NAOJ GW Elog Logbook 3.2
Tabletop experiment of speed meter
Yohei Nishino & Munetake Otsuka
-Configuration
1)Laser -> 2)QWP -> 3)HWP -> 4)Convex lens -> 5)BS -> 6)HWP
-> 7)FR -> 8)Rotary Disc Shutter -> 9)HWP -> 10)AOM -> 11)Mirror
-> 12)Mirror -> 13)Convex lens -> 14)EOM
-Optical power mesurement of 10)AOM
1st diffraction light 1.06 mW
Transmitted light 0.69 mW
while input light 1.81 mW
-Optical component specifications of 10)AOM
Diffraction efficiency : 60% (=1.06/(1.06+0.69)
-Configuration
9)HWP : 51
Function generator for 10)AOM : 40MHz 5V(0 to peak)
-Beam profile after 13)lens
position x-width y-width
700mm 0.29mm 0.38mm
750mm 0.28mm 0.20mm
800mm 0.33mm 0.34mm
850mm 0.49mm 0.60mm
Using data from the long polarization states generation of elog3308 I saved all the voltages required to generate a rotating HWP with 0.1deg error in ellipticity.
The 1st column is V1 and second column V2. They are sorted so that the HWP rotates from -90 to +90 deg.
The file is 'rotatingHWP.csv'
I set a monotor camera tentatively in the center room of TAMA at the request of the safety office. If the test operation was reasonable, I would set more cameras at some points.
Today, I disassembled the setup for cleaning.
I did not have the CAD drawing, so I winged the disassembly and only broke the electrostatic actuator.
The components are now on the optical table in plastic bags (fig. 2).
I will clean them and reassemble the setup next week (probably Wednesday).
I made double pendulum made by inverted pendulum and Roberts' Linkage.
So I measured the transfer function from IP actuator power to IP displacement and Roberts' Linkage displacement to check how it is.
Each set up pictures are attached(Fig1, Fig2).
IP displacement are measured by LVDT mounted on IP top stage.
Roberts' Linkage displacement are measured by photo displacement sensor.
Each transfer function are attached(Fig3, Fig4).
From Fig3 gain plot, I can realzed that common mode and differrential mode peak and characteristical dip.
I can also realized that points from Fig3 phase plot.
At 400mHz peak is IP yaw motion, and at 1Hz peak is Roberts' Linkage paw motion.
I mentioned about Fig4 lately.
We suspend Roberts' Linkage from IP top (image: Fig1, picture: Fig2). The parameters were as follows.
| high [m] | 0.4 |
| weight [kg] | 28.73 |
I also measured the power spectra with using Photo displacement sensor(Fig3). During measuering, we locked IP physically using meal fitting.
I attached the results (Fig4).
I can realized that there are two peak. These peaks are also from resonant frequency.
I thought it from difference between wire lenghts.
The optics are in the small desiccator (fig.2), and the pedestals and mounts are in a plastic box (fig.3).
I found some mirrors had scratches on them. In hindsight, I should have marked them. (Anyhow, it can be sorted later by the user.)
I will clean up the surrounding area after moving the optical table.
The option to choose a specified configuration from a pre-saved file has been added. As a result you can load the list of LC voltages required for any particular scenario.
[Shalika, Gabriella]
The setup was modified and two beam splitters were added (before and after LCs). The measurement to calibrate the beam splitters (both reflection & transmission) were taken as well (done without LCs).
What I did:
I fited the resuts of IP's resonant frequency to theoretical value.
The equation used for fitting was atttached.
f is IP's resonant frequency.
k_e is spring constant of IP's flex joint .
k is s anti spring constant.
g is gravity constant.
l is length of IP.
m_0 is mass of IP.
M is weight on IP.
From fitting, m_0 is 1.08 kg, k_e is 131 N/m, l is 2.48 m.
The results and fitting picture was as follows.
X-axis is weight on IP top, y-axis is IP's resonant frequency.
What we did:
We diagonalized the matrix of IP, and checked how much diagonalization was.
The matrix's values form LVDTs to imaginally L-axis and T-axis were as follows.
| L | T | |
| H1 | +0.20 | -0.04 |
| H2 | -0.07 | -0.12 |
| H3 | -0.09 | +0.13 |
The matrix's vakues from L-axis and T-axis to actuators were as follows.
| A1 | A2 | A3 | |
| L | -2.20 | +2.45 | +0.46 |
| T | +0.45 | +3.21 | -2.71 |
I measured the trasfer function from L,T to L,T respectively.
