Fiber-based systems offer apparent advantages over
electrical methods in large plants and factories where
the harsh environment threatens data reliability and
security. Noise is the most common problem. High voltage
and motor feeders, motors and motor drives, and even
fluorescent lighting can generate sufficient electromagnetic
interference (EMI) to corrupt wire or radio-based systems.
In a distributed control system (DCS) or supervisory
control and data acquisition (SCADA) installation, this
problem will at least render the system unreliable.
An unstable factory network can result in the uncontrolled
operation of equipment and possible personnel injury
or facility damage. If such an environment is unavoidable,
fiber optics will effectively eliminate this problem.
Fiber optics have come a long way since 1854, when
John Tyndall established that a jet of water streaming
from a container could transmit light. Later nineteenth
century experimenters used parabolic mirrors and internally
metalized tubes to guide light. In the 1920s, scientists
used the first bundles of glass tubes to transmit crude
but recognizable images.
Back to Basics
The concept of refractive index is basic to understanding
fiber-optic technology. To effect the most complete
transfer of energy through a 1nedlure, the light energy
has to be completely contained within the medium. This
containment is done by making use of the medium's refractive
Every material capable of transmitting light has a
characteristic refractive index, defined as the ratio
of the velocity of light in the material to the velocity
of light in a vacuum. When light encounters a boundary
between materials of differing refractive indices (e.g.,
air and water), a part of the light is bent while a
part of it is reflected.
If the angle at which the light ray (also known as
a mode) striking the boundary exceeds some angle of
total ray is reflected. light ray enters the other material.
critical angle is the maximum angle of refraction.
Fiber-optic cable consists of three parts: the core,
the cladding, and the coating. The core transmits the
light and has a high refractive index. The cladding
contains the light within the core because its lower
refractive index causes all the light rays to reflect
back into the core. This "total internal reflection"
or "fiber-optic effect" is the technology's
underlying principle. The coating, usually an acrylate
polymer, protects the core/cladding assembly.
Optical fiber is typically made from high-purity silica
glass. Plastic fiber of varying configurations is also
available. but the attenuation of light energy can approach
one thousand times that of glass fiber.
You identify optical fiber by its core/ cladding size.
A singlemode, step-index fiber with a core diameter
of 8 microns and a cladding diameter of 125 um would
have a designation of 8/125. Similarly, a multimode,
step-index fiber could be designated 200/230. (The term
step index means that there is a sharp change in refractive
index from the core to the cladding.)
The length and integrity of the transmission path
and the core/cladding arrangement affect the bandwidth,
or the frequency range that the optical fiber transmits.
Fiber bandwidth is expressed in megahertz-kilometers
Attenuation in optical fiber is due mainly to scattering
and absorption within the core material. Dirty or poorly
made connections will diminish the light intensity.
Attenuation, or loss, is expressed in decibels per kilometer;
thus, loss is introduced over distance and at connections.
Any connection, regardless of quality, will introduce
a loss or "insertion loss."
Three core/cladding arrangements are in common use.
Each exhibits characteristics that affect the bandwidth
and, thus, the application. Basically, the refractive
indices and diameters of the core and cladding are manipulated
relative to each other, producing fibers that are optimized
for specific applications.
The single-mode, step-index fiberoptic cable has the
highest bandwidth (exceeding 25 GHz-km) and the lowest
loss. A typical fiber of this type is designated 8/125.
The single-mode fiber core allows one mode of transmission.
This means that only one ray of light is transmitted.
The transmitted light's change in velocity over its
wavelength range limits the bandwidth. This "chromatic
dispersion" is an inherent property of the core
material and increases with distance.
Single-mode cables are difficult to connect or splice
because the small core diameter makes the alignment
of the fibers critical. Attaching connectors or splicing
requires specialized equipment and trained personnel.
Single-mode fibers are mostly used in long-distance
At the other end of the bandwidth scale is the multimode,
step-index fiber. A typical designation is 200/230.
The core diameter is large relative to the light rays
transmitted; therefore, more than one ray, or mode,
can be transmitted.
