'; window.popUpWin.document.write(zhtm); window.popUpWin.document.close(); // Johnny Jackson 4/28/98 } //--> Upgrading & Repairing PCs Eighth Edition -- Ch 3 -- System Teardown and Inspection


Upgrading & Repairing PCs Eighth Edition

Previous chapterNext chapterContents

Chapter 3

System Teardown and Inspection

This chapter examines procedures for tearing down and inspecting a system. The chapter describes the types of tools required, the procedure for disassembling the system, and the various components that make up the system. A special section discusses some of the test equipment you can use when troubleshooting a system; another section covers some problems you may encounter with the hardware (screws, nuts, bolts, and so on).

Using the Proper Tools

To troubleshoot and repair PC systems properly, you need a few basic tools. If you intend to troubleshoot and repair PCs professionally, there are many more specialized tools you will want to purchase. These advanced tools allow you to more accurately diagnose problems and make the jobs easier and faster. The basic tools that should be in every troubleshooter's toolbox are:

Some environments may also have the resources to purchase the following, although it's not required for most work:

In addition, an experienced troubleshooter will probably want to have soldering and desoldering tools to fix bad serial cables. These tools are discussed in more detail in the following section. Diagnostics software and hardware are discussed in Chapter 21, "Software and Hardware Diagnostic Tools."

Hand Tools

When you work with PC systems, it immediately becomes apparent that the tools required for nearly all service operations are very simple and inexpensive. You can carry most of the required tools in a small pouch. Even a top-of-the-line "master mechanics" set fits inside a briefcase-size container. The cost of these tool kits ranges from about $20 for a small service kit to $500 for one of the briefcase-size deluxe kits. Compare these costs with what might be necessary for an automotive technician. Most automotive service techs spend $5,000 to $10,000 or more for the tools they need. Not only are PC tools much less expensive, but I can tell you from experience that you don't get nearly as dirty working on computers as you do working on cars.

In this section, you learn about the tools required to make a kit that is capable of performing basic, board-level service on PC systems. One of the best ways to start such a set of tools is a small kit sold especially for servicing PCs.

The following list shows the basic tools that you can find in one of the small PC tool kits that sell for about $20:

NOTE: Some tools aren't recommended because they are of limited use. However, they normally come with these types of kits.

You use nut drivers to remove the hexagonal-headed screws that secure the system-unit covers, adapter boards, disk drives, power supplies, and speakers in most systems. The nut drivers work much better than conventional screwdrivers.

Because some manufacturers have substituted slotted or Phillips-head screws for the more standard hexagonal-head screws, standard screwdrivers can be used for those systems.

You use the chip-extraction and insertion tools to install or remove memory chips (or other smaller chips) without bending any pins on the chip. Usually, you pry out larger chips, such as some microprocessors or ROMs, with the small screwdriver. Larger processors such as the 486, Pentium, or Pentium Pro chips require a chip extractor if they are in a standard LIF (Low Insertion Force) socket. These chips have so many pins on them that a large amount of force is required to remove them, despite the fact that they call the socket "low insertion force." If you use a screwdriver on a large physical-size chip like a 486 or Pentium, you risk cracking the case of the chip and permanently damaging it. The chip extractor tool for removing these chips has a very wide end with tines that fit between the pins on the chip to distribute the force evenly along the chip's underside. This will minimize the likelihood of breakage. Most of these types of extraction tools must be purchased specially for the chip you're trying to remove.

Fortunately, motherboard designers have seen fit to use mostly ZIF (Zero Insertion Force) sockets on systems with 486 and larger processors. The ZIF socket has a lever which when raised releases the grip on the pins of the chip, allowing it to be easily lifted out with your fingers.

The tweezers and parts grabber can be used to hold any small screws or jumper blocks that are difficult to hold in your hand. The parts grabber is especially useful when you drop a small part into the interior of a system; usually, you can remove the part without completely disassembling the system.

Finally, the Torx driver is a special, star-shaped driver that matches the special screws found in most Compaq systems and in many other systems as well.

Although this basic set is useful, you should supplement it with some other small hand tools, such as:

Pliers are useful for straightening pins on chips, applying or removing jumpers, crimping cables, or grabbing small parts.

Hemostats are especially useful for grabbing small components, such as jumpers.

The wire cutter or stripper, obviously, is useful for making or repairing cables or wiring.

The metric nut drivers can be used in many clone or compatible systems as well as in the IBM PS/2 systems, all of which use metric hardware.

The tamperproof Torx drivers can be used to remove Torx screws with the tamper- resistant pin in the center of the screw. A tamperproof Torx driver has a hole drilled in it to allow clearance for the pin.

You can use a vise to install connectors on cables and to crimp cables to the shape you want, as well as to hold parts during delicate operations. In addition to the vise, Radio Shack sells a nifty "extra hands" device which has two movable arms with alligator clips on the end. This type of device is very useful for making cables or for other delicate operations where an extra set of hands to hold something might be useful.

You can use the file to smooth rough metal edges on cases and chassis, as well as to trim the faceplates on disk drives for a perfect fit.

The flashlight can be used to illuminate system interiors, especially when the system is cramped and the room lighting is not good. I consider this tool to be essential.

Another consideration for your tool kit is an ESD (electrostatic discharge) protection kit. This kit consists of a wrist strap with a ground wire and a specially conductive mat, also with its own ground wire. Using a kit like this when working on a system will help to ensure that you never accidentally zap any of the components with a static discharge.

NOTE: You can work without an ESD protection kit, if you're disciplined and careful about working on systems. If you don't have an ESD kit available, you should leave the computer plugged in, so that the power cord connects the chassis of the PC to ground. Then make sure that you remain in constant or nearly constant contact with the case. It's easy to rest an arm or elbow on some part of the case while working inside the computer.

The ESD kits, as well as all the other tools and much more, are available from a variety of tool vendors. Specialized Products Company and Jensen Tools are two of the most popular vendors of computer and electronic tools and of service equipment. Their catalogs show an extensive selection of very high-quality tools. (These companies and several others are listed in Appendix A, "Vendor List.") With a simple set of hand tools, you will be equipped for nearly every PC repair or installation situation. The total cost of these tools should be less than $150, which is not much considering the capabilities they give you.

Soldering and Desoldering Tools

In certain situations--such as repairing a broken wire, making cables, reattaching a component to a circuit board, removing and installing chips that are not in a socket, or adding jumper wires or pins to a board--you must use a soldering iron to make the repair.

Although virtually all repairs these days are done by simply replacing the entire failed board, you may need a soldering iron in some situations. The most common case would be where there was physical damage to a system, such as where somebody had ripped the keyboard connector off of a motherboard by pulling on the cable improperly. Simple soldering skills could save the motherboard in this case.

Most motherboards these days include I/O components such as serial and parallel ports. Many of these ports are fuse-protected on the board; however, the fuse is usually a small soldered-in component. These fuses are designed to protect the motherboard circuits from damage from an external source. If a short circuit or static charge from an external device blows these fuses, the board can be saved if you can replace them.

To perform minor repairs such as these, you need a low-wattage soldering iron--usually about 25 watts. More than 30 watts generates too much heat and can damage the components on the board. Even with a low-wattage unit, you must limit the amount of heat to which you subject the board and its components. You can do this with quick and efficient use of the soldering iron, as well as with the use of heat-sinking devices clipped to the leads of the device being soldered. A heat sink is a small metal clip-on device designed to absorb excessive heat before it reaches the component that the heat sink is protecting. In some cases, you can use a pair of hemostats as an effective heat sink when you solder a component.

To remove components that originally were soldered into place from a printed circuit board, you can use a soldering iron with a solder sucker. This device normally is constructed as a small tube with an air chamber and a plunger-and-spring arrangement. (I do not recommend the squeeze-bulb type of solder sucker.) The unit is "cocked" when you press the spring-loaded plunger into the air chamber. When you want to remove a device from a board, you use the soldering iron from the underside of the board, and heat the point at which one of the component leads joins the circuit board until the solder melts. As soon as melting occurs, move the solder-sucker nozzle into position, and press the actuator. This procedure allows the plunger to retract and creates a momentary suction that inhales the liquid solder from the connection and leaves the component lead dry in the hole.

Always do the heating and suctioning from the underside of a board, not from the component side. Repeat this action for every component lead joined to the circuit board. When you master this technique, you can remove a small component in a minute or two with only a small likelihood of damage to the board or other components. Larger chips that have many pins can be more difficult to remove and resolder without damaging other components or the circuit board.

TIP: These procedures are intended for Through-Hole devices only. These are components whose pins extend all the way through holes in the board to the underside. Surface mount devices are removed with a completely different procedure, using much more expensive tools. Working on surface-mounted components is beyond the capabilities of all but the most well-equipped shops.

If you intend to add soldering and desoldering skills to your arsenal of capabilities, you should practice. Take a useless circuit board and practice removing various components from the board, then reinstall the components. Try to remove the components from the board by using the least amount of heat possible. Also, perform the solder-melting operations as quickly as possible, limiting the time that the iron is applied to the joint. Before you install any components, clean out the holes through which the leads must project and mount the component in place. Then apply the solder from the underside of the board, using as little heat and solder as possible.

Attempt to produce joints as clean as the joints that the board manufacturer performed by machine. Soldered joints that do not look clean may keep the component from making a good connection with the rest of the circuit. This "cold-solder joint" normally is created because you have not used enough heat. Remember that you should not practice your new soldering skills on the motherboard of a system that you are attempting to repair! Don't attempt to work on real boards until you are sure of your skills. I always keep a few junk boards around for soldering practice and experimentation.

TIP: When first learning to solder, you may be tempted to set the iron on the solder and leave it there until the solder melts. If the solder doesn't melt immediately when applying the iron to it, you're not transferring the heat from the iron to the solder efficiently. This means that either the iron is dirty, or there is debris between it and the solder. To clean the iron, take a wet sponge and drag it across the tip of the iron.

If after cleaning the iron there's still some resistance, try to scratch the solder with the iron when it's hot. Generally, this removes any barriers to heat flow and will instantly melt the solder.

No matter how good you get at soldering and desoldering, some jobs are best left to professionals. Components that are surface-mounted to a circuit board, for example, require special tools for soldering and desoldering, as do other components that have high pin densities.

I upgraded an IBM P75 portable system by replacing the 486DX-33 processor with a 486DX2-66 processor. This procedure normally would be simple (especially if the system uses a ZIF socket), but in this particular system, the 168-pin 486DX chip was soldered into a special processor card. To add to the difficulty, there were surface-mounted components on both sides of the card--even the solder side.

Needless to say, this was a very difficult job that required a special piece of equipment called a hot air rework station. The hot air rework station uses blasts of hot air to solder or desolder all of the pins on a chip simultaneously. To perform this replacement job, the components on the solder side of the board were protected with special heat-resistant masking tape, while the hot air was directed at the 168 pins of the 486 chip, allowing it to be removed. Then the replacement chip was inserted into the holes in the board, a special solder paste was applied to the pins, and the hot air was used again to solder all 168 pins simultaneously.

The use of professional equipment such as this resulted in a perfect job that cannot be told from the factory original. Attempting a job like this with a conventional soldering iron probably would have damaged the expensive processor chips, as well as the even more expensive multilayer processor card.

Using Proper Test Equipment

In some cases, you must use specialized devices to test a system board or component. This test equipment is not expensive or difficult to use, but it can add much to your troubleshooting abilities. I consider a voltmeter to be required gear for proper system testing. A multimeter can serve many purposes, including checking for voltage signals at different points in a system, testing the output of the power supply, and checking for continuity in a circuit or cable. An outlet tester is an invaluable accessory that can check the electrical outlet for proper wiring. This capability is useful if you believe that the problem lies outside the computer system itself.

Wrap Plugs (Loopback Connectors)

For diagnosing serial- and parallel-port problems, you need wrap plugs (also called loopback connectors), which are used to circulate, or wrap, signals. The plugs enable the serial or parallel port to send data to itself for diagnostic purposes.

Several types of wrap plugs are available. You need one for the 25-pin serial port, one for the 9-pin serial port, and one for the 25-pin parallel port (see Table 3.1). Many companies, including IBM, sell the plugs separately. IBM also sells a special version that includes all three types in one plug.

Table 3.1  Wrap Plug Types

Description IBM Part Number
Parallel-port wrap plug 8529228
Serial-port wrap plug, 25-pin 8529280
Serial-port wrap plug, 9-pin (AT) 8286126
Tri-connector wrap plug 72X8546

The handy tri-connector unit contains all commonly needed plugs in one compact unit. The unit costs approximately $30 from IBM. Be aware that most professional diagnostics packages (especially the ones that I recommend) include the three types of wrap plugs in the package; you may not need to purchase them separately. If you're handy, you can even make your own wrap plugs for testing. I include wiring diagrams for the three types of wrap plugs in Chapter 11, "Communications and Networking." In that chapter, you also will find a detailed discussion of serial and parallel ports.