The pictures was attached.
I concluded it was good at some extend.
The measurement seems to go fine and is mostly finished.
However, because of the new failure of air conditioning the temperature and humidity increased quite a lot in TAMA.
It's about 30deg in main building and quite more inside clean room.
just for safety I turned off PCI laser.
Mitushahshi
I summarized the tabletop experiments of Roberts' Linkage.
What I did
The followings are what I did.
- I measured the resonant frequency of Roberts' Linkage with changing COM.
- I checked the difference of resonant frequencys between x-axis and y-axis(The definitions of these axes are written on below sentence).
Set up
I attached the picture(Fig1).
The Roberts' Linkage was shaked by coil magnet actuator powered by FFT analyzer.
The displacement of it's was detected by photo sensor.
I measured the transfer function from coil magnet actuator voltage to photo sensor read out voltage.
Measurement results
Firstly I mentioned resonat frequency and COM position. I attached the picture(Fig2).
The results says that the resonant frequency get lower and lower when Roberts' Linkage was tuned at hight COM points.
The gap are there between first mesurement and second measuremt. I thought the reasons are some coupling by different tension on wires and by human error with tuning COM(this is most properly).
Secondary, I mentioned the strange peak. I attached the power spectra density plot.
The peak at 71mHz is the resonant, and harmonic of it also are there(rabeled blue).
On the other hands, the strange peak that it is not reasonable by above mention also are there(label green).
The paper "F. Garoi et al., Passive vibration isolation using a Roberts Linkage, Rev Sci Instrum 74, 3487-3491(2003)" reported that there are cross coupling.
So I measured the x-axis and y-axis resonat frequency. The definition of x-axis is following on the road of 20m interference site(Fig3).
I attached the results. The red one is x-axis, and the blue one is y-axis.
I realized that these axes have a difference resonat frequencys, and conclued that the strange peak is from y-axis' resonant frequency (cross coupling).
[Aso, Marc, Michael, Rishabh]
We installed spare ETMY with ears up (ie reversed top/down with respect to KAGRA installation) on the translation stage.
It helped a lot to remove the front top right metal support looking from the laser side for the installation.
After small realignment of the 2 photodetectors I checked the mirror center by trying to find the position where the signals drop to 0 (ie edge of the mirror).
X_center = 393.5 mm and Y_center = 155.5 mm.
Then I started measurement with Spol at the input (HWP = 351.2 deg) with maximum AC and DC dynamic range (1V) to take map over 70 mm radius with 1 mm step.
I'll keep these settings during holydays to avoid saturation.
The measurement seems to go fine and is mostly finished.
However, because of the new failure of air conditioning the temperature and humidity increased quite a lot in TAMA.
It's about 30deg in main building and quite more inside clean room.
just for safety I turned off PCI laser.
I realigned the beam at normal incidence using the razor blades using input power of 1.59mW.
After realignment, I got angle of incidences of 0.009 deg and 0.0059 deg for vertical and horizontal direction respectively.
I tuned the input HWP and QWP to have input s polarized light by minimizing the p polarization component.
With current configuration, 'DC' corresponds to S pol and 'AC' to P pol.
I took 10mn of data with each polarization to finalized the setup calibration.
I was planning to measure the ILM sample birefringence but could not find suitable holder (30mm diameter ; 15 mm thickness).
We plan to start instead with spare ETMY measurement.
I reinstalled the output cross-polarizer and also installed a HWP (WPQ-10640-2M from SigmaKoki) just after LC2.
I am now scanning the LCs voltages with various step to measure birefringence in 'electronic Soleil-Babinet' mode.
The measurement started in elog 3295 finished.
It consists of about 6.3 millions couples of votlages applied to LC1 and LC2.
Figures 1 to 3 reports the resulting polarization states which is similar to previous measurement despite the 90 deg rotation of LC1.
Somehow we have an overall phase difference of 90 deg as before. I'll cross-check simulations.
When measurement finished, the LCs temperature was about 31degC. This indicates a stronge heating as the usual LC temperature is about 24 degC before heating for temperature control.
