Sheila, Jenne

Background:

We took a little bit more commissioning time to try to address the slow yaw oscillations we have been having since changing the SR3 heater.  (49160 49077 48948 48857) These seem somewhat similar to some oscillations seen in January: 46170  We know that the phasing of AS36 (used for BS control only) can reduce the oscillations, but it seems like this isn't a long term solution since phase changes of only ~10 degrees seem important.  Yesterday we learned that the oscillations go away when the INP1 Y loop was off.  We are currently using an ASC input matrix for the REFL WFS that is rather different for pit than yaw (1st sreenshot): for pitch we aren't using any REFL9 sensors, while yaw does use REFL9 mixed with REFL45 for CHARD and INP1.  

What we did:

We prepared to measure the sensing matrix using the script in userapps/h1/asc/scripts/sensingMatrix/run_sensmat.py  We updated the locations of the 8Hz notches and Jenne ran the script for one DOF, and was able to run the analysis script to get a result.  In order to actually do the full sensing matrix measurement we need to check the amplitudes of all the exctiations.

We noticed that the oscillation was showing up strongly in the REFL Q ASC signals, second attached screenshot, for both 45 and 9, with almost equal signal in I+Q for RF9.  We decided to check the phasing before measuring the sensing matrix.  We notched the ASC loops at 8Hz, and made a frequency line at 8.125Hz using 30 counts on LSC-EXTRA_AO_2, which is connected to the CM board excitation.  For REFL45, the excitation was mostly showing up in the I signal, so we didn't adjust it's phase.  For REFL9, we rotated the phases by about 30 degrees for all quadrants (3rd screenshot shows diffs).  In the attached tiem series, we phased REFL9B (only in loop for INP1) at about -1100 seconds, and we phased REFL9A (INP1+CHARD) at about -800 seconds.  The oscillation stopped after the phasing, and has not reappeared in the hour since.  

We went back to observing without finishing the sensing matrix measurement since LLO was back to observing and we want to preserve triple coincidence with Virgo.

Observation about frequency excitation:

While we were injecting the 8Hz frequency line (starting around 17:30 UTC May 10th) we noticed a large comb in DARM, which could be interesting to investigate further.  Pep also has some interesting comments about lines that changed after we phased REFL.