Your Brain is Hungry.
InformIT
ProgramITDatabaseITWebITNetworkITConfigureIT


Enter search string and press enter!

You are here : Home : Upgrading & Repairing PCs, Eighth Edition

Upgrading & Repairing PCs, Eighth Edition

Add To MyInformIT

 

< Back Contents Next >

Author: Scott Mueller
Retail Price: $49.99
Publisher: Que
ISBN: 0789712954
Publication Date: 9/16/97
Pages: 1168


Chapter 3 - System Teardown and Inspection

Best-selling author Scott Mueller tells you all you need to know about disassembling and inspecting your PC hardware - the tools you need, preventive maintenance, right down to the nuts and bolts!

 
1x1blue.gif (43 bytes)
 

 

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:

  • Simple hand tools for basic disassembly and reassembly procedures, including a flat blade and Phillips screwdrivers (both medium and small sizes), tweezers, an IC extraction tool, and a parts grabber or hemostats

  • Diagnostics software and hardware for testing components in a system

  • A multimeter that allows accurate measurement of voltage and resistance

  • Chemicals, such as contact cleaners, component freeze sprays, and compressed air for cleaning the system

  • Foam swabs, or cotton swabs if foam isn't available

  • Small wire ties for "dressing" or organizing wires

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

  • Memory testing machines, which are used to evaluate the operation of SIMMs (Single Inline Memory Modules), DIP (Dual Inline Pin) chips, and other memory modules

  • Serial and parallel wrap plugs to test serial and parallel ports

  • A network cable scanner, if a network is used

  • A serial breakout box if a lot of the systems operate over serial cables, such as UNIX dumb terminals

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:

  • 3/16-inch nut driver

  • Chip extractor

  • Chip inserter

  • Tweezers

  • Claw-type parts grabber

  • T10 and T15 Torx drivers

  • 1/4-inch nut driver

  • Small Phillips screwdriver

  • Small flat-blade screwdriver

  • Medium Phillips screwdriver

  • Medium flat-blade screwdriver


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:

  • Needlenose pliers

  • Vise or clamp

  • Small flashlight

  • File

  • Hemostats

  • Wire cutter or wire stripper

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:

  • Identify the type of memory

  • Identify the memory speed

  • Identify whether the memory has parity or is using bogus parity emulation

  • Vary the refresh timing and access speed timing

  • Locate single-bit failures

  • Detect power and noise-related failures

  • Detect solder opens and shorts

  • Isolate timing-related failures

  • Detect data-retention errors

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 ej jected 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:

  • Case or cover assembly

  • Power supply

  • Motherboard

  • Adapter boards

  • Disk drives

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 s