Today many communications companies are replacing their copper carrier wires with fiber optic cables. A fiber optic cable is capable of transmitting laser light across thousands of miles and can carry many more messages at the same time than the copper wire of equivalent diameter. With the relentless pursuit of bandwidth, fiber optic cabling is being deployed at an ever increasing rate. This cable, which uses glass to carry light pulses, poses both advantages and challenges. The intent of this paper is to explain the how’s and why’s of fiber optic cabling and to provide a set of solutions to the challenges faced with it’s use and give you an understanding of fiber optic cable technology and its applications. Fiber optic cabling has much to offer, and in most cases, its use will provide benefits which justify the implementation.
Since the invention of the telegraph by Samuel Morse in 1838, there has been a constant push to provide data at higher and higher rates. Today, the push continues. Just as RS-232 attached terminals gave way to 10Mbps Ethernet and 4 and 16 Mbps Token Ring, these are giving way to Fast Ethernet (100Mbps), FDDI (100Mbps), ATM (155Mbps), Fiber Channel (1062Mbps) , Gigabit Ethernet (1000Mbps). With each of these increases in speed, the physical layer of the infrastructure is placed under more stress and more limitations. The cabling installed in many environments today cannot support the demands of Fast Ethernet let alone ATM, Fiber Channel, or Gigabit Ethernet. Fiber Optic cabling provides a viable alternative to copper. Unlike its metallic counterpart, fiber cabling does not have the severe speed and distance limitations that plague network administrators wishing to upgrade their networks. Because it is transmitting light, the limitations are on the devices driving it more than on the cable itself. By installing fiber optic cabling, the high cost of labor and the time associated with the cabling plant can be expected to provide service for the projected future.
Plastic Optical Fiber (POF) technology is making fiber even more affordable and easier to install. Because the core is plastic instead of glass, terminating the cable is easier. The trade-off for this lower cost and ease of installation is shorter distance capabilities and bandwidth limitations.
Fiber optic cabling has the following components (starting in the center and working out): core, cladding, coating, strength member, and jacket. The design and function of each of these will be defined. The core is in the very center of the cable and is the medium of propagation for the signal. The core is made of silica glass or plastic (in the case of POF) with a high refractive index. The actual core is very small (compared to the wire gauges we are used to). Typical core sizes range from 8 microns (millionth of a meter) for single mode silica glass cores up to 1000 microns for multi mode POF. The cladding is a material of lower index of refraction which surrounds the core. This difference in index forms a mirror at the boundary of the core and cladding. Because of the lower index, it reflects the light back into the center of the core, forming an optical wave guide. This is the same effect as looking out over a calm lake and noting the reflection, while looking straight down you see through the water. It is this interaction of core and cladding that is at the heart of how optical fiber works. The coating (also referred to as buffer or buffer coating) is a protective layer around the outside of the cladding. It is typically made of a thermoplastic material for tight buffer construction and a gel material for loose buffer construction. As the name implies, in tight buffer construction, the buffer is extruded directly onto the fiber, tightly surrounding it. Loose buffer construction uses a gel filled tube which is larger than the fiber itself. Loose buffer construction offers a high degree of isolation from external mechanical forces such as vibration. Tight buffer construction on the other hand provides for a smaller bend radius, smaller overall diameter, and crush resistance. To further protect the fiber from stretching during installation, and to protect it from expansion and contraction due to temperature changes, strength members are added to the cable construction. These members are made from various materials from steel (used in some multi – strand cables) to Kevlar. In single and double fiber cables, the strength members are wrapped around the coating. In some multi-strand cables, the strength member is in the center of the bundle. The jacket is the last item in the construction, and provides the final protection from the environment in which the cable is installed. Of concern here is the intended placement of the cable. Different jackets provide different solutions for indoor, outdoor, aerial, and buried installations.
The most common size of multi mode fiber used in networking is 62.5/125 fiber. This fiber has a core of 62.5 microns and a cladding of 125 microns. This is ideally suited for use with 850nm and 1300nm wavelength drivers and receivers. For single mode networking applications, 8.3/125 is the most common size. It’s smaller core is the key to single mode operation.