This seems to be related to the air conditioning failing noticed by Takahashi-san that could also explain the strange power fluctuations of the FDS laser reported in elogs 3306 and 3307.
8/1 [Chien-Ming and Hsun-Chung]
We measured around 5.4dB squeezing and 11.4dB antisqueezing at 25mW green pumped,
and 5.7dB squeezing and 16.7dB antisqueezing at 40mW green pumped.
Without green pump, OPO T=7.107 kOhm; BAB=192mV, p-pol PLL: 225MHz
At 25mW pump, OPO T=7.118 kOhm; BAB max=1.27V, min=80mV, p-pol PLL:165MHz
Nonlinear gain=6.61
At 40mW pump, OPO T=7.120 kOhm; BAB max=3.28V, min=40mV, p-pol PLL:130MHz
Nonlinear gain=17.08
The SHG power in front of MZ drops to 151.5mW today.
In addition, we found the LO power to HOM also drops to 1.68mW. We can't increase it by alignment, so we have to replace the OD0.3 ND filter before IRMC to OD0.1.
As a result, the LO power is restored to 1.94mW.
We suspect that the output power of the main laser may be continuously decaying, causing the SHG and LO power to drop!
Attached are two squeezing measurement at different pump powers. It is worth mentioning that when the HOM DC cable is too close to the oscilloscope, some noise peaks will appear obviously, especially on dark noise.
The units setting parameters of SR785 was also attached.
7/31 [Chien-Ming and Hsun-Chung]
We investigate the reasons for the squeezing level decline on 28th July http://www2.nao.ac.jp/~gw-elog/osl/?r=3305.
Originally we suspected it was caused by poor cleaning of the optics, but today we found out that the phase-matching temperature of OPO is not at the optimal value.
On 7/28 we set OPO T=7.159 kOhm with 25mW green pump and T=7.184 kOhm with 40mW green respectively in order to lock the p-pol at 270MHz.
Today we check that the optimal value of OPO T should be set near 7.107 kOhm to achieve the maximum nonlinear gain.
OPOT=7.107k; 25mW green; BAB Max=1.26V; p-pol=33MHz*5=165MHz.
Without green; BAB=196mV-6mV(Background)=190mV; p-pol=49MHz*5=245MHz.
Therefore the nonlinear gain=6.63 at 25mW pump.
HOM visibility:
LO at AMC: TEM00=5.96v; hom=2.4mV+2.4mV; mode-matching: 99.92%
BAB at AMC: TEM00= 196mV; hom=6.4mV+1.8mV; mode-matching: 95.98%
SHG output power in front of MZ drops to 165.5mW even if the SHG temperature and alignment are optimized (194.9mW on 7/28)
We preset the SR785, however, the parameters are incorrect so the squeezing measurement is weird today.
At first, the SHG is 119.7mW in front of MZ.
We correct the HOM DC balanced from -10mV to 0mV by fine-tuning HOM BS Yaw.
Hsun-Chung bought some glass dropper bottles to hold acetone (buy from HANDS).
We cleaned some optics on the optical path of HOM and SQ and then checked the BAB alignment at AMC.
TEM00=206mV; hom=5.6mV+5.6mV; mode-matching:94.8%
In the beginning, the OPO T was 7.117k(p-pol is 270MHz) with pump 25mW, the BAB max=52mV, min=24mV.
After setting the OPO T to 7.159k(p-pol is 270MHz), the BAB max= 952mV, min= 104mV
At the 25mW green pump, we measured 4.47dB squeezing and 8.43dB anti-squeezing which is worse than yesterday's result.
Therefore, those optics may not be cleaned well.
Then we measured twice FIS at the 40mW pump, got 5.28dB SQ and 12.13dB Anti-SQ at first, and 5.29dB SQ and 11.65dB ASQ at second.
Before the second measurement, the MZ failed to lock again. The reason is that the output power of the SHG seems insufficient (100mW in front of MZ).
By setting the temperature of SHG from 3.158k to 3.123k, the green light in front of MZ becomes 194.9mW.
The BAB max=1.54V and min=152mV at 40mW pump and OPO T= 7.182k (p-pol is 270MHz).
Without green pump, the BAB=234mV at OPO T=7.121k (p-pol is still 270MHz)
Therefore, the nonlinear gain is 6.58 at the 40mW pump.
The attachment is the comparison of the measurement data for two days
