WDM Components: All-Band Spectral Characterization
A whole family of subsystems and components is needed to enable and facilitate the development, manufacturing, and deployment of systems using dense WDM technology. Specifications and measurement techniques have been developed to characterize each of these elements. This section describes the integrated test solutions offered by EXFO and explores key measurement and testing techniques used to characterize the major passive network elements of a dense WDM system.
As dense WDM technology moves to closer and closer wavelength spacing, the requirements and performance specifications for wavelength-selective components become increasingly demanding, and test procedures become equally complex.
A device under test (DUT) must be injected with an optical signal that has known characteristics, and its output must be analyzed in detail to determine how it differs from the input. Both the measurement source and the analyzing instrument should be carefully selected to test and analyze the intrinsic value of the parameter being measured without introducing extraneous effects.
Typical Component Testing Combinations: The Multiplexer
Multiplexers (MUX) and demultiplexers (DEMUX) are important components of DWDM systems, and their bandwidth, crosstalk, insertion loss, return loss, isolation, and polarization properties are critical to network performance. The desirable spectral attributes of several other DWDM components are very similar to those needed in multiplexers and demultiplexers (or, more particularly, in individual MUX/DEMUX channels), which is why test procedures for these devices are presented in this section.
Insertion Loss
Figure 1 shows a typical setup to determine the insertion loss of a DWDM MUX/DEMUX. A flat, broadband ASE source is used, whose output covers the DWDM band. The OSA output will, therefore, directly indicate the shape of the insertion loss. A suitable reference measurement, made by connecting the source directly to the OSA, can be used to convert this relative loss curve to an absolute one. In this case, the spectral resolution of the measurement is limited by the OSA RBW and its shape. The 1xN switch is included to ease or even automate the testing of multichannel devices. All measurement components must have low polarization state sensitivity.
Figure 1: The ASE Source shoots through the multiplexer, while the OSA gets the signal for each output port
Mueller-Stokes PDL Measurement Method
It is possible to represent an optical wave as mathematical vectors called Stokes vectors (S).The elements constituting the vectors are power measurements easily obtained with a source, polarizer and detector. The input signal (input vector) will be changed by the DUT and will give a new Stokes vector. The complete DUT effect on the polarization of light will then be represented in the Mueller matrix. This 4x4 matrix represents the transmission characteristic of the DUT.
Figure 2: Three different waveplates are used to create different polarization states.
Four different input polarization states are needed to determine the four parameters of the Mueller matrix. The polarization controller (Figure 2) generates those four polarization states. Once a reference measurement between the polarizer and the detector is taken, the DUT is inserted after the polarizer. The power is measured by the detector with respect to every polarization state, and then the Mueller element is calculated with the transmission coefficient (T).
This method is fast and only requires low-cost equipment. PDL could also be measured as a function of wavelength by using a tunable laser in sweep mode. However, the Mueller-Stokes method is sensitive to patchcord manipulation. This method is ideal for automated test systems. A related parameter, polarization-dependent bandwidth (PDBW), can be similarly measured. This spectral characteristic of an optical filtering device is the difference between the minimum and maximum bandwidth when the channel is measured under all states of polarization:
Measurement accuracy increases with the number of polarization states examined, but there is obviously a practical limit to this number; several minutes of continuous measurement normally suffice.
Crosstalk
Crosstalk between DWDM channels is measured by replacing the ASE source used in the previous setups with a narrow-linewidth, tunable laser source. Repeated OSA scans are performed on all of the multiplexer or demultiplexer channels while stepping the tunable source through the wavelength region of interest at the desired resolution. The measurement resolution is determined by the tunable-laser steps, not by the OSA, but loss curves with a resolution as high as 0.001 nm can be obtained with excellent dynamic range. The procedure is time-consuming, however, especially for multichannel DUTs, unless a multichannel power meter is also available. A wavelength meter should be used for calibration.
Figure 3. Crosstalk can be measured using either a tunable laser with an OSA, or a tunable laser with power meters.
Directivity
Directivity, which is sometimes referred to as near-end crosstalk, is a measure of the isolation between the inputs ports of a multi-input device. It is particularly important in a multiplexer, where power returned to the system transmitters must be kept to a minimum.
The users must provide a non-reflective termination on the output side of the device when measuring directivity. If not, reflections at the fiber end-face may degrade the measurement. Often, it is sufficient to perform directivity measurements only at nominal channel wavelengths. For example, a signal could be inserted into Channel 1 at its design wavelength and the power that is returned to the other inputs could then be measured. The difference between the power inserted and that returned to other inputs is the directivity for that particular channel combination. Measurements should be taken for all channel combinations. This makes it difficult to automate directivity measurements, especially for high channel count devices.
Figure 4: Illustration of undesired signal leakage from input 1 to other inputs
Optical Return Loss
Optical return loss (ORL) is measured using a combination of a source, coupler and photodetector, often referred to as an optical continuous-wave reflectometer (OCWR). After a calibration step in which a component with a known reflectance is substituted for the DUT, the DUT is inserted. The detector then measures the power it reflects, a correction factor based on the calibrating step is applied, and the ORL is displayed. With a high-power, non-coherent optical source and a high-sensitivity, high-resolution detection system, return losses of 70 dB or lower can be detected and measured with this configuration.
ORL can be wavelength-dependent. If this dependence must be characterized, either a high-power, moderate-coherence tunable laser or a wide-band source (an ASE laser) can be used as the source in the above configuration, with an OSA used as a detector (Figure 5). Because of the limited dynamic range of an OSA, however, it will be difficult to track ORLs lower than 40 dB.
Figure 5: Wavelength-dependent ORL measurement
Bandwidth
Optical bandwidth is measured using a procedure very similar to that used to measure insertion loss. The spectral characteristics of the measurement system can be removed with a reference measurement. A typical setup is illustrated in Figure 6. If the band edges must be characterized preciselyto a spectral resolution less than 0.1 nm, for instancean alternative configuration should be used, with a tunable source and a power meter or OSA. Here, the spectral resolution is limited by the tuning resolution of the source (probably about 0.01 nm), and the dynamic range by the sideband rejection of the tunable source. The sensitivity is determined by that of the power meter (possibly 100 dBm).
Many of the measurements already discussed share test equipment in similar configurations, so they can readily be combined. This approach is particularly interesting when test methods are being automated. The configuration shown in Figure 6, for example, uses a low-noise tunable laser source (TLS) that scans across the wavelengths of interest, while at the same time multichannel power meters acquire transmission data for each channel of the DUT. With this type of instrument setup, the measurement time is independent of the number of device channels. This particularly suits automated testing of multichannel components. The upper portion of the figure indicates how incorporating an optical switch in the measurement setup can allow other devices to be prepared for testing while a particular test is underway, using much of the same equipment. The resolution of the measurement is determined by the TLS, and the measurements dynamic range is determined by the spontaneous emission (or noise) of the source, as well as the dynamic range and sensitivity of the power meter. By using a TLS, such as a fiber laser, a dynamic range of more than 65 dB is easily attained. The diagram also shows how ORL and PDL measurements are integrated into the process. With this arrangement, it is easy to add extra power meters to deal with multichannel devices.
Figure 6: Automated test setup that measures PDL insertion loss and ORL