Beyond a simple wrap plug, you can use a breakout box. These are usually DB25 connector devices which allow you to make custom temporary cables or even to monitor signals on a cable. For most PC troubleshooting use, one of the "mini" breakout boxes works well and is inexpensive.

Meters

Many troubleshooting procedures require that you measure voltage and resistance. You take these measurements by using a handheld Digital Multi-Meter (DMM). The meter can be an analog device (using an actual meter) or a digital-readout device. The DMM has a pair of wires called test leads or probes. The test leads make the connections so that you can take readings. Depending on the meter's setting, the probes measure electrical resistance, direct-current (DC) voltage, or alternating-current (AC) voltage.

Usually, each system-unit measurement setting has several ranges of operation. DC voltage, for example, usually can be read in several scales, to a maximum of 200 millivolts (mv), 2v, 20v, 200v, and 1,000v. Because computers use both +5 and +12v for various operations, you should use the 20v maximum scale for making your measurements. Making these measurements on the 200mv or 2v scale could "peg the meter" and possibly damage it because the voltage would be much higher than expected. Using the 200v or 1,000v scale works, but the readings at 5v and 12v are so small in proportion to the maximum that accuracy is low.

If you are taking a measurement and are unsure of the actual voltage, start at the highest scale and work your way down. Most of the better meters have autoranging capability: The meter automatically selects the best range for any measurement. This type of meter is much easier to operate. You just set the meter to the type of reading you want, such as DC volts, and attach the probes to the signal source. The meter selects the correct voltage range and displays the value. Because of their design, these types of meters always have a digital display rather than a meter needle.

CAUTION: Whenever using a multimeter to test any voltage that could potentially be 110v or above, always use one hand to do the testing, not two. Either clip one lead to one of the sources and probe with the other, or hold both leads in one hand.

If you are holding a lead in each hand and accidentally slip, you can very easily become a circuit, allowing power to conduct or flow through you. When the power is flowing from arm to arm, the path of the current is directly across the heart. Hearts have a tendency to quit working when subjected to high voltages. They're funny that way.

I prefer the small digital meters; you can buy them for only slightly more than the analog style, and they're extremely accurate, as well as much safer for digital circuits. Some of these meters are not much bigger than a cassette tape; they fit in a shirt pocket. Radio Shack sells a good unit (made for Radio Shack by Beckman) in the $25 price range; the meter is a half-inch thick, weighs 3 1/2 ounces, and is digital and autoranging as well. This type of meter works well for most, if not all, PC troubleshooting and test uses.

CAUTION: You should be aware that many analog meters can be dangerous to digital circuits. These meters use a 9v battery to power the meter for resistance measurements. If you use this type of meter to measure resistance on some digital circuits, you can damage the electronics, because you essentially are injecting 9v into the circuit. The digital meters universally run on 3 to 5v or less.

Logic Probes and Logic Pulsers

A logic probe can be useful for diagnosing problems in digital circuits. In a digital circuit, a signal is represented as either high (+5v) or low (0v). Because these signals are present for only a short time (measured in millionths of a second) or oscillate (switch on and off) rapidly, a simple voltmeter is useless. A logic probe is designed to display these signal conditions easily.

Logic probes are especially useful for troubleshooting a dead system. By using the probe, you can determine whether the basic clock circuitry is operating and whether other signals necessary for system operation are present. In some cases, a probe can help you cross-check the signals at each pin on an Integrated Circuit chip. You can compare the signals present at each pin with the signals that a known-good chip of the same type would show--a comparison that is helpful in isolating a failed component. Logic probes can be useful for troubleshooting some disk drive problems by enabling you to test the signals present on the interface cable or drive-logic board.

A companion tool to the probe is the logic pulser. A pulser is designed to test circuit reaction by delivering into a circuit a logical high (+5v) pulse, usually lasting 1 1/2 to 10 millionths of a second. Compare the reaction with that of a known-functional circuit. This type of device normally is used much less frequently than a logic probe, but in some cases, it can be helpful for testing a circuit.

Outlet Testers

Outlet testers are very useful test tools. These simple, inexpensive devices, which are sold at hardware stores, are used to test electrical outlets. You simply plug the device in, and three LEDs light in various combinations, indicating whether the outlet is wired correctly.

Although you may think that badly wired outlets would be a rare problem, I have seen a large number of installations in which the outlets were wired incorrectly. Most of the time, the problem seems to be in the ground wire. An improperly wired outlet can result in flaky system operation, such as random parity checks and lockups. With an improper ground circuit, currents can begin flowing on the electrical ground circuits in the system. Because the system uses the voltage on the ground circuits as a comparative signal to determine whether bits are 0 or 1, a floating ground can cause data errors in the system.

Once, while running one of my PC troubleshooting seminars, I was using a system that I literally could not approach without locking it up. Whenever I walked past the system, the electrostatic field generated by my body interfered with the system, and the PC locked up, displaying a parity-check error message. The problem was that the hotel I was using was very old and had no grounded outlets in the room. The only way I could prevent the system from locking up was to run the class in my stocking feet, because my leather-soled shoes were generating the static charge.

Other symptoms of bad ground wiring in electrical outlets are continuous electrical shocks when you touch the case or chassis of the system. These shocks indicate that voltages are flowing where they should not be. This problem also can be caused by bad or improper grounds within the system itself. By using the simple outlet tester, you can quickly determine whether the outlet is at fault.

If you just walk up to a system and receive an initial shock, it's probably just static electricity. Touch the chassis again without moving your feet. If you receive another shock, there is something very wrong. In this case, the ground wire actually has voltage applied to it. You should have a professional electrician come out immediately.

If you don't like being a human rat in an electrical experiment, you can test the outlets with your multimeter. First, remember to hold both leads in one hand. Test from one blade hole to another. This should read between 110-125v depending upon the electrical service in the area. Then check from each blade to the ground (the round hole). One blade hole, the smaller one, should show a voltage almost identical to the one that you got from the blade hole-to-blade hole test. The larger blade hole when measured to ground should show less than 0.5v.

Because ground and neutral are supposed to be tied together at the electrical panel, much difference indicates that they are not tied together. However, small differences can be accounted for by the fact that there may be current from other outlets down the line flowing on the neutral, and there isn't any on the ground.

If you don't get the results you expect, call an electrician to test the outlets for you. More weird computer problems are caused by improper grounding, and other power problems, than people like to believe.

SIMM Testers

I now consider a SIMM test machine a virtually mandatory piece of gear for anybody serious about performing PC troubleshooting and repair as a profession. These are basically small test machines designed to evaluate SIMM and other types of memory modules including individual chips such as cache memory. They can be somewhat expensive, costing upwards of $1,000 to $2,500 or more, but these types of machines are the only truly accurate way to test memory.

Without one of these testers, you are relegated to testing memory by running a diagnostic program on the PC and testing the memory as it is installed. This can be very problematic, as the memory diagnostic program can do only two things to the memory: write and read. A SIMM tester can do many things that a memory diagnostic running in a PC cannot do, such as:

No conventional memory diagnostic software can do these things because it has to rely on the fixed-access parameters set up by the memory controller hardware in the motherboard chipset. This prevents the software from being able to alter the timing and methods used to access the memory. You end up with memory that will fail in one system and work in another, when it is in fact actually bad. This can allow intermittent problems to occur, and be almost impossible to detect.

The bottom line is that there is no way that truly accurate memory testing can be done in a PC; a SIMM tester is required for comprehensive and accurate testing of memory. With the price of a typical 32M memory module at more than $200, the price of a SIMM tester can be justified very easily in a shop environment where a lot of PCs will be tested. One of the SIMM testers I recommend the most is the SIGMA LC by Darkhorse Systems. See the vendor list in Appendix A for more information. Also see Chapter 7, "Memory," for more information on memory in general.

Chemicals

Chemicals can be used to help clean, troubleshoot, and even repair a system. For the most basic function--cleaning components, electrical connectors, and contacts--one of the most useful chemicals was 1,1,1 trichloroethane. This substance was a very effective cleaner. This chemical was used to clean electrical contacts and components, and did not damage most plastics and board materials. In fact, trichloroethane could be very useful for cleaning stains on the system case and keyboard. Electronic chemical-supply companies are now offering several replacements for trichloroethane because it is being regulated as a chlorinated solvent, along with CFCs (chlorofluorocarbons) such as freon.

A unique type of contact enhancer and lubricant called Stabilant 22 is currently on the market. This chemical, which is applied to electrical contacts, greatly enhances the connection and lubricates the contact point; it is much more effective than conventional contact cleaners or lubricants.

Stabilant 22 is a liquid-polymer semiconductor; it behaves like liquid metal and conducts electricity in the presence of an electric current. The substance also fills the air gaps between the mating surfaces of two items that are in contact, making the surface area of the contact larger and also keeping out oxygen and other contaminants that can oxidize and corrode the contact point.

This chemical is available in several forms. Stabilant 22 itself is the concentrated version, whereas Stabilant 22a is a version diluted with isopropanol in a 4:1 ratio. An even more diluted 8:1-ratio version is sold in many high-end stereo and audio shops under the name Tweek. Just 15ml of Stabilant 22a sells for about $40, whereas a liter of the concentrate costs about $4,000!

As you can plainly see, Stabilant 22 is fairly expensive, but very little is required in an application, and nothing else has been found to be as effective in preserving electrical contacts. (NASA uses the chemical on spacecraft electronics.) An application of Stabilant can provide protection for up to 16 years, according to its manufacturer, D.W. Electro-chemicals. You will find the company's address and phone number in the vendor list in Appendix A.

Stabilant is especially effective on I/O slot connectors, adapter-card edge and pin connectors, disk drive connectors, power-supply connectors, and virtually any connector in the PC. In addition to enhancing the contact and preventing corrosion, an application of Stabilant lubricates the contacts, making insertion and removal of the connector easier.

Compressed air often is used as an aid in system cleaning. Normally composed of freon or carbon dioxide (CO2), compressed gas is used as a blower to remove dust and debris from a system or component. Be careful when you use these devices: Some of them can generate a tremendous static charge as the compressed gas leaves the nozzle of the can. Be sure that you are using the kind approved for cleaning or dusting computer equipment, and consider wearing a static grounding strap as a precaution. Freon TF is known to generate these large static charges; Freon R12 is less severe.

Of course, because both chemicals damage the ozone layer, most suppliers are phasing them out. Expect to see new versions of these compressed-air devices with CO2 or some other less-harmful propellant.

When using these compressed air products, make sure you hold the can upright so that only gas is ejected from the nozzle. If you tip the can, the raw propellant will come out, which is wasteful. This operation should be performed on equipment which is powered off to minimize any chance of damage through short circuiting or bumping anything.

CAUTION: If you use any chemical that contains the propellant Freon R12 (dichlorodifluoromethane), do not expose the gas to an open flame or other heat source. If you burn this substance, a highly toxic gas called phosgene is generated. Phosgene, used as a choking gas in World War I, can be deadly.

Freon R12 is the substance that was used in most automobile air-conditioning systems before 1995. Automobile service technicians are instructed never to smoke near air-conditioning systems. By 1996, the manufacture and use of these types of chemicals have been either banned or closely regulated by the government, and replacements have been found. For example, virtually all new car automobile air-conditioning systems have been switched to a chemical called R-134a. The unfortunate side effect of this situation is that all the replacement chemicals are much more expensive than freon.

Related to compressed-air products are chemical-freeze sprays. These sprays are used to cool a suspected failing component quickly so as to restore it to operation. These substances are not used to repair a device, but to confirm that you have found the failed device. Often, a component's failure is heat-related; cooling it temporarily restores it to normal operation. If the circuit begins operating normally, the device that you are cooling is the suspect device.

A Word About Hardware

This section discusses some problems that you may encounter with the hardware (screws, nuts, bolts, and so on) used in assembling a system.

Types of Hardware

One of the biggest aggravations that you encounter in dealing with various systems is the different hardware types and designs that hold the units together.

For example, most system hardware types use screws that can be driven with 1/4-inch or 3/16-inch hexagonal drivers. IBM used these screws in all its original PC, XT, and AT systems, and most compatible systems use this standard hardware as well. Some manufacturers use different hardware. Compaq, for example, uses Torx screws extensively in most of its systems. A Torx screw has a star-shape hole driven by the correct-size Torx driver. These drivers carry size designations: T-8, T-9, T-10, T-15, T-20, T-25, T-30, T-40, and so on.

