Fiber-Optic Component Testing
By David W. Fisher
Designers need to understand the standards and examine data defining environmental and mechanical conditions to
ensure they are focusing on the proper system design issues.
Optical fiber is pushing into nearly every segment of the communication industry — from cellular systems to telecommunication systems, from home automation to LANs and WANs, and from on-line learning environments to cutting-edge medical systems. As more information technology managers look for ways to integrate optical technology into their data networks, designers at OEMs are altering existing electronic designs or creating new designs
to accommodate optical fiber. Demands for high performance, reliability, easy upgradeability, and long lifespan require careful systems design, and choosing components to go into these systems requires thorough research.
Because optical technology is relatively young, fiber-optic components do not have an extensive history to help designers forecast what actual field performance will be. Instead, designers must rely heavily on the test results offered by component suppliers. However, evaluating test
results cannot be properly done without evaluating the testing process itself.
Standards bodies such as the Telecommunications Industry Association/Electronic Industries Association (TIA/EIA) and the International Electrotechnical Commission (IEC) define basic testing procedures for optical fiber components in data-networking systems. However, to provide components that meet or exceed more stringent demands put forth by the market, component suppliers will create their own set of performance specifications.
Additionally, these components will need to meet performance specifications set by other industry forces such as leading data-networking equipment providers. Because each supplier offers components that meet different specifications, it becomes difficult for designers to make comparisons. To make an informed, accurate purchasing decision, designers need to be familiar with fiber-optic testing issues and differences in testing practices that can greatly affect performance results.
Testing issues and
criteria
Tests will differ between single-mode and multimode fiber-optic components. Single-mode systems are widely used in service-provider installations, such as cable TV and telecommunications, in which the primary concern is component operation under harsh environmental conditions. Multimode systems are more often used in short-haul, office environments, and tests are tailored more to individual applications and are more complex. Because multimode systems are more often used for data-networking and
premises-cabling systems, we will focus on the testing of multimode components.
When testing components for systems available throughout an industry or that can be used in a variety of systems, the test engineer is more often estimating what industry demands are, what requirements will likely be, and to what conditions the components will be exposed. Because the lab independently defines performance specifications (including critical factors such as temperature extremes, mating cycles, and other
variables that will affect the performance of the component during implementation) testing conditions and performance specifications will vary from supplier to supplier.
There are cases in which designers specify their own performance criteria for a component, especially when they can closely pinpoint environmental conditions and mechanical stresses to which the component will be exposed in their industry segment. There are also cases in which designers may want the lab to measure factors not covered by a
standard. In these instances, the lab will test the component to these specifications. Therefore, it is important for the designer to be intimately familiar with the conditions under which the equipment will be used and to get feedback from the customer base about critical performance factors.
Insertion loss
One of the most important fundamental measurements for fiber-optic components is insertion loss. Insertion loss is taken as a baseline measurement to determine the performance of a component or
cable. It is then remeasured after each test to determine the effects of testing. Several variables can give a skewed or misleading insertion-loss result. This makes it essential for the procedures specified by the test lab to minimize or eliminate the effects of factors such as different fiber-core size and subtle nuances in test procedures.
First, the insertion loss of a component, such as a connector or splice, can be affected if there is a fiber-core size mismatch. For example, if light is traveling
from a smaller fiber core into a larger fiber core, the observed insertion loss may be very near zero because the larger receiving fiber acts like a “bucket” to catch the light. However, if light is traveling in the opposite direction, the smaller receiving fiber would not be able to accommodate all light from the larger fiber and would yield a higher insertion loss. Depending on the size mismatch, insertion loss could be in the order of several dB. For example, transmitting from a
100-µm/140-µm fiber into a 62.5-µm/125-µm fiber yields a loss of approximately 4 dB due to the core-size mismatch.
Secondly, the light source is a critical factor. Tests that use a light source that underfills a fiber often yield lower insertion-loss results than tests that use a light source that overfills a fiber. Use of a component based on misleading insertion-loss results may negatively affect the performance of a communication system and, in extreme cases, may render the system nonfunctional.
Different light-launch conditions affect how light is coupled between two fibers and impacts insertion loss. This condition can be seen in a far-field pattern measurement, which shows how the mode content is distributed in the fiber core.
Whether the light source overfills or underfills the fiber, mode content of long-length multimode fiber systems generally settle to steady-state conditions, also known as equilibrium mode distribution (EMD), after about 1 km to 2 km from the light source. Insertion-loss
readings taken a short distance away from the source can be considerably different from insertion-loss readings 1 km or 2 km away from the source, which can also be misleading.
The TIA and IEC standards both prescribe light-launching conditions under which components should be tested. One of the prescribed launch conditions is TIA’s FOTP-50 (often referred to as “70/70”), provides an alternative means of generating a steady-state condition in fiber optics.
