The problem always cited with optical time domain reflectometer (OTDR) measurements, especially on multimode premises cable plants, is they generally do not agree with insertion loss measurements made with a light source and power meter. The indirect measurement of the OTDR depends on the backscatter of the fiber, which may not be a constant from fiber to fiber. In addition, the laser source of the OTDR does not fill the modes of the fiber in quite the same way as most test sources, and loss of both the fiber and connectors is highly dependent on mode fill.
But the renewed interest in using OTDRs in premises applications has spurred some research into the problem that may help bring OTDR measurements closer to insertion loss measurements. At a recent international standards meeting, I received some information from work done in the United Kingdom that indicates that some changes in test methods may finally offer hope for correlation of OTDR and insertion loss measurements.
First, one must deal with the modal distribution of the test source. Mode fill is a way of describing how light is carried in the core of the fiber. Light traveling in the center of the fiber travels a shorter path than light carried in modes that go all the way to the outside of the core and travel longer paths.
The longer path the light takes going through the fiber, the higher the loss. To measure loss consistently, it is necessary to have the test source conditioned to always launch light the same way.
Even insertion loss measurements cannot be consistent unless the source launch power is properly conditioned, so multimode test standards specify the source characteristics to ensure all sources are similar.
The same methods used to condition light sources for insertion loss measurements, mode scramblers and filters, usually implemented by a mandrel wrap of the launch fiber, can be used with OTDRs. The mandrel wrap method is not as consistent as one would like, as its effect may depend on the cable design of the launch cable, including the stiffness of the jacket and amount of strength member fill in the cable. A more consistent method, using mode conditioning patch cords, is now being used in Europe and is becoming available in the United States.
Modal conditioning, however, only affects the outgoing test pulse. At the current time, we can only speculate on the mode fill of the backscatter signal that sends light back to the OTDR for measurement.
Educated guesses are that the backscatter light has a higher mode fill, making loss in the return direction higher than the conditioned outgoing pulse. Enough interest in this question exists that it should be researched and answered in the near future, but while we may be able to understand it, we have no way of controlling it.
Having dealt with mode conditioning, we now have to deal with the different backscatter coefficients of different fibers. Scattering is the primary loss mechanism of fiber and the light scattered back to the OTDR provides the mechanism for OTDR testing.
The backscatter coefficient is a result of two major fiber characteristics, core diameter (or more correctly, mode fill in the core) and the material itself. In addition, backscatter is wavelength dependent, but the wavelength of the test source is controlled by the manufacturer of the test equipment, which chooses sources in the proper range.
A difference in backscatter coefficient in two fibers causes a measurement error at joints (connectors or splices) between the two fibers. If more light is scattered after a joint, the measured loss will be less, or even show as a “gainer.” If less light is scattered, the measured loss will be higher. Either case can cause significant measurement error.
But when testing a single cable with the OTDR, one has a launch cable on the OTDR and a receive cable on the far end of the cable under test, with both launch and receive cables made from a single fiber cut to length; the differences in backscatter coefficient cancel out. Simply cutting both the launch and receive cables, one after the other, from the same spool of cable allows the OTDR to test individual cables end-to-end without worrying about the backscatter coefficient errors.
All this may sound complex, but it is no more complicated than correctly setting up a light source and power meter for insertion loss testing. Test results I have seen from the United Kingdom indicate that on premises length runs, OTDRs can offer reasonably good correlation to insertion loss tests.
That does not mean that we recommend using an expensive OTDR to test your installed cable plant when a light source and power meter will make measurements with less uncertainty, but it does mean that the confusion caused by OTDR and insertion loss measurement differences will be less for those who choose to additionally use OTDRs for their troubleshooting capability. EC
HAYES is a VDV writer and trainer and the president of The Fiber Optic Association. Find him at www.JimHayes.com.