A variation on the Torx screw is the tamperproof Torx screw found in power supplies and other assemblies. These screws are identical to the regular Torx screws, except that a pin sticks up from the middle of the star-shape hole in the screw. This pin prevents the standard Torx driver from entering the hole to grip the screw; a special tamperproof driver with a corresponding hole for the pin is required. An alternative is to use a small chisel to knock out the pin in the screw. Usually, a device that is sealed with these types of screws is considered to be a complete replaceable unit that rarely, if ever, needs to be opened.

Many manufacturers also use the more standard slotted-head and Phillips-head screws. Using tools on these screws is relatively easy, but tools do not grip these fasteners as well as hexagonal head or Torx screws do, and the heads can be rounded off more easily than other types can. Extremely cheap versions tend to lose bits of metal as they're turned with a driver, and the metal bits can fall onto the motherboard. Stay away from cheap fasteners whenever possible; the headaches of dealing with stripped screws aren't worth it.

Some case manufacturers are making cases which snap together or use thumb screws. These are usually advertised as "no-tool" cases because you literally do not need any tools to take out the cover and many of the major assemblies.

Curtis sells special nylon plastic thumb screws that fit most normal cases and can be used to replace the existing screws to make opening the case a no-tool proposition. You should still always use metal screws to install internal components such as adapter cards, disk drives, power supplies, and the motherboard because the metal screws provide a ground point for these devices.

English versus Metric

Another area of aggravation with hardware is the fact that two types of thread systems are available: English and metric. IBM used mostly English-threaded fasteners in its original line of systems, but many other manufacturers used metric-threaded fasteners in their systems.

The difference becomes apparent especially with disk drives. American-manufactured drives sometimes use English fasteners; drives made in Japan or Taiwan usually use metric fasteners. Whenever you replace a floppy drive in an older PC-compatible unit, you encounter this problem. Try to buy the correct screws and any other hardware, such as brackets, with the drive, because they may be difficult to find at a local hardware store. Many of the drive manufacturers offer retail drive kits that include these components. The OEM's drive manual lists the correct data about a specific drive's hole locations and thread size.

Hard disks can use either English or metric fasteners; check your particular drive to see which type it uses. Most drives today seem to use metric hardware.

CAUTION: Some screws in a system may be length-critical, especially screws that are used to retain hard disk drives. You can destroy some hard disks by using a mounting screw that's too long; such a screw can puncture or dent the sealed disk chamber when you install the drive and fully tighten the screw. When you install a new drive in a system, always make a trial fit of the hardware to see how far the screws can be inserted into the drive before they interfere with components of the drive. When you're in doubt, the drive manufacturer's OEM documentation will tell you precisely what screws are required and how long they should be.

Disassembly and Reassembly Procedures

The process of physically disassembling and reassembling systems isn't difficult. Because of marketplace standardization, only a couple of different types and sizes of screws (with a few exceptions) are used to hold the systems together. Also, the physical arrangement of the major components is similar even among systems from different manufacturers. In addition, a typical system does not contain many components today.

This section covers the disassembly and reassembly procedure in the following sections:

This section discusses how to remove and install these components for several different types of systems. With regard to assembly and disassembly, it is best to consider each system by the type of case that the system uses. All systems that have AT-type cases, for example, are assembled and disassembled in much the same manner. Tower cases basically are AT-type cases turned sideways, so the same basic instructions apply to those cases as well. Most Slimline and XT style cases are similar; these systems are assembled and disassembled in much the same way.

The following section lists disassembly and reassembly instructions for several case types, including those for all standard IBM-compatible systems.

Disassembly Preparation

Before you begin disassembling any system, you must be aware of several issues. One issue is ESD (electrostatic discharge) protection. The other is recording the configuration of the system, with regard to the physical aspects of the system (such as jumper or switch settings and cable orientations) and to the logical configuration of the system (especially in terms of elements such as CMOS settings).

ESD Protection

When you are working on the internal components of a system, you need to take the necessary precautions to prevent accidental static discharges to the components. At any time, your body can hold a large static voltage charge that can easily damage components of your system. Before I ever put my hands into an open system, I first touch a grounded portion of the chassis, such as the power supply case. This action serves to equalize the charges that the device and I would be carrying. The key here is to leave the device computer in. By leaving the computer plugged in, you're allowing the static electricity to drain off safely to ground, rather than forcing the components of the system to accept the jolt.

In past editions of this book, it's been recommended that you unplug your systems. This is still true where you're concerned that you may accidentally power on the system while working on it. However, if this is not a concern, you should leave the computer plugged in.

High-end workbenches at repair facilities have the entire bench grounded, so it's not as big a problem; however, you need something to be a good ground source to prevent current from building up in you, and the best source is in the power cord that connects the computer to the wall.

A more sophisticated way to equalize the charges between you and any of the system components is to use an ESD protection kit. These kits consist of a wrist strap and mat, with ground wires for attachment to the system chassis. When you are going to work on a system, you place the mat next to or partially below the system unit. Next, you clip the ground wire to both the mat and the system's chassis, tying the grounds together. Then you put on the wrist strap and attach that wire to a ground as well. Because the mat and system chassis are already wired together, you can attach the wrist-strap wire to the system chassis or to the mat itself. If you are using a wrist strap without a mat, clip the wrist-strap wire to the system chassis. When clipping these wires to the chassis, be sure to use an area that is free of paint so that a good ground contact can be achieved. This setup ensures that any electrical charges are carried equally by you and any of the components in the system, preventing the sudden flow of static electricity that can damage the circuits.

As you remove disk drives, adapter cards, and especially delicate items such as the entire motherboard, as well as SIMMs or processor chips, you should place these components on the static mat. I see some people putting the system unit on top of the mat, but the unit should be alongside the mat so that you have room to lay out all the components as you remove them. If you are going to remove the motherboard from a system, be sure that you leave enough room for it on the mat.

If you do not have such a mat, simply place the removed circuits and devices on a clean desk or table. Always pick up a loose adapter card by the metal bracket used to secure the card to the system. This bracket is tied into the ground circuitry of the card, so by touching the bracket first, you prevent a discharge from damaging the components of the card. If the circuit board has no metal bracket (a motherboard, for example), handle the board carefully by the edges, and try not to touch any of the components.

CAUTION: Some people have recommended placing loose circuit boards and chips on sheets of aluminum foil. This procedure is absolutely not recommended and can actually result in an explosion! Many motherboards, adapter cards, and other circuit boards today have built-in lithium or ni-cad batteries. These batteries react violently when they are shorted out, which is exactly what you would be doing by placing such a board on a piece of aluminum foil. The batteries will quickly overheat and possibly explode like a large firecracker (with dangerous shrapnel). Because you will not always be able to tell whether a board has a battery built into it somewhere, the safest practice is to never place any board on any conductive metal surface, such as foil.

Recording Setup and Configuration

Before you power off the system for the last time to remove the case, you should learn, and record, several things about the system. Often when working on a system, you intentionally or accidentally wipe out the CMOS Setup information. Most systems use a special battery-powered CMOS clock and data chip that is used to store the system's configuration information. If the battery is disconnected, or if certain pins are accidentally shorted, you can discharge the CMOS memory and lose the setup. The CMOS memory in most systems is used to store simple things such as how many and what type of floppy drives are connected, how much memory is in the system, and the date and time.

A critical piece of information is the hard disk type settings. Although you or the system can easily determine the other settings the next time you power on the system, the hard disk type information is another story. Most modern BIOS software can read the type information directly from most IDE and all SCSI drives. With older BIOS software, however, you have to explicitly tell the system the parameters of the attached hard disk. This means that you need to know the current settings for cylinders, heads, and sectors per track.

Some BIOS software indicates the hard disk only by a type number, usually ranging from 1-50. Be aware that most BIOS programs use type 47 or higher for what is called a user-definable type, which means that the cylinder, head, and sector counts for this type were entered manually and are not constant. These user-definable types are especially important to write down, because this information may be very difficult to figure out later when you need to start the system.

Modern Enhanced IDE drives will also have additional configuration items that should be recorded. These include the translation mode and transfer mode. With drives larger than 528M, it is important to record the translation mode, which will be expressed differently in different BIOS versions. Look for settings like CHS (Cylinder Head Sector), ECHS (Extended CHS), Large (which equals ECHS), or LBA (Logical Block Addressing). If you reconfigure a system and do not set the same drive translation as was used originally with that drive, then all the data may be inaccessible. Most modern BIOS have an autodetect feature that automatically reads the drive's capabilities and sets the CMOS settings appropriately. Even so, there have been some problems with the BIOS not reading the drive settings properly, or where someone had overridden the settings in the previous installation. With translation, you have to match the setting to what the drive was formatted under previously if you want to read the data properly.

The speed setting is a little more straightforward. Older IDE drives can run up a speed of 8.3M/sec, which is called PIO (Programmed I/O) mode 2. Newer EIDE drives can run PIO Mode 3 (11.1M/sec) or PIO Mode 4 (16.6M/sec). Most BIOSes today allow you to set the mode specifically, or you can use the autodetect feature to automatically set the speed. For more information on the settings for hard disk drives, refer to Chapter 15, "Hard Disk Interfaces."

If you do not enter the correct hard disk type information in the CMOS setup program, you will not be able to access the data on the hard disk. I know of several people who lost some or all of their data because they did not enter the correct type information when they reconfigured their systems. If this information is incorrect, the usual results are a Missing operating system error message when the system starts and the inability to access the C drive.

Some of you may be thinking that you can just figure out the parameters by looking up the particular hard disk in a table. Unfortunately, this method works only if the person who set up the system originally entered the correct parameters. I have encountered a large number of systems in which the hard disk parameters were not entered correctly; the only way to regain access to the data is to determine, and then use, the same incorrect parameters that were used originally. As you can see, no matter what, you should record the hard disk information from your setup program.

Most systems have the setup program built right into the ROM BIOS software itself. These built-in setup programs are activated by a hot-key sequence such as Ctrl+Alt+Esc or Ctrl+Alt+S if you have a Phoenix ROM. Other ROMs prompt you for the setup program every time the system boots, such as with the popular AMI BIOS. With the AMI, you simply press the Delete key when the prompt appears on-screen during a reboot.

When you get the setup program running, record all the settings. The easiest way to do this is to print it out. If a printer is connected, press Shift+Print Screen; a copy of the screen display will be sent to the printer. Some setup programs have several pages of information, so you should record the information on each page as well.

Many setup programs, such as those in the AMI BIOS, allow for specialized control of the particular chipset used in the motherboard. These complicated settings can take up several screens of information, all of which should be recorded. Most systems will return all these settings to a default state when the battery is removed, and you lose any custom settings that were made.

MCA and EISA bus systems have a very sophisticated setup program that stores not only the motherboard configuration but also configurations for all the adapter cards. Fortunately, the setup programs for these systems have the capability to save the settings to a file on a floppy disk so that they can be restored later.

To access the setup program for most of these systems, you need the Setup or Reference Disk for the particular system. Some systems, such as various IBM PS/2 and Compaq systems, store a complete copy of the Reference or Setup Disk in a hidden partition of the hard disk. When these systems boot up, the cursor jumps to the right side of the screen for a few seconds. During this time, if you press Ctrl+Alt+Insert for IBM systems or F10 for Compaq systems, the hidden setup programs should load. Other manufacturers will use different keystrokes to activate the setup program or hidden partition, so consult your documentation to find the correct keystrokes for your particular system.

There are programs which advertise the ability to save all of your CMOS settings to disk, allowing you to reload them later. This works as long as the program is designed to work with your specific BIOS, but unfortunately there is no such program that works properly with all of the different BIOS versions on the market today. If you use such a program, be sure it is compatible with your system before proceeding.

Recording Physical Configuration

While you are disassembling a system, it is a good idea to record all the physical settings and configurations within the system, including jumper and switch settings, cable orientations and placement, ground-wire locations, and even adapter-board placement. Keep a notebook handy for recording these items, and write down all the settings in the book.

It is especially important to record all the jumper and switch settings on every card that you remove from the system, as well as those on the motherboard itself. If you accidentally disturb these jumpers or switches, you will know how they were originally set. This knowledge is very important if you do not have all the documentation for the system handy. Even if you do, undocumented jumpers and switches often do not appear in the manuals but must be set a certain way for the item to function. It is very embarrassing, to say the least, if you take apart somebody's system and then cannot make it work again because you disturbed something. If you record these settings, you will save yourself the embarrassment.

