close to zero.
During the throes of particularly frustrating spells, I personally have at various times yelled at PCs, kicked PC case covers, slammed the desk with my fist, crunched up floppy disks in my hands, smashed keyboards, smacked the side of a monitor, and hit the power switch rather hard more than once. As a result of this foolishness, I have on rare occasion damaged a component or two, and even hurt my hands. This is one area where I definitely do not recommend that my readers follow my example. :^)
Try as much as possible to keep your cool when working on your machine. If you feel yourself getting really pissed off, this is likely a good signal that you've been working on the problem too long. Take a break, do something else for a while. A fresh look the next day will be better for both you and the PC.
Steps To Take First When Troubleshooting
There are many different tricks and techniques that I outline in the Troubleshooting Guide, but the ones in this section are the few that I always recommend be tried first before proceeding into a full troubleshooting effort. The reason is simple: they can often solve the problem you are having before you spend a lot of energy looking into it seriously.
Let the Vendor Do the Work!
You may not have to use the troubleshooting information on this site at all. If the system is exhibiting strange behavior, especially if it appears that it has any failed or failing components, and you purchased this system recently, you may want to just return it and let the vendor figure out what the problem is. In many cases this is the best thing to do when your PC is still under warranty, since this also eliminates any chance of you voiding the warranty, and saves you time as well. That's what a warranty is for. See here for information on using your warranty, and repair in general.
The problem with doing this is that you have to incur the hassle of bringing the PC back for service. If the PC was purchased mail order, the problem is even worse. Sometimes looking on the site first is a good idea anyway, to rule out any possible simple causes that you can address yourself.
Scan For Viruses
If you experience any strange behavior on your system, especially either as it is booting, or shortly after it completes, you should always use whatever antivirus software you own to do a full scan of the system before you do anything else. You should use a clean boot floppy if possible.
Do not fall into the "this can't be a virus" trap too easily. Obviously if you turn the power on and nothing happens at all, or if your hard disk won't spin up, or if your system seizes while it is doing the initial BIOS power-on test of your system memory, these are hardware problems that are not caused by a virus. However virtually any other symptom that appears to be a hardware issue can be either a hardware problem or simply a virus that is trying to masquerade as a hardware problem. To remove the complicating factor of always having to wonder "is this a virus?", scan your system before you spend a lot of time chasing ghosts.
Read the Documentation
I never cease to be amazed at how people refuse, steadfastly, to read the documentation that comes with their new hardware or software. They are so excited about their new toy, whatever it is, that they are dying to get it running right away--who has time to start reading? So they start installing whatever-it-is and then have problems. Half the time they end up actually taking more time to get the thing running than they would have if they had just read the instructions in the first place. Meanwhile they often waste a lot of their time and others' asking for technical assistance.
Many, if not most problems are a result of hardware or software that has just been installed into the PC. If you are installing something new, read the documentation that came with the product. Look for files called something like "READ.ME" or "README.1ST" on installation or driver disks, and read them thoroughly before you start. Refer back in the event of trouble. You may be surprised to find that the answer to your dilemma is right there, just waiting for you to read it.
Double-Check Any Recent Changes
Ask yourself: when did the problem I am experiencing first start? If you have just made any change to your system, and are now having a problem with your PC that was not present before the change, the chances are probably 99% that the change is the cause of the problem. This is true even if the problem seems to have absolutely nothing to do with what you changed. If you hunt around, eventually you will find something that links the problem to the change. While it is certainly possible for something to coincidentally stop working at the same time that you make a change, the odds are greatly against this ever happening to you.
Here's an example of something that happened to me, to illustrate what I mean. I opened up my PC to do some work on my system memory. I removed two SIMMs and replaced them with another pair of higher capacity. When I turned on the system next, my floppy disk drive was malfunctioning. Had my floppy drive failed just then? Almost certainly not. As it turned out, my motherboard has the header for the floppy drive cable located directly adjacent to the SIMM slots. While I was working on the memory I had loosened the floppy drive cable.
Sometimes the problems can be even more indirectly related than this. You should in most cases stick to pursuing how the change affected the system, as this is likely to eventually lead you to the cause of the problem.
Simplify, Simplify, Simplify
An absolutely crucial rule to keep in mind when troubleshooting PC system problems is that if there are too many unknowns, it is impossible to determine which one is causing the problem. If you have many possible causes for some difficulty with your system, it can be extremely difficult to narrow down the cause of the problem to any one of them. If you are using a just-installed hard disk with an unfamiliar shareware file utility running on an operating system you just upgraded last week, and now your system has problems, how on earth are you ever going to know what is causing them?
In order to have a fighting chance at figuring out what is going on, you must simplify the situation as much as possible so that it becomes much more obvious what is responsible for the difficulty. This means reducing the number of variables to whatever degree possible. One important way of doing this is undoing or double-checking any changes made to the system.
In addition, I have identified the following items as often being responsible for erratic behavior that can complicate troubleshooting. I would recommend that they be eliminated or temporarily disabled when trying to diagnose a system problem:
- Power Management: Power management is a great idea in theory but in many ways is just "not ready for prime time". Power management routines can cause symptoms that appear to be hardware malfunctions, such as screens that turn off unexpectedly or hard disks that spin down. They also can cause crashes of software that doesn't know how to deal with them. If you want to use power management, turning it off until the problem is resolved is wise.
- Overclocked Hardware and Aggressive BIOS Settings: I do not believe in overclocking. If you insist on doing it, don't be surprised if you have system problems! Scale things back until you can figure out what the problem is. Similarly, if you are "pushing the envelope" in trying to squeeze maximum performance by tuning your BIOS memory timings and other settings very aggressively, try resetting them to more conservative values when troubleshooting.
- Experimental or Beta Software: This software is still in the test process and is likely to have bugs--that is why it is labeled as "beta"! For an end application this is usually no big deal, since any crashes or other problems will be limited to that application and therefore somewhat obvious. Running beta operating systems, drivers or other low-level software however is asking for trouble, and you should try to eliminate these possible sources of confusion when trying to debug your system.
- "Creative" Configurations: The more "unusual" things that you have going on in your system, the more likely that you are going to have a conflict caused by one of these strange pieces of hardware or software. A system that is loaded with unusual utilities, terminate-and-stay-resident programs, an old 8-bit network card salvaged from a 286, etc. will often have more problems than a stock Pentium box with a normal Windows 95 installation. To whatever extent possible, disable these items while troubleshooting. Also try to avoid using unusual low-level software whenever possible.
- Excessive Connections: If the PC is on a network or is connected to a large number or variety of peripheral devices, you may want to try to disconnect those and see if there is any impact on the problem.
In general you want to avoid the unusual or the unknown when troubleshooting. One way to simplify the software environment during diagnosis is to use a boot floppy to "boot clean" and bypass the special drivers and software that you normally load when you boot from your hard disk. You can also use the {F8} key when DOS or Windows 95 are booting to bypass your startup files, basically accomplishing the same thing. You want to be especially wary of software that sits in the background and activates without you specifically telling it to, as this can confound your troubleshooting efforts.
To whatever extent possible, disable as much as you can when trying to figure out a problem. The more funky software utilities, screen savers and cute peripherals you disable now, the more chance you have of finding out which one it is that is causing the problem later on.
General Diagnostic Techniques
This section discusses the rules of thumb that I recommend you employ when you are troubleshooting. These are the general techniques and ideas that can help you solve all sorts of different problems with your PC (and even in other parts of your life...) It often is worth trying these general tips before attempting to find a specific answer to your problem in the Troubleshooting Expert.
Get to the Root Cause: Ask "Why" Five Times
One of the keys to truly solving a problem is to first truly understand it. It is easy to observe that too often, people are quick to jump to solve a problem before they really understand what it is. It is easy to see a symptom and think that you know what is causing it, but surprisingly often, if you take the time to explore more deeply you will find that what you thought was the cause is in fact just another symptom, and that the problem lies much deeper within.
One technique that you can employ is borrowed from Japanese manufacturing management theory. This method is used to help identify the real causes of problems that occur on the manufacturing floor. The goal is not to simply correct the effects of the problem, but to find out the root cause of why the problem is occurring so that we can ensure that it will not happen in the future.
One simple way to do this is called asking why' five times. The idea is that by the time you have asked "why" the fifth time, you will be at the root cause. It isn't always that simple, but the exercise can be surprisingly insightful in helping you figure out what is really going on, and can help you avoid "quick fix" solutions that are really just band-aids and don't resolve anything. It is especially useful for tackling chronic problems that show up over and over again in a system; it is less useful for problems that are unlikely to recur.
Here's an example. Let's suppose your hard disk is having a problem with bad sectors showing up. The knee-jerk reaction to this happening is "the hard disk is bad, replace it". Instead, ask yourself:
- "Why are bad sectors showing up when they didn't before?"
- "The hard disk is being corrupted somehow; something has changed."
- "Why?"
- "The timing of the system probably has been changed in some way."
- "Why?"
- "Hmm, I installed a new hard drive in that system just last week, perhaps it is causing the problem."
- "Why?"
- "The hard drive may not have been installed correctly, or perhaps I should have used a different channel. Or maybe it was the bus-mastering driver that came with the new drive, the old drive may not be able to support it".
You can see the general idea; the answers will differ in every case, but it is the process itself that is useful. Another example: imagine an office that has just had a catastrophic data loss due to a PC crashing after an electrical storm. The initially identified cause of this problem was a lack of a UPS on the PC, which would have protected the system from the electrical storm (usually). So the MIS department starts drafting purchase requisitions for UPSes for all the PCs. Instead, if they looked at the problem carefully, they may have employed the following process:
- "Remember that the problem here was the data loss, not the crashing of the PC. Why was the data lost?"
- "The PC's hard disk crashed."
- "Why would that cause the data to be lost?"
- "Because we didn't have a backup."
- "Why wasn't there a backup?"
- "There is a tape backup unit on every PC in the building but very few of them are being used."
- "Why?"
- "Because most of the PC users do not know how to use them."
- "Why?"
- "They have never been given any training in their use."
- "Why?"
- "Because we don't have a budget for PC hardware training."
As you can see, the root cause that we came up with here is very different than the cause we saw when we only scratched the surface. This means of course that the solution will be very different as well; it might well make a lot more sense here to spend a small amount of money on training instead of a large amount of money on UPSes for every PC.
Be Observant and Look For Evidence
It is sometimes the subtle signs that can provide the triggering piece of information that leads to discovering the source of a problem. In particular, you want to carefully investigate anything that seems "unusual", "wrong" or "surprising". These are often the clues that will get you going down the right track to figuring out what is giving you trouble. If you are observant and careful about what you are doing, you will be much more likely to pick up on these things than if you are in a hurry or you are quick to discount your findings as "unrelated to the problem".
Use the Process of Elimination
Virtually all problems with PCs involve more than one component or subsystem. The difficulty is usually in figuring out which component is responsible for the problem. Using the process of elimination, however, you can usually narrow the problem down rather quickly by making small logical changes and observing the impact on the problem. Your objective is to isolate the cause of the problem so you can correct it.
The key is to make only one change at a time and then see if the problem goes away; if it does, then whatever you changed is likely responsible for the problem (although it could be fixing the problem indirectly in some cases.) If you make more than one change at a time, you cannot readily discern which change was responsible for fixing the problem.
You will want to first check the most probable sources of the problem, and also the things that are easiest to change. For example, if you are having a problem with your disk drive being recognized, it's a lot easier and cheaper to explore things like double-checking jumpers and connections or replacing the interface cable, than it is to try replacing the drive itself. That is something you'd only do after you had eliminated all the other possibilities (or if the evidence implicated the hard disk directly).
Here's a simple example. Let's suppose one morning your PC will not turn on. You hit the switch and nothing happens. There could be many different possible causes for this problem: the power to the house could be out; there could be a malfunction in the wall socket; the surge suppressor that the system is plugged into might have blown; the electrical cord may be loose; the power supply could be damaged. To figure out what is going on you need to eliminate these variables by making small changes and seeing what happens. For example:
- Change the wall socket you are using. If the PC now boots, you have isolated the cause to the electrical wiring in the house.
- If the problem persists, examine the surge suppressor. Change it, or temporarily bypass it and plug the PC into the wall directly. If it now works, the surge suppressor is the problem.
- If the problem still isn't fixed, try changing the power cord.
- If the problem persists still, you may then have to open up the box and look at the power supply unit to see if it might need replacing.
Realize that the key here is making these changes one at a time. If you approach this problem by changing the wall socket you use, bypassing the surge suppressor, and changing the power cord all at once, your problem may go away but how will you know what caused it? This is a valid way to troubleshoot if you have to get the system back up immediately, however. You can then undo the changes one at a time later on to find out what the cause is, in effect doing the same single-change-at-a-time process, but in reverse.
Most problems with the PC can be diagnosed using this sort of process; it's very powerful. You will notice many sequences like the one above in The Troubleshooting Expert.
Do One Upgrade or Assembly Step At a Time
Changes made to the system are the most frequent cause of problems; this is the nature of change. Much the way many problems can be diagnosed by using the process of elimination through making single changes to the system, you can avoid or detect problems with upgrades or new installations by going "one step at a time" in your changes as well. New system installations or major upgrades often can have problems that are very difficult to diagnose simply because there are so many modifications being performed at the same time.
When you build a new PC you will probably be assembling a large number of components that have never been used or tested before. To whatever extent possible, try to go as slowly as possible when assembling the machine. For example, when assembling a new PC from scratch, it is always best to first make sure the basic system is working, with the only expansion card in the system the video card (if applicable). Adding the sound card, network card and other devices should be done later on, and one at a time if possible. Similarly, do not try to do major software or operating system upgrades at the same time that you make hardware changes. Doing this can make it very difficult for you to troubleshoot any system problems.
If you do make multiple changes at once, try retracing your steps. Undo the changes you have made one at a time and see if you can identify the change that caused the problem that way.
Make Use of Components That You Know Work
One valuable resource while troubleshooting is a "stockpile" of extra components. I put "stockpile" in quotes because it doesn't have to be an actual stock of components, just a resource that will let you borrow components that you know work while you are troubleshooting your system. Another functional PC that uses the same or similar components is normally quite adequate. Being in an office environment can be very helpful this way since there are many PCs around to work with, and often they are similar to one another; at home it may be more difficult if you only have the one machine.
These extra components can be useful when you are employing the process of elimination to resolve problems. For example, suppose that there is no video on your monitor when you boot up the PC. After eliminating the simple stuff (cable disconnected, brightness turned down far too much, etc.) there are two components that are most likely to be bad: the video card or the monitor. An easy thing to try is to swap monitors with another machine that is working properly. If the second monitor also doesn't work, and your original monitor works on the other machine, you can feel quite confident that it isn't the monitor that is the problem here.
You can sometimes avoid problems in assembling a new PC by testing out some of the components you are going to use before you begin. For example, take your new mouse, keyboard and monitor and attach them to a PC that is known to work, and test them out. This way if you have problems when you build your new system you will know it isn't due to these components. At my office, I keep an extra keyboard, mouse and monitor just for debugging and assembling PCs. Since I know that they work, I reduce the number of variables that I have to deal with when I experience a problem. Once the main PC box is assembled and working, I connect the new keyboard, mouse and monitor. If a problem shows up then, I know that it is not likely the PC, but the peripherals I just connected.
Write Things Down
Whenever you discern a piece of evidence, or you figure out something about the current situation your PC is in, document it. Keeping a history of what you discover and what your symptoms are will help you not only with this problem but also with any future difficulties. A log book of some sort is a good idea.
Determine Repeatability
Most problems with a PC fall into one of two categories: either they are repeatable or they are intermittent. A repeatable problem is one where the problem occurs all the time, or always in response to a specific user action. For example, a PC that has a problem that prevents it from booting will probably always fail to boot no matter how many times you reset it. Or you may have an application that whenever you try to run, will crash with an error. You may find that your PC hangs, but only when you move the mouse at the same time that you are communicating using your modem.
In contrast, some problems are intermittent and not repeatable. In some cases, you may have a PC that will usually boot up fine, but one day a month will fail to boot for some reason. An application may work most of the time but occasionally crash. The PC may lock up at seemingly random intervals. Your mouse may work almost all of the time, but one day out of five or ten may give you trouble.
It is helpful to determine if the problem you are experiencing is repeatable, because intermittent difficulties are much more difficult to resolve than repeatable ones. If a problem is repeatable, and there is a specific set of actions that cause the problem, this gives you at least some initial clues about how to find the cause. In addition, you have a way of testing to see if you have resolved the problem when you are trying different solutions. Intermittent problems are much more difficult to deal with.
Determining if a problem is repeatable is pretty simple: try to duplicate the conditions that caused the problem and see if it happens again.
Dealing With Intermittent Problems
Intermittent problems are ones which appear to happen spontaneously or randomly. They seem not to be caused by anything obvious and are not repeatable. They can be extremely difficult and frustrating to diagnose.
Sometimes problems that seem intermittent really aren't; it's just that the specific set of circumstances that causes the problem to occur may be obscure or hard to notice. Spend some time trying to determine what the circumstances are when the problem arises. For example, many problems with crashes or lockups will occur only after the system has been on for more than an hour; some may occur only within the first few minutes that the PC is turned on. You may find that program A only crashes when run at the same time as program B. Or that a particular behavior is associated with a peripheral of some sort.
You will have to be patient when dealing with these sorts of issues. Since the problem is not something you can duplicate at will, you may not be able to systematically work your way towards the ultimate cause. In this situation you may have to employ trial-and-error, making a change and then waiting to see if the problem recurs. It can take days (or longer) sometimes since you have to wait before seeing if the problem happens again. Be patient.
Correlation May Not Imply Causation
I use the word correlation, perhaps slightly incorrectly, to refer to two behaviors or symptoms that appear at the same time. Causation refers to two events where one is responsible for the other's appearance or existence. So what I am saying here is that if you see two strange things happening at the same time on your PC, this does not necessarily mean that one of them has caused the other. They could be coincidental, or it could be that where you think A is causing B, B is really causing A.
For example, you may find two symptoms on your machine: you have frequent file system corruption on your hard disk, and you have a lot of system lockups (hangs, crashes). You may think that the crashes are causing the file system errors, and in many cases this is true. However it is possible for the file system errors to cause the lockups. And it is also possible that both are just symptoms of another underlying cause.
Scan For File System Problems
If the problem you are looking at is anything that involves the hard disk, files, operating system problems, crashes or similar behavior, scan the file system for errors. This is a simple step (which you should be performing regularly as part of a preventive maintenance program) that may tell you immediately that there is a problem with your hard disk. This can also be a useful piece of evidence in troubleshooting some sorts of problems.
Use Diagnostic Tools
If you own diagnostic hardware and/or software tools, they can be useful in pinpointing some kinds of problems. If you feel that you are having a problem that can be detected by a tool, or you feel that running a diagnostic utility may shed more light on what's happening, then make use of the tool to help guide your analysis of the problem.
Next: Diagnostic, Troubleshooting and Repair Tools
Diagnostic, Troubleshooting and Repair Tools
Troubleshooting a problem with a PC can be very difficult and frustrating. Fortunately, there are places to turn for assistance, and there are also tools at your disposal that can make the job much easier. These can be used to identify specific problems within the PC, or just give you more information about the system so you understand better what you are dealing with. This chapter takes a brief look at these tools, which can make your life easier if you are having trouble.
Diagnostic Software
The use of diagnostic software tools can save you a great deal of time. While in my experience these tools do not usually identify what the cause of a system problem is, they often provide hints or at least valuable information about what is in the system and how it is working. Some of these are free or are included with common operating systems at no additional charge, while others are commercial products that range from affordable to rather pricey.
Here are some common software diagnostic tools that you will want to consider including as part of your troubleshooting arsenal:
- Power-On Self Test (POST): I am cheating a bit here because this isn't a separate diagnostic utility; it is in fact built into your system BIOS and it runs every time you start up your PC automatically. It is often the best indicator of system problems; make sure you pay attention to its audio and video messages. Don't disable its error-reporting functions unless you really need to.
- MEM.EXE: This simple utility, built into recent versions of DOS and also Windows 95, provides you with details about your memory configuration, as well as what is currently using your memory. It is especially useful when run with the "/C" parameter (use the "/P" parameter as well to make the output pause when it is scrolling).
- Microsoft Diagnostics: Better known as "MSD.EXE", this is a small DOS utility that takes a brief inventory of the contents of your PC and shows them to you in a text-based format. This is very useful for seeing what disks are in the system, how much memory is installed, and also for checking system resource usage such as LPT ports and IRQs. It will show you what type of BIOS you are using and also what UART chip you have in your serial ports. MSD.EXE is included in later versions of DOS.
Tip: MSD.EXE will work on a Windows 95 machine but is not normally installed as part of the normal Windows 95 installation. You can find it on the Windows 95 CD however and then copy it to the C:\WINDOWS\COMMAND directory to make it work.
Warning: MSD.EXE can give incorrect information when run from within Windows. Shut down to DOS mode before running it ("Restart the computer in MS-DOS mode")
- The Windows 95 Device Manager: This is the most useful tool for identifying system configuration and resource usage information under Windows 95. To access it, open the Control Panel and select the "System" icon. Then select the "Device Manager" tab. You will see a graphical "tree" structure showing you all of your PC hardware. If you select "Properties" while "Computer" (the top-level item) is selected, you will be able to see all the IRQs, DMA channels and I/O addresses in use in your PC; very useful for resolving resource conflicts! The same "Properties" button, pressed after selecting a specific hardware device, will show you driver information, resource settings for the hardware item chosen, and much more.
- Norton System Information: This utility is similar to the Microsoft Diagnostics, only more detailed in its later versions. SI shows a great deal of information about what is in the PC, going well beyond what MSD gives you, but really is still an information utility as opposed to a true diagnostic. This program is part of Symantec's Norton Utilities.
- Microsoft ScanDisk and Norton Disk Doctor: These programs are used to check for hard disk problems. This includes file system corruption and hard disk read errors. They should be used when hard disk problems are suspected.
- Norton Diagnostics: This utility is meant to go beyond the System Information program and actually perform tests on the hardware to identify problems. It includes tests of the processor and motherboard and system memory, and will identify some types of resource conflicts. In reality it is still quite limited in terms of the numbers of problems it will find.
- QAPlus: QAPlus from DiagSoft is a more advanced diagnostic suite that comes in several flavors, depending on what you need to do and how you want to do it. This is a more expensive package but can give you much more detailed information about your system and help identify problem situations as well.
Diagnostic Hardware
While not as commonly used as diagnostic software, there are available several pieces of hardware that can be very helpful in troubleshooting some specific hardware problems. These units are not nearly as popular because they are usually more expensive and are used for specific types of problems, making them less universally applicable than the general software utilities that are widely encountered.
If you are a serious home builder or troubleshooter, you may want to consider one or more of the following:
- Loop-Back Plugs: These are small plugs that go onto the serial and parallel ports on your computer and connect the receive to transmit lines to simulate a connection (they connect the port to itself). Using these with a program like Norton Diagnostics allows for full testing of the operation of these ports, whereas without them only the internal portions of the port can be tested.
Note: The type of plugs used by a particular diagnostic utility may be different than the type used by another program made by another company. Check it out before you buy.
- Multimeter/Ohmmeter: An ohmmeter is a device that measures electrical resistance; a multimeter can measure resistance, voltage and current. These devices are used by electricians, electronics designers and repairpeople. They have come down greatly in price and simple ones can be had inexpensively (while top-end ones of course are still quite a bit of money). A simple ohmmeter is useful primarily for checking for short circuits or open circuits (broken connections, damaged cables, etc.) Multimeters can be used for more extensive electronics testing.
- BIOS POST Cards: It is well-known that the BIOS will, if it finds a problem during its power-on self-test (POST), produce audio "beep codes" and/or video messages that indicate what the trouble is that it found. What many do not realize is that the BIOS of most PCs is designed to send a stream of test codes to a special memory location as it performs its tests, usually 80h. By using a special card designed to capture and display these codes, you can pinpoint exactly where in the power-on self-test a system is hanging up. This can be extremely helpful in debugging very stubborn systems. Scott Wainner has more about one of these cards on his site.
- Test Bed: Some experienced PC repair people keep around an older system that they can use as a test bed for components. It can be very useful to be able to test an unknown device with other components that are known to work, to cut down on the guesswork.
There is also more advanced test hardware available, including devices such as logic probes, oscilloscopes, and many types of specialized component testers. These items are generally very expensive and require specific training to be used properly; they are not for the home PC user or even for most hobbyists. Professional test hardware costs thousands of dollars, which is one reason why doing it yourself is often not an option for many kinds of repairs.
The PC Tool Kit
If you are a very casual PC user you can probably get by hacking around with an old screwdriver and a pocket knife when you need to work on your PC. However, if you plan to do any reasonable amount of work upgrading, building, troubleshooting or repairing PCs, a proper tool kit is a good investment. There are many different types of tool kits available for the computer enthusiast, with tools selected specifically to make the lives of those who work with PCs easier. The right tools can save you a lot of time and help you avoid damage to your equipment (and your hands!)
As with most things, tool kits range widely in size, quality and price. At the low end are $5 kits that are barely any better than the screwdriver and pocket knife. At the high end are kits costing over $100 that include everything but the kitchen sink; these are overkill for most people (including me). Most people can do just fine with a toolkit costing between $15 and $50. These are often available pre-assembled into "genuine artificial leather" carrying cases, sold at most general computer stores.
Here are the items that I would recommend you definitely have in your toolkit; note that some of these don't generally come in pre-assembled toolkits so you will want to add them yourself to the carrying case. Every item in this list is one that I use personally on a regular basis:
- A Good Screwdriver: You want a good screwdriver with magnetic, interchangeable bits and a comfortable handle. At a minimum you should have two different sizes each of Phillips and flat bits. If you've never used a magnetic screwdriver before, you'll wonder how you ever did without one once you do; the magnetic tip will prevent screws falling into the PC, and will let you get access to drives mounted deep in your system case.
- Additional Screwdrivers: An additional Philips screwdriver and an additional flat blade screwdriver, for when you misplace the good one. :^) Actually, it's sometimes useful to have two screwdrivers anyway.
- Needle-Nose Pliers: These are useful for grasping small items and for removing and replacing jumpers on circuit boards.
- Wire Snips: A pair of wire snips for cutting wire and stripping insulation. Some kits include wire strippers for the latter purpose.
- A Small Flashlight: Very useful; the insides of PC boxes are quite dark and there are lots of very small things you will need to see, such as the "pin 1" marking on a connector for example.
- Tweezers: Or even better, a part retriever. A retriever is like a tiny set of retractable claws with a spring-loaded handle. Useful stuff.
- A Large Crowbar: For making adjustments to stubborn hardware. (Just kidding. :^) )
- A Roll of Black Electrical Tape: They never include this in the kits but they should. Used for wrapping wire ends and insulating components.
- A Can of Compressed Air: Very handy for cleaning things hands-free and without using any hazardous liquids.
- A Soft, Lint-Free Cloth: For cleaning the monitor and other components.
The following items I categorize as "optional, but useful". They are tools that are used less frequently, or perhaps only in certain circumstances. They are more for someone who does a fair bit of tinkering or who maintains a number of machines:
- A 3/16" Nut Driver: In fact, tool kits often come with several different sizes of nut drivers (using exchangeable bits), but 3/16" is the really important one, because this is the size of nut most commonly used on PCs. These hexagonal nuts are used as mounting hardware for motherboards and serial and parallel ports. A necessity for anyone who builds their own PCs; if you don't have one you'll be attaching port connectors using a pair of poor-fitting pliers.
- Vice Grips: These are incredibly neat tools that have all sorts of uses.
- Torx Screwdrivers or Bits: These are the star-shaped screwdriver heads that are used to make equipment "tamper-proof". You need these drivers if you want to, well, "tamper".
- An ESD (Electrostatic Discharge) Wrist Strap: This is more of a safety device than a tool; it is used to greatly reduce the chances of static damage to components.
- A Knife: A cutting blade or utility knife of some sort.
The following items are sometimes included in tool kits but in my experience are rarely if ever used:
- Soldering Irons: Essential if you are going to try to do repair of circuit boards, but virtually nobody does this (and with good reason). They are also quite dangerous if used improperly.
- Hex (Allen) Keys: They include these in virtually every commercial tool kit I've ever encountered. I've never seen anything PC related that used them but of course that doesn't mean that there isn't something out there that does...
- Chip Extractors: These are used to remove chips from regular (non-ZIF) sockets, but modern PCs do not generally use these sockets any more. Useful for upgrading much older systems in some cases.
Finally, it is a good idea to accumulate a stockpile of spare parts if you are going to be doing work on PCs. I would recommend the following:
- Screws: I have a "big ol' bag o' screws" of all shapes and sizes that comes in handy when working on various machines.
- Expansion Card Inserts: Save the metal inserts that come out of the back of the case when you put a modem or other card into the PC, as you may need them again later on.
- Drive Faceplates: Similarly, save the plastic faceplates that you remove from the front of cases so you can replace them later if needed.
- Mounting Kits: These sometimes come with retail hard disks and allow you to put a 3.5" drive into a 5.25" bay. Useful when your case has more free 5.25" bays than 3.5" ones.
- Cables: Save any power, IDE, floppy, CD-ROM, or other cables that you accumulate in case you need them later on.
- Keyboard, Mouse, 3.5" Floppy Drive: Keep an extra one of each of these components around to aid in troubleshooting problems by swapping. (Actually, keeping an extra of every major component in the PC is helpful, but the three I mention are here because they are cheap. Most people can't afford to keep an extra hard disk lying around as a spare.)
Don't Jump to the Solution Before You Understand the Problem
I have noticed in my studies of problem-solving techniques that there seem to be differences in the way that many Westerners approach problems and the way that many of those of the East seem to. In Japan, for example, when a problem arises, the focus is generally kept on analyzing the problem for a lengthy period of time, and using specific techniques. In the U.S. and many other Western countries, when a problem arises the focus seems to shift very quickly from the problem to implementing the solution.
The difficulty with jumping quickly to the solution to a problem is that if you don't take enough time to really understand what the problem is, you may end up applying the wrong fix to the situation and making things worse. You will also make it more difficult to figure out what caused the problem so that you can prevent it from recurring. Even if you resolve the situation this time, you may find a similar issue cropping up again in the near future.
The key is to exercise patience. If possible, you want to take enough time to make sure that you understand what you are seeing and what the root cause is. (Although with some sorts of intermittent problems you may not be able to explore the root cause easily). Only by addressing the root cause of the problem will you be able to take the most effective corrective steps and ensure that the problem will go away permanently.
Don't Be Stubborn
I'm a very stubborn person, and I know many others are as well, but believe me when I tell you that nobody is going to out-stubborn a computer. If you are having a persistent problem and everything that you have tried isn't working, try to think of something else you haven't tried. Change tracks. Clear your mind for a while and then try to think of a totally different possible explanation for what you are seeing. You may discover what the problem is this way.