The 70/70 condition is
produced with beam optics and is typically done on an optical bench, which does not lend itself to multicomponent testing in a test-lab environment. Instead, many test labs use the other condition that FOTP-50 specifies, the mandrel wrap, which is a smooth, round device around which the fiber is wound in order to filter the light so that it closely approximates steady-state conditions. TIA standards specify different diameter mandrels depending on fiber-core size.
Table 1
shows the
insertion loss for three test sample groups of an SC-type ceramic fiber-optic connector. Tests were conducted from November 1996 to March 1997. The first two columns show the maximum allowed insertion-loss average for the components, according to criteria established by the connector manufacturer. The last two columns show the actual results of insertion-loss testing, showing the average of each group of samples after ten mating/unmating cycles, as well as the maximum value for any single sample.
Ten
mating/unmating cycles are recommended by the TIA, especially to obtain a baseline reading for nonkeyed connectors. While keyed connectors allow only one mating orientation, nonkeyed connectors do not have safeguards to govern the position of the connector in relation to the fiber or other connectors. As a result, they are more prone to mismatch and, consequently, higher average insertion loss. This requires more careful insertion-loss baseline testing.
Once the baseline measurements are taken, mechanical
and environmental tests are run. Changes in optical transmittance are then measured again to determine the effects on the component.
Reflectance
Reflectance, often expressed as return loss, is the ratio of reflected energy to incident energy. When light traveling down an optical fiber hits a reflective surface, such as a connector endface that is open to air, some amount of light energy bounces back to the source. Thus, a reflectance or return loss of 20 dB means that the reflected energy is 20
dB less than the incident energy. Reflectance has long been a concern in single-mode systems and is becoming a greater concern in multimode systems.
Today, most tests of multimode components do not include reflectance testing. Because multimode optical fiber is a graded-index medium, reflectance is difficult to measure and does not lend itself easily to automated multichannel testing.
The TIA and various component suppliers are working to create a technically sound measurement process for
reflectance. Because reflectance has not historically been a great concern, designers need not be suspicious of test results that do not include reflectance testing. However, in a few years reflectance testing will likely be as common, and as necessary, as insertion-loss testing.
Environmental testing
Environmental testing is the most challenging battery of tests to define because it is difficult to predict the types of environments in which components will be used. More often, component suppliers find
themselves running tests that are more harsh than may be required to ensure that the component will not fail due to environmental conditions. Environmental tests may include temperature cycling, thermal aging, thermal shock, humidity, and combinations of temperature/humidity cycling.
Multimode components are most often used in office environments and don’t generally experience extreme environmental conditions or excessive mechanical stresses. Therefore, environmental and mechanical testing procedures
do not need to be as demanding as those for, say, single-mode components in service provider installations.
Environmental tests are usually run on a single set of test components in succession. For example,
Table 2
shows the test results for the fiber-optic connectors. This connector is intended for use in or between wiring closets or in the back of computer systems. Environmental tests were conducted on test group 1, mechanical tests were conducted on test group 2, and
off-axial pull strength was tested on test group 3.
Both the tests and testing sequence tell a great deal about how a component will perform in the field. In this test, the temperature cycling tests induce thermal stresses that cause component parts to “move around.” In extreme cases, these differential stresses can actually lead to fiber fracture. After exposure to these conditions, components were put into humidity testing, which attempted to introduce moisture — at an elevated temperature.
Moisture is deleterious to glass fiber and, over the long term, is known to induce fiber “decay” or embrittlement. In
Table 2
, the first columns show the results that must be achieved while the last columns show the actual results. In each instance, the components met and usually were better than performance requirements. Testing notes also state that no physical damage to the connector was observed during and after each of the tests.
The better a designer is able to
pinpoint the environment in which the component will be placed, the easier it is to examine these test results. Components that will be used under harsher conditions will need to undergo more extreme testing procedures. For example, for a single-mode component intended for use in an outdoor service-provider installation, the procedure may specify extreme environmental conditions. The component may be exposed to serveral types of tests. The first may be thermal-age testing, which dries out the materials. The
second may be humidity testing, which attempts to introduce moisture to the system. And the third may be thermal cycling, which causes the system to expand, contract/ freeze, and thaw any moisture that may have been introduced by the previous test.
For thermal cycling, it is necessary to examine the range of temperatures at which the component was cycled, the number of cycles, the length of time at which the component was ramped down, and how long the component remained at extreme temperatures. For
humidity testing, primary concerns include length of time for which the component was preconditioned, temperature and humidity to which the component was exposed, and the length of exposure.
