Summary:

By looking at the XARM HWS signals, I can see that we're overestimating the response of the CO2 central heating. As a result, we are underapplying the preheating to stablize the total value.

Also: the H1:TCS-SIM_XARM_POWER estimate is now wrong following some recalibration work on the X_TR SUM signals the other day.

Details:

I looked at 24 hours of HWS and SIM data from two days ago. Aside from glitches, there were systematic discrepancies between the measured HWS signal (when scaled by 0.5 to convert it from double pass to single pass) and the sum of the CO2 and SELF heating SIM channels.

• SELF: I spoke to Kiwamu and he suggested that some whitening filter changes may be responsible for the change in calibration of X_TR_B_SUM (here?).
• CO2: it's clear by comparing the lenses measured by the HWS and predicted by the CO2 SIM in the period between 650 and 750 minutes that the CO2 lens is overestimated
  • A best fit between the SIM and HWS suggested that this needs to be dropped by a factor 0.573
  • So: 6.23E-5 diopters per Watt * 0.573 = 3.57E-5 diopters per Watt in H1:TCS-SIM_ITMX_SUB_DEFOCUS_CO2_GAIN

This means that the TCS Guardian function for calculating CO2 power will need to be adjusted. 

The attached plots show (a) the time series from the HWS, the CO2-SIM and SELF-SIM, (b) the HWS measurement minus the optimally scaled CO2-SIM measurement and (c) the second plot minus SELF-SIM calibrated to give the smallest RMS on the whole time series.