Troubleshooting Techniques

When you diagnose a problem, or simply add a new component to your computer, is it essential to make sure you have all the required information before you start -- always keep all the documentation associated with your computer and its parts.  It is also important to take your time and to keep detailed notes when you are working.  Don't assume that you'll remember which cable went where, I either put masking tape on a cable and write down on it the "what and where" or write on the cable itself.  If your computer is giving you error messages, write them down word for word.  Always start with the easiest, most obvious possibilities (power, cables, etc.) and move to the more difficult.  Be methodical, and make any changes one at a time.

Preventive Maintenance

Keep your computer clean and dust free.  Be careful if eating or drinking when using your computer - water and electronic devices do not mix well.  Try to keep your computer in a relatively dust-free and temperature controlled environment.  Buy a can of compressed air at your local computer store (it's well worth the money, but follow the directions for use carefully) and blow out the fans and vents regularly.  If you keep your computer in a dusty place, open the cover and gently spray compressed air over the components inside as well, every so often.  At the same time make sure all your cables are not frayed or crimped, and that their connectors are snug.

Keep your computer and its peripherals on a surge protection strip, if possible.  Surge protectors are inexpensive devices that can absorb certain types of power fluctuations.  If you don't have a surge protector, unplug your computer from the wall when it will go unused for an extended period, and never leave it plugged in during an electrical storm.

=========================

Computer Start-Up

The computer start-up process can be divided into three basic steps: Power On Self Test, BIOS, and Boot (or Bootstrap).

Troubleshooting system start-up problems can involve looking at one or more of these steps.

Although many things can go wrong with a computer, perhaps the most frustrating is when your computer won't start at all.  This is true, in part, because when your computer fails to start you can not use diagnostic tools to help you determine the problem.

If your computer won't start at all, the power light does not come on and the cooling fan is not running:

Check if the wall outlet you are using will it run another appliance such as a small lamp. Change the wall outlet you are using -- perhaps there is inadequate voltage.

Check to make sure that the computer's electrical cord is securely plugged in at both ends.  Switch it with another cord if you have one available to you.

If none of these steps help, it is likely your power supply has gone bad. Power supplies are potentially dangerous units and a professional should replace your power supply if you need a new one.  Never open up a power supply box.

Having a bad power supply can be the cause of many problems that may not initially point to that component.  For example, spontaneous reboots or freezes and even memory parity errors can be traced to faulty or inadequate power supplies.

It is almost always easier, and often cheaper, to replace a defective power supply rather than repair it.  If you do have to replace your power supply, be mindful of the physical requirements (shape, location, screw-hole positions and so on).

If there is no power light, but your cooling fan is running:

Open up your computer and make sure that your 6-pin power supply cables (there are two) are securely attached to the motherboard (note that such 6-pin cables are unique to your system's power supply harness).

If none of these steps help, you likely need a new power supply.

If your power light is on and your cooling fan is running, but there is no computer activity (you hear no beeps at start-up):

Open your computer and make sure your power supply cables are securely attached to the motherboard.

Examine your motherboard for metal pieces that might be touching other metal pieces (such as screws or the motherboard's seatings).  These could be shorting out your motherboard.

Check the seating on all the other expansion boards in your computer.  An expansion board that is not seated securely can cause a short.

Make sure all other cables inside your computer are securely seated and in good condition. Look for frayed or crimped edges.

Check that your CPU is inserted properly in its socket, that it is cool, that the CPU heat-sink/fan works and that it is fitted correctly.

If your CPU has a tension lever, make sure it is closed and locked.

If you hear two or more beeps at start-up, but there is no video:

Open your computer and make sure your video board is seated properly it its expansion slot.

Power On Self Test

The Power On Self Test (POST) is the first set of instructions executed during the start-up of your computer.  You can recognize it during the RAM test, which is one of the system components that the POST checks, along with the oher key components on the motherboard. The POST order of execution looks like the following:

BIOS

Information about the graphics adapter

Information about the BIOS (name, version)

Information about the RAM (counted)

If the POST detects an error or errors in the system, it will write the error messages on the screen.  However, if the monitor is not ready, or if there are video card errors, it provides an audible code.  If a catastrophic fault has been detected in the Power On Self Test that your computer does at start-up, the number of these audible beeps can be indicative of the type of failure.

These beeps (or beep code) can be useful in troubleshooting problems prior to the initialization of your computer's video system.  The number of beeps is dependent on your computer's particular BIOS and is unfortunately not standardized across manufacturers.  If you can match your beep code to your particular BIOS and version, you can accurately diagnose problems with your CPU, RAM, motherboard, video board, or drive controller.

Your system's beep codes should be in the written documentation for your motherboard, or obtainable from the manufacturer's web site.

The following is an example of beep codes from American Megatrends: 1s (short) System RAM refresh failure

2s Memory parity error

3s Base 64-kbytes memory failure

4s System timer failure

5s CPU failure

6s Keyboard controller Gate A20 failure

7s Virtual mode exception error

8s Display memory read/write error

9s ROM BIOS checksum error

10s CMOS shutdown register read/write error

11s Cache memory error

1l (long) - 3s Memory test failure (non fatal)

1l - 8s Display test failure (non fatal)

Your system documentation or the manufacturer's web site may include recommended solutions to these error codes.  For example, from American Megatrends:

For 1-, 2-, and 3-beep errors try reseating the system's memory modules.  If that fails, it is possible that 1 or more memory modules are bad.

For a 6-beep error, try substituting a different keyboard, or reseating the keyboard controller chip. For an 8-beep error, try reseating your video card or a different video card (if your system has a separate video card).

For 4-, 5-, 7-, 9-, and 10-beep errors, it is likely that you will need to have your motherboard repaired or replaced.

CMOS Setup

The CMOS RAM chip maintains a small amount of information that is essential for your computer to start-up properly.  This information includes data about your floppy drive and hard disks, your keyboard, your CPU, cache, boot sequence and much more.  These data have to be set up correctly in order for your computer to operate properly.  In modern systems much of this set-up is automatic, but occasionally you may need to manually configure your CMOS Setup.  In general, you should leave these settings alone unless you are familiar with what you are doing, or you have no other choice.

Keep track of any changes you make to your CMOS Setup settings.  Because your CMOS settings are dependent upon your computer's battery, it is useful to make a hard copy (write them down).  This way, in the event that your battery fails, you have a backup copy.

If your video works but you see a message indicating a setup problem, you will need to enter into your CMOS Setup in order to look for differences in your CMOS system parameters and your hardware configuration.

How you enter your Setup routine will be dependent on your BIOS manufacturer -- for example, American Megatrends' CMOS Setup is entered by pressing the delete key during the POST.  Your system documentation will tell you how your particular CMOS Setup can be entered.

Once you are in CMOS Setup:

Carefully check each entry and make sure that they are reflective of your actual hardware configuration -- in particular check that installed memory and drive parameters are correct.  Some parameters, such as ones for the hard disk are usually automatically set on modern system, however, you may need to enable the drive in CMOS Setup in order to have the system recognize it. Make sure your CMOS battery is good by examining your setup parameters, shutting down and turning off your computer for ten minutes, then going into Setup again.  If your CMOS is not retaining changes, you will need to replace the battery. Many modern systems will display a warning at start up that the CMOS battery is getting low.

If changes you make to your CMOS Setup are not saved after rebooting:

Check to make sure that you are exiting the CMOS routine properly.  Often the default exit will not include a "Save." Video Problems

If your computer appears to be starting normally but you see no video activity:

Make sure your monitor is plugged in and turned on.

Make sure your monitor is working by trying it on a known, good, working system.

Make sure your monitor's cable is securely seated in the video board slot.

Make sure your video card is securely seated in its expansion slot.

Try a known, good, monitor.  If it does not work, you may need to replace your video card.

