Test and measurement equipment must stay one step ahead of next-generation components, placing higher demands on the source.
By MICHAEL LARSSON and TRAJAN BADJU, Radians Innova AB
When brewing coffee, the aroma of the coffee beans is what matters to the coffee drinker, not the impurities in the water used. The same principal holds true when testing components or systems; the characteristics of the device under test are what matter to the technician, not the shortcomings of the source.
The high demands put on fiber-optic equipment require reliable testing. The source should not impose limits on what is measured. It must enable fast and cost-effective testing.
The most versatile and widely used sources in test and measurement equipment are tunable lasers. However, not every tunable laser is suitable for all applications. Specific tunable laser characteristics correlate with the testing of certain components.
The reliability of any system, or network, depends on the individual components. Given that optical communication networks incorporate a number of components, the risk of problems is high. Thus, it is vital to test and characterize the performance of each component during its development, manufacturing, and when it is in use in the network. The various levels of component testing, such as performance assessment and conformance testing, need to be accurate, cheap, and fast.
Quality results depend on quality test equipment
Whatever component measured or method used, the quality of the test result is directly dependent on the quality of the test equipment. An essential part of a typical test setup is the optical source.
In most cases, a wavelength-selective source (tunable laser) is preferable to a broadband source and a wavelength-selective detector such as optical spectrum analyzer (OSA). A tunable laser is usually cheaper, faster, and delivers higher resolution. In some cases there is no alternative to using a tunable laser, and this will be even more prevalent in the future. The ITU grid in DWDM systems is getting denser migrating from today's channel separations of 100 GHz and 50 GHz to 25 GHz and even 12.5 GHz. Better wavelength-selective components will be required and they will need to be tested with a much higher accuracy.
Higher transmission rates per channel is another evolution from 2.5 Gbits/sec to 10 and 40 Gbits/sec; rates at which the chromatic and polarization-mode dispersion of components and fibers must be tested in detail. Measuring these parameters throughout the whole wavelength band requires a tunable source. This is especially true with new, more effective measurement techniques such as interferometry.
A sun without spots
The ideal source should deliver light with a very defined "sharp" wavelength, with no noise or side modes. It should be able to sweep smoothly and quickly over a wide range without any sudden jumps in wavelength (mode hops) and with perfect repeatability. An external cavity laser can come very close to meeting these requirements but it still has restraints that limit the quality and reliability of the measurements.
There is a quite straightforward translation between parameters to be measured and the specifications of a tunable laser source; the problem is that such specifications are not always stated by the tunable laser manufacturers. Examples are ripple in tuning speed, tuning linearity, and polarization variations. Different vendors often define parameters in different ways, which makes it difficult for users to compare tunable lasers. A good example of this is spontaneous source emission (SSE) and total SSE from the laser diode (see Figure).
Insertion loss is an important parameter for every optical component, especially filters. When measuring insertion loss and its dependence on wavelength, it is important that the power measured is the "true" power at the specific wavelength. This measurement can be achieved by using an OSA, or a wavemeter, in combination with an optical power meter. Using an OSA is sufficient in most cases, but this method is expensive and slow.
Measuring "true" power with a wavemeter and a power meter is cheaper, but this method could still be too slow, and the tunable laser source requires low SSE. If the wavemeter is excluded, the measuring time is greatly reduced. However, this approach will require that the source has a high wavelength accuracy as well as low SSE. When ITU channel spacing is decreased to 25 GHz in the future, the tunable laser will need an absolute wavelength accuracy of 1 - 3 pm and a relative accuracy of parts of a pm.
The following list summarizes the tunable laser requirements when these devices are used as sources in testing passive components such as optical filters and add/drop multiplexers:
•Low SSE, At least -65 dB/nm (allows measuring a filter with 65-dB stop band)
•Wavelength accuracy, ±10 pm and soon ±2 pm
•High tuning speed, >100 nm/sec
•A linear wavelength sweep
Fiber and transmission systems in operation
An important parameter for fibers and transmission systems--one that will be even more important in the future at higher transmission speeds--is dispersion: chromatic dispersion (CD) and polarization-mode dispersion (PMD).
