J. Kissel

The other day, we had an inadvertent temperature impulse test at the end stations (see LHO aLOG 55325), and while trending our measure of the drift of the low-frequency sensing function "f_s^2," i.e. H1:GDS-CALIB_F_S_SQUARED for other reasons, I thought -- "Huh -- I wonder if the wiggle here in f_s^2 was related to the wiggles there in VEA temperatures?"

As such, I did a crude correlation analysis lining up ~1 weeks worth of trends between H1:GDS-CALIB_F_S_SQUARED and (for each station)
    - The room temperature, as measured by facilities channels (H0:FMC-EX_VEA_AVTEMP_DEGC, H0:FMC-EY_VEA_AVTEMP_DEGC, H0:FMC-CS_LVEA_ZONE1B_DEGC)
    - The core optic vertical position (our most expensive thermometers; H1:SUS-ETMX_M0_DAMP_V_INMON, H1:SUS-ETMY_M0_DAMP_V_INMON, H1:SUS-BS_M1_DAMP_V_INMON, H1:SUS-SR3_M1_DAMP_V_INMON)
    - The core optic pitch position (which *should* correspond to a temperature change due to any vertical-to-pitch coupling; H1:SUS-ETMX_M0_DAMP_P_INMON, H1:SUS-ETMY_M0_DAMP_P_INMON, H1:SUS-BS_M1_DAMP_P_INMON, H1:SUS-SR3_M1_DAMP_P_INMON)

Attached are screenshots of the results in the following order: ETMX, ETMY, SR3, and BS.

While the ETM vertical and pitch positions are *definitely* correlated to the temperature excursion (with the usual delay attributed to the "low-pass filter" that the vacuum system provides), I'm not convinced that either are correlated to the ~few Hz^2 drift seen in f_s^2. 
In fact, my eyes are most drawn to what coincidences I see in the drift of the Beam Splitter's pitch position and f_s^2. 

Would @DetChar be willing to sign someone up to run a LASSO-like analysis using f_s^2 as their witness, and using core-optic vertical and/or alignment positions as potentially correlated motivators for change? 
That would be awesome. I'm also open to an analysis on other likely candidates (a more careful study of the above temperature channels, expanding the list of optics to include IM4, PRM, PR2, PR3, ITMX, ITMY, SR2, SRM, OM1, and OM2, TCS Laser Power, SR3 disc heater power, and any other channels you think might be good metrics to track things that might be causing mode matching drifts between the arm cavities and the signal recycling cavity.)