===============================================

There are numerous types of drives used for storing computer data; these include floppy drives, hard drives, CD-ROM drives and so on.  These drives use differing types of technologies to store information, including magnetic and optical storage.  Modern hard disks consist of one or more magnetic disks contained in a box that is typically 3.5 inches in diameter.  Size and speed of the hard drive are decisive factors in a computer's performance.

A disk drive is made up of a disk that stores information, a motor that rotates the disk, a magnet, and an actuator with read/write heads attached to suspension arms that magnetically and electronically read and write information on the disk surface.

When the motor is powered up the disk rotates and causes the read/write heads to fly above the disk surface on a cushion of air.  They are two or three microns above the surface, and the disk can rotate up to 7200 RPM or more, which means that the heads can be travelling at speeds in excess of 64 miles per hour.  Under normal operation there is never any head to disk contact.

Hard Disk
Magnetic Disk
EIDE Cable
IDE Sockets
EIDE

Hard drives can be distinguished by their interface types.  Today's systems generally have either EIDE controllers or SCSI controllers.  EIDE (Enhanced Integrated Device Electronics) is the current standard for inexpensive, high performance hard disks.  It is the enhanced version of the old IDE standard.  EIDE controllers are integrated onto motherboards.  The acronym ATA (Advanced Technology Attachment) is synonymous with EIDE. Most modern BIOS software can autodetect an EIDE disk, allowing it to work immediately once it is installed.

The hard disk is connected to the motherboard by means of an EIDE signal cable.  Each motherboard has sockets for two EIDE cables (a primary channel and a secondary channel), and has connectors for two units (a master and a slave) on each channel.

The EIDE interface is designed for more than just hard disks.  It can be used for a variety of devices, including CD-ROM drives, CR-RW drives, DVD drives and other types of storage media.  A new PC will usually have two EIDE units connected, a hard disk on the primary master, and a CD-ROM drive on the secondary master.  A second hard drive must be placed on the primary EIDE channel's slave connector.  Additionally, you also must set jumpers on the disk drive itself to specify if it is to be a master or slave disk.  Jumper pins are used to detail the jumper setting that specifies how the drive is to be used.

A typical setup with four EIDE devices might look like this:

Primary master = Hard disk 1
Primary slave = CD-ROM
Secondary master = Hard disk 2
Secondary slave = LS120 diskette

As shown, on an EIDE host controller there are sockets for two EIDE cables.  For a system with a single hard disk, it should be assigned as primary on the master channel.  In general, to achieve optimal performance, if you have two hard disk drives, each should be assigned as masters on the primary and secondary channels.

This is because the slave and master sub-channels on a given channel can not multi-task; only one operation is performed at a time, as opposed to the primary and secondary channels, which are capable of multitasking.

EIDE Controller

Juggling four (or more!) EIDE units on a system can be a difficult task.

One possible solution is to expand your system's EIDE capabilities with a PCI-based EIDE controller.  These expansion cards work side-by-side with your existing EIDE controller and can be connected to additional hard disks, CD-ROMs, ZIP or other ATA-based drives.

SCSI

The SCSI (Small Computer System Interface) interface is typically found in high end systems and servers.  The main difference between SCSI and EIDE is that a single SCSI adapter can handle 7 or 15 devices, of varying types (internal or external hard disks, scanners, CD-ROMs, zip drives, etc.).  Another difference is that while most modern systems have their EIDE controller as part of the motherboard, SCSI controllers are often a separate expansion card.  SCSI drives are faster and generally more robust than EIDE drives.  The SCSI system holds its own computer power, thus freeing the CPU.

Hard Disk Upgrades

Hard disks have become progressively faster and cheaper, and with their increased capacity can hold ever-growing amounts of data.  Upgrading your hard disk is a common improvement you can make to your system, as is adding a second hard disk.  However, as drives get faster, they require improved cooling.  Many modern systems now come with multiple cooling fans to stop drives from overheating.  It is important to keep your system's cooling fans clean and dust-free, and to make sure they are running properly. Keep a new hard disk in its electrostatic discharge protection container until you are ready to install it.  Handle hard disks carefully, and never handle more than one at a time. Never stack them or place them on edge.  Avoid dropping a drive -- even a 1/4 inch drop can cause damage.

Before you can physically install a new hard disk into your computer you must make sure that it is configured properly.  For an IDE drive this means specifying (using jumpers) whether it is a master or a slave device.

There is no standard way to set jumper pins.  It varies across manufacturers, and even across different drives from a single manufacturer.  Consult your drive's documentation, or the manufacturer's web site for jumper setting configuration.  Many drives have the proper jumper settings printed directly on the drive itself.

For a SCSI drive, configuration entails setting the device's SCSI ID and the SCSI bus termination state.

Each device on the SCSI chain must have a unique SCSI ID.  On most systems the SCSI controller is set as the highest available ID (ID 7 on non-wide systems and ID 14 on wide and ultra-wide systems).  Typically the boot drive is set as 0.  Some SCSI controllers allow you to skip IDs (for example, you can have ID 0, ID 1 and ID 3), other controllers do not permit this.

If you are installing a new SCSI drive in a system that has other SCSI devices installed (for example, a scanner or a tape backup unit), terminate only the end devices on the SCSI chain.

Use all four screws to securely mount your drive.  Using less than four screws may be tempting, but it may cause the drive to vibrate excessively, and this can shorten its life.  Do not tighten the screws, however, as this can warp the drive's internal frame.

After physically installing the drive you will need to configure it.  First, you will need to run FDISK in order to partition it, thereby defining areas of the disk for an operating system to use as a volume.  FDISK writes a master partition boot sector on the first sector of the hard disk.  This partitioning also prepares the drive for formatting, and tells the ROM BIOS which of the partitions are bootable.  If you are unfamiliar with FDISK, we encourage you to read Using the FDISK Utility.

Second, you will need to format your disk.  There are two format procedures that are required prior to your using a new hard disk:

Low-level or physical formatting (LLF)

High-level or logical formatting (HLF)

When you format a floppy diskette, both types of formatting are done simultaneously, but on a hard disk they are separate procedures.  The LLF is performed by the manufacturer prior to sale, and for EIDE drives this is usually never done by the end user.  The low-level format writes the tracks and sectors on the disk.

The HLF is what is done by the DOS FORMAT command. If you are unfamiliar with the DOS FORMAT command, we encourage you to read Using DOS FORMAT.  During the high-level format the operating system writes the structures required for managing data on the disk, basically creating a table of contents. Hard Drive Problems

Your computer's hard drive holds your data, and so is an extremely important and valuable part of your computer. Many things can go wrong with a hard drive -- some catastrophic, like total failure of the media (which can prevent your system from starting), while others are not catastrophic, but still serious, like sectors going bad (which can lead to the loss of a file or files). For many of these problems the troubleshooting activities vary only slightly, depending on, for example, if your drive is receiving power or not.

Warning: Anytime you do hard drive troubleshooting you take the risk of destroying any data on the drive.  It is always wise to attempt to backup any data prior to troubleshooting, if possible.  Never reformat or partition a hard drive except when all else fails, and always do regular backups of your data.

If you do not see any drive light or hear any drive activity and you see a message that the drive cannot be located:

Try booting from a boot diskette.  If you can, then the problem is with your hard drive.

Try switching to the C: drive.  If you can access it, you may have a boot sector problem.  You should check for a boot sector virus immediately.  If you can not access the drive, continue troubleshooting.

Make sure the 4-pin power connector is inserted securely into your drive.  If your drive is on a Y-connector make sure any other connections are secure.

Make sure the signal cable from the motherboard to the hard drive is securely seated.  Check for frayed edges or other damage to the cable.  If you have a spare, try replacing it.  If it is an EIDE cable, make sure that the red stripe that you see on one side of the cable is aligned properly with pin 1 on the disk. Enter your CMOS Setup and confirm that the parameters for your hard drive (heads, cylinders, sectors per track, etc.) have been entered correctly.  In most modern systems these parameters are automatically detected, but you need to make sure that the drive is enabled in the CMOS Setup.

