We replaced the coil driver for the accelerometers from HPC (S1604763) to LPC (S1707379).
And we changed the gain of servo filter by 17.3db for H1, by 17.4db for H2, by 18.2db for H3.
This log report covers two aspects:
(1) We measured the transfer functions from IP's actuators to the geophones and the accelerometers.
(2) We switched the signals to be used for controlling EX IP and checked whether IP control works well or not.
We measured the transfer functions from IP's actuators to the geophones (ETMX_IP_ACCBLEND_ACCL, T, Y) and the accelerometers (ETMX_IP_ACCBLEND_FLDACCL, T, Y).
And we checked the differences of phases between the geophones and the accelerometers.
The transfer functions were attached.
We switched the signals to be used for controlling EX IP from the geophones (ETMX_IP_ACCBLEND_ACCL, T, Y) to the accelerometers (ETMX_IP_ACCBLEND_FLDACCL, T, Y) and checked whether IP control works well or not.
The spectra were attached when we used the signals from accelerometers.
The spectra's name are ETMY_IP_ACCBLEND_ACCL,T,Y and ETMY_IP_ACCBLEND_FLDACCL,T,Y.
The picture of the spectra was attached.
1) I tried to match the values of power spectra measured by the accelerometers and the geophones (nemed ETMX_IP_ACCINF and ETMX_IP_FLDACC).
2) I tried to match the values of power spectra measured by the accelerometers and the geophones (nemed ETMX_IP_ACCBLEND_ACCL, T, Y and ETMX_IP_FLDACC_FLDACCL, T, Y)
I tried to match the values of power spectra measured by the accelerometer and the geophones (nemed ETMX_IP_ACCINF and ETMX_IP_FLDACC).
To do this, I changed the servo filter named ETMX IP FLDACCINF FILTER.
The picture of results was attached.
I tried to match the values of power spectra measured by the accelerometer and the geophones (nemed ETMX_IP_ACCBLEND_ACCL, T, Y and ETMX_IP_FLDACC_FLDACCL, T, Y).
To do this, I changed the matrix named ETMX IP FLDACCALIGN MATRIX.
The picture of results was attached.
,H2,H3 and ETMY_IP_FLDACCINF_H1,H2,H3) measured by the geophones and the accelerometers.
The picture of the power spectra was attached.
The spectra's name are ETMY_IP_ACCBLEND_ACCL,T,Y and ETMY_IP_ACCBLEND_FLDACCL,T,Y.
The picture of the spectra was attached.
This log report covers one aspects:
(1)We switched the signals measured by the geophones to the signals measured by the accelerometers and checked weather IP control works well or not.
We switched the signals named ITMX IP ACCBLEND ACCL,T,Y to the signals named ITMX ACCBLEND FLDACCL,T,Y temporary and checked weather IP's feed back control works well or not.
Before switching these signals, we measured the transfer functions from ITMX IP test_L,T,Y to ITMX ACCBLEND ACCL,T,Y and ITMX ACCBLEND FLDACCL,T,Y respectively.
The pictures were attached.
We changed the gain of ITMX IP ACCBLEND FLDACCY to match the phase of ACCY and FLDACCY.
After switching the signals, we measured the power specrtas' of ITMX IP ACCBLEND measured by the geophones and the accelerometers.
The picture was attached.
This log report covers two aspects:
(1) I changed the gain of servo filter named ITMX IP FLDACCINF by 0.18 for H2
(2) I changed the matrix's values named ITMX IP FLDACC2EUL MATRIX temporary.
I changed the gain of servo filter named ITMX IP FLDACCINF by 0.18 for H2 to match the power spectra's value(ITMX_IP_ACCINF_H2 and ITMX_IP_FLDACCINF_H2) measured by the geophones and the accelerometers.
The picture of the power spectra was attached.
I changed the matrix's values temporary named ITMX IP FLDACC2EUL MATRIX to match the power spectra's value(ITMX_ACCBLEND_ACCL,T,Y and ITMX_ACCBLEND_FLDACCL,T,Y) measured by the geophones and the accelerometers.
I can't match these value, and finally I set the matrix's value that is set originally.
The picture of the power spectra was attached.
The servo filter of FLDACCINF's location is medm / ITMX / ITMX TWR / Pre-Isolater / FLDACC / FLDACCINF.
This log report covers two aspects:
(1) We changed the value of ITMY IP FLDACC2EUL MATRIX. We did it 2023/1/18. I should write down this article immediately. I'm sorry.
