General Computer Repair Procedures

With the hand tools and utilities described in the preceding sections, you have everything you need to upgrade or repair a PC except for the new components. Before you start, take a few minutes to read through the following sections, which describe the common procedures and general knowledge you need to work on PCs. These sections describe the common tasks involved in working on a PC things like opening the case, setting jumpers, manipulating cables, and adding or removing expansion cards. Instructions for specific tasks like replacing a motherboard, disk drive, or power supply are given in the relevant section.

Wallets Aren't Just for Money Anymore

The best way we've found to organize and protect CDs and DVDs is to lose the jewel cases and store them or, better still, copies of them in one of those zippered vinyl or Cordura disc wallets you can buy for a few dollars at Wal-Mart or Best Buy. These wallets use plastic or Tyvek sleeves to protect the discs, hold from half a dozen to several dozen discs, and make it easy to find the one you want. If the disc has a serial number or activation key on the original jewel case, make sure to record it on the CD, using a soft permanent marker on the label side. It's also a good idea to record the serial number or initialization (init) key on the disc sleeve or a small card so that the number is accessible when the disc is already in the drive.

We stock one of these wallets with essential discs Windows and Linux distribution CDs, applications, various diagnostics, and so on and always have it at hand. We also buy a disc wallet for each PC we buy or build. New PCs usually arrive with several discs, as do individual components. Storing these discs in one place, organized by the system they belong to, makes it much easier to locate the one you need.

Although you may be raring to get in there and fix something, taking the time to prepare properly before you jump in pays big dividends later. When your system has problems, do the following before you open the case:

Weird things can happen with cables. Disconnect all unessential cables, leaving only the mouse, keyboard, and display attached. Unplug the printer, USB hub, and any other attached peripherals to give them a chance to reset themselves. Turn your computer off, then restart it. If the problem is gone, try reattaching the cables one at a time to see if it comes back.

The old saying that "if all you have is a hammer, everything looks like a nail" is nowhere more true than with PC repairs. Before you assume that it's a hardware problem, make sure that the problem isn't caused by an application, by Windows, or by a virus. Use Knoppix and your virus/malware scanners before you assume the hardware is at fault and start disconnecting things. If the system boots and runs Knoppix successfully, defective hardware is very unlikely to be the problem.

The reliability of electrical power varies by where you live, which individual circuit you are connected to, and even from moment to moment as other loads on the circuit kick in and out. Sporadic problems such as spontaneous reboots are often caused by poor-quality power. Before you start tearing your system down, make sure the problem isn't caused by bad electrical power. At a minimum, use a surge protector to smooth incoming power. Better still, connect the system to a UPS (Uninterruptible Power Supply). If you don't have a UPS, connect the system to a power receptacle on a different circuit.

Modern systems particularly 'high-performance models run very hot. Sporadic problems, or those that occur only after a system has been running for some time, are often caused by excessive heat. Most modern motherboards include built-in temperature sensors generally one embedded in the processor socket to report CPU temperature and one or more others near the memory, chipset, and other critical components.

Most motherboard manufacturers supply utility programs that report and log temperature readings, as well as such other critical information as the speeds of the CPU and other system fans, the voltages on specific voltage rails, and so on. If no such utility is available for your operating system, simply reboot the computer, run BIOS Setup, and navigate the Setup menus until you find the option for Hardware Monitoring or something similar. Because the built-in temperature, voltage, and fan-speed sensors report their readings to the BIOS, you can read and record those values directly from the BIOS Setup screen. It's best to reboot and take the reading after the computer's been up and running for a while, and preferably just after it's exhibited the problems you are trying to resolve.

It's useful to establish baseline values for temperature readings, because "normal" temperatures vary significantly depending on the type and speed of the processor, the type of heatsink/fan unit used, the number and type of supplementary case fans, ambient temperature, degree of system load, and so on. For example, a processor that normally idles at 35 C may reach 60 C or higher when it runs a CPU-intensive program. The idle and loaded temperatures are both important. An increase in idle temperature probably indicates a cooling problem, such as clogged air inlets or a failing CPU fan, while very high loaded temperatures may result in system errors, processor slow-downs due to "thermal clamping," or, in the worst case, actual damage to the processor.

MONITOR YOUR MOTHERBOARD

To protect your system against thermal problems, we recommend installing and activating the monitoring utility supplied with the motherboard. Most such utilities allow you to set user-defined "tripwire" values that produce an alarm if the temperature becomes too high, the voltages are out of tolerance, or the fans are running too slowly. Most of these utilities can also shut down the system to prevent damage if the readings exceed the limits you've set. To determine the proper range of settings, refer to the documentation included with your system, motherboard, or processor.

Inexperienced technicians dive in willy-nilly without thinking things through first. Experienced ones first decide what is the most likely cause of the problem, what can be done to resolve it, in what order they should approach the repair, and what they'll need to complete it. Medical students have a saying, "when you hear thundering hooves, don't think about zebras." In other words, most of the time it'll be horses, and you can waste a lot of time looking for nonexistent zebras. Determine the most likely causes of the problem in approximate ranked order, decide which are easy to check for, and then eliminate the easy ones first. In order, check easy/likely, easy/unlikely, hard/likely, and finally hard/unlikely. Otherwise, you may find yourself tearing down your PC and removing the video card before you notice that someone unplugged the monitor.

