PC Boot Process

To troubleshoot a broken system, it helps to understand how it is supposed to work in the first place. In this section, we will look at what is involved in getting the computer from the time that you turn it on, to the point that it is ready to start doing work. There are two phases to this process. The first phase is carried out by the computer itself, and the second phase is carried out by the Operating System software that is installed.

When you press the power button on a modern computer, a signal is sent to the power supply to turn on. Keep in mind that the power may not have actually been completely off up to this point, but rather in a mostly off mode. Do not confuse this with the stand-by, hibernate, or sleep, modes offered on some computers. This low-power mode is more like the stand-by mode found on most common home electronics.

When the power supply is activated, the very first thing the computer does is test for a 'Power Good' signal. This is supposed to indicate that the power supply is stable, and ready for the computer to start up. Once the Power Good signal is present, the computer begins a very basic set of tests to make sure that what it considers to be key components are present and working. This set of tests is called POST (Power On Self Test). What components are considered “key” is determined by the BIOS (Basic Input Output System) software, which is on a chip on the Main System Board (sometimes called the Motherboard).

You can examine what is controlled by BIOS, and even make adjustments by running the BIOS setup software. On most computers, you access this software by hitting a specified key during the POST process. The key may be identified on screen. Otherwise, you may need to check your systems documentation. If you get a screen with the computer manufacturer logo, sometimes hitting the 'ESC' key will allow you to see the BIOS information that is hidden behind it. If nothing else, you can try the following common keys during POST to see if it takes you into BIOS setup. The Common keys are F2, DEL, F10, F12, ESC.

The components are tested by POST in order of importance. For example, if the processor does not check out, there is no point in continuing since the process is needed to preform the rest of the tests. The same thing is true for the system memory (RAM), and therefore it is tested second. Since it is difficult for the system to communicate problems to you without Video, the video system is generally checked next. If there is a failure before video passes its tests, the computer obviously cannot use the display to tell you what is wrong. Therefore, it will generally use the built-in speaker, and a series of beeps. You will need to find documentation for your specific BIOS to know what the beeps mean. Some modern systems will actually use a voice to try to tell you the problem, but beeps are still more likely right now.

Once the video system passes its tests, the monitor is brought up and you can see the results of the remaining tests. This may include testing (or just counting) additional memory, and taking an inventory of other devices. The system may look to be sure a keyboard and mouse are connected. It will look to see what types of Disk Drives are supposed to be present, and then check for them. If there are any problems, an error will be displayed on the screen.

Once BIOS has completed testing all of the devices it knows about (generally only the built in ones), it will then look for add-in cards. BIOS will inventory the add-in cards, and initialize each one. As each card is initialized, it may have it's own BIOS, with it's own Setup program, and it's own POST. You can sometimes tell this is happening if you see messages that offer different key combinations for Setup, in addition to the first such offer.

As each add-in card finishes initialization, BIOS moves on to the next. After the last card is initialized, BIOS begins looking for an OS (Operating System) to load. It may search different drives until it finds an OS. The list of places it will look, as well as the order in which it searches, is controlled from the BIOS setup program.

As soon as BIOS finds an OS, it loads it into memory, and turns control of the computer over to the OS. We will talk about what the OS does a little later.

POST Failures

Any errors that occur up to this point are likely hardware based. All of the above tests were performed before the Software has been loaded and run. This may be a bit confusing, since I said that BIOS, and POST are both software. This is where a new term Firmware, comes in. BIOS and POST are software that is stored as part of the Main System board. A failure in that software might be viewed as a failure of the Main Board. This software does not change with regular usage. It is possible to update this software, but generally, this software is frozen. When we talk about Software, such as the OS software, we are talking about programs that are stored on the Hard Drive or some other disk. What we commonly refer to as software is prone to change all the time. Every time you start and shut down Windows, the Windows software has changed. It never starts the same twice.

If POST completes successfully, this does not mean that all of the hardware is fine. This just means that there are no glaring errors in any of the devices that POST checks.

If POST indicates that a device (such as memory) is bad, it may just mean that the computer cannot reliably communicate with that device. The first thing you should do if a device is identified as bad is shut down the computer, remove the device, inspect it, maybe clean it, and re-install it. Frequently this will make the problem go away.

Basic Maintenance can often fix a problem of this type. If you have not performed maintenance on this system recently, give that a try first.

If after these basic steps, you still get the same error, odds are you will not be able to do any testing. Generally, if there is a POST error, the computer will not proceed past it.

