Windows Knowledge Base

[NOC Tutor] Troubleshooting Cabling Problems

Cabling problems account for a substantial number of network-support calls; some authorities
say as many as 40 to 50 percent. Whether or not the figure is accurate, any network administrator
will nevertheless experience network-communication problems that can be attributed to
no other cause than the network cabling. The type of cable your network uses and how it is
installed will have a big effect on the frequency and severity of cabling problems.
For example, a coaxial thin Ethernet network allowed to run wild on floors and behind furniture
is far more likely to experience problems than a 10Base-T network installed inside the

walls and ceilings. This is true not only because the coaxial cables are exposed and more liable
to be damaged but also because the bus topology is more sensitive to faults and the BNC connectors
are more easily loosened. Once cabling is installed in the walls and verified for performance,
there is very little that will go wrong with it. This goes to show that you can take steps
toward minimizing the potential for cable problems by selecting the right products and installing
them properly.

Establishing a Baseline

The symptoms of many cable problems are similar to symptoms of software problems, so it can
often be difficult to determine when the cable causes a problem. The first step in simplifying
the isolation of the source of network problems is to make sure that all of your cables are functioning
properly at the outset. You do this by testing all of your cable runs as you install them,
as described earlier in this chapter, and by documenting your network installation.
If you use a multifunction cable tester, you can usually store the results of your tests by retaining
them in the tester’s memory, copying them to a PC, or printing them out. You thus establish
a performance baseline against which you can compare future test results. For example, by
recording the lengths of all your cable runs at the time of the installation, you can tell if a cable
break has occurred later by retesting and seeing if the length results are different than before.
In the same way, you can compare the levels of crosstalk, outside noise, and other characteristics
that may have changed since the cable was installed. Even if your tester does not have these
data-storage features, you should manually record the results for future reference.
Another good idea is to create and maintain a map of all your cable runs on a floor plan of
your site. Sometimes cable problems can be the result of outside factors, such as interference
from electrical equipment in the building. A problem that affects multiple cable runs might be
traced to a particular location where new equipment was installed or existing equipment modified.
When you install your cables inside walls and ceilings (and especially when outside contractors
do it for you), it can be difficult to pinpoint the routes that individual cables take. A
map serves as a permanent record of your installation, both for yourself and any future people
working on the network.

Locating the Problem

Troubleshooting your network’s cable plant requires many of the same common-sense skills as
other troubleshooting. You try to isolate the cause of the problem by asking questions like the
following:
- Has the cable ever worked properly?
- When did the malfunctions start?
- Do the malfunctions occur at specific times?
- What has changed since the cable functioned properly?

Once you’ve gathered all the answers to such questions, the troubleshooting consists of steps
like the following:

1. Split the system into its logical elements.
2. Locate the element that is most likely the cause of the problem.
3. Test the element or install a substitute to verify it as the cause of the problem.
4. If the suspected element is not the cause, move on to the next likely element.
5. After locating the cause of the problem, repair or replace it.

You might begin troubleshooting by determining for sure that the cable run is the source of
the problem. You can do this by connecting different devices to both ends of the cable to see
if the problem continues to occur. Once you verify that the cable is at fault, you can logically
break it down into its component elements. For example, a typical cable run might consist of
two patch cables (one at each end), a wall plate, a punch-down block, and the permanently
installed cable.

In this type of installation, it is easiest to test the patch cables, either by replacing them or
testing them with a cable scanner. Replacing components can be a good troubleshooting
method, as long as you know that the replacements are good. If, for example, you purchase a
box of 100 cheap patch cables that are labeled Category 5e when they actually are Category 3
cable, replacing one with another won’t do any good.
The most accurate method is to test the individual components with a cable scanner. If the
patch cables pass, then proceed to test the permanent link. If you don’t have a scanner available,
you can examine the connectors at either end of the cable run and even reconnect or replace
them to verify that they were installed correctly. However, there’s little you can do if the problem
is inside a wall or in some other inaccessible place. If you do have a scanner, the results of
the tests should provide you with the information you need to proceed.

Resolving Specific Problems

Cable testers, no matter how elaborate, can’t tell you what to do to resolve the problems they
disclose. The following sections examine some of the courses of action you can take to address
the most common cabling problems.

Wire-Map Faults
Wire-map faults are the result of an improper installation. When the wires within a
twisted-pair cable are attached to the wrong pins, the cable is no longer wired straight
through. If the pairs used to carry network data are involved, then signals won’t reach their
destination. In most cases, this fault occurs on a permanent link, although it is possible for
a patch cable to be miswired.
The possible causes of wire-map faults are simple errors made during the installation or the
use of different pinouts (T568-A and T568-B) at each end of the cable. Whatever the cause,
however, the remedy is to rewire the connectors on one or both ends so that each pin at one
end is connected to its equivalent pin at the other end.

Excessive Length
Cable lengths should be carefully planned before network installation and tested immediately
after installation to make sure that the cables are not longer than the recommended maximum
according to the ANSI/TIA/EIA-568-B Standard. Cable runs that are too long can cause problems
like late collisions on an Ethernet network or excessive retransmissions due to attenuated
signals. Most protocols have some leeway built into them that permit a little excess, so don’t be
overly concerned if the maximum allowable length for a cable segment is 95 meters and you
have one run that is 96 meters long.