If you have a separate drive controller board (most modern systems do not unless they are SCSI systems) make sure it is seated securely in its expansion slot.

Try switching the non-working drive with a known, good, drive (remember you may need to change the CMOS Setup settings if you do this).  If the good drive works as you expected, then the replaced drive is bad.  If it does not work as expected, you may need to replace the disk controller or the motherboard.

If you see or hear drive activity (constant or intermittent) but your boot drive can not be located (you might see a "No Fixed Disk Present" error message on the monitor):

Check to make sure the signal cable is inserted with the proper orientation (that it is not reversed at one of the two ends).

Check that the signal cable is in good condition.

Check that your CMOS Setup has the correct parameters for the drive.  In most modern systems these parameters are automatically detected, but you need to make sure that the drive is enabled in Setup.

Boot from a boot floppy and attempt to access the drive.  If you can, it is likely that your boot files have become corrupt. Check for boot sector viruses immediately.

Try switching the non-working drive with a known, good, drive (remember to change the CMOS Setup settings if you do this).  If the good drive works as you expected, then the replaced drive is bad.  If it does not work as expected, you may need to replace the controller board.

If you hear your disk spinning up but it is not recognized by your computer (you may get a "Hard-disk error" or "Hard-disk controller failure" message on your monitor):

Check to make sure the signal cable is inserted with the proper orientation (that it is not reversed at one of the two ends).

Check that the signal cable is in good condition.

Check that the jumpers are set correctly on your drive(s) and that the primary is set as primary, and the secondary or slave, if existent, is set as secondary.

Check that your CMOS Setup has the correct parameters for the drive.  In most modern systems these parameters are automatically detected, but you need to make sure that the drive is enabled in Setup.

Boot from a boot floppy and run FDISK to confirm that there is at least one DOS partition and that if it is your boot drive, that it is active and bootable.

Boot from a boot floppy and attempt to access the drive.  If you can, it is likely that your boot files have become corrupt. Check for boot sector viruses immediately.

Try switching the non-working drive with a known, good, drive (remember to change the CMOS Setup settings if you do this).  If the good drive works as you expected, then the replaced drive is bad.  If it does not work as expected, you may need to replace the disk controller or the motherboard.

If you hear your drive spinning up at start-up but then it immediately spins down:

Check to make sure the signal cable is inserted with the proper orientation (that it is not reversed at one of the two ends).

Check that the signal cable is in good condition.

Check that the jumpers are set correctly on your drive(s) and that the primary drive is set as master, and the secondary or slave, if existent, is set as slave.

Try switching the non-working drive with a known, good, drive (remember to change the CMOS Setup settings if you do this).  If the good drive works as you expected, then the replaced drive is bad.  If it does not work as expected, you may need to replace the disk controller or the motherboard.

=====================

If your drive light remains on continuously but the drive is not accessible:

The boot drive can not be located.  Check to see that the signal cable is inserted with the proper orientation (that it is not reversed at one of the two ends).

Replace the drive with a known, good, drive (remember to change the CMOS Setup settings if you do this).  If the good drive works as you expected, then the replaced drive is bad.  If it does not work as expected, you may need to replace the disk controller or the motherboard.

If your drive spins up when power is applied, but then rapidly spins down again:

Make sure the 4-pin power connector is inserted securely into your drive.  If your drive is on a Y-connector make sure any other connections are secure.

Make sure the signal cable from the motherboard to the hard drive is securely seated.  Check for frayed edges or other damage to the cable.  If it is an EIDE cable, make sure that the red stripe that you see on one side of the cable is aligned properly with pin 1 on the disk.

Try a new signal cable.

Check that the jumpers are set correctly on your drive(s) and that the primary drive is set as master, and the secondary or slave, if existent, is set as slave.

Replace the drive with a known, good, drive (remember to change the CMOS Setup settings if you do this).  If the good drive works as you expected, then the replaced drive is bad.

If, after completing the installation process for a SCSI drive, the drive's LED does not show on/off activity when the host is trying to communicate with the drive:

A duplicate SCSI ID setting may be the problem.  If this is the case, change the ID so that each device on the SCSI chain has its own unique ID. If the SCSI IDs are all unique, make sure that your SCSI chain is terminated properly.

If you see a "Sector not found" error message on your computer monitor you likely have a non-catastrophic media problem with your hard drive, mostly likely some sectors are going bad.  This commonly happens to older drives:

Use a disk utility to evaluate the drive and then locate and map out any bad sectors. If you have a large number of bad sectors, you might consider getting a low-level format utility from your drive manufacturer, backing up your drive, doing a low-level format, and then reformatting and restoring your drive.  This may help you to recover some of the bad sectors.

If you suspect that your boot sector has been corrupted or destroyed, and that this has not been caused by a virus, you should try running FDISK with the /MBR parameter.  This command will replace your master boot record without altering the partition table at the end of the sector.

If running FDISK /MBR does not work, you will need to run a full FDISK and then DOS FORMAT on your drive.  This is a drastic step and will destroy the data on your hard disk.  Before you do so, you can boot off of a start-up floppy and try to backup any critical data that may not be on your most recent system backup.

=======================

If your computer fails to load the operating system on a cold boot, but seems fine when you do a reset (warm boot):

It is possible that your hard drive is not powering up in time for the cold boot to execute properly.

Check your CMOS Setup and make sure the POST has not been disabled.  If it has, the lack of the RAM test may be giving your disk inadequate time to spin up before the operating system is loaded.

Your CMOS Setup may give you an option to enable or disable the "Quick" POST.  Make sure you are running the full POST (as opposed to the quick).  This longer POST includes a more in-depth memory test, and may provide your disk sufficient time to spin up before the operating system is loaded.

Alternatively, your CMOS may have a Power On Delay/Boot Delay option that will allow you to specify a time in seconds to delay the boot process.  This is usually disabled, and should be unless you are experiencing this problem.

Your hard drive's performance seems to be degrading over time, getting slower and slower.

Boot from a boot floppy and make sure there are no drivers or Terminate but Stay Resident (TSR) programs being loaded.

Check for viruses.
Run a defragmentation.

====================================

The Motherboard

The motherboard (sometimes referred to as system or main board) in your computer is a highly complicated piece of hardware.  It is the basis of any computer, and possibly its single most important part.  The typical motherboard contains a number of individual components, including:

CPU socket (or slot)
CPU voltage regulators
Chip set
Memory SIMM or DIMM sockets
Level 2 cache
Bus slots
ROM BIOS
Level 2 cache
Super I/O chip
Clock/CMOS battery

All the critical subsystems run directly off of your system's motherboard, and it manages all the data transactions between the CPU and the computer's peripherals.
Motherboard

Major Components

It's a good idea to identify the crucial parts of your system's motherboard.  Although each motherboard is designed differently, its actually quite easy to identify most of the major components by their size, shape and placement.  Of course, having your system's manual on hand will make this task a lot easier.

The physical dimensions (shape and size) of the motherboard dictate the type of case it can fit into.  There are several common dimension standards for motherboards.   Because non-standard boards are hard to upgrade, it is wise to avoid them.  A truly standardized motherboard is interchangeable with other boards of its type.

A few of the most common standards for motherboards are: Full-sized AT, Baby-AT, ATX and NLX.  Systems that use proprietary motherboards, such as Compaq and Packard Bell, can be very expensive or even impossible to upgrade or repair after the warranty has expired.

Sockets and Slots

When examining your motherboard, you'll likely find it designated something like  "Socket 7," "Socket 8," or "Slot 1."   These classifications refer to the type of CPU that the motherboard can support.

For example:

Socket 7 motherboards are generally designed for Pentium and Penium MMX CPUs
Socket 8 motherboards are made for PentiumPro CPUs.
Slot 1 motherboards are for Pentium II systems.

