by rotating the polarization before a pbs or a tilted beam sampler we should be able to tune the reflected (or transmitted) power.

However the typical change we saw (0.3%) is far lower than expected.