Keep an Open Mind
Very often the real cause of the problem you are seeing is the last thing that you can possibly think would be it; problems simply are not always cause-and-effect. Keep in mind the famous quote of Sherlock Holmes (Sir Arthur Conan Doyle): "When you have eliminated the impossible, whatever remains, however improbable, must be the truth". It is very common for a symptom to be the result of a problem in a completely different part of the PC, which is why you need to make sure that you fully explore the problem before looking for a solution. In addition, bear in mind that changes to a system usually are the cause of problems that arise shortly thereafter, even if they are in unrelated areas.
Give Yourself Time
Starting a troubleshooting session on a difficult problem one half-hour before you have to go away for the weekend is not a great idea. Since you haven't allowed yourself nearly enough time to do a proper job of diagnosing the situation, you are going to feel rushed. It's likely that you will misdiagnose the problem, have to stop half-way before you are finished, and that you will generally feel uncomfortable during the entire process (even more than usual).
Give yourself a fighting chance to resolve the problem properly by allowing yourself enough time to do the job right.
Be Wary of Fatigue
Problem-solving, which is what troubleshooting is at its core, requires concentration. Concentration is hard to hold when you are feeling tired. If you start to feel fatigued, you may be much better off to call it a night and try again the next day, unless there is a rush to get the problem addressed for some reason. (While you are taking this rest, as well, your subconscious mind may even help to solve the problem for you.)
A fresh look after a good night's rest has been enough to solve a problem that seemed "unresolvable" the night before more than once for me.
If You're Stuck, Try Using Your "Background Processing Capabilities"
An amazing capability of the human brain is that under certain circumstances you can put your own subconscious to work for you, to help you solve a problem without expending any energy at all. I am no expert in the field of psychology, but I have used this technique myself with great success on many occasions.
Have you ever tried to remember something, some detail of a past conversation, or a phone number of an old friend, and felt that you knew it but just couldn't remember the exact words or numbers you were looking for? And then a few hours or a day or two later the answer will "pop into your mind" apparently for no reason, while you are doing something else? This is what I am talking about, and sometimes it can help you to resolve problems with your machine.
Some day, you may have a particularly difficult problem that you have been trying to solve for quite a while. It may seem that the answer is obvious, yet you can't quite grasp it. If this is the case, try just going on and doing something else temporarily. Take your conscious mind off the task at hand and try to do something enjoyable, relaxing or distracting for a few hours or the rest of the night. Let your subconscious work on the problem "in the background".
Then come back the next day and look at the problem again. You may find it much easier to resolve the problem now--I've even had the answer come to me in a matter of minutes after spending the previous entire day trying to figure it out. You may even find the answer popping into your head while doing something completely unrelated.
Two Heads are Better Than One
If you have at your disposal someone who knows about PCs, say a peer in your office or a family member, have them take a look at a difficult problem. Even if you are great at troubleshooting, there will be the occasional problem that for one reason or another, you just won't be able to crack. Sometimes another person who can look at a problem from a fresh perspective (who hasn't been staring at the same thing for days trying to resolve it) will be able to set you on the right path to fixing the problem.
Don't be afraid of asking for help. There is no shame in needing assistance, and it can be an excellent way to learn.
Tip: You may be able to get assistance from others with your PC problem by using The PC Guide Discussion Forums.
It Isn't Personal...
OK, this section hits me very close to home. Before you take umbrage at any of the comments that I make here, I will start by saying that I am one of the worst people I know in doing all of the things that I say in this section one should not do. :^) I have a pretty bad temper and tend to take things too seriously in general, so it's not surprising that I tend to lose it sometimes when I am stuck with a particularly enigmatic situation with a PC.
What I am saying here is that despite the fact that it may seem otherwise when you are feeling frustrating about some sort of difficulty, the PC is not doing whatever it is doing intentionally to annoy you. It is not alive, it has no feelings, no motivations, no agenda. There is a logical reason for everything that a PC does--it is just hard sometimes to understand what it is. It can seem like the PC is "out to get you", and then to start feeling very frustrated and angry. Usually when this happens your ability to actually solve the problem will drop pretty close to zero.
During the throes of particularly frustrating spells, I personally have at various times yelled at PCs, kicked PC case covers, slammed the desk with my fist, crunched up floppy disks in my hands, smashed keyboards, smacked the side of a monitor, and hit the power switch rather hard more than once. As a result of this foolishness, I have on rare occasion damaged a component or two, and even hurt my hands. This is one area where I definitely do not recommend that my readers follow my example. :^)
Try as much as possible to keep your cool when working on your machine. If you feel yourself getting really pissed off, this is likely a good signal that you've been working on the problem too long. Take a break, do something else for a while. A fresh look the next day will be better for both you and the PC.
Steps To Take First When Troubleshooting
There are many different tricks and techniques that I outline in the Troubleshooting Guide, but the ones in this section are the few that I always recommend be tried first before proceeding into a full troubleshooting effort. The reason is simple: they can often solve the problem you are having before you spend a lot of energy looking into it seriously.
Let the Vendor Do the Work!
You may not have to use the troubleshooting information on this site at all. If the system is exhibiting strange behavior, especially if it appears that it has any failed or failing components, and you purchased this system recently, you may want to just return it and let the vendor figure out what the problem is. In many cases this is the best thing to do when your PC is still under warranty, since this also eliminates any chance of you voiding the warranty, and saves you time as well. That's what a warranty is for. See here for information on using your warranty, and repair in general.
The problem with doing this is that you have to incur the hassle of bringing the PC back for service. If the PC was purchased mail order, the problem is even worse. Sometimes looking on the site first is a good idea anyway, to rule out any possible simple causes that you can address yourself.
Scan For Viruses
If you experience any strange behavior on your system, especially either as it is booting, or shortly after it completes, you should always use whatever antivirus software you own to do a full scan of the system before you do anything else. You should use a clean boot floppy if possible.
Do not fall into the "this can't be a virus" trap too easily. Obviously if you turn the power on and nothing happens at all, or if your hard disk won't spin up, or if your system seizes while it is doing the initial BIOS power-on test of your system memory, these are hardware problems that are not caused by a virus. However virtually any other symptom that appears to be a hardware issue can be either a hardware problem or simply a virus that is trying to masquerade as a hardware problem. To remove the complicating factor of always having to wonder "is this a virus?", scan your system before you spend a lot of time chasing ghosts.
Read the Documentation
I never cease to be amazed at how people refuse, steadfastly, to read the documentation that comes with their new hardware or software. They are so excited about their new toy, whatever it is, that they are dying to get it running right away--who has time to start reading? So they start installing whatever-it-is and then have problems. Half the time they end up actually taking more time to get the thing running than they would have if they had just read the instructions in the first place. Meanwhile they often waste a lot of their time and others' asking for technical assistance.
Many, if not most problems are a result of hardware or software that has just been installed into the PC. If you are installing something new, read the documentation that came with the product. Look for files called something like "READ.ME" or "README.1ST" on installation or driver disks, and read them thoroughly before you start. Refer back in the event of trouble. You may be surprised to find that the answer to your dilemma is right there, just waiting for you to read it.
Double-Check Any Recent Changes
Ask yourself: when did the problem I am experiencing first start? If you have just made any change to your system, and are now having a problem with your PC that was not present before the change, the chances are probably 99% that the change is the cause of the problem. This is true even if the problem seems to have absolutely nothing to do with what you changed. If you hunt around, eventually you will find something that links the problem to the change. While it is certainly possible for something to coincidentally stop working at the same time that you make a change, the odds are greatly against this ever happening to you.
Here's an example of something that happened to me, to illustrate what I mean. I opened up my PC to do some work on my system memory. I removed two SIMMs and replaced them with another pair of higher capacity. When I turned on the system next, my floppy disk drive was malfunctioning. Had my floppy drive failed just then? Almost certainly not. As it turned out, my motherboard has the header for the floppy drive cable located directly adjacent to the SIMM slots. While I was working on the memory I had loosened the floppy drive cable.
Sometimes the problems can be even more indirectly related than this. You should in most cases stick to pursuing how the change affected the system, as this is likely to eventually lead you to the cause of the problem.
Simplify, Simplify, Simplify
An absolutely crucial rule to keep in mind when troubleshooting PC system problems is that if there are too many unknowns, it is impossible to determine which one is causing the problem. If you have many possible causes for some difficulty with your system, it can be extremely difficult to narrow down the cause of the problem to any one of them. If you are using a just-installed hard disk with an unfamiliar shareware file utility running on an operating system you just upgraded last week, and now your system has problems, how on earth are you ever going to know what is causing them?
In order to have a fighting chance at figuring out what is going on, you must simplify the situation as much as possible so that it becomes much more obvious what is responsible for the difficulty. This means reducing the number of variables to whatever degree possible. One important way of doing this is undoing or double-checking any changes made to the system.
In addition, I have identified the following items as often being responsible for erratic behavior that can complicate troubleshooting. I would recommend that they be eliminated or temporarily disabled when trying to diagnose a system problem:
- Power Management: Power management is a great idea in theory but in many ways is just "not ready for prime time". Power management routines can cause symptoms that appear to be hardware malfunctions, such as screens that turn off unexpectedly or hard disks that spin down. They also can cause crashes of software that doesn't know how to deal with them. If you want to use power management, turning it off until the problem is resolved is wise.
- Overclocked Hardware and Aggressive BIOS Settings: I do not believe in overclocking. If you insist on doing it, don't be surprised if you have system problems! Scale things back until you can figure out what the problem is. Similarly, if you are "pushing the envelope" in trying to squeeze maximum performance by tuning your BIOS memory timings and other settings very aggressively, try resetting them to more conservative values when troubleshooting.
- Experimental or Beta Software: This software is still in the test process and is likely to have bugs--that is why it is labeled as "beta"! For an end application this is usually no big deal, since any crashes or other problems will be limited to that application and therefore somewhat obvious. Running beta operating systems, drivers or other low-level software however is asking for trouble, and you should try to eliminate these possible sources of confusion when trying to debug your system.
- "Creative" Configurations: The more "unusual" things that you have going on in your system, the more likely that you are going to have a conflict caused by one of these strange pieces of hardware or software. A system that is loaded with unusual utilities, terminate-and-stay-resident programs, an old 8-bit network card salvaged from a 286, etc. will often have more problems than a stock Pentium box with a normal Windows 95 installation. To whatever extent possible, disable these items while troubleshooting. Also try to avoid using unusual low-level software whenever possible.
- Excessive Connections: If the PC is on a network or is connected to a large number or variety of peripheral devices, you may want to try to disconnect those and see if there is any impact on the problem.
In general you want to avoid the unusual or the unknown when troubleshooting. One way to simplify the software environment during diagnosis is to use a boot floppy to "boot clean" and bypass the special drivers and software that you normally load when you boot from your hard disk. You can also use the {F8} key when DOS or Windows 95 are booting to bypass your startup files, basically accomplishing the same thing. You want to be especially wary of software that sits in the background and activates without you specifically telling it to, as this can confound your troubleshooting efforts.
To whatever extent possible, disable as much as you can when trying to figure out a problem. The more funky software utilities, screen savers and cute peripherals you disable now, the more chance you have of finding out which one it is that is causing the problem later on.
General Diagnostic Techniques
This section discusses the rules of thumb that I recommend you employ when you are troubleshooting. These are the general techniques and ideas that can help you solve all sorts of different problems with your PC (and even in other parts of your life...) It often is worth trying these general tips before attempting to find a specific answer to your problem in the Troubleshooting Expert.
Get to the Root Cause: Ask "Why" Five Times
One of the keys to truly solving a problem is to first truly understand it. It is easy to observe that too often, people are quick to jump to solve a problem before they really understand what it is. It is easy to see a symptom and think that you know what is causing it, but surprisingly often, if you take the time to explore more deeply you will find that what you thought was the cause is in fact just another symptom, and that the problem lies much deeper within.
One technique that you can employ is borrowed from Japanese manufacturing management theory. This method is used to help identify the real causes of problems that occur on the manufacturing floor. The goal is not to simply correct the effects of the problem, but to find out the root cause of why the problem is occurring so that we can ensure that it will not happen in the future.
One simple way to do this is called asking why' five times. The idea is that by the time you have asked "why" the fifth time, you will be at the root cause. It isn't always that simple, but the exercise can be surprisingly insightful in helping you figure out what is really going on, and can help you avoid "quick fix" solutions that are really just band-aids and don't resolve anything. It is especially useful for tackling chronic problems that show up over and over again in a system; it is less useful for problems that are unlikely to recur.
Here's an example. Let's suppose your hard disk is having a problem with bad sectors showing up. The knee-jerk reaction to this happening is "the hard disk is bad, replace it". Instead, ask yourself:
- "Why are bad sectors showing up when they didn't before?"
- "The hard disk is being corrupted somehow; something has changed."
- "Why?"
- "The timing of the system probably has been changed in some way."
- "Why?"
- "Hmm, I installed a new hard drive in that system just last week, perhaps it is causing the problem."
- "Why?"
- "The hard drive may not have been installed correctly, or perhaps I should have used a different channel. Or maybe it was the bus-mastering driver that came with the new drive, the old drive may not be able to support it".
You can see the general idea; the answers will differ in every case, but it is the process itself that is useful. Another example: imagine an office that has just had a catastrophic data loss due to a PC crashing after an electrical storm. The initially identified cause of this problem was a lack of a UPS on the PC, which would have protected the system from the electrical storm (usually). So the MIS department starts drafting purchase requisitions for UPSes for all the PCs. Instead, if they looked at the problem carefully, they may have employed the following process:
- "Remember that the problem here was the data loss, not the crashing of the PC. Why was the data lost?"
- "The PC's hard disk crashed."
- "Why would that cause the data to be lost?"
- "Because we didn't have a backup."
- "Why wasn't there a backup?"
- "There is a tape backup unit on every PC in the building but very few of them are being used."
- "Why?"
- "Because most of the PC users do not know how to use them."
- "Why?"
- "They have never been given any training in their use."
- "Why?"
- "Because we don't have a budget for PC hardware training."
As you can see, the root cause that we came up with here is very different than the cause we saw when we only scratched the surface. This means of course that the solution will be very different as well; it might well make a lot more sense here to spend a small amount of money on training instead of a large amount of money on UPSes for every PC.
Be Observant and Look For Evidence
It is sometimes the subtle signs that can provide the triggering piece of information that leads to discovering the source of a problem. In particular, you want to carefully investigate anything that seems "unusual", "wrong" or "surprising". These are often the clues that will get you going down the right track to figuring out what is giving you trouble. If you are observant and careful about what you are doing, you will be much more likely to pick up on these things than if you are in a hurry or you are quick to discount your findings as "unrelated to the problem".
Use the Process of Elimination
Virtually all problems with PCs involve more than one component or subsystem. The difficulty is usually in figuring out which component is responsible for the problem. Using the process of elimination, however, you can usually narrow the problem down rather quickly by making small logical changes and observing the impact on the problem. Your objective is to isolate the cause of the problem so you can correct it.
The key is to make only one change at a time and then see if the problem goes away; if it does, then whatever you changed is likely responsible for the problem (although it could be fixing the problem indirectly in some cases.) If you make more than one change at a time, you cannot readily discern which change was responsible for fixing the problem.
You will want to first check the most probable sources of the problem, and also the things that are easiest to change. For example, if you are having a problem with your disk drive being recognized, it's a lot easier and cheaper to explore things like double-checking jumpers and connections or replacing the interface cable, than it is to try replacing the drive itself. That is something you'd only do after you had eliminated all the other possibilities (or if the evidence implicated the hard disk directly).
Here's a simple example. Let's suppose one morning your PC will not turn on. You hit the switch and nothing happens. There could be many different possible causes for this problem: the power to the house could be out; there could be a malfunction in the wall socket; the surge suppressor that the system is plugged into might have blown; the electrical cord may be loose; the power supply could be damaged. To figure out what is going on you need to eliminate these variables by making small changes and seeing what happens. For example:
- Change the wall socket you are using. If the PC now boots, you have isolated the cause to the electrical wiring in the house.
- If the problem persists, examine the surge suppressor. Change it, or temporarily bypass it and plug the PC into the wall directly. If it now works, the surge suppressor is the problem.
- If the problem still isn't fixed, try changing the power cord.
- If the problem persists still, you may then have to open up the box and look at the power supply unit to see if it might need replacing.
Realize that the key here is making these changes one at a time. If you approach this problem by changing the wall socket you use, bypassing the surge suppressor, and changing the power cord all at once, your problem may go away but how will you know what caused it? This is a valid way to troubleshoot if you have to get the system back up immediately, however. You can then undo the changes one at a time later on to find out what the cause is, in effect doing the same single-change-at-a-time process, but in reverse.
Most problems with the PC can be diagnosed using this sort of process; it's very powerful. You will notice many sequences like the one above in The Troubleshooting Expert.
Do One Upgrade or Assembly Step At a Time
Changes made to the system are the most frequent cause of problems; this is the nature of change. Much the way many problems can be diagnosed by using the process of elimination through making single changes to the system, you can avoid or detect problems with upgrades or new installations by going "one step at a time" in your changes as well. New system installations or major upgrades often can have problems that are very difficult to diagnose simply because there are so many modifications being performed at the same time.
When you build a new PC you will probably be assembling a large number of components that have never been used or tested before. To whatever extent possible, try to go as slowly as possible when assembling the machine. For example, when assembling a new PC from scratch, it is always best to first make sure the basic system is working, with the only expansion card in the system the video card (if applicable). Adding the sound card, network card and other devices should be done later on, and one at a time if possible. Similarly, do not try to do major software or operating system upgrades at the same time that you make hardware changes. Doing this can make it very difficult for you to troubleshoot any system problems.
If you do make multiple changes at once, try retracing your steps. Undo the changes you have made one at a time and see if you can identify the change that caused the problem that way.
Make Use of Components That You Know Work
One valuable resource while troubleshooting is a "stockpile" of extra components. I put "stockpile" in quotes because it doesn't have to be an actual stock of components, just a resource that will let you borrow components that you know work while you are troubleshooting your system. Another functional PC that uses the same or similar components is normally quite adequate. Being in an office environment can be very helpful this way since there are many PCs around to work with, and often they are similar to one another; at home it may be more difficult if you only have the one machine.
These extra components can be useful when you are employing the process of elimination to resolve problems. For example, suppose that there is no video on your monitor when you boot up the PC. After eliminating the simple stuff (cable disconnected, brightness turned down far too much, etc.) there are two components that are most likely to be bad: the video card or the monitor. An easy thing to try is to swap monitors with another machine that is working properly. If the second monitor also doesn't work, and your original monitor works on the other machine, you can feel quite confident that it isn't the monitor that is the problem here.
You can sometimes avoid problems in assembling a new PC by testing out some of the components you are going to use before you begin. For example, take your new mouse, keyboard and monitor and attach them to a PC that is known to work, and test them out. This way if you have problems when you build your new system you will know it isn't due to these components. At my office, I keep an extra keyboard, mouse and monitor just for debugging and assembling PCs. Since I know that they work, I reduce the number of variables that I have to deal with when I experience a problem. Once the main PC box is assembled and working, I connect the new keyboard, mouse and monitor. If a problem shows up then, I know that it is not likely the PC, but the peripherals I just connected.
Write Things Down
Whenever you discern a piece of evidence, or you figure out something about the current situation your PC is in, document it. Keeping a history of what you discover and what your symptoms are will help you not only with this problem but also with any future difficulties. A log book of some sort is a good idea.
Determine Repeatability
Most problems with a PC fall into one of two categories: either they are repeatable or they are intermittent. A repeatable problem is one where the problem occurs all the time, or always in response to a specific user action. For example, a PC that has a problem that prevents it from booting will probably always fail to boot no matter how many times you reset it. Or you may have an application that whenever you try to run, will crash with an error. You may find that your PC hangs, but only when you move the mouse at the same time that you are communicating using your modem.
In contrast, some problems are intermittent and not repeatable. In some cases, you may have a PC that will usually boot up fine, but one day a month will fail to boot for some reason. An application may work most of the time but occasionally crash. The PC may lock up at seemingly random intervals. Your mouse may work almost all of the time, but one day out of five or ten may give you trouble.
It is helpful to determine if the problem you are experiencing is repeatable, because intermittent difficulties are much more difficult to resolve than repeatable ones. If a problem is repeatable, and there is a specific set of actions that cause the problem, this gives you at least some initial clues about how to find the cause. In addition, you have a way of testing to see if you have resolved the problem when you are trying different solutions. Intermittent problems are much more difficult to deal with.
Determining if a problem is repeatable is pretty simple: try to duplicate the conditions that caused the problem and see if it happens again.
Dealing With Intermittent Problems
Intermittent problems are ones which appear to happen spontaneously or randomly. They seem not to be caused by anything obvious and are not repeatable. They can be extremely difficult and frustrating to diagnose.
Sometimes problems that seem intermittent really aren't; it's just that the specific set of circumstances that causes the problem to occur may be obscure or hard to notice. Spend some time trying to determine what the circumstances are when the problem arises. For example, many problems with crashes or lockups will occur only after the system has been on for more than an hour; some may occur only within the first few minutes that the PC is turned on. You may find that program A only crashes when run at the same time as program B. Or that a particular behavior is associated with a peripheral of some sort.
You will have to be patient when dealing with these sorts of issues. Since the problem is not something you can duplicate at will, you may not be able to systematically work your way towards the ultimate cause. In this situation you may have to employ trial-and-error, making a change and then waiting to see if the problem recurs. It can take days (or longer) sometimes since you have to wait before seeing if the problem happens again. Be patient.
Correlation May Not Imply Causation
I use the word correlation, perhaps slightly incorrectly, to refer to two behaviors or symptoms that appear at the same time. Causation refers to two events where one is responsible for the other's appearance or existence. So what I am saying here is that if you see two strange things happening at the same time on your PC, this does not necessarily mean that one of them has caused the other. They could be coincidental, or it could be that where you think A is causing B, B is really causing A.
For example, you may find two symptoms on your machine: you have frequent file system corruption on your hard disk, and you have a lot of system lockups (hangs, crashes). You may think that the crashes are causing the file system errors, and in many cases this is true. However it is possible for the file system errors to cause the lockups. And it is also possible that both are just symptoms of another underlying cause.
Scan For File System Problems
If the problem you are looking at is anything that involves the hard disk, files, operating system problems, crashes or similar behavior, scan the file system for errors. This is a simple step (which you should be performing regularly as part of a preventive maintenance program) that may tell you immediately that there is a problem with your hard disk. This can also be a useful piece of evidence in troubleshooting some sorts of problems.
Use Diagnostic Tools
If you own diagnostic hardware and/or software tools, they can be useful in pinpointing some kinds of problems. If you feel that you are having a problem that can be detected by a tool, or you feel that running a diagnostic utility may shed more light on what's happening, then make use of the tool to help guide your analysis of the problem.
Next: Diagnostic, Troubleshooting and Repair Tools
Diagnostic, Troubleshooting and Repair Tools
Troubleshooting a problem with a PC can be very difficult and frustrating. Fortunately, there are places to turn for assistance, and there are also tools at your disposal that can make the job much easier. These can be used to identify specific problems within the PC, or just give you more information about the system so you understand better what you are dealing with. This chapter takes a brief look at these tools, which can make your life easier if you are having trouble.
Diagnostic Software
The use of diagnostic software tools can save you a great deal of time. While in my experience these tools do not usually identify what the cause of a system problem is, they often provide hints or at least valuable information about what is in the system and how it is working. Some of these are free or are included with common operating systems at no additional charge, while others are commercial products that range from affordable to rather pricey.
Here are some common software diagnostic tools that you will want to consider including as part of your troubleshooting arsenal:
- Power-On Self Test (POST): I am cheating a bit here because this isn't a separate diagnostic utility; it is in fact built into your system BIOS and it runs every time you start up your PC automatically. It is often the best indicator of system problems; make sure you pay attention to its audio and video messages. Don't disable its error-reporting functions unless you really need to.
- MEM.EXE: This simple utility, built into recent versions of DOS and also Windows 95, provides you with details about your memory configuration, as well as what is currently using your memory. It is especially useful when run with the "/C" parameter (use the "/P" parameter as well to make the output pause when it is scrolling).
- Microsoft Diagnostics: Better known as "MSD.EXE", this is a small DOS utility that takes a brief inventory of the contents of your PC and shows them to you in a text-based format. This is very useful for seeing what disks are in the system, how much memory is installed, and also for checking system resource usage such as LPT ports and IRQs. It will show you what type of BIOS you are using and also what UART chip you have in your serial ports. MSD.EXE is included in later versions of DOS.
Tip: MSD.EXE will work on a Windows 95 machine but is not normally installed as part of the normal Windows 95 installation. You can find it on the Windows 95 CD however and then copy it to the C:\WINDOWS\COMMAND directory to make it work.
Warning: MSD.EXE can give incorrect information when run from within Windows. Shut down to DOS mode before running it ("Restart the computer in MS-DOS mode")
- The Windows 95 Device Manager: This is the most useful tool for identifying system configuration and resource usage information under Windows 95. To access it, open the Control Panel and select the "System" icon. Then select the "Device Manager" tab. You will see a graphical "tree" structure showing you all of your PC hardware. If you select "Properties" while "Computer" (the top-level item) is selected, you will be able to see all the IRQs, DMA channels and I/O addresses in use in your PC; very useful for resolving resource conflicts! The same "Properties" button, pressed after selecting a specific hardware device, will show you driver information, resource settings for the hardware item chosen, and much more.
- Norton System Information: This utility is similar to the Microsoft Diagnostics, only more detailed in its later versions. SI shows a great deal of information about what is in the PC, going well beyond what MSD gives you, but really is still an information utility as opposed to a true diagnostic. This program is part of Symantec's Norton Utilities.
- Microsoft ScanDisk and Norton Disk Doctor: These programs are used to check for hard disk problems. This includes file system corruption and hard disk read errors. They should be used when hard disk problems are suspected.
- Norton Diagnostics: This utility is meant to go beyond the System Information program and actually perform tests on the hardware to identify problems. It includes tests of the processor and motherboard and system memory, and will identify some types of resource conflicts. In reality it is still quite limited in terms of the numbers of problems it will find.
- QAPlus: QAPlus from DiagSoft is a more advanced diagnostic suite that comes in several flavors, depending on what you need to do and how you want to do it. This is a more expensive package but can give you much more detailed information about your system and help identify problem situations as well.
Diagnostic Hardware
While not as commonly used as diagnostic software, there are available several pieces of hardware that can be very helpful in troubleshooting some specific hardware problems. These units are not nearly as popular because they are usually more expensive and are used for specific types of problems, making them less universally applicable than the general software utilities that are widely encountered.
If you are a serious home builder or troubleshooter, you may want to consider one or more of the following:
- Loop-Back Plugs: These are small plugs that go onto the serial and parallel ports on your computer and connect the receive to transmit lines to simulate a connection (they connect the port to itself). Using these with a program like Norton Diagnostics allows for full testing of the operation of these ports, whereas without them only the internal portions of the port can be tested.
Note: The type of plugs used by a particular diagnostic utility may be different than the type used by another program made by another company. Check it out before you buy.
- Multimeter/Ohmmeter: An ohmmeter is a device that measures electrical resistance; a multimeter can measure resistance, voltage and current. These devices are used by electricians, electronics designers and repairpeople. They have come down greatly in price and simple ones can be had inexpensively (while top-end ones of course are still quite a bit of money). A simple ohmmeter is useful primarily for checking for short circuits or open circuits (broken connections, damaged cables, etc.) Multimeters can be used for more extensive electronics testing.
- BIOS POST Cards: It is well-known that the BIOS will, if it finds a problem during its power-on self-test (POST), produce audio "beep codes" and/or video messages that indicate what the trouble is that it found. What many do not realize is that the BIOS of most PCs is designed to send a stream of test codes to a special memory location as it performs its tests, usually 80h. By using a special card designed to capture and display these codes, you can pinpoint exactly where in the power-on self-test a system is hanging up. This can be extremely helpful in debugging very stubborn systems. Scott Wainner has more about one of these cards on his site.
- Test Bed: Some experienced PC repair people keep around an older system that they can use as a test bed for components. It can be very useful to be able to test an unknown device with other components that are known to work, to cut down on the guesswork.
There is also more advanced test hardware available, including devices such as logic probes, oscilloscopes, and many types of specialized component testers. These items are generally very expensive and require specific training to be used properly; they are not for the home PC user or even for most hobbyists. Professional test hardware costs thousands of dollars, which is one reason why doing it yourself is often not an option for many kinds of repairs.
The PC Tool Kit
If you are a very casual PC user you can probably get by hacking around with an old screwdriver and a pocket knife when you need to work on your PC. However, if you plan to do any reasonable amount of work upgrading, building, troubleshooting or repairing PCs, a proper tool kit is a good investment. There are many different types of tool kits available for the computer enthusiast, with tools selected specifically to make the lives of those who work with PCs easier. The right tools can save you a lot of time and help you avoid damage to your equipment (and your hands!)
As with most things, tool kits range widely in size, quality and price. At the low end are $5 kits that are barely any better than the screwdriver and pocket knife. At the high end are kits costing over $100 that include everything but the kitchen sink; these are overkill for most people (including me). Most people can do just fine with a toolkit costing between $15 and $50. These are often available pre-assembled into "genuine artificial leather" carrying cases, sold at most general computer stores.
Here are the items that I would recommend you definitely have in your toolkit; note that some of these don't generally come in pre-assembled toolkits so you will want to add them yourself to the carrying case. Every item in this list is one that I use personally on a regular basis:
- A Good Screwdriver: You want a good screwdriver with magnetic, interchangeable bits and a comfortable handle. At a minimum you should have two different sizes each of Phillips and flat bits. If you've never used a magnetic screwdriver before, you'll wonder how you ever did without one once you do; the magnetic tip will prevent screws falling into the PC, and will let you get access to drives mounted deep in your system case.
- Additional Screwdrivers: An additional Philips screwdriver and an additional flat blade screwdriver, for when you misplace the good one. :^) Actually, it's sometimes useful to have two screwdrivers anyway.
- Needle-Nose Pliers: These are useful for grasping small items and for removing and replacing jumpers on circuit boards.
- Wire Snips: A pair of wire snips for cutting wire and stripping insulation. Some kits include wire strippers for the latter purpose.
- A Small Flashlight: Very useful; the insides of PC boxes are quite dark and there are lots of very small things you will need to see, such as the "pin 1" marking on a connector for example.
- Tweezers: Or even better, a part retriever. A retriever is like a tiny set of retractable claws with a spring-loaded handle. Useful stuff.
- A Large Crowbar: For making adjustments to stubborn hardware. (Just kidding. :^) )
- A Roll of Black Electrical Tape: They never include this in the kits but they should. Used for wrapping wire ends and insulating components.
- A Can of Compressed Air: Very handy for cleaning things hands-free and without using any hazardous liquids.
- A Soft, Lint-Free Cloth: For cleaning the monitor and other components.
