Today I installed the razor blade that cuts the beam vertically and measured the pump beam parameter.

The blade is at 90 mm from the edge of the breadboard and I used about 100 mW of input power.

I measured a waist size of 22 um at 58 mm (in Manuel's unit) while we expect 36 um at 61 mm.

I started the tuning of the telescope but it is not concluded yet.

I subtracted the dark noise from the locking accuracy measurement done in elog2864. I calculated sqrt(locking accuracy^2-dark noise^2). In the point where the dark noise is larger than the locking accuracy, the subtraction is set to 0. The rms after the subtraction is almost the same as one before subtraction.

We have two iris inside filter cavity arm to check the direction of green laser beam. To make sure the beam can arrive at the filter cavity end mirror, we need to make the green beam hit on a point checked in the past when filter cavity is aligned.

Remember that, for the first step, we should make the green beam hit on several reference points on PR and BS chambers.

The first attached picture shows the beam position on the first iris. In this case, the iris needs to be rotated so that the stick is not visible (hidden on the back side of iris). At this moment, the beam position should be just above the hole.

For the first iris, when the stick is visible, the situation is shown in Fig. 2. At this moment, the beam is above the hole and located a bit the left side.

If the beam hits on the first iris like what is shown here, it should be not diffcult to find the green beam on the second iris. If not visible on the second iris, we should move the green beam around with little adjustment. Then what we need to do is to just make the green beam go through the hole of it.

Information

Date: 14th Apr 2022

Members: Dan Chen, Satoru Ikeda

Place: Kamioka Hokubu-kaikan 1F meeting room

Background

The mirror substrates made for KAGRA O5 may have high internal scattering, and we are developing a simple measurement method for this purpose.
The idea is using a camera to take a picture.
At first we used a laser pointer (Green, 1mW) as a laser beam source and a sapphire sample from Mitaka to check the measurement principle.

Result

ASI camera (ASI224MC) is not suitable for this measurement because of the noise.
The reason can be the insufficient noise reduction in the camera or poor lens we used or both.

A digital camera (Canon EOS Kiss M2) successfully captured the scattered light from the sapphire sample with 61s exposure time and ISO6400.

Detail

Laser source

Model: UC-S1
Wave length: 532nm
Max power: 1mW

ASI camera

Camera: ASI 224MC
Lens: CCTV LENS 2.1mm 3MP
Application: Planetary Imager

Result: we tried several conditions for exposure time and gain, but the scattered light in the sample was not clear because of noise.

Digital camera

Camera: Canon EOS Kiss M2
Lens: EF-M15-45 F3.5-6.3 IS STM
Exposure time: 61s
ISO: 6400
Color temperature: 6000K
F: 16
Focus length: 45mm
Recorde image quality: best (about 15MB/picture)
Output image format: jpeg

Result:
With the above condition, the scattered light was observed. In order to make it clear, we took a picture with the Green laser ON and OFF, then we made a diff image using "imageJ". The diff image shows the internal scattered light clear.

What we learned:
Because the scattered light we want to measure is very faint, we need to avoid environmental light and stray/scattered light generated outside of the sapphire sample interfering with the measurement.

Additional maps.

Here the position and average.
Z=105.3, mean absorption: (61.6 +/- 13.95)
XZ map, mean absorption: (50.1 +/- 18.1)
YZ map, mean absorption: (56.8 +/- 19.1)

[Aritomi, Michael]

After opening the gate valves between input/end and arm, we tried to recover FC. We centered the PR reference and first target, but we could see only the scattered light at second target... We checked that the green beam is roughly center of the gate valve between BS/input with the PR reference.

Note: After the gate valve between BS/input is used as PR reference, don't forget to open the gate valve.

In elog2865, it is reported that the RF amplifier (ZHL-2) for AOM stopped outputting. I checked the output of the RF amplifier and confirmed that it is working. I injected the input of 5.5dBm and the output of the amplifier is 23.4dBm, which is reasonable.

I opened the gate valves between input/end and arm. Before I opened the gate valve between input/arm, the pressure of input and arm were 1.2e-6 mbar and 3.7e-8 mbar, respectively. After I opened the gate valve, both of them became 1e-6 mbar. Before I opened the gate valve between end/arm, the pressure of end and arm were 2e-7 mbar and 1.5e-7 mbar, respectively. After I opened the gate valve between end/arm, the pressure of end and arm became 4e-7 mbar and 5.4e-7 mbar, respectively. 

[Aritomi, Michael]

We started the evacuation of input/end chambers. After evacuation with a rotary pump, we opened small gate valves close to input/end chambers. The current pressure of end chamber and arm are 3.7e-4 mbar and 3.4e-8 mbar, respectively. The current pressure of input chamber and arm are 4.1e-4 mbar and 3.1e-8 mbar, respectively. We will open small gate valves close to arm, and large gate valves between input/end and arm next week.

