Network+ Online Course
Lesson 3 - The Physical Network
B. Cable types
- Coaxial cable
- Twisted pair
- Fiber optic cable
Each network allows some flexibility in choosing cable. For example, Ethernet supports
coaxial cable, UTP or fiber optic. Token Ring , on the other hand, supports UTP, STP and
fiber optic cable. It is up to the network designer to decide which of these cables to choose.
Many factors affect this decision including the length of the cable runs, throughput of the
network, possible interference and security needs. In the following pages, you will examine
the cable types along these parameters and at the end, decide for yourself which cable type
is appropriate in various situations.
Coaxial cable was the first major cable used for Local Area Networks. The cable was already
extensively used in mainframe systems and its characteristics made it ideal for early LANs.
It has now fallen out of favor and is being replaced with cable that is more flexible and
less expensive, such as unshielded twisted pair or by cable which can offer greater distances
or more capacity such as fiber optic cable.
Figure 19: Coaxial cable
Coaxial cable is composed of a center conductor, either solid or stranded, copper or maybe
aluminum. Surrounding it is an insulation, called the dielectric. It is normally foam. Around
the dielectric is a copper wire mesh or aluminum foil, which is the second conductor. This
layer also give the wire some protection from interference. Finally, there is a covering
called the cladding. It may be made from PVC (PolyVinyl Chloride) or Teflon.
Notice that the two conductors run parallel to each other or on the same axis, hence the
name “coaxial” cable.
Coaxial cable varieties
Ethernet has two kinds of coaxial cable available. Thinnet is less expensive, hence its
nickname “cheapernet”, easier to work with and install. Thicknet has a larger diameter, is
very expensive and hard to work with. However, it has much greater resistance to interference
and can extend to a 500 meter distance. Thinnet has a maximum length of 185 M.
Pros and cons
Coaxial cable can go longer distances and has better protection from EMI than twisted pair.
On the negative side, it is harder to work with and does not lend itself to the modern
practice of running cable from every office to a central location, usually the wiring closet.
Nevertheless, when you need to go just that extra distance, coax may solve the problem and it
does a good job of connecting networks on different floors together.
Twisted Pair Cable
- Standard cable used for LANS
- Various grades, called categories, determine bandwidth
- Shielded twisted pair no longer common
- Conductors must be placed in correct order by colour
Twisted Pair cable
The most popular network cable is twisted pair. It developed from the cable used in telephone
systems. That cable, however, wasn’t twisted, nor did it need to be because voice signals
traveled along it fine. When early developers attempted to put digital signals on it, it was
deemed a failure because of the crosstalk. Twisting the conductors around each other
alleviated the crosstalk and the modern twisted pair data cable was born. Early unshielded
twisted pair (UTP) cable was restricted to low throughput networks such as ARCnet and 4Mbps
Token Ring. There was despair that high speed networks could ever use UTP. With advances in
quality construction and by increasing the number of twists per foot, UTP can now successfully
be used for networks with throughput as high as 1Gbps.
Figure 20: UTP cable
Shielded or unshielded
Each conductor is composed of a copper core, either solid or stranded, surrounded by an
insulating sheath. Two of these conductors are twisted about each other to form a twisted pair.
A varied number of pairs are bundled together to make the final cable. Modern LANs have
standardized on a four pair cable.
Twisted pair cable is available in two varieties, shielded or unshielded. Shielded twisted
pair (STP) cable has a shield made of copper braid, aluminum braid or aluminum foil
surrounding the inner conductors. This gives it much greater resistance to electrical
interference than unshielded twisted pair (UTP) cable. STP can go three times the distance
of UTP because of this characteristic. Unfortunately, STP, like coaxial cable, is much
heavier than UTP and is harder to work with. More to the point, it has a difference impedance
than UTP and therefore cannot be substituted without making special provisions. Ethernet
uses UTP while Token Ring can use either.
Plenum or patch cable
Twisted pair specifications
The plenum is the space above a dropped down ceiling in a modern building. It is the natural
space in which to run cabling between rooms. UTP plenum cable is stiff and holds its shape
when bent. The construction of the cable is the reason. It uses a solid, copper core for
each conductor. This type of cable would break if handled excessively, which is the case for
patch cables. These are used to connect the computer to the wall plates in a room and are
expected to be flexed many times in their life. Patch cables use stranded copper wire as their
conductors. Therefore always use the correct cabling for the job: patch cables to attach
computers to the wall, plenum cable for the interior of walls and ceilings.
Originally, there was no standard for twisted pair cables used for data transmission.
However, when IBM designed their Token Ring network, they also created specifications for
the cable which would be used. Enough cables were specified so that all important wiring
conditions would be met. These included shielded and non-shielded cable, patch cables and
cables for both the plenum (in the ceiling) as well as under carpet. These cable types are
summarized in the following table. Subsequently, the Electrical
Industry Association and the Telephone Industry Association (EIA/TIA) codified the
specifications for UTP into categories called the EIA/TIA 568 Commercial Building Wiring
Standard. The most modern practice is to wire buildings with category 5 (Cat 5e) UTP cable
which is suitable for transmission up to 100Mbps.
