Optical Connector Tuning and Testing WHITE PAPER With the ever-increasing demand for higher performance in the optical network, there is a need for higher performance in optical connectors. Connector performance is typically defined as a measurement of insertion loss that occurs when the connector is mated to a reference connector. Manufacturers use reference connectors for production testing of insertion loss because they provide a repeatable reference point. For a connector with a cylindrical ferrule, a reference connector is one selected for low insertion loss when mated with another reference connector regardless of rotational orientation. The low insertion loss indicates that the cores of the fibers are highly concentric relative to the outside diameter of the ferrules. In the real-life network, service providers are more concerned with the insertion loss of randomly mated connectors. When randomly mated, connector pairs can produce a greater insertion loss than the stated specification of either connector to a reference connector. This occurs when the slight coaxial misalignment of the fiber ends in the individual connectors are in opposite directions and thereby create a greater total misalignment. One of the primary ways to achieve a significant performance improvement in randomly mated cylindrical ferrule connectors is tuning. Tuning decreases insertion loss by improving the coaxial alignment of the fiber ends in intermated connectors. The accuracy of the tuning process has a direct effect on randomly intermated connector performance and is improved through an automated process. Optical Connector Tuning and Testing Optical Connector: Tuning and Testing Page 3 Introduction Transverse, longitudinal, and angular misalignments of optical fibers are the sources of insertion loss in fiber optic connections. Transverse insertion loss is caused by lateral mismatching of the fiber optic core and increases with the square of the distance between the centers of the cores. Sources of transverse insertion loss include ferrule, fiber and fiber through hole concentricity. Longitudinal offsets occur in the axial direction of the fiber and are a logarithmic function of endface separation. Sources of longitudinal insertion loss include high apex offset, fiber withdrawal in a ferrule, and mechanical separation of ferrules under load. Angular insertion losses are a small contributor to connector performance, but if ferrule inside diameter runout becomes large, angular insertion losses will occur. One of the biggest challenges in optical connectorization is the accurate alignment of the fiber ends in a mated connector pair. Microscopic misalignment will produce insertion loss that is measurable and of concern to the end user. The final alignment is driven by the accuracy of the components in the individual connectors and the mating adapter. These components are held to the tightest dimensional tolerances allowed by state of the art manufacturing equipment and processes. Even with these controls, individual components will produce connectors that have fiber alignment that varies from the nominal fiber end position, though the variation is on the order of a micron. When graphed, this variation will produce a gamma distribution. The performance of a mated pair of connectors will depend on which individual connectors happen to be chosen out of a given population. Consider the following example of four fiber pigtail cables terminated with SC ultra physical contact (UPC) connectors in which the lateral offset of each connector is within specification according to insertion loss testing to a master cable. In Figure 1, the fiber end position of each connector in relation to the true center is shown. As stated earlier, there is a direct correlation between lateral offset and insertion loss. The insertion loss for connectors A, B, C, and D is 0.13, 0.03, 0.16, and 0.07 dB respectively if the connectors were tested at 1310 nm. Figure 1. Lateral Offset Connector Fiber core offset from nominal (Microns) Fiber core offset intermated (Microns) Resulting insertion loss A 0.8 0.1 (connector A and C together) 0.002 dB C 0.9 B 0.4 1.0 (connector B and D together) 0.20 dB D 0.6 Optical Connector: Tuning and Testing Page 4 Connectors B and D have the least lateral offset from nominal and would perform the best when tested with a reference connector, but if B and D were intermated, they would have a total lateral offset of 1.0 microns which is equal to an insertion loss of 0.20 dB. This creates a higher insertion loss than would be found when connectors A and C were intermated with a total lateral offset of 0.1 microns or an insertion loss of 0.002 dB (see Figure 2). The Tuning Process From the previous example it can be seen that an improvement in overall intermated connector fiber end alignment can be achieved if the core position was controlled in a single direction for all connectors. Tuning is the aligning of the connector ferrule such that any offset from nominal position of the fiber end is aligned within an angular tolerance zone around a key feature on the connector body (as defined by TIA Fiber Optic Connector Intermateability standards). Therefore, this offset would be in the same direction on all connectors, and the insertion loss between connector pairs would be minimized. A tunable connector has a mechanism for rotating the ferrule in the connector body and locking it into a given position. Most tuned connector designs allow for six or eight ferrule positions. During manufacturing, the connector is mated with a special tuning cable (adjustment cable) that has a fiber end that is deliberately offset a large amount from the nominal position in a specified orientation to the connector key. The mated pair is then tested for insertion