There are several methods of measuring CD and PMD such as optical interferometry and modulation phase-shift. Both require a tunable laser source. A big advantage to using an optical interferometer to test and characterize optical components is that it supplies many different optical parameters at the same time using only one measurement. It enables rapid measurements, that is, if the tunable laser source used is fast. When measuring PMD on fibers exposed to the "real world," a short measurement time can even be a necessity. Furthermore, the tunable laser source must fulfill the following optical specifications: smooth and truly continuous tuning speed, singlemode operation, slow or no polarization fluctuations and wavelength accuracy of ±10 pm, and soon ±2 pm.
Of particular importance is the smooth tuning, since the interferometry method will not work if the tuning speed fluctuates. This specific behavior is difficult to find on tunable laser data sheets.
Important parameters for amplifiers are gain and noise. When measuring these for an optical amplifier it is vital that the source does not add noise to the measurement and cause an overestimation of the amplifier noise figure. The SSE level for a distributed feedback laser can be approximately -50 dBm/nm, but for accurate measurements a value below -65 dBm/nm is advisable¹.Usually several tunable lasers set to different wavelengths (ITU-channels) are used to test an optical amplifier. Another method is to use a broadband laser together with only a few tunable lasers.
Important features for the tunable laser sources when testing amplifiers are low SSE (-65 dB/nm), wavelength stability of better than ±10pm, and a stable polarization state. When using multiple sources, cost and size are additional factors to be considered.
Lasers and transceivers
Lasers and transceivers are modulated in order to send data, but this modulation broadens the optical spectra of the source (chirp). The spectral broadening of the source in combination with wavelength dispersion in optical fibers will cause the shape of the data pulse to erode.
The chirp can be measured by using optical heterodyne. The quality of the heterodyne method is set by the performance of the internal oscillator, i.e. the tunable laser. It is the optical line width of the tunable laser, together with its absolute and relative wavelength accuracy (frequency jitter), which determines the resolution. When scanning the laser (DUT), the tunable laser source must change the wavelength continuously. To be able to fully examine the laser's dynamics, the tunable laser source requires a low SSE.
A source of profit
The quality of the light is just one aspect. Equally important is the cost-effectiveness. High yield of a production line is necessary to ensure a low cost per piece, and it is crucial to identify defective components before they are built into subsystems or networks. Not only do the components and systems need to be more thoroughly tested, but the testing must be faster and cheaper. Thus, when selecting the right tunable laser source, consider the following features:
•Wavelength range: The wider the range, the lower the risk of needing more than one source, meaning that a tunable laser should cover at least the C- and L- bands.
•Tuning speed: The faster the sweep, the more units can be tested under a certain period. Today there are sources with speeds over 100 nm/sec available.
•Power: High output means the possibility of sharing the light among many simultaneous devices under test.
•Price: Today the laser is often built into a specialized test instrument or delivered as a module in a general test platform. It is too expensive and impractical to use separate tunable-laser-source bench tops. Therefore, the tunable laser itself must feature a modular design so that it can be easily integrated into the test equipment.
A tunable laser source must meet all of these requirements to be truly versatile and accommodate future needs. To be used for testing all different types of components, the tunable laser source must meet the following specifications:
•Low SSE: ≥65 dB/nm
•Wavelength accuracy: < ±10pm
•Wavelength stability: < ±10pm
•High tuning speed: <100 nm/s
•Narrow effective line width
•Smooth and truly continuous tuning speed
•Single mode operation
•Slow or no polarization fluctuations during tuning
•Small physical size
•High output power
Test and measurement equipment must stay one step ahead of the components being developed, leading to higher demands on the source. Increasing numbers of parameters must also be measured on installed optical network systems. Parts of the test and measurement systems, which are currently used in labs and production facilities will move to the field or into monitoring systems. This means that the equipment will have to be cheaper, smaller, more robust and self-contained. Field applications put extra demands on the tunable lasers in the instruments such as size, low power consumption, reliability and ruggedness, while still requiring speed and accuracy. Let's not forget that the field engineers will need a good cup of coffee, too.
Michael Larsson is a product manager and Trajan Badju is vice president of sales and marketing at Radians Innova AB (Gothenburg, Sweden ).