The following items I categorize as "optional, but useful". They are tools that are used less frequently, or perhaps only in certain circumstances. They are more for someone who does a fair bit of tinkering or who maintains a number of machines:
- A 3/16" Nut Driver: In fact, tool kits often come with several different sizes of nut drivers (using exchangeable bits), but 3/16" is the really important one, because this is the size of nut most commonly used on PCs. These hexagonal nuts are used as mounting hardware for motherboards and serial and parallel ports. A necessity for anyone who builds their own PCs; if you don't have one you'll be attaching port connectors using a pair of poor-fitting pliers.
- Vice Grips: These are incredibly neat tools that have all sorts of uses.
- Torx Screwdrivers or Bits: These are the star-shaped screwdriver heads that are used to make equipment "tamper-proof". You need these drivers if you want to, well, "tamper".
- An ESD (Electrostatic Discharge) Wrist Strap: This is more of a safety device than a tool; it is used to greatly reduce the chances of static damage to components.
- A Knife: A cutting blade or utility knife of some sort.
The following items are sometimes included in tool kits but in my experience are rarely if ever used:
- Soldering Irons: Essential if you are going to try to do repair of circuit boards, but virtually nobody does this (and with good reason). They are also quite dangerous if used improperly.
- Hex (Allen) Keys: They include these in virtually every commercial tool kit I've ever encountered. I've never seen anything PC related that used them but of course that doesn't mean that there isn't something out there that does...
- Chip Extractors: These are used to remove chips from regular (non-ZIF) sockets, but modern PCs do not generally use these sockets any more. Useful for upgrading much older systems in some cases.
Finally, it is a good idea to accumulate a stockpile of spare parts if you are going to be doing work on PCs. I would recommend the following:
- Screws: I have a "big ol' bag o' screws" of all shapes and sizes that comes in handy when working on various machines.
- Expansion Card Inserts: Save the metal inserts that come out of the back of the case when you put a modem or other card into the PC, as you may need them again later on.
- Drive Faceplates: Similarly, save the plastic faceplates that you remove from the front of cases so you can replace them later if needed.
- Mounting Kits: These sometimes come with retail hard disks and allow you to put a 3.5" drive into a 5.25" bay. Useful when your case has more free 5.25" bays than 3.5" ones.
- Cables: Save any power, IDE, floppy, CD-ROM, or other cables that you accumulate in case you need them later on.
- Keyboard, Mouse, 3.5" Floppy Drive: Keep an extra one of each of these components around to aid in troubleshooting problems by swapping. (Actually, keeping an extra of every major component in the PC is helpful, but the three I mention are here because they are cheap. Most people can't afford to keep an extra hard disk lying around as a spare.)
Keep an Open Mind
Very often the real cause of the problem you are seeing is the last thing that you can possibly think would be it; problems simply are not always cause-and-effect. Keep in mind the famous quote of Sherlock Holmes (Sir Arthur Conan Doyle): "When you have eliminated the impossible, whatever remains, however improbable, must be the truth". It is very common for a symptom to be the result of a problem in a completely different part of the PC, which is why you need to make sure that you fully explore the problem before looking for a solution. In addition, bear in mind that changes to a system usually are the cause of problems that arise shortly thereafter, even if they are in unrelated areas.
Give Yourself Time
Starting a troubleshooting session on a difficult problem one half-hour before you have to go away for the weekend is not a great idea. Since you haven't allowed yourself nearly enough time to do a proper job of diagnosing the situation, you are going to feel rushed. It's likely that you will misdiagnose the problem, have to stop half-way before you are finished, and that you will generally feel uncomfortable during the entire process (even more than usual).
Give yourself a fighting chance to resolve the problem properly by allowing yourself enough time to do the job right.
Be Wary of Fatigue
Problem-solving, which is what troubleshooting is at its core, requires concentration. Concentration is hard to hold when you are feeling tired. If you start to feel fatigued, you may be much better off to call it a night and try again the next day, unless there is a rush to get the problem addressed for some reason. (While you are taking this rest, as well, your subconscious mind may even help to solve the problem for you.)
A fresh look after a good night's rest has been enough to solve a problem that seemed "unresolvable" the night before more than once for me.
If You're Stuck, Try Using Your "Background Processing Capabilities"
An amazing capability of the human brain is that under certain circumstances you can put your own subconscious to work for you, to help you solve a problem without expending any energy at all. I am no expert in the field of psychology, but I have used this technique myself with great success on many occasions.
Have you ever tried to remember something, some detail of a past conversation, or a phone number of an old friend, and felt that you knew it but just couldn't remember the exact words or numbers you were looking for? And then a few hours or a day or two later the answer will "pop into your mind" apparently for no reason, while you are doing something else? This is what I am talking about, and sometimes it can help you to resolve problems with your machine.
Some day, you may have a particularly difficult problem that you have been trying to solve for quite a while. It may seem that the answer is obvious, yet you can't quite grasp it. If this is the case, try just going on and doing something else temporarily. Take your conscious mind off the task at hand and try to do something enjoyable, relaxing or distracting for a few hours or the rest of the night. Let your subconscious work on the problem "in the background".
Then come back the next day and look at the problem again. You may find it much easier to resolve the problem now--I've even had the answer come to me in a matter of minutes after spending the previous entire day trying to figure it out. You may even find the answer popping into your head while doing something completely unrelated.
Two Heads are Better Than One
If you have at your disposal someone who knows about PCs, say a peer in your office or a family member, have them take a look at a difficult problem. Even if you are great at troubleshooting, there will be the occasional problem that for one reason or another, you just won't be able to crack. Sometimes another person who can look at a problem from a fresh perspective (who hasn't been staring at the same thing for days trying to resolve it) will be able to set you on the right path to fixing the problem.
Don't be afraid of asking for help. There is no shame in needing assistance, and it can be an excellent way to learn.
Tip: You may be able to get assistance from others with your PC problem by using The PC Guide Discussion Forums.
It Isn't Personal...
OK, this section hits me very close to home. Before you take umbrage at any of the comments that I make here, I will start by saying that I am one of the worst people I know in doing all of the things that I say in this section one should not do. :^) I have a pretty bad temper and tend to take things too seriously in general, so it's not surprising that I tend to lose it sometimes when I am stuck with a particularly enigmatic situation with a PC.
What I am saying here is that despite the fact that it may seem otherwise when you are feeling frustrating about some sort of difficulty, the PC is not doing whatever it is doing intentionally to annoy you. It is not alive, it has no feelings, no motivations, no agenda. There is a logical reason for everything that a PC does--it is just hard sometimes to understand what it is. It can seem like the PC is "out to get you", and then to start feeling very frustrated and angry. Usually when this happens your ability to actually solve the problem will drop pretty close to zero.
During the throes of particularly frustrating spells, I personally have at various times yelled at PCs, kicked PC case covers, slammed the desk with my fist, crunched up floppy disks in my hands, smashed keyboards, smacked the side of a monitor, and hit the power switch rather hard more than once. As a result of this foolishness, I have on rare occasion damaged a component or two, and even hurt my hands. This is one area where I definitely do not recommend that my readers follow my example. :^)
Try as much as possible to keep your cool when working on your machine. If you feel yourself getting really pissed off, this is likely a good signal that you've been working on the problem too long. Take a break, do something else for a while. A fresh look the next day will be better for both you and the PC.
Steps To Take First When Troubleshooting
There are many different tricks and techniques that I outline in the Troubleshooting Guide, but the ones in this section are the few that I always recommend be tried first before proceeding into a full troubleshooting effort. The reason is simple: they can often solve the problem you are having before you spend a lot of energy looking into it seriously.
Let the Vendor Do the Work!
You may not have to use the troubleshooting information on this site at all. If the system is exhibiting strange behavior, especially if it appears that it has any failed or failing components, and you purchased this system recently, you may want to just return it and let the vendor figure out what the problem is. In many cases this is the best thing to do when your PC is still under warranty, since this also eliminates any chance of you voiding the warranty, and saves you time as well. That's what a warranty is for. See here for information on using your warranty, and repair in general.
The problem with doing this is that you have to incur the hassle of bringing the PC back for service. If the PC was purchased mail order, the problem is even worse. Sometimes looking on the site first is a good idea anyway, to rule out any possible simple causes that you can address yourself.
Scan For Viruses
If you experience any strange behavior on your system, especially either as it is booting, or shortly after it completes, you should always use whatever antivirus software you own to do a full scan of the system before you do anything else. You should use a clean boot floppy if possible.
Do not fall into the "this can't be a virus" trap too easily. Obviously if you turn the power on and nothing happens at all, or if your hard disk won't spin up, or if your system seizes while it is doing the initial BIOS power-on test of your system memory, these are hardware problems that are not caused by a virus. However virtually any other symptom that appears to be a hardware issue can be either a hardware problem or simply a virus that is trying to masquerade as a hardware problem. To remove the complicating factor of always having to wonder "is this a virus?", scan your system before you spend a lot of time chasing ghosts.
Read the Documentation
I never cease to be amazed at how people refuse, steadfastly, to read the documentation that comes with their new hardware or software. They are so excited about their new toy, whatever it is, that they are dying to get it running right away--who has time to start reading? So they start installing whatever-it-is and then have problems. Half the time they end up actually taking more time to get the thing running than they would have if they had just read the instructions in the first place. Meanwhile they often waste a lot of their time and others' asking for technical assistance.
Many, if not most problems are a result of hardware or software that has just been installed into the PC. If you are installing something new, read the documentation that came with the product. Look for files called something like "READ.ME" or "README.1ST" on installation or driver disks, and read them thoroughly before you start. Refer back in the event of trouble. You may be surprised to find that the answer to your dilemma is right there, just waiting for you to read it.
Double-Check Any Recent Changes
Ask yourself: when did the problem I am experiencing first start? If you have just made any change to your system, and are now having a problem with your PC that was not present before the change, the chances are probably 99% that the change is the cause of the problem. This is true even if the problem seems to have absolutely nothing to do with what you changed. If you hunt around, eventually you will find something that links the problem to the change. While it is certainly possible for something to coincidentally stop working at the same time that you make a change, the odds are greatly against this ever happening to you.
Here's an example of something that happened to me, to illustrate what I mean. I opened up my PC to do some work on my system memory. I removed two SIMMs and replaced them with another pair of higher capacity. When I turned on the system next, my floppy disk drive was malfunctioning. Had my floppy drive failed just then? Almost certainly not. As it turned out, my motherboard has the header for the floppy drive cable located directly adjacent to the SIMM slots. While I was working on the memory I had loosened the floppy drive cable.
Sometimes the problems can be even more indirectly related than this. You should in most cases stick to pursuing how the change affected the system, as this is likely to eventually lead you to the cause of the problem.
Simplify, Simplify, Simplify
An absolutely crucial rule to keep in mind when troubleshooting PC system problems is that if there are too many unknowns, it is impossible to determine which one is causing the problem. If you have many possible causes for some difficulty with your system, it can be extremely difficult to narrow down the cause of the problem to any one of them. If you are using a just-installed hard disk with an unfamiliar shareware file utility running on an operating system you just upgraded last week, and now your system has problems, how on earth are you ever going to know what is causing them?
In order to have a fighting chance at figuring out what is going on, you must simplify the situation as much as possible so that it becomes much more obvious what is responsible for the difficulty. This means reducing the number of variables to whatever degree possible. One important way of doing this is undoing or double-checking any changes made to the system.
In addition, I have identified the following items as often being responsible for erratic behavior that can complicate troubleshooting. I would recommend that they be eliminated or temporarily disabled when trying to diagnose a system problem:
- Power Management: Power management is a great idea in theory but in many ways is just "not ready for prime time". Power management routines can cause symptoms that appear to be hardware malfunctions, such as screens that turn off unexpectedly or hard disks that spin down. They also can cause crashes of software that doesn't know how to deal with them. If you want to use power management, turning it off until the problem is resolved is wise.
- Overclocked Hardware and Aggressive BIOS Settings: I do not believe in overclocking. If you insist on doing it, don't be surprised if you have system problems! Scale things back until you can figure out what the problem is. Similarly, if you are "pushing the envelope" in trying to squeeze maximum performance by tuning your BIOS memory timings and other settings very aggressively, try resetting them to more conservative values when troubleshooting.
- Experimental or Beta Software: This software is still in the test process and is likely to have bugs--that is why it is labeled as "beta"! For an end application this is usually no big deal, since any crashes or other problems will be limited to that application and therefore somewhat obvious. Running beta operating systems, drivers or other low-level software however is asking for trouble, and you should try to eliminate these possible sources of confusion when trying to debug your system.
- "Creative" Configurations: The more "unusual" things that you have going on in your system, the more likely that you are going to have a conflict caused by one of these strange pieces of hardware or software. A system that is loaded with unusual utilities, terminate-and-stay-resident programs, an old 8-bit network card salvaged from a 286, etc. will often have more problems than a stock Pentium box with a normal Windows 95 installation. To whatever extent possible, disable these items while troubleshooting. Also try to avoid using unusual low-level software whenever possible.
- Excessive Connections: If the PC is on a network or is connected to a large number or variety of peripheral devices, you may want to try to disconnect those and see if there is any impact on the problem.
In general you want to avoid the unusual or the unknown when troubleshooting. One way to simplify the software environment during diagnosis is to use a boot floppy to "boot clean" and bypass the special drivers and software that you normally load when you boot from your hard disk. You can also use the {F8} key when DOS or Windows 95 are booting to bypass your startup files, basically accomplishing the same thing. You want to be especially wary of software that sits in the background and activates without you specifically telling it to, as this can confound your troubleshooting efforts.
To whatever extent possible, disable as much as you can when trying to figure out a problem. The more funky software utilities, screen savers and cute peripherals you disable now, the more chance you have of finding out which one it is that is causing the problem later on.
General Diagnostic Techniques
This section discusses the rules of thumb that I recommend you employ when you are troubleshooting. These are the general techniques and ideas that can help you solve all sorts of different problems with your PC (and even in other parts of your life...) It often is worth trying these general tips before attempting to find a specific answer to your problem in the Troubleshooting Expert.
Get to the Root Cause: Ask "Why" Five Times
One of the keys to truly solving a problem is to first truly understand it. It is easy to observe that too often, people are quick to jump to solve a problem before they really understand what it is. It is easy to see a symptom and think that you know what is causing it, but surprisingly often, if you take the time to explore more deeply you will find that what you thought was the cause is in fact just another symptom, and that the problem lies much deeper within.
One technique that you can employ is borrowed from Japanese manufacturing management theory. This method is used to help identify the real causes of problems that occur on the manufacturing floor. The goal is not to simply correct the effects of the problem, but to find out the root cause of why the problem is occurring so that we can ensure that it will not happen in the future.
One simple way to do this is called asking why' five times. The idea is that by the time you have asked "why" the fifth time, you will be at the root cause. It isn't always that simple, but the exercise can be surprisingly insightful in helping you figure out what is really going on, and can help you avoid "quick fix" solutions that are really just band-aids and don't resolve anything. It is especially useful for tackling chronic problems that show up over and over again in a system; it is less useful for problems that are unlikely to recur.
Here's an example. Let's suppose your hard disk is having a problem with bad sectors showing up. The knee-jerk reaction to this happening is "the hard disk is bad, replace it". Instead, ask yourself:
- "Why are bad sectors showing up when they didn't before?"
- "The hard disk is being corrupted somehow; something has changed."
- "Why?"
- "The timing of the system probably has been changed in some way."
- "Why?"
- "Hmm, I installed a new hard drive in that system just last week, perhaps it is causing the problem."
- "Why?"
- "The hard drive may not have been installed correctly, or perhaps I should have used a different channel. Or maybe it was the bus-mastering driver that came with the new drive, the old drive may not be able to support it".
You can see the general idea; the answers will differ in every case, but it is the process itself that is useful. Another example: imagine an office that has just had a catastrophic data loss due to a PC crashing after an electrical storm. The initially identified cause of this problem was a lack of a UPS on the PC, which would have protected the system from the electrical storm (usually). So the MIS department starts drafting purchase requisitions for UPSes for all the PCs. Instead, if they looked at the problem carefully, they may have employed the following process:
- "Remember that the problem here was the data loss, not the crashing of the PC. Why was the data lost?"
- "The PC's hard disk crashed."
- "Why would that cause the data to be lost?"
- "Because we didn't have a backup."
- "Why wasn't there a backup?"
- "There is a tape backup unit on every PC in the building but very few of them are being used."
- "Why?"
- "Because most of the PC users do not know how to use them."
- "Why?"
- "They have never been given any training in their use."
- "Why?"
- "Because we don't have a budget for PC hardware training."
As you can see, the root cause that we came up with here is very different than the cause we saw when we only scratched the surface. This means of course that the solution will be very different as well; it might well make a lot more sense here to spend a small amount of money on training instead of a large amount of money on UPSes for every PC.
Be Observant and Look For Evidence
It is sometimes the subtle signs that can provide the triggering piece of information that leads to discovering the source of a problem. In particular, you want to carefully investigate anything that seems "unusual", "wrong" or "surprising". These are often the clues that will get you going down the right track to figuring out what is giving you trouble. If you are observant and careful about what you are doing, you will be much more likely to pick up on these things than if you are in a hurry or you are quick to discount your findings as "unrelated to the problem".
Use the Process of Elimination
Virtually all problems with PCs involve more than one component or subsystem. The difficulty is usually in figuring out which component is responsible for the problem. Using the process of elimination, however, you can usually narrow the problem down rather quickly by making small logical changes and observing the impact on the problem. Your objective is to isolate the cause of the problem so you can correct it.
The key is to make only one change at a time and then see if the problem goes away; if it does, then whatever you changed is likely responsible for the problem (although it could be fixing the problem indirectly in some cases.) If you make more than one change at a time, you cannot readily discern which change was responsible for fixing the problem.
You will want to first check the most probable sources of the problem, and also the things that are easiest to change. For example, if you are having a problem with your disk drive being recognized, it's a lot easier and cheaper to explore things like double-checking jumpers and connections or replacing the interface cable, than it is to try replacing the drive itself. That is something you'd only do after you had eliminated all the other possibilities (or if the evidence implicated the hard disk directly).
Here's a simple example. Let's suppose one morning your PC will not turn on. You hit the switch and nothing happens. There could be many different possible causes for this problem: the power to the house could be out; there could be a malfunction in the wall socket; the surge suppressor that the system is plugged into might have blown; the electrical cord may be loose; the power supply could be damaged. To figure out what is going on you need to eliminate these variables by making small changes and seeing what happens. For example:
- Change the wall socket you are using. If the PC now boots, you have isolated the cause to the electrical wiring in the house.
- If the problem persists, examine the surge suppressor. Change it, or temporarily bypass it and plug the PC into the wall directly. If it now works, the surge suppressor is the problem.
- If the problem still isn't fixed, try changing the power cord.
- If the problem persists still, you may then have to open up the box and look at the power supply unit to see if it might need replacing.
Realize that the key here is making these changes one at a time. If you approach this problem by changing the wall socket you use, bypassing the surge suppressor, and changing the power cord all at once, your problem may go away but how will you know what caused it? This is a valid way to troubleshoot if you have to get the system back up immediately, however. You can then undo the changes one at a time later on to find out what the cause is, in effect doing the same single-change-at-a-time process, but in reverse.
Most problems with the PC can be diagnosed using this sort of process; it's very powerful. You will notice many sequences like the one above in The Troubleshooting Expert.
Do One Upgrade or Assembly Step At a Time
Changes made to the system are the most frequent cause of problems; this is the nature of change. Much the way many problems can be diagnosed by using the process of elimination through making single changes to the system, you can avoid or detect problems with upgrades or new installations by going "one step at a time" in your changes as well. New system installations or major upgrades often can have problems that are very difficult to diagnose simply because there are so many modifications being performed at the same time.
When you build a new PC you will probably be assembling a large number of components that have never been used or tested before. To whatever extent possible, try to go as slowly as possible when assembling the machine. For example, when assembling a new PC from scratch, it is always best to first make sure the basic system is working, with the only expansion card in the system the video card (if applicable). Adding the sound card, network card and other devices should be done later on, and one at a time if possible. Similarly, do not try to do major software or operating system upgrades at the same time that you make hardware changes. Doing this can make it very difficult for you to troubleshoot any system problems.
If you do make multiple changes at once, try retracing your steps. Undo the changes you have made one at a time and see if you can identify the change that caused the problem that way.
Make Use of Components That You Know Work
One valuable resource while troubleshooting is a "stockpile" of extra components. I put "stockpile" in quotes because it doesn't have to be an actual stock of components, just a resource that will let you borrow components that you know work while you are troubleshooting your system. Another functional PC that uses the same or similar components is normally quite adequate. Being in an office environment can be very helpful this way since there are many PCs around to work with, and often they are similar to one another; at home it may be more difficult if you only have the one machine.
These extra components can be useful when you are employing the process of elimination to resolve problems. For example, suppose that there is no video on your monitor when you boot up the PC. After eliminating the simple stuff (cable disconnected, brightness turned down far too much, etc.) there are two components that are most likely to be bad: the video card or the monitor. An easy thing to try is to swap monitors with another machine that is working properly. If the second monitor also doesn't work, and your original monitor works on the other machine, you can feel quite confident that it isn't the monitor that is the problem here.
You can sometimes avoid problems in assembling a new PC by testing out some of the components you are going to use before you begin. For example, take your new mouse, keyboard and monitor and attach them to a PC that is known to work, and test them out. This way if you have problems when you build your new system you will know it isn't due to these components. At my office, I keep an extra keyboard, mouse and monitor just for debugging and assembling PCs. Since I know that they work, I reduce the number of variables that I have to deal with when I experience a problem. Once the main PC box is assembled and working, I connect the new keyboard, mouse and monitor. If a problem shows up then, I know that it is not likely the PC, but the peripherals I just connected.
Write Things Down
Whenever you discern a piece of evidence, or you figure out something about the current situation your PC is in, document it. Keeping a history of what you discover and what your symptoms are will help you not only with this problem but also with any future difficulties. A log book of some sort is a good idea.
Determine Repeatability
Most problems with a PC fall into one of two categories: either they are repeatable or they are intermittent. A repeatable problem is one where the problem occurs all the time, or always in response to a specific user action. For example, a PC that has a problem that prevents it from booting will probably always fail to boot no matter how many times you reset it. Or you may have an application that whenever you try to run, will crash with an error. You may find that your PC hangs, but only when you move the mouse at the same time that you are communicating using your modem.
In contrast, some problems are intermittent and not repeatable. In some cases, you may have a PC that will usually boot up fine, but one day a month will fail to boot for some reason. An application may work most of the time but occasionally crash. The PC may lock up at seemingly random intervals. Your mouse may work almost all of the time, but one day out of five or ten may give you trouble.
It is helpful to determine if the problem you are experiencing is repeatable, because intermittent difficulties are much more difficult to resolve than repeatable ones. If a problem is repeatable, and there is a specific set of actions that cause the problem, this gives you at least some initial clues about how to find the cause. In addition, you have a way of testing to see if you have resolved the problem when you are trying different solutions. Intermittent problems are much more difficult to deal with.
Determining if a problem is repeatable is pretty simple: try to duplicate the conditions that caused the problem and see if it happens again.
Dealing With Intermittent Problems
Intermittent problems are ones which appear to happen spontaneously or randomly. They seem not to be caused by anything obvious and are not repeatable. They can be extremely difficult and frustrating to diagnose.
Sometimes problems that seem intermittent really aren't; it's just that the specific set of circumstances that causes the problem to occur may be obscure or hard to notice. Spend some time trying to determine what the circumstances are when the problem arises. For example, many problems with crashes or lockups will occur only after the system has been on for more than an hour; some may occur only within the first few minutes that the PC is turned on. You may find that program A only crashes when run at the same time as program B. Or that a particular behavior is associated with a peripheral of some sort.
You will have to be patient when dealing with these sorts of issues. Since the problem is not something you can duplicate at will, you may not be able to systematically work your way towards the ultimate cause. In this situation you may have to employ trial-and-error, making a change and then waiting to see if the problem recurs. It can take days (or longer) sometimes since you have to wait before seeing if the problem happens again. Be patient.
Correlation May Not Imply Causation
I use the word correlation, perhaps slightly incorrectly, to refer to two behaviors or symptoms that appear at the same time. Causation refers to two events where one is responsible for the other's appearance or existence. So what I am saying here is that if you see two strange things happening at the same time on your PC, this does not necessarily mean that one of them has caused the other. They could be coincidental, or it could be that where you think A is causing B, B is really causing A.
For example, you may find two symptoms on your machine: you have frequent file system corruption on your hard disk, and you have a lot of system lockups (hangs, crashes). You may think that the crashes are causing the file system errors, and in many cases this is true. However it is possible for the file system errors to cause the lockups. And it is also possible that both are just symptoms of another underlying cause.
Scan For File System Problems
If the problem you are looking at is anything that involves the hard disk, files, operating system problems, crashes or similar behavior, scan the file system for errors. This is a simple step (which you should be performing regularly as part of a preventive maintenance program) that may tell you immediately that there is a problem with your hard disk. This can also be a useful piece of evidence in troubleshooting some sorts of problems.
Use Diagnostic Tools
If you own diagnostic hardware and/or software tools, they can be useful in pinpointing some kinds of problems. If you feel that you are having a problem that can be detected by a tool, or you feel that running a diagnostic utility may shed more light on what's happening, then make use of the tool to help guide your analysis of the problem.
Next: Diagnostic, Troubleshooting and Repair Tools
Diagnostic, Troubleshooting and Repair Tools
Troubleshooting a problem with a PC can be very difficult and frustrating. Fortunately, there are places to turn for assistance, and there are also tools at your disposal that can make the job much easier. These can be used to identify specific problems within the PC, or just give you more information about the system so you understand better what you are dealing with. This chapter takes a brief look at these tools, which can make your life easier if you are having trouble.
Diagnostic Software
The use of diagnostic software tools can save you a great deal of time. While in my experience these tools do not usually identify what the cause of a system problem is, they often provide hints or at least valuable information about what is in the system and how it is working. Some of these are free or are included with common operating systems at no additional charge, while others are commercial products that range from affordable to rather pricey.
Here are some common software diagnostic tools that you will want to consider including as part of your troubleshooting arsenal:
- Power-On Self Test (POST): I am cheating a bit here because this isn't a separate diagnostic utility; it is in fact built into your system BIOS and it runs every time you start up your PC automatically. It is often the best indicator of system problems; make sure you pay attention to its audio and video messages. Don't disable its error-reporting functions unless you really need to.
- MEM.EXE: This simple utility, built into recent versions of DOS and also Windows 95, provides you with details about your memory configuration, as well as what is currently using your memory. It is especially useful when run with the "/C" parameter (use the "/P" parameter as well to make the output pause when it is scrolling).
- Microsoft Diagnostics: Better known as "MSD.EXE", this is a small DOS utility that takes a brief inventory of the contents of your PC and shows them to you in a text-based format. This is very useful for seeing what disks are in the system, how much memory is installed, and also for checking system resource usage such as LPT ports and IRQs. It will show you what type of BIOS you are using and also what UART chip you have in your serial ports. MSD.EXE is included in later versions of DOS.
Tip: MSD.EXE will work on a Windows 95 machine but is not normally installed as part of the normal Windows 95 installation. You can find it on the Windows 95 CD however and then copy it to the C:\WINDOWS\COMMAND directory to make it work.
Warning: MSD.EXE can give incorrect information when run from within Windows. Shut down to DOS mode before running it ("Restart the computer in MS-DOS mode")
- The Windows 95 Device Manager: This is the most useful tool for identifying system configuration and resource usage information under Windows 95. To access it, open the Control Panel and select the "System" icon. Then select the "Device Manager" tab. You will see a graphical "tree" structure showing you all of your PC hardware. If you select "Properties" while "Computer" (the top-level item) is selected, you will be able to see all the IRQs, DMA channels and I/O addresses in use in your PC; very useful for resolving resource conflicts! The same "Properties" button, pressed after selecting a specific hardware device, will show you driver information, resource settings for the hardware item chosen, and much more.
- Norton System Information: This utility is similar to the Microsoft Diagnostics, only more detailed in its later versions. SI shows a great deal of information about what is in the PC, going well beyond what MSD gives you, but really is still an information utility as opposed to a true diagnostic. This program is part of Symantec's Norton Utilities.
- Microsoft ScanDisk and Norton Disk Doctor: These programs are used to check for hard disk problems. This includes file system corruption and hard disk read errors. They should be used when hard disk problems are suspected.
- Norton Diagnostics: This utility is meant to go beyond the System Information program and actually perform tests on the hardware to identify problems. It includes tests of the processor and motherboard and system memory, and will identify some types of resource conflicts. In reality it is still quite limited in terms of the numbers of problems it will find.
- QAPlus: QAPlus from DiagSoft is a more advanced diagnostic suite that comes in several flavors, depending on what you need to do and how you want to do it. This is a more expensive package but can give you much more detailed information about your system and help identify problem situations as well.
Diagnostic Hardware
While not as commonly used as diagnostic software, there are available several pieces of hardware that can be very helpful in troubleshooting some specific hardware problems. These units are not nearly as popular because they are usually more expensive and are used for specific types of problems, making them less universally applicable than the general software utilities that are widely encountered.
If you are a serious home builder or troubleshooter, you may want to consider one or more of the following:
- Loop-Back Plugs: These are small plugs that go onto the serial and parallel ports on your computer and connect the receive to transmit lines to simulate a connection (they connect the port to itself). Using these with a program like Norton Diagnostics allows for full testing of the operation of these ports, whereas without them only the internal portions of the port can be tested.
Note: The type of plugs used by a particular diagnostic utility may be different than the type used by another program made by another company. Check it out before you buy.
- Multimeter/Ohmmeter: An ohmmeter is a device that measures electrical resistance; a multimeter can measure resistance, voltage and current. These devices are used by electricians, electronics designers and repairpeople. They have come down greatly in price and simple ones can be had inexpensively (while top-end ones of course are still quite a bit of money). A simple ohmmeter is useful primarily for checking for short circuits or open circuits (broken connections, damaged cables, etc.) Multimeters can be used for more extensive electronics testing.
- BIOS POST Cards: It is well-known that the BIOS will, if it finds a problem during its power-on self-test (POST), produce audio "beep codes" and/or video messages that indicate what the trouble is that it found. What many do not realize is that the BIOS of most PCs is designed to send a stream of test codes to a special memory location as it performs its tests, usually 80h. By using a special card designed to capture and display these codes, you can pinpoint exactly where in the power-on self-test a system is hanging up. This can be extremely helpful in debugging very stubborn systems. Scott Wainner has more about one of these cards on his site.
- Test Bed: Some experienced PC repair people keep around an older system that they can use as a test bed for components. It can be very useful to be able to test an unknown device with other components that are known to work, to cut down on the guesswork.
