Transcript Calibration of the gravitational wave signal in the LIGO

Calibration of the gravitational
wave signal in the
LIGO detectors
Gabriela Gonzalez (LSU), Mike Landry
(LIGO-LHO), Patrick Sutton (PSU)
with the calibration team: Rana Adhikari
(LIGO-MIT), Brian O’Reilly (LIGO-LLO), and
many others…
Gravitational Wave Signal
• Originates in a photodetector at the
antisymmetric port.
• It is a demodulated photocurrent:
(L1,H1,H2):LSC-AS_Q
• It is proportional to the differential length of
the arm cavities, and thus the gw strain.
• In the frequency domain, AS_Q(f)=C(f)h(f)
• C(f): calibration function (complex)
Detector control system
interferometer
digital filters
pendulum
Gravitational wave strain: h=Xext/L
H(f)=open loop gain=G(f)A(f)C(f)
calibration
Detected signal:
Calibrated spectrum
Calibration lines
Open Loop Gain: Fits
Simulink model (R. Adhikari)
Open Loop Gain: Fits
Parameters in the fit:
- Time delays
- Optical gain
Changing optical gain
a=scale factor for
open loop gain
Changing optical gain
Scale factor for open loop gain
Scale factor for Open Loop Gain
Interferometer sensitivity and
alignment
Carrier power in the arm cavities
Sideband Power in the recycling cavity
Signal  carrier x sideband
Changing optical gain
LIGO Autocalibrator
• Tcl script to expedite calibration
• Calls to DTT (ifo transfer function, psd estimate),
Root and/or Matlab manipulation
• Output (web-navigable) data in ascii, xml, pdf
(transfer functions, noise data, olg, sensing
functions)
• Pole-zero fitting useful for rapid commissioning
• New version interpolates TF, no fits
Autocalibrator vs S1 analyses
V2 Autocalibrator
•Latest version does
Interpolation of transfer
Function (no fit problems)
•Not as flexible for
Commissioning
•Adding matlab fit and
DTT cal record shortly
Autocalibrator for S2
• S2 run ~per day per ifo
• In addition to noise curves, output open
loop gain and sensing function as a check
on point calibration/alpha propagation
method
• Brian O’Reilly has version running at LLO
online resources
• online resources include the calibration homepage
– http://blue.ligo-wa.caltech.edu/engrun/Calib_Home/
• history (ascii strain and displacement vs time, elog pointers)
• commissioning (online autocalibration automatic storage of data)
• science (S1) and engineering (E2-E9) runs
• And the official page on LIGO sensitivity
– http://www.ligo.caltech.edu/~lazz/distribution/LSC_Data/
E9 & S2 Calibrations from
SenseMonitor
• SenseMonitor: A DMT program which
provides real-time estimates of the average
range to which an IFO is sensitive with
SNR > 8 to the inspiral of a 1.4-1.4Mo
neutron-star binary.
• Runs continuously at LLO, LHO.
• Tracks the response function of the IFO to
make its range estimate, therefore is a
natural tool for providing near-real-time
calibration information.
What SenseMonitor Needs
Reference calibration information supplied
to SenseMonitor:
–
–
–
–
Measured/modelled open-loop gain H(f) and
sensing function C(f) of the IFO, supplied in
ASCII data files.
RMS amplitude of the 1kHz calibration line in
AS_Q during calibration run.
Injected amplitude of the 1kHz calibration line
in EXC during calibration run.
Digital gain (`GW_K' ® b during S1) during
calibration run.
What SenseMonitor Gives
For each 60 sec of AS_Q data,
SenseMonitor:
–
–
–
measures the current amplitude of the 1kHz
calibration line.
using the current line amplitude and the
reference data, computes the new value of
calibration parameter `a'.
exports a, b, range data to DMTViewer.
Once per hour:
–
exports a, b, range data to trend frames.
Reference Calibration:
AS_Q Data:
amplAS_Q = 0.00160
amplEXC = 0.25
beta = 1.0
H(f), C(f):
at start 1/minute
3.00e1 5.28e+0 4.47e+0
...
2.00e3 2.45e-1 2.13e-2
SenseMonitor
1/hour
1/minute
Trend frames:
DMTViewer:
a, b, range, ...
a, b, range, ...
Caveats
Up to 1-hour latency in writing trend frames.
Reference calibration data (open loop gain,
sensing function, etc.) not available in trend
frames.
Calibration lines are not robust (sometimes
crash); alpha is not updated if the line
amplitude is <50% or >150% nominal.
Alpha, beta set to zero if IFO is not locked
in common mode.
Expectations for S2
• Good models
• Frequent checks (autocalibrators)
• Automatic on-line calibration and reach
measurement (SenseMon)