(2) We switched the signals of ITMY IP ACCBLEND temporary to checked weather IP control works well or not. We did it 2023/1/20.
We changed the value of ITMY IP FLDACC2EUL MATRIX to match the power spectras' value of ITMY_IP_ACCBLEND_ACCL, ACCT, ACCY, and ITMY_IP_ACCBLEND_FLDACCL, FLDACCT, FLDACCY respectively.
We changed the value of the matrix as follows.
0.6439 → -0.6439 matrix[1,1]
0.1725 → -0.1725 matrix[2,1]
0.5635 → -0.5635 matrix[3,1]
Fig1 is the matrix.
And we also attached the picture of the power spectras. Fig2 is before changing the matrix, and Fig 3 is after changing it.
We switched the signals of ITMY IP ACCBLEND temporary from the geophone to the accelerometer to checked weather IP control works well or not.
The signals from the geophone are ITMY IP ACCBLEND ACCL, ACCT, and ACCY. The signals from the accelerometers are ITMY ACCBLEND FLDACCL, FLDACCY, and FLDACCY.
Before switching the signals, we checked the transfer functions and phases of the geophone and the accelerometer. The transfer functions are same, but the phases were different by 90 degrees. So, we adjust the accelerometer's gain of ITMY IP ACCBLEND to match the phase of the geophone and the accelerometer.
We attached the pictures of the power spectra.
Fig4 is before switching the signal, and Fig5 is after switching the signal.
These pictures were taken before adjusting gain of the accelerometers.
These pictures were taken before adjusting gain of the accelerometers.
We measured the power spectra after replacing the coil driver for the accelerometers from the LPC for Cryo-payload(S1707386) to the LPC(S1604943).
And we adjusted the gain of the servo filter named FLDACCINF by 0.17dB for H1, 0.17dB for H2, 0.2dB for H3 to match the power spectra's value measured by the geophones and the accelerometers.
We attached the pictures of the power spectra.
The first one is measured before replacing the recoil driver, and the second one is measured after replacing the coil driver.
The servo filter of FLDACCINF's location is medm / ITMY / ITMY TWR / Pre-Isolater / FLDACC / FLDACCINF.
Here is the calibrated result of the long-term measurement of the modulaiton index
== Parameters ==
measurement date: 0502
frequency: f3
phase diff: 40 [deg]
== Note ==
This is the end of the long-term measurement of modulation indices.
I'll analyze the whole of data and get more informatioin, like the standard deviation depending on the phase difference.
5/3
- Crontab stopped at 12:55
- This is the end of the long-term measurement of modulation indices.
Here is the calibrated result of the long-term measurement of the modulaiton index
== Parameters ==
measurement date: 0501
frequency: f3
phase diff: 0 [deg]
== Note ==
Because f3 is designed to have modulation index of 0 at 0 deg, it's almost on noise level.
* In a sense, this is a good confirmation for the functionality of MZM.
Here is the calibrated result of the long-term measurement of the modulaiton index
== Parameters ==
measurement date: 0430
frequency: f3
phase diff: 120 [deg]
== Note ==
Because we calibrate the PM index like...
(PM index) = sqrt((total index)^2-(AM index)^2))
the PM index is sensitive at the phase difference where the AM index and total index are designed to be identical in principle, means PM index is 0.
*Reference: f1 at 120 deg (8767)
Because f3 is designed to be under this condition at any phase difference, PM index of f3 is difficult to calibrate with the high precision.
Here is the calibrated result of the long-term measurement of the modulaiton index
== Parameters ==
measurement date: 0426
frequency: f1
phase diff: 120 [deg]
== Notes ==
For this plot, I removed the data from 14:40, starting time, to 18:45 on 4/26 because the data during ths period sometimes looks weird. I imagine it had lock loss.
In the end, the data from 18:50 on 4/26 to 15:55 on 4/27 is used.
Here is the calibrated result of the long-term measurement of the modulation index
== Parameters ==
measurement date: 0425
frequency: f1
phase diff: 0 [deg]
Here is the calibrated result of the long-term measurement of the modulation index
== Parameters ==
measurement date: 0424
frequency: f1
phase diff: 80 [deg]
I measured the long-term trend of the modulation index of f1 with 0 [deg].
Additionally, we need to measure with..
f1: 80 and 120 [deg]
f3: 0, 50 and 120 [deg]
Taiki Tanaka, Kohei Yamamoto
Related Link: 8577
==Background==
1.As pointed out in 8577, the clipped beam affected the measurement of the modulation indices.
