As I've mentioned here a number of times, I work for a small fiber optics firm in New Hampshire. When I first joined the company I knew very little about fiber optics. My background was in radar and microwave. I knew about electrons and waveguides and frequency sources and transmitters and receivers and magnetrons and traveling wave tubes and phased array antennas and phase discriminators and YIG filters and microwave striplines and DROs. I also had some exposure to infrared target tracking systems as well as hybrid IC construction and test. But I knew squat about fiber optics other than it used glass fibers thinner than a human hair.
The past twelve years have been an education.
I got into fiber optics just as the Internet really started its boom and the need for more and more bandwidth climbed. I've learned so much, and yet I still know not nearly enough. But I know just enough to clue the curious amongst you in to how the whole thing works. I won't delve deeply into theory, but I will try to include as many links as I can for those of you wanting to know more than I will cover here.
There are many kinds of optical fiber, though we are more interested in fiber made from glass.
There is optical fiber made from plastic, much like the ones used to hook up DVD players or satellite tuners to Dolby Digital Surround receivers, for those of you fortunate enough to have them. Plastic optical fiber is also used for illumination.
There are also specialty fibers used for such things as temperature sensors, strain gauges, fluid level sensors, and high power laser waveguides. Many of these are made from sapphire or are metal coated in order to withstand high temperatures or other harsh environmental conditions.
With glass fiber, there are a number of different applications. Most are used for telecommunications, be it voice, data, or video. Others are used for imaging. One such imaging application is the medical field. Sometimes diagnosis or surgical procedures require taking a look inside the body without cutting it open. Fiber optic imaging scopes make this possible, though sometimes a small incision is necessary to allow the fiber scope access to the area in interest. Other imaging applications include use for inspection of machinery in tight spaces where one would not normally be able to look without dismantling the machine. Some glass fiber is also used for illumination, much like plastic fiber, but it gives better overall spot illumination than plastic.
So what is it about glass optical fiber that makes it so useful to carry communications? There are two answers to that question.
Optical fiber can carry incredible amounts of data, more than any other technology being used today.
Bandwidth is the big attraction of fiber optics. The technology to allow this large amount of bandwidth has already been deployed. A single fiber is capable of carrying up to 16 terabits of data per second. That's 16 trillion bits of data every second, enough to transmit the entire contents of the Library of Congress in less than a tenth of a second, or in excess of 250 million phone conversations simultaneously. That's a lot of phone calls.
I will delve into how this bit of technological legerdemain is done later in this post.
Second, it has minimal signal loss over distance as compared to copper wire, coaxial cable, or wireless technology (i.e., radio).
The big attraction to using fiber optics for communication is that an optical signal can travel a long way before it needs to be amplified or regenerated. It's not uncommon to have fiber optic links hundreds of miles long that don't require the use of amplifiers or regenerators. Copper wire or coaxial cable, on the other hand, require amplifiers and repeaters every couple of miles or so in order to maintain signal integrity. They use a considerable amount of equipment and power to run signals between two points. That's one reason why the phone companies and other telecommunications companies all use fiber to connect their central offices together. Copper wire is primarily used to connect what is called the 'last mile' to their switching systems. It's what connects you to your phone company's central office, using the same technology as one hundred years ago.
So how is it that optical fiber has such low loss? The simple answer is chemistry. Though the fiber is made from glass, it isn't the same as the window glass in your home or in your car. Optical fiber glass is a very pure formulation that exhibits incredible clarity, particularly at wavelengths of light used in fiber optic communications. In comparison, dry air is a murky, hazy curtain. And just to complicate things a little, optical fiber actually uses two different glass formulations - one for the inner 'core' and a different one for the outer coating, or cladding. The reason for this is something called refraction.
Refraction is defined as the bending of light as it passes between materials of different densities. (For a good demonstration of refraction, click here or here.)
Do you remember looking into a pool of water, poking a stick into that pool and seeing that the stick 'bent' where it entered the water? The apparent bend was due to refraction, the bending of light at the interface between the air and water. Different materials will bend light to different degrees. How much a material will bend light is defined by the index of refraction. The index of refraction is a ratio between the speed of light in a vacuum and the speed of light through the material in question. The index of refraction is almost always a number equal to or greater than 1. (Unless Einstein and Hawking are wrong, light cannot travel faster through a material faster than it can through a vacuum. Therefore the index of refraction can never be a number less than 1 unless the material is something called a metamaterial, a man-made substance that exhibits negative refraction, but we won't go into that now.)