This type of fiber has a much lower bandwidth than
the single-mode fiber, usually topping out at about
20 MHz-km. The usable bandwidth is even lower, in the
vicinity of 5 MHz-km. The sharp difference in refractive
indices (step index) causes some of the transmitted
light rays to be reflected many times as they travel
the length of the fiber.
This results in the transmitted light rays arriving
at the receiving end at different times. A sharp pulse
at the transmitted side, for example, would arrive at
the receiving end in a "spread." This "modal
dispersion" limits this type of fiber to low-end
process and specialized medical applications.
Between the single-mode and multimode step-index fibers
is the multimode, graded-index optical fiber. In contrast
to the other two types of fiber, the difference in core/cladding
refractive indices is gradual, not stepped. The core
has a high index that becomes gradually lower as it
approaches the outer diameter of the cladding. Instead
of producing a sharp boundary between the two components.
with resultant degradation in performance, light rays
in a graded-index fiber are reflected to make the path
lengths of all rays equal.
The gradation of refractive indices produces a focusing
effect in the core. Consequently, modal dispersion does
not greatly affect multimode, graded-index fibers. A
common fiber of this type is designated 62.5/125 and
has a bandwidth between 100 and 800 MHz-km. This type
of fiber finds wide use in commercial and industrial
data communications systems.
Light sources for multimode fibers are lasers or high-performance
LEDs. Because of the large core size of this type of
fiber, alignment when connecting is not as critical
as it is with single-mode fibers.
Each type of fiber will only transmit light that enters
the fiber core's "cone of acceptance." If
a ray hits the core at an angle outside of this cone,
the light will not be internally reflected. This parameter
is the fiber's numerical aperture (NA). Larger cones
have a larger NA, which also indicates the relative
light-gathering power of the fiber. Thus, a 200/230
fiber would have a larger NA than a 8/125 fiber.
Depending on the application, the distance involved,
and the location, several types of cable configurations
and connector types are available. Optical fiber is
fragile and must be protected, mostly from mechanical
stresses such as bending, crushing, thermal effects,
and pulling during installation.
Even though tensile strength typically exceeds 600,000
pounds per square inch, glass fiber can lose a significant
amount of strength when exposed to water. Excessive
bending will also cause degradation in performance.
With a reasonable amount of care, you can install optical
fiber easily, and it will last for about 30 years or
more. The most common mode of failure in optical cable
is the "backhoe syndrome." As with any underground
installation, accurate drawings, close supervision,
and the use of warning tapes serve to avoid costly outages.
A tight-tube (or tight-buffer) design has a PVC coating,
which tightly bonds to the fiber, limiting movement
(see Figure 3). This cable
type can have strength members, which you pull through
conduit and cable trays. This design, however, has low
crush resistance and is susceptible to deformation due
to thermal expansion; thus, it is recommended for indoor
A loose-tube design gives a fiber free movement
(see Figure 4). Each component of the cable (the
sheath or outer coating, the strength member, and the
buffer tubes that carry the fibers) has different thermal
characteristics. By allowing the fibers and the components
surrounding them free movement, deformation is avoided.
Loose-tube construction has much better crush resistance
than tight tube because of the buffer-tube protection
of the fibers. Loose-tube cables have a strength member,
which is used as the pulling member for conduit installation.
In some installations, the strength member replaces
the messenger used in an aerial installation.
The strength member is made either of a high tensile
strength plastic or steel. Note that adding a steel
strength member renders the cable conductive. Loose-tube
cables are usually filled with a gel, which surrounds
the fibers and increases protection from water. This
also improves crash resistance because of the gel's
You mostly use this cable type for outdoor applications,
but you can also use it in harsh industrial environments.
A drawback to this type of cable is the difficulty in
handling individual fibers. The fiber coating does not
have to be as thick as in tight-tube construction; thus,
attaching connectors is difficult.
Breakout kits are available for loosetube cables.
These provide each fiber with a sheath that is color-coded
and equipped with a strength member. The thicker sheath
lets you attach any connectors.
Breakout cables are a hybrid solution (see
Figure 5). In a breakout cable, each fiber is treated
as a separate unit, complete with a sheath and strength
member. This design eliminates the need for a breakout
kit, because the sheath lets you attach connectors easily.