Also record all cable orientations. Most name-brand systems use cables and connectors that are keyed so that they cannot be plugged in backward, but most generic compatibles do not have this added feature. In addition, it is possible to mix up hard disk and floppy cables. You should mark or record what each cable was plugged into and its proper orientation. Ribbon cables usually have an odd-colored (red, green, blue, or black) wire at one end that indicates pin 1. There may also be a mark on the connector such as a triangle or even the number 1. The devices into which the cables are plugged also are marked in some way to indicate the orientation of pin 1. Often there will be a dot next to the pin 1 side of the connector, or there may be a 1 or other mark. In some cases, the cables and connectors will be keyed such that they can only go in one way.

Although cable orientation and placement seem to be very simple, we rarely get through the entire course of my PC troubleshooting seminars without at least one group of people having cable-connection problems. Fortunately, in most cases (excepting power cables), plugging in any of the ribbon cables inside the system backward rarely causes any permanent damage.

Power and battery connections are exceptions; plugging them in backward in most cases will cause damage. In fact, plugging the motherboard power connectors in backward or in the wrong plug location will put 12v where only 5v should be--a situation that can cause components of the board to violently explode. I know of several people who have facial scars caused by shrapnel from exploding components caused by improper power supply connections! As a precaution, I always like to turn my face away from the system when I power it on for the first time.

Plugging the CMOS battery in backward can damage the CMOS chip itself, which usually is soldered into the motherboard; in such a case, the motherboard itself must be replaced.

Finally, it is a good idea to record miscellaneous items such as the placement of any ground wires, adapter cards, and anything else that you may have difficulty remembering later. Some configurations and setups are particular about which slots the adapter cards are located in; it usually is a good idea to put everything back exactly the way it was originally, especially in MCA and EISA bus systems.

Now that you have made the necessary preparations and taken the necessary precautions, you can actually begin working on the systems.

System Disassembly

Disassembling most systems normally requires only a few basic tools: a 1/4-inch nut driver or Phillips-head screwdriver for the external screws that hold the cover in place, and a 3/16-inch nut driver or Phillips-head screwdriver for all the other screws. A needle-nose pliers can also help in removing motherboard standoffs, jumpers, and stubborn cable connectors.

An antistatic mat is useful for placing components on while they are out of the system chassis to protect them from static. If you don't have a mat, then any nonmetallic static-free surface can suffice as a work area.

Removing the Cover

To remove the case cover, follow these steps:

1. Power off the system. Disconnect all of the cables at the back of the case, including the power cable.

2. Examine your case to determine how to remove the cover. Remove the screws holding the case cover on the chassis. These are normally around the rim of the cover and are normally in the rear; however, in some cases, the screws are behind the front plastic faceplate or bevel (see Figure 3.1).

FIG. 3.1  Remove the screws that hold the cover in place.

3. Once the screws are removed, grasp the cover and slide or lift it off. Some covers slide toward the back and some to the front; some lift straight up.

4. Now is a good time to connect the wrist strap or antistatic mat if you have one. Wear the wrist strap on one wrist and clip the wire end to a metal part of the case. Clip the wire from the mat to the case as well. This will keep you and the equipment at the same electrical potential and prevent any damage due to static electricity.

Removing Adapter Boards

To remove all the adapter boards from the system unit, first remove the system-unit cover, as described in the previous section. Then proceed as follows for each adapter:

1. Note which slots each adapter is in; if possible, make a diagram or drawing.

2. Remove the screw that holds the adapter in place (see Figure 3.2).

3. Note the positions of any cables that are plugged into the adapter before you remove them. In a correctly wired system, the colored stripe on one side of the ribbon cable always denotes pin 1. Some connectors have keys that enable them to be inserted only the correct way.

FIG. 3.2  Remove the screw that holds the adapter in place.

4. Remove the adapter by lifting with even force at both ends.

5. Note the positions of any jumpers or switches on the adapter, especially when documentation for the adapter is not available. Even when documentation is available, manufacturers often use undocumented jumpers and switches for special purposes, such as testing or unique configurations. It's a good idea to know the existing settings, in case they are disturbed later.

Removing Disk Drives

Removing drives is very easy. The procedure is similar for all types of drives, such as floppy, hard disk, CD-ROM, and even tape drives. Special rails or brackets are attached to the sides of the drives, and the drives slide into the system-unit chassis on these rails or brackets. The chassis has guide tracks for the rails, which enable you to remove the drive from the front of the unit without having to access the side to remove any mounting screws. The rails normally are made of plastic or fiberglass, but they can be made of metal in some systems. It should be noted that any brackets or rails should remain attached to the drive while you are removing or installing it.

Always back up hard disks completely before removing disks from the system. A backup is important because the possibility always exists that data will be lost or the drive damaged by rough handling.

To remove the drives, first remove the cover as described earlier. Then proceed as follows:

1. Depending on the specific case design, drives may be mounted using special brackets or rails. Locate the screws holding each drive bracket or drive assembly in the case and remove them. Some drives will slide out on rails, or be removed with the bracketry still attached to the drive itself.

2. Disconnect from the drives the power cables, data cables, and any ground wires if present (see Figures 3.3 and 3.4). In a correctly wired system, the colored stripe on one side of the ribbon cable always denotes pin 1. The power connector is shaped so that it can be inserted only the correct way.

3. Slide the drive completely out of the unit.

FIG. 3.3  Disconnect the hard drive power cable, signal and data cables, and ground wire.

Removing the Power Supply

The power supply is mounted in the system unit with several (normally four) screws in the rear and sometimes interlocking tabs on the bottom. Removing the power supply may require that you slide the disk drives forward for clearance when you remove the supply. To remove the power supply, first remove the cover, then proceed as follows:

1. Remove the power-supply retaining screws from the rear of the system-unit chassis (see Figure 3.5).

FIG. 3.4  Disconnect the floppy drive power cable, signal cable, and ground wire.

FIG. 3.5  Remove the power-supply retaining screws from the rear of the chassis.

2. Disconnect the cables from the power supply to the motherboard (see Figure 3.6), and then disconnect the power cables from the power supply to the disk drive. Always grasp the connectors themselves; never pull on the wires.

3. Lift the power supply out of the chassis.

FIG. 3.6  Disconnect the cables from the power supply to the motherboard.

Removing the Motherboard

After all adapter cards are removed from the unit, you can remove the motherboard. The motherboard is normally held in place by several screws, and may also use plastic standoffs that elevate the board from the metal chassis so that the bottom of the board does not touch the chassis and cause a short.

You should not separate the standoffs from the motherboard before removing it from the chassis; instead, remove the board and the standoffs as a unit. The standoffs normally slide into slots in the chassis. When you reinstall the motherboard, make sure that the standoffs are located in their slots properly. If one or more standoffs do not engage the chassis properly, you may crack the motherboard when you tighten the screws or install adapter cards.

To remove the motherboard, first remove all adapter boards from the system unit, as described earlier. Then proceed as follows:

1. If the motherboard has onboard floppy, hard disk, serial, or parallel ports, document those cable connections and mark them before disconnecting them.

2. There are numerous small wires that go from the front panel on the case to the motherboard. Before disconnecting them, document the wire connections to the motherboard, and use some masking tape or some kind of label to mark the small wire connectors as you take them off the motherboard. Marking these wires will save you a lot of time later during the installation.

3. If there is an active heat sink in the CPU which incorporates a fan, unplug the power lead to the CPU fan.

4. If you have not already removed the power supply, document how the power supply cables are plugged into the motherboard, and disconnect the power supply connections to the board.

5. Locate and remove the motherboard retaining screws, making sure you save any plastic washers that might be used.

6. Slide the motherboard away from the power supply about a half-inch, until the standoffs have disengaged from their mounting slots (see Figure 3.7).

7. Lift the motherboard up and out of the chassis. Place it on a static-free surface, such as an antistatic mat.

8. Remove the CPU and any memory modules you want to reuse later.

FIG. 3.7  Disengage standoffs from their mounting slots.

Removing SIMMs or DIMMs

One benefit of using single or dual inline memory modules (SIMMs or DIMMs) is that they're easy to remove or install. When you remove memory modules, remember that because of physical interference, you must remove the memory-module package that is closest to the disk drive bus-adapter slot before you remove the package that is closest to the edge of the motherboard. This procedure describes removing a SIMM device; note that a DIMM is also removed in exactly the same way. The only difference is that a DIMM is slightly longer and has more contacts than a SIMM device.

To remove a SIMM (or DIMM) properly, follow this procedure:

1. Gently pull the tabs on each side of the SIMM or DIMM socket outward.

2. Rotate or pull the SIMM or DIMM up and out of the socket (see Figure 3.8).

FIG. 3.8  Removing a SIMM or DIMM.

CAUTION: Be careful not to damage the connector. If you damage the motherboard memory connector, you could be looking at an expensive repair. Never force the module; it should come out easily. If it doesn't, you are doing something wrong.

Motherboard Installation

If you are installing a new motherboard, unpack the motherboard and check to make sure you have everything that should be included. If you purchase a new board, normally you should get at least the motherboard itself, some I/O cables, and a manual. If you ordered the motherboard with a processor or memory, it will normally be installed on the board for you, but may also be included separately. Some board kits will include an antistatic wrist strap to help prevent damage due to static electricity when installing the board.

Prepare the New Motherboard

Before a new motherboard can be installed, it must be set up properly to accept the processor. Most newer motherboards have jumpers which control both the CPU speed as well as the voltage supplied to it. If these are set incorrectly, the system may not operate at all, operate erratically, or possibly even damage the CPU. If you have any questions about the proper settings, contact the vendor who sold you the board before making any jumper changes.

Most processors today run hot enough to require some form of heat sink to dissipate heat from the processor. To install the processor and heat sink, use the following procedures:

1. Take the new motherboard out of the antistatic bag it was supplied in, and set it on the bag or the antistatic mat if you have one.

2. Refer to the motherboard manufacturer's manual to set the jumpers to match the CPU you are going to install. Look for the diagram of the motherboard to find the jumper location, and look for the tables for the right settings for your CPU. If the CPU was supplied already installed on the motherboard, then the jumpers should already be correctly set for you, but it is still a good idea to check them.

3. Find pin 1 on the processor; it is usually denoted by a corner of the chip that is marked by a dot or a bevel in that corner. Next, find the corresponding pin 1 of the ZIF socket for the CPU on the motherboard. Pin 1 is usually marked on the board, or there is a bevel in one corner of the socket. Insert the CPU into the ZIF socket by lifting the release lever, aligning the pins on the processor with the holes in the socket, and pushing it down into place. If the processor does not go all the way into the socket, then check for possible interference or pin alignment problems. When the processor is fully seated in the socket, push the locking lever on the socket down to secure the processor.

4. If the CPU does not already have a heat sink attached to it, then attach it now. Most heat sinks will either clip directly to the CPU or to the socket with one or more retainer clips. Be careful when attaching the clip to the socket; you don't want it to scrape against the motherboard, which might damage circuit traces or components. In most cases, it is a good idea to put a dab of heat sink thermal transfer compound (normally a white-colored grease) on the CPU before installing the heat sink. This prevents any air gaps and allows the heat sink to work more efficiently.

Install Memory Modules

In order to function, the motherboard must have memory installed on it. Modern motherboards use either SIMMs or DIMMs. Depending on the module type, it will have a specific method of sliding into and clipping to the sockets. Normally, you install modules in the lowest numbered sockets or banks first. Note that some boards will require modules to be installed in pairs or even four at a time. Consult the motherboard documentation for more information on which sockets to use first and in what order, and how to install the specific modules the board uses.

Memory modules are normally keyed to the sockets by a notch on the side or on the bottom, so they can only go in one way. Reverse the procedure discussed in the section "Removing SIMMs or DIMMs" for removing the modules to install them instead.

Mount the New Motherboard in the Case

The motherboard will attach to the case with one or more screws and often several plastic standoffs. If you are using a new case, you may have to attach one or more metal spacers or plastic standoffs in the proper holes before you can install the motherboard. Use the following procedure to install the new motherboard in the case:

1. If plastic standoffs were used, you may need to take them out of the old mother-board. Using a needlenose pliers, carefully squeeze the top of each standoff and push it carefully through the board.

2. Find the corresponding holes in the new motherboard for the metal spacers and plastic standoffs. You should use metal spacers wherever there is a ring of solder around the hole. Use plastic standoffs where there is not the ring of solder. Screw any metal spacers into the new case in the proper positions to align with the screw holes in the motherboard.

