M. Ball, J. Driggers, S. Dwyer, J. Kissel We've gathered a standard set of sensing function measurements this week in the nominal IFO configuration. What may be of note -- after yesterday's computer process restart of the h1asc model (LHO aLOG 50567), and during recovery, we re-acquired the IFO *without* the SRC1 Yaw offset we've been running with since May (see LHO aLOG 49393). You'll notice in the plots below, the sensing function has changed shape to something more like a physical spring unlike what we've seen in the past. We lost lock before we were able to explore this further, but we will likely be running with this offset OFF from now on, so it'll be something interesting to keep an eye on (i.e. to see if next week's sensing function is consistent with this week's, and we'll also try actively changing this offset to see if it is yet another knob beyond spot position and ASC gain). Attached is a collection of 4 plots: Page (1): The sensing function (i.e. ratio of PCAL2DARM and DARM Loop Suppression transfer functions), compared against a new MCMC fit to establish this measurement's parameters. Note that although the cavity pole frequency is reported low, this is not the only time it's been low like this (see later, Page (4) comments). Page (2): The same sensing function measurement compared against the *reference* model Page (3): The individual measurements going in to the sensing function. Page (4): a time series trend of the MCMC fit values for the parameters of sensing function measurements throughout O3 up to today. We also began to phase and balance the quadrants of the AS_A_45 WFS (i.e. the error signal for DHARD), using the PCAL calibration line at 17.1 Hz as our "pure longitudinal / DARM" drive, minimizing that line in the I phase. Unfortunately, we lost lock just after measuring the DARM loop suppression and did not get the PCAL2DARM transfer function, so we can't show a plot of how the sensing function changed. However, from our previous work (see LHO aLOG 50498, G1901353), we have found that only the DARM Loop Suppression changes under these ASC-like changes -- and we saw no change. Raw data templates: /ligo/svncommon/CalSVN/aligocalibration/trunk/Runs/O3/H1/Measurements/FullIFOSensingTFs 2019-07-17_H1_DARM_OLGTF_LF_SS_5to1100Hz_15min.xml :: standard measurement of 1/(1+G) 2019-07-17_H1_PCALX2DARMTF_LF_SS_5t1100Hz_10min.xml :: new-ish standard measure of C/(1+G) with PCAL X 2019-07-17_H1_PCALY2DARMTF_LF_SS_5t1100Hz_10min.xml :: standard measurement of C/(1+G) 2019-07-17_H1_PCALX2DARMTF_BB.xml :: new-ish standard measure of broad-band C/(1+G) with PCAL X for testing GDS-CALIB Strain, Start Time: 2019-07-17 20:42:25 UTC 2019-07-17_H1_PCALY2DARMTF_BB.xml :: standard measure of broad-band C/(1+G) with PCAL Y for testing GDS-CALIB Strain, Start Time: 2019-07-17 20:40:19 UTC 2019-07-17_H1_OMCDCPDSUM_to_DARMIN1.xml :: standard measure of the conversion between DARM IN1 error signal in [ct] and (roughly calibrated) DCPD current in [mA] 2019-07-17_H1_ASATuneUp_DARM_OLGTF_LF_SS_5to1100Hz_15min.xml :: standard measure of 1/(1+G), but after modifications of AS_A_45 WFS quadrant phases and gains. processing scripts: /ligo/svncommon/CalSVN/aligocalibration/trunk/Runs/O3/H1/Scripts/FullIFOSensingTFs/ process_sensingmeas_20190717.py :: first three plots plotMCMC_vs_GDSTDCFs.py :: last plot using rev 8028 of function library, /ligo/svncommon/CalSVN/aligocalibration/trunk/Common/pyDARM/src/ sensing.py
The first attachment here is a screenshot of the changes to AS A 45 phasing and Q gain balancing.
The second screenshot shows how we phased the AS45, the right panels show before transfer functions at the pcal Y line frequency (17.1Hz) before phasing with dashed lines and after with solid lines. The spectrum of the individual quadrants after phasing shows that the calibration lines are now in the Q phase.
A more quantitative statement about how much we reduce the length to angle signals will be added soon....
I've finally gotten around to looking at the magnitude of the angular response to the calibration length actuation before and after this AS A rephasing.
In the attached screenshot, the top panel is the AS A RF45 pitch signals versus the Pcal, and the bottom panel is the AS A RF45 yaw signals. Dashed traces are before the rephasing, solid are after. Red traces are the I signals (not used for the IFO), and blue traces are the Q phase which are used for DHARD.
Looking primarily at the blue traces, we see that for a constant length drive (we're assuming that the pcal actuation is (a) constant, and (b) pure length) the Q-phase angular response is lower after we rephased the WFS. However, the ratio isn't very large - we see a 5dB improvement in pitch and only a 3dB improvement in yaw.