We'll say it again: before you start upgrading or repairing a system, back up the important data on its hard drive. Every time you pop the cover of a PC, there's a small but ever-present risk that something that used to work won't work when you put everything together again. One of the wires in a cable may be hanging by a thread, or the hard drive may be teetering on the edge of failure. Just opening the case may cause a marginal component to fail irreversibly. So, before you even think of doing PC surgery, make sure that the hard drive is backed up.

It may seem obvious, but you need to disconnect all external cables before you can move the PC itself to the operating room. Many PCs are under desks or somewhere that otherwise makes it difficult to see the rear panel. If necessary, get down on the floor and crawl behind the PC with a flashlight to make sure it isn't still tethered to something. We've dragged modems, keyboards, and mice off desks because we weren't paying attention, and we once came within inches of pulling a $2,000 monitor onto the floor. Check the cables or pay the price.

CRT displays are not only fragile, but can cause serious injuries if the tube implodes. Flat-panel LCD displays aren't dangerous in that respect, but it's easy to do a lot of expensive damage very quickly if you don't take care. A display on the floor is an accident waiting to happen. If you're not moving the display to the work area, keep it on the desk out of harm's way. If you must put it on the floor, at least turn the screen toward the wall.

You can eliminate most of the risk of damaging components by static electricity simply by making it a habit to touch the case chassis or power supply to ground yourself before touching the processor, memory modules, or other static-sensitive components. It's also a good idea to avoid rubber-soled shoes and synthetic clothing and to work in an uncarpeted area.

MISTER SAFETY

If the air is particularly dry, use one of those spray/mister bottles that you can buy at any hardware store or supermarket. Fill it with water and add a few drops of dishwashing liquid or fabric softener. Before you begin work, mist the work area liberally, both air and surfaces. The goal isn't to get anything wet. Just the added humidity is enough to all but eliminate static electricity.

It sounds stupid, but it's not always immediately obvious how to get the cover off the chassis. We've worked on hundreds of different PCs from scores of manufacturers over the years, and we're still sometimes stumped. Manufacturers use an endless variety of fiendish ways to secure the cover to the chassis. Some were intended to allow tool-free access, others to prevent novice users from opening the case, and still others were apparently designed just to prove that there was yet one more way to do it.

We've seen novice upgraders throw up their hands in despair, figuring that if they couldn't even get the case open they weren't destined to become PC technicians. Nothing could be further from the truth. It just sometimes takes a while to figure it out.

The most evil example we ever encountered was a mini-tower case that had no screws visible except those that secured the power supply. The cover appeared seamless and monolithic. The only clue was a two-inch long piece of silver "warranty void if removed" tape that wrapped from the top of the cover to one side, making it clear that the separation point was there. We tried everything we could think of to get that cover off. We pulled gently on the front of the case, thinking that perhaps it would pop off and reveal screws underneath. We pressed in gently on the side panels, thinking that perhaps they were secured by a spring latch or friction fit. Nothing worked.

Finally, we turned the thing upside down and examined the bottom. The bottom of computer cases is almost always unfinished metal, but this one was finished beige material that looked just like the other parts of the cover. That seemed odd, so we examined the four rubber feet closely. They had what appeared to be center inserts, so we pried gently on one of these with our small screwdriver. Sure enough, it popped off and revealed a concealed screw within the rubber foot. Once we removed those four screws, the cover slid off easily, bottom first.

The moral is that what one person can assemble, another person can disassemble. It sometimes just takes determination, so keep trying. Your first resort should be the manual or, lacking that, the web site of the system or case manufacturer. Fortunately, most cases don't use such convoluted methods, so opening the case is usually straightforward.

Oddball Cables

Rather than using pins and holes, the connectors used on some cables for example, modular telephone cables and 10/100/1000BaseT Ethernet cables use other methods to establish the connection. The connector that terminates a cable may mate with a connector on the end of another cable, or it may mate with a connector that is permanently affixed to a device, such as a hard disk or a circuit board. Such a permanently affixed connector is called a socket, and may be male or female.

When you pop the cover of a PC, the first thing you'll notice is cables all over the place. These cables carry power and signals between various subsystems and components of the PC. Making sure they're routed and connected properly is no small part of working on PCs.

The cables used in PCs terminate in a variety of connectors. By convention, every connector is considered either male or female. Many male connectors, also called plugs or headers, have protruding pins, each of which maps to an individual wire in the cable. The corresponding female connector, also called a jack, has holes that match the pins on the mating male connector. Matching male and female connectors are joined to form the connection.

Some cables use unsheathed wires joined to a connector. Three cables of this sort are common in PCs those used to supply power to the motherboard and drives; those that connect front-panel LEDs, switches, and (sometimes) USB, FireWire, and audio ports to the motherboard; and those that connect audio-out on an optical drive to a sound card or motherboard audio connector. Figure 2-5 shows the front-panel power LED cable already connected to the motherboard, and the female jack of the front-panel reset switch cable being seated against the male motherboard header-pin connector for that cable.