Mechanical testing
Mechanical tests examine various mechanical stresses on fiber components. These tests might examine the effects of physical shock, vibration, tension, compression, durability, twist, and flexing.
Table 2
shows that the connector was tested for coupling mechanism
strength, cable flexing, twist, and durability. Again, because this is a component for use inside an office environment, extreme mechanical testing is not required. These tests were run according to TIA standards.
To test the coupling mechanism strength, the detector side of the test-bed samples was secured by wrapping cable around a 3-in mandrel at a minimum distance of 8 in behind the strain relief. A 7.5-lb tensile load was applied to the detector side of the sample and was removed after 1 minute. Samples
were unmated, inspected, cleaned, and remated. Optical-transmittance measurements were taken before the load was applied, 30 seconds after the load was removed, and after cleaning and inspection.
To test the component performance after the cable to which it is attached is flexed, the cable was flexed for 500 cycles at rate of 15 cycles per minute, with a tensile load of 1.1 lb applied to the cable. Measurements were taken before testing and after every fifty cycles with the load removed.
To test the
component performance when the cable it’s attached to is twisted, the cable was twisted for 10 cycles at a rate of 30 cycles per minute, with a load of 5.51 lb applied to the cable on the source side. Measurements were taken before and after testing with the load removed.
To test the durability of the component, the sample groups were exposed to 500 mating/unmating cycles at a rate of less than 300 cycles per minute. Measurements were taken before the test and after every fifty cycles.
On the
third group of samples, the lab tested the off-axial pull strength of the connectors. This is a side load test in which the cable, attached to the connector, is pulled toward the floor and measurements are taken for insertion loss.
Usually all mechanical testing is conducted on a single sample group of components. However, tests may be conducted on a separate group if the lab determines that it is unreasonable for any component to be expected to be put through such severe mechanical stress. Designers may
also see that a separate sample group is used if the test was not believed necessary until all other testing was completed.
Test documents will detail required results and actual results. A separate product specification details the test procedure and the standards by which those procedures were determined. Designers need to inquire about how the test lab determined the required results, for which applications these results are intended, and how components meeting these requirements will operate in your
system under expected operating conditions.
Design assurance department
The testing lab is an important technical resource, but just as important is the design assurance department, which oversees the procedures and tests that are run by the lab.
The design assurance department helps to determine test procedures depending on standards, system requirements, customer demand, and industry feedback. Most importantly, it ensures that tests are conducted on components that are pulled from the
normal production process so that the sample group comprises an unbiased representation of products, rather than products that have been specially altered or designed for the test. The sign-off of the design assurance department on testing results assures designers that the components they purchase can be expected to perform similarly to those tested.
Designers should be aware of how the design assurance department functions, how heavily involved it is in the testing process, whether it outlines the
procedures that are run, and how it ensures that testing samples are normal products taken from the assembly line. Designers should also ask a representative of this department about how it determines whether requalification testing of new versions of components is required.
Study the standards
To begin to gain an understanding of the issues involved, designers can study the standards by which a component should be tested. Standards and test procedures from the TIA and, on a global level, the IEC are
the most critical for components for fiber-optic data networking systems. The standards specify requirements that the product must meet and usually refer to test procedures that may be used. Other agency bodies such as military-based organizations also contribute to written processes for fiber-optic testing (see Resources, “Key Standards Organizations for Fiber-Optic Testing”).
However, standards alone do not a test procedure make. Designers — and the test labs — also need to examine
data defining environmental and mechanical conditions in their field to ensure that they are focusing on the proper issues. It is also important to share experiences with other designers who have experience implementing fiber-based technology in their systems.
David W. Fisher is the supervisor of quality control/design assurance for the optical cable and accessories division of AMP Inc. He has a BSEE from Penn State University, PA and has 14 years of experience in the testing of fiber-optic
components. Fisher represents AMP in TIA FO6.3. He is also a member of the American Society for Quality and is a certified reliability engineer through that organization. He can be reached at david.fisher@amp.com.
Resources
Key Standards Organizations for Fiber-Optic Testing
American National Standards Institute (ANSI), www.ansi.org
American Society for Testing and Materials (ASTM), www.astm.org
Bell Communications Research (Bellcore), www.bellcore.com
Electronic Industries Association
(EIA), www.eia.org
Insulated Cable Engineers Association (ICEA), (508) 394-4424
International Telecommunications Union — Telecommunications Standards Sector, ITU-TSS
International Electrotechnical Commission (IEC), www.iec.com
National Fire Protection Association (NFPA), www.nfpa.org
Rural Utilities Service, www.rurdev.vsda.gov/rus/
Telecommunications Industry Association (TIA), www.tiaonline.org
Underwriters Laboratories (UL), www.ul.org
Puzzled by a network processing design issue?