3. Insert any plastic standoffs directly into the new motherboard from underneath until they snap into place.

4. Install the new motherboard into the case by setting it down so any standoffs engage the case. Often you will have to set the board into the case and slide it sideways to engage the standoffs into the slots in the case. When the board is in the proper position, the screw holes in the board should be aligned with any of the metal spacers or screw holes in the case.

5. Take the screws and any plastic washers that were previously used, and screw the motherboard to the case.

Connect the Power Supply

Newer ATX-style motherboards have a single power connector which can only go on one way. Baby-AT and other style board designs usually have two separate six-wire power connectors from the power supply to the board, which may not be keyed and therefore may be interchangeable. Even though it may be possible to insert them several ways, only one way is correct! These power leads are usually labeled P8 and P9 in most systems. The order in which they are put on the board is crucial; if you install them backwards, you might cause damage to the motherboard when you power it up. Many systems also use a CPU cooling fan which should be connected as well. To attach the power connectors from the power supply to the motherboard:

1. If the system uses a single ATX-style power connector, then plug it in; it can only go on one way.

If two separate six-wire connectors are used, the two black ground wires on the ends of the connectors must meet in the middle. Align the power connectors such that the black ground wires are adjacent to one another and plug the connectors in. Consult the documentation with your board to make sure the power supply connection is correct.

2. Plug in the power lead for the CPU fan if one is used. The fan will either connect to the power supply via a disk drive power connector, or it may connect directly to a fan power connector directly on the motherboard.

Connect I/O and Other Cables to the Motherboard

There are several connections that must be made between a motherboard and the case. These include LEDs for the hard disk and power, an internal speaker connection, a reset button, and a de-turbo button on some systems. Most modern motherboards also have several built-in I/O ports which have to be connected. This would include dual IDE host adapters, a floppy controller, dual serial ports, and a parallel port. Some boards will also include additional items such as built-in video, sound, or SCSI adapters.

If the board is an ATX type, the connectors for all of the external I/O ports are already built into the rear of the board. If you are using a Baby-AT type board, then you may have to install cables and brackets to run the serial, parallel, and other external I/O ports to the rear of the case. If your motherboard has on-board I/O, use the following procedures to connect the cables:

1. Connect the floppy cable between the floppy drives and the 34-pin floppy controller connector on the motherboard.

2. Connect the IDE cables between the hard disk, IDE CD-ROM, and IDE tape drives and the 40-pin primary and secondary IDE connectors on the motherboard. Normally, you will use the primary IDE channel connector for hard disks only and the secondary IDE channel connector to attach an IDE CD-ROM or tape drive.

3. On non-ATX boards, a 25-pin female cable port bracket is normally used for the parallel port. There are usually two serial ports: a 9-pin, and either another 9-pin or a 25-pin male connector port. Align pin 1 on the serial and parallel port cables with pin 1 on the motherboard connector and plug them in.

4. If the ports don't have card slot type brackets or if you need all of your expansion slots, there may be port knockouts on the back of the case you can use instead. Find ones that fit the ports, and push them out, removing the metal piece covering the hole. Unscrew the hex nuts on either side of the port connector and position it in the hole. Install the hex nuts back in through the case to hold the port connector in place.

5. Most newer motherboards also include a built-in mouse port. If the connector for this port is not built into the back of the motherboard (usually next to the keyboard connector), then you will probably have a card bracket type connector to install. In that case, plug the cable into the motherboard mouse connector and then attach the external mouse connector bracket to the case.

6. Attach the front panel switch, LED, and internal speaker wires from the case front panel to the motherboard. If they are not marked on the board, check where each one is on the diagram in the motherboard manual.

Install Bus Expansion Cards

Most systems use expansion cards for video, network interface, sound, and SCSI adapters. These cards will be plugged into the bus slots present on the motherboard. To install these cards:

1. Insert each card by holding it carefully by the edges, making sure not to touch the chips and circuitry. Put the bottom edge finger connector into a slot that fits. Firmly press down on the top of the card, exerting even pressure, until it snaps into place.

2. Secure each card bracket with a screw.

3. Attach any internal cables you may have removed earlier from the cards.

Replace the Cover and Connect External Cables

Now the system should be nearly assembled. All that remains is installing the cover assembly and connecting any external devices which are cabled to the system. I usually don't like to install the case cover screws until I have tested the system and I am sure everything is working properly! Often I will find that a cable has been connected improperly, or some jumper setting is not correct, requiring that I remove the cover to repair the problem. Use the following procedure to complete the assembly:

1. Slide the cover back on the case.

2. Before powering up the system, connect any external cables. Most of the connectors are D-shaped and only go in one way.

3. Plug the 15-pin monitor cable into the video card female connector.

4. Attach the phone cord to the modem, if any.

5. Plug the round keyboard cable into the keyboard connector, and the mouse into the mouse port or serial port if a serial mouse is being used.

6. If you have any other external cabling such as joystick or audio jacks to a sound card, attach them now as well.

Run the Motherboard BIOS Setup Program (CMOS Setup)

Now that everything is connected, you can power up the system and run the system configuration or setup program. This will allow you to configure the motherboard to the installed devices, as well as to set the system date and time. The system will also test itself to determine if there are any problems:

1. Power on the monitor first, and then the system unit. Observe the operation via the screen and listen for any beeps from the system speaker.

2. The system should automatically go through a Power On Self Test (POST) consisting of video BIOS checking, RAM test, and usually an installed component report. If there is a fatal error during the POST, you may not see anything on the screen and the system may beep several times indicating a specific problem. Check the motherboard or BIOS documentation to determine what the beep codes mean.

3. If there are no fatal errors, you should see the POST display on the screen. Depending on the type of motherboard BIOS, such as Phoenix, AMI, Award, or others, you need to press a key or series of keys to interrupt the normal boot sequence and get to the setup program screens which allow you to enter important system information. Normally the system will indicate via the on-screen display which key to press to activate the BIOS setup program during the POST, but if not, then check the motherboard manual for the key(s) to press to enter the BIOS Setup.

4. Once the setup program is running, use the setup program menus to enter the current date and time, your hard drive settings, floppy drive types, video cards, keyboard settings, and so on. Most newer motherboard BIOS can autodetect the hard drive, so you should not have to manually enter any parameters for it.

5. Follow the instructions on the screen or in the motherboard manual to save the settings and exit the setup menu.

At this point, the system should boot normally from either a floppy disk or hard disk. If there are any problems, turn to the troubleshooting section of the motherboard manual. When you are sure the system is up and running successfully, power it off and screw the chassis cover securely to the case.

Now your new motherboard should be installed and your system successfully upgraded or repaired.

Preventive Maintenance and Backups

Preventive maintenance is the key to obtaining years of trouble-free service from your computer system. A properly administered preventive maintenance program pays for itself by reducing problem behavior, data loss, component failure, and by ensuring a long life for your system. In several cases, I have "repaired" an ailing system with nothing more than a preventive maintenance session. Preventive maintenance also increases your system's resale value because it will look and run better.

You will also learn the importance of creating backup files of data and the various backup procedures available. A sad reality in the computer repair and servicing world is that hardware can always be repaired or replaced, but data cannot. Most hard disk troubleshooting and service procedures, for example, require that a low-level format be performed. This low-level format overwrites any data on the disk.

Most of the discussion of backing up systems in this chapter is limited to professional solutions that require special hardware and software. Backup solutions that employ floppy disk drives, such as the DOS backup software, are insufficient and too costly in most cases for hard disk backups. It would take 2,867 1.44M floppy disks, for example, to back up the 4G hard disk in my portable system! That would cost more than $1,000 in disks, not to mention the time involved. A DAT (Digital Audio Tape) or Travan tape system, on the other hand, can put 4G to 8G or more on a single $5 to $30 tape.

Although the tape drive itself will cost up to $500 or more, the media costs are really far more significant than the cost of the drive. If you are doing responsible backups, you will have at least three sets of media for each system you are backing up. The media sets would be used on a rotating basis, and one of them should be moved offsite. The media sets should be changed out at an interval of a year or less to prevent excessive wear. If you are backing up more than one system, these media costs will add up quickly.

You should also factor in the cost of time. If a backup requires manual intervention to change the media during the backup, I don't recommend it. A backup system should be able to fit a complete backup on a single tape so that the backup can be performed unattended. If someone has to hang around to switch tapes every so often, the backup becomes a real chore and is less likely to be performed. Also, every time a media change occurs, there is a substantial increase in the likelihood of errors and problems that may not be realized until a restore is performed. A backup is far more important than most people realize, and spending a little more on a quality piece of hardware like a Travan or DAT drive will pay off in the long run with greater reliability, lower media costs, higher performance, and unattended backups that contain the entire system file structure.

Developing a Preventive Maintenance Program

Developing a preventive maintenance program is important to everyone who uses or manages personal computer systems. Two types of preventive maintenance procedures exist--active and passive.

Active preventive maintenance includes steps you apply to a system that promote a longer, trouble-free life. This type of preventive maintenance primarily involves periodic cleaning of the system and its components. This section describes several active preventive maintenance procedures, including cleaning and lubricating all major components, reseating chips and connectors, and reformatting hard disks.

Passive preventive maintenance includes steps you can take to protect a system from the environment, such as using power-protection devices; ensuring a clean, temperature-controlled environment; and preventing excessive vibration. In other words, passive preventive maintenance means treating your system well. This section also describes passive preventive maintenance procedures.

Active Preventive Maintenance Procedures

How often you should implement active preventive maintenance procedures depends on the system's environment and the quality of the system's components. If your system is in a dirty environment, such as a machine shop floor or a gas station service area, you might need to clean your system every three months or less. For normal office environments, cleaning a system every one to two years is usually fine. However, if you open your system after one year and find dust bunnies inside, you should probably shorten the cleaning interval.

Other hard disk preventive maintenance procedures include making periodic backups of critical areas, such as boot sectors, file allocation tables (FATs), and directory structures on the disk. Also defragmenting a hard disk should be performed periodically, normally right after any backups.

Cleaning a System

One of the most important operations in a good preventive maintenance program is regular and thorough cleaning of the system. Dust buildup on the internal components can lead to several problems. One is that the dust acts as a thermal insulator, which prevents proper system cooling. Excessive heat shortens the life of system components and adds to the thermal stress problem caused by wider temperature changes between power-on and power-off states. Additionally, the dust may contain conductive elements that can cause partial short circuits in a system. Other elements in dust and dirt can accelerate corrosion of electrical contacts and cause improper connections. In all, the removal of any layer of dust and debris from within a computer system benefits that system in the long run.

Most non-ATX PC-compatible systems use a forced-air cooling system that allows for even cooling inside the system. A fan is mounted in, on, or near the power supply and pushes air outside. This setup depressurizes the interior of the system relative to the outside air. The lower pressure inside the system causes outside air to be drawn into openings in the system chassis and cover. This draw-through, or depressurization, is the most efficient cooling system that can be designed without an air filter. Air filters typically are not used with depressurization systems because there is no easy way to limit air intake to a single port that can be covered by a filter.

Some industrial computers and ATX systems use a forced-air system that uses the fan to pressurize, rather than to depressurize, the case. This system forces air to exhaust from any holes in the chassis and case or cover. The key to the pressurization system is that all air intake for the system is at a single location--the fan. The air flowing into the system therefore can be filtered by simply integrating a filter assembly into the fan housing. The filter must be cleaned or changed periodically. Because the interior of the case is pressurized relative to the outside air, airborne contaminants are not drawn into the system even though it may not be sealed. Any air entering the system must pass through the fan and filter housing, which removes the contaminants. Pressurization cooling systems are used primarily in industrial computer models designed for extremely harsh environments.

Most systems you have contact with are depressurization systems. Mounting any sort of air filter on these types of systems is impossible because air enters the system from too many sources. With any cooling system in which incoming air is not filtered, dust and other chemical matter in the environment is drawn in and builds up inside the computer. This buildup can cause severe problems if left unchecked.

One problem that can develop is overheating. The buildup of dust acts as a heat insulator, which prevents the system from cooling properly. Some of the components in a modern PC can generate an enormous amount of heat that must be dissipated for the component to function. The dust also might contain chemicals that conduct electricity. These chemicals can cause minor current shorts and create electrical signal paths where none should exist. The chemicals also cause rapid corrosion of cable connectors, socket-installed components, and areas where boards plug into slots. All can cause intermittent system problems and erratic operation.

TIP: Cigarette smoke contains chemicals that can conduct electricity and cause corrosion of computer parts. The smoke residue can infiltrate the entire system, causing corrosion and contamination of electrical contacts and sensitive components such as floppy drive read/write heads and optical drive lens assemblies. You should avoid smoking near computer equipment and encourage your company to develop and enforce a similar policy.