There is also more advanced test hardware available, including devices such as logic probes, oscilloscopes, and many types of specialized component testers. These items are generally very expensive and require specific training to be used properly; they are not for the home PC user or even for most hobbyists. Professional test hardware costs thousands of dollars, which is one reason why doing it yourself is often not an option for many kinds of repairs.
The PC Tool Kit
If you are a very casual PC user you can probably get by hacking around with an old screwdriver and a pocket knife when you need to work on your PC. However, if you plan to do any reasonable amount of work upgrading, building, troubleshooting or repairing PCs, a proper tool kit is a good investment. There are many different types of tool kits available for the computer enthusiast, with tools selected specifically to make the lives of those who work with PCs easier. The right tools can save you a lot of time and help you avoid damage to your equipment (and your hands!)
As with most things, tool kits range widely in size, quality and price. At the low end are $5 kits that are barely any better than the screwdriver and pocket knife. At the high end are kits costing over $100 that include everything but the kitchen sink; these are overkill for most people (including me). Most people can do just fine with a toolkit costing between $15 and $50. These are often available pre-assembled into "genuine artificial leather" carrying cases, sold at most general computer stores.
Here are the items that I would recommend you definitely have in your toolkit; note that some of these don't generally come in pre-assembled toolkits so you will want to add them yourself to the carrying case. Every item in this list is one that I use personally on a regular basis:
- A Good Screwdriver: You want a good screwdriver with magnetic, interchangeable bits and a comfortable handle. At a minimum you should have two different sizes each of Phillips and flat bits. If you've never used a magnetic screwdriver before, you'll wonder how you ever did without one once you do; the magnetic tip will prevent screws falling into the PC, and will let you get access to drives mounted deep in your system case.
- Additional Screwdrivers: An additional Philips screwdriver and an additional flat blade screwdriver, for when you misplace the good one. :^) Actually, it's sometimes useful to have two screwdrivers anyway.
- Needle-Nose Pliers: These are useful for grasping small items and for removing and replacing jumpers on circuit boards.
- Wire Snips: A pair of wire snips for cutting wire and stripping insulation. Some kits include wire strippers for the latter purpose.
- A Small Flashlight: Very useful; the insides of PC boxes are quite dark and there are lots of very small things you will need to see, such as the "pin 1" marking on a connector for example.
- Tweezers: Or even better, a part retriever. A retriever is like a tiny set of retractable claws with a spring-loaded handle. Useful stuff.
- A Large Crowbar: For making adjustments to stubborn hardware. (Just kidding. :^) )
- A Roll of Black Electrical Tape: They never include this in the kits but they should. Used for wrapping wire ends and insulating components.
- A Can of Compressed Air: Very handy for cleaning things hands-free and without using any hazardous liquids.
- A Soft, Lint-Free Cloth: For cleaning the monitor and other components.
The following items I categorize as "optional, but useful". They are tools that are used less frequently, or perhaps only in certain circumstances. They are more for someone who does a fair bit of tinkering or who maintains a number of machines:
- A 3/16" Nut Driver: In fact, tool kits often come with several different sizes of nut drivers (using exchangeable bits), but 3/16" is the really important one, because this is the size of nut most commonly used on PCs. These hexagonal nuts are used as mounting hardware for motherboards and serial and parallel ports. A necessity for anyone who builds their own PCs; if you don't have one you'll be attaching port connectors using a pair of poor-fitting pliers.
- Vice Grips: These are incredibly neat tools that have all sorts of uses.
- Torx Screwdrivers or Bits: These are the star-shaped screwdriver heads that are used to make equipment "tamper-proof". You need these drivers if you want to, well, "tamper".
- An ESD (Electrostatic Discharge) Wrist Strap: This is more of a safety device than a tool; it is used to greatly reduce the chances of static damage to components.
- A Knife: A cutting blade or utility knife of some sort.
The following items are sometimes included in tool kits but in my experience are rarely if ever used:
- Soldering Irons: Essential if you are going to try to do repair of circuit boards, but virtually nobody does this (and with good reason). They are also quite dangerous if used improperly.
- Hex (Allen) Keys: They include these in virtually every commercial tool kit I've ever encountered. I've never seen anything PC related that used them but of course that doesn't mean that there isn't something out there that does...
- Chip Extractors: These are used to remove chips from regular (non-ZIF) sockets, but modern PCs do not generally use these sockets any more. Useful for upgrading much older systems in some cases.
Finally, it is a good idea to accumulate a stockpile of spare parts if you are going to be doing work on PCs. I would recommend the following:
- Screws: I have a "big ol' bag o' screws" of all shapes and sizes that comes in handy when working on various machines.
- Expansion Card Inserts: Save the metal inserts that come out of the back of the case when you put a modem or other card into the PC, as you may need them again later on.
- Drive Faceplates: Similarly, save the plastic faceplates that you remove from the front of cases so you can replace them later if needed.
- Mounting Kits: These sometimes come with retail hard disks and allow you to put a 3.5" drive into a 5.25" bay. Useful when your case has more free 5.25" bays than 3.5" ones.
- Cables: Save any power, IDE, floppy, CD-ROM, or other cables that you accumulate in case you need them later on.
- Keyboard, Mouse, 3.5" Floppy Drive: Keep an extra one of each of these components around to aid in troubleshooting problems by swapping. (Actually, keeping an extra of every major component in the PC is helpful, but the three I mention are here because they are cheap. Most people can't afford to keep an extra hard disk lying around as a spare.)
Be Wary of Fatigue
Problem-solving, which is what troubleshooting is at its core, requires concentration. Concentration is hard to hold when you are feeling tired. If you start to feel fatigued, you may be much better off to call it a night and try again the next day, unless there is a rush to get the problem addressed for some reason. (While you are taking this rest, as well, your subconscious mind may even help to solve the problem for you.)
A fresh look after a good night's rest has been enough to solve a problem that seemed "unresolvable" the night before more than once for me.
If You're Stuck, Try Using Your "Background Processing Capabilities"
An amazing capability of the human brain is that under certain circumstances you can put your own subconscious to work for you, to help you solve a problem without expending any energy at all. I am no expert in the field of psychology, but I have used this technique myself with great success on many occasions.
Have you ever tried to remember something, some detail of a past conversation, or a phone number of an old friend, and felt that you knew it but just couldn't remember the exact words or numbers you were looking for? And then a few hours or a day or two later the answer will "pop into your mind" apparently for no reason, while you are doing something else? This is what I am talking about, and sometimes it can help you to resolve problems with your machine.
Some day, you may have a particularly difficult problem that you have been trying to solve for quite a while. It may seem that the answer is obvious, yet you can't quite grasp it. If this is the case, try just going on and doing something else temporarily. Take your conscious mind off the task at hand and try to do something enjoyable, relaxing or distracting for a few hours or the rest of the night. Let your subconscious work on the problem "in the background".
Then come back the next day and look at the problem again. You may find it much easier to resolve the problem now--I've even had the answer come to me in a matter of minutes after spending the previous entire day trying to figure it out. You may even find the answer popping into your head while doing something completely unrelated.
Two Heads are Better Than One
If you have at your disposal someone who knows about PCs, say a peer in your office or a family member, have them take a look at a difficult problem. Even if you are great at troubleshooting, there will be the occasional problem that for one reason or another, you just won't be able to crack. Sometimes another person who can look at a problem from a fresh perspective (who hasn't been staring at the same thing for days trying to resolve it) will be able to set you on the right path to fixing the problem.
Don't be afraid of asking for help. There is no shame in needing assistance, and it can be an excellent way to learn.
Tip: You may be able to get assistance from others with your PC problem by using The PC Guide Discussion Forums.
It Isn't Personal...
OK, this section hits me very close to home. Before you take umbrage at any of the comments that I make here, I will start by saying that I am one of the worst people I know in doing all of the things that I say in this section one should not do. :^) I have a pretty bad temper and tend to take things too seriously in general, so it's not surprising that I tend to lose it sometimes when I am stuck with a particularly enigmatic situation with a PC.
What I am saying here is that despite the fact that it may seem otherwise when you are feeling frustrating about some sort of difficulty, the PC is not doing whatever it is doing intentionally to annoy you. It is not alive, it has no feelings, no motivations, no agenda. There is a logical reason for everything that a PC does--it is just hard sometimes to understand what it is. It can seem like the PC is "out to get you", and then to start feeling very frustrated and angry. Usually when this happens your ability to actually solve the problem will drop pretty close to zero.
During the throes of particularly frustrating spells, I personally have at various times yelled at PCs, kicked PC case covers, slammed the desk with my fist, crunched up floppy disks in my hands, smashed keyboards, smacked the side of a monitor, and hit the power switch rather hard more than once. As a result of this foolishness, I have on rare occasion damaged a component or two, and even hurt my hands. This is one area where I definitely do not recommend that my readers follow my example. :^)
Try as much as possible to keep your cool when working on your machine. If you feel yourself getting really pissed off, this is likely a good signal that you've been working on the problem too long. Take a break, do something else for a while. A fresh look the next day will be better for both you and the PC.
Steps To Take First When Troubleshooting
There are many different tricks and techniques that I outline in the Troubleshooting Guide, but the ones in this section are the few that I always recommend be tried first before proceeding into a full troubleshooting effort. The reason is simple: they can often solve the problem you are having before you spend a lot of energy looking into it seriously.
Let the Vendor Do the Work!
You may not have to use the troubleshooting information on this site at all. If the system is exhibiting strange behavior, especially if it appears that it has any failed or failing components, and you purchased this system recently, you may want to just return it and let the vendor figure out what the problem is. In many cases this is the best thing to do when your PC is still under warranty, since this also eliminates any chance of you voiding the warranty, and saves you time as well. That's what a warranty is for. See here for information on using your warranty, and repair in general.
The problem with doing this is that you have to incur the hassle of bringing the PC back for service. If the PC was purchased mail order, the problem is even worse. Sometimes looking on the site first is a good idea anyway, to rule out any possible simple causes that you can address yourself.
Scan For Viruses
If you experience any strange behavior on your system, especially either as it is booting, or shortly after it completes, you should always use whatever antivirus software you own to do a full scan of the system before you do anything else. You should use a clean boot floppy if possible.
Do not fall into the "this can't be a virus" trap too easily. Obviously if you turn the power on and nothing happens at all, or if your hard disk won't spin up, or if your system seizes while it is doing the initial BIOS power-on test of your system memory, these are hardware problems that are not caused by a virus. However virtually any other symptom that appears to be a hardware issue can be either a hardware problem or simply a virus that is trying to masquerade as a hardware problem. To remove the complicating factor of always having to wonder "is this a virus?", scan your system before you spend a lot of time chasing ghosts.
Read the Documentation
I never cease to be amazed at how people refuse, steadfastly, to read the documentation that comes with their new hardware or software. They are so excited about their new toy, whatever it is, that they are dying to get it running right away--who has time to start reading? So they start installing whatever-it-is and then have problems. Half the time they end up actually taking more time to get the thing running than they would have if they had just read the instructions in the first place. Meanwhile they often waste a lot of their time and others' asking for technical assistance.
Many, if not most problems are a result of hardware or software that has just been installed into the PC. If you are installing something new, read the documentation that came with the product. Look for files called something like "READ.ME" or "README.1ST" on installation or driver disks, and read them thoroughly before you start. Refer back in the event of trouble. You may be surprised to find that the answer to your dilemma is right there, just waiting for you to read it.
Double-Check Any Recent Changes
Ask yourself: when did the problem I am experiencing first start? If you have just made any change to your system, and are now having a problem with your PC that was not present before the change, the chances are probably 99% that the change is the cause of the problem. This is true even if the problem seems to have absolutely nothing to do with what you changed. If you hunt around, eventually you will find something that links the problem to the change. While it is certainly possible for something to coincidentally stop working at the same time that you make a change, the odds are greatly against this ever happening to you.
Here's an example of something that happened to me, to illustrate what I mean. I opened up my PC to do some work on my system memory. I removed two SIMMs and replaced them with another pair of higher capacity. When I turned on the system next, my floppy disk drive was malfunctioning. Had my floppy drive failed just then? Almost certainly not. As it turned out, my motherboard has the header for the floppy drive cable located directly adjacent to the SIMM slots. While I was working on the memory I had loosened the floppy drive cable.
Sometimes the problems can be even more indirectly related than this. You should in most cases stick to pursuing how the change affected the system, as this is likely to eventually lead you to the cause of the problem.
Simplify, Simplify, Simplify
An absolutely crucial rule to keep in mind when troubleshooting PC system problems is that if there are too many unknowns, it is impossible to determine which one is causing the problem. If you have many possible causes for some difficulty with your system, it can be extremely difficult to narrow down the cause of the problem to any one of them. If you are using a just-installed hard disk with an unfamiliar shareware file utility running on an operating system you just upgraded last week, and now your system has problems, how on earth are you ever going to know what is causing them?
In order to have a fighting chance at figuring out what is going on, you must simplify the situation as much as possible so that it becomes much more obvious what is responsible for the difficulty. This means reducing the number of variables to whatever degree possible. One important way of doing this is undoing or double-checking any changes made to the system.
In addition, I have identified the following items as often being responsible for erratic behavior that can complicate troubleshooting. I would recommend that they be eliminated or temporarily disabled when trying to diagnose a system problem:
- Power Management: Power management is a great idea in theory but in many ways is just "not ready for prime time". Power management routines can cause symptoms that appear to be hardware malfunctions, such as screens that turn off unexpectedly or hard disks that spin down. They also can cause crashes of software that doesn't know how to deal with them. If you want to use power management, turning it off until the problem is resolved is wise.
- Overclocked Hardware and Aggressive BIOS Settings: I do not believe in overclocking. If you insist on doing it, don't be surprised if you have system problems! Scale things back until you can figure out what the problem is. Similarly, if you are "pushing the envelope" in trying to squeeze maximum performance by tuning your BIOS memory timings and other settings very aggressively, try resetting them to more conservative values when troubleshooting.
- Experimental or Beta Software: This software is still in the test process and is likely to have bugs--that is why it is labeled as "beta"! For an end application this is usually no big deal, since any crashes or other problems will be limited to that application and therefore somewhat obvious. Running beta operating systems, drivers or other low-level software however is asking for trouble, and you should try to eliminate these possible sources of confusion when trying to debug your system.
- "Creative" Configurations: The more "unusual" things that you have going on in your system, the more likely that you are going to have a conflict caused by one of these strange pieces of hardware or software. A system that is loaded with unusual utilities, terminate-and-stay-resident programs, an old 8-bit network card salvaged from a 286, etc. will often have more problems than a stock Pentium box with a normal Windows 95 installation. To whatever extent possible, disable these items while troubleshooting. Also try to avoid using unusual low-level software whenever possible.
- Excessive Connections: If the PC is on a network or is connected to a large number or variety of peripheral devices, you may want to try to disconnect those and see if there is any impact on the problem.
In general you want to avoid the unusual or the unknown when troubleshooting. One way to simplify the software environment during diagnosis is to use a boot floppy to "boot clean" and bypass the special drivers and software that you normally load when you boot from your hard disk. You can also use the {F8} key when DOS or Windows 95 are booting to bypass your startup files, basically accomplishing the same thing. You want to be especially wary of software that sits in the background and activates without you specifically telling it to, as this can confound your troubleshooting efforts.
To whatever extent possible, disable as much as you can when trying to figure out a problem. The more funky software utilities, screen savers and cute peripherals you disable now, the more chance you have of finding out which one it is that is causing the problem later on.
General Diagnostic Techniques
This section discusses the rules of thumb that I recommend you employ when you are troubleshooting. These are the general techniques and ideas that can help you solve all sorts of different problems with your PC (and even in other parts of your life...) It often is worth trying these general tips before attempting to find a specific answer to your problem in the Troubleshooting Expert.
Get to the Root Cause: Ask "Why" Five Times
One of the keys to truly solving a problem is to first truly understand it. It is easy to observe that too often, people are quick to jump to solve a problem before they really understand what it is. It is easy to see a symptom and think that you know what is causing it, but surprisingly often, if you take the time to explore more deeply you will find that what you thought was the cause is in fact just another symptom, and that the problem lies much deeper within.
One technique that you can employ is borrowed from Japanese manufacturing management theory. This method is used to help identify the real causes of problems that occur on the manufacturing floor. The goal is not to simply correct the effects of the problem, but to find out the root cause of why the problem is occurring so that we can ensure that it will not happen in the future.
One simple way to do this is called asking why' five times. The idea is that by the time you have asked "why" the fifth time, you will be at the root cause. It isn't always that simple, but the exercise can be surprisingly insightful in helping you figure out what is really going on, and can help you avoid "quick fix" solutions that are really just band-aids and don't resolve anything. It is especially useful for tackling chronic problems that show up over and over again in a system; it is less useful for problems that are unlikely to recur.
Here's an example. Let's suppose your hard disk is having a problem with bad sectors showing up. The knee-jerk reaction to this happening is "the hard disk is bad, replace it". Instead, ask yourself:
- "Why are bad sectors showing up when they didn't before?"
- "The hard disk is being corrupted somehow; something has changed."
- "Why?"
- "The timing of the system probably has been changed in some way."
- "Why?"
- "Hmm, I installed a new hard drive in that system just last week, perhaps it is causing the problem."
- "Why?"
- "The hard drive may not have been installed correctly, or perhaps I should have used a different channel. Or maybe it was the bus-mastering driver that came with the new drive, the old drive may not be able to support it".
You can see the general idea; the answers will differ in every case, but it is the process itself that is useful. Another example: imagine an office that has just had a catastrophic data loss due to a PC crashing after an electrical storm. The initially identified cause of this problem was a lack of a UPS on the PC, which would have protected the system from the electrical storm (usually). So the MIS department starts drafting purchase requisitions for UPSes for all the PCs. Instead, if they looked at the problem carefully, they may have employed the following process:
- "Remember that the problem here was the data loss, not the crashing of the PC. Why was the data lost?"
- "The PC's hard disk crashed."
- "Why would that cause the data to be lost?"
- "Because we didn't have a backup."
- "Why wasn't there a backup?"
- "There is a tape backup unit on every PC in the building but very few of them are being used."
- "Why?"
- "Because most of the PC users do not know how to use them."
- "Why?"
- "They have never been given any training in their use."
- "Why?"
- "Because we don't have a budget for PC hardware training."
As you can see, the root cause that we came up with here is very different than the cause we saw when we only scratched the surface. This means of course that the solution will be very different as well; it might well make a lot more sense here to spend a small amount of money on training instead of a large amount of money on UPSes for every PC.
Be Observant and Look For Evidence
It is sometimes the subtle signs that can provide the triggering piece of information that leads to discovering the source of a problem. In particular, you want to carefully investigate anything that seems "unusual", "wrong" or "surprising". These are often the clues that will get you going down the right track to figuring out what is giving you trouble. If you are observant and careful about what you are doing, you will be much more likely to pick up on these things than if you are in a hurry or you are quick to discount your findings as "unrelated to the problem".
Use the Process of Elimination
Virtually all problems with PCs involve more than one component or subsystem. The difficulty is usually in figuring out which component is responsible for the problem. Using the process of elimination, however, you can usually narrow the problem down rather quickly by making small logical changes and observing the impact on the problem. Your objective is to isolate the cause of the problem so you can correct it.
The key is to make only one change at a time and then see if the problem goes away; if it does, then whatever you changed is likely responsible for the problem (although it could be fixing the problem indirectly in some cases.) If you make more than one change at a time, you cannot readily discern which change was responsible for fixing the problem.
You will want to first check the most probable sources of the problem, and also the things that are easiest to change. For example, if you are having a problem with your disk drive being recognized, it's a lot easier and cheaper to explore things like double-checking jumpers and connections or replacing the interface cable, than it is to try replacing the drive itself. That is something you'd only do after you had eliminated all the other possibilities (or if the evidence implicated the hard disk directly).
Here's a simple example. Let's suppose one morning your PC will not turn on. You hit the switch and nothing happens. There could be many different possible causes for this problem: the power to the house could be out; there could be a malfunction in the wall socket; the surge suppressor that the system is plugged into might have blown; the electrical cord may be loose; the power supply could be damaged. To figure out what is going on you need to eliminate these variables by making small changes and seeing what happens. For example:
- Change the wall socket you are using. If the PC now boots, you have isolated the cause to the electrical wiring in the house.
- If the problem persists, examine the surge suppressor. Change it, or temporarily bypass it and plug the PC into the wall directly. If it now works, the surge suppressor is the problem.
- If the problem still isn't fixed, try changing the power cord.
- If the problem persists still, you may then have to open up the box and look at the power supply unit to see if it might need replacing.
Realize that the key here is making these changes one at a time. If you approach this problem by changing the wall socket you use, bypassing the surge suppressor, and changing the power cord all at once, your problem may go away but how will you know what caused it? This is a valid way to troubleshoot if you have to get the system back up immediately, however. You can then undo the changes one at a time later on to find out what the cause is, in effect doing the same single-change-at-a-time process, but in reverse.
Most problems with the PC can be diagnosed using this sort of process; it's very powerful. You will notice many sequences like the one above in The Troubleshooting Expert.
Do One Upgrade or Assembly Step At a Time
Changes made to the system are the most frequent cause of problems; this is the nature of change. Much the way many problems can be diagnosed by using the process of elimination through making single changes to the system, you can avoid or detect problems with upgrades or new installations by going "one step at a time" in your changes as well. New system installations or major upgrades often can have problems that are very difficult to diagnose simply because there are so many modifications being performed at the same time.
When you build a new PC you will probably be assembling a large number of components that have never been used or tested before. To whatever extent possible, try to go as slowly as possible when assembling the machine. For example, when assembling a new PC from scratch, it is always best to first make sure the basic system is working, with the only expansion card in the system the video card (if applicable). Adding the sound card, network card and other devices should be done later on, and one at a time if possible. Similarly, do not try to do major software or operating system upgrades at the same time that you make hardware changes. Doing this can make it very difficult for you to troubleshoot any system problems.
If you do make multiple changes at once, try retracing your steps. Undo the changes you have made one at a time and see if you can identify the change that caused the problem that way.
Make Use of Components That You Know Work
One valuable resource while troubleshooting is a "stockpile" of extra components. I put "stockpile" in quotes because it doesn't have to be an actual stock of components, just a resource that will let you borrow components that you know work while you are troubleshooting your system. Another functional PC that uses the same or similar components is normally quite adequate. Being in an office environment can be very helpful this way since there are many PCs around to work with, and often they are similar to one another; at home it may be more difficult if you only have the one machine.
These extra components can be useful when you are employing the process of elimination to resolve problems. For example, suppose that there is no video on your monitor when you boot up the PC. After eliminating the simple stuff (cable disconnected, brightness turned down far too much, etc.) there are two components that are most likely to be bad: the video card or the monitor. An easy thing to try is to swap monitors with another machine that is working properly. If the second monitor also doesn't work, and your original monitor works on the other machine, you can feel quite confident that it isn't the monitor that is the problem here.
You can sometimes avoid problems in assembling a new PC by testing out some of the components you are going to use before you begin. For example, take your new mouse, keyboard and monitor and attach them to a PC that is known to work, and test them out. This way if you have problems when you build your new system you will know it isn't due to these components. At my office, I keep an extra keyboard, mouse and monitor just for debugging and assembling PCs. Since I know that they work, I reduce the number of variables that I have to deal with when I experience a problem. Once the main PC box is assembled and working, I connect the new keyboard, mouse and monitor. If a problem shows up then, I know that it is not likely the PC, but the peripherals I just connected.
Write Things Down
Whenever you discern a piece of evidence, or you figure out something about the current situation your PC is in, document it. Keeping a history of what you discover and what your symptoms are will help you not only with this problem but also with any future difficulties. A log book of some sort is a good idea.
Determine Repeatability
Most problems with a PC fall into one of two categories: either they are repeatable or they are intermittent. A repeatable problem is one where the problem occurs all the time, or always in response to a specific user action. For example, a PC that has a problem that prevents it from booting will probably always fail to boot no matter how many times you reset it. Or you may have an application that whenever you try to run, will crash with an error. You may find that your PC hangs, but only when you move the mouse at the same time that you are communicating using your modem.
In contrast, some problems are intermittent and not repeatable. In some cases, you may have a PC that will usually boot up fine, but one day a month will fail to boot for some reason. An application may work most of the time but occasionally crash. The PC may lock up at seemingly random intervals. Your mouse may work almost all of the time, but one day out of five or ten may give you trouble.
It is helpful to determine if the problem you are experiencing is repeatable, because intermittent difficulties are much more difficult to resolve than repeatable ones. If a problem is repeatable, and there is a specific set of actions that cause the problem, this gives you at least some initial clues about how to find the cause. In addition, you have a way of testing to see if you have resolved the problem when you are trying different solutions. Intermittent problems are much more difficult to deal with.
Determining if a problem is repeatable is pretty simple: try to duplicate the conditions that caused the problem and see if it happens again.
Dealing With Intermittent Problems
Intermittent problems are ones which appear to happen spontaneously or randomly. They seem not to be caused by anything obvious and are not repeatable. They can be extremely difficult and frustrating to diagnose.
Sometimes problems that seem intermittent really aren't; it's just that the specific set of circumstances that causes the problem to occur may be obscure or hard to notice. Spend some time trying to determine what the circumstances are when the problem arises. For example, many problems with crashes or lockups will occur only after the system has been on for more than an hour; some may occur only within the first few minutes that the PC is turned on. You may find that program A only crashes when run at the same time as program B. Or that a particular behavior is associated with a peripheral of some sort.
You will have to be patient when dealing with these sorts of issues. Since the problem is not something you can duplicate at will, you may not be able to systematically work your way towards the ultimate cause. In this situation you may have to employ trial-and-error, making a change and then waiting to see if the problem recurs. It can take days (or longer) sometimes since you have to wait before seeing if the problem happens again. Be patient.
Correlation May Not Imply Causation
I use the word correlation, perhaps slightly incorrectly, to refer to two behaviors or symptoms that appear at the same time. Causation refers to two events where one is responsible for the other's appearance or existence. So what I am saying here is that if you see two strange things happening at the same time on your PC, this does not necessarily mean that one of them has caused the other. They could be coincidental, or it could be that where you think A is causing B, B is really causing A.
For example, you may find two symptoms on your machine: you have frequent file system corruption on your hard disk, and you have a lot of system lockups (hangs, crashes). You may think that the crashes are causing the file system errors, and in many cases this is true. However it is possible for the file system errors to cause the lockups. And it is also possible that both are just symptoms of another underlying cause.
Scan For File System Problems
If the problem you are looking at is anything that involves the hard disk, files, operating system problems, crashes or similar behavior, scan the file system for errors. This is a simple step (which you should be performing regularly as part of a preventive maintenance program) that may tell you immediately that there is a problem with your hard disk. This can also be a useful piece of evidence in troubleshooting some sorts of problems.
Use Diagnostic Tools
If you own diagnostic hardware and/or software tools, they can be useful in pinpointing some kinds of problems. If you feel that you are having a problem that can be detected by a tool, or you feel that running a diagnostic utility may shed more light on what's happening, then make use of the tool to help guide your analysis of the problem.
Next: Diagnostic, Troubleshooting and Repair Tools
Diagnostic, Troubleshooting and Repair Tools
Troubleshooting a problem with a PC can be very difficult and frustrating. Fortunately, there are places to turn for assistance, and there are also tools at your disposal that can make the job much easier. These can be used to identify specific problems within the PC, or just give you more information about the system so you understand better what you are dealing with. This chapter takes a brief look at these tools, which can make your life easier if you are having trouble.
Diagnostic Software
The use of diagnostic software tools can save you a great deal of time. While in my experience these tools do not usually identify what the cause of a system problem is, they often provide hints or at least valuable information about what is in the system and how it is working. Some of these are free or are included with common operating systems at no additional charge, while others are commercial products that range from affordable to rather pricey.
Here are some common software diagnostic tools that you will want to consider including as part of your troubleshooting arsenal:
- Power-On Self Test (POST): I am cheating a bit here because this isn't a separate diagnostic utility; it is in fact built into your system BIOS and it runs every time you start up your PC automatically. It is often the best indicator of system problems; make sure you pay attention to its audio and video messages. Don't disable its error-reporting functions unless you really need to.
- MEM.EXE: This simple utility, built into recent versions of DOS and also Windows 95, provides you with details about your memory configuration, as well as what is currently using your memory. It is especially useful when run with the "/C" parameter (use the "/P" parameter as well to make the output pause when it is scrolling).
- Microsoft Diagnostics: Better known as "MSD.EXE", this is a small DOS utility that takes a brief inventory of the contents of your PC and shows them to you in a text-based format. This is very useful for seeing what disks are in the system, how much memory is installed, and also for checking system resource usage such as LPT ports and IRQs. It will show you what type of BIOS you are using and also what UART chip you have in your serial ports. MSD.EXE is included in later versions of DOS.
Tip: MSD.EXE will work on a Windows 95 machine but is not normally installed as part of the normal Windows 95 installation. You can find it on the Windows 95 CD however and then copy it to the C:\WINDOWS\COMMAND directory to make it work.
Warning: MSD.EXE can give incorrect information when run from within Windows. Shut down to DOS mode before running it ("Restart the computer in MS-DOS mode")
- The Windows 95 Device Manager: This is the most useful tool for identifying system configuration and resource usage information under Windows 95. To access it, open the Control Panel and select the "System" icon. Then select the "Device Manager" tab. You will see a graphical "tree" structure showing you all of your PC hardware. If you select "Properties" while "Computer" (the top-level item) is selected, you will be able to see all the IRQs, DMA channels and I/O addresses in use in your PC; very useful for resolving resource conflicts! The same "Properties" button, pressed after selecting a specific hardware device, will show you driver information, resource settings for the hardware item chosen, and much more.
- Norton System Information: This utility is similar to the Microsoft Diagnostics, only more detailed in its later versions. SI shows a great deal of information about what is in the PC, going well beyond what MSD gives you, but really is still an information utility as opposed to a true diagnostic. This program is part of Symantec's Norton Utilities.
- Microsoft ScanDisk and Norton Disk Doctor: These programs are used to check for hard disk problems. This includes file system corruption and hard disk read errors. They should be used when hard disk problems are suspected.
- Norton Diagnostics: This utility is meant to go beyond the System Information program and actually perform tests on the hardware to identify problems. It includes tests of the processor and motherboard and system memory, and will identify some types of resource conflicts. In reality it is still quite limited in terms of the numbers of problems it will find.
- QAPlus: QAPlus from DiagSoft is a more advanced diagnostic suite that comes in several flavors, depending on what you need to do and how you want to do it. This is a more expensive package but can give you much more detailed information about your system and help identify problem situations as well.
Diagnostic Hardware
While not as commonly used as diagnostic software, there are available several pieces of hardware that can be very helpful in troubleshooting some specific hardware problems. These units are not nearly as popular because they are usually more expensive and are used for specific types of problems, making them less universally applicable than the general software utilities that are widely encountered.