Figure 2-5: Typical unsheathed cables

Some PC cables contain many individual wires packaged as a ribbon cable, so called because individually insulated conductors are arranged side-by-side in a flat array that resembles a ribbon. Ribbon cables provide a way to organize the wires required to connect devices like drives and controllers, whose interfaces require many conductors. Ribbon cables are used primarily for low-voltage signals, although they are also used to conduct low voltage/low current power in some applications. Ribbon cables are normally used only inside the case, because their electrical characteristics cause them to generate considerable RF emissions, which can interfere with nearby electronic components.

Square Peg, Round Hole

System designers attempt to avoid two potential dangers with regard to PC cables. Most important is to prevent connecting a cable to the wrong device. For example, connecting 12-volt power to a device that expects only 5 volts might have a catastrophic result. This goal is achieved by using unique connectors that physically prevent the cable from connecting to a device not designed to receive it. The second potential error is connecting a cable upside-down or backward. Most PC cables prevent this by using unsymmetrical connectors that physically fit only if oriented correctly, a process called keying.

Two keying methods are commonly used for PC cables, either individually or in conjunction. The first uses mating connectors whose bodies connect only one way, and is used for all power cables and some ribbon cables. The second, used by some ribbon cables, blocks one or more holes on the female connector and leaves out the corresponding pin on the male connector. Such a ribbon cable can be installed only when oriented so that missing pins correspond to blocked holes.

Ideal PC cables use unambiguous keyed connectors. You can't connect these cables to the wrong thing because the connector only fits the right thing; you can't connect them backwards, because the connector only fits the right way. Fortunately, most of the dangerous cables in PCs the ones that could damage a component or the PC itself if they were misconnected are of this sort. Power cables for disk drives and ATX motherboards, for example, fit only the correct devices and cannot be connected backwards.

Some PC cables, on the other hand, require careful attention. Their connectors may physically fit a component that they're not intended to connect to, and/or they may not be keyed, which means you can easily connect them backwards if you're not paying attention. Connecting one of these cables wrong usually won't damage anything, but the system may not work properly, either. The cables that link front-panel switches and indicator LEDs to the motherboard are of this variety.

Figure 2-6 shows a 40-wire ATA ribbon cable connected to the secondary ATA interface on an ASUS K8N-E Deluxe motherboard. The 40 individual wires are visible as raised ridges in the ribbon cable assembly. ASUS has provided a pull tab on the motherboard end of the cable to make it easier to remove, and has labeled the pull tab to recommend using it with optical drives. (Hard drives use the 80-wire version of the cable, shown later in Figure 2-7.)

Figure 2-6: A 40-wire ATA cable connected to the secondary motherboard ATA interface

All ribbon cables appear similar. They're often light gray, although some newer motherboards targeted at gamers and other enthusiasts include cables that are black, a bright primary color, or rainbow-colored. All of them use a contrasting colored stripe to indicate pin 1 red on standard gray cables; white on the cable shown here; brown on rainbow cables. But there are the following differences among ribbon cables:

Two for One

With one exception, the number of wires in a cable matches the number of pins on the connector, or very nearly so. The exception is Ultra-ATA hard drive cables, which use 40-pin connectors with 80-wire cables. The "extra" 40 wires are ground wires that are placed between the signal wires to reduce interference. Although the physical connectors are identical, if you connect an Ultra- ATA hard drive with a 40-wire ATA cable drive, performance will be significantly slower than if you use the proper 80-wire cable.

Common ribbon cable connectors range from the 10-pin connectors on the cables that are often used to extend serial, USB, FireWire, and audio ports from the motherboard header-pin connector to the front or back panel, through 34-pin floppy drive connectors, 40-pin ATA (IDE) drive connectors, to 50-, 68-, and 80-pin SCSI connectors.

Some ribbon cables have only two connectors, one at either end. ATA cables, used to connect hard drives and optical drives, have three connectors, a motherboard connector at one end, a connector for the master drive at the other end, and a connector for the slave drive in the middle (but located nearer the master drive connector). SCSI cables, used in servers and high-end workstations, may have five or more drive connectors.

Some ATA drive cables, called cable-select or CS cables, cut one conductor between the two device connectors. That is, while all 40 signal wires connect to the drive connector in the middle of the cable, only 39 of those signal wires are routed to the drive connector on the end of the cable. This missing conductor allows the position of the device on the cable to determine whether that device functions as a master or slave device, without requiring jumpers to be set.

FLAT VERSUS ROUND

So-called "round" ribbon cables have recently become popular, particularly with makers who cater to gamers and other enthusiasts. A round ribbon cable is simply a standard cable that has been sliced longitudinally into smaller groups of wires. For example, a standard flat 40-wire IDE ribbon cable might be sliced into ten 4-wire segments, which are then bound with cable ties or otherwise secured into a more or less round package. The advantage to round ribbon cables is that they reduce clutter inside the case and improve air flow. The disadvantage is that doing this reduces signal integrity on the individual wires because signal-bearing wires are put into closer proximity than intended. We recommend you avoid round ribbon cables, and replace any you find in any of your systems with flat ribbon cables. Note, however, that some round cables, such as Serial ATA cables, are designed to be round, and need not be replaced.