Floppy disk drives are particularly vulnerable to the effects of dirt and dust. Floppy drives are a large "hole" within the system through which air continuously flows. Therefore, they accumulate a large amount of dust and chemical buildup within a short time. Hard disk drives do not present quite the same problem. Because the head disk assembly (HDA) in a hard disk is a sealed unit with a single barometric vent, no dust or dirt can enter without passing through the barometric vent filter. This filter ensures that contaminating dust or particles cannot enter the interior of the HDA. Thus, cleaning a hard disk requires simply blowing the dust and dirt off from outside the drive. No internal cleaning is required.

Disassembly and Cleaning Tools

To properly clean the system and all the boards inside requires certain supplies and tools. In addition to the tools required to disassemble the unit, you should have these items:

You also might want to acquire these optional items:

These simple cleaning tools and chemical solutions will allow you to perform most common preventive maintenance tasks.

Chemicals

You can use several different types of cleaning solutions with computers and electronic assemblies. Most fall into the following categories:

TIP: The makeup of many of the chemicals used for cleaning electronic components has been changing because many of the chemicals originally used are now considered environmentally unsafe. They have been attributed to damaging the earth's ozone layer. Chlorine atoms from chlorofluorocarbons (CFCs) and chlorinated solvents attach themselves to ozone molecules and destroy them. Many of these chemicals are now strictly regulated by federal and international agencies in an effort to preserve the ozone layer. Most of the companies that produce chemicals used for system cleaning and maintenance have had to introduce environmentally safe replacements. The only drawback is that many of these safer chemicals cost more and usually do not work as well as those they replace.

Many specific chemicals are used in cleaning and dusting solutions, but five types are of particular interest. The EPA has classified ozone-damaging chemicals into two classes, Class I and Class II. Chemicals that fall into these two classes have their usage regulated. Other chemicals are nonregulated. Class I chemicals include:

Class I chemicals can only be sold for use in professional service and not to consumers. A law that went into effect on May 15, 1993, requires that the containers for Class I chemicals be labeled with a warning that the product "contains substances that harm public health and the environment by destroying ozone in the atmosphere." Additionally, electronics manufacturers and other industries must also apply a similar warning label to any products that use Class I chemicals in the production process. This means that any circuit board or computer that is manufactured with CFCs will have this label.

The most popular Class I chemicals are the various forms of freon, which are CFCs. A very popular cleaning solution called 1,1,1 Trichloroethane is a chlorinated solvent and also is strictly regulated. Up until 1995, virtually all computer or electronic cleaning solutions contained one or both of these chemicals. While you can still purchase them, regulations and limited production have made them more expensive and more difficult to find.

Class II chemicals include hydrochlorofluorocarbons (HCFCs). These are not as strictly regulated as Class I chemicals because they have a lower ozone depletion potential. Many cleaning solutions have switched to HCFCs because they do not require the restrictive labeling required by Class I chemicals and are not as harmful. Most HCFCs have only one-tenth the ozone damaging potential of CFCs.

Other nonregulated chemicals include volatile organic compounds (VOCs) and hydro- fluorocarbons (HFCs). These chemicals do not damage the ozone layer but actually contribute to ozone production which, unfortunately, appears in the form of smog or ground level pollution. Pure isopropyl alcohol is an example of a VOC that is commonly used in electronic part and contact cleaning. HFCs are used as a replacement for CFCs because the HFCs do not damage the ozone layer.

The Environmental Protection Agency has developed a method to measure the ozone damaging capability of a chemical. The Ozone Depletion Potential (ODP) of a chemical solution is the sum of the depletion potentials of each of the chemicals used in the solution by weight. The ODP of Freon R12 (Automotive Air Conditioning Freon) is 1.0 on this scale. Most modern CFC replacement chemicals have an ODP rating of 0.0 to 0.1, as opposed to those using CFCs and chlorinated solvents that usually have ODP ratings of 0.75 or higher.

Standard Cleaners

Standard cleaning solutions are available in a variety of types and configurations. You can use pure isopropyl alcohol, acetone, freon, trichloroethane, or a variety of other chemicals. Most board manufacturers and service shops are now leaning toward the alcohol, acetone, or other chemicals that do not cause ozone depletion and comply with government regulations and environmental safety.

You should be sure that your cleaning solution is designed to clean computers or electronic assemblies. In most cases, this means that the solution should be chemically pure and free from contaminants or other unwanted substances. You should not, for example, use drugstore rubbing alcohol for cleaning electronic parts or contacts because it is not pure and could contain water or perfumes. The material must be moisture-free and residue-free. The solutions should be in liquid form, not a spray. Sprays can be wasteful, and you almost never spray the solution directly on components. Instead, wet a foam or chamois swab used for wiping the component. These electronic-component cleaning solutions are available at any good electronics parts stores.

Contact Cleaner/Lubricants

These are very similar to the standard cleaners but include a lubricating component. The lubricant eases the force required when plugging and unplugging cables and connectors, which reduces strain on the devices. The lubricant coating also acts as a conductive protectant that insulates the contacts from corrosion. These chemicals can greatly prolong the life of a system by preventing intermittent contacts in the future.

Contact cleaner/lubricants are especially effective on I/O slot connectors, adapter card edge and pin connectors, disk drive connectors, power supply connectors, and virtually any connectors in the PC. Refer to the earlier section "Chemicals" for more information on contact enhancers and lubricants.

Dusters

Compressed gas often is used as an aid in system cleaning. The compressed gas is used as a blower to remove dust and debris from a system or component. Originally, these dusters used CFCs such as freon, while modern dusters now use HFCs or carbon dioxide, neither of which is damaging to the ozone layer. Be careful when you use these devices because some of them can generate a static charge when the compressed gas leaves the nozzle of the can. Be sure that you are using the kind approved for cleaning or dusting off computer equipment, and consider wearing a static grounding strap as a precaution. The type of compressed-air cans used for cleaning camera equipment can sometimes differ from the type used for cleaning static-sensitive computer components.

Related to compressed-air products are chemical-freeze sprays. These sprays are used to quickly cool down a suspected failing component, which often temporarily restores it to operation. These substances are not used to repair a device, but to confirm that you have found a failed device. Often, a component's failure is heat-related and cooling it temporarily restores it to function. If the circuit begins operating normally, the device you are cooling is the suspect device.

Vacuum Cleaners

Some people prefer to use a vacuum cleaner instead of canned gas dusters for cleaning a system. Canned gas is usually better for cleaning in small areas. A vacuum cleaner is more useful when you are cleaning a system loaded with dust and dirt. You can use the vacuum cleaner to suck out dust and debris instead of blowing them on other components, which sometimes happens with canned air. For outbound servicing (when you are going to the location of the equipment instead of the equipment coming to you), canned air is easier to carry in a tool kit than a small vacuum cleaner. There are also tiny vacuum cleaners available for system cleaning. These small units are easy to carry and may serve as an alternative to compressed air cans.

There are special vacuum cleaners specifically designed for using on and around electronic components. They are designed to minimize ESD while in use. If you are using a regular vacuum cleaner and not one specifically designed with ESD protection, then you should take precautions such as wearing a grounding wrist strap. Also be careful if the cleaner has a metal nozzle not to touch it to the circuit boards or components you are cleaning.

Brushes and Swabs

A small brush (makeup, photographic, or paint) can be used to carefully loosen accumulated dirt and dust before spraying with canned air or using the vacuum cleaner. Be careful about generating static electricity. In most cases, the brushes should not be used directly on circuit boards but should be used instead on the case interior and other parts such as fan blades, air vents, and keyboards. Wear a grounded wrist strap if you are brushing on or near any circuit boards, and brush slowly and lightly to prevent static discharges from occurring.

Use cleaning swabs to wipe off electrical contacts and connectors, disk drive heads, and other sensitive areas. The swabs should be made of foam or synthetic chamois material that does not leave lint or dust residue. Unfortunately, proper foam or chamois cleaning swabs are more expensive than the typical cotton swabs. Do not use cotton swabs because they leave cotton fibers on everything they touch. Cotton fibers are conductive in some situations and can remain on drive heads, which can scratch disks. Foam or chamois swabs can be purchased at most electronics supply stores.

One item to avoid is an eraser for cleaning contacts. Many people (including myself) have recommended using a soft pencil-type eraser for cleaning circuit board contacts. Testing has proven this to be bad advice for several reasons. One is that any such abrasive wiping on electrical contacts generates friction and an ESD. This ESD can be damaging to boards and components, especially with newer low-voltage devices made using new technology. These devices are especially static-sensitive, and cleaning the contacts without a proper liquid solution is not recommended. Also, the eraser will wear off the gold coating on many contacts, exposing the tin contact underneath, which will rapidly corrode when exposed to air. Some companies sell premoistened contact cleaning pads that are soaked in a proper contact cleaner and lubricant. These pads are safe to wipe on conductor and contacts with no likelihood of ESD damage or abrasion of the gold plating.

Silicone Lubricants

Silicone lubricants are used to lubricate the door mechanisms on floppy disk drives and any other part of the system that may require clean, non-oily lubrication. Other items you can lubricate are the disk drive head slider rails or even printer-head slider rails, which allow for smooth operation.

Using silicone instead of conventional oils is important because silicone does not gum up and collect dust and other debris. Always use the silicone sparingly. Do not spray it anywhere near the equipment as it tends to migrate and will end up where it doesn't belong (such as on drive heads). Instead, apply a small amount to a toothpick or foam swab and dab the silicone on the components where needed. You can use a lint-free cleaning stick soaked in silicone lubricant to lubricate the metal print-head rails in a printer.

Remember that some of the cleaning operations described in this section might generate a static charge. You may want to use antistatic grounding strap in cases in which static levels are high to ensure that you do not damage any boards as you work with them.

Obtaining Required Tools and Accessories

Most cleaning chemicals and tools can be obtained from a number of electronics supply houses, or even your local Radio Shack. A company called Chemtronics specializes in chemicals for the computer and electronics industry. These and other companies that supply tools, chemicals, and other computer and electronic cleaning supplies are listed in the vendor list in Appendix A. With all these items on hand, you should be equipped for most preventive maintenance operations.

Disassembling and Cleaning Procedures

To properly clean your system, it must be at least partially disassembled. Some people go as far as to remove the motherboard. Removing the motherboard results in the best possible access to other areas of the system; but in the interest of saving time, you probably need to disassemble the system only to where the motherboard is completely visible.

All plug-in adapter cards must be removed, along with the disk drives. Although you can clean the heads of a floppy drive with a cleaning disk without opening the system unit's cover, you probably will want to do more thorough cleaning. In addition to the heads, you also should clean and lubricate the door mechanism and clean any logic boards and connectors on the drive. This procedure usually requires removing the drive.

Next, do the same procedure with a hard disk: Clean the logic boards and connectors, as well as lubricate the grounding strap. To do so, you must remove the hard disk assembly. As a precaution, be sure it is backed up before removal.

Reseating Socketed Chips

A primary preventive maintenance function is to undo the effects of chip creep. As your system heats and cools, it expands and contracts, and the physical expansion and contraction causes components that are plugged into sockets to gradually work their way out of those sockets. This process is called chip creep. To correct its effects, you must find all socketed components in the system and make sure that they are properly reseated.

In most systems, all the memory chips are socketed or are installed in socketed SIMMs or DIMMs. SIMM/DIMM devices are retained securely in their sockets by a positive latching mechanism and cannot creep out. Memory SIPP (Single Inline Pin Package) devices (SIMMs with pins rather than contacts) are not retained by a latching mechanism and therefore can creep out of their sockets. Standard socketed memory chips are prime candidates for chip creep. Most other logic components are soldered in. You can also expect to find the ROM chips, the main processor or CPU, and the math coprocessor in sockets. In most systems, these items are the only components that are socketed; all others are soldered in.

Exceptions, however, might exist. A socketed component in one system might not be socketed in another--even if both are from the same manufacturer. Sometimes this difference results from a parts-availability problem when the boards are manufactured. Rather than halt the assembly line when a part is not available, the manufacturer adds a socket instead of the component. When the component becomes available, it is plugged in and the board is finished. Many newer systems place the CPU in a ZIF socket, which has a lever that can release the grip of the socket on the chip. In most cases, there is very little creep with a ZIF socket.

To make sure that all components are fully seated in their sockets, place your hand on the underside of the board and then apply downward pressure with your thumb (from the top) on the chip to be seated. For larger chips, seat the chip carefully in two movements, and press separately on each end of the chip with your thumb to be sure that the chip is fully seated. (The processor and math coprocessor chips can usually be seated in this manner.) In most cases, you hear a crunching sound as the chip makes its way back into the socket. Because of the great force sometimes required to reseat the chips, this operation is difficult if you do not remove the board.

