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 |