Yesterday we were sitting waiting for the violin modes to damp before changing the OMC whitening, so I took advantage of the time to do some scans of the RF3 demod phase (which is labeled RF6 in beckhoff) with different SRCL offsets since we suspected that the SQZ angle servo might not work well during thermalization because of the changin SRC detuning. However, it seems that doesn't have much impact on the SQZ angle readback, and this servo should work fine with changes in SRC detuning.

The first subplot shows the ADF ellipse, and then the channels after the normalization that is used to turn it into a circle, these look the same for the two offsets.  The next plot shows the readback of the SQZ angle, which is arctan2 of the normalized signals from the first plot.  This relation ship is not linear, it should be described by the relationship Daniel wrote here: 49026

The bottom plot shows the sqz blrms, we want to operate the sevo where this is minimzed, where it seems that the sqz angle readback is a good error signal to use.