J. Kissel
Using the canned function that produces the official sensitivity spectra for the public,
/ligo/svncommon/CalSVN/aligocalibration/trunk/Runs/O2/Common/Scripts/DARMASDs/produceofficialstrainasds_O2_C01.m
I grabbed the amplitude spectral density of DCS / GDS-CALIB_STRAIN and PCALY RX PD, with
- C01 data from Dec 13 2016 @ 10:00 UTC for H1 (because CAL line heights have not changed all run)
- C00 data from Jan 30 2017 @ 10:00 UTC for L1 (because C01 isn't ready after Jan 24th, but L1 last changed line heights on Jan 27th)
and recast the calibration line heights into displacement amplitude (in [m_pk]) of all calibration lines as the stand currently in O2 for future reference.
The tables below lists the function of these lines, which actuator is used to drive the lines, the frequency, their amplitude in [cts] at the the drive point, and the amplitude it displaces the test mass at its given frequency.
H1:
Used To Characterize C (f_s,Q_s) A (K_TST) A (REF) A (K_PU) C (K_C, f_cc) C (f_cc, t_C)
Actuator PCAL SUS PCAL SUS PCAL PCAL
Frequency [Hz] 7.93 35.9 36.7 37.3 331.9 1083.7
Amplitude [drive ct] 5000 0.55 500 0.44 5000 15000
Amplitude [m_pk] 4.78e-15 2.71e-17 2.21e-17 1.36e-19 2.71e-18 7.71e-19
L1: (edit-- there are several errors in the table below. Will fix by the end of the day)
Used To Characterize A (K_TST) A (REF) A (K_PU) C (K_C, f_cc) C (f_cc, t_C)
Actuator SUS PCAL SUS PCAL PCAL
Frequency [Hz] 23.3 22.7 331.3 23.9 1083.1
Amplitude [drive ct] 0.052 400 0.08 7200 16160
Amplitude [m_pk] 2.87e-17 8.73e-20 2.55e-18 3.62e-17 5.42e-19
where the parameters of the sensing function, C, are as listed in T1600278, and the scalar gains in the actuation function, A, and described in T1500377.
On the conversion between raw [m] of each data stream into [m] of amplitude at the calibration line frequency:
produceofficialstrainasds_O2_C01.m runs the matlab function pwelch over the time series of the data stored in the frames,
[ifo ':' subSystemTag '-CALIB_STRAIN' channelSuffix ]
[ifo ':CAL-PCALY_RX_PD_OUT_DQ']
where subSystemTag is either GDS or DCS, and channelSuffix is either C00 or C01. I used all of the same pwelch/FFT parameters as are used normally in producing G1500623, except reducing the number of averages by a factor of 10 and increasing the frequency resolution,
nAvgs = 10;
detrendOrder = 1;
windowType = @hann;
freqRes = 0.01; % [Hz]
percentOverlap = 50; % %
Making sure that the chosen frequency resolution results in the calibration lines being bin-centered (I chose 0.01 [Hz] bin width), pwelch spits out a power spectral density in units of [(power)_{RMS} / Hz]. So to convert to (amplitude)_{pk}, I do the following (since, after a little pre-processing of the frame data, both channels are in start in [m]):
sqrt( pwelch output ) * sqrt(ENBW) * sqrt(2)
where pwelch output is in [m^2/Hz], ENBW is the effective noise bandwidth for a Hann window (with normalized effective noise band width NENBW = 1.5) with a frequency resolution of 0.01 [Hz] i.e. ENBW = NENBW*f_res = 0.015 [Hz], and the sqrt(2) is to convert from RMS to Peak amplitude.
Did you get a VERBAL alarm or DIAG_MAIN message about the Noise Eater? Just curious. (Or should operators know to look for the Noise Eater when the IMC isn't locking.)