The Agilent 83438A #009/71452B/ 11896A setup as shown in Figure 13 characterizes components for dense wave length division multiplexing (DWDM) applications fast and accurately versus wavelength. Under remote control, it can measure insertion loss, crosstalk and PDL automatically. Other characteristics can be calculated too, such as the polarization dependence of the center wavelength or bandwidth of a filter.
The Erbium ASE Source Agilent 83438A with option 009 provides polarized light in the range 1525 to 1565 nm. Its polarization state is automatically randomized using the Agilent 11896A Polarization Controller. If the Agilent 11896A runs fast and the Agilent 71452B Optical Spectrum Analyzer averages 100 or more sweeps, then the resulting trace represents the loss versus wavelength of unpolarized light. However, if the Agilent 11896A runs slowly and the Agilent 71452B Optical Spectrum Analyzer sweeps quickly, then trace A of the OSA can capture the maxima and trace B the minima of each sweep. The difference between trace A and trace B shows the polarization dependent loss versus wavelength. For low PDL values (<1 dB), the aerage of the two traces is approximately the insertion loss for unpolarized light.
This procedure enables PDL measurements to be performed over a specified wavelength span which is equal to or less than the width of the source used in the test system. The best setting for the Agilent 11896A depends on the OSA. The following explains the procedure and the relationship between the Agilent 11896A and the OSA for the best measurement results.
Agilent 11896A settings:
In this setup, the Agilent 11896A AUTO SCAN mode is used. However, what scan rate is the best? To minimize the overall measurement time, the scan rate should not be too slow. To measure the PDL at one state of polarization, that state should stay long enough for the OSA to complete a sweep; therefore the OSA sweep speed should be appropriately fast.
Table 8 lists the relationship between displayed scan rate and the time Tc required for one great circle on the Poincaré sphere. The Agilent 11896A randomizes the SOP. Therefore, the time Tc is only a statistical measure. The actual time may be longer or shorter. The SOP rarely moves exactly along a great circle. Rather it rather moves in a figure “8” as well as other figures along the whole surface of the Poincaré sphere.
Agilent recommends that Tc be about ten to forty times larger than the sweep time of the OSA. In this case the state of polarization changes little during a single OSA sweep. The total measurement time should exceed twenty to fifty times Tc, so that the Agilent 11896A exposes the DUT to all SOPs during hundreds of OSA sweeps.
In the example below, the OSA sweeps in 110 ms. From Table 8 we selected the scan rate 3 (Tc =3 sec is greater than 20 * 110 ms) and let the measurement run for two minutes (40 * Tc).
Agilent 71452B settings:
Set wavelength and amplitude settings to match the device under test. In the example shown here, START is 1542 nm, STOP is 1562 nm, and resolution bandwidth (RES BW) is 0.2 nm. Our DUT has approximately 7 dB insertion loss.
1. Connect the DUT as shown in Figure 13. Activate the light source and let the Agilent 11896A scan.
2. AUTO ALIGN the OSA, then set its wavelength range and resolution as desired.
3. Select an appropriate vertical scale: press PEAK SEARCH, then TO REF LEVEL, then Ampt, LOG dB/DIV and enter 1 dB.
4. Set SENSitivity to –60 dBm. If necessary, increase the sensitivity. Watch the sweep time (ST).
5. Set the SCAN RATE of the Agilent 11896A Polarization Controller as discussed above (e.g., scan rate 3).
6. Press Traces, CLEAR WRT A. Wait until at least one sweep has been completed, then press MAX HOLD A.
7. Activate trace B and CLEAR WRT B. Wait until at least one sweep has been completed, then press MIN HOLD B. Then activate trace C.
8. Adjust the vertical scale (e.g.,Ampt, LOG dB/DIV, 0.2 dB)
9. Wait a sufficient time (2 minutes in this example), then stop the sweep (BW, Swp, SINGLE SWEEP).
Calculate the PDL: press Traces, MORE, trace logmath, A <- A-B. Trace A now contains the peakto- peak PDL, but it may be out of the display. Press Ampt, REV LVL 0 dB, then MORE, REF LVL POSN 0 <enter>. This sequence causes the OSA to display 0 dB PDL at the bottom of the screen. Use Ampt, LOG dB/DIV to adjust the scale.
The uncertainty of the measurement result depends on the DUT. Many components and materials have less than 2 dB PDL. The following analysis considers two examples. One has 2 dB PDL and the other 0.2 dB (see Table 9). The measurement uncertainty for such devices is ±0.12 and ±0.06 dB, respectively.
The degree of polarization of the Agilent 83438A with option 009 is 95% to 99%. For devices with high PDL (>2 dB), the amount of un-polarized light contributes most to measurement errors. In the range 2 to 10 dB PDL, the error is up to 10% of the PDL measured. Above 10 dB it increases even more; therefore Agilent does not recommend this setup any more.
It is possible to further reduce the measurement uncertainty for devices with very low PDL. However, it requires a second Agilent 11896A and sophisticated software running on a remote computer1 to get up to ±0.025 dB accuracy. The necessary setup and the procedures are very complex and therefore not discussed in this document.