These designations mean that the motherboard supports a given class of processor, as opposed to any such processor.   Some Socket 7 motherboards can support all Pentium and Pentium MMX CPUs, but older Socket 7 motherboards might only support Pentium CPUs up to 120MHz.

Chip Sets

If the CPU is the brain of your computer, the chip set is its heart.  It controls the data that flow between your computer's CPU, system memory and the system bus.  Chip sets are intelligent controller chips that are located on the motherboard.  They are a critical component that is closely related to the CPU because they control the buses around it.  Chip sets are the mechanism that allows the RAM and the I/O buses to work with the CPU.

Comprised of one or more chips, chip sets integrate numerous functions, including the Clock Generator, Bus Controller, System Timer, Interrupt Controller, DMA Controller, CMOS RAM Clock and Keyboard Controller.

There are many different chip sets and new, improved (faster, added capabilities) chip sets are introduced continuously.   Historically, Intel has been the leader in chip set technology, releasing successive generations with new and more powerful features, but a number of other companies (Acer Laboratories, VIA Technologies and Silicon integrated Systems to name three) manufacture chip sets as well.

One warning: swapping a new Pentium II board into an older system chassis may be hazardous to your new CPU.  The Pentium II (Slot 1) cooling requirements are substantially different than the Pentium (Socket 7).

Motherboard Problems

Because your computer's motherboard is a complex piece of electronics that handles numerous system resources, troubleshooting problems can be difficult and time consuming.  And because it is expandable -- allowing for the addition of "expansion devices" such as sound cards and the like -- resource conflicts can result when you add new devices (there are three types of resources provided by the motherboard for expansion devices: Interrupt Requests, DMA channels, and I/O port addresses).

As expansion devices are added to a PC, resources are assigned to each particular device and no two devices may use the same resources -- if this happens, conflicts occur and the system will work improperly or not at all.   You can get more help with such resource conflict in our I/O & Expansion Cards troubleshooting section.

When problems are indicated with your motherboard, it is wise to do the following general checks to make sure the problem isn't a basic one.  Always remember to unplug the system when working on it and to use proper ESD protection.

Check all cables and connectors.  Make sure all cables are seated securely.   Make sure all of your cables are in good working order, do not have frayed or damaged edges or ends, and are not crimped by the cover when replacing it.

Check all socket-mounted components such as the CPU.  Make sure they are securely seated, and that any locking levers are tightly locked down.

Check power levels, including the wall outlet and other devices (such as printers or coffee pots) that may be using the same circuit and causing power fluctuations.

Check the motherboard for foreign objects such as loose screws, loose wires, or other objects that may be short circuiting it.  Make sure there are no cracks.

Use your manual to confirm that all the jumpers and dip switches on the motherboard are correctly set.

Before taking the drastic step of replacing your motherboard, go to the next level of troubleshooting steps:

Remove all optional devices from the motherboard, including expansion boards, external peripherals, etc. so that you just the minium in the machine required to make it work (CPU, full bank of memory, video card and drive).  See if that resolves the problem.

Reset all CMOS BIOS settings to the default, conservative values to make sure an overly aggressive BIOS setting isn't causing the problem.  Set all cache, memory and hard disk timing to as slow as possible and turn off BIOS shadowing.

Try replacing the video card with a known-good one.
Try replacing the power supply with a newer, more powerful one.

=======================

If your computer spontaneously reboots or crashes for no apparent reason:

Check your system for viruses immediately.  This is a symptom of memory-resident viruses.

Check your power supply cables -- both external and internal.  Make sure your internal 4-pin cables are seated securely on the motherboard.

Move your computer to another power outlet.

Make sure your CPU is seated securely and that the heat-sink/fan is operating properly to keep it cool.

Make sure your computer's memory modules are seated securely.

Check that all your expansion boards are seated securely and evenly on the motherboard.

Examine your motherboard for metal pieces that might be touching other metal pieces (such as screws or the board's seatings).

If you recently added a new expansion board, make sure there are no interrupt (IRQ), DMA channel or I/O address conflicts.

If you can not get your system to boot up, but the problem goes away when the PC's outer cover is removed:

Try placing the cover on the system and booting PRIOR to replacing any of the cover's screws.
If the PC boots without any screws securing the cover, carefully insert and tighten each screw with the system on.

This way you may find the screw that is causing an intermittent connection -- leave it out for the short-run, but in the long-term, you will need to replace the motherboard.

If you are unable to get a parallel-port device (for example, a printer) to work with your motherboard:

Check that the CMOS Setup parallel-port mode (i.e., SPP/ECP/EPP) is configured correctly for the peripheral device you are trying to use.

Spontaneous Rebooting Could Be Caused By:

As far as I know, these main culprits are heat (overclocking or faulty fans/ventilation)

1. Defective power supply
2. Bad power input (from the electrical socket or power company)
3. Defective memory (or memory installation) (related to a short between the Ram and
the mainboard)
4. Defective mainboard (internal short)
5. Bad case installation (something is a hair away from shorting)
6. Having DMA on when your Motherboard or Bios doesn't fully support it.
7. Your PC is plugged into a AC plug that may be marginal or have a cycling appliance
on it like a heater or air conditioner
8. I've seen anti-virus programs cause this also, in the last problem I worked on,
PC-Cillin, was the source of the spontaneous rebooting.

Reboots are usually caused by electrical shorts.

=================================

CPUs

Your computer's CPU (or processor) is centrally located on its motherboard, and it is the brain of your PC.  The CPU carries out much of the work on your system, and is constantly passing data and receiving instructions.  There are many brand names and
many manufacturers of CPUs, and the manufacturers are often producing several models at once, which overlap.  This can make for a difficult time in identifying your computer's particular CPU.

CPUs are usually specified by two main parameters:

Speed: cycles per second counted in megahertz (MHz).
Width: I/O bus, internal registers and memory address bus.

For example, a Pentium II 333MHz has a  64-bit I/O bus width, a 32-bit internal register size, and a 36-bit memory address bus width.  An older 486 DX4 had a CPU speed of 100 MHz, and 32-bit I/O bus, internal register and memory address bus widths.

These specifications, along with the manufacturer and model are what uniquely identifies a given CPU.

CPU Fan

All of today's CPUs share a common need for cooling.  The cooling mechanisms (usually a fan and a heat sink) are matched to the CPU size, and must be attached properly to provide maximum cooling.  Most modern systems have an improved motherboard design that provides for superior cooling of the CPU because of its position near the power supply.

Modern systems have a Socket or Slot Number that facilitates interchangeability of CPUs.  These numbers (Socket 1, Socket 2, Socket 3 and so forth) are directly related to the processor chips that they can support.  Pentium II processors and beyond use a slot where the processor card or cartridge is installed (Slot 1 or Slot 2).

Upgrading Your CPU

One very common way of improving your system is to upgrade its CPU, thereby increasing clock frequency, cache and other attributes that have a direct impact on your computer's performance.

CPU upgrades are a relatively easy and common system improvement.  Upgrading the CPU processor is often the fastest and easiest way to gain performance without upgrading the entire motherboard.  To maximize your motherboard you can almost always upgrade to the fastest CPU that its processor socket will support.  Some upgrades will require a voltage regulator adapter.  For example, if your motherboard has a Pentium Socket 5 with a 100MHz CPU, you can upgrade to a Pentium MMX 233MHz, as long as you add in a 2.8v adapter.

You need to stay within the same class when upgrading your CPU.  You can replace a slow 486 with a faster one, even possibly from a different vendor, but you can not replace it with a Pentium without replacing your entire motherboard or purchasing a special version of the new processor designed CPU Fan specifically for this type of upgrade.

Be very careful when you replace your new CPU that you install it properly aligned.   If you accidentally put it in sideways it is likely you will ruin your new CPU and your motherboard -- an expensive mistake.  From that point on your motherboard may ruin any chip you install.

