Ryan S, Ibrahim, Oli, Sheila, Camilla, Mayank

Problem: The ADF_OMC_TRANS_SQZ_ANG (ADF readback of Sqz angle) was not linear with the SQZ angle (~ SQZ-CLF_REFL_RF6_PHASE_PHASEDEG).

What we tried so far?

1. We played with different combinations of ADF_VCXO phase and ADF demod phases to see if we can get linearity between these two channels however it did not work. (See Figure 1)
2. We plotted parametric ADF_I and ADF_Q for various Demod phases. Ideally this curve should be an ellipse centered around (0,0). However the ellipse we got was shifted from (0,0). (Figure 2 part A). { The ellipticity of the ellipse depends on the amount of squeezing and the location on the ellipse determines the squeezing angle}

What we did ?

1. We calculated the required offsets (ADF_OFFSET_I and ADF_OFFSET_Q) so that the ellipse can be centered around zero.
2. We calculated the G (4.5) value from the ellipse and used this normalization constant of ADF_Q to turn the ellipse into a circle.
3. Calculated the new ADF readback of sqz angle values after numerically adjusting the offsets and normalization and verified that it indeed is linear over a wide range of squeezing angle. (It is matches well with squeezing angle, see Figure 3 )   
4. Added all these changes to the model and scanned the RF6 demod phase to verify that the ADF readback values now closely follow the squeezing angle. (see Figure 4) and Figure 5 shows the ADF_I and ADF_Q demod plot (Ideally it should be a circle )
5. We modified the SQZ_ANG_ADJUST.py  (Fig 6):  We found that the ADF_OMC_TRANS_SQZ_ANG = -95 for the optimum Squeezing and added it as the setpoint. The servo locked with these changes .

We may further tweak the parameters to improve the ADF performance.