All ribbon cables used in current and recent systems use a header-pin connector similar to the ones shown in Figures 2-6 and 2-7. (Very old systems those from the days of 5.25" floppy drives used another type of connector called a card-edge connector, but that connector has not been used in new systems for more than a decade.) Header-pin connectors are used on cables for hard drives, optical drives, tape drives, and similar components, as well as for connecting embedded motherboard ports to external front or rear panel jacks.

The female header-pin connector on the cable has two parallel rows of holes that mate to a matching array of pins on the male connector on the motherboard or peripheral. On all but the least-expensive drives and other peripherals, these pins are enclosed in a plastic socket designed to accept the female connector. On inexpensive motherboards and adapter cards, the male connector may be just a naked set of pins. Even high-quality motherboards and adapter cards often use naked pins for secondary connectors (like USB ports or feature connectors).

Figure 2-7 shows an Ultra-ATA hard drive cable compare the 80-wire cable shown here with the 40-wire cable shown in the preceding image and two ATA interfaces on the motherboard. This cable uses two keying methods. The raised tab visible at the top of the cable connector mates to the slot visible on the lower edge of the connector shroud of the blue primary ATA interface on the motherboard. The blocked hole in the lower row of holes on the cable connector matches the missing pin visible in the top row of pins on the motherboard connector. Although there are 80 conductors, there are still only 40 pins. The 80-conductor cables have a grounded wire running between each pair of signal wires, which reduces electrical crosstalk, thus permitting higher data rates with greater reliability.

Figure 2-7: An 80-wire Ultra-ATA cable and two motherboard interfaces, showing keying

Also note the keying arrangements for the black secondary ATA motherboard connector. Like the primary motherboard connector, the secondary connector is keyed with a missing pin. But the secondary connector lacks the cut-out slot present in the primary motherboard connector, which means that this cable cannot be inserted into the secondary connector. That's by design. Although the 80-wire cable would function properly with the secondary connector, ASUS has chosen to key this Ultra-ATA cable to ensure that it can be connected only to the primary motherboard ATA interface connector, which is typically used to connect a hard drive. The secondary motherboard ATA connector, which is usually used to connect an optical drive, requires a cable that doesn't have the keying tab, such as the one shown in Figure 2-6.

Some header-pin connectors, male and female, are not keyed. Others use connector body keying, pin/hole keying, or both. This diversity means that it is quite possible to find that you cannot use a particular header-pin cable for its intended purpose. For example, we once attempted to use the ATA cable supplied with a drive to connect that drive to the secondary ATA header pin connector on the motherboard. The motherboard end of that cable was keyed by a blocked hole, but the header-pin connector on the motherboard had all pins present, which prevented the cable from seating. Fortunately, the cable that came with the motherboard fit both the motherboard and the drive connectors properly, allowing us to complete the installation.

If you run into such a keying problem, there are four possible solutions:

The IDE and other header-pin cables that most computer stores sell use connectors that use neither connector body nor pin/hole keying. You can use one of these cables of the proper size to connect any device, but the absence of all keying means that you must be especially careful not to connect it backwards.

If you don't have an unkeyed cable available, you may be able to remove the key from the existing cable. Most keyed cables use a small bit of plastic to block one of the holes. You may be able to use a needle to pry the block out far enough that you can extract it with your needlenose pliers. Alternatively, try pushing a pin into the block at an angle, then bending the top of the pin over and pulling both bent pin and block out with your pliers. If the key is a solid, integral part of the cable (which is rarely the case), you may be able to use a heated needle or pin to melt the key out of the hole far enough for the pin to seat.

Heat a needle with a pair of pliers over a flame and carefully insert to a depth of 3/8" to bore open the offending plug.

Sometimes you have no choice. If the stores are closed, the only cable you have uses pin/hole keying with a solid block that you can't get out, and you must connect that cable to a header-pin connector that has all pins present, you have to go with what you have. You can use diagonal cutters to nip off the pin that prevents you from connecting the cable. Obviously, this is drastic. If you nip the wrong pin, you'll destroy the motherboard or expansion card, or at least render that interface unusable. Before you cut, see if you can swap cables within the PC to come up with an unkeyed cable for the problem connector. If not, you can sometimes bend the offending pin slightly enough to allow the female connector to partially seat. This may be good enough to use as a temporary connection until you can replace the cable. If all else fails and you need to cut the pin, before doing so, align the keyed female connector with the pin array and verify just which pin needs to be cut. Also, check the manual for a detailed list of signal/pin assignments on that interface. The pin you are about to remove should be labeled No Connection or N/C in that list. Use the old carpenter's maxim here measure twice and cut once.

Connector and keying issues aside, the most common mishap with header-pin connectors occurs when you install the cable offset by a column or a row. The shrouded male connectors used on most drives make this impossible to do, but the male connectors used on some cheap motherboards are an unshrouded double row of pins, making it very easy to install the connector with the pins and holes misaligned. Working in a dark PC, it's very easy to slide a connector onto a set of header pins and end up with an unconnected pair of pins at one end and an unconnected pair of holes at the other. It's just as easy to misalign the connector the other way, and end up with an entire row of pins and holes unconnected. One of our reviewers did this and fried a client's hard drive. If you need reading glasses, this isn't the time to find out the hard way.