Don't forget that you should buy a new fan when you upgrade your CPU.  Better yet, get a boxed CPU with an integrated fan.  That way you'll know you are getting the proper size of fan.  Make sure your fan has a ball-bearing motor; the cheaper sleeve bearing motors often freeze after a very short life.  If you've max-ed out on your CPU then your next step is a motherboard upgrade.

CPU Problems

When a CPU fails, it generally fails in a big way.  A System that has not been changed recently, but suddenly does not boot, or freezes during the boot process, is a good CPU candidate for a CPU problem.

If your computer powers up briefly, and perhaps even your hard drive spins up, but then your system dies suddenly and completely, you may have a problem with your CPU:

Make sure that the CPU is installed securely in its socket or slot.  Make sure all system cables and connectors are secure.

Check first for hardware conflicts or a defective peripheral device by removing all such peripherals and expansion boards and attempting to start the system.  If it CPU starts, one of the expansion devices is interrupting your system's operation.  Re-install each device, one at a time, and check the system, to determine which device is bad.

Replace your CPU with a known, good, CPU in the motherboard.  If it works, it is probable that your CPU is bad, if it does not, you may have to replace your motherboard.

If your system starts up without a problem but then crashes or freezes a few minutes after operation your CPU may be over-heating:

Make sure that the CPU is installed securely in its socket or slot.  Make sure all cables and connectors are secure.

CPU Socket

Make sure your CPU heat-sink/fan is working properly.  If you do not have a fan, consider adding one.

If your system starts up without a problem but crashes or freezes when a certain application(s) is run:

Try a known, good CPU in the motherboard.

=====================

If you receive a beep code that indicates possible CPU fault:

Check that the motherboard's 6-pin power connector is securely inserted.
CPU Fan
Power Cable

Try a known, good CPU in the motherboard.  If the new CPU fails to correct the problem, you may have to replace the motherboard.

If your system runs fine but your BIOS reports the wrong type of CPU:

The motherboard BIOS is not written to support your particular CPU directly.
You may need a BIOS upgrade, if available.  Contact your system or motherboard manufacturer.

Overclocking

Overclocking is when a CPU is run at a speed higher than it is rated for.  The concept was not invented by high-tech users trying to get the most from their hardware, but by unscrupulous vendors who, to save money, would install a CPU rated for less than the actual operating speed.

If you suspect that your vendor may have done this, you can look at your chip's markings.  Usually the part number will end in a suffix of -xxx where xxx is the number indicating the maximum speed.  For example -333 indicates that the chip was rated for 333MHz operation.  One word of caution -- some vendors actually remarked the CPUs, obscuring the original part number.

You can set jumpers on your motherboard to force your CPU to run faster than it is designed to run.  It is easy to be dazzled by the concept because the CPU is one of the most expensive components on the motherboard, and faster-rated chips cost more money.   In addition, it is widely speculated that the manufacturers rate their chips conservatively.

Despite this, I do not recommend overclocking here as the chip will run hotter than it would normally.  Symptoms of overheating include random lockups, glitches and headaches.  If you do decide to try to overclock your CPU and experience such problems, it would be wise to return it to the original speed.

=======================================

RAM - Working Memory Storage

All the data that your system uses and works with during operation are stored in its RAM.  In order for your computer to use information stored on a drive, it must be read into working memory storage, which is made up of RAM chips.  This memory is temporary because the data and programs are there only as long as the system is powered.

When a system is shut down properly (as opposed to being powered off while still in Windows) any changes made to the data are copied back to disk.  This is one reason why it is important to shut down your system gracefully by using the shut down command and then waiting for the "It's now safe to turn off  your computer" message before turning it off.

RAM is a very complicated and highly technical subject.  I can touch the most important points that you will need to know if you are troubleshooting RAM problems.

The traditional RAM type is called DRAM (dynamic RAM).  The other type is SRAM (static RAM).  While SRAM can respond much faster than DRAM, the latter is cheaper and is becoming faster all the time.  Types of DRAM include:

FPM (Fast Page Mode) - The RAM traditionally used for PC's prior to the introduction of EDO RAM.  Mounted in SIMM modules of 2, 4, 8, 16, or 32 MB.  Typically found in 70 ns (nano seconds) and 60 ns (faster) versions.  You can not mix different speeds on the same Pentium motherboard.

ECC (Error Correcting Code) - often used for servers.

EDO (Extended Data Output) - Faster than FPM, it is usually sold as 60 ns.

SDRAM (Synchronous DRAM) - The first DRAM to synchronize itself to the timing of the CPU.  With today's faster CPUs this type of RAM provides significant speed improvement over EDO.  Found as 168 pin DIMMs.

The type of RAM you can install on your system is controlled by the motherboard and its chip sets, and must match their specifications.  In addition, RAM chips come in different sizes, which must also match the motherboard's requirements.

In the early days, memory chips were installed individually, then someone figured out to put the individual chips on an expansion card.  These cards were called SIMM (single inline memory module) modules, and at first they had only a few MB of RAM on a 30-pin card.  Later, as their capacity increased, so did the number of pins (from 30 to 72 to 168).  SIMMs can have 2 or 4 or 8 chips on each side, and can be double-sided.

Many early computers required that SIMM modules be installed in pairs, in what were called memory "banks" of sockets, which would hold two modules.

SDRAM are made on DIMM (dual Inline memory module) modules.  They have a 168-pin edge connector and require modern motherboards.  Unlike SIMMs, you can install one module at a time, and their advantage is significantly increased speed.  Modern systems can support from 4 to 8 DIMM sockets.

A quick test of whether you are looking at a SIMM or a DIMM is to count the number of notches at the bottom.  SIMMs have one notch, DIMMs have two.  SIMM slots are white and held in place by small clips at both ends.  DIMM slots are black and have much larger clips on both ends.  DIMMs are inserted upright, with the clips snapping up to lock the module into place.  SIMMs are inserted at a 45 degree angle and tilted forward until they are upright and the clips snap into place.

Upgrading Memory

PC memory speeds range from about 10ns to 200ns.  When you replace a failed memory module, you must install a module of the same type and speed, or faster.  Substituting a faster module, however, doesn't necessarily improve your system's performance because it will operate the new memory at the slower speed.

In addition, it is important to know if your memory sockets are tin-plated or gold-plated, and to install memory that is of the same metal as the socket.  Mixing the plating can result in eventual memory failure.

Adding memory to a system is an excellent way to improve the performance of your computer, and as memory prices get cheaper, adding memory is cost-effective as well.  To do this you can either add memory to empty slots on your motherboard, or replace the current memory with higher-capacity modules.

If you need help determining how much memory your system currently holds, you can simply reboot your computer and watch for the POST memory test, which counts up your memory and displays the count as it does so.   Or you can right click on the My Computer icon on your desktop, go to Properties, and look at your System Properties, which displays the current memory near the bottom of the window.

Always follow the motherboard manufacturer's recommendations on which memory you can install.  In addition, some chip sets and BIOS can not support more than 64MB of RAM, adding more than that will actually decrease your computer's performance.

Memory must also be installed in banks, and some systems require multiple memory chips (generally two or four SIMMs) in a given bank.  On other systems (for example, Pentiums and above) a single DIMM represents an individual bank, and they can be added or removed one at a time.  In general, modern machines have from 2 to 4 memory banks, and a bank does not work if it is only partially filled.  The documentation for your computer's motherboard describes its particular requirements in detail.

The physical orientation and numbering of SIMMs or DIMMs on a motherboard is not standardized across manufacturers.  Use your system documentation to identify memory bank layout.

It is critical to prevent electrostatic discharge when you install memory modules, and to install them according to the accompanying instructions.  Make sure your system is turned off and unplugged from its power source.  Memory modules are keyed (notched) to ensure proper installation in the expansion socket.  Never use force to install a module - if it does not slip easily into place and lock, it may not be aligned properly.

