Sheila, Kiwamu, Rana, Lisa This plot corresponds to one of tonight PRMI carrier locks (Sep 11 around 3.50 UTC). The recycling power fluctuations are larger than 50%. PR3 PIT is causing the motion at 700 mHz, ITMX YAW is largely dominating around 100 mHz .
From the video of the BS camera that I posted last night (see alog 13875) and the size of the BS elliptical baffle (T1300408), I roughly estimated the vertical spot motion on the BS to be 22 mm in pk-pk (at about 0.7 Hz). This is incredibly big.
According to an ABCD matrix model for the PRMI, the most effective optic for the BS spot motion is, of course, PR3. The PR3-to-BS coefficient, which is obtained from the matrix analysis, is about 1.7 x 105 [m/rad]. In order for PR3 to cause a 22 mm motion on the BS, PR3 should move by 22 mm / 1.7 x 105 = 0.13 urad. Looking at the rms of the PR3 oplev pit, indeed a peak at 0.7 Hz was dominating the rms which pushes the rms to something like 0.05 urad which is consistent with the estimated PR3 motion from the BS spot motion. So it was clearly from PR3.