Two weeks ago Sheila locked PRMI and DRMI_1F (alog 42200).  We would like to understand how the PRMI sideband gains have changed since O2, using the POPAIR_B 2f RFPDs.
I looked at thirteen times from September 1 through September 12, 2017 for when ISC_LOCK was in state PRMI_LOCKED, and compared the gain then to Sheila's May 27, 2018 lock. 

PRMI 1f Sideband Relative Gains (Now/PostO2)
 
      

POPAIR_B_LF was also ~1.25 larger for now as opposed to PostO2, but POP_A_NSUM and POP_B_NSUM are about half of what they were, and POP_X_DC_NSUM is about a fifth.  The above results could be a bad POPAIR_B alignment during PostO2 locks manifesting itself as increased sideband gain now.

Calculation Details
POPAIR 2f signals measure the 1f sideband power beating with one another (in the case where low and high sidebands are equally resonant in the cavity, the carrier and 2f sideband beat cancels):

where  is the PRC gain for the 1f sideband, and  is the 1f sideband modulation depth.  

We calculate the relative sideband gain , because there is no light on POPAIR_B when PRMI is not resonant.  

The input power  is normalized using H1:IMC_IM4_TRANS_SUM.  
The modulation depths are changed from PostO2 to now:  Koji's alog 41435 and Daniel's alog 41889 report the alterations:

Modulation Depth [rad]         O2          New
----------------------------------------------
9.1 MHz                     0.187        0.210         
45.5 MHz                    0.287        0.259        


All filter modules, digital gains, and whitening gains for POPAIR_B 2f signals were compared and found to be the same.  The I and Q error signals were summed in quadrature to avoid phase differences.  Dark offsets were calculated from the nearest DOWN time.