For motherboards, forcibly seating chips can be dangerous if you do not directly support the board from the underside with your hand. Too much pressure on the board can cause it to bow or bend in the chassis, and the pressure can crack it before seating takes place. The plastic standoffs that separate and hold the board up from the metal chassis are spaced too far apart to properly support the board under this kind of stress. Try this operation only if you can remove and support the board adequately from underneath.

You may be surprised to know that, even if you fully seat each chip, they might need reseating again within a year. The creep usually is noticeable within a year or less.

Cleaning Boards

After reseating any socketed devices that may have crept out of their sockets, the next step is to clean the boards and all connectors in the system. For this step, use the cleaning solutions and the lint-free swabs. First, clean the dust and debris off the board and then clean any connectors on the board. To clean the boards, it is usually best to use a vacuum cleaner designed for electronic assemblies and circuit boards or a duster can of compressed gas. The dusters are especially effective at blasting any dust and dirt off the boards.

Also, blow any dust out of the power supply, especially around the fan intake and exhaust areas. You do not need to disassemble the power supply to do this; just use a duster can and blast the compressed air into the supply through the fan exhaust port. This will blow the dust out of the supply and clean off the fan blades and grill, which will help with system airflow.

CAUTION: Be careful with ESD, which can damage components when cleaning electronic components. Take extra precautions in the dead of winter in an extremely dry, high-static environment. You can apply antistatic sprays and treatments to the work area to reduce the likelihood of ESD damage.

An antistatic grounding wrist strap is recommended. This should be connected to a ground on the card or board you are wiping. This strap ensures that no electrical discharge occurs between you and the board. An alternative method is to keep a finger or thumb on the ground of the motherboard or card as you wipe it off. It is easier to ensure proper grounding while the motherboard is still installed in the chassis, so it is a good idea not to remove it.

Cleaning Connectors and Contacts

Cleaning the connectors and contacts in a system promotes reliable connections between devices. On a motherboard, you will want to clean the slot connectors, power supply connectors, keyboard connector, and speaker connector. For most plug-in cards, you will want to clean the edge connectors that plug into slots on the motherboard as well as any other connectors, such as external ones mounted on the card bracket. Submerge the lint-free swabs in the liquid cleaning solution. If you are using the spray, hold the swab away from the system and spray a small amount on the foam end until the solution starts to drip. Then, use the soaked foam swab to wipe the connectors on the boards. Presoaked wipes are the easiest to use. Simply wipe them along the contacts to remove any accumulated dirt and leave a protective coating behind.

On the motherboard, pay special attention to the slot connectors. Be liberal with the liquid; resoak the foam swab repeatedly, and vigorously clean the connectors. Don't worry if some of the liquid drips on the surface of the motherboard. These solutions are entirely safe for the whole board and will not damage the components.

Use the solution to wash the dirt off the gold contacts in the slot connectors, and then douse any other connectors on the board. Clean the keyboard connector, the grounding positions where screws ground the board to the system chassis, power-supply connectors, speaker connectors, battery connectors, and so on.

If you are cleaning a plug-in board, pay special attention to the edge connector that mates with the slot connector on the motherboard. When people handle plug-in cards, they often touch the gold contacts on these connectors. Touching the gold contacts coats them with oils and debris, which prevents proper contact with the slot connector when the board is installed. Make sure that these gold contacts are free of all finger oils and residue. It is a good idea to use one of the contact cleaners that has a conductive lubricant, which both allows connections to be made with less force and also protects the contacts from corrosion.

You will also want to use the swab and solution to clean the ends of ribbon cables or other types of cables or connectors in a system. Clean the floppy drive cables and connectors, the hard disk cables and connectors, and any others you find. Don't forget to clean the edge connectors that are on the disk drive logic boards, as well as the power connectors to the drives.

Cleaning the Keyboard and Mouse

Keyboards and mice are notorious for picking up dirt and garbage. If you have ever opened up an older keyboard, you often will be amazed at the junk you will find in there. To prevent problems, it is a good idea to periodically clean the keyboard with a vacuum cleaner. An alternative method is to turn the keyboard upside down and shoot it with a can of compressed gas. This will blow out the dirt and debris that has accumulated inside the keyboard and possibly prevent future problems with sticking keys or dirty keyswitches.

If a particular key is stuck or making intermittent contact, you can soak or spray that switch with contact cleaner. The best way to do this is to first remove the keycap and then spray the cleaner into the switch. This usually does not require complete disassembly of the keyboard. Periodic vacuuming or compressed gas cleaning will prevent more serious problems with sticking keys and keyswitches.

Most mice are easily cleaned. In most cases, there is a twist-off locking retainer that keeps the mouse ball retained in the body of the mouse. By removing the retainer, the ball will drop out. After removing the ball, you should clean it with one of the electronic cleaners. I would recommend a pure cleaner instead of a contact cleaner with lubricant because you do not want any lubricant on the mouse ball. Then you should wipe off the rollers in the body of the mouse with the cleaner and some swabs.

Periodic cleaning of a mouse in this manner will eliminate or prevent skipping or erratic movement that can be frustrating. I also recommend a mouse pad for most ball-type mice because the pad will prevent the mouse ball from picking up debris from your desk.

Mice often need frequent cleaning before they start sticking and jumping, which can be frustrating. If you never want to clean a mouse again, I suggest you look into the Honeywell mouse. These mice have a revolutionary new design that uses two external wheels rather than the conventional ball and roller system. The wheels work directly on the desk surface and are unaffected by dirt and dust. Because the body of the mouse is sealed, dirt and dust cannot enter it and gum up the positional sensors. I find this mouse excellent to use with my portable system because it works well on any surface. This mouse is virtually immune to the sticking and jumping that plagues ball and roller designs and never needs to be cleaned, so it is less frustrating than conventional mice.

Other pointing devices requiring little or no maintenance are the IBM designed Trackpoint and similar systems introduced by other manufacturers, such as the Glidepoint by Alps. These devices are totally sealed, and use pressure transducers to control pointer movement. Because they are sealed, cleaning need only be performed externally, and is as simple as wiping the device off with a mild cleaning solution to remove oils and other deposits that have accumulated from handling them.

Hard Disk Maintenance

Certain preventive maintenance procedures protect your data and ensure that your hard disk works efficiently. Some of these procedures actually minimize wear and tear on your drive, which will prolong its life. Additionally, a high level of data protection can be implemented by performing some simple commands periodically. These commands provide methods for backing up (and possibly later restoring) critical areas of the hard disk that, if damaged, would disable access to all your files.

Defragmenting Files

Over time, as you delete and save files to a hard disk, the files become fragmented. This means that they are split into many noncontiguous areas on the disk. One of the best ways to protect both your hard disk and the data on it is to periodically defragment the files on the disk. This serves two purposes. One is that by ensuring that all of the files are stored in contiguous sectors on the disk, head movement and drive wear and tear will be minimized. This has the added benefit of improving the speed at which files will be retrieved from the drive by reducing the head thrashing that occurs every time a fragmented file is accessed.

The second major benefit, and in my estimation the more important of the two, is that in the case of a disaster where the FATs and root directory are severely damaged, the data on the drive can usually be recovered very easily if the files are contiguous. On the other hand, if the files are split up in many pieces across the drive, it is virtually impossible to figure out which pieces belong to which files without an intact FAT and directory system. For the purposes of data integrity and protection, I recommend defragmenting your hard disk drives on a weekly basis, or immediately after you perform any major backup.

Three main functions are found in most defragmenting programs:

Defragmentation is the basic function, but most other programs also add file packing. Packing the files is optional on some programs because it usually takes additional time to perform. This function packs the files at the beginning of the disk so that all free space is consolidated at the end of the disk. This feature minimizes future file fragmentation by eliminating any empty holes on the disk. Because all free space is consolidated into one large area, any new files written to the disk will be able to be written in a contiguous manner with no fragmentation necessary.

The last function, file sorting, is not usually necessary and is performed as an option by many defragmenting programs. This function adds a tremendous amount of time to the operation, and has little or no effect on the speed at which information is accessed. It can be somewhat beneficial for disaster recovery purposes because you will have an idea of which files came before or after other files if a disaster occurs. These benefits would be minimal compared to having the files be contiguous no matter what their order. Not all defragmenting programs offer file sorting, and the extra time it takes is probably not worth any benefits you will receive. Other programs can sort the order that files are listed in directories, which is a quick and easy operation compared to sorting the file ordering the disk.

Checking for Virus Programs

Both Microsoft and IBM now provide standard antivirus software in DOS. The Microsoft Anti-Virus program is actually a reduced function version of the Central Point Anti-Virus software. IBM has written a package called the IBM Anti-Virus program. Many aftermarket utility packages are available that will scan for and remove virus programs. One of the best known is the McAfee Associates' SCAN program, which is also one of the easiest to run because it is a command-line utility. The McAfee program is also distributed through BBS systems and is often site-licensed to large companies. Because Windows 95 does not include an antivirus program, you may want to invest in one of the aftermarket utilities such as the McAfee program.

No matter which of these programs you use, it is a good idea to perform a scan for virus programs periodically, especially before making hard disk backups. This will help to ensure that you catch any potential virus problem before it spreads and becomes a major hassle.

Passive Preventive Maintenance Procedures

Passive preventive maintenance involves taking care of the system in an external manner: basically, providing the best possible environment--both physical as well as electrical--for the system to operate in. Physical concerns are conditions such as ambient temperature, thermal stress from power cycling, dust and smoke contamination, and disturbances such as shock and vibration. Electrical concerns are items such as ESD, power-line noise, and radio-frequency interference. Each of these environmental concerns is discussed in this section.

Examining the Operating Environment

Oddly enough, one of the most overlooked aspects of microcomputer preventive maintenance is protecting the hardware--and the sizable financial investment it represents--from environmental abuse. Computers are relatively forgiving, and they are generally safe in an environment that is comfortable for people. Computers, however, are often treated with no more respect than desktop calculators. The result of this type of abuse is many system failures. Before you acquire a system, prepare a proper location for your new system, free of airborne contaminants such as smoke or other pollution. Do not place your system in front of a window: The system should not be exposed to direct sunlight or temperature variations. The environmental temperature should be as constant as possible. Power should be provided through properly grounded outlets, and should be stable and free from electrical noise and interference. Keep your system away from radio transmitters or other sources of radio frequency energy. This section examines these issues in more detail.

Heating and Cooling

Thermal expansion and contraction from temperature changes place stress on a computer system. Therefore, keeping the temperature in your office or room relatively constant is important to the successful operation of your computer system. Temperature variations can lead to serious problems. You might encounter excessive chip creep, for example. If extreme variations occur over a short period, signal traces on circuit boards can crack and separate, solder joints can break, and contacts in the system undergo accelerated corrosion. Solid-state components such as chips can be damaged also, and a host of other problems can develop.

Temperature variations can play havoc with hard disk drives. Writing to a disk at different ambient temperatures can, on some drives, cause data to be written at different locations relative to the track centers. Read and write problems then might accelerate later.

To ensure that your system operates in the correct ambient temperature, you first must determine your system's specified functional range. Most manufacturers provide data about the correct operating temperature range for their systems. Two temperature specifications might be available, one indicating allowable temperatures during operation and another indicating allowable temperatures under nonoperating conditions. IBM, for example, indicates the following temperature ranges as acceptable for most of its systems:

System on: 60 to 90° Fahrenheit

System off: 50 to 110° Fahrenheit

For the safety of the disk and the data it contains, avoid rapid changes in ambient temperatures. If rapid temperature changes occur--for example, when a new drive is shipped to a location during the winter and then brought indoors--let the drive acclimate to room temperature before turning it on. In extreme cases, condensation forms on the platters inside the drive HDA--disastrous for the drive if you turn it on before the condensation can evaporate. Most drive manufacturers specify a timetable to use as a guide in acclimating a drive to room temperature before operating it. You usually must wait several hours to a day before a drive is ready to use after it has been shipped or stored in a cold environment. They normally advise that you leave the drive in the packing until it is acclimatized. Removing the drive from a shipping carton when extremely cold will increase the likelihood of condensation forming as the drive warms up.

Most office environments provide a stable temperature in which to operate a computer system, but some do not. Be sure to give some consideration to the placement of your equipment.