Locating Pin 1

If you upgrade your system and it fails to boot or the new device doesn't work, chances are that you connected a ribbon cable backwards. This can't happen if all connectors and cables are keyed, but many systems have at least some unkeyed connectors. The good news is that connecting ribbon cables backwards almost never damages anything. We're tempted to say "never" without qualification, but there's a first time for everything. If your system doesn't boot after an upgrade, go back and verify the connections for each cable. Better yet, verify them before you restart the system.

To avoid connecting a ribbon cable backwards, locate pin 1 on each device and then make sure that pin 1 on one device connects to pin 1 on the other. This step is sometimes easier said than done. Nearly all ribbon cables use a colored stripe to indicate pin 1, so there's little chance of confusion there. However, not all devices label pin 1. Those that do usually use a silk-screened numeral 1 on the circuit board itself. If pin 1 is not labeled numerically, you can sometimes determine which is pin 1 in one of the following ways:

• Instead of a numeral, some manufacturers print a small arrow or triangle to indicate pin 1.
• The layout of some circuit boards allows no space for a label near pin 1. On these boards, the manufacturer may instead number the last pin. For example, rather than pin 1 being labeled on an ATA connector, pin 40 may be labeled on the other side of the connector.
• If there is no indication of pin 1 on the front of the board, turn it over (this is tough for an installed motherboard) and examine the reverse side. Some manufacturers use round solder connections for all pins other than 1, and a square solder connection for pin 1.
• If all else fails, you can make an educated guess. Many disk drives place pin 1 closest to the power supply connector. On a motherboard, pin 1 is often the one closest to the memory or processor. We freely admit that we use this method on occasion to avoid having to remove a disk drive or motherboard to locate pin 1 with certainty. We've never damaged a component using this quick-and-dirty method, but we use it only for ATA drives, rear-panel port connectors, and other cables that do not carry power. Don't try this with SCSI particularly differential SCSI.

Once you locate an unmarked or unclearly marked pin 1, use nail polish or some other permanent means to mark it so that you won't have to repeat the process the next time. Wite-Out is really handy for this. Make a single stripe across both cable connector and plug and you'll have a visual confirmation that they align correctly.

For many years, most PCs used only the types of cables we've already described. In 2003, motherboards and drives began shipping that used a new standard called Serial ATA (often abbreviated S-ATA or SATA). For clarity, old-style ATA drives are now sometimes called Parallel ATA (P-ATA or PATA), although the formal name of the older standard has not changed.

The obvious difference between ATA devices and SATA devices is that they use different cables and connectors for power and data. Rather than the familiar wide 40-pin data connector and large 4-pin Molex power connector used by ATA devices (shown in Figure 2-8), SATA uses a 7-pin thin, flat data connector and a similar 15-pin power connector (shown inFigure 2-9).

Figure 2-8: PATA data connector (left) and power connector

Figure 2-9: SATA power connector (left) and data connector

Missing Voltages

The SATA power cable shown in Figure 2-9 supplies only +5V on the red wire and +12V on the yellow wire, with two black ground wires. A fully compliant SATA power connector adds an orange +3.3V wire.

Perhaps coincidentally, the 15-pin SATA power connector is exactly the same width as the 4-pin Molex PATA power connector, although the SATA power connector is considerably thinner. At 8 mm wide, the 7-pin SATA data connector is much narrower than the 40-pin PATA data connector. This reduced overall width and thickness made SATA a natural for 2.5" notebook hard drives, which are becoming increasingly common in desktop systems as well.

FRAGILE CONNECTORS

Be very careful when you install or remove SATA data and power cables. The thinness of SATA connectors means they are fragile, although recent SATA connectors seem more robust than early models. Do not twist or torque the connector as you install or remove it. Install a connector by aligning the cable connector with the device connector and pressing straight inward until the connector seats. Remove a connector by pulling it straight outward, without putting any sideways force on it. Otherwise, you may snap off the connector.

The relatively large number of pins in the SATA power connector accommodates two SATA design goals. First, additional connectors are required to support hot-plugging installing or removing drives without turning off the system which is a part of the SATA standard. Second, SATA power connectors are designed to provide voltages of +3.3V, +5V, and +12V, rather than just the +5V and +12V provided by the PATA power connector. The lower +3.3V voltage is a forward-looking provision for smaller, quieter, cooler-running drives that will be introduced over the coming years.

Figure 2-10: A group of four SATA data connectors on a motherboard, showing the L-shaped keying

Although all PATA power connectors are keyed, the same cannot be said for PATA data connectors. One of the design goals of SATA was to use unambiguous keying. SATA uses L-shaped contact bodies, as shown in Figure 2-10, which prevent a cable from being installed upside-down or backward. (While there's no Pin 1 to worry about, you may find it handy to use a Wite-Out pen to label the UP position of the SATA cable and the connector, or to run a stripe across both.)