If you are a serious home builder or troubleshooter, you may want to consider one or more of the following:
- Loop-Back Plugs: These are small plugs that go onto the serial and parallel ports on your computer and connect the receive to transmit lines to simulate a connection (they connect the port to itself). Using these with a program like Norton Diagnostics allows for full testing of the operation of these ports, whereas without them only the internal portions of the port can be tested.
Note: The type of plugs used by a particular diagnostic utility may be different than the type used by another program made by another company. Check it out before you buy.
- Multimeter/Ohmmeter: An ohmmeter is a device that measures electrical resistance; a multimeter can measure resistance, voltage and current. These devices are used by electricians, electronics designers and repairpeople. They have come down greatly in price and simple ones can be had inexpensively (while top-end ones of course are still quite a bit of money). A simple ohmmeter is useful primarily for checking for short circuits or open circuits (broken connections, damaged cables, etc.) Multimeters can be used for more extensive electronics testing.
- BIOS POST Cards: It is well-known that the BIOS will, if it finds a problem during its power-on self-test (POST), produce audio "beep codes" and/or video messages that indicate what the trouble is that it found. What many do not realize is that the BIOS of most PCs is designed to send a stream of test codes to a special memory location as it performs its tests, usually 80h. By using a special card designed to capture and display these codes, you can pinpoint exactly where in the power-on self-test a system is hanging up. This can be extremely helpful in debugging very stubborn systems. Scott Wainner has more about one of these cards on his site.
- Test Bed: Some experienced PC repair people keep around an older system that they can use as a test bed for components. It can be very useful to be able to test an unknown device with other components that are known to work, to cut down on the guesswork.
There is also more advanced test hardware available, including devices such as logic probes, oscilloscopes, and many types of specialized component testers. These items are generally very expensive and require specific training to be used properly; they are not for the home PC user or even for most hobbyists. Professional test hardware costs thousands of dollars, which is one reason why doing it yourself is often not an option for many kinds of repairs.
The PC Tool Kit
If you are a very casual PC user you can probably get by hacking around with an old screwdriver and a pocket knife when you need to work on your PC. However, if you plan to do any reasonable amount of work upgrading, building, troubleshooting or repairing PCs, a proper tool kit is a good investment. There are many different types of tool kits available for the computer enthusiast, with tools selected specifically to make the lives of those who work with PCs easier. The right tools can save you a lot of time and help you avoid damage to your equipment (and your hands!)
As with most things, tool kits range widely in size, quality and price. At the low end are $5 kits that are barely any better than the screwdriver and pocket knife. At the high end are kits costing over $100 that include everything but the kitchen sink; these are overkill for most people (including me). Most people can do just fine with a toolkit costing between $15 and $50. These are often available pre-assembled into "genuine artificial leather" carrying cases, sold at most general computer stores.
Here are the items that I would recommend you definitely have in your toolkit; note that some of these don't generally come in pre-assembled toolkits so you will want to add them yourself to the carrying case. Every item in this list is one that I use personally on a regular basis:
- A Good Screwdriver: You want a good screwdriver with magnetic, interchangeable bits and a comfortable handle. At a minimum you should have two different sizes each of Phillips and flat bits. If you've never used a magnetic screwdriver before, you'll wonder how you ever did without one once you do; the magnetic tip will prevent screws falling into the PC, and will let you get access to drives mounted deep in your system case.
- Additional Screwdrivers: An additional Philips screwdriver and an additional flat blade screwdriver, for when you misplace the good one. :^) Actually, it's sometimes useful to have two screwdrivers anyway.
- Needle-Nose Pliers: These are useful for grasping small items and for removing and replacing jumpers on circuit boards.
- Wire Snips: A pair of wire snips for cutting wire and stripping insulation. Some kits include wire strippers for the latter purpose.
- A Small Flashlight: Very useful; the insides of PC boxes are quite dark and there are lots of very small things you will need to see, such as the "pin 1" marking on a connector for example.
- Tweezers: Or even better, a part retriever. A retriever is like a tiny set of retractable claws with a spring-loaded handle. Useful stuff.
- A Large Crowbar: For making adjustments to stubborn hardware. (Just kidding. :^) )
- A Roll of Black Electrical Tape: They never include this in the kits but they should. Used for wrapping wire ends and insulating components.
- A Can of Compressed Air: Very handy for cleaning things hands-free and without using any hazardous liquids.
- A Soft, Lint-Free Cloth: For cleaning the monitor and other components.
The following items I categorize as "optional, but useful". They are tools that are used less frequently, or perhaps only in certain circumstances. They are more for someone who does a fair bit of tinkering or who maintains a number of machines:
- A 3/16" Nut Driver: In fact, tool kits often come with several different sizes of nut drivers (using exchangeable bits), but 3/16" is the really important one, because this is the size of nut most commonly used on PCs. These hexagonal nuts are used as mounting hardware for motherboards and serial and parallel ports. A necessity for anyone who builds their own PCs; if you don't have one you'll be attaching port connectors using a pair of poor-fitting pliers.
- Vice Grips: These are incredibly neat tools that have all sorts of uses.
- Torx Screwdrivers or Bits: These are the star-shaped screwdriver heads that are used to make equipment "tamper-proof". You need these drivers if you want to, well, "tamper".
- An ESD (Electrostatic Discharge) Wrist Strap: This is more of a safety device than a tool; it is used to greatly reduce the chances of static damage to components.
- A Knife: A cutting blade or utility knife of some sort.
The following items are sometimes included in tool kits but in my experience are rarely if ever used:
- Soldering Irons: Essential if you are going to try to do repair of circuit boards, but virtually nobody does this (and with good reason). They are also quite dangerous if used improperly.
- Hex (Allen) Keys: They include these in virtually every commercial tool kit I've ever encountered. I've never seen anything PC related that used them but of course that doesn't mean that there isn't something out there that does...
- Chip Extractors: These are used to remove chips from regular (non-ZIF) sockets, but modern PCs do not generally use these sockets any more. Useful for upgrading much older systems in some cases.
Finally, it is a good idea to accumulate a stockpile of spare parts if you are going to be doing work on PCs. I would recommend the following:
- Screws: I have a "big ol' bag o' screws" of all shapes and sizes that comes in handy when working on various machines.
- Expansion Card Inserts: Save the metal inserts that come out of the back of the case when you put a modem or other card into the PC, as you may need them again later on.
- Drive Faceplates: Similarly, save the plastic faceplates that you remove from the front of cases so you can replace them later if needed.
- Mounting Kits: These sometimes come with retail hard disks and allow you to put a 3.5" drive into a 5.25" bay. Useful when your case has more free 5.25" bays than 3.5" ones.
- Cables: Save any power, IDE, floppy, CD-ROM, or other cables that you accumulate in case you need them later on.
- Keyboard, Mouse, 3.5" Floppy Drive: Keep an extra one of each of these components around to aid in troubleshooting problems by swapping. (Actually, keeping an extra of every major component in the PC is helpful, but the three I mention are here because they are cheap. Most people can't afford to keep an extra hard disk lying around as a spare.)
Two Heads are Better Than One
If you have at your disposal someone who knows about PCs, say a peer in your office or a family member, have them take a look at a difficult problem. Even if you are great at troubleshooting, there will be the occasional problem that for one reason or another, you just won't be able to crack. Sometimes another person who can look at a problem from a fresh perspective (who hasn't been staring at the same thing for days trying to resolve it) will be able to set you on the right path to fixing the problem.
Don't be afraid of asking for help. There is no shame in needing assistance, and it can be an excellent way to learn.
Tip: You may be able to get assistance from others with your PC problem by using The PC Guide Discussion Forums.
It Isn't Personal...
OK, this section hits me very close to home. Before you take umbrage at any of the comments that I make here, I will start by saying that I am one of the worst people I know in doing all of the things that I say in this section one should not do. :^) I have a pretty bad temper and tend to take things too seriously in general, so it's not surprising that I tend to lose it sometimes when I am stuck with a particularly enigmatic situation with a PC.
What I am saying here is that despite the fact that it may seem otherwise when you are feeling frustrating about some sort of difficulty, the PC is not doing whatever it is doing intentionally to annoy you. It is not alive, it has no feelings, no motivations, no agenda. There is a logical reason for everything that a PC does--it is just hard sometimes to understand what it is. It can seem like the PC is "out to get you", and then to start feeling very frustrated and angry. Usually when this happens your ability to actually solve the problem will drop pretty close to zero.
During the throes of particularly frustrating spells, I personally have at various times yelled at PCs, kicked PC case covers, slammed the desk with my fist, crunched up floppy disks in my hands, smashed keyboards, smacked the side of a monitor, and hit the power switch rather hard more than once. As a result of this foolishness, I have on rare occasion damaged a component or two, and even hurt my hands. This is one area where I definitely do not recommend that my readers follow my example. :^)
Try as much as possible to keep your cool when working on your machine. If you feel yourself getting really pissed off, this is likely a good signal that you've been working on the problem too long. Take a break, do something else for a while. A fresh look the next day will be better for both you and the PC.
Steps To Take First When Troubleshooting
There are many different tricks and techniques that I outline in the Troubleshooting Guide, but the ones in this section are the few that I always recommend be tried first before proceeding into a full troubleshooting effort. The reason is simple: they can often solve the problem you are having before you spend a lot of energy looking into it seriously.
Let the Vendor Do the Work!
You may not have to use the troubleshooting information on this site at all. If the system is exhibiting strange behavior, especially if it appears that it has any failed or failing components, and you purchased this system recently, you may want to just return it and let the vendor figure out what the problem is. In many cases this is the best thing to do when your PC is still under warranty, since this also eliminates any chance of you voiding the warranty, and saves you time as well. That's what a warranty is for. See here for information on using your warranty, and repair in general.
The problem with doing this is that you have to incur the hassle of bringing the PC back for service. If the PC was purchased mail order, the problem is even worse. Sometimes looking on the site first is a good idea anyway, to rule out any possible simple causes that you can address yourself.
Scan For Viruses
If you experience any strange behavior on your system, especially either as it is booting, or shortly after it completes, you should always use whatever antivirus software you own to do a full scan of the system before you do anything else. You should use a clean boot floppy if possible.
Do not fall into the "this can't be a virus" trap too easily. Obviously if you turn the power on and nothing happens at all, or if your hard disk won't spin up, or if your system seizes while it is doing the initial BIOS power-on test of your system memory, these are hardware problems that are not caused by a virus. However virtually any other symptom that appears to be a hardware issue can be either a hardware problem or simply a virus that is trying to masquerade as a hardware problem. To remove the complicating factor of always having to wonder "is this a virus?", scan your system before you spend a lot of time chasing ghosts.
Read the Documentation
I never cease to be amazed at how people refuse, steadfastly, to read the documentation that comes with their new hardware or software. They are so excited about their new toy, whatever it is, that they are dying to get it running right away--who has time to start reading? So they start installing whatever-it-is and then have problems. Half the time they end up actually taking more time to get the thing running than they would have if they had just read the instructions in the first place. Meanwhile they often waste a lot of their time and others' asking for technical assistance.
Many, if not most problems are a result of hardware or software that has just been installed into the PC. If you are installing something new, read the documentation that came with the product. Look for files called something like "READ.ME" or "README.1ST" on installation or driver disks, and read them thoroughly before you start. Refer back in the event of trouble. You may be surprised to find that the answer to your dilemma is right there, just waiting for you to read it.
Double-Check Any Recent Changes
Ask yourself: when did the problem I am experiencing first start? If you have just made any change to your system, and are now having a problem with your PC that was not present before the change, the chances are probably 99% that the change is the cause of the problem. This is true even if the problem seems to have absolutely nothing to do with what you changed. If you hunt around, eventually you will find something that links the problem to the change. While it is certainly possible for something to coincidentally stop working at the same time that you make a change, the odds are greatly against this ever happening to you.
Here's an example of something that happened to me, to illustrate what I mean. I opened up my PC to do some work on my system memory. I removed two SIMMs and replaced them with another pair of higher capacity. When I turned on the system next, my floppy disk drive was malfunctioning. Had my floppy drive failed just then? Almost certainly not. As it turned out, my motherboard has the header for the floppy drive cable located directly adjacent to the SIMM slots. While I was working on the memory I had loosened the floppy drive cable.
Sometimes the problems can be even more indirectly related than this. You should in most cases stick to pursuing how the change affected the system, as this is likely to eventually lead you to the cause of the problem.
Simplify, Simplify, Simplify
An absolutely crucial rule to keep in mind when troubleshooting PC system problems is that if there are too many unknowns, it is impossible to determine which one is causing the problem. If you have many possible causes for some difficulty with your system, it can be extremely difficult to narrow down the cause of the problem to any one of them. If you are using a just-installed hard disk with an unfamiliar shareware file utility running on an operating system you just upgraded last week, and now your system has problems, how on earth are you ever going to know what is causing them?
In order to have a fighting chance at figuring out what is going on, you must simplify the situation as much as possible so that it becomes much more obvious what is responsible for the difficulty. This means reducing the number of variables to whatever degree possible. One important way of doing this is undoing or double-checking any changes made to the system.
In addition, I have identified the following items as often being responsible for erratic behavior that can complicate troubleshooting. I would recommend that they be eliminated or temporarily disabled when trying to diagnose a system problem:
- Power Management: Power management is a great idea in theory but in many ways is just "not ready for prime time". Power management routines can cause symptoms that appear to be hardware malfunctions, such as screens that turn off unexpectedly or hard disks that spin down. They also can cause crashes of software that doesn't know how to deal with them. If you want to use power management, turning it off until the problem is resolved is wise.
- Overclocked Hardware and Aggressive BIOS Settings: I do not believe in overclocking. If you insist on doing it, don't be surprised if you have system problems! Scale things back until you can figure out what the problem is. Similarly, if you are "pushing the envelope" in trying to squeeze maximum performance by tuning your BIOS memory timings and other settings very aggressively, try resetting them to more conservative values when troubleshooting.
- Experimental or Beta Software: This software is still in the test process and is likely to have bugs--that is why it is labeled as "beta"! For an end application this is usually no big deal, since any crashes or other problems will be limited to that application and therefore somewhat obvious. Running beta operating systems, drivers or other low-level software however is asking for trouble, and you should try to eliminate these possible sources of confusion when trying to debug your system.
- "Creative" Configurations: The more "unusual" things that you have going on in your system, the more likely that you are going to have a conflict caused by one of these strange pieces of hardware or software. A system that is loaded with unusual utilities, terminate-and-stay-resident programs, an old 8-bit network card salvaged from a 286, etc. will often have more problems than a stock Pentium box with a normal Windows 95 installation. To whatever extent possible, disable these items while troubleshooting. Also try to avoid using unusual low-level software whenever possible.
- Excessive Connections: If the PC is on a network or is connected to a large number or variety of peripheral devices, you may want to try to disconnect those and see if there is any impact on the problem.
In general you want to avoid the unusual or the unknown when troubleshooting. One way to simplify the software environment during diagnosis is to use a boot floppy to "boot clean" and bypass the special drivers and software that you normally load when you boot from your hard disk. You can also use the {F8} key when DOS or Windows 95 are booting to bypass your startup files, basically accomplishing the same thing. You want to be especially wary of software that sits in the background and activates without you specifically telling it to, as this can confound your troubleshooting efforts.
To whatever extent possible, disable as much as you can when trying to figure out a problem. The more funky software utilities, screen savers and cute peripherals you disable now, the more chance you have of finding out which one it is that is causing the problem later on.
General Diagnostic Techniques
This section discusses the rules of thumb that I recommend you employ when you are troubleshooting. These are the general techniques and ideas that can help you solve all sorts of different problems with your PC (and even in other parts of your life...) It often is worth trying these general tips before attempting to find a specific answer to your problem in the Troubleshooting Expert.
Get to the Root Cause: Ask "Why" Five Times
One of the keys to truly solving a problem is to first truly understand it. It is easy to observe that too often, people are quick to jump to solve a problem before they really understand what it is. It is easy to see a symptom and think that you know what is causing it, but surprisingly often, if you take the time to explore more deeply you will find that what you thought was the cause is in fact just another symptom, and that the problem lies much deeper within.
One technique that you can employ is borrowed from Japanese manufacturing management theory. This method is used to help identify the real causes of problems that occur on the manufacturing floor. The goal is not to simply correct the effects of the problem, but to find out the root cause of why the problem is occurring so that we can ensure that it will not happen in the future.
One simple way to do this is called asking why' five times. The idea is that by the time you have asked "why" the fifth time, you will be at the root cause. It isn't always that simple, but the exercise can be surprisingly insightful in helping you figure out what is really going on, and can help you avoid "quick fix" solutions that are really just band-aids and don't resolve anything. It is especially useful for tackling chronic problems that show up over and over again in a system; it is less useful for problems that are unlikely to recur.
Here's an example. Let's suppose your hard disk is having a problem with bad sectors showing up. The knee-jerk reaction to this happening is "the hard disk is bad, replace it". Instead, ask yourself:
- "Why are bad sectors showing up when they didn't before?"
- "The hard disk is being corrupted somehow; something has changed."
- "Why?"
- "The timing of the system probably has been changed in some way."
- "Why?"
- "Hmm, I installed a new hard drive in that system just last week, perhaps it is causing the problem."
- "Why?"
- "The hard drive may not have been installed correctly, or perhaps I should have used a different channel. Or maybe it was the bus-mastering driver that came with the new drive, the old drive may not be able to support it".
You can see the general idea; the answers will differ in every case, but it is the process itself that is useful. Another example: imagine an office that has just had a catastrophic data loss due to a PC crashing after an electrical storm. The initially identified cause of this problem was a lack of a UPS on the PC, which would have protected the system from the electrical storm (usually). So the MIS department starts drafting purchase requisitions for UPSes for all the PCs. Instead, if they looked at the problem carefully, they may have employed the following process:
- "Remember that the problem here was the data loss, not the crashing of the PC. Why was the data lost?"
- "The PC's hard disk crashed."
- "Why would that cause the data to be lost?"
- "Because we didn't have a backup."
- "Why wasn't there a backup?"
- "There is a tape backup unit on every PC in the building but very few of them are being used."
- "Why?"
- "Because most of the PC users do not know how to use them."
- "Why?"
- "They have never been given any training in their use."
- "Why?"
- "Because we don't have a budget for PC hardware training."
As you can see, the root cause that we came up with here is very different than the cause we saw when we only scratched the surface. This means of course that the solution will be very different as well; it might well make a lot more sense here to spend a small amount of money on training instead of a large amount of money on UPSes for every PC.
Be Observant and Look For Evidence
It is sometimes the subtle signs that can provide the triggering piece of information that leads to discovering the source of a problem. In particular, you want to carefully investigate anything that seems "unusual", "wrong" or "surprising". These are often the clues that will get you going down the right track to figuring out what is giving you trouble. If you are observant and careful about what you are doing, you will be much more likely to pick up on these things than if you are in a hurry or you are quick to discount your findings as "unrelated to the problem".
Use the Process of Elimination
Virtually all problems with PCs involve more than one component or subsystem. The difficulty is usually in figuring out which component is responsible for the problem. Using the process of elimination, however, you can usually narrow the problem down rather quickly by making small logical changes and observing the impact on the problem. Your objective is to isolate the cause of the problem so you can correct it.
The key is to make only one change at a time and then see if the problem goes away; if it does, then whatever you changed is likely responsible for the problem (although it could be fixing the problem indirectly in some cases.) If you make more than one change at a time, you cannot readily discern which change was responsible for fixing the problem.
You will want to first check the most probable sources of the problem, and also the things that are easiest to change. For example, if you are having a problem with your disk drive being recognized, it's a lot easier and cheaper to explore things like double-checking jumpers and connections or replacing the interface cable, than it is to try replacing the drive itself. That is something you'd only do after you had eliminated all the other possibilities (or if the evidence implicated the hard disk directly).
Here's a simple example. Let's suppose one morning your PC will not turn on. You hit the switch and nothing happens. There could be many different possible causes for this problem: the power to the house could be out; there could be a malfunction in the wall socket; the surge suppressor that the system is plugged into might have blown; the electrical cord may be loose; the power supply could be damaged. To figure out what is going on you need to eliminate these variables by making small changes and seeing what happens. For example:
- Change the wall socket you are using. If the PC now boots, you have isolated the cause to the electrical wiring in the house.
- If the problem persists, examine the surge suppressor. Change it, or temporarily bypass it and plug the PC into the wall directly. If it now works, the surge suppressor is the problem.
- If the problem still isn't fixed, try changing the power cord.
- If the problem persists still, you may then have to open up the box and look at the power supply unit to see if it might need replacing.
Realize that the key here is making these changes one at a time. If you approach this problem by changing the wall socket you use, bypassing the surge suppressor, and changing the power cord all at once, your problem may go away but how will you know what caused it? This is a valid way to troubleshoot if you have to get the system back up immediately, however. You can then undo the changes one at a time later on to find out what the cause is, in effect doing the same single-change-at-a-time process, but in reverse.
Most problems with the PC can be diagnosed using this sort of process; it's very powerful. You will notice many sequences like the one above in The Troubleshooting Expert.
Do One Upgrade or Assembly Step At a Time
Changes made to the system are the most frequent cause of problems; this is the nature of change. Much the way many problems can be diagnosed by using the process of elimination through making single changes to the system, you can avoid or detect problems with upgrades or new installations by going "one step at a time" in your changes as well. New system installations or major upgrades often can have problems that are very difficult to diagnose simply because there are so many modifications being performed at the same time.
When you build a new PC you will probably be assembling a large number of components that have never been used or tested before. To whatever extent possible, try to go as slowly as possible when assembling the machine. For example, when assembling a new PC from scratch, it is always best to first make sure the basic system is working, with the only expansion card in the system the video card (if applicable). Adding the sound card, network card and other devices should be done later on, and one at a time if possible. Similarly, do not try to do major software or operating system upgrades at the same time that you make hardware changes. Doing this can make it very difficult for you to troubleshoot any system problems.
If you do make multiple changes at once, try retracing your steps. Undo the changes you have made one at a time and see if you can identify the change that caused the problem that way.
Make Use of Components That You Know Work
One valuable resource while troubleshooting is a "stockpile" of extra components. I put "stockpile" in quotes because it doesn't have to be an actual stock of components, just a resource that will let you borrow components that you know work while you are troubleshooting your system. Another functional PC that uses the same or similar components is normally quite adequate. Being in an office environment can be very helpful this way since there are many PCs around to work with, and often they are similar to one another; at home it may be more difficult if you only have the one machine.
These extra components can be useful when you are employing the process of elimination to resolve problems. For example, suppose that there is no video on your monitor when you boot up the PC. After eliminating the simple stuff (cable disconnected, brightness turned down far too much, etc.) there are two components that are most likely to be bad: the video card or the monitor. An easy thing to try is to swap monitors with another machine that is working properly. If the second monitor also doesn't work, and your original monitor works on the other machine, you can feel quite confident that it isn't the monitor that is the problem here.
You can sometimes avoid problems in assembling a new PC by testing out some of the components you are going to use before you begin. For example, take your new mouse, keyboard and monitor and attach them to a PC that is known to work, and test them out. This way if you have problems when you build your new system you will know it isn't due to these components. At my office, I keep an extra keyboard, mouse and monitor just for debugging and assembling PCs. Since I know that they work, I reduce the number of variables that I have to deal with when I experience a problem. Once the main PC box is assembled and working, I connect the new keyboard, mouse and monitor. If a problem shows up then, I know that it is not likely the PC, but the peripherals I just connected.
Write Things Down
Whenever you discern a piece of evidence, or you figure out something about the current situation your PC is in, document it. Keeping a history of what you discover and what your symptoms are will help you not only with this problem but also with any future difficulties. A log book of some sort is a good idea.
Determine Repeatability
Most problems with a PC fall into one of two categories: either they are repeatable or they are intermittent. A repeatable problem is one where the problem occurs all the time, or always in response to a specific user action. For example, a PC that has a problem that prevents it from booting will probably always fail to boot no matter how many times you reset it. Or you may have an application that whenever you try to run, will crash with an error. You may find that your PC hangs, but only when you move the mouse at the same time that you are communicating using your modem.
In contrast, some problems are intermittent and not repeatable. In some cases, you may have a PC that will usually boot up fine, but one day a month will fail to boot for some reason. An application may work most of the time but occasionally crash. The PC may lock up at seemingly random intervals. Your mouse may work almost all of the time, but one day out of five or ten may give you trouble.
It is helpful to determine if the problem you are experiencing is repeatable, because intermittent difficulties are much more difficult to resolve than repeatable ones. If a problem is repeatable, and there is a specific set of actions that cause the problem, this gives you at least some initial clues about how to find the cause. In addition, you have a way of testing to see if you have resolved the problem when you are trying different solutions. Intermittent problems are much more difficult to deal with.
Determining if a problem is repeatable is pretty simple: try to duplicate the conditions that caused the problem and see if it happens again.
Dealing With Intermittent Problems
Intermittent problems are ones which appear to happen spontaneously or randomly. They seem not to be caused by anything obvious and are not repeatable. They can be extremely difficult and frustrating to diagnose.
Sometimes problems that seem intermittent really aren't; it's just that the specific set of circumstances that causes the problem to occur may be obscure or hard to notice. Spend some time trying to determine what the circumstances are when the problem arises. For example, many problems with crashes or lockups will occur only after the system has been on for more than an hour; some may occur only within the first few minutes that the PC is turned on. You may find that program A only crashes when run at the same time as program B. Or that a particular behavior is associated with a peripheral of some sort.
You will have to be patient when dealing with these sorts of issues. Since the problem is not something you can duplicate at will, you may not be able to systematically work your way towards the ultimate cause. In this situation you may have to employ trial-and-error, making a change and then waiting to see if the problem recurs. It can take days (or longer) sometimes since you have to wait before seeing if the problem happens again. Be patient.
Correlation May Not Imply Causation
I use the word correlation, perhaps slightly incorrectly, to refer to two behaviors or symptoms that appear at the same time. Causation refers to two events where one is responsible for the other's appearance or existence. So what I am saying here is that if you see two strange things happening at the same time on your PC, this does not necessarily mean that one of them has caused the other. They could be coincidental, or it could be that where you think A is causing B, B is really causing A.
For example, you may find two symptoms on your machine: you have frequent file system corruption on your hard disk, and you have a lot of system lockups (hangs, crashes). You may think that the crashes are causing the file system errors, and in many cases this is true. However it is possible for the file system errors to cause the lockups. And it is also possible that both are just symptoms of another underlying cause.
Scan For File System Problems
If the problem you are looking at is anything that involves the hard disk, files, operating system problems, crashes or similar behavior, scan the file system for errors. This is a simple step (which you should be performing regularly as part of a preventive maintenance program) that may tell you immediately that there is a problem with your hard disk. This can also be a useful piece of evidence in troubleshooting some sorts of problems.
Use Diagnostic Tools
If you own diagnostic hardware and/or software tools, they can be useful in pinpointing some kinds of problems. If you feel that you are having a problem that can be detected by a tool, or you feel that running a diagnostic utility may shed more light on what's happening, then make use of the tool to help guide your analysis of the problem.
Next: Diagnostic, Troubleshooting and Repair Tools
Diagnostic, Troubleshooting and Repair Tools
Troubleshooting a problem with a PC can be very difficult and frustrating. Fortunately, there are places to turn for assistance, and there are also tools at your disposal that can make the job much easier. These can be used to identify specific problems within the PC, or just give you more information about the system so you understand better what you are dealing with. This chapter takes a brief look at these tools, which can make your life easier if you are having trouble.
Diagnostic Software
The use of diagnostic software tools can save you a great deal of time. While in my experience these tools do not usually identify what the cause of a system problem is, they often provide hints or at least valuable information about what is in the system and how it is working. Some of these are free or are included with common operating systems at no additional charge, while others are commercial products that range from affordable to rather pricey.
Here are some common software diagnostic tools that you will want to consider including as part of your troubleshooting arsenal:
- Power-On Self Test (POST): I am cheating a bit here because this isn't a separate diagnostic utility; it is in fact built into your system BIOS and it runs every time you start up your PC automatically. It is often the best indicator of system problems; make sure you pay attention to its audio and video messages. Don't disable its error-reporting functions unless you really need to.
- MEM.EXE: This simple utility, built into recent versions of DOS and also Windows 95, provides you with details about your memory configuration, as well as what is currently using your memory. It is especially useful when run with the "/C" parameter (use the "/P" parameter as well to make the output pause when it is scrolling).
- Microsoft Diagnostics: Better known as "MSD.EXE", this is a small DOS utility that takes a brief inventory of the contents of your PC and shows them to you in a text-based format. This is very useful for seeing what disks are in the system, how much memory is installed, and also for checking system resource usage such as LPT ports and IRQs. It will show you what type of BIOS you are using and also what UART chip you have in your serial ports. MSD.EXE is included in later versions of DOS.
Tip: MSD.EXE will work on a Windows 95 machine but is not normally installed as part of the normal Windows 95 installation. You can find it on the Windows 95 CD however and then copy it to the C:\WINDOWS\COMMAND directory to make it work.
Warning: MSD.EXE can give incorrect information when run from within Windows. Shut down to DOS mode before running it ("Restart the computer in MS-DOS mode")
- The Windows 95 Device Manager: This is the most useful tool for identifying system configuration and resource usage information under Windows 95. To access it, open the Control Panel and select the "System" icon. Then select the "Device Manager" tab. You will see a graphical "tree" structure showing you all of your PC hardware. If you select "Properties" while "Computer" (the top-level item) is selected, you will be able to see all the IRQs, DMA channels and I/O addresses in use in your PC; very useful for resolving resource conflicts! The same "Properties" button, pressed after selecting a specific hardware device, will show you driver information, resource settings for the hardware item chosen, and much more.
- Norton System Information: This utility is similar to the Microsoft Diagnostics, only more detailed in its later versions. SI shows a great deal of information about what is in the PC, going well beyond what MSD gives you, but really is still an information utility as opposed to a true diagnostic. This program is part of Symantec's Norton Utilities.
- Microsoft ScanDisk and Norton Disk Doctor: These programs are used to check for hard disk problems. This includes file system corruption and hard disk read errors. They should be used when hard disk problems are suspected.
- Norton Diagnostics: This utility is meant to go beyond the System Information program and actually perform tests on the hardware to identify problems. It includes tests of the processor and motherboard and system memory, and will identify some types of resource conflicts. In reality it is still quite limited in terms of the numbers of problems it will find.
- QAPlus: QAPlus from DiagSoft is a more advanced diagnostic suite that comes in several flavors, depending on what you need to do and how you want to do it. This is a more expensive package but can give you much more detailed information about your system and help identify problem situations as well.
Diagnostic Hardware
While not as commonly used as diagnostic software, there are available several pieces of hardware that can be very helpful in troubleshooting some specific hardware problems. These units are not nearly as popular because they are usually more expensive and are used for specific types of problems, making them less universally applicable than the general software utilities that are widely encountered.
