MICRO CENTER: COMPUTERS AND ELECTRONICS
Random Access   chris, kp & rob
In The Lab
PC Troubleshooting Part 2: A Real-world Example
by chris

Last month, I defined a five-step approach to Troubleshooting. The process steps I identified for our readers were:
1. Define the problem.
2. Identify the possible cause.
3. Research the problem.
4. Select and Test Solutions.
5. Prevention.

One of the most popular clinics we have held in our stores has been on Building-Your-Own PC. We also find that a majority of the questions being asked in our Support departments have to do with this subject either directly, by people trying to assemble a system for the first time, or indirectly, by people asking about upgrades, product compatibility, and troubleshooting installation issues. Let's take an example of building a new PC from scratch. Say that you collect a new case, motherboard, CPU, memory, video card, hard disk and optical drives, maybe a TV tuner card, lighting and fan kits (sounds like a typical quick weekend project, right?) You put everything together and press the power button... and nothing happens. Time to start troubleshooting your new toy, but you are already at a disadvantage because it never worked; this is not the same as an upgrade, where you made one or two changes that can be reversed in the process to troubleshoot.

Let's apply the troubleshooting steps to see what the process might look like. First, define the problem ("that's easy, it doesn't work...") While that is certainly descriptive, it does seem to be a bit vague. Start asking yourself questions that more specifically describe the problem. Does the system power on at all? Is there any activity from the fans, LEDs, etc.? If the answer is yes, then we do have power, but still may get a blank screen and no apparent activity. Try it again, listening carefully; is there a system beep or other indicator of a start-up problem? Does the optical or hard drive LEDs come on at first, and then go off after a short period, or do they come on and stay on? All of these questions help define the problem and gives us the clues for the next step.

To "identify the possible cause" on a new system, you need to pay attention and answer all of the subtle little questions I brought up in the example. If there is no LED or fan activity, the problem could be with the power supply, or caused by a short from something connected to it. First check the obvious things like checking that the power cable is firmly pushed into the wall socket, power strip, and rear power connector. If all of these check out, are the switches turned on? Do other devices on the power strip come on? Is there a rear power switch on the power supply and is it in the "on" position? Is there a voltage selector switch and is it set correctly to your 110/220 Volt source?

If everything checks out so far, then you need to start working your way into the case systematically. Check that all power connections are secure to the motherboard and that you didn't forget any. Some motherboards may have two or more power connections, and if any one of them is missing, the system is not going to start. Besides the 20-pin (or new 24-pin) ATX connector, you probably have a second 4-pin ATX 12 volt connection. RAMBUS motherboards had a special 6-pin connection, and some of the new PCI Express motherboards or video adapters may need a connection as well.

With hardware troubleshooting, the process to identify the possible cause sometimes tends to merge with researching the problem; the bottom line is that we need to try to isolate the problem. One way to do this is to remove anything that is not required to test the system. Remove all expansion cards and disconnect the floppy, hard disk and optical drives from the motherboard. Take the system down to the very minimum parts of motherboard, CPU and memory. If your system board has onboard video, connect it to the monitor, if not, you can try it with the video card installed, but keep in mind that this is one more item that could be causing the problem. If the video card is not installed firmly into its slot, it could be shorting out the system. If you have a brand new motherboard that only supports 8x AGP video and you have an old 2x/4x video adapter, the adapter may not work with the motherboard. If you have onboard video and a video card in the expansion slot, where is the video going to come out? Look at it this way, if there is no video on the motherboard, and you do not have a video card installed, the system should beep to indicate a problem when you apply power.

BIOS Beep Codes
Most systems will generate a series of "beep codes" during the start-up process. These are not always consistent between manufacturers and different BIOS versions, but there are generally a few common signals. You should find information on the beep codes following the configuration section in your motherboard documentation. Beep codes for the Award / Phoenix BIOS used in many of the current systems translate something like this:

Beep Meaning
One short beep when displaying logo No errors during POST. This is normal
Long beeps in an endless loop No DRAM installed, DRAM not detected, or base 64MB memory bad
One long beep followed by three short beeps Video card not found or video memory bad. (Note: Some video cards can generate this code if the monitor is not attached and/or turned on.)
High frequency beep when system is working CPU overheated. System running at a lower frequency.