Remember that tight-tube construction does not allow
for free movement and provides low protection against
mechanical stress. It does, however, have a thick coating
for ease of handling. Loose-tube construction, on the
other hand, allows free movement and provides a good
degree of protection.
Breakout cables let fiber subunits move freely, and
they protect each fiber by virtue of their thicker coating/strength
member arrangement. Each fiber subunit is configured
as a tight tube. Breakout cables also come equipped
with a separate strength member just like the loose-tube
Know Your Connections
Several connector types, such as the generic connector
shown in the sidebar "Kinds of Connectors,"
are available. As the technology matures, the cost and
difficulty of making a connection will decrease. At
present, attaching connectors requires specialized equipment
and trained personnel. Attaching connectors is time
consuming, and even the best connectors introduce an
optical, or insertion, loss.
Connectors are classified by the shape and size of
the alignment ferrule and by the means of mating the
housings. The ferrule is a sleeve that holds the fiber
in position inside the connector assembly. The fiber-to-fiber
alignment inside the connector coupling assembly is
critical for the most complete and efficient transfer
of light energy.
In a single-mode fiber with a core of only 8 um, any
misalignment results in a significant or complete loss
of signal. In a 200/230 fiber, even as much as a 15%
misalignment can be tolerated. If the connector's mechanical
tolerances can be tightly controlled, then proper installation
of fiber into ferrule is the only variable.
Each connector requires a different technique for
proper installation. You should consult the manufacturer's
literature for the latest information on these connectors.
Connect for Life
When a permanent connection between fibers is desired,
optical fibers can be physically spliced (see
Figure 7). Splicing, while limiting flexibility,
is cheaper and results in a stronger, permanent connection.
The following two methods are used:
Mechanical splice: This splice requires treatment
and precise alignment of the fiber ends. It is a low-cost,
long-term alternative to a connector that introduces
less loss in most cases. This type requires a significant
investment in equipment, and some training is required.
Fusion splice: This is the most reliable and low-loss
splice. Trained personnel fuse together the fiber ends
using specialized equipment. This type of splice will
stand up to a high degree of physical stress and introduces
the lowest optical loss.
Patch panels are used as distribution points, storage
locations for spare fiber, and splice cases. Strategically
located patch panels will allow easy configuration and
reconfiguration of the network.
As you add and remove devices, the patch panel serves
as a convenient point at which fibers can be terminated,
tagged, and made available for use. You can store unterminated,
spare fibers for future use. Panels also provide protection
from environmental conditions, and many include splicing
There are several methods for testing fiber-optic
cables. The most basic is the pocket flashlight. This
is a quick and dirty method for checking continuity
or for "ringing out" a single fiber. A note
of caution: If you are unaware of what is happening
at the other end of the fiber, never look into the end
of a fiber-optic cable! One or more fibers may be carrying
potentially harmful levels of light energy. Point the
cable at the wall while testing.
Somewhat more sophisticated is the optical time domain
reflectometer (OTDR).The OTDR measures backscatter,
reflections caused by a break or some other fault in
the optical fiber. A laser pulse is injected into the
cable and, when a fault is detected, a portion of the
pulse is reflected back to the OTDR. On well-documented
systems, the fault can pinpoint the location of any
problems with the cable.
For periodic checkouts, the optical-loss test set
is used. This is basically a light source of specific
or selectable frequency and power output and a receiver
that indicates optical power. The test set is inexpensive
compared with the OTDR.
It is good practice to test optical cables before
and after installation. Testing before will spot any
manufacturing or shipping problems and eliminate having
to rip out the installation later. It also provides
a baseline for later comparison. You should always test
the cable after installation. This also provides a baseline
and can spot a problem developed during installation.
Transmit and Receive
An optical modem converts electrical energy into light
energy. The modem receives voltage levels from the data
terminal. It then translates these voltages into light
pulses of varying duration arranged to conform to one
of several data communications protocols.
For single-mode fiber, used mostly in 1ong-d'lstance
communications, a semiconductor laser is used to transmit
light pulses in the 1,300 to 1,550 nm range. These devices
produce high-quality, coherent light and can support
transmission rates of over 2 Gbps at distances in excess
of 30 km. For multimode fibers, high-performance LED
modems in the 850 to 1,300 nm range are generally used
in transmission rates over low to medium distances.