On older systems you may have to run the CMOS Setup to add the new amount of memory (modern systems will do this automatically).  When you power up the first time watch your Power On Self Test to make sure the memory test agrees with the total amount of memory you now have installed.

NOTE: Sometimes adding a lot of memory to a computer can increase the demand on the power supply, leading to power supply failure.  The only way to solve this problem and keep your new memory is to replace the power supply.

Errors and faults in system memory can crash a system, make it behave erratically, or even stop you from installing an operating system.  Symptoms of bad memory can manifest themselves in many ways, including "Parity Check" errors, "201" errors, "164" errors, "Incorrect Memory Size" errors, "ROM" errors, "Memory Address" errors, and more.

Although in modern systems the address of a fault is generally presented in the error message, it is usually impossible to correlate a specified bad address with a particular component, so trial and error testing is your best option.  Generally, when you suspect memory problems you can try the following:

If you have just installed a new peripheral device, check Device Manager to make sure there are no conflicts (look for a yellow exclamation mark or a red "X" symbol).  If there is no hardware conflict specified, take out all devices and reinstall them one at a time to determine the device that is causing the problem.

Check your memory by reseating each unit and restarting your system each time to be sure that each unit is inserted properly and working.

Check your CMOS Setup settings to make sure the correct amount of RAM has been entered.

Check your motherboard jumpers if necessary, especially when you add more memory, to make sure that bank 0 is defined properly, along with the number of active banks.

Check your memory speed to make sure that the RAM in your system is fast enough to support your CPU.  Never mix memory speeds within the same bank, and do not mix memory types.  It is inadvisable to mix parity and non-parity RAM.

Try a known, good memory unit by rotating it through each occupied memory socket in sequence.  If the error disappears when the known, good memory is in a given slot, then the memory that was displaced is likely bad.

If you see the number "164" displayed on the monitor:

This is a generic memory-size error - the amount of memory found during the POST does not match the amount of memory listed in the system's CMOS Setup.

Check your CMOS Setup settings to make sure the listed memory matches the actual memory amount.

Adjust the figure in CMOS Setup if necessary.

If CMOS Setup parameters do not remain in the system after repowering the computer, try replaing the battery.

Note that some modern CMOS Setup routines automatically detect the amount of RAM so do not list it.  If this is the case, try recalculating memory by entering CMOS Setup, saving changes and then exiting.

If you see an "Incorrect Memory Size" error message displayed on the monitor:

This message can be displayed if there is a CMOS Setup error, or if there is an actual memory failure.

Check your CMOS Setup as described above.

If CMOS Setup parameters are correct, try a known, good memory unit by rotating it through each occupied memory socket in sequence.  If the error disappears when the known, good memory is in a given slot, then the memory that was displaced is likely bad.

If you see a "ROM Error" message displayed on the monitor:

The POST system ROM test has failed.  You may need to replace the system BIOS ROM(s).

If you see a general "Memory address line failure" error or similar error on the monitor:

Try a known, good memory unit by rotating it through each occupied memory socket in sequence.  If the error disappears when the known, good memory is in a given slot, then the memory that was displaced is likely bad.
If, after rotating through all occupied memory the problem persists, you may have a motherboard problem.

You see a "Decreasing Available Memory" error message on the monitor:

This message indicates that a failure has been detected.  In order to allow the system to function, all memory after the failure has been disabled.

Check your memory by reseating each unit and restarting your system each time to be sure that each unit is inserted properly and working.

Try a known, good memory unit by rotating it through each occupied memory socket in sequence.  If the error disappears when the known, good memory is in a given slot, then the memory that was displaced is likely bad.

Many CMOS Setup programs allow you disable the "Quick" Power On Self Test to get a more detailed test that counts down your memory several times as different areas are tested.

The terminology used to describe these test options differs between manufacturers.  Consult your documentation if you are unsure of the proper setting for the detailed POST on your system.

=======================

The full memory test can be very helpful in troubleshooting memory problems:

Disable your "Quick" POST or enable the "Full" POST (depending on your BIOS).

If you have just one memory module, run the full POST on it and note any POST errors.

If you have multiple modules, remove them, then run the full POST on each separate chip, rotating through all, and note any POST errors.  Remember you may need to rotate sets of multiple modules so that you have a complete bank.

If Device Manager indicates that there is a memory conflict, you likely have two expansion boards that have overlapping ROM or RAM addresses.  In these cases typically neither board will function properly when both are enabled; but disabling one will often cause the other to work properly.

Memory conflicts can be a particularly difficult problem to troubleshoot, and you need to examine the documentation for each adapter to find out what memory addresses it uses and how to change them.  In order for the two boards to co-exist you will need to reconfigure one of the boards by either changing jumpers, switching settings or software-driver parameters.

In a Plug-and-Play (PnP) system with PnP adapters, you may be able to use Device Manager's Edit Properties  to change memory usage on a conflicting board.

===================================

I/O and Expansion Cards

If the brain of your computer is the CPU, its expansion card slots (or buses) are the nervous system.  They are the common pathways by which the various components of your computer send and receive data.  Thinking about buses can be difficult, because unlike many of your computer's other components, its buses are more virtual -- not having a real physical piece of hardware to see, but rather being made up of wires and connectors.

There is a hierarchy of different buses in a computer, with slower ones connecting to faster ones above them.  I/O (input/output) buses connect the all the various components of the system except RAM (sound board, graphics board, network card, keyboard, parallel and serial ports, and so forth) to the CPU.  The system or processor bus connects the CPU to the motherboard's chip set.  A subset of the processor bus, the memory bus, connects your system's CPU with its main memory.

In a modern desktop computer there are generally two or three different I/O buses (or slot types): the ISA (industry standard architecture) bus, which is an older and slower bus (If you have different sized slots, these would be the longer ones.  If the slots are all one size, they are all ISA);  the EISA (enhanced industry standard architecture) bus, which are typically used on servers, not desktops; the PCI (peripheral component interconnect) bus, which is a newer and higher speed bus; and the USB (universal serial bus) bus, which is a new, lower speed bus.  Another type of I/O bus, the AGP (accelerated graphics port) bus, is used exclusively for the graphics card.

In laptops, you'll see PCMCIA (personal computer memory card international association) sockets in which you can plug removable credit-card size devices.  These cards can contain extra memory, hard drives, modems, network adapters, sound cards, and more.

==========================

Buses can lead to internal or external ports (devices) or other bus controllers.  For example:

ISA internal controller LPT1 can lead to a printer or a parallel port scanner.
ISA internal controller COM1 can lead to a mouse, modem or other serial device.
ISA internal controllers can also lead to a floppy drive or a keyboard.
ISA adapters can lead to a sound card or other peripheral.
PCI EIDE controller can lead to a hard disk, CD-ROM, or zip drive.
PCI adapter can lead to a network card, sound card, modem, graphics accelerator, SCSI controller, hard disk or scanner.
USB ports can lead to a keyboard, digital camera or scanner.

An expansion board, such as a PCI network interface card, is used to extend the functionality of your computer.  Some of the most common expansion boards include video cards, sound cards, modems and video capture cards.  These cards provide a link between your CPU and various peripherals.  Most motherboards provide several ISA and PCI slots for expansion boards.  The different types of I/O cards have their own slot configurations; you can not accidentally put an ISA card in an PCI bus slot.  Because there is a limited number of each type of slot, you are limited to the number of such expansion devices you can add to your system.  The adapter card fits into the appropriate slot and it gets connected to the I/O bus.
Resource Conflicts

Expansion cards and the devices they support require the assignment of low-level system resources so that they can communicate with the bus.  These resources typically include the following:

Interrupt Request channels
Direct Memory Access channels
I/O Port addresses

As you add more expansion cards to your system you will find that the potential for resource conflicts increases.  And although Plug-and-Play systems should automatically resolve resource conflicts, they sometimes do not.  The most common symptom of a resource conflict is that one of your computer's devices simply stops working.  Other possible symptoms might be:

You can not format a floppy disk.
Your mouse does not work.
Your system hangs or freezes frequently.
Your sound card does not sound right.
Your printer prints garbage.

