Fiber optics

The principle of fiber-optic cables has been used in the medical and imaging industries since the 1930s and is based on a simple physical property that was first demonstrated back as far as the early 1840s. So while we consider fiber optics a modern technology, the fundamentals have been understood and in use for more than 150 years. But what makes fiber optics so appealing to our industry, and what considerations need to be taken in to account when applying this technology to media delivery systems?

Construction of fiber-optic cables

Fiber-optic cable is formed with a number of layers, each performing a specific task. (See Figure 1.) Working from the center outwards, the first component is the core. This is a solid cylinder made from either plastic or glass through which the light passes.

The next layer is called the cladding, which again can be made from either plastic or glass. There is often a gel or powder coating surrounding the outside of the cladding, which helps to protect the inner cores from moisture and water ingress.

The third layer is a thin protective tube commonly made from hard plastic and is illustrated in Figure 1 as a thin black line between the cladding and coating. Finally, the coating forms the outer layer and is primarily used as a protective barrier against impacts and the rigors of installation.

These layers form one unidirectional cable. Where bidirectional communications are required, you will often find duplex pairs of cable where two of the above are joined together and form parallel cores with light travelling in opposite directions. It is also possible to have many cores inside the same multicore cable to allow multiple transmission paths for different data types all contained in a single, easy to manage cable.

Total internal refraction

The basic principle of fiber-optic technology is to rely on total internal refraction within the core material. This phenomenon describes how light is reflected when it hits the outer edge of the core. If the angle to the perpendicular is shallow enough, the light cannot exit the core and is instead bounced back inside the core. The angle required to guarantee total internal refraction is determined by the refractive index of the material used for the core, as well as the interaction between the outer edge of the core and its surrounding material — whether that be air, water or glass.

In order to guarantee total internal refraction with modern fiber-optic cables, the cladding is designed to have a lower refraction index than the material used for the core. This means that light can never pass from the core to the cladding, thus keeping all of the light in the core. This relationship between core and cladding is called the step index where there is an abrupt change in refractive index between the two layers.

It is possible to manufacture class cores that have a gradually reducing refraction index as you move from the center toward the outside. The effect this has on the path of the light is that instead of bouncing off the edge of the core/cladding boundary, the light is instead gradually bent back in toward the center of the core in a curve. The outcome is lower loss of power over distance, as well as more accurate transport of the data. The refraction property of this cable type is called graded index.

Multimode vs. single-mode

Fiber-optic technology is used as a transport medium for digital data. The two major types of cable in use today are multimode and single-mode. (See Figure 2 on page 16.) The physical differences between these cables relate to the diameter of the core where a typical step index multimode core would be 200µm, whereas both 62.5µm and 50µm would be common core diameters with graded index multimode. Single-mode cables have cores with a diameter of just 9µm. Each of these cable types has different data transmission properties that make them suitable for different applications.

If you look at the light emitted from a torch, it's not a direct beam of light. A central hot spot comes out perpendicular to the end face of the torch lens, and spilled light comes out at varying angles from the lens. The same thing happens when you shine an LED transmitter into a multimode cable. You get the central hotspot (or mode), as well as light entering the cable at varying angles to the perpendicular. The steeper the angle when the light enters the core, the more times it will bounce off the core/ladding boundary as it passes down the cable.

Consequently, it has further to travel than the primary mode. As light travels at a constant speed, an impulse of light will be spread across the time domain as it passes through the cable because some of the light has further to travel than the rest; this is called time dispersion. If you are sending multiple impulses, as you would with a digital communications system, you risk losing the transition between light and dark at the output if the time dispersion is sufficiently long over the length of cable. For that reason, there are maximum data rates per length of cable for different cable types.

Conversely, with a high-quality light source such as a laser, where the core is sufficiently small to only allow one beam of light (or mode) through, the time dispersion problems are effectively eradicated. Thus, you can have much higher data rates over much longer distances.

Energy loss

Fiber-optic cable absorbs some of the light energy presented at its input. This happens mostly at the refraction nodes, where the light bounces in step index cable but also in the curves generated by a graded index core. The core material has different absorption properties depending upon the wavelength of the light used. Some typical values for wavelengths used are 850nm, 1300nm and 1550nm.