SATA differs from PATA in two other respects. First, PATA allows two devices to be connected to each interface, one jumpered as master and the other as slave. An SATA interface supports only one device, eliminating the need for configuring the device as master or slave. In effect, all SATA devices are master devices. Second, PATA limits the length of data cables to 18" (45.7 cm), while SATA allows data cables as long as 1 meter (39.4"). The thinness and additional length of SATA data cables makes it much easier to route and dress the cables in the case particularly in a full-tower case and contributes to improved air flow.

Expansion cards are circuit boards that you install in a PC to provide functions that the PC motherboard itself does not provide. Figure 2-11 shows an ATI All-In-Wonder 9800 Pro AGP graphics adapter and video capture card, a typical expansion card.

Figure 2-11: ATI All-In-Wonder 9800 Pro, a typical expansion card

Years ago, most PCs had several expansion cards installed. A typical vintage-2000 PC might have had a video card, a sound card, a LAN adapter, an internal modem, and perhaps a communications adapter of some sort or a SCSI host adapter. It wasn't uncommon for PCs back then to have all of their expansion slots filled.

Things are different nowadays. Nearly all recent motherboards include embedded audio and LAN adapters. Many include embedded video, and some include less common features such as embedded FireWire, modems, SCSI host adapters, and other devices. Because so many features are routinely incorporated in modern motherboards, it's not unusual for a relatively new PC to have no expansion cards installed at all.

Still, installing an expansion card is an easy, inexpensive way to upgrade an older system. You might, for example, install an AGP graphics card to upgrade the on-board video, a video capture card to turn your PC into a digital video recorder, an SATA controller to add support for SATA drives, a USB adapter to add more USB 2.0 ports, or an 802.11g card to add wireless networking.

Each expansion card plugs into an expansion slot located on the motherboard or on a riser card that attaches to the motherboard. The rear panel of the PC chassis includes a cutout for each expansion slot, which provides external access to the card. The cutouts for vacant expansion slots are covered by thin metal slot covers that are secured to the chassis. These covers prevent dust from entering through the cutout and also preserve the cooling air flow provided by the power supply fan and any auxiliary fans installed in the system.

Don't Leave Holes in Your Case

Cheap cases sometimes have slot covers that must be twisted off to be removed and are destroyed in the process. If you need to cover an open slot in such a case and don't have a spare slot cover, ask your local computer store, which probably has a stack of them in back. Or just use duct tape to cover the gap. (Put it on the outside of the case, where it won't goop up a slot that you may need to use later.) If you're worried about RF leakage, 3M makes some metal tapes that are conductive through the adhesive, but you'd have to put them inside the case, of course, to take advantage of that.

To install an expansion card, remove the slot cover, which may be secured by a small screw or may simply be die-stamped into the surrounding metal. In the latter case, carefully twist off the slot cover using a screwdriver or your needlenose pliers. (Be careful! The edges can be quite sharp.) If you need to replace the slot cover later, secure it to the chassis using a small screw that fits a notch in the top portion of the slot cover. The back of the expansion card forms a bracket that resembles a slot cover and is secured to the chassis in the same way. Depending on the purpose of the card, this bracket may contain connectors that allow you to connect external cables to the card.

There is frequently a need to install and remove expansion cards when you work on a PC. Even if you are not working on a particular expansion card, you must sometimes remove it to provide access to the section of the PC that you do need to work on. Installing and removing expansion cards can be hard or easy, depending on the quality of the case, the motherboard, and the expansion card itself. High-quality cases, motherboards, and expansion cards are built to tight tolerances, making expansion cards easy to insert and remove. Cheap cases, motherboards, and expansion cards have such loose tolerances that you must sometimes literally bend sheet metal to force them to fit.

People often ask whether it matters which card goes into which slot. Beyond the obvious there are different kinds of expansion slots, and a card can be installed only in a slot of the same type there are four considerations that determine the answer to this question:

Depending on the size of the card and the design of the motherboard and case, a given card may not physically fit a particular slot. For example, the case design may prevent a particular slot from accepting a full-length card. If this occurs, you may have to juggle expansion cards, moving a shorter card from a full-length slot to a short slot and then using the freed-up full-length slot for the new expansion card. Also, even if a card physically fits a particular slot, a connector protruding from that card may interfere with another card, or there may not be enough room to route a cable to it.

There are several variables, including slot type, card type, BIOS, and operating system, that determine whether a card is position sensitive.

For this reason, although it may not always be possible, it's good general practice to reinstall a card into the same slot that you removed it from. If you do install the card in a different slot, don't be surprised if Windows forces you to reinstall the drivers. If you're really lucky, you might even have the pleasure of going through Product Activation again.

Two's a Crowd

Rather than installing two expansion cards in adjacent slots, try to space them out as far as possible to improve air flow and cooling and to make any connectors or jumpers on the cards as accessible as possible.

ON THE GRIPPING HAND

Although interrupt conflicts are rare with PCI motherboards and modern operating systems, they can occur. In particular, PCI motherboards with more than four PCI slots share interrupts between slots, so installing two PCI cards that require the same resource in two PCI slots that share that interrupt may cause a conflict. If that occurs, you can eliminate the conflict by relocating one of the conflicting expansion cards to another slot. Even in a system with all PC slots occupied, we have frequently eliminated a conflict just by swapping the cards around. See your motherboard manual for details.

