Building Blocks

As the closest thing to the ideal communication medium, optical fiber has become one of the hottest commodities around. In Part 1 of this three-part series, I'll introduce some optical transmission basics.

First and foremost - I want to offer my apologies for question #2 that appeared in the October "Building Blocks" Quiz (check it out online at www.csdmag.com/quiz.htm). Several folks noticed, and I am sure there are more to come, that the right answer was not given. I like my perceived answer, (a), much better than the actual one, but the all-incorrect choices given may have caused confusion in the ranks. My only excuse is that I spent all of my cable modem acronym-ese getting MCNS straight before they changed the name. In any event, nobody will be penalized for this question. And I promise, no one will lose out on a DVD player because of it!

Tour de fiber

This fiber-optics industry is starting to get the "wireless" feel to it. Which means, expect the usual glut of books to follow from unheard of authors, the new buzzwords to appear on professional resumes, the creation of new trade magazines, or, as in the case of some trade journal, name changes.

The proceedings from NFOEC should be interesting collector's items in a few years, as the demise and buy up of the onslaught of start-ups grinds down the enormous list of companies on site represented by their contingent of booth warriors. You can tell which companies are start-ups by the cheesy freebies, including lots of keychains and pens. Everyone is looking to become the next Cerent (not necessarily meaning the desire to be absorbed by Goliath) or invent the next breakthrough technology.

Interestingly enough, if you think hard, recent breakthrough commercial technologies (not laboratory novelties) in optics are relatively few - precisely the reason the field is ripe with so many wanna-bes.

In my most productive booth visit, I asked when and where the company would begin to show product, and what its outstanding features would be. The answer? We aren't ready to announce that kind of information just yet, we're just here to introduce ourselves to people and listen to the industry's needs.

I don't know, but in my book, you aren't ready to even be at a trade show, much less a techie one, if you can't even tell people what you plan to do. Not so in this - and I hate to use this new millennium cliche - space.

All you need to know about this enthusiasm can be summed up in my experience at the after-hours mini-sessions at OptiCon. I wandered over to the session devoted to new advances and ongoing research - basically the work I'd expect to be laboratory novelties at this point, and the nerdy stuff I like. Not only was the session weakly attended, but the moderator and presenters never even showed. We learned, after waiting around for ten minutes, that it had been cancelled. No wonder - next door was a jam-packed, standing-room only session being led by a venture capitalist giving his view on this space from the investor standpoint. I suspect the R&D presenters and the moderator were the folks sitting in the front row.

Investing aside, lets learn a little about the technology. A strand of optical fiber has some of the best qualities we can hope for in a communication medium. It is very low loss, particularly relative to wired alternatives. It has tremendous bandwidth, such that the capacity limitations are driven by what connects to it on either end. It is immune to external interference sources, unlike electrical wiring. And, finally, it has become cost effective to fabricate and deploy, albeit there are still some unique aspects of fiber handling that require more sensitive craftsmanship than shown by the average telephone or cable guy in a house call.

In fact, the low-cost angle can be applied to the devices at the ends of the fiber as well, where sources that drive light into the fiber are dropping in cost just as rapidly.

Light ideas

What this means is that the brightness of light that travels down the fiber corresponds to the strength of the signal that you are transporting. We'll talk about this more in next month's column. However, when you read about optics in textbooks and articles, if it is not qualified otherwise, the discussion assumes the medium is being used as a digital transmission channel to transmit ones and zeroes.

This can be envisioned in a most straightforward fashion - the light is on or the light is off. Of course, there are a bunch of numbers and fancy names of specifications that say the same thing. Extinction Ratio is my particular favorite parameter name. I keep thinking this must have been stolen from a National Geographic special on dinosaurs, or perhaps from a chapter from meteorologist/Internet pioneer Al Gore's environmental catastrophe book. I will talk about some of these terms in a later column, but the theme is as stated above for digital optics: light on = one, light off = zero.

Now, consider that as an undergraduate engineering student, it took two semesters before you understood exactly what a volt was, and that signals were actually carried by electrons (which make up current not voltage) between two different potentials. Finally, remember (do you?) that all of our conventions about current are, in fact, backwards, because it was found, after it was too late, that electrons, and not their positively charged buddies, are the current carriers.

Now that I have successfully made electronic signaling sound way more complicated than it is, I'm sure you will agree that it's easier to conceive of optical rather than electrical transport.

On the same wavelength

The ranges are known as windows because of the way at which these low attenuation portions of the curve sit between higher attenuation obstacles created by properties of the glass used in the fiber. These windows are centered around 850 nanometers (nm), 1300, and 1550 nm. While the 660-nm range is also used, the curve is moving quickly into higher attenuation ranges at this point.

