Modular instrumentation speeds BER and insertion-loss testing

Modular instrumentation can be defined as hardware and software platforms that allow a variety of functionally specific modules to be rapidly and easily incorporated to deliver a unique measurement or process function. In the fiberoptic industry, modular test platforms have many product formats, technologies, and performance variants, including single-port passives, multiport passives, amplifiers, sources, and modulators. However, the core building blocks in test and measurement are common—tunable lasers, power meters, switches, optical-to-electrical converters, bit-error-rate testers (BERTs), and time-domain analyzers. In the current business climate, this instrumentation must be able to respond to changes in production volumes and product mix.

The strength of a modular platform is its ability to rapidly reconfigure to meet specific measurement requirements. In many cases, a test set can use most, if not all, of the modules from a previous deployment or development phase. As a result, time to market, cost of test, and capital budgets can adapt to the current business environment.

Modular test-set design is not a new concept. Test engineers have long been preoccupied with efforts to efficiently redeploy software elements and capital equipment. When used successfully, capabilities within the test-bay environment can be added and removed without substantially affecting other test processes. While some forward planning remains key, modular test instrumentation can be an enabler of this functionality.

Investments in measurement methods, software algorithms, and capital are made during the entire R&D and production life cycle of a product. Modular software-design philosophies have reduced development times, and decreased maintenance cost. Deploying modular instrumentation can enable similar benefits.

As the product phase changes, the key focus of the test engineer also changes (see table, p. 24). With a modular test instrumentation platform, the test engineer can develop a hardware evolution plan. A broad platform that enables test automation and meets reliability criteria efficiently transfers much of the investment in development and preproduction to production.

For example, extra slot capacity allows incremental instrument additions without dramatically altering rack configurations. Losing a test bay for a mechanical retrofit can cause significant schedule slippages. From an up-time standpoint, hot swappable modules, and field upgradeable and replaceable power supplies ensure test bays are always operational and minimally affected by failures or recalibrations. Finally, the software command commonality provided by the platform drivers can substantially aid in the development of test scripts.

Through proper expansion planning, software and hardware choices can work together to provide improvements in capability as well as capacity. For the purposes of this examination, a base system capacity of eight test modules per mainframe will be assumed. Additional modules can be added by increasing the number of main frames deployed in the test bay.

A potential test configuration for a 10-Gbit/s transponder is a loop-back configuration (see Fig. 1). An optical-pattern generator delivers a test pattern to the diagnostic input through an optical attenuator. The test pattern is retransmitted by the transponder through an additional attenuator. The signal can be studied on both the optical error detector and a communication analyzer by way of a 50:50 splitter. With this configuration, receiver sensitivity on the transponder input can be verified, in addition to inspecting the quality of the transmit eye. The majority of these test elements can be provided by modular platforms available on the market.

New features and capabilities can easily be added (see Fig. 2), such as the hardware needed for a dispersion penalty test. From a software perspective, two simple sequential BER tests must be performed at the output of the transponder; once without the fiber in the path and once with it in the path. With the script for the first BER already developed from the earlier version of the test set, it is really a matter of test sequencing. The dispersion power penalty can be calculated from the two BER curves generated.

Adding two 1 × N switches achieves further expansion of the test capability. At this point, the base test capacity can be scaled to meet production throughput and capital cost objectives. Again, only incremental software is needed from a test method and algorithm perspective, freeing up time for the test engineer to address data archiving and database integration.

The test-bay design should specify at least three additional module slots in the first configuration. This allows quick and easy introduction to the test stand without dramatically taking it offline for mechanical retrofit, or spending a significant amount of time in re-engineering software.

A similar process applies to a typical insertion loss test on a multiplexer/demultiplexer (see Fig. 3). In this case, a tunable laser is stepped across the region of interest with power measurements recorded before and after the device under test. While cost-effective and easily deployed, the system suffers from a significant amount of manual reconnection. This can add uncertainty and will obviously affect test time.

One possible migration path adds multiple power meters to reduce reconnection time and allows all ports of the multiplexer/demultiplexer to be tested (see Fig. 4). The addition of two switches enhances throughput. The use of a four-output ganged switch configuration simplifies switch integration.

Matthew Adams is manager of application development at JDS Uniphase Instrumentation, 3000 Merivale Rd., Ottawa, Ontario K2G 6N7 Canada. He can be reached at [email protected].