IRQ

The IRQ (Interrupt Request) is used to regulate traffic across I/O buses.  These are sometimes called hardware interrupts.  These interrupts allow the peripherals to communicate with the CPU in an orderly fashion.  Although PCI buses allow for shared IRQs, ISA buses do not.  So there is the potential for conflict between peripherals assigned the same IRQ.  A new ISA network card may have been preset by the manufacturer to have an IRQ of 9.  If you install it in your system and IRQ is already assigned to another device, you will have a conflict and your system may not even start.

In modern systems the bus will support 16 interrupts, 15 of which are "available" to the system.

DMA

The second possible source of conflict is the DMA (Direct Memory Access) channel.  This allows an adapter to transfer data to the RAM directly, bypassing the CPU.  These channels are used by high-speed communications devices such as sound cards and SCSI adapters.

There are eight DMA channels supported by modern systems.  Typically five or six these are available.  Conflict can result if two devices request the same DMA channel.

I/O  Ports

I/O ports enable communication between your computer's devices and software.  The CPU has assigned address "ranges" for its adapters, ports, and so forth.  Driver programs are what interact with devices via I/O ports.  The driver must know what ports a device is using, otherwise it can not work with it.  Similarly, the device must know what ports the driver is using.

There are over 65,000 such ports (numbered using hexadecimal addresses 0000h to FFFFh).  Because there are so many ports, and because port assignments are widely standardized, I/O port conflicts are rarer than IRQ and DMA conflicts. Motherboard and chip set devices are typically set to use I/O port addresses from 0h to FFh; all other devices use from the 100h to FFFFh range.  This table includes some adapter card examples.

Device conflicts can be easily seen using Device Manager in System Properties.  Just  look for yellow exclamation marks or red "X" symbols.  By looking at the properties of a conflicted device you can tell which resource assignment is causing the problem.

Troubleshooting Resource Conflicts

Checking for IRQ, DMA channel, and I/O address conflicts can be difficult.  Of the three, the most common source of problems is the IRQ, because there are few of them and they are in high demand.   As we've shown, your computer has only a limited number of these resources for use by various devices, and no two devices can use the same resource without causing conflict.  Classic signs of resource conflict are:

System fails to boot.
System freezes during boot.
System freezes during operation for no apparent reason.
Erratic device behavior.

Fortunately, resource conflicts are almost always the result of an incorrect upgrade.  If the following rules are all true, it is likely you have a resource conflict:

A piece of hardware and/or software was recently added to the system.

The trouble occurred after a piece of hardware and/or software was added to the system.

The system was working fine before the hardware and/or software was added.

If you experience such symptoms after you installed a new expansion card, it is likely that the new card is causing the conflict.  Similarly, if the symptoms occurred after installing new software, it is probable that the software is causing the conflict.  If two similar devices are not working (like your system's keyboard and mouse) it is likely that they are conflicting with each other.  By far the most common causes of resource conflicts are sound cards and SCSI adapters.  This is because both (especially sound cards) use an enormous amount of resources.

The following set of procedures can help you to troubleshoot such device conflicts when they occur after installing a new expansion device.  Before starting, have all your system and adapter card manuals handy, and make sure that you write down your system's current settings, so that you can go back to the original configuration if necessary.   Follow the cardinal rule of only making one change at a time, then testing the result.

Some computer viruses may appear to be resource conflicts.  Always run a virus check before going on to more difficult conflict resolution steps.

Right-click on the My Computer icon on your desktop and go to Properties.
Click on Device Manager and select the new device.  You may need to expand the device's entry by clicking on the [+] symbol to the left of the component's category.

If the new device has a yellow exclamation mark or red "X" symbol flagging it, that means it is conflicting with another device.

Use the Properties Resource tab to see what is causing the conflict.

Change the conflicting resource.  This change may be a hardware-based physical change (for example, the IRQ on your new device as set by jumpers) or a software change made in the setup program.

Shut down your system, re-install the device, and run the setup software.

Window's Device Manager is a good tool to help you troubleshoot IRQ, DMA and I/O address conflicts.  By double-clicking on the Computer entry at the top of the Device Manager listing, you can View Resources and easily see all the IRQs, I/O ports, DMA channels and memory addresses in your system, and the devices using them.

If the newly installed expansion device is a PCI adapter:

Use Device Manager in System Properties to look for a yellow exclamation mark or red "X" symbol indicating a device conflict.
Use the Properties of a conflicted device to see which type of system resource is causing the problem.  For most PCI cards this will be an IRQ conflict.

Some motherboards let you use your CMOS BIOS to change IRQs assigned to PCI slots.  If you have a conflict between an ISA card and a PCI slot, you can try fixing it in your CMOS BIOS Setup.

A bargain-basement or non-standard PCI card (especially in combination with a cheap motherboard) can cause problems with IRQs and lead to IRQ conflicts.  If you just can't get a new PCI controller to work on your system, you might try it on another computer.  If your problems persist, you may need to replace the PCI card.

Plug-and-Play

Plug-and-Play (PnP) was a real revolution in computer technology.  Despite the jokes that call it "Plug-and-Pray", it has reduced the need for computer users to mess with dip switches and jumpers each time they add a new device to their system.  Most new systems come ready to support Plug-and-Play with PnP compatible hardware, BIOS, and operating system (all three are necessary to full PnP compliance).

But even Plug-and-Play hardware can cause IRQ conflicts.  Sometimes PnP can misdetect, or simply not detect new hardware -- especially the first time you try to install it.  If you get a conflict during the initial install for a PnP device:

Go to Device Manager, remove the device manually and then try again by restarting your system and going through the device detection routine.

Make sure you have the most current drivers for your new device.  Even if you just bought it, the box may have been sitting on the shelf for some time and chances are good that newer drivers may be available.  It is always best to use the most current device drivers that you can obtain.

If the new device is not automatically detected, you will need to use the Add New Hardware wizard.

If this doesn't work you can try the following steps to change the automatically set resource assignments to those you enter manually:

Make sure you have the most current drivers for your new device.  Even if you just bought it, the box may have been sitting on the shelf for some time and chances are good that newer drivers is available.  It is always best to use the most current device drivers you can obtain.

Double-click on the new device's entry in Device Manager.  Go to the Resources tab and clear the Use Automatic Settings box.
Double-click on the IRQs, I/O ports, and UMBs (upper memory blocks) that the Device Manager says are in conflict with other devices on your computer.

Use the Edit Resource dialog to make new entries -- Device Manager should limit your choices to values that are supported by your device, and warn you of further conflicts between the newly entered values and any other device assignments.  When you have successfully changed the conflicting assignments, exit out of the dialog by clicking on "OK" and reboot your system.

If you disable a PnP device in CMOS Setup and its resources are not freed.

Disabling the device in your system's CMOS Setup is not enough, you must first make sure you disable it from Windows Device Manager:

In modern systems the bus will support 16 interrupts, 15 of which are "available" to the system.

=========================

The Video Subsystem

Video problems may not stop your computer from working, but they can make using it a miserable experience.  Your computer's video subsystem is comprised of its monitor, which is connected by a cable to the video card; the graphics (or video) card, which is an expansion card that generates electric signals to the monitor; and a device driver, which the operating system uses to control the video card and ensure that it sends the correct signals to the monitor.

These three pieces work together to create the image you see on your monitor's display.  Getting them to work together properly can be more of an art than a science, and fine-tuning the process can try the most patient person.

Moreover, the best monitor will provide a poor quality image if it is connected with a cheap video card or the wrong driver.