Refraction is this property fiber optics exploits in order to keep the light traveling down the fiber contained in the fiber. The inner core has one index of refraction and the outer cladding has a slightly different one. This difference is what guides the light down the fiber and keeps it from escaping even if it is bent somewhat. However, there are limits to how much one can bend an optical fiber. Bend one far enough and it will break. For the kind of fiber used in telecommunications, a bend tighter than a certain radius will allow light to leak out of the fiber core and cladding but generally won't harm the fiber itself.
-Optical Transmitters and Receivers-
Okay, we have a medium that can be used to carry optical signals a long distance without the need for amplifiers or regenerators. But what is it that generates or detects the optical signal in a fiber?
Transmitters and receivers, of course.
In fiber optics, transmitters primarily consist of a laser and a modulator. The laser generates the light and the modulator changes some characteristic of the laser light in order to couple the data to be transmitted. In effect, the modulator piggybacks the data on to the laser light.
Many of us are familiar with lasers because we see them in use almost every day. The checkout at your local supermarket uses a laser scanner to tally your purchases (that's the 'doot doot' sound you hear at the checkout line). Those red lines you see flowing over the package of Pop-Tarts you wanted as it passes over the scanner come from a red laser diode. You're probably also familiar with the pen-like red or green laser pointers lecturers and teachers use.
The lasers used in fiber optic communications are similar in many ways to those used at supermarkets and lecture halls. Both types are made in a similar fashion, using semiconductor materials much like the silicon used to make most electronic IC components. However, the one big difference between them is that communications lasers do not generate visible light, but rather infrared light. The 'colors' of the infrared spectrum lie just below that of visible light, below the color red. Coincidentally, optical fibers of the type mentioned above actually pass infrared light far better than visible light.
The light from the laser diode is coupled into the optical fiber by means of a miniature lens that focuses the light into the core of the fiber. From there it travels along the fiber to the modulator (assuming the modulator isn't built in to the package). The modulator is usually a crystal that has properties that make it absorb laser light or shift its phase when an electric current is applied to it. The digital data stream that consists all of the phone calls and Internet traffic and so on is what actually controls the electric current feeding the modulator crystal.
Data can be transmitted by a laser by directly modulating the laser by turning it on and off, but there are limitations to how fast a laser can be switched like that. Therefore, most high speed systems use an external modulator. Most external modulators run at a data rate of up to 10 gigabits per second (10 billion bits per second). There are some modulators that can run as high as 40 or even 100 gigabits per second, but they are not as numerous or as widely deployed as 10 gigabit systems.
Now that you have all of that data modulating the laser, you need to have some way of detecting it and turning it back into an electronic data stream. That's where receivers come in to play.
At their most basic, a fiber optic receiver is made up of two parts: a photodiode and an amplifier.
A photodiode is a semiconductor that does one of two things, depending on how it is used. It either generates an electric current when illuminated by light (i.e. photovoltaic effect), or it allows a current to flow under the same conditions (i.e. photoconductive effect). Both configurations are used in fiber optic receivers.
The amplifier is connected to the output of the photodiode and amplifies the signal detected by it. In most cases, the amount of light being detected by the photodiode in a fiber optic receiver can be as little as 10 microwatts (that's 10 millionths of a watt). The amplifier raises the level of the detected signal so that other circuitry can shape and clean up the signal and turn it back into data that can be read and routed by the switching systems, whether they switch data or phone calls. Sometimes the data is turned back into an optical signal and sent down another optical fiber to yet another switching system.
-Optical Amplifiers and Regenerators-
For fiber optic communications over relatively short distances (less than 80 km), there is rarely a need to amplify or regenerate an optical signal. But for longer distances or instances when there is higher than normal loss, optical signals need to be restored to their original strength and the pulse waveforms corrected. This is done in one of two ways - optical amplification and regeneration.
The first is optical amplification. An optical signal comes into one end of the amplifier and comes out the other end many times stronger than the input. How is this accomplished?
Simply, it's magic.
Well, not really magic, but it might seem that way to the uninitiated. Optical amplifiers use something called laser pumping in order to amplify optical signals in a fiber. There are three different methods used to achieve amplification via laser pumping: Erbium Doped Fiber Amplifiers, or EDFAs; Semiconductor Optical Amplifiers, or SOAs; and Raman Amplification.
EDFAs and SOAs use similar methods to achieve amplification. Only the medium used differs.
