Last month, I discussed one of the most important tools in the fiber tech’s toolbox: the fiber optic power meter. This month, let’s examine the other tool that is equally important: the test source used with the power meter to measure the loss of the cable plant. And, while we’re at it, let’s cover the role of reference cables in testing insertion loss.


Insertion-loss testing needs a test source and power meter to simulate the way a fiber optic link works to measure the cable plant loss. The source should be similar to the source used in the actual communications link. I write “similar” because there will be great variation among sources used in link transceivers.


It’s easier if you are testing single-mode fiber, since the test source will always be a laser of either 1,310 or 1,550 nanometers (nm) like the system sources. For multimode fiber, standards have generally specified using light-emitting diode (LED) sources for testing at either 850 or 1,300 nm, even though actual systems may use LEDs or several types of lasers.


The reason for using LEDs for multi-mode fiber testing is that lasers have problems with noise due to their coherence (that “graininess” you see in laser pointer light) and the way they launch light into the fiber. Generally, LEDs launch light into the multimode fiber filling the core with light while lasers only launch light into the middle of the core, which affects the measured loss of the fiber optic cable plant.


If you have been keeping up with cabling standards, you have probably heard the ongoing debates over “mandrel wrap” and “encircled flux” mode conditioning. Using an LED source with a launch cable wrapped tightly around a small-diameter mandrel will create launch conditions that are repeatable and will give consistent test results. Encircled flux is nothing but a high-tech way of specifying mode-conditioning methods, such as the mandrel wrap, and may not even be applicable to most multimode systems.


Interestingly, with all that concern over modal conditions, a more basic specification of sources can be a major source of error—the wavelength of the source. An 850-nm LED is allowed to have a wavelength in the range of 820 to 880 nm. Over that wavelength range, the difference in attenuation of multimode fiber is large enough that the potential error in loss measurement due to variations in source wavelength is as much as, if not more than, the errors caused by modal conditions.


The wavelength issue is also a problem in longer single-mode cable plants, say more than 10 kilometers (roughly 6 miles). Ideally, the manufacturer should calibrate all sources for wavelength so the actual value of the cable-plant loss could be corrected for the nominal wavelength, or the LEDs and lasers used in test sources could be selected to be closer to the nominal wavelength. Typically, the most expensive test sources will allow you to use either option.


When using test sources, the source needs to be turned on and warmed up for a few minutes until the output is stabilized. The battery on portable sources should be kept charged so the source continues to operate for the duration of tests and the output power does not change.


The other important factor in using a test source is the launch reference cables. Reference cables must match the type of fiber being tested (single-mode or multimode, 50/125 or 62.5/125), and the connectors on the cables must mate to the cables being tested. One should not use reference cables made with the new bend-insensitive fibers because their performance can be unpredictable.


Since the launch cable is going to mate with the cable under test to measure connection loss, it is mandatory that it have a low-loss connector that is cleaned often to keep dirt from affecting the measurements. It should also be tested occasionally to verify it has not been damaged.


Two things to remember about the source and launch cable: After you set the 0-decibel (dB) reference on the fiber optic power meter with the launch cable attached to the source, do not remove the cable because it may change the reference value. Additionally, never stress the reference cable as it will cause loss and change the 0-dB reference.


There is one final device to consider: mating adapters to connect your source and reference cable to the cable to be tested. That adapter is a major factor in connection loss. Therefore, only the best quality adapters should be used (preferably those rated for single-mode, even when testing multimode), and they must be kept very clean.