[Shingo, Dan, YamaT]
Quick estimation of calibration parameters is as follows.
We haven't found strange results.
Detailed analysis (mainly for Low-Latency calibration) will be done later.
We haven't finished measurements of the difference between no dewhitening filter and 3 dewhitening filters on ITMX yet.
This is made 0.9dB bias on DARM calibration by Free-swing and PRMI calibration by Pcal in theoretically.
(DARM by pcal and PRMI by Free-swing aren't affected.)
So some remaining measurements will be done tomorrow.
- Optical gain of mich by Free-swing Mich
1.1475e+09 +/- 0.0062e+09
20min (10min before and after mich OLTF measurement) average.
POP1DC shows sometimes negative value and contrast becomes larger than 100%.
So dark offset seems to be overestimated.
- Actuator efficiency of BS (no dewhite). Fig.1-2
6.128e-11 +/- 0.035e-13 (+1.1%)
1Hz normalized
- Actuator efficiency ratio between ITMX (no dewhite) and BS (no dewhite) Fig.3-4
0.04965 +/- 0.00033 (+3.2%)
- Actuator efficiency of ETMX_TM by xpcal-1 Fig.5-6
3.8818e-12 +/- 0.0043e-12 (-0.7%)
1Hz normalized
- Actuator efficiency of ETMX_IM by xpcal-1 Fig.7-8
1.84e-10 +/- 0.10e-10 (+9.6%)
1Hz normalized
- Actuator efficiency ratio of ETMX_IM and ETMX_TM Fig.9-10
45.7 +/- 2.3 (+6.4%)
- Optical gain of DARM Fig.11-12
2.2574e+12 +/-0.0072e+09 (-3.7%)
- xpcal in the 1st lock vs one in the 2nd lock Fig.13-14
0.99828 +/- 0.00077
Optical gain is ~0.17% stability in two locks.
It's nice reproducibility!
- xpcal vs ypcal in the 2nd lock Fig.15-16
0.98663 +/- 0.00090
Xpcal and Ypcal have ~1.34% inconsistency each other.
It seems slightly larger than accuracy of integration sphere calibration of both pcals.
Does this inconsistency become dominant source of errors on pcal itself?
- BS_TM (dewhite 1) vs BS_TM (no dewhite) Fig.17-18
1.0376 +/- 0.0035
It's consistent with low frequency zero/pole effect of disabled dewhtiening (~0.3dB per stage)
- Actuator efficiency of BS_TM (dewhite 1)
6.358e-11 +/- 0.021e-11
- Optical gain of MICH Fig.19-20
4.215e+08 +/- 0.014e+06 (-5.9%)
- PRM_TM vs BS_TM (dewhite 1) Fig.21-22
26.95 +/- 0.10 (-2.0~+10.0%)
use only above 70Hz
accurate analysis with correcting MICH OLG effect will be done later.
past results is no enough consideration about MICH OLG
- Actuator efficiency of PRM_TM (dewhite 1)
8.568e-10 +/- 0.043e-10 (+15.8%)
1Hz normalized
- Optical gain of PRCL Fig.23-24
1.9714e+09 +/- 0.0099e+06 (-11.2%)
I performed transfer function measurements from ITMX_TM to DARM under two conditions: with no dewhitening filter and with all three dewhitening filters engaged, while the interferometer was in the PRFPMI_RF_LOCKED state.
Figure 1 shows the comparison between the WF ON and WF OFF cases. The result is consistent with the expected effect from the low-frequency zero/pole, showing approximately a 0.9 dB (~10.9%) difference.
I estimated calibration parameters including pcal beam spot effect (klog#34039).
All DARM-related parameters were updated as of 6/3 at 13:25.
1. Parameters in sensing
Optical_gain = 2.1994e12 (-6 from % klog#33643)
Cavity_pole = 18.078 (+0.43 %)
2. Actuator efficiencies
H_etmxtm = 3.8756e-14 at 10Hz (-0.85 from % klog#33944)
H_etmxim = 1.6451e-14 at 10Hz (-1.95 from % klog#33609)
The difference in the results is considered to arise from the PCAL correction and the frequency range used for fitting. After discussing with Yamamoto-san, we decided to update the values.
Modified IFO guardian
We raised CFC check latch for testing READY state.
IFO guardian might speak repeatedly for a while, so sound notification in READY state of IFO guardian was disabled.
It will be enabled again later.
Missed DARM related filters in actuation path
Because some actuation filters hadn't copied from VIS model to CAL model (see Fig.1) when optical gain and actuator efficiency were updated, I copied them. Copied filters are shown in Fig.2. Fortunately, they are not engaged in the observation state, so there is no impact on calibration even if they are not copied. But remaining non-copied filters may hides important changes, they should be copied even if they are not used in the observation state.
By copying filters, CAL_PROC guardian dropped CFC_LATCH because filter updates were detected and then IFO guardian backed to CALIB_NOT_READY (Fig.3 -> Fig.4). In this time, we knew this was not a impact on calibration and observation, I loaded filter coefficients on k1calcs.
Update of MICH/PRCL calibration
MICH and PRCL calibration parameters were also update as values in klog#34033. Detailed changes in foton filters as shown in Fig.5.
Finally, I raised CFC_LATCH flag and keep IFO guardian in READY state.