Let's say you have your system stripped down to just the power supply, memory and CPU, but you still get no activity (fans or LEDs) when you hit the power button. How can you possibly isolate things even further? If you have extra parts, you can always try swapping things around; test a different power supply, memory and so forth. But if this is your new dream system, you may not have any spare parts to test with. Before taking the next step, you might want to do some research and double check that the CPU is compatible with your motherboard. For example, installing an 800MHz FSB processor in a motherboard designed for 400/533 MHz CPUs may not start and could damage either the CPU or motherboard if the voltage levels are wrong. Double check the case wires to make sure that they are attached correctly to the front panel header on the motherboard. You can also try disconnecting all of the front connector wires except for the power switch. If LED connections are backwards, they generally will not light up. If you have a bad or shorted reset switch, the system will never leave its Power On Self Test (POST) stage. If you still get no activity, try removing the motherboard from the chassis and set it on a piece of cardboard or other insulated surface. Do not try this on metal surfaces or on plastic surfaces that could generate a static charge that could damage the electronics. If you have isolated things to this point and still get no activity, something is probably dead or incompatible, and you will either have to start swapping things, or take your pile of parts to someone who can. On the other hand, if your motherboard, memory, and CPU do power on and give one or more beeps, then you have succeeded in isolating the problem to the case assembly.

Researching this type of problem may amount to nothing more than verifying that all metal stand-offs match up with holes in the motherboard (you did use stand-offs, right?) You also need to look for any other projections or extra posts that could be making contact with the motherboard and shorting out the power. If you find posts in the wrong position or extra posts installed, add or remove as needed until you have support posts only where they are supposed to be; then, install the motherboard back into the case. One alternative may be to replace metal stand-offs with insulated nylon ones that can't short out. If there are raised projections in the case itself, you may need to place some sort of insulated pad or spacer between the sheet metal and the bottom of the motherboard.

In this example, something about the installation or case construction was shorting out our system. Once we have isolated the cause, and taken some steps to correct it, we are ready to test our solution (and hope that nothing got fried when it shorted out.) Once you have placed the motherboard back in the case, make only the same minimal connections you had outside and test that your system still starts up. To minimize introducing new problems, install one item at a time and double check that the system still powers on at each step. For example, attach the rest of the front panel connections and try to power on. If the LEDs light up and the system starts, then whatever you found was probably the source of the problem. Be sure to cut power completely before installing any expansion cards or memory; remember power is flowing from the power supply to the motherboard whenever the supply is connected to the outlet and turned on. Next, add your video card and check that you can now get a video display during the POST process. Keep adding items, one at a time, testing that the system still starts at each step. The only reason for this is to prevent the introduction of new problems during the build. When installing your operating system from CD, avoid problems by running the setup with only the basic components assembled. Once the OS has been copied to the hard drive and identified your base hardware, then you can add your remaining components, one at a time, detecting and/or installing the necessary drivers as you go.

If the system comes on with fans running and LEDs illuminated, but no other start-up activity or video, then you might need to research the motherboard more. Some motherboards may have jumpers to set CPU or bus speed, and most have a "Clear CMOS" jumper that must be in a specific position for normal operation. Some Intel motherboards have a position that forces a maintenance or configuration mode in the Setup. This feature protects a CPU from operating outside of its specified speed or voltage range. In some cases, this setting may need to be used when installing a new processor the first time.

You can see from this scenario that troubleshooting steps are not always precisely defined between some of the stages. Depending on your situation, you might encounter different issues that would create similar symptoms. For example, if your motherboard can support the CPU you have, but it has an old version of the BIOS, you may need to flash update the system board with the latest BIOS image before it will even start with that CPU installed. To flash the BIOS you may need to create a special BIOS recovery disk (i.e. some Intel boards support this) or you might have to install a slower processor to flash the system, then replace the processor with your faster one. If this is an AMD system, check for jumpers that set the CPU clock speed. If these are set wrong, your system my not start or could damage the processor when power is applied.

Here are some additional warnings when building a system from pieces: make sure you have a correct style heat sink for your CPU, use thermal compound or remove the protective film from a pre-applied thermal pad on the heat sink, and have both installed correctly and securely before turning on power. A mistake at this very basic assembly stage may result in a dead CPU and/or a dead motherboard.

One last method of researching your problem and selecting solutions may be to talk with one of our Service or Technical Support Associates. Their product knowledge and experience may help narrow down possible problems or at least suggest some additional things to try. The more details you can provide, the better their suggestions will be.

Get Random Access

Understanding Tech

Print this article

Shop Online

Send-to-a-
Friend

Your Name:

Your Email:

Your Friend's Name:

Your Friend's Email:


 © Micro Center