The final stage in the installation of a fiber-optic cabling system is qualification testing. Historically, such testing has consisted of optical time-domain reflectometry tracing and optical-loss or attenuation testing.
The cost of this testing is an important consideration, especially when fiber counts rise. In many larger installations, hundreds--and in some cases, thousands--of fibers must be tested. In which of these cases should you use an optical time-domain reflectometer rather than an optical-loss test set, and why is it wise--in some cases--to leave tradition behind?
Justification for the use of optical time-domain reflectometry seems obvious. A properly interpreted reflectometer trace indicates overall length, individual localized stress points and an estimate of the attenuation of the fiber from end to end. The hard-copy printout of the trace becomes a map of the fiber link and may be used in the future for measuring any degradation of the system.
While all these justifications may seem reasonable, there may be good reasons to re-evaluate the need for all this data.
In earlier days of fiber installation, there were fewer experienced installers, fewer high-quality cable suppliers and little historical information on which to build confidence. Most fiber installations were for telephone, involving at least several kilometers of fiber that required field splicing. Because optical time-domain reflectometers were routinely used to monitor splicing operations during installation, getting the equipment on site required no additional expense. The traces could be taken during installation instead of being done as an add-on test at the end of the project. Finally, the lengths of the systems were great enough to minimize the inherent dead-zone problems in the reflectometer readings.
Many installations today, however, involve fiber lengths of less than 1 kilometer. More often, the intrabuilding networks use fibers of only a few hundred feet. Minimal field splicing is required except in telephony systems. Therefore, reflectometer testing may not be economically justifiable.
It typically takes at least 3 minutes for a skilled OTDR technician with high-quality equipment to obtain one trace. Printing out the trace adds another 2 to 3 minutes per fiber. In a system of only 48 fibers, this adds approximately 5 man-hours, including set-up time, to the testing procedure. If the testing is performed at two wavelengths and from both ends of the fibers, the time and cost increase fourfold.
Optical time-domain reflectometer testing reveals:
End-to-end fiber length
An approximation of attenuation
Continuity of the fiber
Kinks or pinches in the fibers, beyond the reflectometer dead zone.
Furthermore, if the tests are performed from one end only, and the technician is sufficiently skilled, then the relative quality of one of the system connectors is known (the connector on the launch end). In addition, users have a hard copy of the trace for their files.
OTDR testing does not reveal:
The throughput loss of the system at the operating wavelength
The absolute quality of the connectors on either end of the fiber.
Optical-loss, or attenuation, testing is necessary for qualifying the system. It is the only test that indicates how the fibers will perform in the network under actual power conditions. Unlike an OTDR reading, the optical-loss test set measures the difference between a known quantity of power input and a measured quantity of power output. When the proper instruments are used for this test, the quality of fibers, connectors and connector adapters in the link is ensured.
Optical-loss test sets come in various styles, but the higher-quality sets provide light sources and receivers that operate at the same wavelengths as the optical networks--850, 1300 or 1550 nanometers. Many popular sets have dual-wavelength capability at 850 and 1300 nm. The receiver (detector) unit in the test sets usually includes a digital readout with a 0.1-decibel resolution.
Using this type of tester, you can obtain enough information to ensure that the installed fiber system meets qualification standards without the need for optical time-domain reflectometer testing. Because actual-loss testing is necessary--whether or not reflectometer testing is performed--design-loss testing is the least costly method to qualify the installation to design specifications.
Optical-loss testing is less time-consuming and less technically challenging. Based on field experience, it requires two technicians to perform the test--one on each end of the system. Each fiber test takes approximately 30 seconds. Set-up time for each termination site takes almost 10 minutes. If the theoretical 48-fiber system consists of one main distribution frame and four subdistribution frames, and the testing is performed at two wavelengths, the estimated time for labor is approximately 1.6 man-hours for testing and 1.6 man-hours for setup.
Optical-loss testing reveals:
Throughput loss in decibels of the fiber link at the wavelength of system operation
Continuity of the fiber
Quality of the connections at each end of each fiber
Quality of the installed panel adapters.
The optical-loss test does not reveal:
Pinches or kinks in the fibers.
The missing data about the presence of pinches or kinks in the fibers is important. Properly applying loss data to the specifications for the system fiber and the connectors can determine if pinches or kinks exist.
Accurate optical-loss testing of the completed fiber network, therefore, is sufficient to ensure that the installation, including connectors, cable and couplers, is satisfactory. If done properly, this system-qualification test can substitute for the more-expensive optical time-domain reflectometer test.
Attenuation Testing and the OTDR
Historically, an optical time-domain reflectometer has been used to measure end-to-end loss on installed optical fibers. This practice should never be considered a substitute for actual power-in/power-out attenuation measurements. Although the instrument may provide a usable approximation of true fiber loss, variables in the method can affect results. Following are some issues to consider:
Optical loss is wavelength-dependent. If the reflectometer`s light source is not centered on the actual operating wavelength of the fiber system, the attenuation readings will be inaccurate. The percentage of error is not easily predictable because the actual source wavelength and the loss characteristics of the fiber at non-specified wavelengths are often unknown.
An unavoidable dead zone at the reflectometer-to-fiber interconnection is caused by the Fresnel reflection at the interface. It may be minimized but not eliminated. Because of this phenomenon, the attenuation measurements obtained on the test instrument do not include the loss attributable to the input connector or to any anomalies that may exist in the dead-zone area. This area may include 3 to 200 meters of fiber, depending on the reflectometer and the conditions of the interface. These factors, especially the input-end connector, may be major factors contributing to loss in the fiber link.
(The quality of the input-end, or launch, connector may be noticed by a very skilled technician, but only in relation to connectors of other fibers being tested on that end. There is a drop in the input-power amplitude if the input connector is bad, but it may not be readily apparent.)
The optical time-domain reflectometer is a one-end tester; it launches pulses of light into the fiber and analyzes the energy that returns from the fiber. For this reason, the trace gives no information concerning the loss at the connector on the far end of the fiber under test. As with the launch-end connector, this can be a major loss factor in the fiber link.
The reflectometer is a specialized testing tool. Although the manufacturers of these instruments have added many easy-to-use features to the latest testers, they are still complex. The information obtained and reported from this device requires proper interpretation, and many network fiber users do not have the training to do so.