Video Card

Pixels and Other Jargon

As with most of the other components of your computer, the video subsystem has its own special jargon.  This jargon revolves around the art and science of images, and primarily focuses on pixels, resolution, and refresh rate.

Pixels

The image you see on your screen is made up of pixels (or small dots).  The pixels are arranged in columns and rows across the screen.  Typical screens range from a total of 480,000 to 1,920,000 pixels.

In a 480,000 pixel screen image there are 800 dots in each horizontal line, and 600 lines from the top to the bottom of the screen.  This results in a 800 x 600 image.   (800 x 600=480,000)

In most monitors each pixel consists of three sub-parts, each part representing one of the primary colors (red, green or blue).  By varying the intensity of these three sub-parts, the various other colors of the rainbow can be created.  The number of possible colors that a given pixel can display is referred to as color depth.  Color depth is directly correlated with the amount of video RAM in your video card - the greater the amount of RAM, the greater the possible color depth you can achieve.

Refresh Rate

The image on your screens is refreshed many times per second.  The rate at which it refreshes itself is referred to in Hertz (Hz), which means times per second.   Typical refresh rates are from 70 to 85 times per second.

One way to think of it is that the higher the refresh rate, the more stable and steady the screen image will appear.  Low refresh rates in the range of 60 Hz will seem to flicker in Windows.  Higher refresh rates of 70 Hz are acceptable, but may be taxing to your eyes after a few hours of steady use.  The current common wisdom is that for for extended work periods, a refresh rate of 80 or 80 Hz is desirable.

Of course, to have a high refresh rate you must have a good monitor.  You can try to boost the refresh rate of your monitor beyond its normal capability, but your resolution will then suffer.  For example, you can run a high quality Trinitron 17" monitor at 110 Hz at 800 x 600, but when you push the resolution to 1600 x 1200, you must settle for a refresh rate of 85 Hz: a significant difference.

Resolution

The greater the number of pixels in your screen image, the better its resolution, and the sharper the images on your display will appear.

The screen size of your monitor is closely related to the resolutions it can accommodate and the number of pixels that it holds:

Screen Size Pixels Resolution
15", 16" 480,000 800 x 600
17", 19" 786,432 1024 x 728
17", 19", 21" 995,328 1152 x 864
19", 21" 1,310,720 12808 x 1024

The easiest way to think of it is that the greater the resolution, the clearer and more detailed the image on the screen.  Someone who just runs DOS would not need a resolution higher than 800 x 600, but a person who runs modern games would need something much larger.

Moreover, if you like to fill your desktop with icons, you can get up to three times as many at 1280 x 1024 resolution than at 800 x 600.

If you desire both high resolution and high refresh rate, you will need a good monitor and a good video card.  There is no cutting corners here.

The Video Card

As stated earlier, your computer's video card (sometimes referred to as the graphics card) is one of the three main components of your its video subsystem.

In many computers the video card is an expansion board that can be replaced.  In other systems it is built right onto your motherboard (as is the case in laptops).  For your desktop system it may be preferable to make sure that it is an adapter, and thus is upgradable, as opposed to an integrated video card.

You can think of your video card as similar to your CPU in that it is a processor that is designed to control the images you see on your monitor's display.  In fact, your video card contains a processor chip (made perhaps by ATI, Matrox, or Cirrus Logic, for example), which creates the electric signals that are received by the screen in order for it to form an image.

Your video card also contains RAM, which it uses to "remember" the display's entire screen image.  As we've seen, the amount of RAM is directly related to the quality of the image you see on your screen (for example, the greater the RAM the greater the color depth and the finer the resolution).

The Video Driver

The difference between a terrific video card and an average one is more than just physical - it has to do with the software, too.

A good video card will have a driver that allows it to perform optimally, thus providing you with good screen performance.  For example, the Matrox driver will determine for you what the ideal refresh rate is for your monitor at a given resolution.  Another good sign of a quality video driver is that it will allow you to adjust the refresh rate.

Making It All Work

The video driver is what makes your monitor and video card work with the operating system.  Often times if you allow Windows to install the default drivers for your video card and monitor, the result will be less than optimal.

It is therefore advisable to make sure that you have the current drivers for both your monitor and video card and that you install them both.  This will insure that you get the optimal performance out of your display.  Be warned that the drivers that ship with the hardware may be out of date.  It is always wise to check the manufacturer's web site for the latest drivers.

Installing both monitor and video card drivers is done by using the Advanced button in the Settings tab of the Control Panel's Display Properties.  In the Adapter Properties you can change the video driver and the video card's refresh rate.  You can update the monitor driver in the adjacent Monitor tab.

In the Settings tab of the Display Properties window you can set your color depth and resolution.

=======================

Troubleshooting Video Problems

Most problems with video the subsystem are related either to setup issues with the video card or to image quality issues associated with the monitor.

Always make sure that you have the correct drivers for your video card and monitor before doing any troubleshooting.

Warning: Never open up a monitor!  There are high voltages present inside the case that can injure or kill you.  Only a trained professional should work on a monitor.

If your computer is refusing to boot and you suspect that the video card is the problem:

Make sure the video card is tightly inserted into its slot.  Try removing it and reinserting and then rebooting.

If there are chips on the video card, try pressing down on them gently (supporting the card from the back with your fingers) to make sure they are seated snuggly.

Check for resource conflicts.  PCI video cards will often take an IRQ channel via the PCI slot they are using.

If you have an onboard (integrated) video card on your motherboard that you have replaced with a new video adapter board, you must go into the CMOS BIOS and disable the integrated video card.

Try swapping the suspect card with another identical card, if possible.  If the problem is resolved, your video card will need to be replaced.

Try placing the suspect video card in another system.  If it doesn't work, the card is likely bad.  If it works, it is likely that there is a resource conflict in the original system.

If you hear two or more beeps at start-up, but there is no video:

Open your computer and make sure your video board is seated properly it its expansion slot.

If your monitor appears to have no power (the power light does not turn on):

Make sure the electrical outlet is working properly by plugging another appliance into it.

If the monitor is plugged into a powerstrip, remove the powerstrip and plug the monitor directly into the outlet.

Try another outlet.

If none of these options works, you likely have a bad monitor.  You will need to decide whether to replace it or repair it.

If your monitor does appear to have power (for example the power-on light is on) but it is not working properly:

Check that the brightness and contrast controls are not turned down too low.

Check any other controls that you have with your monitor; make sure you go over each one to be certain one of them isn't set incorrectly or is not conflicting with another control.

Try a different monitor with your computer.

Check that your video card is seated properly and that the monitor's cables are securely attached.  Make sure the cable isn't crimped or damaged.

If none of these options works, you likely have a bad monitor.

You get a "no signal" error on your monitor, indicating there is a problem with the signal connection between the video card and the monitor:

Make sure your monitor is plugged in and turned on.

Make sure the PC is on and running.

Check that your video card is seated properly and that the monitor's cables are securely attached.  Make sure the cable isn't crimped or damaged.

Try the monitor with a different computer, if possible.  If it works you may have a video card problem.

If your screen goes blank after a period of inactivity:

The most common cause of this behavior is that power management features have been enabled in the system BIOS.  Go into your CMOS BIOS and look for Power Management Settings and disable them.

The video output is a bunch of garbage on your monitor's display, you see wavy lines or other junk:

Make sure you have the correct and most up-to-date driver for your video card.

Your resolution or refresh rate may be set too high for the video driver.  Boot into Safe Mode by hitting F8 at start-up and see if the video is improved with the standard driver's lower resolution and refresh rate settings.

Check any controls that you have with your monitor; make sure you go over each one to be certain one of them isn't set incorrectly or is not conflicting with another control.   Pay particular attention to position and size adjustments.

Make sure the video card is tightly inserted into its slot.  Try removing it and reinserting and then rebooting.

Make sure the monitor's cables are securely attached and free from any crimps or damage.