Yesterday I checked again the polarisation of BAB and LO beams entering the homodybne detector. I obtained different results than in entry 53129.

I placed a power meter on the beam reflected by HD_MIR, the BAB reflected from the HD BS and/or the LO transmitted through the BS. To measure the fraction of p-polarisation, I inserted a PBS in front of the power meter.

With the LO at maximum power (about 2 mW on HD, i.e. 1 mW on power meter) I measured a residual p-polarisation of about 0.05% (i.e. 500 nW over 1 mW of s-polarisation).

With the BAB at maximum power (about 2 mW on HD, i.e. 1 mW on power meter) I measured a residual p-polarisation of about 1.3% (i.e. 13 uW over 1 mW of s-polarisation).

Then I placed a waveplate at different locations along the BAB path, and I rotated the waveplate to minimise the PBS transmission. The PBS and power meter were at the same location as before, i.e. after HD_MIR.

By placing the waveplate before the delay line I could reduce the residual p-polarisation component to 0.2%.

Placing it after the first or the second mirror of the delay line I could reduce the residual p-polarisation component to 0.3%. I could reduce the residual p-polarisation component to 0.2%.

By placing the waveplate after the third mirror of the delay line I could only reduce the residual p-polarisation component to 1%.

I observed that the beam is not pefectly in plane along the delay line, i.e. the height w.r.t. the optical table changes by about 2 mm p-p along the path.

I left the waveplate between the first and second mirror of the delay line, tuned to minimize residual p-pol at HD.