Calibration/validation of the AS_Q_FAST channels
Rick Savage - LHO
Stefanos Giampanis – Univ. Rochester
( Daniel Sigg – LHO )
LIGO-G060106-00-W March 2006 LSC meeting - LHO 2
High frequency length response
1FSR 2FSR
LIGO-G060106-00-W March 2006 LSC meeting - LHO 3
Length response TF detail
DC
1FSR
LIGO-G060106-00-W March 2006 LSC meeting - LHO 4
Comparison with single pole approximation
1% error at 1 kHz
12% error at 10 kHz
LIGO-G060106-00-W March 2006 LSC meeting - LHO 5
“FAST” channels
All three interferometers have FAST (262 kHz) data acquisition systems generating AS_Q_FAST signals with 100 kHz bandwidths
The signals are also digitally heterodyned at 1024 Hz and 37.504 kHz and downsampled to 2048 Hz to generate the 0FSR and 1FSR channels» AS_Q_0FSR.mag AS_Q_1FSR.mag
» AS_Q_0FSR.phs AS_Q_1FSR.phs
» AS_Q_0FSR.real AS_Q_1FSR.real
» AS_Q_0FSR.img AS_Q_1FSR.img
At LHO the full 262 kHz AS_Q_FAST channels are written to the frames. Not yet implemented at LLO.
LIGO-G060106-00-W March 2006 LSC meeting - LHO 6
AS signal flow
AS_Q path
Whitening filter
ICS 130 ADC (262 kHz)
Monitor outputs a b a b a b a b
AS portphotodetectors
DAQ channels AS_Q_FAST and 0FSR and 1FSR channels
AS1 demod AS2 demod AS3 demod AS4 demod
AS_Q_FAST path
Pentium front-end processor (e.g. h1adcufast)
LIGO-G060106-00-W March 2006 LSC meeting - LHO 7
Hardware – AS demod end
ASnoutputs
ASn-monoutputs(x 2 gain)
whitening filters for ICS 130
LIGO-G060106-00-W March 2006 LSC meeting - LHO 8
Hardware – ADC end
ICS 130 – 8 channelsrunning at 262,144 Hz
Pentek code generates the AS_Q_FAST and hetrodyned channels
LIGO-G060106-00-W March 2006 LSC meeting - LHO 9
Whitening filter transfer function
100
101
102
103
104
105
-20
-10
0
10
20
30
40
50
Freqency (Hz)
Mag
nitu
de (
dB)
H1 AS_Q_FAST whitening board transfer function
100
101
102
103
104
105
-200
-100
0
100
200
Freqency (Hz)
Phas
e (
deg
.)
gain of 0.2, four zeros at 20 Hz, four poles at 100 Hz, poles at 150 kHz and 300 kHz
LIGO-G060106-00-W March 2006 LSC meeting - LHO 10
ASPD summation for AS_Q and AS_Q_FAST
» Similar calculation done in adcufast
» Presently, startup.cmd must be updated manually every time medm settings are changed
» LLO adjusts ASPD matrix coefficients for each lock stretch
» Working toward reading epics settings on the fly.
LIGO-G060106-00-W March 2006 LSC meeting - LHO 11
AS_Q : AS_Q_FAST comparison
LIGO-G060106-00-W March 2006 LSC meeting - LHO 12
AS_Q_FAST signals during full ifo. lock
LIGO-G060106-00-W March 2006 LSC meeting - LHO 13
0FSR signals
LIGO-G060106-00-W March 2006 LSC meeting - LHO 14
1FSR signals
LIGO-G060106-00-W March 2006 LSC meeting - LHO 15
Calibration options
Drive test mass at 37.5 kHz» Attempted during Oct. 2005
– Factor of five larger value than expected– Electronics not designed to be driven at this frequency– Uncertainty in force/displacement too high
Photon calibrator» Now driving at ~ 1600 Hz
– To resolve peak in ASD factor at the level of a factor of a few above the background requires hundreds of seconds of integration
– At 37.5 kHz one data point would require ~ one year integration time
Thermally-excited TM internal mode vibrations» Modeling (and more thought) required
LIGO-G060106-00-W March 2006 LSC meeting - LHO 16
“Blind” extrapolation of low-freq. calibration
Use cal line at ~1150 Hz Replace cavity pole approximation with length
response given by HL(f) Extrapolate to 100 kHz without HF measurement
» Analytic model used to date for single FP cavity - probably gets differential mode right , but not leakage of common mode effects into AS port – doesn’t model noise floor.
» Use full interferometer model such as Finesse to study noise couplings at high frequencies.
Effort underway to implement “blind” calibration
Top Related