Last month’s article dealt with testing copper cabling systems.This month, we examine the performance and testing requirements for installed optical fiber systems. Testing and performance requirements for optical fiber cabling systems were established in Annex H of TIA/EIA-568-A in 1995. Unlike copper, optical fiber established performance requirements for not only horizontal cabling, but also for backbone cabling. With the publication of TIA/EIA-568-B.3 “Optical Fiber Cabling Component Standard” and the approval of TIA/EIA-568-B.1 (SP-4425) “Commercial Building Tele-communications Cabling Standard,” these initial requirements have been refined and enhanced.

Unlike copper systems that require eight electrical performance tests, optical fiber systems only require one test—attenuation. Additionally, copper testing specifies both “permanent link” and “channel” testing, whereas optical fiber testing only requires “link segment” testing. The optical fiber “link segment” is equivalent to the copper “permanent link,” but given the consistent performance and limited impact associated with TIA/EIA-568-B.3- compliant patch cords to the channel, the standard allows the “link segment test” to represent the resulting channel performance when the system patch cords are installed. This does not mean that non-compliant patch cords will not affect the channel, because they can have dramatic effects; however, complaint patch cords will provide similar results to those obtained with the test patch cords used during the link segment testing.

One change that has occurred with the revision has been the requirement on the light launch condition of the test equipment. Previously, the launch condition was for what is known as a Category 1 (overfilled) launch. Now the requirement is for launch conditions, as established by ANSI/EIA/TIA-455-50B (mode-conditioned) launch.

This launch condition is typically obtained by mandrel wrapping the launch fiber prior to and during test and removes the transit higher-order modes of light that can produce erroneous higher attenuation results. This launch can either be obtained inside the test equipment or by mandrel wrapping the test equipment cords. This is accomplished by wrapping the test equipment jumper attached to the light source five times around in non-overlapping warps around a smooth mandrel. For 50-micon systems, the mandrel diameter is specified as 0.9 inches and for 62.5-micron as 0.7 inches. By following this requirement, the test results will not only be more consistent and accurate but also better represent the system’s performance, such as whether it uses light-emitting diodes (LEDs) (such as Fast Ethernet) or uses vertical cavity surface emitting lasers (VCSELs) (such as Gigabit Ethernet). This is because single-mode fiber operates with only one mode of light, that this mode conditioning and corresponding mandrel warp is not applicable.

The performance and test procedures for optical fiber systems remain relatively unchanged with TIA/EIA-568-B.1 approval. First, the test procedure still specifies a one-jumper reference and a two-jumper test method. This procedure ensures that the system under test will contain two more connector pairs than the reference test, which represents what the system electronics will experience in actual operation. This sort of reference is easily established when the connector interface on the test equipment is the same as the system under test. But unfortunately, this is not always the case, especially now with the allowance of small-form factor (SFF) connectors; however, this exists also with just ST-Style and SC connectors.

Many fiber optic system providers and test equipment suppliers have established acceptable modifications to the test reference set-up to accommodate unlike connectors between the test equipment and the system. The over-guiding principle is that the system under test should contain two more connector pairs than the reference. Therefore, the reference could be a two-jumper (one-connector pair) reference with a three-jumper (three-connector pair) test. Additionally, some manufacturers have introduced test equipment with selected SFF interfaces that simplifies the reference set-up and reduces the number of variables in the test.

Both backbone and horizontal link segments are to be tested in only one direction. For dual-fiber systems, specifically SFF systems, the common practice is to test one fiber in one direction and the second fiber in the opposite direction. This is entirely acceptable and actually preferred because it also then tests for proper polarity of the fibers. Horizontal and centralized cabling systems need only to be tested at one wavelength (either 850nm or 1300nm for multimode) because of the short length of the system resulting in only minor differences between the two wavelengths. In the case of both horizontal and centralized cabling systems, a single attenuation performance value is given for system acceptance. These values are stated below and are dependent upon whether the system is implemented with a consolidation point (additional connector pair) used in open office cabling configurations. These numbers are based on acceptable attenuation results resulting from complaint cable and connectors for these various configurations.

Because optical fiber is the predominant media in the backbone, the committee felt it essential to provide system performance and test requirements for the backbone, unlike copper. The only difference between testing the backbone from the horizontal is that the backbone should be tested at both wavelengths (850nm and 1,300nm for multimode or 1310nm and 1,550nm for single-mode). The backbone performance requirements are actually based on a formula versus a single number as with the horizontal and centralized cabling requirement. The formula takes into consideration the length variations of the backbone as well as the possibility of splices being contained in the backbone. Simply put, the acceptable link attenuation is equal to the sum of the attenuation of the cable, connector pairs, and splices.

Connector and splice attenuation is a straightforward calculation. For connectors, it is the number of connector pairs times 0.75 dB and for splices it is the number of splices times 0.3 dB. Cable attenuation varies based on the length of the system and the attenuation coefficient (dB/km) at the appropriate wavelength. Figures 1 and 2 provide a graph of this equation for a multimode (50 or 62.5 micron) and single-mode backbone systems that contain two connector pairs and no splices. The values would simply be increased at corresponding 0.75 dB for each additional connector pair or 0.3 dB for each splice.

While optical fiber systems are quick and easy to test, care must be taken to properly establish the reference. This requires high-quality test jumpers and that you maintain the quality during the test by cleaning. Also, when testing systems with a different interface from the test equipment, ensure you have the proper jumpers.

BEAM is director of systems marketing at AMP NETCONNECT Systems. He can be reached at (336) 727-5784 or tebeam@tycoelectonics.com.