2.From the perspective of the security management, the situation was not good that we needed to open the shutter for the measurement of the modulation indices.
==What we did==
1. Picked off the beam reflected by the 2-inch lens, named L8
2. Aligned the beam to RAM monitor and AM/PM monitor
3. Measured the modulation index with the phase difference b/w EOMs shifted
==Result==
1. The new configuration is shown in the first figure.
2. The measured modulation indices look what is expected from the theory.
3. We don't have to care about the shutter for the measurement of the modulation indices any more
==Next==
With this configuration, we are going to finally measure the good data of the long-term trend of the modulation indices.
I measured the long-term trend of the modulation index of f1 with 0 [deg].
Additionally, we need to measure with..
f1: 80 and 120 [deg]
f3: 0, 50 and 120 [deg]
Kokeyama, Ge, Kohei Yamamoto
Related Link: 8576 (Dependency on beam pointing), 8329 (Clipping at the periscope), 8441 (Long-term measurment of modulation indices)
In 8576, we investigate the dependency of the AM index on the beam pointing.
Furthermore, we also checked how the clipping of the laser happened at the periscope, see 8329, affected the measurement of the AM index.
= What we did (with PM monitor)=
1. Picked off the laser before the periscope so that non-clipped laser hits on the PM monitor, see 1st figure.
2. Measure the modulation indices with changing the phase difference b/w EOMs in the 1st MZI, see 2nd figure.
3. Remove the beam sampler and the mirror for the pick off, and aligned the laser clipped at the periscope to the PM monitor.
4. Measure the AM index of the clipped laser with the phase difference with which the AM index was maximized, see 3rd figure.
= Condition =
As mentioned in 8576, we tuned the beam pointing so that the laser power is maximized.
= Results =
The non-clipped laser showed the behaivor expected by the theory, while the AM index of the clipped laser is so high that it is larger than the total of the modulation index, sqrt(AM^2+PM^2).
(We don't know how this can happen tho.)
= Conclusion =
It is necessary to fix the clipping for the long-term measurment of the modulation indices.
Kokeyama, Kohei Yamamoto
These days, we have faced the wierd behavior of the AM index, so we investigated this.
=What we did (with RAM monitor)=
1. Check the behavior of the AM index with changing the beam pointing
2. Check the behavior of the AM index with changing the laser power by HWP.
=What turned out=
From the first figure, we concluded that the AM index highly depends on the beam pointing. We don't really understand the reason tho.
In the second figure, we showed the result of the same investigation with the PM monitor, and there is the flat region around the pointing where the laser power is maximized.
For now, we think the value of this region is the "Actual AM index".
=What remains to be investigated=
We should understand what causes this spacial dependency of the AM index.
From yesterday's night to today's morning for almost 14hours, I measured the modulation indices of the f1 AM and PM.
The result is shown in the figure.
== setting ==
1. I remotely controlled the Moku:Lab in PSL room.
2. I introduce ZERO phase difference b/w EOMs in 1st MZI, leading to the pure PM in principle
3. I got the data every 5 mins.
== to do ==
1. Need to check the consistency b/w the results of Moku:Lab and the double demodulation.
2. Do the same measurement with NON-ZERO phase difference, where the modulation index get more sensitive to the phase difference b/w EOMs.
Yesterday, I changed the connection of the signals a bit as shown in the picture.
==Background==
Although we planned to get the V80 with "RF MON" on the back side of the I/Q demod according to the schematic in JGW-G1909628, we found that the DC signal is cut by the directional coupler and RF (power) detector in the circuit, and it is only the sum of the power of RF signals which we can measure with the port.
see: https://gwdoc.icrr.u-tokyo.ac.jp/DocDB/0024/D1402413/001/IQ_demod.PDF
==What I did==
1. I just splitted the RF signal before the I/Q demod so that the one goes to the I/Q demod.
2. The other goes to ADC after passing through a 20dB attenuator and LPF.
==Note==
From "MIF box", I borrowed two for each of SMA-to-BNC connector, BNC-to-SMA connector, SMA gender changer (Male-Male), 20dB attenuator and LPF shown in the picture.
Ge, Kohei Yamamoto
Today, we gave a minor change to the cabling.
1. Connect the RFPD to a new RF splitter
2. One of the output of the splitter goes to Moku:Lab
3. The other to the system of the double demodulation
With this setting, we can simultaneously measure the modulation indices with the Moku:Lab and the double-demodulation system.
