Naval Weapons Engineering
When multiple signals are sent over a single transmission channel, the process
that keeps the signals from interfering with one another is called multiplexing.
Most communications systems are capable of transmitting and receiving. These are
two independent signals are therefore must be multiplexed. There are three possible
combinations of transmit and receive capability:
Simplex: a receive only system. A television set is a simplex system.
Half-duplex: can receive and transmit but not at the same time. Requires only one
individual channel. Hand-held radios are half-duplex. You must push a button to
talk, let go to receive. You can't do both at the same time.
Full-duplex: can receive and transmit at the same time. A telephone is an example
of full duplex.
The next logical step is to send multiple independent communications
signals over the same transmission medium. For example, thousands of telephone
conversations can be carried on the same wire. There are several methods of
Frequency Domain Multiplexing
In the chapter on analog modulation, the information signal was used to
modulate a carrier wave. In frequency domain multiplexing, each signal is given
a unique carrier frequency. These frequencies must be chosen so that adjacent
signals do not overlap. Therefore they must be separated by a frequency interval
equal at least to the signal bandwidth. For example, commercial AM radio stations
are separated by 10 kHz, which is the signal bandwidth. The multiplexing capacity,
which is the maximum number of independent signals that can be transmitted
simultaneously, is limited by both the bandwidth of the signals and the available
For broadcasts in air, the available bandwidth is determined by the
allocated frequency band. For example, AM radio signals must be in the 535-1605
kHz band. Therefore there are 107 possible carrier frequencies that may be
assigned, at 10 kHz intervals. Likewise FM radio stations are at 0.2 MHz intervals
in the range of 88-108 MHz allowing 100 stations in a local area. For any system
using frequency domain multiplexing the channel capacity may be determined by
channel capacity = available bandwidth/signal bandwidth.
The available bandwidth may also be a physical limitation of the medium in
which your are transmitting. Coaxial cable cannot transmit frequencies above 1 GHz
without significant losses. Therefore coaxial cable has an available bandwidth from
0 to 1 GHz, and therefore can carry 167 separate TV signals at 6 MHz each over a
Time Domain Multiplexing
In this method, each signal can occupy the entire bandwidth, but is given
a time slot of limited duration, which is available at regular intervals. This is
a common method for digital signals, so it is natural to talk about the digital
rate of transmission, measured in bits per second or (baud rate). Theoretically
there is no maximum number of signals that can be multiplexed, but there is a
limit on the overall transmission rate (i.e. the capacity in bits per second as
determined by the Hartley-Shannon law). Suppose we have a digital transmission
channel with a capacity of 100 Mbps. If there is only one user, they could
occupy the entire channel and transmit their signal at 100 Mbps. If there are
20 users, each will only have transmission rate of 5 Mbps.
Some systems have predetermined time slots which sets a fixed rate for
any user. If the capacity is not filled, the other time slots go empty. This
is the case for synchronous systems which rely on a fixed timing schedule.
Asynchronous systems (like ATM) can dynamically allocate capacity as the demand
varies. ATM systems have a much better performance when demand is low, because
the few users have the entire bandwidth available. However at maximum capacity,
the synchronous system is faster because less space is wasted framing the data
(see discussion of input/output in the chapter on computer basics).
Spread Spectrum Multiplexing
This is a special technique in which every signal occupies the full
bandwidth of the channel simultaneously. The signals are intentionally spread
out until their bandwidth matches the channel. Signals are distinguished from
one another by a special code which is applied when the signal is spread out.
The signals are increased in bandwidth by mixing them with a pseudo-random code,
which is a noise-like signal that repeats a unique code over a very long period.
There are two main advantages to spread spectrum multiplexing:
It reduces the effect of interference. Recall that interference is narrowband,
meaning strong association with a particular frequency. By spreading out
the power of the signal over a large range of frequency, the effect of
noise at any particular frequency is reduced accordingly. Therefore spread
spectrum is often used in high interference situations, like satellite
communications or cellular phone networks.
The use of a code makes the signal inherently secure. The receiver must
know the code to "un-spread" the spectrum and recover the signal.
Local Area Networks (LAN)
Local area networks link communication system users together within the
immediate vicinity. An example is the intercom system on an office phone. The
most obvious example is of course the local computer network. There are many
features about local networks that are general.
The way the users, or stations, in a LAN are connected together is called
the topology. It should be thought of as the schematic diagram of the network.
It does not necessarily represent the physical interconnections between them.
There are four main topologies:
Star. This features a central hub to which all the stations are connected.