If POST complains about memory, and you have more than one memory module installed, try removing extra memory so that you have minimum amount installed. In most computers this will be one module. Some computers may require two modules at a time, your documentation will tell you what is needed. If your computer still complains about memory, turn it off and swap the installed module for one of the ones you removed. Unless the problem is in the System Board, it is not likely that multiple memory modules failed at the same time.

If POST reports a problem with some other part, and you happen to have an extra one available, try swapping it out. For example, if POST complains about the hard drive, and you happen to have a spare hard drive lying around (old computer, maybe you upgraded an old drive earlier, etc..), try installing the spare part to see if POST will pass.

Keep in mind that the above tactics (removing memory, swapping parts) are not intended to fix the computer, only to help identify the problem. If, for example, placing an old hard drive in the computer causes it to get past POST, you do not necessarily want to let it start up fully with that hard drive. The OS on that hard drive might not be configured for this computer. However, you now at least know the problem is definitely the hard drive.

If you never get anything to display, and instead the computer just keeps beeping at you, you will need to look up the beep codes for your computer. Hopefully, this will be in the documentation for your computer. If it is not, you may need to get on the Internet from a different computer, and try to find a list of codes. Not all computers use the same codes, so make sure the list you find goes with you make and model of computer.

Power Supply Problems

If you do not get beeps, and you do not get a display, do you get any signs of life? Is the fan running? If there are no signs of life, make sure you have a working outlet, and that the cords are all connected. If there is still no sign of life, that can mean any of the following;

The power supply is dead
The power switch is not working
There is a short circuit somewhere in the system
There is another power switch somewhere.

The Power Supply used in computers can be damaged, or may refuse to power on at all, unless they are connected to something that will use the power that is generated. With this in mind, you cannot just disconnect it, turn it on, and see if it works. There are tools available that you can connect to the power supply and test it. These are generally around $20, and can be ordered on the Internet. I do not know if any local stores sell them. I do not include this in the common list of tools since you will rarely need this. If you check everything else, and you still think the power supply is bad, you can either order a tester, or just buy a new supply and see if it fixes the problem

Since the power button on most modern computers is not an actual switch, it is often just connected to the main board. It is possible the connection has come loose. Take a look at the wires as they come away from the power button and see where they are supposed to be connected. Make sure the connection is good. You might even remove and reconnect this cable. WARNING! If the computer is plugged in, and you re-connect this cable, the computer could turn on unexpectedly. Make sure you are not touching anything dangerous, and that you are not obstructing any fans when doing this.

Examine the system for any extra switches. Do not limit this search to just the computer itself.

Is the outlet we are plugged into on a light switch?
Is the computer plugged into a power strip with a switch? a fuse or circuit breaker?
Have you tried plugging into a different outlet? Directly into an outlet without a power strip?
Is there a switch on the power cord?

If there is a short circuit in the system, it may prevent the system from powering on. To test for this, first examine the inside of the computer for anything that is obviously out of place. If there is nothing obvious, then we need to move on to testing with the most basic configuration.

Basic Configuration

There are two ways to approach the Basic Configuration tests.

One view suggests that you should remove everything, and try to get a predictable error condition. This would mean removing all cables from the Main Board, except for power, and a power button cable, if it exists. You would remove all add-in cards, memory, any other upgrades. Ideally, you would only have the Processor, Power, and a system speaker, connected. You would set the Main board on a NON-conductive surface, and try to power on the system. The idea is that you should get beep codes indicating that memory failed POST. This would tell you that at least the power supply, and the main board, were both working to some degree. At this point you would just install the memory, and try again. On older computers, before video was built in, you would now get beep codes indicating video failed. Most modern computers have video built into the main board so this step is skipped. Once you had beep codes for video, you would re-install the video board and try again. If all was well, you would now at least get something on the display. If you did not, then either the video board, or the monitor itself, is the cause of the problem.

The alternate view goes the opposite direction. Instead of removing everything and building back up, You can remove items, one at a time, and see if the system begins working. The first thing you would remove would be any add-in cards, other than video. Second, you would disconnect all drive cables from the main board. You would continue removing items until the system responded in a way that made sense (For example, if it would not power on at all, and you disconnect the drives, a sensible response from the computer would be to power on, but complain that it cannot find the drives.) You may keep removing things, right down to the memory, which should result in beep codes.

If you use the first method, then the bad part is the last one you added when the system stopped working. If you use method two, then the bad part is probably the last one you removed. The key thing here is the last change is likely the problem.

Method two probably makes more sense now since more and more things are built into the main board. In the way-past, nothing was built into the main board so you could really start from scratch with troubleshooting. Now, it is not uncommon for things such as video, audio, I/O Ports (Input/Output, like USB and printer ports), and networking, to be built into the main board. You may choose whichever method makes the most sense to you.