Power Cycling (On/Off)

As you have just learned, the temperature variations a system encounters greatly stress the system's physical components. The largest temperature variations a system encounters, however, are those that occur during system warm-up when you initially turn it on. Turning on (also called powering on) a cold system subjects it to the greatest possible internal temperature variations. For these reasons, limiting the number of power-on cycles a system is exposed to greatly improves its life and reliability.

If you want a system to have the longest and most trouble-free life possible, you should limit the temperature variations in its environment. You can limit the extreme temperature cycling in two simple ways during a cold startup: Leave the system off all the time or leave it on all the time. Of these two possibilities, of course, you want to choose the latter option. Leaving the power on is the best way I know to promote system reliability. If your only concern is system longevity, the simple recommendation would be to keep the system unit powered on (or off!) continuously. In the real world, however, there are more variables to consider, such as the cost of electricity, the potential fire hazard of unattended running equipment, and other concerns as well.

If you think about the way light bulbs typically fail, you can begin to understand that thermal cycling can be dangerous. Light bulbs burn out most often when you first turn them on, because the filament must endure incredible thermal stresses as it changes temperature--in less than one second--from ambient to several thousands of degrees. A bulb that remains on continuously lasts longer than one that is turned on and off repeatedly.

Where problems can occur immediately at power-on is in the power supply. The startup current draw for the system and for any motor during the first few seconds of operation is very high compared to the normal operating-current draw. Because the current must come from the power supply, the supply has an extremely demanding load to carry for the first few seconds of operation, especially if several disk drives will be started. Motors have an extremely high power-on current draw. This demand often overloads a marginal circuit or component in the supply and causes it to burn or break with a "snap." I have seen several power supplies die the instant a system was powered up. To enable your equipment to have the longest possible life, try to keep the temperature of solid-state components relatively constant, and limit the number of startups on the power supply. The only way I know to do so is to leave the system on.

Although it sounds as though I am recommending that you leave all of your computer equipment on 24 hours a day, seven days a week, I no longer recommend this type of operation. A couple of concerns have tempered my urge to leave everything running continuously. One is that an unattended system that is powered on represents a fire hazard. I have seen monitors start themselves on fire after internally shorting, and systems whose cooling fans have frozen, enabling the power supply and entire system to overheat. I do not leave any system running in an unattended building. Another problem is wasted electrical power. Many companies have adopted austerity programs that involve turning lights and other items off when not in use. The power consumption of some of today's high-powered systems and accessories is not trivial. Also, an unattended operating system is more of a security risk than one that is powered off and locked.

Realities--such as the fire hazard of unattended systems running during night or weekend hours, security problems, and power-consumption issues--might prevent you from leaving your system on all the time. Therefore, you must compromise. Power on the system only one time daily. Don't power the system on and off several times every day. This good advice is often ignored, especially when several users share systems. Each user powers on the system to perform work on the PC and then powers off the system. These systems tend to have many more problems with component failures.

If you are in a building with a programmable thermostat, you have another reason to be concerned about temperatures and disk drives. Some buildings have thermostats programmed to turn off the heat overnight or over the weekend. These thermostats are programmed also to quickly raise the temperature just before business hours every day. In Chicago, for example, outside temperatures in the winter can dip to -20° (not including a wind-chill factor). An office building's interior temperature can drop as low as 50° during the weekend. When you arrive Monday morning, the heat has been on for only an hour or so, but the hard disk platters might have not yet reached even 60° when you turn on the system unit. During the first 20 minutes of operation, the disk platters rapidly rise in temperature to 120° or more. If you have an inexpensive stepper motor hard disk and begin writing to the disk at these low temperatures, you are setting yourself up for trouble. Again, many systems with these "cheap" drives don't even boot properly under these circumstances and must be warmed up before they even boot DOS.

TIP: If you do not leave a system on continuously; at least give it 15 minutes or more to warm up after a cold start before writing to the hard disk. Power up the system and go get a cup of coffee, read the paper, or do some other task. This practice does wonders for the reliability of the data on your disk, especially cheaper units.

If you do leave your system on for long periods of time, make sure that the screen is blank or displays a random image if the system is not in use. The phosphor on the picture tube can burn if a stationary image is left on-screen continuously. Higher- persistence phosphor monochrome screens are most susceptible, and the color displays with low-persistence phosphors are the least susceptible. If you ever have seen a monochrome display with the image of some program permanently burned in--even with the display off--you know what I mean. Look at the monitors that display flight information at the airport--they usually show some of the effects of phosphor burn.

Most modern displays that have power-saving features can automatically enter a standby mode on command by the system. If your system has these power-saving functions, enable them as they will help to reduce energy costs as well as preserve the monitor.

Screen savers or blankers will either blank the screen completely or display some sort of moving random image to prevent burn-in. This can be accomplished by either a manual or automatic procedure as follows:

Screen savers are obsolete in a modern green PC that features power management capabilities. In fact, in these systems, using a screen saver can defeat the power management functions by keeping the hard disk drive and the screen fully powered up at all times. With the built-in power management found in most system BIOS as well as Windows 95, screen savers are more for entertainment than to serve a practical purpose.

Static Electricity

Static electricity can cause numerous problems within a system. The problems usually appear during the winter months when humidity is low or in extremely dry climates where the humidity is low year-round. In these cases, you might need to take special precautions to ensure that the system functions properly.

Static discharges outside a system-unit chassis are rarely a source of permanent problems within the system. The usual effect of a static discharge to the case, keyboard, or even in close proximity to a system is a parity check (memory) error or a locked-up system. In some cases, I have been able to cause parity checks or system lockups by simply walking past a system. Most static-sensitivity problems such as this are caused by improper grounding of the system power. Be sure that you always use a three-prong, grounded power cord plugged into a properly grounded outlet. If you are unsure about the outlet, you can buy an outlet tester at most electronics supply or hardware stores for only a few dollars.

Whenever you open a system unit or handle circuits removed from the system, you must be much more careful with static. You can permanently damage a component with a static discharge if the charge is not routed to a ground. I usually recommend handling boards and adapters first by a grounding point such as the bracket to minimize the potential for static damage.

An easy way to prevent static problems is with good power-line grounding, which is extremely important for computer equipment. A poorly designed power-line grounding system is one of the primary causes of poor computer design. The best way to prevent static damage is to prevent the static charge from getting into the computer in the first place. The chassis ground in a properly designed system serves as a static guard for the computer, which redirects the static charge safely to ground. For this ground to be complete, therefore, the system must be plugged into a properly grounded three-wire outlet.

If the static problem is extreme, you can resort to other measures. One is to use a grounded static mat underneath the computer. Touch the mat first before you touch the computer to ensure that any static charges are routed to ground and away from the system unit's internal parts. If problems still persist, you might want to check out the electrical building ground. I have seen installations in which three-wire outlets exist but are not grounded properly. You can use an outlet tester to be sure that the outlet is wired properly.

Power-Line Noise

To run properly, a computer system requires a steady supply of clean noise-free power. In some installations, however, the power line serving the computer serves heavy equipment also, and the voltage variations resulting from the on-off cycling of this equipment can cause problems for the computer. Certain types of equipment on the same power line also can cause voltage spikes--short transient signals of sometimes 1,000v or more--that can physically damage a computer. Although these spikes are rare, they can be crippling. Even a dedicated electrical circuit used only by a single computer can experience spikes and transients, depending on the quality of the power supplied to the building or circuit. During the site-preparation phase of a system installation, you should be aware of these factors to ensure a steady supply of clean power:

Air conditioners, coffee makers, copy machines, laser printers, space heaters, vacuum cleaners, and power tools are some of the worst corrupters of a PC system's power. Any of these items can draw an excessive amount of current and play havoc with a PC system on the same electrical circuit. I've seen offices in which all the computers begin to crash at about 9:05 A.M. daily, which is when all the coffee makers are turned on!

Also, try to ensure that copy machines and laser printers do not share a circuit with other computer equipment. These devices draw a large amount of power.

Another major problem in some companies is partitioned offices. Many of these partitions are prewired with their own electrical outlets and are plugged into one another in a sort of power-line daisy chain, similar to chaining power strips together. I pity the person in the cubicle at the end of the electrical daisy chain, who will have very flaky power!

As a real-world example of too many devices sharing a single circuit, I can describe several instances in which a personal computer had a repeating parity-check problem. All efforts to repair the system had been unsuccessful. The reported error locations from the parity-check message also were inconsistent, which normally indicates a problem with power. The problem could have been the power supply in the system unit or the external power supplied from the wall outlet. This problem was solved one day as I stood watching the system. The parity-check message was displayed at the same instant someone two cubicles away turned on a copy machine. Placing the computers on a separate line solved the problem.

By following the guidelines in this section, you can create the proper power environment for your systems and help to ensure trouble-free operation.

Radio-Frequency Interference

Radio-frequency interference (RFI) is easily overlooked as a problem factor. The interference is caused by any source of radio transmissions near a computer system. Living next door to a 50,000-watt commercial radio station is one sure way to get RFI problems, but less powerful transmitters cause problems, too. I know of many instances in which portable radio-telephones have caused sporadic random keystrokes to appear, as though an invisible entity were typing on the keyboard. I also have seen RFI cause a system to lock up. Solutions to RFI problems are more difficult to state because every case must be handled differently. Sometimes, reorienting a system unit eliminates the problem because radio signals can be directional in nature. At other times, you must invest in specially shielded cables for cables outside the system unit, such as the keyboard cable.

One type of solution to an RFI noise problem with cables is to pass the cable through a toroidal iron core, a doughnut-shaped piece of iron placed around a cable to suppress both the reception and transmission of electromagnetic interference (EMI). If you can isolate an RFI noise problem in a particular cable, you often can solve the problem by passing the cable through a toroidal core. Because the cable must pass through the center hole of the core, it often is difficult, if not impossible, to add a toroid to a cable that already has end connectors installed.

Radio Shack sells a special snap-together toroid designed specifically to be added to cables already in use. This toroid looks like a thick-walled tube that has been sliced in half. You just lay the cable in the center of one of the halves, and snap the other half over the first. This type of construction makes it easy to add the noise-suppression features of a toroid to virtually any existing cable.

IBM also makes a special 6-foot long PS/2 keyboard cable with a built-in toroid core (part number 27F4984) that can greatly reduce interference problems. This cable has the smaller six-pin DIN (PS/2 style) connector at the system end and the standard SDL (Shielded Data Link) connector at the keyboard end; it costs about $40.

The best, if not the easiest, way to eliminate the problem probably is to correct it at the source. You likely won't convince the commercial radio station near your office to shut down, but if you are dealing with a small radio transmitter that is generating RFI, sometimes you can add to the transmitter a filter that suppresses spurious emissions. Unfortunately, problems sometimes persist until the transmitter is either switched off or moved some distance away from the affected computer.

Dust and Pollutants

Dirt, smoke, dust, and other pollutants are bad for your system. The power-supply fan carries airborne particles through your system, and they collect inside. If your system is used in an extremely harsh environment, you might want to investigate some of the industrial systems on the market designed for harsh conditions. IBM used to sell industrial-model XT and AT systems but discontinued them after introducing the PS/2. IBM has licensed several third-party companies to produce industrial versions of PS/2 systems.

Compatible vendors also have industrial systems; many companies make special hardened versions of their systems for harsh environments. Industrial systems usually use a different cooling system from the one used in a regular PC. A large cooling fan is used to pressurize the case rather than depressurize it, as most systems do. The air pumped into the case passes through a filter unit that must be cleaned and changed periodically. The system is pressurized so that no contaminated air can flow into it; air flows only outward. The only way air can enter is through the fan and filter system.

These systems also might have special keyboards impervious to liquids and dirt. Some flat-membrane keyboards are difficult to type on, but are extremely rugged; others resemble the standard types of keyboards but have a thin, plastic membrane that covers all the keys. You can add this membrane to normal types of keyboards to seal them from the environment.

A new breed of humidifier can cause problems with computer equipment. This type of humidifier uses ultrasonics to generate a mist of water sprayed into the air. The extra humidity helps cure problems with static electricity resulting from a dry climate, but the airborne water contaminants can cause many problems. If you use one of these systems, you might notice a white ash-like deposit forming on components. The deposit is the result of abrasive and corrosive minerals suspended in the vaporized water. If these deposits collect on the system components, they can cause all kinds of problems. The only safe way to run one of these ultrasonic humidifiers is with pure distilled water. If you use a humidifier, be sure it does not generate these deposits.

If you do your best to keep the environment for your computer equipment clean, your system will run better and last longer. Also, you will not have to open your unit as often for complete preventive maintenance cleaning.

Previous chapterNext chapterContents


Macmillan Computer Publishing USA

© Copyright, Macmillan Computer Publishing. All rights reserved.