The majority of premises fiber optic cables use multimode fiber. Many of these cables are backbone cables for high-speed local area networks or data centers, which operate at 1 gigabit per second (Gbps) or higher. Many cables installed today contain high-bandwidth OM3 or OM4 fiber intended for upgrades to 10, 40 or even 100 Gbps in the future.

All these high-speed systems have one thing in common: they require low-loss cable plants. This means they should use the best connectors with the lowest loss and accurate testing to confirm proper installation. High-speed networks also use lasers, mostly inexpensive 850 nm vertical cavity surface-emitting lasers (VCSELs), since light-emitting diodes (LEDs) are too slow for gigabit and faster networks.

Accurate testing of these links, or in fact, any multimode link, requires more care than just about any other testing format in fiber optics. Multimode cable attenuation is highly dependent on the way light is carried in the fiber. It is called mode power distribution, since the rays of light carried in the fiber are called modes. The figure shows the difference between a LED and a VCSEL. Most LEDs launch light into the whole core of the fiber, while a laser only uses part of the core at its center.

The center of the fiber’s core has lower loss than the outer layers of the core, so a laser naturally has lower loss than an LED in any fiber. Furthermore, if the core of the fiber is filled with LED light, it will be more sensitive to alignment at connections, making connector losses much higher. The difference in a short link with several connections can be more than 0.2–0.3 decibels (dB), a significant error when total link loss may only be around 2 dB.

Remember last month’s article on insertion-loss testing? You must connect a test source to a launch cable to measure the loss of the cable plant. Why not use a VCSEL test source—i.e., what the network will use? VCSEL sources can have a lot of variability, and if alignment is not straight down the fiber core, the offset alignment can lead to high and unpredictable loss. LEDs are simply more predictable, and of course, there are tens of thousands of them already in the field being used to test these cables.

Thus, international standards groups have worked for decades on ways to define mode power distribution including coupled power ratio (CPR), equilibrium modal distribution, mode power distribution and the newly approved encircled flux (EF). With the exception of CPR, these methods are so technical that even many of the test equipment companies who need to comply with them don’t understand them, so you should not worry about using those methods.

Most fiber optic testing standards require controlling the test source launch power distribution. The most familiar standard, TIA-568, has called for sources to meet CPR specifications and has added a mandrel wrap mode conditioner on the launch cable. The mandrel wrap acts as a mode filter to remove light in the outer part of the fiber’s core and mixes the remaining light to make the distribution of light in the core more consistent.

Most installers are familiar with this method and know it can make a big difference in loss in some instances, with the measured loss significantly lower than without the mandrel. If you are unfamiliar with it, try this simple experiment. Measure the loss of a cable plant with several connections that have 2–5 dB loss. Then wrap your launch cable around a half-inch (13 mm) dowel five times. Reset your 0 dB reference and retest the cable. I predict you will be amazed—and pleasantly surprised—at the result.

TIA has recently approved a change in the multimode cable plant test standard, OFSTP-14; the change replaces OFSTP-14 with an international standard. This new standard uses the same test methods of OFSTP-14, but also allows optical time domain reflectometer testing under special conditions that will be covered in future articles in this series. It also replaces the CPR and mandrel wrap with an EF requirement. While it certainly makes sense today to use international standards, EF is, in fact, still in development and most test equipment manufacturers are generally not in compliance with it, if they are familiar with it at all.

In the meantime, you should continue using the same mandrel-wrap test methods you have been using.

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