This type of modem is used in most LAN applications.
Typical bandwidths for 1,300-nm LEDs would be about
100 Mbps/10 km. IEEE 802.3J (Ethernet) will support
10 Mbps/4 km, while fiber distributed data interface
(FDDI) will run at 100 Mbps/2.2 kin.
Cost differences between laser and LED modems are
significant. LEDs are used for the majority of LAN or
factory applications. Lasers are simply unnecessary.
Adequate bandwidth, reliability, and lower cost make
any of the high-quality LED-based modems suitable for
most communications applications.
Detectors come in two flavors: the PIN photodiode
and the avalanche photodiode (APD). The PIN photodiode
(named for its semiconductor layers) generates charge
carriers when exposed to light. These charge carriers,
or photocarriers, form a photo current that is then
translated into usable data. PIN devices require low
bias voltages and are abundantly available matched to
the common LED source ranges.
The APD also generates photocarriers when struck by
light. These charges cause an avalanche of charge carriers
and are, therefore, capable of forming a much larger
output current. APDs are much more sensitive than PINs.
APDs are used primarily for laser sources over long-distance
links. But APDs are more costly, require bias voltages
in excess of 100 volts, and are sensitive to temperature
With a good understanding of the equipment, the next
step is to lay out the network. Fundamental to proper
operation of optical links is the need for transmitted
power to reach the receiver. This sounds obvious, but
it is often overlooked.
Every component in an optical communications system
introduces loss. Before specifying modems, the designer
must have a thorough understanding of the proposed system's
present and future performance goals. The first step
is to calculate the optical budget and the total system
Optical fiber has a characteristic loss per unit distance.
This is usually expressed in dB/km and is dependent
on the wavelength of the light source. For a typical
8/125 general-purpose fiber, the attenuation is 0.8
dB/km @ 1,550 nm. Higher quality fiber of the same size
can approach 0.4 dB/km. A good attenuation for a 100/140
fiber is 3.5 dB/km @ 1,300 nm.
Each connector introduces a loss (the insertion loss)
into the link. This is expressed in dB/connector or
simply dB. Connector insertion losses range from approximately
1.5 dB for an SMA-905 to 0.3 dB for an SC connector.
Splices also introduce an insertion loss, which varies
widely depending on type and quality. For the sake of
calculating the system loss, you should use a figure
of 0.10 dB/splice unless you've tested the splice and
know its insertion loss.
Assume that we have a link with four connectors, two
splices, and a total of 6 km of fiber. We are using
a 62.5/125 fiber with a loss of 1.0 dB/km. Transmitter
output power into the fiber is -12 dBm (decibels as
referenced to 1 milliwatt). Our receiver's sensitivity
is -27 dBm. Subtracting the transmitter's output power
from the receiver's sensitivity gives us an optical
budget of 15 dB.
Assuming a source of 1,300 nm, we can now summarize
the system or link losses:
6 km fiber @ 1.0 dB/km = 6.0 dB
4 connectors = 1.2 dB
(ceramic ST) @ 0.3 dB
2 splices @ 0.10 dB = 0.2 dB
Total system loss = 7.4 dB
Subtracting the system loss from the optical budget
gives us a margin of 7.6 dB. This is a comfortable margin,
which will allow future expansion of the link. If no
expansion is planned, we can use a less powerful and
less costly modem.
We should also consider changes in fiber size over
the length of the link or a bypassed node. Furthermore,
a 1-dB loss should be allowed for single-mode fiber
links to allow for dispersion and absorption. In all
cases, a margin of at least 3 dB should be maintained.
A margin of 1 dB is not acceptable, considering that
LEDs and other components degrade with time, which could
render such a link susceptible to erratic performance
or complete failure.
Fiber-optics technology is still maturing. A good
grounding in the basics is a necessity for anyone who
wishes to design, implement, or even effectively use
a fiber-based system. There are numerous short courses
and literally tons of documentation available for those
who wish to learn more about the technology described
in this article. For a quick review of basic networking
technology, see the December 1995 issue of Industrial