These are generally found in a system where the stations are terminals,
and which are connected to the central, or main frame, computer. Star topology
uses a significant amount of wire. If one of the stations is faulty, the
overall system is unaffected.
Ring. All of the stations are connected in a closed loop. The fiber optic
network, FDDI (fiber distributed data interface), uses a ring topology.
Each station must pass all of the information through it, even if it will
not use any of it.
Bus. All of the stations are connected to a common line, called the bus.
The ethernet is a common example of the bus topology. Any signal put on
the bus is heard by all stations. Individual stations are identified by
Mesh. All of the stations have separate connections to every other station.
Signals may be sent directly, station-to-station, or via one or more other
stations. This topology is rarely used except for critical connections,
usually between central computer systems. It requires the most wires and
This applies when sending data between stations on any network. A protocol
defines the rules. Here are three common ones:
Carrier Sense Multiple Access with Collision Detection (CSMA/CD). Used
in bus topologies. Stations desiring to send data, first must listen to
the bus to determine if another station is already transmitting. If so,
it must wait. In the unlikely event that two stations begin to transmit
at the same time, a collision will be detected and they must both stop
and try again later.
Token ring. This applies to the ring topology. Signals are identified to
particular stations on the ring by something called a token. When a particular
message is received by the intended user, a different token is put on the
ring to indicate receipt, and is subsequently taken off by the originating
Polling. This is used in the star topology. The master station which controls
the network, will send a message to each station is turn which asks them
if there is anything to transmit. If so, the operation will be carried
out. If not, the other stations will be polled until the cycle is complete
at which point it will start over.
Wide Area Networks (WANs)
When the stations in a network at separated by a significant distance,
the physical connections of a LAN cannot work. In many cases local networks are
linked together by a larger system.
Fixed Wide Area Networks
There are two extensive fixed networks in place in the United States:
the telephone system and cable TV. Fiber optic systems are generally replacing
the backbone of the telephone system, but most household connections use
simple telephone wire. Both these fixed networks can serve many purposes.
The telephone system interconnects most of the computers all over the world by
what is called the Internet. The cable TV system connects most of the households
in a simplex mode (receive only) for TV signals.
The capacity of these networks is ultimately limited by the physical
medium of transmission. There are three main types:
Unshielded Twisted Pair (UTP) Wire. This is the ordinary telephone wire
probably connecting to your phone. There are, however, many different grades
of UTP wire, some of which has very high capacity, up to 100 MHz. But it
can only be used for short distances. Most LANs use UTP of some variety.
Coaxial cable. This has a bandwidth of up to 1 GHz, and is very good at
shielding the connection from electromagnetic interference because of its
construction. It can be used for longer distance connections than UTP,
but is bulkier and more expensive.
Fiber Optics. This has a bandwidth of over 2 GHz. It can be used for very
long ranges. Making connections to fiber optics is fairly difficult.
The cellular network uses 800-900 MHz radio waves to connect stations to
central receiving antennae and base station. The region of coverage by that base
station defines the cell. The base stations are interconnected by fixed media
similar to the telephone network. A central system takes care of locating the
recipient and the signal is sent to the appropriate antenna in that cell for
broadcast, or alternatively passed onto the fixed telephone network.
When a station passes out of one cell into another, the base stations perform
a handoff and transfers the call to the next base station. Even when not
placing calls, cellular phones send low power signals in order for the system
to determine the phone's location for incoming calls.
Somewhere around 50 MHz of bandwidth is available divided between the transmit
and receive side. Most cellular phone signals are MSK (minimum shift keying,
a variety of FSK) and have a signal bandwidth of 200 KHz. That gives a
capacity of about 125 channels per cell.
There are over 750 satellites, many of which are available for use
in communications. The orbit of the satellite either polar, meaning it
travels around the earth over the poles, or geo-synchronous, where the
position is fixed (mostly) somewhere over the equator. Geo-synchronous
orbits are at 22,500 miles in altitude, which is over five Earth radii.
The time delay for the signal to travel to the satellite and back is about
0.3 sec and is noticeable. Polar orbits are much lower, so there is no delay.
Polar orbit satellites travel around the Earth anywhere from 1-12 hrs, and
require many satellites for complete coverage. Satellite communications must
be UHF and higher in order to penetrate the ionosphere, and are commonly in the
SHF range. They are generally either C band (3.75-7.5 GHz) or Ku band (12-17 GHz).
Bandwidths are generally in the GHz range. Furthermore, circular polarization
is predominately used since it is difficult to properly orient the receiving
antenna to match the satellite antenna if linear polarization where used.