The Erbium Doped Fiber Amplifier uses an optical fiber doped with erbium atoms as the amplification medium. Lasers with wavelengths below those being amplified are coupled to the doped fiber, causing excitation of the erbium atoms in the fiber (the erbium atoms absorb energy from the pump lasers), charging them to a higher energy state. The optical signals enter the doped fiber at one end. The excited erbium atoms donate their energy to the optical signals as they pass through the fiber, and the energy level of those signals increases. Depending upon the length of the doped fiber and the power of the pump lasers, the optical signals leaving the doped fiber can be anywhere from 10 to 50 times stronger than were when they entered the EDFA. All of this amplification takes place without the need to convert the optical signal to an electronic signal, and back again. All of the amplification is done in the optical domain. Neat, huh? EDFAs are used primarily on long haul fiber optic links like undersea cables and long run terrestrial links.
The SOA works on a similar principle as the EDFA, but rather than using a doped fiber as the amplification medium it uses an optical waveguide. The SOA doesn't provide nearly as much amplification as an EDFA, but then it isn't designed for that. SOAs are used in shorter fiber links, usually in what are called MANs, or Metro Area Networks.
The third amplification method uses something called Raman Amplification. Like the EDFA and SOA, it uses pump lasers to generate the 'donor' energy for amplification. But in this case no special amplification medium is used, just the fiber in the communications link. A high powered laser operating on the same wavelengths one would use in an EDFA is coupled into the communications fiber. The energy from this laser couples to the photons of the communications laser pulses also in the fiber, and the power of the comm laser pulses are amplified. One of the big advantages to Raman Amplification is that it generates less noise than an EDFA. The downside is that it doesn't provide as much amplication as an EDFA. In some cases, EDFAs are used in conjunction with Raman Amplification in long haul fiber links which combines the best of both - high amplification with lower noise.
The second method used to extend range in a fiber optic system is regeneration. A regenerator is basically a fiber optic receiver coupled to a fiber optic transmitter with control, filtering, and sometimes switching electronics in between. The filtering and 'shaping' circuitry can correct for signal defects created by the characteristics of the fiber being used. The regenerator receives the optical signal and converts it to an electrical signal. The electrical signal is then filtered to clean up any noise or pulse distortions in the signal. It is then amplified and the amplified signal is used to drive the optical transmitter.
Each method has its advantages and disadvantages. One big advantage of optical amplification over regeneration is that it can be used to amplify multiple optical signals simultaneously.Remmeber, a single optical fiber can carry multiple wavelengths, or colors, of infrared light at the same time. Optical amplifiers amplify all of those wavelengths at the same time. Regenerators can do this too, but it requires separating each wavelength and sending each one to its own regenerator. That's a lot of receivers, electronics, and lasers, as well as power to run them all.
-Wavelength Division Multiplexing-
As mentioned earlier, one way to increase the amount of data a single fiber can carry is to use more than one wavelength of light as a carrier. This is called Wavelength Division Multiplexing, or WDM. The more wavelengths in a fiber, the more data that can be transmitted. There are a number of different versions of WDM being used today - simple, coarse, and dense. All of them use multiplexers and demultiplexers. These devices allow multiple wavelengths to be coupled to and from a single optical fiber.
Simple WDM is the oldest form of WDM. It uses only two wavelengths, 1310nm and 1550nm, to transmit data through a fiber. This form of WDM uses inexpensive uncooled lasers and inexpensive multiplexers and demultiplexers. In some cases specially tuned optical couplers are all that is necessary for this function. Simple DWM has been in use for well over a decade.
Coarse WDM uses up to 18 wavelengths between 1270nm and 1610nm. CWDM is inexpensive compared to DWDM. It can use uncooled lasers and inexpensive multiplexers and demultiplexers.
Dense WDM packs up to 160 wavelengths into a fiber, allowing up to 16 terabits per second of data to be transmitted on a single fiber. DWDM systems presently deployed use nowhere near this many wavelengths, but the capability is there. DWDM is rather expensive to implement because of the tight wavelength tolerances required by the lasers. The lasers are temperature controlled, which minimizes wavelength shifts due to temperature changes. The multiplexers and demultiplexers are also quite expensive and have to exhibit the same tight tolerances as that of the lasers. DWDM is used primarily for long haul connections, particularly in undersea cables.
A good tutorial for DWDM and all of the subjects covered so far can be found here or downloaded in PDF form here.