The development of the windows used is a story of chronology. The first systems used 850 nm, the first window, because light emitting diodes (LEDs) - those tiny lights that you see more of than anything else at a trade show - were well developed. In addition, low-cost detectors of light could also be implemented.

Eventually as needs progressed, folks were limited by the loss characteristic, which was close to 3 dB/km. After some initial dueling efforts at wavelengths of 1060 nm and 1300 nm, the lower loss 1300-nm window became the next biggie, dropping losses into the 0.5-dB/km range. Following this was the development of the 1550-nm window, with that effort aimed at exploiting the minimum attenuation region on silica-based fiber on the order of 0.25 dB/km.

If you just add up some simple numbers, you can begin to see the power of optical signaling. For a laser with 0 dBm of optical output power in the 1550-nm window (a pedestrian output level) and a receiver with input sensitivity of -20 dBm (again, mediocre performance), we can afford 20 dB of attenuation of light. At 1550 nm, 20 dB at 0.25 dB/km means an 80-km link (which conveniently rounds to 50 miles). Sure I've ignored some things - namely connector losses and possible dispersion impacts - but you get the idea. That's still a long way to go with a modest equipment set on a connected medium with one transmit and receive pair.

Public transmission

For those of you who read this column regularly, and have some recollection of past material, you may recall pieces I have written about multipath, equalizers, filters, and group delay (Check out the April 1998, July 1997 and August 1997, April 2000 and May 2000, and November 2000 back issues of Communication Systems Design). In those columns, I discussed dispersion at length. With digital optics, the situation is analogous.

Multipath everywhere

In optics, there are two basic fiber types: multimode fiber and single-mode fiber. Multimode came out first, because it was easier to make. It has a big fat glass core, relatively speaking, allowing many different optical modes to successfully travel its length.

For you microwave types, optical fiber is simply a guided wave phenomenon, using different means and materials. But, the idea is the same - create boundaries in the material that cause the signal energy to be contained within the medium, so that it bounces its way down the path.

For microwave waveguides, the barriers are formed by clunky metal plumbing, typically surrounding an air medium. For fiber optics, the barrier is a glass core surrounded by what is known as cladding material to create the proper material boundary needed to reflect light down the glass.

Since we always hear about the need for a fat broadband pipe, one might suspect that multimode is better and provides more bandwidth. In fact, this is not the case, because multiple modes are not desirable, in general. They represent transmission, at a different speed, of the same information using another mode of guided wave through the same fiber. Multiple versions of the same information impinging on the same receiver is not good.

This situation is qualitatively analogous to multipath dispersion in wireless systems, where multiple receiver inputs may be received as signals bounce off of buildings and other obstacles in addition to the direct path to the wireless receiver.

In optics, it is multiple versions of the signal via excitation of different modes in the fiber, a situation that makes things more difficult for the receiver. This is called multimode dispersion. When it exists, it is the dominant dispersion mechanism.

Group delay

The analogous effect in the optical domain is that different wavelengths across the fiber traverse the fiber at different speeds. Since different wavelengths correspond to different colors, this is called chromatic dispersion.

Digging deeper into those group delay and filter discussions, recall that it was signals with bandwidth stretched across the bandpass of a channel that were at the most risk, particularly for a wideband channel, because it is less likely that the delay will remain constant over a wide passband.

Similarly, in optics it is desirable for the source to have a narrow line-width, meaning that the wavelength of the light is highly concentrated around the wavelength window of design and not spread out, the latter being a characteristic of a lower-quality source.

It is also significant to realize that standard fiber, either single mode or multimode, has a zero dispersion wavelength at 1300 nm. In other words, pulse spreading is naturally zero at that point, so it is desirable to operate in that window on that type of fiber. Thus, low loss (1550 nm) versus low dispersion (1300 nm) can become a trade-off. Special fiber (read: more expensive) exists that moves the zero dispersion point into the 1550-nm window.

And that's not all

I also had hoped to tell you all about some of the latest laboratory novelties in this series. That wasn't meant to be from my experience at OptiCon, so maybe I will share with you some of my investment tips instead. I'll even share one ahead of time - buy low, sell high.

About the Author

Rob Howald is the director of systems engineering in the transmission network systems group at the Motorola Broadband Communications sector in Horsham, PA. He has a BSEE and an MSEE from Villanova University, and received his PhD from Drexel University. He can be reached at rhowald@gi.com..

Figure 1: Optical loss as a function of wavelength

1. Goff, D., Fiber Optic Reference Guide, Focal Press, 1996

Return to the Table of Contents