If you are a serious home builder or troubleshooter, you may want to consider one or more of the following:
- Loop-Back Plugs: These are small plugs that go onto the serial and parallel ports on your computer and connect the receive to transmit lines to simulate a connection (they connect the port to itself). Using these with a program like Norton Diagnostics allows for full testing of the operation of these ports, whereas without them only the internal portions of the port can be tested.
Note: The type of plugs used by a particular diagnostic utility may be different than the type used by another program made by another company. Check it out before you buy.
- Multimeter/Ohmmeter: An ohmmeter is a device that measures electrical resistance; a multimeter can measure resistance, voltage and current. These devices are used by electricians, electronics designers and repairpeople. They have come down greatly in price and simple ones can be had inexpensively (while top-end ones of course are still quite a bit of money). A simple ohmmeter is useful primarily for checking for short circuits or open circuits (broken connections, damaged cables, etc.) Multimeters can be used for more extensive electronics testing.
- BIOS POST Cards: It is well-known that the BIOS will, if it finds a problem during its power-on self-test (POST), produce audio "beep codes" and/or video messages that indicate what the trouble is that it found. What many do not realize is that the BIOS of most PCs is designed to send a stream of test codes to a special memory location as it performs its tests, usually 80h. By using a special card designed to capture and display these codes, you can pinpoint exactly where in the power-on self-test a system is hanging up. This can be extremely helpful in debugging very stubborn systems. Scott Wainner has more about one of these cards on his site.
- Test Bed: Some experienced PC repair people keep around an older system that they can use as a test bed for components. It can be very useful to be able to test an unknown device with other components that are known to work, to cut down on the guesswork.
There is also more advanced test hardware available, including devices such as logic probes, oscilloscopes, and many types of specialized component testers. These items are generally very expensive and require specific training to be used properly; they are not for the home PC user or even for most hobbyists. Professional test hardware costs thousands of dollars, which is one reason why doing it yourself is often not an option for many kinds of repairs.
The PC Tool Kit
If you are a very casual PC user you can probably get by hacking around with an old screwdriver and a pocket knife when you need to work on your PC. However, if you plan to do any reasonable amount of work upgrading, building, troubleshooting or repairing PCs, a proper tool kit is a good investment. There are many different types of tool kits available for the computer enthusiast, with tools selected specifically to make the lives of those who work with PCs easier. The right tools can save you a lot of time and help you avoid damage to your equipment (and your hands!)
As with most things, tool kits range widely in size, quality and price. At the low end are $5 kits that are barely any better than the screwdriver and pocket knife. At the high end are kits costing over $100 that include everything but the kitchen sink; these are overkill for most people (including me). Most people can do just fine with a toolkit costing between $15 and $50. These are often available pre-assembled into "genuine artificial leather" carrying cases, sold at most general computer stores.
Here are the items that I would recommend you definitely have in your toolkit; note that some of these don't generally come in pre-assembled toolkits so you will want to add them yourself to the carrying case. Every item in this list is one that I use personally on a regular basis:
- A Good Screwdriver: You want a good screwdriver with magnetic, interchangeable bits and a comfortable handle. At a minimum you should have two different sizes each of Phillips and flat bits. If you've never used a magnetic screwdriver before, you'll wonder how you ever did without one once you do; the magnetic tip will prevent screws falling into the PC, and will let you get access to drives mounted deep in your system case.
- Additional Screwdrivers: An additional Philips screwdriver and an additional flat blade screwdriver, for when you misplace the good one. :^) Actually, it's sometimes useful to have two screwdrivers anyway.
- Needle-Nose Pliers: These are useful for grasping small items and for removing and replacing jumpers on circuit boards.
- Wire Snips: A pair of wire snips for cutting wire and stripping insulation. Some kits include wire strippers for the latter purpose.
- A Small Flashlight: Very useful; the insides of PC boxes are quite dark and there are lots of very small things you will need to see, such as the "pin 1" marking on a connector for example.
- Tweezers: Or even better, a part retriever. A retriever is like a tiny set of retractable claws with a spring-loaded handle. Useful stuff.
- A Large Crowbar: For making adjustments to stubborn hardware. (Just kidding. :^) )
- A Roll of Black Electrical Tape: They never include this in the kits but they should. Used for wrapping wire ends and insulating components.
- A Can of Compressed Air: Very handy for cleaning things hands-free and without using any hazardous liquids.
- A Soft, Lint-Free Cloth: For cleaning the monitor and other components.
The following items I categorize as "optional, but useful". They are tools that are used less frequently, or perhaps only in certain circumstances. They are more for someone who does a fair bit of tinkering or who maintains a number of machines:
- A 3/16" Nut Driver: In fact, tool kits often come with several different sizes of nut drivers (using exchangeable bits), but 3/16" is the really important one, because this is the size of nut most commonly used on PCs. These hexagonal nuts are used as mounting hardware for motherboards and serial and parallel ports. A necessity for anyone who builds their own PCs; if you don't have one you'll be attaching port connectors using a pair of poor-fitting pliers.
- Vice Grips: These are incredibly neat tools that have all sorts of uses.
- Torx Screwdrivers or Bits: These are the star-shaped screwdriver heads that are used to make equipment "tamper-proof". You need these drivers if you want to, well, "tamper".
- An ESD (Electrostatic Discharge) Wrist Strap: This is more of a safety device than a tool; it is used to greatly reduce the chances of static damage to components.
- A Knife: A cutting blade or utility knife of some sort.
The following items are sometimes included in tool kits but in my experience are rarely if ever used:
- Soldering Irons: Essential if you are going to try to do repair of circuit boards, but virtually nobody does this (and with good reason). They are also quite dangerous if used improperly.
- Hex (Allen) Keys: They include these in virtually every commercial tool kit I've ever encountered. I've never seen anything PC related that used them but of course that doesn't mean that there isn't something out there that does...
- Chip Extractors: These are used to remove chips from regular (non-ZIF) sockets, but modern PCs do not generally use these sockets any more. Useful for upgrading much older systems in some cases.
Finally, it is a good idea to accumulate a stockpile of spare parts if you are going to be doing work on PCs. I would recommend the following:
- Screws: I have a "big ol' bag o' screws" of all shapes and sizes that comes in handy when working on various machines.
- Expansion Card Inserts: Save the metal inserts that come out of the back of the case when you put a modem or other card into the PC, as you may need them again later on.
- Drive Faceplates: Similarly, save the plastic faceplates that you remove from the front of cases so you can replace them later if needed.
- Mounting Kits: These sometimes come with retail hard disks and allow you to put a 3.5" drive into a 5.25" bay. Useful when your case has more free 5.25" bays than 3.5" ones.
- Cables: Save any power, IDE, floppy, CD-ROM, or other cables that you accumulate in case you need them later on.
- Keyboard, Mouse, 3.5" Floppy Drive: Keep an extra one of each of these components around to aid in troubleshooting problems by swapping. (Actually, keeping an extra of every major component in the PC is helpful, but the three I mention are here because they are cheap. Most people can't afford to keep an extra hard disk lying around as a spare.)
Steps To Take First When Troubleshooting
There are many different tricks and techniques that I outline in the Troubleshooting Guide, but the ones in this section are the few that I always recommend be tried first before proceeding into a full troubleshooting effort. The reason is simple: they can often solve the problem you are having before you spend a lot of energy looking into it seriously.
Let the Vendor Do the Work!
You may not have to use the troubleshooting information on this site at all. If the system is exhibiting strange behavior, especially if it appears that it has any failed or failing components, and you purchased this system recently, you may want to just return it and let the vendor figure out what the problem is. In many cases this is the best thing to do when your PC is still under warranty, since this also eliminates any chance of you voiding the warranty, and saves you time as well. That's what a warranty is for. See here for information on using your warranty, and repair in general.
The problem with doing this is that you have to incur the hassle of bringing the PC back for service. If the PC was purchased mail order, the problem is even worse. Sometimes looking on the site first is a good idea anyway, to rule out any possible simple causes that you can address yourself.
Scan For Viruses
If you experience any strange behavior on your system, especially either as it is booting, or shortly after it completes, you should always use whatever antivirus software you own to do a full scan of the system before you do anything else. You should use a clean boot floppy if possible.
Do not fall into the "this can't be a virus" trap too easily. Obviously if you turn the power on and nothing happens at all, or if your hard disk won't spin up, or if your system seizes while it is doing the initial BIOS power-on test of your system memory, these are hardware problems that are not caused by a virus. However virtually any other symptom that appears to be a hardware issue can be either a hardware problem or simply a virus that is trying to masquerade as a hardware problem. To remove the complicating factor of always having to wonder "is this a virus?", scan your system before you spend a lot of time chasing ghosts.
Read the Documentation
I never cease to be amazed at how people refuse, steadfastly, to read the documentation that comes with their new hardware or software. They are so excited about their new toy, whatever it is, that they are dying to get it running right away--who has time to start reading? So they start installing whatever-it-is and then have problems. Half the time they end up actually taking more time to get the thing running than they would have if they had just read the instructions in the first place. Meanwhile they often waste a lot of their time and others' asking for technical assistance.
Many, if not most problems are a result of hardware or software that has just been installed into the PC. If you are installing something new, read the documentation that came with the product. Look for files called something like "READ.ME" or "README.1ST" on installation or driver disks, and read them thoroughly before you start. Refer back in the event of trouble. You may be surprised to find that the answer to your dilemma is right there, just waiting for you to read it.
Double-Check Any Recent Changes
Ask yourself: when did the problem I am experiencing first start? If you have just made any change to your system, and are now having a problem with your PC that was not present before the change, the chances are probably 99% that the change is the cause of the problem. This is true even if the problem seems to have absolutely nothing to do with what you changed. If you hunt around, eventually you will find something that links the problem to the change. While it is certainly possible for something to coincidentally stop working at the same time that you make a change, the odds are greatly against this ever happening to you.
Here's an example of something that happened to me, to illustrate what I mean. I opened up my PC to do some work on my system memory. I removed two SIMMs and replaced them with another pair of higher capacity. When I turned on the system next, my floppy disk drive was malfunctioning. Had my floppy drive failed just then? Almost certainly not. As it turned out, my motherboard has the header for the floppy drive cable located directly adjacent to the SIMM slots. While I was working on the memory I had loosened the floppy drive cable.
Sometimes the problems can be even more indirectly related than this. You should in most cases stick to pursuing how the change affected the system, as this is likely to eventually lead you to the cause of the problem.
Simplify, Simplify, Simplify
An absolutely crucial rule to keep in mind when troubleshooting PC system problems is that if there are too many unknowns, it is impossible to determine which one is causing the problem. If you have many possible causes for some difficulty with your system, it can be extremely difficult to narrow down the cause of the problem to any one of them. If you are using a just-installed hard disk with an unfamiliar shareware file utility running on an operating system you just upgraded last week, and now your system has problems, how on earth are you ever going to know what is causing them?
In order to have a fighting chance at figuring out what is going on, you must simplify the situation as much as possible so that it becomes much more obvious what is responsible for the difficulty. This means reducing the number of variables to whatever degree possible. One important way of doing this is undoing or double-checking any changes made to the system.
In addition, I have identified the following items as often being responsible for erratic behavior that can complicate troubleshooting. I would recommend that they be eliminated or temporarily disabled when trying to diagnose a system problem:
- Power Management: Power management is a great idea in theory but in many ways is just "not ready for prime time". Power management routines can cause symptoms that appear to be hardware malfunctions, such as screens that turn off unexpectedly or hard disks that spin down. They also can cause crashes of software that doesn't know how to deal with them. If you want to use power management, turning it off until the problem is resolved is wise.
- Overclocked Hardware and Aggressive BIOS Settings: I do not believe in overclocking. If you insist on doing it, don't be surprised if you have system problems! Scale things back until you can figure out what the problem is. Similarly, if you are "pushing the envelope" in trying to squeeze maximum performance by tuning your BIOS memory timings and other settings very aggressively, try resetting them to more conservative values when troubleshooting.
- Experimental or Beta Software: This software is still in the test process and is likely to have bugs--that is why it is labeled as "beta"! For an end application this is usually no big deal, since any crashes or other problems will be limited to that application and therefore somewhat obvious. Running beta operating systems, drivers or other low-level software however is asking for trouble, and you should try to eliminate these possible sources of confusion when trying to debug your system.
- "Creative" Configurations: The more "unusual" things that you have going on in your system, the more likely that you are going to have a conflict caused by one of these strange pieces of hardware or software. A system that is loaded with unusual utilities, terminate-and-stay-resident programs, an old 8-bit network card salvaged from a 286, etc. will often have more problems than a stock Pentium box with a normal Windows 95 installation. To whatever extent possible, disable these items while troubleshooting. Also try to avoid using unusual low-level software whenever possible.
- Excessive Connections: If the PC is on a network or is connected to a large number or variety of peripheral devices, you may want to try to disconnect those and see if there is any impact on the problem.
In general you want to avoid the unusual or the unknown when troubleshooting. One way to simplify the software environment during diagnosis is to use a boot floppy to "boot clean" and bypass the special drivers and software that you normally load when you boot from your hard disk. You can also use the {F8} key when DOS or Windows 95 are booting to bypass your startup files, basically accomplishing the same thing. You want to be especially wary of software that sits in the background and activates without you specifically telling it to, as this can confound your troubleshooting efforts.
To whatever extent possible, disable as much as you can when trying to figure out a problem. The more funky software utilities, screen savers and cute peripherals you disable now, the more chance you have of finding out which one it is that is causing the problem later on.
General Diagnostic Techniques
This section discusses the rules of thumb that I recommend you employ when you are troubleshooting. These are the general techniques and ideas that can help you solve all sorts of different problems with your PC (and even in other parts of your life...) It often is worth trying these general tips before attempting to find a specific answer to your problem in the Troubleshooting Expert.
Get to the Root Cause: Ask "Why" Five Times
One of the keys to truly solving a problem is to first truly understand it. It is easy to observe that too often, people are quick to jump to solve a problem before they really understand what it is. It is easy to see a symptom and think that you know what is causing it, but surprisingly often, if you take the time to explore more deeply you will find that what you thought was the cause is in fact just another symptom, and that the problem lies much deeper within.
One technique that you can employ is borrowed from Japanese manufacturing management theory. This method is used to help identify the real causes of problems that occur on the manufacturing floor. The goal is not to simply correct the effects of the problem, but to find out the root cause of why the problem is occurring so that we can ensure that it will not happen in the future.
One simple way to do this is called asking why' five times. The idea is that by the time you have asked "why" the fifth time, you will be at the root cause. It isn't always that simple, but the exercise can be surprisingly insightful in helping you figure out what is really going on, and can help you avoid "quick fix" solutions that are really just band-aids and don't resolve anything. It is especially useful for tackling chronic problems that show up over and over again in a system; it is less useful for problems that are unlikely to recur.
Here's an example. Let's suppose your hard disk is having a problem with bad sectors showing up. The knee-jerk reaction to this happening is "the hard disk is bad, replace it". Instead, ask yourself:
- "Why are bad sectors showing up when they didn't before?"
- "The hard disk is being corrupted somehow; something has changed."
- "Why?"
- "The timing of the system probably has been changed in some way."
- "Why?"
- "Hmm, I installed a new hard drive in that system just last week, perhaps it is causing the problem."
- "Why?"
- "The hard drive may not have been installed correctly, or perhaps I should have used a different channel. Or maybe it was the bus-mastering driver that came with the new drive, the old drive may not be able to support it".
You can see the general idea; the answers will differ in every case, but it is the process itself that is useful. Another example: imagine an office that has just had a catastrophic data loss due to a PC crashing after an electrical storm. The initially identified cause of this problem was a lack of a UPS on the PC, which would have protected the system from the electrical storm (usually). So the MIS department starts drafting purchase requisitions for UPSes for all the PCs. Instead, if they looked at the problem carefully, they may have employed the following process:
- "Remember that the problem here was the data loss, not the crashing of the PC. Why was the data lost?"
- "The PC's hard disk crashed."
- "Why would that cause the data to be lost?"
- "Because we didn't have a backup."
- "Why wasn't there a backup?"
- "There is a tape backup unit on every PC in the building but very few of them are being used."
- "Why?"
- "Because most of the PC users do not know how to use them."
- "Why?"
- "They have never been given any training in their use."
- "Why?"
- "Because we don't have a budget for PC hardware training."
As you can see, the root cause that we came up with here is very different than the cause we saw when we only scratched the surface. This means of course that the solution will be very different as well; it might well make a lot more sense here to spend a small amount of money on training instead of a large amount of money on UPSes for every PC.
Be Observant and Look For Evidence
It is sometimes the subtle signs that can provide the triggering piece of information that leads to discovering the source of a problem. In particular, you want to carefully investigate anything that seems "unusual", "wrong" or "surprising". These are often the clues that will get you going down the right track to figuring out what is giving you trouble. If you are observant and careful about what you are doing, you will be much more likely to pick up on these things than if you are in a hurry or you are quick to discount your findings as "unrelated to the problem".
Use the Process of Elimination
Virtually all problems with PCs involve more than one component or subsystem. The difficulty is usually in figuring out which component is responsible for the problem. Using the process of elimination, however, you can usually narrow the problem down rather quickly by making small logical changes and observing the impact on the problem. Your objective is to isolate the cause of the problem so you can correct it.
The key is to make only one change at a time and then see if the problem goes away; if it does, then whatever you changed is likely responsible for the problem (although it could be fixing the problem indirectly in some cases.) If you make more than one change at a time, you cannot readily discern which change was responsible for fixing the problem.
You will want to first check the most probable sources of the problem, and also the things that are easiest to change. For example, if you are having a problem with your disk drive being recognized, it's a lot easier and cheaper to explore things like double-checking jumpers and connections or replacing the interface cable, than it is to try replacing the drive itself. That is something you'd only do after you had eliminated all the other possibilities (or if the evidence implicated the hard disk directly).
Here's a simple example. Let's suppose one morning your PC will not turn on. You hit the switch and nothing happens. There could be many different possible causes for this problem: the power to the house could be out; there could be a malfunction in the wall socket; the surge suppressor that the system is plugged into might have blown; the electrical cord may be loose; the power supply could be damaged. To figure out what is going on you need to eliminate these variables by making small changes and seeing what happens. For example:
- Change the wall socket you are using. If the PC now boots, you have isolated the cause to the electrical wiring in the house.
- If the problem persists, examine the surge suppressor. Change it, or temporarily bypass it and plug the PC into the wall directly. If it now works, the surge suppressor is the problem.
- If the problem still isn't fixed, try changing the power cord.
- If the problem persists still, you may then have to open up the box and look at the power supply unit to see if it might need replacing.
Realize that the key here is making these changes one at a time. If you approach this problem by changing the wall socket you use, bypassing the surge suppressor, and changing the power cord all at once, your problem may go away but how will you know what caused it? This is a valid way to troubleshoot if you have to get the system back up immediately, however. You can then undo the changes one at a time later on to find out what the cause is, in effect doing the same single-change-at-a-time process, but in reverse.
Most problems with the PC can be diagnosed using this sort of process; it's very powerful. You will notice many sequences like the one above in The Troubleshooting Expert.
Do One Upgrade or Assembly Step At a Time
Changes made to the system are the most frequent cause of problems; this is the nature of change. Much the way many problems can be diagnosed by using the process of elimination through making single changes to the system, you can avoid or detect problems with upgrades or new installations by going "one step at a time" in your changes as well. New system installations or major upgrades often can have problems that are very difficult to diagnose simply because there are so many modifications being performed at the same time.
When you build a new PC you will probably be assembling a large number of components that have never been used or tested before. To whatever extent possible, try to go as slowly as possible when assembling the machine. For example, when assembling a new PC from scratch, it is always best to first make sure the basic system is working, with the only expansion card in the system the video card (if applicable). Adding the sound card, network card and other devices should be done later on, and one at a time if possible. Similarly, do not try to do major software or operating system upgrades at the same time that you make hardware changes. Doing this can make it very difficult for you to troubleshoot any system problems.
If you do make multiple changes at once, try retracing your steps. Undo the changes you have made one at a time and see if you can identify the change that caused the problem that way.
Make Use of Components That You Know Work
One valuable resource while troubleshooting is a "stockpile" of extra components. I put "stockpile" in quotes because it doesn't have to be an actual stock of components, just a resource that will let you borrow components that you know work while you are troubleshooting your system. Another functional PC that uses the same or similar components is normally quite adequate. Being in an office environment can be very helpful this way since there are many PCs around to work with, and often they are similar to one another; at home it may be more difficult if you only have the one machine.
These extra components can be useful when you are employing the process of elimination to resolve problems. For example, suppose that there is no video on your monitor when you boot up the PC. After eliminating the simple stuff (cable disconnected, brightness turned down far too much, etc.) there are two components that are most likely to be bad: the video card or the monitor. An easy thing to try is to swap monitors with another machine that is working properly. If the second monitor also doesn't work, and your original monitor works on the other machine, you can feel quite confident that it isn't the monitor that is the problem here.
You can sometimes avoid problems in assembling a new PC by testing out some of the components you are going to use before you begin. For example, take your new mouse, keyboard and monitor and attach them to a PC that is known to work, and test them out. This way if you have problems when you build your new system you will know it isn't due to these components. At my office, I keep an extra keyboard, mouse and monitor just for debugging and assembling PCs. Since I know that they work, I reduce the number of variables that I have to deal with when I experience a problem. Once the main PC box is assembled and working, I connect the new keyboard, mouse and monitor. If a problem shows up then, I know that it is not likely the PC, but the peripherals I just connected.
Write Things Down
Whenever you discern a piece of evidence, or you figure out something about the current situation your PC is in, document it. Keeping a history of what you discover and what your symptoms are will help you not only with this problem but also with any future difficulties. A log book of some sort is a good idea.
Determine Repeatability
Most problems with a PC fall into one of two categories: either they are repeatable or they are intermittent. A repeatable problem is one where the problem occurs all the time, or always in response to a specific user action. For example, a PC that has a problem that prevents it from booting will probably always fail to boot no matter how many times you reset it. Or you may have an application that whenever you try to run, will crash with an error. You may find that your PC hangs, but only when you move the mouse at the same time that you are communicating using your modem.
In contrast, some problems are intermittent and not repeatable. In some cases, you may have a PC that will usually boot up fine, but one day a month will fail to boot for some reason. An application may work most of the time but occasionally crash. The PC may lock up at seemingly random intervals. Your mouse may work almost all of the time, but one day out of five or ten may give you trouble.
It is helpful to determine if the problem you are experiencing is repeatable, because intermittent difficulties are much more difficult to resolve than repeatable ones. If a problem is repeatable, and there is a specific set of actions that cause the problem, this gives you at least some initial clues about how to find the cause. In addition, you have a way of testing to see if you have resolved the problem when you are trying different solutions. Intermittent problems are much more difficult to deal with.
Determining if a problem is repeatable is pretty simple: try to duplicate the conditions that caused the problem and see if it happens again.
Dealing With Intermittent Problems
Intermittent problems are ones which appear to happen spontaneously or randomly. They seem not to be caused by anything obvious and are not repeatable. They can be extremely difficult and frustrating to diagnose.
Sometimes problems that seem intermittent really aren't; it's just that the specific set of circumstances that causes the problem to occur may be obscure or hard to notice. Spend some time trying to determine what the circumstances are when the problem arises. For example, many problems with crashes or lockups will occur only after the system has been on for more than an hour; some may occur only within the first few minutes that the PC is turned on. You may find that program A only crashes when run at the same time as program B. Or that a particular behavior is associated with a peripheral of some sort.
You will have to be patient when dealing with these sorts of issues. Since the problem is not something you can duplicate at will, you may not be able to systematically work your way towards the ultimate cause. In this situation you may have to employ trial-and-error, making a change and then waiting to see if the problem recurs. It can take days (or longer) sometimes since you have to wait before seeing if the problem happens again. Be patient.
Correlation May Not Imply Causation
I use the word correlation, perhaps slightly incorrectly, to refer to two behaviors or symptoms that appear at the same time. Causation refers to two events where one is responsible for the other's appearance or existence. So what I am saying here is that if you see two strange things happening at the same time on your PC, this does not necessarily mean that one of them has caused the other. They could be coincidental, or it could be that where you think A is causing B, B is really causing A.
For example, you may find two symptoms on your machine: you have frequent file system corruption on your hard disk, and you have a lot of system lockups (hangs, crashes). You may think that the crashes are causing the file system errors, and in many cases this is true. However it is possible for the file system errors to cause the lockups. And it is also possible that both are just symptoms of another underlying cause.
Scan For File System Problems
If the problem you are looking at is anything that involves the hard disk, files, operating system problems, crashes or similar behavior, scan the file system for errors. This is a simple step (which you should be performing regularly as part of a preventive maintenance program) that may tell you immediately that there is a problem with your hard disk. This can also be a useful piece of evidence in troubleshooting some sorts of problems.
Use Diagnostic Tools
If you own diagnostic hardware and/or software tools, they can be useful in pinpointing some kinds of problems. If you feel that you are having a problem that can be detected by a tool, or you feel that running a diagnostic utility may shed more light on what's happening, then make use of the tool to help guide your analysis of the problem.
Next: Diagnostic, Troubleshooting and Repair Tools
Diagnostic, Troubleshooting and Repair Tools
Troubleshooting a problem with a PC can be very difficult and frustrating. Fortunately, there are places to turn for assistance, and there are also tools at your disposal that can make the job much easier. These can be used to identify specific problems within the PC, or just give you more information about the system so you understand better what you are dealing with. This chapter takes a brief look at these tools, which can make your life easier if you are having trouble.
Diagnostic Software
The use of diagnostic software tools can save you a great deal of time. While in my experience these tools do not usually identify what the cause of a system problem is, they often provide hints or at least valuable information about what is in the system and how it is working. Some of these are free or are included with common operating systems at no additional charge, while others are commercial products that range from affordable to rather pricey.
Here are some common software diagnostic tools that you will want to consider including as part of your troubleshooting arsenal:
- Power-On Self Test (POST): I am cheating a bit here because this isn't a separate diagnostic utility; it is in fact built into your system BIOS and it runs every time you start up your PC automatically. It is often the best indicator of system problems; make sure you pay attention to its audio and video messages. Don't disable its error-reporting functions unless you really need to.
- MEM.EXE: This simple utility, built into recent versions of DOS and also Windows 95, provides you with details about your memory configuration, as well as what is currently using your memory. It is especially useful when run with the "/C" parameter (use the "/P" parameter as well to make the output pause when it is scrolling).
- Microsoft Diagnostics: Better known as "MSD.EXE", this is a small DOS utility that takes a brief inventory of the contents of your PC and shows them to you in a text-based format. This is very useful for seeing what disks are in the system, how much memory is installed, and also for checking system resource usage such as LPT ports and IRQs. It will show you what type of BIOS you are using and also what UART chip you have in your serial ports. MSD.EXE is included in later versions of DOS.
Tip: MSD.EXE will work on a Windows 95 machine but is not normally installed as part of the normal Windows 95 installation. You can find it on the Windows 95 CD however and then copy it to the C:\WINDOWS\COMMAND directory to make it work.
Warning: MSD.EXE can give incorrect information when run from within Windows. Shut down to DOS mode before running it ("Restart the computer in MS-DOS mode")
- The Windows 95 Device Manager: This is the most useful tool for identifying system configuration and resource usage information under Windows 95. To access it, open the Control Panel and select the "System" icon. Then select the "Device Manager" tab. You will see a graphical "tree" structure showing you all of your PC hardware. If you select "Properties" while "Computer" (the top-level item) is selected, you will be able to see all the IRQs, DMA channels and I/O addresses in use in your PC; very useful for resolving resource conflicts! The same "Properties" button, pressed after selecting a specific hardware device, will show you driver information, resource settings for the hardware item chosen, and much more.
- Norton System Information: This utility is similar to the Microsoft Diagnostics, only more detailed in its later versions. SI shows a great deal of information about what is in the PC, going well beyond what MSD gives you, but really is still an information utility as opposed to a true diagnostic. This program is part of Symantec's Norton Utilities.
- Microsoft ScanDisk and Norton Disk Doctor: These programs are used to check for hard disk problems. This includes file system corruption and hard disk read errors. They should be used when hard disk problems are suspected.
- Norton Diagnostics: This utility is meant to go beyond the System Information program and actually perform tests on the hardware to identify problems. It includes tests of the processor and motherboard and system memory, and will identify some types of resource conflicts. In reality it is still quite limited in terms of the numbers of problems it will find.
- QAPlus: QAPlus from DiagSoft is a more advanced diagnostic suite that comes in several flavors, depending on what you need to do and how you want to do it. This is a more expensive package but can give you much more detailed information about your system and help identify problem situations as well.
Diagnostic Hardware
While not as commonly used as diagnostic software, there are available several pieces of hardware that can be very helpful in troubleshooting some specific hardware problems. These units are not nearly as popular because they are usually more expensive and are used for specific types of problems, making them less universally applicable than the general software utilities that are widely encountered.
If you are a serious home builder or troubleshooter, you may want to consider one or more of the following:
- Loop-Back Plugs: These are small plugs that go onto the serial and parallel ports on your computer and connect the receive to transmit lines to simulate a connection (they connect the port to itself). Using these with a program like Norton Diagnostics allows for full testing of the operation of these ports, whereas without them only the internal portions of the port can be tested.
Note: The type of plugs used by a particular diagnostic utility may be different than the type used by another program made by another company. Check it out before you buy.
- Multimeter/Ohmmeter: An ohmmeter is a device that measures electrical resistance; a multimeter can measure resistance, voltage and current. These devices are used by electricians, electronics designers and repairpeople. They have come down greatly in price and simple ones can be had inexpensively (while top-end ones of course are still quite a bit of money). A simple ohmmeter is useful primarily for checking for short circuits or open circuits (broken connections, damaged cables, etc.) Multimeters can be used for more extensive electronics testing.
- BIOS POST Cards: It is well-known that the BIOS will, if it finds a problem during its power-on self-test (POST), produce audio "beep codes" and/or video messages that indicate what the trouble is that it found. What many do not realize is that the BIOS of most PCs is designed to send a stream of test codes to a special memory location as it performs its tests, usually 80h. By using a special card designed to capture and display these codes, you can pinpoint exactly where in the power-on self-test a system is hanging up. This can be extremely helpful in debugging very stubborn systems. Scott Wainner has more about one of these cards on his site.
- Test Bed: Some experienced PC repair people keep around an older system that they can use as a test bed for components. It can be very useful to be able to test an unknown device with other components that are known to work, to cut down on the guesswork.
There is also more advanced test hardware available, including devices such as logic probes, oscilloscopes, and many types of specialized component testers. These items are generally very expensive and require specific training to be used properly; they are not for the home PC user or even for most hobbyists. Professional test hardware costs thousands of dollars, which is one reason why doing it yourself is often not an option for many kinds of repairs.
The PC Tool Kit
If you are a very casual PC user you can probably get by hacking around with an old screwdriver and a pocket knife when you need to work on your PC. However, if you plan to do any reasonable amount of work upgrading, building, troubleshooting or repairing PCs, a proper tool kit is a good investment. There are many different types of tool kits available for the computer enthusiast, with tools selected specifically to make the lives of those who work with PCs easier. The right tools can save you a lot of time and help you avoid damage to your equipment (and your hands!)