loss, rotated to the next position in the connector housing, and tested again. Once all the possible ferrule positions are tested, the ferrule is rotated and locked into the position that had the lowest insertion loss. The ferrule is now oriented such that any fiber end offset is aligned with the offset in the tuning cable (see Figure 3). Figure 2. Intermated Insertion Loss Figure 3. Aligning Fiber Core Offset Optical Connector: Tuning and Testing Page 5 When a tuned connector is intermated with any other tuned connector, any offset of the fiber cores will be in a common direction and sector. Therefore, the insertion loss of the randomly intermated pairs is minimized. It is important to note that the performance of a tuned connector to a reference connector is typically no better than a connector that is not tuned to a reference connector. This is due to the fact that the distance between the core and the true center has not changed. The improved performance is only seen in randomly intermated connectors. In Figure 4, a typical insertion loss improvement for a population of connectors is shown before and after tuning. Figure 4. Tuning Results Optical Connector: Tuning and Testing Page 6 Automated Tuning The final performance of a tuned connector ultimately depends on the accuracy of the tuning process. Two steps must be performed to tune a connector: first, locate the connector under test core relative to a known core position and second, rotate the core to the desired pie shaped region. An automated process is necessary to achieve the required tuning accuracy and eliminate the possibility of human error in choosing the tuned position of the ferrule. Locating the core manually is very time consuming and has limited accuracy, since only the lowest insertion loss out of a set of several readings is used to determine the core position. An automated system can take advantage of microprocessing to perform a non-linear regression and fit the theoretical sinusoidal curve to the data. Regression minimizes measurement uncertainty and allows one to interpolate to find the most likely core position. A typical example is shown in Figure 5. A manual tuning process would have had the 120° angle chosen for the core position, but a curve fit reveals that a better angle would be 150°. Connectors that have fine tuning capability, like ADC’s 12 position tunable SC connector, can therefore maximize tuning accuracy. A non-linear regression analysis also provides numerous statistics that can be used to determine if the connector or test setup is in error. Another advantage of an automated tuning process is that it eliminates the possibility of any damage to the connector endface from any movement of the connector’s ferrule while mated. In a manual tuning process, there is a possibility that the individual performing the tuning operation could try to rotate the ferrule to the next position without fully disengaging the ferrule, thereby damaging the connector and the tuning cable. An automated machine can reliably and quickly perform the second step of the tuning process of rotating the connector to the upper pie region and locking it into place. Manual tuning relies on an operator to rotate the connector in the proper direction and rotate it the proper number of tuning increments. Some connectors pose additional sources of error due to their locking mechanism. For example, the operator is required to press a key onto the FC connector to lock it in place. An automated tuning process integrates this step into the machine, which greatly eases the difficulty of tuning FC connectors. Figure 5. Curve Fit Optical Connector: Tuning and Testing Page 7 Summary The performance of optical connectors when mated to a reference connector is not a true test of how the connectors will perform when randomly mated in a service provider’s network. When randomly mated, connector pairs can produce a greater insertion loss than the stated specification of either connector to a reference connector. This occurs when the fiber ends of individual connectors are misaligned in opposite directions from the nominal position and thereby create a greater total misalignment. An automated tuning process can significantly decrease insertion loss in paired connectors by accurately aligning the fiber end offset of each connector in a common direction. Therefore, when a tuned connector is intermated with any other tuned connector, the insertion loss of the randomly intermated pair is minimized. Web Site: www.adc.com From North America, Call Toll Free: 1-800-366-3891 • Outside of North America: +1-952-938-8080 Fax: +1-952-917-3237 • For a listing of ADC’s global sales office locations, please refer to our web site. ADC Telecommunications, Inc., P.O. Box 1101, Minneapolis, Minnesota USA 55440-1101 Specifications published here are current as of the date of publication of this document. Because we are continuously improving our products, ADC reserves the right to change specifications without prior notice. At any time, you may verify product specifications by contacting our headquarters office in Minneapolis. ADC Telecommunications, Inc. views its patent portfolio as an important corporate asset and vigorously enforces its patents. Products orfeatures contained herein may be covered by one or more U.S. or foreign patents. An Equal Opportunity Employer 102257AE 3/06 Revision © 2002, 2006 ADC Telecommunications, Inc. All Rights Reserved WHITE PAPER . automated process. Optical Connector Tuning and Testing Optical Connector: Tuning and Testing Page 3 Introduction Transverse, longitudinal, and angular misalignments. 0.8 0.1 (connector A and C together) 0.002 dB C 0.9 B 0.4 1.0 (connector B and D together) 0.20 dB D 0.6 Optical Connector: Tuning and Testing Page 4 Connectors