American Wire Gauge (AWG)
When reading specifications, you will often run across the term AWG. This represents the
thickness of the metal conductor. Smaller numbers represent thicker wire. Telephone wire is
typically 22 AWG, 14 is thicker, 26 is thinner.
||Telephone, RS-232C, only two pair
||less than 1 Mbps
IBM Type 3
||ARCnet, 4 Mbps Token Ring
||less than 10 Mbps
||Ethernet/802.3, 10BaseT 100Base4
||16 Mbps Token Ring, 802.5
|Category 5, 5e
||CDDI, FastEthernet, ATM
|Shielded Twisted Pair
IBM Type 1
IBM Type 2
IBM Type 6
IBM Type 8
IBM Type 9
||Token Ring, 802.5
2 pairs shielded
2 pairs shielded, 2 pairs not shielded (data+voice)
Data patch cable
Under carpet grade
Pros and cons
Unshielded twisted pair cable has become the most popular LAN cabling because it is
inexpensive and it is easy to install. It lends itself to modern topologies in which
offices are pre-wired and the cable is terminated in a central location, the wiring closet.
Because it can be used by both Ethernet and Token Ring, it is common practice to install
CAT5 wiring during construction without worrying about which LAN will eventually be used.
UTP cable segments are limited to 100 meters or about 328 feet. Since it is estimated that
90% of all cable runs fall within this distance, UTP is sufficient. If however, a longer
cable segment is required, other cable must be used. With Ethernet, coaxial cable or fiber
optic cable can be used. With Token Ring, shielded twisted pair or fiber optic cable is an
option. UTP is also the most prone to interference and is the easiest to tap into. If these
are concerns, fiber optic cable may be the best choice although shielded cable may be
considered. Because of the different categories of UTP, it may be possible to have the wrong
kind, necessitating an expensive retrofitting of the cable system. An organization with 10
Mbps Ethernet running on CAT3 cable may desire to upgrade to fast Ethernet at 100Mbps. Since
this requires CAT5 cable, the old cabling will need to be replaced. Since 1995, CAT5 cable
has enhanced performance characteristics, hence the CAT5e designation.
Fiber Optic Cable
- Uses light instead of electricity to transmit signals
- Fiber is glass or plastic
- Light source is laser or LEDs
- Can transmit huge distances compared to copper cable
- No EMI, hard to intercept
A data transmission medium consisting of glass or plastic fibers. Light-emitting diodes (LEDs)
or lasers send light through the fiber to a detector, which then converts the light back into
electrical signals. Fiber-optics offer immense bandwidth, as well as protection from
eavesdropping, electromagnetic interference, and radioactivity. Fiber-optic cabling employs
one or more optical fibers, consisting of a central glass or plastic fiber, glass cladding,
and a plastic outer sheath. Fiber-optic cable carries information as light instead of
electricity, and can carry much more information over greater distances than copper cabling.
Figure 21: The structure of fiber optic cable
Fiber can transmit light a long distance because it keeps it in a channel. This is
accomplished by surrounding the inner fiber with a cladding that has a different refractive
index. This keeps the light bouncing along from side to side until the end.
Single mode cable only has one light beam in it. The cable has a small diameter core and is
made of the purest glass. A laser is usually the light source and the cable can extend
several kilometers without a repeater.
Multimode cable has greater capacity because several light beams are sent down it. The beams
enter at slightly different angles and therefore don’t interfere with each other. The core is
wider than used in single-mode and often made of fine plastic. Instead of lasers, the light
source is usually light emitting diodes (LEDs). The system has greater capacity and is less
expensive than single-mode. On the other hand, multimode doesn’t have the same reach, often
restricted to hundreds of meters instead of kilometers. For this reason, multimode is
usually used inside buildings, whereas single-mode is used between buildings and long
distances. If the cable needs to be extended, optical repeaters are used.
The conductors in the cable only carry signals in one direction, therefore two conductors are
used, each carrying signals in the opposite direction. The two cables are connected together
with Kevlar webbing.
Pros and cons
Fiber optic cable has the longest range and should be used when long distance is a
requirement. It has extremely high bandwidth, in the Gbps range. It is also immune to
electrical interference, nor does it give off any of its own. This would be a factor in
dangerous environments where an electrical spark could set off an explosion, say in an
explosives factory. Finally, it is extremely difficult, but not impossible, to tap into a
fiber optic system.
On the negative side, fiber optic cable is the most expensive to install. The cable and
connectors are more expensive plus the installation requires highly skilled craftsmen. This
is particularly true when two fibers need to be spliced together.
Connectors used for fiber cable are SC connectors which are a push and pull connector and ST
connectors which are a bayonet mount.
Exercise 3-3: Cable exercise
Exercise: Which cable?
Your are the network architect and need to specify the most appropriate cable. For the
following situations, which cable would you recommend?
1. Cable for an explosive’s factory
2. A network which will be running Ethernet at 100Mbps
3. Cable for a building which is being used by a national security agency
4. A cable must be run between two buildings facing each other across a courtyard. The
distance between the wiring closet in the basement of one building to the wiring closet in
the basement of the other building is 150M. The network is a 10Mbps Ethernet.
5. Cable which must connect switches in the wiring closets of each floor of a 10 story
building. The cable will be routed down the elevator shaft.
6. Wiring for a test network that will be using Ethernet running at 1,000Mbps.
7. Cable that will run between a signal controller and the control room in a locomotive
switching yard. The distance is 2km.
[Top of page][On to next section]