To address the problem, you can start by using shorter patch cables, if possible. In some
cases, you may find that an installer has left extra cable coiled in a ceiling or wall space that can
be removed, and the end can be reconnected to the wall plate or punch-down block. Sometimes
a more efficient cable route can enable you to rewire the run using less cable. If, however,
you find that bad planning caused the problem and the wall plate is too far away from the
punch-down block, you can still take actions.
The first and easiest action is to test the attenuation and NEXT on the cable run to see if they
exceed the requirements for the protocol. These characteristics are the primary reasons for these maximum-length specifications. If you have installed cable that is of a higher quality than is required, you may be able to get away with the additional length, but if you are having network
problems, chances are this isn’t the answer.

Opens and Shorts
Opens and shorts can be caused by improper installation, or they can occur later if a cable is
damaged or severed. If the cable’s length is correct but one or more wires are open or shorted,
then a connector is likely faulty or has come loose and needs repairing or replacing. If all of the
wires in a cable are reported as open in the same place or if the length of all the wires is suddenly
shorter than it should be, the cable may have been accidentally cut by nearby equipment or by
someone working in the area. Cables damaged but not completely severed may show up with
drastically different lengths for the wire pairs or as shorts at some interim point.
Cable scanners usually display the distance to the open or short so that you can more easily
locate and repair it. For cables installed in walls and ceilings, the cable map you (we hope) created
during the installation can come in handy. If you don’t know the cable’s route, you can use
a tone generator and probe to trace the cable to the point of the break.
It is tempting to try to splice the ends of the severed wires. Don’t do it. You’ll create a nexus
for all sorts of potential transmission problems, including SRL (structural return loss) reflections,
higher attenuation, and increased crosstalk. You must completely replace the permanent
cable run. Broken or damaged patch cables should always be discarded.

Excessive Attenuation
A cable run can exhibit excessive attenuation for several different reasons, most of which are
attributable to improper installation practices. The most obvious cause is excessive length. The
longer the cable, the more the signals attenuate. Address this problem as you would any other

excessive-length condition.
Another possible cause is that the cable used in the run is not suitable for the rate at which
data will be transmitted. If, for example, you try to run a 100Base-TX network using Category 3
cable, one of the reasons it will fail is that the specified attenuation level for Category 3 allows
the signal to decay more than 100Base-TX can handle. In this case, there is no other alternative
than to replace the cable with the proper grade. Inferior or untwisted patch cables are a frequent
cause of this type of problem. These are easily replaced, but if your permanent links are
not of an appropriate performance grade, the only alternative is to replace the cabling.
Excessive attenuation can also be caused by other components that are of an inferior grade,
such as connectors or punch-down blocks. Fortunately, these are generally easier to replace
than the entire cable.Environmental factors, such as a conductor stretched during installation or a high-heat environment, also cause excessive attenuation.

Excessive Crosstalk

Crosstalk is a major problem that can have many different causes, including the following:
Inferior cable Cables not of the grade required for a protocol can produce excessive
crosstalk levels. The only solution is to replace the cable with the appropriate grade.
Inferior components All the components of a cable run should be rated at the same grade,
including all connectors. Using Category 3 connectors on a Category 5e network can introduce
excessive crosstalk and other problems. Replace inferior components with those of the
correct grade.
Improper patch cables Replace inappropriate cables with twisted-pair patch cables that
are rated the same as your permanent links. Silver-satin patch cables used for telephone systems
may appear at first to work with data connections, but the wire pairs in these cables are
not twisted, and the main reason for twisting conductors together in pairs is to minimize
crosstalk.
Split pairs Incorrect pinouts that cause data-carrying wires to be twisted together result in
additional crosstalk, even when both ends are wired in the same way. Split pairs can be the
result of mistakes during the installation or the use of the USOC pinouts. The solution is to
reattach the connectors at both ends using either the T568-A or T568-B pinouts.
Couplers Using couplers to join short lengths of cable generates more crosstalk than using
a single cable segment of the appropriate length. Use one 12-foot patch cable (for example)
instead of two 6-foot cables joined with a coupler. When repairing broken permanent links,
pull a new length of cable rather than using couplers to join the broken ends together.
Twisting The individual wire pairs of every Category 5e cable must remain twisted up to
a point no farther than 1/2 (0.5) inches from any connector. A Category 6 pair must remain
twisted to within 3/8 (0.375) inches of its termination. If the wires are too loosely twisted, reattach
the connectors, making sure that all of the wire pairs are twisted tightly.
Sharing cables Many network protocols use only two of the four wire pairs in a standard
twisted-pair cable, so some people believe they can utilize the other two pairs for voice traffic
or some other application. They shouldn’t, however, because other signals running over the
same cable can produce crosstalk. The problem may be difficult to diagnose in these cases
because the crosstalk only occurs when the other application is using the other wire pairs,
such as when the user is talking on the phone. If two pairs in a wire are used for another application,
you must install new cabling for one application or the other so that they will no
longer share a cable.

Excessive Noise

The potential for noise generated by outside sources should be considered during the planning
phase of a network installation. Cables should be routed away from AC power lines, light fixtures,
electric motors, and other sources of EMI and RFI. Sometimes outside noise sources can
be difficult to detect. You may, for example, test your cables immediately after you install them
and detect no excess noise from outside sources and then find during later testing that your network
performance is severely degraded by noise. It is entirely possible that a new source of
interference has been introduced into the environment, but you also have to consider that your
original tests may not have been valid.
If you installed and tested the cable plant during nights and weekends, your tests for outside
noise may have generated all pass ratings because some sources of noise were not operating.
When lights and machinery are turned on Monday morning, noise levels could be excessive.
Always test your cable runs in the actual environmental conditions in which they’ll be used.
If, after cable installation, a new source generates excessive noise levels, you must either move
the cables and source away from each other or replace the UTP cables with fiber-optic cables.