Simon

Yesterday, I slightly changed the setup and put the QWP after the HWP so to be able to make the input beam circular polarized. The results can be seen in the pictures attached.

Obviously, the map is much more homogeneous than with linear polarized light, although there is a shift towards being more elliptical as the mean angle is ~52 degrees and not 45 degrees as expected with pure circular polarized light.
Interestingly, the prominent region in the upper left part visible in the other maps, vanished completely.