Numerical aperture and acceptance angles are two different ways of expressing the same thing. For the core / cladding boundary to work as a mirror, the light needs to strike at it a small / shallow angle (referred to as the angle of incidence). This angle is specified as the acceptance angle and is the maximum angle at which light can be accepted by the core. Acceptance angle can also be specified as Numerical Aperture, which is the sin of the acceptance angle (Numerical Aperture = sin (acceptance angle)).
With a basic understanding of fiber construction, explanation of transmitters (the devices that put the pulses of light into the fiber) is in order. From a general level, there are three aspects of transmitters to discuss:
Transmitters can be divided into 2 groups, lasers and LEDs. LEDs are by far the most common as they provide low cost and very efficient solutions. Most multi mode transmitters are of the LED variety. When high power is required for extended distances, lasers are used. Lasers provide reliable light and the ability to produce a lot of light energy. The drawbacks to lasers are their cost and electrical power consumption. Equipment using high power lasers must provide cooling and access to a primary power source such as 120V AC.
Transmitter types can also be broken down into single mode versus multi mode transmitters. Multi mode transmitters are used with larger cable (typically 62.5/125 microns for most data networking applications) and emit multiple rays or “modes” of light into the fiber. Each one of these rays enters at a different angle and as such has a slightly different path through the cable. This results in the light reaching the far end at slightly different times. This difference is arrival times are termed modal dispersion and causes signal degradation. Single mode transmitters are used with very small cable (typically 8/125 microns) and emit light in a single ray. Because there is only one mode, all light gets to the far end at the same time, eliminating modal dispersion.
The wavelength of the transmitter is the “color” of the light. The visible light spectrum starts around 750nm and goes to 390nm. The 850nm transmitters common in multi mode Ethernet can be seen because 850nm is the center of their bandwidth and they emit some visible light in the 750nm range giving them their red color. The 1300nm and 1550nm transmitters emit light only in the infrared spectrum. The difference in performance of the various wavelengths is beyond the scope of this paper. What is important is an awareness of the wavelengths and that the equipment on both ends of the fiber needs to be matched. The final characteristic of transmitters is the output power. This is a measure of the optical energy (intensity) launched into the fiber. It is measured in dBm. A typical value for multi mode transmitters used in Ethernet is -15dBm. Single mode transmitters have a wide range in power depending on the application.
With a knowledge of transmitters, what happens at the other end of the cable is important. The light pulses are terminated and detected with a receiver. Receivers have three basic considerations. These are:
Sensitivity is the counterpart to power for transmitters. It is a measurement of how much light is required to accurately detect and decode the data in light stream. It is expressed in dBm and is a negative number. The smaller the number (remember -40 is smaller than -30) the better the receiver. Typical values range from -30dBm to -40dBm. Receive sensitivity and transmitter power are used to calculate the optical power budget available for the cable. This calculation is: Power Budget = Transmitter Power – Receiver Sensitivity, Using the typical values given for multi mode Ethernet above, the power budget would be: 15dBm = -15dBm – (-30dBm) The optical power budget must be greater then all of the cable plant losses (such as attenuation, losses due to splices and connectors, etc.) for the installation to work properly.
Figure A. – SC Connector Figure B. – ST Connector
Many different connector styles have found their way into fiber optic networking. The SC connector (Figure A) has recently been standardized by ANSI TIA/EIA-568A for use in structured wiring installations. Many single mode applications are now only available in the SC style. The ST connector (Figure B) has been the connector of choice for these environments, and continues to be widely used. FDDI uses the MIC connector which is a duplex connector. It is physically larger then the SC connector, and the SC connector is gaining acceptance in the FDDI marketplace.