Although it is relatively uncommon nowadays, some combinations of motherboard and power supply can provide adequate power for power-hungry expansion cards like internal modems only if those cards are installed in the slots nearest the power supply. This was a common problem years ago, when power supplies were less robust and cards required more power than they do now, but you are unlikely to experience this problem with modern equipment. One exception to this is AGP video cards. Many recent motherboards support only AGP 2.0 1.5V video cards and/or AGP 3.0 0.8V video cards, which means that old 3.3V AGP cards are incompatible with that slot.

Another problem that is much less common with recent equipment is that some expansion cards generate enough RF to interfere with cards in adjacent slots. Years ago, the manuals for some cards (notably some disk controllers, modems, and network adapters) described this problem, and suggested that their card be installed as far as possible from other cards. We haven't seen this sort of warning on a new card in years, but you may still encounter it if your system includes older cards.

Figure 2-12: Five white PCI slots and a dark brown AGP slot

Figure 2-13: Two white PCI slots, two PCI Express X1 slots, two more white PCI slots, and a black PCI Express X16 video card slot

Figure 2-14: Seat the expansion card by pressing down evenly

To install an expansion card, proceed as follows:

1. Read the instructions that come with the card. In particular, read carefully any instructions about installing software drivers for the card. For some cards, you must install the driver before you install the card; for other cards, you must install the card first and then the driver.
2. Remove the cover from the chassis and examine the motherboard to determine which expansion slots are free. Locate a free expansion slot of the type required by the expansion card. Recent PCs may have several types of expansion slots available, including 32- and 64-bit PCI general-purpose expansion slots, an AGP video card slot, one or two PCI Express x16 video card slots, and one or more PCI Express x1 feature slots. If more than one slot of the proper type is free, you can reduce the likelihood of heat-related problems by choosing one that maintains spacing between the expansion cards rather than one that clusters the cards. Figure 2-12 shows a standard arrangement of slots for an AGP motherboard, with five white 32-bit PCI slots at the upper left and one dark brown AGP slot below and to the right of the PCI slots. Figure 2-13 shows a standard arrangement of slots for a PCI Express motherboard, with, from left to right, two white 32-bit PCI slots; two short, black PCI Express X1 slots; two more white PCI slots; and one long, black PCI Express X16 slot for a video adapter.'
3. An access hole for each expansion slot is present on the rear of the chassis. For unoccupied slots, this hole is blocked by a thin metal slot cover secured by a screw that threads downward into the chassis. Determine which slot cover corresponds to the slot you chose. This may not be as easy as it sounds. Some types of expansion slots are offset, and the slot cover that appears to line up with that slot may not be the right one. You can verify which slot cover corresponds to a slot by aligning the expansion card itself with the slot and seeing which slot cover the card bracket matches to.
4. Remove the screw that secures the slot cover, slide the slot cover out, and place it and the screw aside.
5. If an internal cable blocks access to the slot, gently move it aside or disconnect it temporarily, noting the proper connections so that you will know where to reconnect it.
6. Guide the expansion card gently into position, but do not yet seat it. Verify visually that the tongue on the bottom of the expansion card bracket will slide into the matching gap in the chassis and that the expansion card bus connector section aligns properly with the expansion slot. With a high-quality case, everything should align properly with no effort. With a cheap case, you may have to use pliers to bend the card bracket slightly to make the card, chassis, and slot all line up. Rather than doing that, we prefer to replace the case.'
7. When you are sure that everything is properly aligned, position your thumbs on the top edge of the card, with one thumb at each end of the expansion slot below the card, and press gently straight down on the top of the card until it seats in the slot, as shown in Figure 2-14. Apply pressure centered on the expansion slot beneath the card, and avoid twisting or torquing the card. Some cards seat easily with little tactile feedback. Others require quite a bit of pressure and you can feel them snap into place. Once you complete this step, the expansion card bracket should align properly with the screw hole in the chassis.
8. Replace the screw that secures the expansion card bracket, and replace any cables that you temporarily disconnected while installing the card. Connect any external cables required by the new card don't tighten the thumbscrews quite yet and give the system a quick once-over to make sure you haven't forgotten to do anything.
9. Turn on the PC and verify that the new card is recognized and that it functions as expected. Once you have done so, power the system down, replace the cover, and reconnect everything. Store the unused slot cover with your spares.

To remove an expansion card, proceed as follows:

1. Remove the system cover and locate the expansion card to be removed. It's surprising how easy it is to remove the wrong card if you're not careful. No wonder surgeons occasionally get it wrong.
2. Once you're sure you've located the right card, disconnect any external cables connected to it. If the card has internal cables connected, disconnect those as well. You may also need to disconnect or reroute other unrelated cables temporarily to gain access to the card. If so, label those you disconnect.
3. Remove the screw that secures the card bracket, and place it safely aside.
4. Grasp the card firmly near both ends and pull straight up with moderate force. If the card will not release, gently rock it from front to back (parallel to the slot connector) to break the connection. Be careful when grasping the card. Some cards have sharp solder points that can cut you badly if you don't take precautions. If there's no safe place to grasp the card and you don't have a pair of heavy gloves handy, try using heavy corrugated cardboard between the card and your skin.
5. If you plan to save the card, place it in an antistatic bag for storage. It's a good idea to label the bag with the date and the make and model of the card for future reference. If you have a driver disk, throw that in the bag as well. If you are not installing a new expansion card in the vacated slot, install a slot cover to ensure proper air flow and replace the screw that secures the slot cover.