As the wavelength of the light is increased, the absorption through the cable decreases. The absorption amount is measured in decibels and is typically quoted in dB/km. For example, a step index multimode cable would absorb 6dB of energy per kilometer when 850nm light is used. Conversely, a step-index single-mode cable with 1550nm light will absorb just 0.2dB/km.

This is important because fiber-optic devices have a parameter called power budget that describes how much loss of power they can cope with before they're unable to correctly decode the data on the incoming cable. Consequently, if you need to go long distances with high data rates, a single-mode 1550nm fiber system would be the best option. If distance and high data rates are not a critical factor, however, step index multimode at 850nm could be sufficient.

Connector types

In just the same way as an analog audio cable, fiber-optic cables are rarely useful without some form of connector. A number of different types are common, but they all revolve around two basic principles. (See Figure 3.) The first type involves butting exposed ends of two connectors together and aligning them in such a way that the light can pass through. The alignment in a connector of this type that will be unplugged and replugged numerous times can never be 100-percent accurate. For this reason, there is some loss in power at the junction. Any dust or imperfections such as scratches will add significantly to the power loss over that link. This is a common cause of problems in fiber-optic systems.

The second type of connector is called the expanded beam format. Several manufacturers produce connectors using this principle, which involves passing light from the core through a lens that expands the beam. The light then passes a small air gap, where it meets another lens that shrinks the beam back down in to the opposite core. These connectors typically feature a larger loss in power compared with coupled connectors.

However, due to the fact that the beam of light is so much larger at the transition point, these connectors are far less susceptible to problems introduced by dirt, dust and moisture. For this reason, they are normally found in outdoor situations, where they are often combined with Kevlar-protected fiber-optic cable. This makes them extremely tough and road-worthy.

Planning your network

There are a few key considerations when choosing the correct fiber-optic infrastructure for a given product or system. The first consideration is largely determined by the fiber-optic devices you are using. These will have a known bandwidth over distance requirement, which effectively stipulates the type of cable that must be used. Some devices will give you the option of either multimode or single-mode transceivers, which offer the user the option of keeping costs down with multimode infrastructure but doing long runs over single-mode where required.

In terms of cable routing, some systems require a star topology, whereas others work as a ring. Both have their pros and cons, but where redundancy is a factor, remember to consider diverse cable runs; try to avoid both primary and backup fiber cores within the same fiber-optic multicore cable.

As we have seen already, every connection in the signal path induces some power loss. Nearly all products buffer the fiber-optic signal on their outputs, so the power budget is on a per link basis. Because of this, careful consideration needs to be given to systems where patch panels and multiple interconnects are required between devices.

It is possible to calculate the loss over cable and through interconnects approximately, but it's always important to measure the actual power loss over the fiber system once installed. Calibrated test equipment should be used that applies a known light level at the input point and measures the light output at the end. This is a key fault-finding tool and would often be the first port of call when investigating problems with a fiber-optic system.

Why use fiber-optic cabling?

As we have seen above, the construction of the cable is very compact, so it is possible to have multiple cores running through multicore cables with very small external diameters. When factoring in the potential bandwidth through the cores, you realize huge bandwidth-to-dimensions ratios. Fiber-optic cable is extremely light when compared with analog multicore cables and significantly lighter than Cat 5 cable. All of this leads to easier installation and transportation, along with higher performance.

In addition, it's important to note that there is no electric connection between the transmitting and receiving devices. This fact alone eliminates any potential for ground loops, buzzes or hums. Fiber-optic cable is immune to any sort of electromagnetic interference, so it can be run alongside HVAC cabling without cause for concern.

Conclusion

The benefits offered by fiber-optic technology in terms of bandwidth, size and weight make it an obvious choice for facilities and rental companies wishing to invest in a future-proof infrastructure. The fact that a fiber-optic infrastructure is simply a protocol-independent transport medium means that all sorts of data — whether it be open standards data such as Ethernet or proprietary network links — is adaptable over time and can provide a basis to build on as technologies and components improve.

The misconceived impression that fiber optics are fragile and difficult to maintain has been shown to be more a learning curve than fact. As time progresses, cable and connector manufacturers will find better ways of protecting the connectors, resulting in more resilient systems and happier engineers.

Martin Barbour is systems support engineer for Optocore.