Many fiber optic specifications and tests refer to wavelength because most parameters have some wavelength dependency. Attenuation, for example, is caused by absorption and scattering. Absorption occurs at certain wavelengths and scattering gets less significant at longer wavelengths. What exactly does all that mean?

Let's start by defining wavelength. Wavelength is the scientific term for what we perceive as color, actually referring to the length of the light wave of that color of light. When we say red, a scientist thinks about light with a wavelength of 650 nm. Green is about 500 nm, yellow is 570 nm and blue is 475 nm.

The term “nm” or “nanometer” is a measure of length equal one billionth of a meter, which compared to the diameter of a glass optical fiber at 125 microns (m) or 125,000 nanometers, is pretty small. Sometimes in fiber optic literature, you will see wavelength expressed in microns, which are 1,000 nm. Thus, red light at 650 nm becomes 0.65 m.

The human eye is limited in its ability to see light to a small fraction of the wavelengths of light. Those are the most important ones to us because they represent the light given off by the sun. That means we can see well by sunlight, greatly increasing our chances of survival during daylight hours or bright moonlit nights.

Below the visible range, we find ultraviolet light, which causes sunburn. At shorter wavelengths we find X-rays and gamma rays-all part of the spectrum of electromagnetic radiation of which light is but a small part. At longer wavelengths, we have infrared light and radio waves.

We use the infrared just above the region of visible light for fiber optics, because of the transmission properties of glass optical fibers. As you should remember from your first basic fiber class, there are two major causes of attenuation in glass fibers: absorption and scattering.

Absorption is caused by materials in the glass fiber absorbing light at certain wavelengths, especially any water left in the glass after processing. Scattering is a bigger problem, caused by light hitting molecules in the material and bouncing off to the sides of the fiber where it is lost in the cladding. Scattering is much lower at longer wavelengths, so it makes sense to use the longest wavelengths of light possible to transmit signals so it has the least attenuation, by choosing wavelengths that are between the absorption peaks.

The result is three low-attenuation windows available for transmission in optical fibers at 850, 1,300 and 1,550 nm. Multimode fiber uses 850 and 1,300 nm, while single-mode fiber is specified for 1,310 and 1,550 nm by designing a small enough core in the fiber that only transmits one mode of light at those longer wavelengths.

Did you ever wonder why multimode fiber is specified at 1,300 nm and single-mode at 1,310 nm? There must be some technical reason, right? Multimode fibers use LEDs, which have broad spectral outputs. Their output peaks at 1,300 nm, but they have significant light output from 1,250 to 1,350 nm. The difference of a few nanometers in center wavelength makes little difference.

Lasers have a very narrow spectral output and are more carefully specified, as they must have a wavelength greater than 1,260 nm to assure single-mode operation and lower than the water peak at 1,340 nm.

Traditionally, lasers are specified as 1,310 +/- 20 nm while LEDs are around 1,300 nm. Exact wavelength specification however, came from the first manufacturers of lasers whose devices turned out to be 1,310 nm. Every device since then has been made to that specification.

Not only is the attenuation of the fiber wavelength-dependent, so is the dispersion or bandwidth of the fiber. Manufacturers optimize both multimode and single-mode fiber to a particular wavelength, so the person specifying the fiber needs to know what kinds of systems are going to use the fiber in order to choose the proper one.

Of course, all testing must be done at the specified wavelengths. Test equipment must be chosen to cover the proper wavelengths and controls set properly when making measurements. Remember that multimode fiber may have a loss of 3 dB/km at 850 nm but only 1 dB/km at 1,300 nm. Likewise, single-mode fiber will have different losses at 1,310 and 1,550 nm. Even small differences in test source wavelength can make a difference in long distance links, where a 1,290 nm laser may have 3 dB more loss than a 1,330 nm laser in a 50 km length of single-mode fiber. EC

HAYES is a VDV writer and trainer and the president of The Fiber Optic Association. Find him at www.JimHayes.com.