Danger, Will Robinson!

You may someday encounter an expansion card that's seated so tightly that it appears to be welded to the motherboard. When this happens, it's tempting to gain some leverage by pressing upwards with your thumb on a connector on the back of the card bracket. Don't do it. The edges of the chassis against which the bracket seats may be razor sharp, and you may cut yourself badly when the card finally gives. Instead, loop two pieces of cord around the card to the front and rear of the slot itself, and use them to "walk" the card out of its slot, as shown in Figure 2-15. Your shoelaces will work if nothing else is at hand. For a card that's well and truly stuck, you may need a second pair of hands to apply downward pressure on the motherboard itself to prevent it from flexing too much and possibly cracking as you pull the card from the slot.

Figure 2-15: Barbara pulls a recalcitrant expansion card the safe way

If you're removing an AGP or PCI Express video card, take particular care. Many motherboards include a video card retention mechanism, shown in Figure 2-16, that physically latches the card into place. When you remove a video card, release the latch and pull gently upward on the card until it comes free. If you attempt to force it, you could damage the video card and/or the motherboard.

Figure 2-16: The AGP retention bracket physically locks an AGP card into the slot

Jumpers are sometimes used to set hardware options on PCs and peripherals. Jumpers allow you to make or break a single electrical connection, which is used to configure one aspect of a component. Jumper or switch settings specify such things as the front-side bus speed of the processor, whether a PATA drive functions as a master or slave device, whether a particular function on an expansion card is enabled or disabled, and so on.

Older motherboards and expansion cards may use dozens of jumpers to set most or all configuration options. Recent motherboards use fewer jumpers, and instead use the BIOS setup program to configure components. In fact, most current motherboards have only one or a few jumpers. You use these jumpers when you install the motherboard to configure static options such as processor speed or to enable infrequent actions such as updating the BIOS.

More properly called a jumper block, a jumper is a small plastic block with embedded metal contacts that can bridge two pins to form an electrical connection. When a jumper block bridges two pins, that connection is called on, closed, shorted, or enabled. When the jumper block is removed, that connection is called off, open, or disabled. The pins themselves are also called a jumper, usually abbreviated JPx, where x is a number that identifies the jumper.

Jumpers with more than two pins may be used to select among more than two states. One common arrangement, shown in Figure 2-17, is a jumper that contains a row of three pins, numbered 1, 2, and 3. You can select among three states by shorting pins 1 and 2, pins 2 and 3, or by removing the jumper block entirely. Note that you cannot jumper pins 1 and 3 because a jumper can be used to close only an adjacent pair of pins. In this example, the USBPW12 and USBPW34 jumpers allow you to set the Wake-on-USB configuration for the four USB ports numbered 1 through 4. These jumpers are shown shorting pins 1 and 2, which configures the motherboard to use +5V for Wake-on-USB. If we moved those jumpers to the 2 3 position, Wake-on-USB would use +5Vsb.

Figure 2-17: Two jumpers shorting the 1 2 pins of 3-pin jumper blocks

You can often use your fingers to install and remove isolated jumpers, but needlenose pliers are usually the best tool. However, jumpers are sometimes clustered so tightly that even needlenose pliers may be too large to grab just the jumper you want to work on. When this happens, use a hemostat or mosquito forceps (available from any drugstore). When you need to set a jumper open, don't remove the jumper block entirely. Instead, install it on just one pin. This leaves the connection open, but ensures that a jumper block will be handy if you later need to close that connection.

Jumper blocks come in at least two sizes that are not interchangeable:

• Standard blocks are the larger and the more commonly used size, and are often dark blue or black. (The jumpers shown in Figure 2-17 are the standard size.)
• Mini jumper blocks are used on some disk drives and boards that use surface-mount components, and are often white or light blue.

New components always come with enough jumper blocks to configure them. If you remove one when configuring a device, tape it to a convenient flat area on the device for possible future use. It's also a good idea to keep a few spares on hand, just in case you need to reconfigure a component from which someone has removed all the "surplus" jumper blocks. Any time you discard a board or disk drive, strip the jumper blocks from it first and store them in your parts tube. (If you don't have an official parts tube, do what we do: use an old aspirin bottle with a snap-on lid.)

We planned to write an overview section here to describe how to install and configure drives. Unfortunately, we found it impossible to condense that information to an overview level. Physical installation procedures vary significantly, and configuration procedures even more, depending on numerous factors, including:

• Drive type
• Physical drive size: both height and width, and (sometimes) depth
• Internal (hard drives) versus externally accessible (floppy, optical, and tape drives)
• Mounting arrangements provided by the particular case
• Drive interface (ATA versus Serial ATA)

For specific information about installing and configuring various drive types, including illustrations and examples, refer to the section that covers that type of device, be it Hard Drives, Optical Drives or External Storage Devices.

More about Working on Computers