Fiber provides several advantages to Ethernet and Fast Ethernet networks. The most common advantage and therefore use of fiber is to overcome the distance limitations of coaxial and twisted pair copper topologies. Ethernet being run on coax (10Base2) has a maximum distance limitation of 185m, and Ethernet being run on twisted pair (10BaseT and 100BaseTX) has a limitation of 100m.Fiber can greatly extend these distances with multi-mode fiber providing 2000m and single-mode fiber supporting 5km in half duplex environments, and much more (depending on transmitter strength and receiver sensitivity) in full duplex installations. Ethernet running at 10Mbps has a limitation of 4 repeaters, providing some leniency in the solutions available for distance, however, Fast Ethernet only allows for 2 repeaters and only 5m of cable between them.
As Fast Ethernet becomes more ubiquitous, the need for fiber optic cabling will grow as well. When distance is an issue, fiber provides what may be the only solution. Even when using coaxial cable or twisted pair (shielded or unshielded), some electrical noise may be emitted by the cable. This is especially true as connectors and ground connections age or weaken. In some environments (medical for example), the potential risk associated with this is just not acceptable, and costs of alternative cable routings too high. Because fiber optic cabling uses light pulses to send the signal, there is NO radiated noise. This makes it perfectly safe to install this cabling in any sensitive environment. Optical fiber adds additional security protection as well. There are no emissions to pick up and decode, and it is not feasible to “tap” into it for the purposes of “eavesdropping”. This makes fiber optic cabling ideal for secure network installations.
Another problem that is common when using copper cabling is other electrical noise getting into the desired electrical networking signal. This can be a problem in noisy manufacturing environments or other heavy industrial applications. The use of optical fiber provides a signal that will be completely unaffected by this noise. In some instances, fiber provides the advantage that it can withstand more tension during the cable pulling. It is also smaller in size then twisted pair cables and therefore takes up less room. Compared to Category 5 UTP, most duplex fiber optical cable can also endure a tighter bend radius while maintaining specified performance.
Fiber optical cabling is not a cure-all however, there are some challenges to be resolved. The first (and probably the best known), is the cost of termination. Because of the need for “perfect” connections, splices and connections must be carefully cut and then polished to preserve the optical characteristics. The connectors must also maintain a very high level of precision to guarantee alignment of the fibers.
The second problem that is encountered when installing fiber cabling is that legacy equipment does not support fiber connections. Very few desktop computers have a fiber network interface, and some critical network equipment does not offer a fiber interface. In Ethernet, the size of the collision domain can effect the use of fiber. In a half duplex (shared media) environment, no 2 devices can be separated by more then 512 bit times. While the transmission of a signal is faster through fiber than copper, only about 11% faster and not enough to make a significant difference. This limitation means that there are times when the signal quality and fiber are sufficient to carry the signal but the distance and network design rule out it’s use.
Fortunately, the problems are not without solutions. As fiber deployment increases, the economy of scale for the manufacturers is driving costs down. Also, much work is being done to further reduce these costs, Plastic Optical Fiber is an example of one such development.
The need to connect to legacy equipment and infrastructure also has a solution. By using copper to fiber media converters, fiber can be connected to almost any legacy environment. Equipment equipped with an AUI port can also make use of fiber transceivers as well. Media converters are devices (usually small enough in size to fit in the palm of your hand) which take in signals from one media type and send it out on another media type. For those instances when collision domain restrictions preclude the use of fiber, a 2 port bridging device (such as Transition Networks Bridging Media Converter) with 10/100-Base-T(X) on one port and fiber on the other can be used. Bridges by definition break collision domains, and when connected to a server, workstation, or another bridge can operate in Full Duplex mode. In this mode, there are no limitations imposed by collision domains, and the distance attainable is solely a function of the fiber cable; and transmitters and receivers.
Fiber optic cabling is rapidly becoming the most viable choice for data networking infrastructure. With the cost of cable, connectors, installation, and equipment becoming competitive with traditional copper solutions, fiber should be given serious consideration. Transition Networks’ complete line of fiber connectivity products are specifically designed to ease this migration to fiber. Once installed, fiber optic cabling will “future proof” your cabling infrastructure, providing support for even the fastest most demanding protocols.