Comparison of two configurations of a Stokes polarimeter measuring three wavelengths simultaneously

In this paper, we study two configurations of a full-Stokes polarimeter that measures three wavelengths simultaneously. A recently developed fitting calibration process is used to reduce the error in the final Stokes vector. The first configuration uses an optimized setup for the central wavelength that fixes the non-optimized setups for the other two wavelengths. The second configuration is the result of a search for the best simultaneous setups for all three wavelengths. It is found that this second setup gives smaller errors in the measured Stokes parameters before calibration, but the first configuration gives smaller errors after calibration. These results demonstrate the need for a method of finding the best polarimeter configuration for this type of measurement. It seems clear that the condition number by itself is not a metric that provides sufficient information to determine whether a particular configuration is better by giving smaller errors. It is also important to point out that the results presented here are for the particular calibration scheme used. Additionally, it may be that a different calibration method is required to give better results, but that is outside the scope of the present work.