As with most things, tool kits range widely in size, quality and price. At the low end are $5 kits that are barely any better than the screwdriver and pocket knife. At the high end are kits costing over $100 that include everything but the kitchen sink; these are overkill for most people (including me). Most people can do just fine with a toolkit costing between $15 and $50. These are often available pre-assembled into "genuine artificial leather" carrying cases, sold at most general computer stores.
Here are the items that I would recommend you definitely have in your toolkit; note that some of these don't generally come in pre-assembled toolkits so you will want to add them yourself to the carrying case. Every item in this list is one that I use personally on a regular basis:
- A Good Screwdriver: You want a good screwdriver with magnetic, interchangeable bits and a comfortable handle. At a minimum you should have two different sizes each of Phillips and flat bits. If you've never used a magnetic screwdriver before, you'll wonder how you ever did without one once you do; the magnetic tip will prevent screws falling into the PC, and will let you get access to drives mounted deep in your system case.
- Additional Screwdrivers: An additional Philips screwdriver and an additional flat blade screwdriver, for when you misplace the good one. :^) Actually, it's sometimes useful to have two screwdrivers anyway.
- Needle-Nose Pliers: These are useful for grasping small items and for removing and replacing jumpers on circuit boards.
- Wire Snips: A pair of wire snips for cutting wire and stripping insulation. Some kits include wire strippers for the latter purpose.
- A Small Flashlight: Very useful; the insides of PC boxes are quite dark and there are lots of very small things you will need to see, such as the "pin 1" marking on a connector for example.
- Tweezers: Or even better, a part retriever. A retriever is like a tiny set of retractable claws with a spring-loaded handle. Useful stuff.
- A Large Crowbar: For making adjustments to stubborn hardware. (Just kidding. :^) )
- A Roll of Black Electrical Tape: They never include this in the kits but they should. Used for wrapping wire ends and insulating components.
- A Can of Compressed Air: Very handy for cleaning things hands-free and without using any hazardous liquids.
- A Soft, Lint-Free Cloth: For cleaning the monitor and other components.
The following items I categorize as "optional, but useful". They are tools that are used less frequently, or perhaps only in certain circumstances. They are more for someone who does a fair bit of tinkering or who maintains a number of machines:
- A 3/16" Nut Driver: In fact, tool kits often come with several different sizes of nut drivers (using exchangeable bits), but 3/16" is the really important one, because this is the size of nut most commonly used on PCs. These hexagonal nuts are used as mounting hardware for motherboards and serial and parallel ports. A necessity for anyone who builds their own PCs; if you don't have one you'll be attaching port connectors using a pair of poor-fitting pliers.
- Vice Grips: These are incredibly neat tools that have all sorts of uses.
- Torx Screwdrivers or Bits: These are the star-shaped screwdriver heads that are used to make equipment "tamper-proof". You need these drivers if you want to, well, "tamper".
- An ESD (Electrostatic Discharge) Wrist Strap: This is more of a safety device than a tool; it is used to greatly reduce the chances of static damage to components.
- A Knife: A cutting blade or utility knife of some sort.
The following items are sometimes included in tool kits but in my experience are rarely if ever used:
- Soldering Irons: Essential if you are going to try to do repair of circuit boards, but virtually nobody does this (and with good reason). They are also quite dangerous if used improperly.
- Hex (Allen) Keys: They include these in virtually every commercial tool kit I've ever encountered. I've never seen anything PC related that used them but of course that doesn't mean that there isn't something out there that does...
- Chip Extractors: These are used to remove chips from regular (non-ZIF) sockets, but modern PCs do not generally use these sockets any more. Useful for upgrading much older systems in some cases.
Finally, it is a good idea to accumulate a stockpile of spare parts if you are going to be doing work on PCs. I would recommend the following:
- Screws: I have a "big ol' bag o' screws" of all shapes and sizes that comes in handy when working on various machines.
- Expansion Card Inserts: Save the metal inserts that come out of the back of the case when you put a modem or other card into the PC, as you may need them again later on.
- Drive Faceplates: Similarly, save the plastic faceplates that you remove from the front of cases so you can replace them later if needed.
- Mounting Kits: These sometimes come with retail hard disks and allow you to put a 3.5" drive into a 5.25" bay. Useful when your case has more free 5.25" bays than 3.5" ones.
- Cables: Save any power, IDE, floppy, CD-ROM, or other cables that you accumulate in case you need them later on.
- Keyboard, Mouse, 3.5" Floppy Drive: Keep an extra one of each of these components around to aid in troubleshooting problems by swapping. (Actually, keeping an extra of every major component in the PC is helpful, but the three I mention are here because they are cheap. Most people can't afford to keep an extra hard disk lying around as a spare.)
Scan For Viruses
If you experience any strange behavior on your system, especially either as it is booting, or shortly after it completes, you should always use whatever antivirus software you own to do a full scan of the system before you do anything else. You should use a clean boot floppy if possible.
Do not fall into the "this can't be a virus" trap too easily. Obviously if you turn the power on and nothing happens at all, or if your hard disk won't spin up, or if your system seizes while it is doing the initial BIOS power-on test of your system memory, these are hardware problems that are not caused by a virus. However virtually any other symptom that appears to be a hardware issue can be either a hardware problem or simply a virus that is trying to masquerade as a hardware problem. To remove the complicating factor of always having to wonder "is this a virus?", scan your system before you spend a lot of time chasing ghosts.
Read the Documentation
I never cease to be amazed at how people refuse, steadfastly, to read the documentation that comes with their new hardware or software. They are so excited about their new toy, whatever it is, that they are dying to get it running right away--who has time to start reading? So they start installing whatever-it-is and then have problems. Half the time they end up actually taking more time to get the thing running than they would have if they had just read the instructions in the first place. Meanwhile they often waste a lot of their time and others' asking for technical assistance.
Many, if not most problems are a result of hardware or software that has just been installed into the PC. If you are installing something new, read the documentation that came with the product. Look for files called something like "READ.ME" or "README.1ST" on installation or driver disks, and read them thoroughly before you start. Refer back in the event of trouble. You may be surprised to find that the answer to your dilemma is right there, just waiting for you to read it.
Double-Check Any Recent Changes
Ask yourself: when did the problem I am experiencing first start? If you have just made any change to your system, and are now having a problem with your PC that was not present before the change, the chances are probably 99% that the change is the cause of the problem. This is true even if the problem seems to have absolutely nothing to do with what you changed. If you hunt around, eventually you will find something that links the problem to the change. While it is certainly possible for something to coincidentally stop working at the same time that you make a change, the odds are greatly against this ever happening to you.
Here's an example of something that happened to me, to illustrate what I mean. I opened up my PC to do some work on my system memory. I removed two SIMMs and replaced them with another pair of higher capacity. When I turned on the system next, my floppy disk drive was malfunctioning. Had my floppy drive failed just then? Almost certainly not. As it turned out, my motherboard has the header for the floppy drive cable located directly adjacent to the SIMM slots. While I was working on the memory I had loosened the floppy drive cable.
Sometimes the problems can be even more indirectly related than this. You should in most cases stick to pursuing how the change affected the system, as this is likely to eventually lead you to the cause of the problem.
Simplify, Simplify, Simplify
An absolutely crucial rule to keep in mind when troubleshooting PC system problems is that if there are too many unknowns, it is impossible to determine which one is causing the problem. If you have many possible causes for some difficulty with your system, it can be extremely difficult to narrow down the cause of the problem to any one of them. If you are using a just-installed hard disk with an unfamiliar shareware file utility running on an operating system you just upgraded last week, and now your system has problems, how on earth are you ever going to know what is causing them?
In order to have a fighting chance at figuring out what is going on, you must simplify the situation as much as possible so that it becomes much more obvious what is responsible for the difficulty. This means reducing the number of variables to whatever degree possible. One important way of doing this is undoing or double-checking any changes made to the system.
In addition, I have identified the following items as often being responsible for erratic behavior that can complicate troubleshooting. I would recommend that they be eliminated or temporarily disabled when trying to diagnose a system problem:
- Power Management: Power management is a great idea in theory but in many ways is just "not ready for prime time". Power management routines can cause symptoms that appear to be hardware malfunctions, such as screens that turn off unexpectedly or hard disks that spin down. They also can cause crashes of software that doesn't know how to deal with them. If you want to use power management, turning it off until the problem is resolved is wise.
- Overclocked Hardware and Aggressive BIOS Settings: I do not believe in overclocking. If you insist on doing it, don't be surprised if you have system problems! Scale things back until you can figure out what the problem is. Similarly, if you are "pushing the envelope" in trying to squeeze maximum performance by tuning your BIOS memory timings and other settings very aggressively, try resetting them to more conservative values when troubleshooting.
- Experimental or Beta Software: This software is still in the test process and is likely to have bugs--that is why it is labeled as "beta"! For an end application this is usually no big deal, since any crashes or other problems will be limited to that application and therefore somewhat obvious. Running beta operating systems, drivers or other low-level software however is asking for trouble, and you should try to eliminate these possible sources of confusion when trying to debug your system.
- "Creative" Configurations: The more "unusual" things that you have going on in your system, the more likely that you are going to have a conflict caused by one of these strange pieces of hardware or software. A system that is loaded with unusual utilities, terminate-and-stay-resident programs, an old 8-bit network card salvaged from a 286, etc. will often have more problems than a stock Pentium box with a normal Windows 95 installation. To whatever extent possible, disable these items while troubleshooting. Also try to avoid using unusual low-level software whenever possible.
- Excessive Connections: If the PC is on a network or is connected to a large number or variety of peripheral devices, you may want to try to disconnect those and see if there is any impact on the problem.
In general you want to avoid the unusual or the unknown when troubleshooting. One way to simplify the software environment during diagnosis is to use a boot floppy to "boot clean" and bypass the special drivers and software that you normally load when you boot from your hard disk. You can also use the {F8} key when DOS or Windows 95 are booting to bypass your startup files, basically accomplishing the same thing. You want to be especially wary of software that sits in the background and activates without you specifically telling it to, as this can confound your troubleshooting efforts.
To whatever extent possible, disable as much as you can when trying to figure out a problem. The more funky software utilities, screen savers and cute peripherals you disable now, the more chance you have of finding out which one it is that is causing the problem later on.
General Diagnostic Techniques
This section discusses the rules of thumb that I recommend you employ when you are troubleshooting. These are the general techniques and ideas that can help you solve all sorts of different problems with your PC (and even in other parts of your life...) It often is worth trying these general tips before attempting to find a specific answer to your problem in the Troubleshooting Expert.
Get to the Root Cause: Ask "Why" Five Times
One of the keys to truly solving a problem is to first truly understand it. It is easy to observe that too often, people are quick to jump to solve a problem before they really understand what it is. It is easy to see a symptom and think that you know what is causing it, but surprisingly often, if you take the time to explore more deeply you will find that what you thought was the cause is in fact just another symptom, and that the problem lies much deeper within.
One technique that you can employ is borrowed from Japanese manufacturing management theory. This method is used to help identify the real causes of problems that occur on the manufacturing floor. The goal is not to simply correct the effects of the problem, but to find out the root cause of why the problem is occurring so that we can ensure that it will not happen in the future.
One simple way to do this is called asking why' five times. The idea is that by the time you have asked "why" the fifth time, you will be at the root cause. It isn't always that simple, but the exercise can be surprisingly insightful in helping you figure out what is really going on, and can help you avoid "quick fix" solutions that are really just band-aids and don't resolve anything. It is especially useful for tackling chronic problems that show up over and over again in a system; it is less useful for problems that are unlikely to recur.
Here's an example. Let's suppose your hard disk is having a problem with bad sectors showing up. The knee-jerk reaction to this happening is "the hard disk is bad, replace it". Instead, ask yourself:
- "Why are bad sectors showing up when they didn't before?"
- "The hard disk is being corrupted somehow; something has changed."
- "Why?"
- "The timing of the system probably has been changed in some way."
- "Why?"
- "Hmm, I installed a new hard drive in that system just last week, perhaps it is causing the problem."
- "Why?"
- "The hard drive may not have been installed correctly, or perhaps I should have used a different channel. Or maybe it was the bus-mastering driver that came with the new drive, the old drive may not be able to support it".
You can see the general idea; the answers will differ in every case, but it is the process itself that is useful. Another example: imagine an office that has just had a catastrophic data loss due to a PC crashing after an electrical storm. The initially identified cause of this problem was a lack of a UPS on the PC, which would have protected the system from the electrical storm (usually). So the MIS department starts drafting purchase requisitions for UPSes for all the PCs. Instead, if they looked at the problem carefully, they may have employed the following process:
- "Remember that the problem here was the data loss, not the crashing of the PC. Why was the data lost?"
- "The PC's hard disk crashed."
- "Why would that cause the data to be lost?"
- "Because we didn't have a backup."
- "Why wasn't there a backup?"
- "There is a tape backup unit on every PC in the building but very few of them are being used."
- "Why?"
- "Because most of the PC users do not know how to use them."
- "Why?"
- "They have never been given any training in their use."
- "Why?"
- "Because we don't have a budget for PC hardware training."
As you can see, the root cause that we came up with here is very different than the cause we saw when we only scratched the surface. This means of course that the solution will be very different as well; it might well make a lot more sense here to spend a small amount of money on training instead of a large amount of money on UPSes for every PC.
Be Observant and Look For Evidence
It is sometimes the subtle signs that can provide the triggering piece of information that leads to discovering the source of a problem. In particular, you want to carefully investigate anything that seems "unusual", "wrong" or "surprising". These are often the clues that will get you going down the right track to figuring out what is giving you trouble. If you are observant and careful about what you are doing, you will be much more likely to pick up on these things than if you are in a hurry or you are quick to discount your findings as "unrelated to the problem".
Use the Process of Elimination
Virtually all problems with PCs involve more than one component or subsystem. The difficulty is usually in figuring out which component is responsible for the problem. Using the process of elimination, however, you can usually narrow the problem down rather quickly by making small logical changes and observing the impact on the problem. Your objective is to isolate the cause of the problem so you can correct it.
The key is to make only one change at a time and then see if the problem goes away; if it does, then whatever you changed is likely responsible for the problem (although it could be fixing the problem indirectly in some cases.) If you make more than one change at a time, you cannot readily discern which change was responsible for fixing the problem.
You will want to first check the most probable sources of the problem, and also the things that are easiest to change. For example, if you are having a problem with your disk drive being recognized, it's a lot easier and cheaper to explore things like double-checking jumpers and connections or replacing the interface cable, than it is to try replacing the drive itself. That is something you'd only do after you had eliminated all the other possibilities (or if the evidence implicated the hard disk directly).
Here's a simple example. Let's suppose one morning your PC will not turn on. You hit the switch and nothing happens. There could be many different possible causes for this problem: the power to the house could be out; there could be a malfunction in the wall socket; the surge suppressor that the system is plugged into might have blown; the electrical cord may be loose; the power supply could be damaged. To figure out what is going on you need to eliminate these variables by making small changes and seeing what happens. For example:
- Change the wall socket you are using. If the PC now boots, you have isolated the cause to the electrical wiring in the house.
- If the problem persists, examine the surge suppressor. Change it, or temporarily bypass it and plug the PC into the wall directly. If it now works, the surge suppressor is the problem.
- If the problem still isn't fixed, try changing the power cord.
- If the problem persists still, you may then have to open up the box and look at the power supply unit to see if it might need replacing.
Realize that the key here is making these changes one at a time. If you approach this problem by changing the wall socket you use, bypassing the surge suppressor, and changing the power cord all at once, your problem may go away but how will you know what caused it? This is a valid way to troubleshoot if you have to get the system back up immediately, however. You can then undo the changes one at a time later on to find out what the cause is, in effect doing the same single-change-at-a-time process, but in reverse.
Most problems with the PC can be diagnosed using this sort of process; it's very powerful. You will notice many sequences like the one above in The Troubleshooting Expert.
Do One Upgrade or Assembly Step At a Time
Changes made to the system are the most frequent cause of problems; this is the nature of change. Much the way many problems can be diagnosed by using the process of elimination through making single changes to the system, you can avoid or detect problems with upgrades or new installations by going "one step at a time" in your changes as well. New system installations or major upgrades often can have problems that are very difficult to diagnose simply because there are so many modifications being performed at the same time.
When you build a new PC you will probably be assembling a large number of components that have never been used or tested before. To whatever extent possible, try to go as slowly as possible when assembling the machine. For example, when assembling a new PC from scratch, it is always best to first make sure the basic system is working, with the only expansion card in the system the video card (if applicable). Adding the sound card, network card and other devices should be done later on, and one at a time if possible. Similarly, do not try to do major software or operating system upgrades at the same time that you make hardware changes. Doing this can make it very difficult for you to troubleshoot any system problems.
If you do make multiple changes at once, try retracing your steps. Undo the changes you have made one at a time and see if you can identify the change that caused the problem that way.
Make Use of Components That You Know Work
One valuable resource while troubleshooting is a "stockpile" of extra components. I put "stockpile" in quotes because it doesn't have to be an actual stock of components, just a resource that will let you borrow components that you know work while you are troubleshooting your system. Another functional PC that uses the same or similar components is normally quite adequate. Being in an office environment can be very helpful this way since there are many PCs around to work with, and often they are similar to one another; at home it may be more difficult if you only have the one machine.
These extra components can be useful when you are employing the process of elimination to resolve problems. For example, suppose that there is no video on your monitor when you boot up the PC. After eliminating the simple stuff (cable disconnected, brightness turned down far too much, etc.) there are two components that are most likely to be bad: the video card or the monitor. An easy thing to try is to swap monitors with another machine that is working properly. If the second monitor also doesn't work, and your original monitor works on the other machine, you can feel quite confident that it isn't the monitor that is the problem here.
You can sometimes avoid problems in assembling a new PC by testing out some of the components you are going to use before you begin. For example, take your new mouse, keyboard and monitor and attach them to a PC that is known to work, and test them out. This way if you have problems when you build your new system you will know it isn't due to these components. At my office, I keep an extra keyboard, mouse and monitor just for debugging and assembling PCs. Since I know that they work, I reduce the number of variables that I have to deal with when I experience a problem. Once the main PC box is assembled and working, I connect the new keyboard, mouse and monitor. If a problem shows up then, I know that it is not likely the PC, but the peripherals I just connected.
Write Things Down
Whenever you discern a piece of evidence, or you figure out something about the current situation your PC is in, document it. Keeping a history of what you discover and what your symptoms are will help you not only with this problem but also with any future difficulties. A log book of some sort is a good idea.
Determine Repeatability
Most problems with a PC fall into one of two categories: either they are repeatable or they are intermittent. A repeatable problem is one where the problem occurs all the time, or always in response to a specific user action. For example, a PC that has a problem that prevents it from booting will probably always fail to boot no matter how many times you reset it. Or you may have an application that whenever you try to run, will crash with an error. You may find that your PC hangs, but only when you move the mouse at the same time that you are communicating using your modem.
In contrast, some problems are intermittent and not repeatable. In some cases, you may have a PC that will usually boot up fine, but one day a month will fail to boot for some reason. An application may work most of the time but occasionally crash. The PC may lock up at seemingly random intervals. Your mouse may work almost all of the time, but one day out of five or ten may give you trouble.
It is helpful to determine if the problem you are experiencing is repeatable, because intermittent difficulties are much more difficult to resolve than repeatable ones. If a problem is repeatable, and there is a specific set of actions that cause the problem, this gives you at least some initial clues about how to find the cause. In addition, you have a way of testing to see if you have resolved the problem when you are trying different solutions. Intermittent problems are much more difficult to deal with.
Determining if a problem is repeatable is pretty simple: try to duplicate the conditions that caused the problem and see if it happens again.
Dealing With Intermittent Problems
Intermittent problems are ones which appear to happen spontaneously or randomly. They seem not to be caused by anything obvious and are not repeatable. They can be extremely difficult and frustrating to diagnose.
Sometimes problems that seem intermittent really aren't; it's just that the specific set of circumstances that causes the problem to occur may be obscure or hard to notice. Spend some time trying to determine what the circumstances are when the problem arises. For example, many problems with crashes or lockups will occur only after the system has been on for more than an hour; some may occur only within the first few minutes that the PC is turned on. You may find that program A only crashes when run at the same time as program B. Or that a particular behavior is associated with a peripheral of some sort.
You will have to be patient when dealing with these sorts of issues. Since the problem is not something you can duplicate at will, you may not be able to systematically work your way towards the ultimate cause. In this situation you may have to employ trial-and-error, making a change and then waiting to see if the problem recurs. It can take days (or longer) sometimes since you have to wait before seeing if the problem happens again. Be patient.
Correlation May Not Imply Causation
I use the word correlation, perhaps slightly incorrectly, to refer to two behaviors or symptoms that appear at the same time. Causation refers to two events where one is responsible for the other's appearance or existence. So what I am saying here is that if you see two strange things happening at the same time on your PC, this does not necessarily mean that one of them has caused the other. They could be coincidental, or it could be that where you think A is causing B, B is really causing A.
For example, you may find two symptoms on your machine: you have frequent file system corruption on your hard disk, and you have a lot of system lockups (hangs, crashes). You may think that the crashes are causing the file system errors, and in many cases this is true. However it is possible for the file system errors to cause the lockups. And it is also possible that both are just symptoms of another underlying cause.
Scan For File System Problems
If the problem you are looking at is anything that involves the hard disk, files, operating system problems, crashes or similar behavior, scan the file system for errors. This is a simple step (which you should be performing regularly as part of a preventive maintenance program) that may tell you immediately that there is a problem with your hard disk. This can also be a useful piece of evidence in troubleshooting some sorts of problems.
Use Diagnostic Tools
If you own diagnostic hardware and/or software tools, they can be useful in pinpointing some kinds of problems. If you feel that you are having a problem that can be detected by a tool, or you feel that running a diagnostic utility may shed more light on what's happening, then make use of the tool to help guide your analysis of the problem.
Next: Diagnostic, Troubleshooting and Repair Tools
Diagnostic, Troubleshooting and Repair Tools
Troubleshooting a problem with a PC can be very difficult and frustrating. Fortunately, there are places to turn for assistance, and there are also tools at your disposal that can make the job much easier. These can be used to identify specific problems within the PC, or just give you more information about the system so you understand better what you are dealing with. This chapter takes a brief look at these tools, which can make your life easier if you are having trouble.
Diagnostic Software
The use of diagnostic software tools can save you a great deal of time. While in my experience these tools do not usually identify what the cause of a system problem is, they often provide hints or at least valuable information about what is in the system and how it is working. Some of these are free or are included with common operating systems at no additional charge, while others are commercial products that range from affordable to rather pricey.
Here are some common software diagnostic tools that you will want to consider including as part of your troubleshooting arsenal:
- Power-On Self Test (POST): I am cheating a bit here because this isn't a separate diagnostic utility; it is in fact built into your system BIOS and it runs every time you start up your PC automatically. It is often the best indicator of system problems; make sure you pay attention to its audio and video messages. Don't disable its error-reporting functions unless you really need to.
- MEM.EXE: This simple utility, built into recent versions of DOS and also Windows 95, provides you with details about your memory configuration, as well as what is currently using your memory. It is especially useful when run with the "/C" parameter (use the "/P" parameter as well to make the output pause when it is scrolling).
- Microsoft Diagnostics: Better known as "MSD.EXE", this is a small DOS utility that takes a brief inventory of the contents of your PC and shows them to you in a text-based format. This is very useful for seeing what disks are in the system, how much memory is installed, and also for checking system resource usage such as LPT ports and IRQs. It will show you what type of BIOS you are using and also what UART chip you have in your serial ports. MSD.EXE is included in later versions of DOS.
Tip: MSD.EXE will work on a Windows 95 machine but is not normally installed as part of the normal Windows 95 installation. You can find it on the Windows 95 CD however and then copy it to the C:\WINDOWS\COMMAND directory to make it work.
Warning: MSD.EXE can give incorrect information when run from within Windows. Shut down to DOS mode before running it ("Restart the computer in MS-DOS mode")
- The Windows 95 Device Manager: This is the most useful tool for identifying system configuration and resource usage information under Windows 95. To access it, open the Control Panel and select the "System" icon. Then select the "Device Manager" tab. You will see a graphical "tree" structure showing you all of your PC hardware. If you select "Properties" while "Computer" (the top-level item) is selected, you will be able to see all the IRQs, DMA channels and I/O addresses in use in your PC; very useful for resolving resource conflicts! The same "Properties" button, pressed after selecting a specific hardware device, will show you driver information, resource settings for the hardware item chosen, and much more.
- Norton System Information: This utility is similar to the Microsoft Diagnostics, only more detailed in its later versions. SI shows a great deal of information about what is in the PC, going well beyond what MSD gives you, but really is still an information utility as opposed to a true diagnostic. This program is part of Symantec's Norton Utilities.
- Microsoft ScanDisk and Norton Disk Doctor: These programs are used to check for hard disk problems. This includes file system corruption and hard disk read errors. They should be used when hard disk problems are suspected.
- Norton Diagnostics: This utility is meant to go beyond the System Information program and actually perform tests on the hardware to identify problems. It includes tests of the processor and motherboard and system memory, and will identify some types of resource conflicts. In reality it is still quite limited in terms of the numbers of problems it will find.
- QAPlus: QAPlus from DiagSoft is a more advanced diagnostic suite that comes in several flavors, depending on what you need to do and how you want to do it. This is a more expensive package but can give you much more detailed information about your system and help identify problem situations as well.
Diagnostic Hardware
While not as commonly used as diagnostic software, there are available several pieces of hardware that can be very helpful in troubleshooting some specific hardware problems. These units are not nearly as popular because they are usually more expensive and are used for specific types of problems, making them less universally applicable than the general software utilities that are widely encountered.
If you are a serious home builder or troubleshooter, you may want to consider one or more of the following:
- Loop-Back Plugs: These are small plugs that go onto the serial and parallel ports on your computer and connect the receive to transmit lines to simulate a connection (they connect the port to itself). Using these with a program like Norton Diagnostics allows for full testing of the operation of these ports, whereas without them only the internal portions of the port can be tested.
Note: The type of plugs used by a particular diagnostic utility may be different than the type used by another program made by another company. Check it out before you buy.
- Multimeter/Ohmmeter: An ohmmeter is a device that measures electrical resistance; a multimeter can measure resistance, voltage and current. These devices are used by electricians, electronics designers and repairpeople. They have come down greatly in price and simple ones can be had inexpensively (while top-end ones of course are still quite a bit of money). A simple ohmmeter is useful primarily for checking for short circuits or open circuits (broken connections, damaged cables, etc.) Multimeters can be used for more extensive electronics testing.
- BIOS POST Cards: It is well-known that the BIOS will, if it finds a problem during its power-on self-test (POST), produce audio "beep codes" and/or video messages that indicate what the trouble is that it found. What many do not realize is that the BIOS of most PCs is designed to send a stream of test codes to a special memory location as it performs its tests, usually 80h. By using a special card designed to capture and display these codes, you can pinpoint exactly where in the power-on self-test a system is hanging up. This can be extremely helpful in debugging very stubborn systems. Scott Wainner has more about one of these cards on his site.
- Test Bed: Some experienced PC repair people keep around an older system that they can use as a test bed for components. It can be very useful to be able to test an unknown device with other components that are known to work, to cut down on the guesswork.
There is also more advanced test hardware available, including devices such as logic probes, oscilloscopes, and many types of specialized component testers. These items are generally very expensive and require specific training to be used properly; they are not for the home PC user or even for most hobbyists. Professional test hardware costs thousands of dollars, which is one reason why doing it yourself is often not an option for many kinds of repairs.
The PC Tool Kit
If you are a very casual PC user you can probably get by hacking around with an old screwdriver and a pocket knife when you need to work on your PC. However, if you plan to do any reasonable amount of work upgrading, building, troubleshooting or repairing PCs, a proper tool kit is a good investment. There are many different types of tool kits available for the computer enthusiast, with tools selected specifically to make the lives of those who work with PCs easier. The right tools can save you a lot of time and help you avoid damage to your equipment (and your hands!)
As with most things, tool kits range widely in size, quality and price. At the low end are $5 kits that are barely any better than the screwdriver and pocket knife. At the high end are kits costing over $100 that include everything but the kitchen sink; these are overkill for most people (including me). Most people can do just fine with a toolkit costing between $15 and $50. These are often available pre-assembled into "genuine artificial leather" carrying cases, sold at most general computer stores.
Here are the items that I would recommend you definitely have in your toolkit; note that some of these don't generally come in pre-assembled toolkits so you will want to add them yourself to the carrying case. Every item in this list is one that I use personally on a regular basis:
- A Good Screwdriver: You want a good screwdriver with magnetic, interchangeable bits and a comfortable handle. At a minimum you should have two different sizes each of Phillips and flat bits. If you've never used a magnetic screwdriver before, you'll wonder how you ever did without one once you do; the magnetic tip will prevent screws falling into the PC, and will let you get access to drives mounted deep in your system case.
- Additional Screwdrivers: An additional Philips screwdriver and an additional flat blade screwdriver, for when you misplace the good one. :^) Actually, it's sometimes useful to have two screwdrivers anyway.
- Needle-Nose Pliers: These are useful for grasping small items and for removing and replacing jumpers on circuit boards.
- Wire Snips: A pair of wire snips for cutting wire and stripping insulation. Some kits include wire strippers for the latter purpose.
- A Small Flashlight: Very useful; the insides of PC boxes are quite dark and there are lots of very small things you will need to see, such as the "pin 1" marking on a connector for example.
- Tweezers: Or even better, a part retriever. A retriever is like a tiny set of retractable claws with a spring-loaded handle. Useful stuff.
- A Large Crowbar: For making adjustments to stubborn hardware. (Just kidding. :^) )
- A Roll of Black Electrical Tape: They never include this in the kits but they should. Used for wrapping wire ends and insulating components.
- A Can of Compressed Air: Very handy for cleaning things hands-free and without using any hazardous liquids.
- A Soft, Lint-Free Cloth: For cleaning the monitor and other components.
The following items I categorize as "optional, but useful". They are tools that are used less frequently, or perhaps only in certain circumstances. They are more for someone who does a fair bit of tinkering or who maintains a number of machines:
- A 3/16" Nut Driver: In fact, tool kits often come with several different sizes of nut drivers (using exchangeable bits), but 3/16" is the really important one, because this is the size of nut most commonly used on PCs. These hexagonal nuts are used as mounting hardware for motherboards and serial and parallel ports. A necessity for anyone who builds their own PCs; if you don't have one you'll be attaching port connectors using a pair of poor-fitting pliers.
- Vice Grips: These are incredibly neat tools that have all sorts of uses.
- Torx Screwdrivers or Bits: These are the star-shaped screwdriver heads that are used to make equipment "tamper-proof". You need these drivers if you want to, well, "tamper".
