I am always surprised to find out how many field technicians have never heard of mandrels for multimode fiber testing, much less the latest encircled flux method.
No, a mandrel is not a type of monkey—that’s a mandrill, a primate from Africa that’s closely related to the baboon. And encircled flux is not something that will put you in the hospital—that’s inflammatory flux, a nasty disease of the intestines.
However, with the lack of knowledge, some technicians working in the data center space might feel like they do in fact have a monkey on their back or a bad case of intestinal flu when confronted with these two fiber testing methods.
Not to worry… you just need to know the basics.
During fiber testing, LED light sources tend to overfill the fiber, leading to pessimistic readings with potentially false link failures due to higher insertion loss. On the other hand, underfilled launch conditions can show artificially low insertion loss values that can later lead to unforeseen network problems.
LED light sources tend to overfill a fiber, while lasers tend to underfill. Consequently, standards specify LED light sources for testing.
Tightly bending the launch cable around a mandrel (or rod) has been considered an effective method for overfilled launch conditions. This essentially removes the unwanted modes outside of the fiber core that are more susceptible to the bending losses, resulting in more accurate results.
The size of the mandrel and number of turns determines the effect. For 50/125 multimode fiber, TIA standards recommend five wraps around a 22 mm mandrel for 3 mm jacketed cable and around a 25 mm mandrel for 900 micron buffered fiber.
For bend-insensitive fibers, it’s important to refer to manufacturer recommendations as these fibers typically require much tighter bends to eliminate the unwanted modes.
Today’s higher data rate fiber links place more stringent requirements on the maximum allowable loss for a channel, calling for a new fiber testing method that more accurately measures insertion loss and narrows the variability.
The encircled flux method involves the use of launch controllers that replace the mandrels to precisely launch the right amount of light that more closely matches the laser launch condition of the actual transceiver.
Encircled flux testing requires commercially-available EF-compliant launch controllers. These are specially constructed test-grade reference cords fitted with modal conditioners that restrict the number of modes launched from the test cord.
If you use the mandrel method, you will likely experience much greater variability due to differences in the cable and mandrel wrapping.
For testing high-speed data center fiber links where loss budgets are tight, standards recommend the EF method for more accurate testing on both standard and bend-insensitive multimode fiber. And EF is the test method now specified for 40 and 100 gigabit applications.
If cost of EF-compliant launch controllers is a concern, the mandrel method can be used for the majority of links with the EF method used for a sanity check, statistical sampling or in the event of inconsistent results with repeated mandrel testing on the same fiber.
When choosing pre-terminated cabling, it is also best to choose a manufacturer that uses EF-compliant factory testing—it will help ensure you get the insertion loss you think you’re getting.
Click here to find out more about Belden preterminated fiber solutions that are all subjected to EF testing in a factory environment.
Greg Funk began his career in 1984 as a distributed computer network analyst. Serving Belden for more than two decades, Greg is one of Belden’s data center experts and a recognized speaker at industry seminars and events. He has fulfilled a variety of functions and today is a systems engineer providing network engineering sales support in his region.