Is your box filled with gems or junk?

Communications News, April, 1998 by David W. Fisher

The testing of the fiber optic components inside fiber optic data network system boxes is critical to the data network's overall performance.

To ensure that new fiber optic networks perform as quickly and efficiently as possible, IT managers conduct several tests on network components, including cables, wall outlets, breakout boxer, and connectors. For mission-critical, time-sensitive applications, they will also search high and low for the best-performing data network systems available.

Putting together these system is no simple task, and much of it depends on using the right components inside the box. There are several components that must perform optimally for the system to perform optimally in a network. These components are typically connectors, connector adapters, and transceivers.

Unlike copper technology, optical fiber technology does not have several decades of history behind it, which makes it difficult for IT managers to project how components will perform under different conditions. Instead, component suppliers establish their own labs for testing component performance and make the results available to equipment designers to help them make comparisons to other components.

To identify procedures and criteria for data network system components, these test labs begin with standards and suggested test procedures specified by national and international standards bodies such as the Telecommunications Industry Association/ Electronic Industries Association (TIA/ EIA) and the International Electrotechnical Commission (IEC, http://www.iec.ch). Bellcore (http://www.bellcore.com) suggests standards for service-provider and telecommunications networks.

Then, to ensure that the components meet or exceed more stringent demands made by both the market and customers, product engineers are often forced to create their own set of performance specifications.

In some cases, the lab also may test the components against specifications set by other industry forces such as leading data networking equipment providers. As a result, components in the industry are tested against different sets of specifications, and it can become difficult to make comparisons between components from different suppliers.

In addition, use of a component based on misleading performance results may negatively affect performance of a system, and may even render a system nonfunctional. It is critical for both the equipment manufacturer and the IT manager to have an understanding of test procedures to avoid varied or misleading performance results.

Single mode fiber systems are most commonly used for cable television and telecommunications. IT managers will find single-mode fiber in intrabuilding campus wide-area networks and even in backbones.

Because most of the network for these applications must endure outdoor conditions, single-model fiber components and equipment must be able to withstand harsh environments and mechanical stresses.

Multimode systems are most commonly used for data network equipment and short-haul communications, such as within an office building. Because these environments are more controlled and predictable, environmental and mechanical stress testing is not as strenuous. Terminating devices are less expensive and are usually simple in design. The larger core size of multimode fiber makes it easier to connect and gives it greater tolerance of components with less precision.

Test labs will run several tests for performance under different environmental and mechanical conditions. Environmental tests can include temperature cycling, thermal aging, thermal shock, and humidity. Mechanical testing can include stresses such as physical shock, vibration, tension, compression, durability, twist, and flexing.

While the performance of the component can be approximated, the conditions to which the components will be exposed is unpredictable. When systems are available throughout an industry in various applications, the test lab must guess the conditions to which the systems will be exposed. Factors such as temperatures, number of mating cycles, and other performance items may not be predictable, so testing of the component is difficult to plan.

Knowing the conditions under which the equipment will function is essential to choosing the proper components to go inside the box, and is therefore essential to choosing the proper box for use within a network.

Important criteria for fiber optic components are:

* Insertion loss. Insertion loss is the optical power lost due to the insertion of an optical component such as a connector or splice. It is taken as a baseline measurement to determine the performance of a component or cable.

Several factors can create different insertion loss results. One factor is fiber core size mismatch. 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 wider receiving fiber acts as a "bucket" to capture the light from the smaller fiber core. However, light traveling in the opposite direction may yield a higher insertion loss because the smaller fiber core is not able to accommodate the light from the larger fiber core.