- An ESD (Electrostatic Discharge) Wrist Strap: This is more of a safety device than a tool; it is used to greatly reduce the chances of static damage to components.
- A Knife: A cutting blade or utility knife of some sort.
The following items are sometimes included in tool kits but in my experience are rarely if ever used:
- Soldering Irons: Essential if you are going to try to do repair of circuit boards, but virtually nobody does this (and with good reason). They are also quite dangerous if used improperly.
- Hex (Allen) Keys: They include these in virtually every commercial tool kit I've ever encountered. I've never seen anything PC related that used them but of course that doesn't mean that there isn't something out there that does...
- Chip Extractors: These are used to remove chips from regular (non-ZIF) sockets, but modern PCs do not generally use these sockets any more. Useful for upgrading much older systems in some cases.
Finally, it is a good idea to accumulate a stockpile of spare parts if you are going to be doing work on PCs. I would recommend the following:
- Screws: I have a "big ol' bag o' screws" of all shapes and sizes that comes in handy when working on various machines.
- Expansion Card Inserts: Save the metal inserts that come out of the back of the case when you put a modem or other card into the PC, as you may need them again later on.
- Drive Faceplates: Similarly, save the plastic faceplates that you remove from the front of cases so you can replace them later if needed.
- Mounting Kits: These sometimes come with retail hard disks and allow you to put a 3.5" drive into a 5.25" bay. Useful when your case has more free 5.25" bays than 3.5" ones.
- Cables: Save any power, IDE, floppy, CD-ROM, or other cables that you accumulate in case you need them later on.
- Keyboard, Mouse, 3.5" Floppy Drive: Keep an extra one of each of these components around to aid in troubleshooting problems by swapping. (Actually, keeping an extra of every major component in the PC is helpful, but the three I mention are here because they are cheap. Most people can't afford to keep an extra hard disk lying around as a spare.)
Double-Check Any Recent Changes
Ask yourself: when did the problem I am experiencing first start? If you have just made any change to your system, and are now having a problem with your PC that was not present before the change, the chances are probably 99% that the change is the cause of the problem. This is true even if the problem seems to have absolutely nothing to do with what you changed. If you hunt around, eventually you will find something that links the problem to the change. While it is certainly possible for something to coincidentally stop working at the same time that you make a change, the odds are greatly against this ever happening to you.
Here's an example of something that happened to me, to illustrate what I mean. I opened up my PC to do some work on my system memory. I removed two SIMMs and replaced them with another pair of higher capacity. When I turned on the system next, my floppy disk drive was malfunctioning. Had my floppy drive failed just then? Almost certainly not. As it turned out, my motherboard has the header for the floppy drive cable located directly adjacent to the SIMM slots. While I was working on the memory I had loosened the floppy drive cable.
Sometimes the problems can be even more indirectly related than this. You should in most cases stick to pursuing how the change affected the system, as this is likely to eventually lead you to the cause of the problem.
Simplify, Simplify, Simplify
An absolutely crucial rule to keep in mind when troubleshooting PC system problems is that if there are too many unknowns, it is impossible to determine which one is causing the problem. If you have many possible causes for some difficulty with your system, it can be extremely difficult to narrow down the cause of the problem to any one of them. If you are using a just-installed hard disk with an unfamiliar shareware file utility running on an operating system you just upgraded last week, and now your system has problems, how on earth are you ever going to know what is causing them?
In order to have a fighting chance at figuring out what is going on, you must simplify the situation as much as possible so that it becomes much more obvious what is responsible for the difficulty. This means reducing the number of variables to whatever degree possible. One important way of doing this is undoing or double-checking any changes made to the system.
In addition, I have identified the following items as often being responsible for erratic behavior that can complicate troubleshooting. I would recommend that they be eliminated or temporarily disabled when trying to diagnose a system problem:
- Power Management: Power management is a great idea in theory but in many ways is just "not ready for prime time". Power management routines can cause symptoms that appear to be hardware malfunctions, such as screens that turn off unexpectedly or hard disks that spin down. They also can cause crashes of software that doesn't know how to deal with them. If you want to use power management, turning it off until the problem is resolved is wise.
- Overclocked Hardware and Aggressive BIOS Settings: I do not believe in overclocking. If you insist on doing it, don't be surprised if you have system problems! Scale things back until you can figure out what the problem is. Similarly, if you are "pushing the envelope" in trying to squeeze maximum performance by tuning your BIOS memory timings and other settings very aggressively, try resetting them to more conservative values when troubleshooting.
- Experimental or Beta Software: This software is still in the test process and is likely to have bugs--that is why it is labeled as "beta"! For an end application this is usually no big deal, since any crashes or other problems will be limited to that application and therefore somewhat obvious. Running beta operating systems, drivers or other low-level software however is asking for trouble, and you should try to eliminate these possible sources of confusion when trying to debug your system.
- "Creative" Configurations: The more "unusual" things that you have going on in your system, the more likely that you are going to have a conflict caused by one of these strange pieces of hardware or software. A system that is loaded with unusual utilities, terminate-and-stay-resident programs, an old 8-bit network card salvaged from a 286, etc. will often have more problems than a stock Pentium box with a normal Windows 95 installation. To whatever extent possible, disable these items while troubleshooting. Also try to avoid using unusual low-level software whenever possible.
- Excessive Connections: If the PC is on a network or is connected to a large number or variety of peripheral devices, you may want to try to disconnect those and see if there is any impact on the problem.
In general you want to avoid the unusual or the unknown when troubleshooting. One way to simplify the software environment during diagnosis is to use a boot floppy to "boot clean" and bypass the special drivers and software that you normally load when you boot from your hard disk. You can also use the {F8} key when DOS or Windows 95 are booting to bypass your startup files, basically accomplishing the same thing. You want to be especially wary of software that sits in the background and activates without you specifically telling it to, as this can confound your troubleshooting efforts.
To whatever extent possible, disable as much as you can when trying to figure out a problem. The more funky software utilities, screen savers and cute peripherals you disable now, the more chance you have of finding out which one it is that is causing the problem later on.
General Diagnostic Techniques
This section discusses the rules of thumb that I recommend you employ when you are troubleshooting. These are the general techniques and ideas that can help you solve all sorts of different problems with your PC (and even in other parts of your life...) It often is worth trying these general tips before attempting to find a specific answer to your problem in the Troubleshooting Expert.
Get to the Root Cause: Ask "Why" Five Times
One of the keys to truly solving a problem is to first truly understand it. It is easy to observe that too often, people are quick to jump to solve a problem before they really understand what it is. It is easy to see a symptom and think that you know what is causing it, but surprisingly often, if you take the time to explore more deeply you will find that what you thought was the cause is in fact just another symptom, and that the problem lies much deeper within.
One technique that you can employ is borrowed from Japanese manufacturing management theory. This method is used to help identify the real causes of problems that occur on the manufacturing floor. The goal is not to simply correct the effects of the problem, but to find out the root cause of why the problem is occurring so that we can ensure that it will not happen in the future.
One simple way to do this is called asking why' five times. The idea is that by the time you have asked "why" the fifth time, you will be at the root cause. It isn't always that simple, but the exercise can be surprisingly insightful in helping you figure out what is really going on, and can help you avoid "quick fix" solutions that are really just band-aids and don't resolve anything. It is especially useful for tackling chronic problems that show up over and over again in a system; it is less useful for problems that are unlikely to recur.
Here's an example. Let's suppose your hard disk is having a problem with bad sectors showing up. The knee-jerk reaction to this happening is "the hard disk is bad, replace it". Instead, ask yourself:
- "Why are bad sectors showing up when they didn't before?"
- "The hard disk is being corrupted somehow; something has changed."
- "Why?"
- "The timing of the system probably has been changed in some way."
- "Why?"
- "Hmm, I installed a new hard drive in that system just last week, perhaps it is causing the problem."
- "Why?"
- "The hard drive may not have been installed correctly, or perhaps I should have used a different channel. Or maybe it was the bus-mastering driver that came with the new drive, the old drive may not be able to support it".
You can see the general idea; the answers will differ in every case, but it is the process itself that is useful. Another example: imagine an office that has just had a catastrophic data loss due to a PC crashing after an electrical storm. The initially identified cause of this problem was a lack of a UPS on the PC, which would have protected the system from the electrical storm (usually). So the MIS department starts drafting purchase requisitions for UPSes for all the PCs. Instead, if they looked at the problem carefully, they may have employed the following process:
- "Remember that the problem here was the data loss, not the crashing of the PC. Why was the data lost?"
- "The PC's hard disk crashed."
- "Why would that cause the data to be lost?"
- "Because we didn't have a backup."
- "Why wasn't there a backup?"
- "There is a tape backup unit on every PC in the building but very few of them are being used."
- "Why?"
- "Because most of the PC users do not know how to use them."
- "Why?"
- "They have never been given any training in their use."
- "Why?"
- "Because we don't have a budget for PC hardware training."
As you can see, the root cause that we came up with here is very different than the cause we saw when we only scratched the surface. This means of course that the solution will be very different as well; it might well make a lot more sense here to spend a small amount of money on training instead of a large amount of money on UPSes for every PC.
Be Observant and Look For Evidence
It is sometimes the subtle signs that can provide the triggering piece of information that leads to discovering the source of a problem. In particular, you want to carefully investigate anything that seems "unusual", "wrong" or "surprising". These are often the clues that will get you going down the right track to figuring out what is giving you trouble. If you are observant and careful about what you are doing, you will be much more likely to pick up on these things than if you are in a hurry or you are quick to discount your findings as "unrelated to the problem".
Use the Process of Elimination
Virtually all problems with PCs involve more than one component or subsystem. The difficulty is usually in figuring out which component is responsible for the problem. Using the process of elimination, however, you can usually narrow the problem down rather quickly by making small logical changes and observing the impact on the problem. Your objective is to isolate the cause of the problem so you can correct it.
The key is to make only one change at a time and then see if the problem goes away; if it does, then whatever you changed is likely responsible for the problem (although it could be fixing the problem indirectly in some cases.) If you make more than one change at a time, you cannot readily discern which change was responsible for fixing the problem.
You will want to first check the most probable sources of the problem, and also the things that are easiest to change. For example, if you are having a problem with your disk drive being recognized, it's a lot easier and cheaper to explore things like double-checking jumpers and connections or replacing the interface cable, than it is to try replacing the drive itself. That is something you'd only do after you had eliminated all the other possibilities (or if the evidence implicated the hard disk directly).
Here's a simple example. Let's suppose one morning your PC will not turn on. You hit the switch and nothing happens. There could be many different possible causes for this problem: the power to the house could be out; there could be a malfunction in the wall socket; the surge suppressor that the system is plugged into might have blown; the electrical cord may be loose; the power supply could be damaged. To figure out what is going on you need to eliminate these variables by making small changes and seeing what happens. For example:
- Change the wall socket you are using. If the PC now boots, you have isolated the cause to the electrical wiring in the house.
- If the problem persists, examine the surge suppressor. Change it, or temporarily bypass it and plug the PC into the wall directly. If it now works, the surge suppressor is the problem.
- If the problem still isn't fixed, try changing the power cord.
- If the problem persists still, you may then have to open up the box and look at the power supply unit to see if it might need replacing.
Realize that the key here is making these changes one at a time. If you approach this problem by changing the wall socket you use, bypassing the surge suppressor, and changing the power cord all at once, your problem may go away but how will you know what caused it? This is a valid way to troubleshoot if you have to get the system back up immediately, however. You can then undo the changes one at a time later on to find out what the cause is, in effect doing the same single-change-at-a-time process, but in reverse.
Most problems with the PC can be diagnosed using this sort of process; it's very powerful. You will notice many sequences like the one above in The Troubleshooting Expert.
Do One Upgrade or Assembly Step At a Time
Changes made to the system are the most frequent cause of problems; this is the nature of change. Much the way many problems can be diagnosed by using the process of elimination through making single changes to the system, you can avoid or detect problems with upgrades or new installations by going "one step at a time" in your changes as well. New system installations or major upgrades often can have problems that are very difficult to diagnose simply because there are so many modifications being performed at the same time.
When you build a new PC you will probably be assembling a large number of components that have never been used or tested before. To whatever extent possible, try to go as slowly as possible when assembling the machine. For example, when assembling a new PC from scratch, it is always best to first make sure the basic system is working, with the only expansion card in the system the video card (if applicable). Adding the sound card, network card and other devices should be done later on, and one at a time if possible. Similarly, do not try to do major software or operating system upgrades at the same time that you make hardware changes. Doing this can make it very difficult for you to troubleshoot any system problems.
If you do make multiple changes at once, try retracing your steps. Undo the changes you have made one at a time and see if you can identify the change that caused the problem that way.
Make Use of Components That You Know Work
One valuable resource while troubleshooting is a "stockpile" of extra components. I put "stockpile" in quotes because it doesn't have to be an actual stock of components, just a resource that will let you borrow components that you know work while you are troubleshooting your system. Another functional PC that uses the same or similar components is normally quite adequate. Being in an office environment can be very helpful this way since there are many PCs around to work with, and often they are similar to one another; at home it may be more difficult if you only have the one machine.
These extra components can be useful when you are employing the process of elimination to resolve problems. For example, suppose that there is no video on your monitor when you boot up the PC. After eliminating the simple stuff (cable disconnected, brightness turned down far too much, etc.) there are two components that are most likely to be bad: the video card or the monitor. An easy thing to try is to swap monitors with another machine that is working properly. If the second monitor also doesn't work, and your original monitor works on the other machine, you can feel quite confident that it isn't the monitor that is the problem here.
You can sometimes avoid problems in assembling a new PC by testing out some of the components you are going to use before you begin. For example, take your new mouse, keyboard and monitor and attach them to a PC that is known to work, and test them out. This way if you have problems when you build your new system you will know it isn't due to these components. At my office, I keep an extra keyboard, mouse and monitor just for debugging and assembling PCs. Since I know that they work, I reduce the number of variables that I have to deal with when I experience a problem. Once the main PC box is assembled and working, I connect the new keyboard, mouse and monitor. If a problem shows up then, I know that it is not likely the PC, but the peripherals I just connected.
Write Things Down
Whenever you discern a piece of evidence, or you figure out something about the current situation your PC is in, document it. Keeping a history of what you discover and what your symptoms are will help you not only with this problem but also with any future difficulties. A log book of some sort is a good idea.
Determine Repeatability
Most problems with a PC fall into one of two categories: either they are repeatable or they are intermittent. A repeatable problem is one where the problem occurs all the time, or always in response to a specific user action. For example, a PC that has a problem that prevents it from booting will probably always fail to boot no matter how many times you reset it. Or you may have an application that whenever you try to run, will crash with an error. You may find that your PC hangs, but only when you move the mouse at the same time that you are communicating using your modem.
In contrast, some problems are intermittent and not repeatable. In some cases, you may have a PC that will usually boot up fine, but one day a month will fail to boot for some reason. An application may work most of the time but occasionally crash. The PC may lock up at seemingly random intervals. Your mouse may work almost all of the time, but one day out of five or ten may give you trouble.
It is helpful to determine if the problem you are experiencing is repeatable, because intermittent difficulties are much more difficult to resolve than repeatable ones. If a problem is repeatable, and there is a specific set of actions that cause the problem, this gives you at least some initial clues about how to find the cause. In addition, you have a way of testing to see if you have resolved the problem when you are trying different solutions. Intermittent problems are much more difficult to deal with.
Determining if a problem is repeatable is pretty simple: try to duplicate the conditions that caused the problem and see if it happens again.
Dealing With Intermittent Problems
Intermittent problems are ones which appear to happen spontaneously or randomly. They seem not to be caused by anything obvious and are not repeatable. They can be extremely difficult and frustrating to diagnose.
Sometimes problems that seem intermittent really aren't; it's just that the specific set of circumstances that causes the problem to occur may be obscure or hard to notice. Spend some time trying to determine what the circumstances are when the problem arises. For example, many problems with crashes or lockups will occur only after the system has been on for more than an hour; some may occur only within the first few minutes that the PC is turned on. You may find that program A only crashes when run at the same time as program B. Or that a particular behavior is associated with a peripheral of some sort.
You will have to be patient when dealing with these sorts of issues. Since the problem is not something you can duplicate at will, you may not be able to systematically work your way towards the ultimate cause. In this situation you may have to employ trial-and-error, making a change and then waiting to see if the problem recurs. It can take days (or longer) sometimes since you have to wait before seeing if the problem happens again. Be patient.
Correlation May Not Imply Causation
I use the word correlation, perhaps slightly incorrectly, to refer to two behaviors or symptoms that appear at the same time. Causation refers to two events where one is responsible for the other's appearance or existence. So what I am saying here is that if you see two strange things happening at the same time on your PC, this does not necessarily mean that one of them has caused the other. They could be coincidental, or it could be that where you think A is causing B, B is really causing A.
For example, you may find two symptoms on your machine: you have frequent file system corruption on your hard disk, and you have a lot of system lockups (hangs, crashes). You may think that the crashes are causing the file system errors, and in many cases this is true. However it is possible for the file system errors to cause the lockups. And it is also possible that both are just symptoms of another underlying cause.
Scan For File System Problems
If the problem you are looking at is anything that involves the hard disk, files, operating system problems, crashes or similar behavior, scan the file system for errors. This is a simple step (which you should be performing regularly as part of a preventive maintenance program) that may tell you immediately that there is a problem with your hard disk. This can also be a useful piece of evidence in troubleshooting some sorts of problems.
Use Diagnostic Tools
If you own diagnostic hardware and/or software tools, they can be useful in pinpointing some kinds of problems. If you feel that you are having a problem that can be detected by a tool, or you feel that running a diagnostic utility may shed more light on what's happening, then make use of the tool to help guide your analysis of the problem.
Next: Diagnostic, Troubleshooting and Repair Tools
Diagnostic, Troubleshooting and Repair Tools
Troubleshooting a problem with a PC can be very difficult and frustrating. Fortunately, there are places to turn for assistance, and there are also tools at your disposal that can make the job much easier. These can be used to identify specific problems within the PC, or just give you more information about the system so you understand better what you are dealing with. This chapter takes a brief look at these tools, which can make your life easier if you are having trouble.
Diagnostic Software
The use of diagnostic software tools can save you a great deal of time. While in my experience these tools do not usually identify what the cause of a system problem is, they often provide hints or at least valuable information about what is in the system and how it is working. Some of these are free or are included with common operating systems at no additional charge, while others are commercial products that range from affordable to rather pricey.
Here are some common software diagnostic tools that you will want to consider including as part of your troubleshooting arsenal:
- Power-On Self Test (POST): I am cheating a bit here because this isn't a separate diagnostic utility; it is in fact built into your system BIOS and it runs every time you start up your PC automatically. It is often the best indicator of system problems; make sure you pay attention to its audio and video messages. Don't disable its error-reporting functions unless you really need to.
- MEM.EXE: This simple utility, built into recent versions of DOS and also Windows 95, provides you with details about your memory configuration, as well as what is currently using your memory. It is especially useful when run with the "/C" parameter (use the "/P" parameter as well to make the output pause when it is scrolling).
- Microsoft Diagnostics: Better known as "MSD.EXE", this is a small DOS utility that takes a brief inventory of the contents of your PC and shows them to you in a text-based format. This is very useful for seeing what disks are in the system, how much memory is installed, and also for checking system resource usage such as LPT ports and IRQs. It will show you what type of BIOS you are using and also what UART chip you have in your serial ports. MSD.EXE is included in later versions of DOS.
Tip: MSD.EXE will work on a Windows 95 machine but is not normally installed as part of the normal Windows 95 installation. You can find it on the Windows 95 CD however and then copy it to the C:\WINDOWS\COMMAND directory to make it work.
Warning: MSD.EXE can give incorrect information when run from within Windows. Shut down to DOS mode before running it ("Restart the computer in MS-DOS mode")
- The Windows 95 Device Manager: This is the most useful tool for identifying system configuration and resource usage information under Windows 95. To access it, open the Control Panel and select the "System" icon. Then select the "Device Manager" tab. You will see a graphical "tree" structure showing you all of your PC hardware. If you select "Properties" while "Computer" (the top-level item) is selected, you will be able to see all the IRQs, DMA channels and I/O addresses in use in your PC; very useful for resolving resource conflicts! The same "Properties" button, pressed after selecting a specific hardware device, will show you driver information, resource settings for the hardware item chosen, and much more.
- Norton System Information: This utility is similar to the Microsoft Diagnostics, only more detailed in its later versions. SI shows a great deal of information about what is in the PC, going well beyond what MSD gives you, but really is still an information utility as opposed to a true diagnostic. This program is part of Symantec's Norton Utilities.
- Microsoft ScanDisk and Norton Disk Doctor: These programs are used to check for hard disk problems. This includes file system corruption and hard disk read errors. They should be used when hard disk problems are suspected.
- Norton Diagnostics: This utility is meant to go beyond the System Information program and actually perform tests on the hardware to identify problems. It includes tests of the processor and motherboard and system memory, and will identify some types of resource conflicts. In reality it is still quite limited in terms of the numbers of problems it will find.
- QAPlus: QAPlus from DiagSoft is a more advanced diagnostic suite that comes in several flavors, depending on what you need to do and how you want to do it. This is a more expensive package but can give you much more detailed information about your system and help identify problem situations as well.
Diagnostic Hardware
While not as commonly used as diagnostic software, there are available several pieces of hardware that can be very helpful in troubleshooting some specific hardware problems. These units are not nearly as popular because they are usually more expensive and are used for specific types of problems, making them less universally applicable than the general software utilities that are widely encountered.
If you are a serious home builder or troubleshooter, you may want to consider one or more of the following:
- Loop-Back Plugs: These are small plugs that go onto the serial and parallel ports on your computer and connect the receive to transmit lines to simulate a connection (they connect the port to itself). Using these with a program like Norton Diagnostics allows for full testing of the operation of these ports, whereas without them only the internal portions of the port can be tested.
Note: The type of plugs used by a particular diagnostic utility may be different than the type used by another program made by another company. Check it out before you buy.
- Multimeter/Ohmmeter: An ohmmeter is a device that measures electrical resistance; a multimeter can measure resistance, voltage and current. These devices are used by electricians, electronics designers and repairpeople. They have come down greatly in price and simple ones can be had inexpensively (while top-end ones of course are still quite a bit of money). A simple ohmmeter is useful primarily for checking for short circuits or open circuits (broken connections, damaged cables, etc.) Multimeters can be used for more extensive electronics testing.
- BIOS POST Cards: It is well-known that the BIOS will, if it finds a problem during its power-on self-test (POST), produce audio "beep codes" and/or video messages that indicate what the trouble is that it found. What many do not realize is that the BIOS of most PCs is designed to send a stream of test codes to a special memory location as it performs its tests, usually 80h. By using a special card designed to capture and display these codes, you can pinpoint exactly where in the power-on self-test a system is hanging up. This can be extremely helpful in debugging very stubborn systems. Scott Wainner has more about one of these cards on his site.
- Test Bed: Some experienced PC repair people keep around an older system that they can use as a test bed for components. It can be very useful to be able to test an unknown device with other components that are known to work, to cut down on the guesswork.
There is also more advanced test hardware available, including devices such as logic probes, oscilloscopes, and many types of specialized component testers. These items are generally very expensive and require specific training to be used properly; they are not for the home PC user or even for most hobbyists. Professional test hardware costs thousands of dollars, which is one reason why doing it yourself is often not an option for many kinds of repairs.
The PC Tool Kit
If you are a very casual PC user you can probably get by hacking around with an old screwdriver and a pocket knife when you need to work on your PC. However, if you plan to do any reasonable amount of work upgrading, building, troubleshooting or repairing PCs, a proper tool kit is a good investment. There are many different types of tool kits available for the computer enthusiast, with tools selected specifically to make the lives of those who work with PCs easier. The right tools can save you a lot of time and help you avoid damage to your equipment (and your hands!)
As with most things, tool kits range widely in size, quality and price. At the low end are $5 kits that are barely any better than the screwdriver and pocket knife. At the high end are kits costing over $100 that include everything but the kitchen sink; these are overkill for most people (including me). Most people can do just fine with a toolkit costing between $15 and $50. These are often available pre-assembled into "genuine artificial leather" carrying cases, sold at most general computer stores.
Here are the items that I would recommend you definitely have in your toolkit; note that some of these don't generally come in pre-assembled toolkits so you will want to add them yourself to the carrying case. Every item in this list is one that I use personally on a regular basis:
- A Good Screwdriver: You want a good screwdriver with magnetic, interchangeable bits and a comfortable handle. At a minimum you should have two different sizes each of Phillips and flat bits. If you've never used a magnetic screwdriver before, you'll wonder how you ever did without one once you do; the magnetic tip will prevent screws falling into the PC, and will let you get access to drives mounted deep in your system case.
- Additional Screwdrivers: An additional Philips screwdriver and an additional flat blade screwdriver, for when you misplace the good one. :^) Actually, it's sometimes useful to have two screwdrivers anyway.
- Needle-Nose Pliers: These are useful for grasping small items and for removing and replacing jumpers on circuit boards.
- Wire Snips: A pair of wire snips for cutting wire and stripping insulation. Some kits include wire strippers for the latter purpose.
- A Small Flashlight: Very useful; the insides of PC boxes are quite dark and there are lots of very small things you will need to see, such as the "pin 1" marking on a connector for example.
- Tweezers: Or even better, a part retriever. A retriever is like a tiny set of retractable claws with a spring-loaded handle. Useful stuff.
- A Large Crowbar: For making adjustments to stubborn hardware. (Just kidding. :^) )
- A Roll of Black Electrical Tape: They never include this in the kits but they should. Used for wrapping wire ends and insulating components.
- A Can of Compressed Air: Very handy for cleaning things hands-free and without using any hazardous liquids.
- A Soft, Lint-Free Cloth: For cleaning the monitor and other components.
The following items I categorize as "optional, but useful". They are tools that are used less frequently, or perhaps only in certain circumstances. They are more for someone who does a fair bit of tinkering or who maintains a number of machines:
- A 3/16" Nut Driver: In fact, tool kits often come with several different sizes of nut drivers (using exchangeable bits), but 3/16" is the really important one, because this is the size of nut most commonly used on PCs. These hexagonal nuts are used as mounting hardware for motherboards and serial and parallel ports. A necessity for anyone who builds their own PCs; if you don't have one you'll be attaching port connectors using a pair of poor-fitting pliers.
- Vice Grips: These are incredibly neat tools that have all sorts of uses.
- Torx Screwdrivers or Bits: These are the star-shaped screwdriver heads that are used to make equipment "tamper-proof". You need these drivers if you want to, well, "tamper".
- An ESD (Electrostatic Discharge) Wrist Strap: This is more of a safety device than a tool; it is used to greatly reduce the chances of static damage to components.
- A Knife: A cutting blade or utility knife of some sort.
The following items are sometimes included in tool kits but in my experience are rarely if ever used:
- Soldering Irons: Essential if you are going to try to do repair of circuit boards, but virtually nobody does this (and with good reason). They are also quite dangerous if used improperly.
- Hex (Allen) Keys: They include these in virtually every commercial tool kit I've ever encountered. I've never seen anything PC related that used them but of course that doesn't mean that there isn't something out there that does...
- Chip Extractors: These are used to remove chips from regular (non-ZIF) sockets, but modern PCs do not generally use these sockets any more. Useful for upgrading much older systems in some cases.
Finally, it is a good idea to accumulate a stockpile of spare parts if you are going to be doing work on PCs. I would recommend the following:
- Screws: I have a "big ol' bag o' screws" of all shapes and sizes that comes in handy when working on various machines.
- Expansion Card Inserts: Save the metal inserts that come out of the back of the case when you put a modem or other card into the PC, as you may need them again later on.
- Drive Faceplates: Similarly, save the plastic faceplates that you remove from the front of cases so you can replace them later if needed.
- Mounting Kits: These sometimes come with retail hard disks and allow you to put a 3.5" drive into a 5.25" bay. Useful when your case has more free 5.25" bays than 3.5" ones.
- Cables: Save any power, IDE, floppy, CD-ROM, or other cables that you accumulate in case you need them later on.
- Keyboard, Mouse, 3.5" Floppy Drive: Keep an extra one of each of these components around to aid in troubleshooting problems by swapping. (Actually, keeping an extra of every major component in the PC is helpful, but the three I mention are here because they are cheap. Most people can't afford to keep an extra hard disk lying around as a spare.)
Get to the Root Cause: Ask "Why" Five Times
One of the keys to truly solving a problem is to first truly understand it. It is easy to observe that too often, people are quick to jump to solve a problem before they really understand what it is. It is easy to see a symptom and think that you know what is causing it, but surprisingly often, if you take the time to explore more deeply you will find that what you thought was the cause is in fact just another symptom, and that the problem lies much deeper within. One technique that you can employ is borrowed from Japanese manufacturing management theory. This method is used to help identify the real causes of problems that occur on the manufacturing floor. The goal is not to simply correct the effects of the problem, but to find out the root cause of why the problem is occurring so that we can ensure that it will not happen in the future. One simple way to do this is called asking why' five times. The idea is that by the time you have asked "why" the fifth time, you will be at the root cause. It isn't always that simple, but the exercise can be surprisingly insightful in helping you figure out what is really going on, and can help you avoid "quick fix" solutions that are really just band-aids and don't resolve anything. It is especially useful for tackling chronic problems that show up over and over again in a system; it is less useful for problems that are unlikely to recur. Here's an example. Let's suppose your hard disk is having a problem with bad sectors showing up. The knee-jerk reaction to this happening is "the hard disk is bad, replace it". Instead, ask yourself:
- "Why are bad sectors showing up when they didn't before?"
- "The hard disk is being corrupted somehow; something has changed."
- "Why?"
- "The timing of the system probably has been changed in some way."
- "Why?"
- "Hmm, I installed a new hard drive in that system just last week, perhaps it is causing the problem."
- "Why?"
- "The hard drive may not have been installed correctly, or perhaps I should have used a different channel. Or maybe it was the bus-mastering driver that came with the new drive, the old drive may not be able to support it".
You can see the general idea; the answers will differ in every case, but it is the process itself that is useful. Another example: imagine an office that has just had a catastrophic data loss due to a PC crashing after an electrical storm. The initially identified cause of this problem was a lack of a UPS on the PC, which would have protected the system from the electrical storm (usually). So the MIS department starts drafting purchase requisitions for UPSes for all the PCs. Instead, if they looked at the problem carefully, they may have employed the following process:
- "Remember that the problem here was the data loss, not the crashing of the PC. Why was the data lost?"
- "The PC's hard disk crashed."
- "Why would that cause the data to be lost?"
- "Because we didn't have a backup."
- "Why wasn't there a backup?"
- "There is a tape backup unit on every PC in the building but very few of them are being used."
- "Why?"
- "Because most of the PC users do not know how to use them."
- "Why?"
- "They have never been given any training in their use."
- "Why?"
- "Because we don't have a budget for PC hardware training."