Yuhang and Michael

We went through the instructions outlined in T2012111, for installing the Debian 10 operating system on the frontend computer for the central station. As described in 2905, this overrides the previous OS installations on the frontend computer. We have a list of errors/ambiguities in the installation procedure which will be sent to Takahiro Yamamoto.

1.1 OS installation
OS was installed in English. Actually I have tried installing Japanese input packages on my Debian 10 workstation in the past and found it extremely poorly documented, so I didn't even bother with that and will just ask Yamamoto-san.

1.2 Additional packages
No issues. We used 1.0.6 version of cdssoft on the ALIGO Caltech repository. Python and Python3 are called separately, just to let it be known (Python version is 2.7.12, Python3 version is 3.7.3)

1.3 OS settings
As mentioned previously, there was an issue with mounting network locations in the /etc/fstb file. We did not follow through with the rest of the instructions since they all seem to be tied to things appearing in the /users/ directory that was created. I feel like there is also some abiguity in the instructions on where these mount directories should be created. I created them inside /etc/, but maybe I should try again just starting at /.

1.4 Bash settings
No issues. Should mention for clarity that .bashrc is a hidden file located in /home/controls.

1.5 miniconda3 installation
We did not follow through with the last step since it seems the conda command doesn't work. In the first step here, a symbolic link is created between /home/controls and a folder that should be created in /users/DGS..., but nothing is present inside the /users/ folder right now.

1.6 ssh settings
No issues

1.7 NTP settings
No issues

1.8 Vinagre setting
No issues

1.9 Guardian setting
We were not sure where the bugfix is supposed to be applied. But upon looking again there is the blue line at the top that indicates which file is changed, so I will check again before sending my questions to Yamamoto-san.

1.10 Camera settings
The procedure says "camera code itself is old version" but I'm not sure what it is referring to or even if anything is installed.

1.11 SDF settings
I guess the blue line at the top indicates which files should be changed. We haven't tried yet.

1.12 z command
No issues

1.13 Zoom setting
No issues

1.14 Skippy-XD installation
We weren't sure what counts as the "1st PC". 

Similar measurement was done in elog2646. In this measurement, the minimum(maximum) CC2 amplitude did not correspond to squeezing(anti-squeezing) quadrature.

Michael and Yuhang

As suggested by Matteo, we took squeezing level and CC2 signal magnitude as a function of CC2 demodulation phase.

They are both normalized, either by shot noise level or CC2 signal without pump.

Fig 1 is squeezing measurements at different CC2 demodulation phase. The phase is changed as linspace(0, 180, 19). It is not understood why sometimes some peaks appear around 5-20kHz.

Fig 2 is squeezing and CC2 signal magnitude as a function of CC2 demodulation phase.

Michael and Yuhang

The optical bench was closed and some squeezing measurements are done with different pump power.

The result is attached. Compared with previous characterization, this measurement has more phase noise of 30mrad. It is not understood why phase noise is increased.

Some pictures of the gluing.

After optimizing homodyne balance, alignment and checking parametric gain, I performed some sqz-asqz measurements. The result is shown in the attached figures. The measurement seems not stable enough to match the theoretical prediction curve in Fig.2. New measurements should be done.

6.19dB squeezing was observed

Following on from the previous report, I ignored the Kerberos prompt since it is not used at NAOJ. When it asked me for the hostname and admin password changing server I just left both options blank. After that, the package installation completed.

I am now working on 1.3 of T2012111 (Installation procedure of new workstations (Debian10)). I'm a bit stuck on mounting the folders as shown. The instruction page doesn't seem to match the format accepted by /etc/fstb (figure 1) so I added some spaces that seem consistent with how the mount command works - i.e. device directory filesystem options 0 0. My guess is that the device 10.68.10.100:/opt/rtcds is being mounted to the director /etc/opt/rtcds but maybe I should understand which device this actually is. I get the error shown in figure 2, so something is wrong here.

Takahashi-san glued the magnet of input mirror and released the input/end mirrors. We will start the evacuation of input/end chambers tomorrow.

MZ offset: 4.3

Pump power: 29.4mW

Parametric gain: 504/51.6 = 9.8 (Done by scanning OPO and check BAB transmission. The gain is the ratio of Airy peaks for BAB without pump and BAB with 29.4mW pump. The pump phase was continuously scanned.)

This parametric gain is consistent with old measurement.

During the weekend the first two XY maps.

The maps are attached.

Here the position and average.
Z= 71.075 (center), mean absorption: (42.4 +/- 16.8)
Z= 36.85, mean absorption: (40.8 +/- 9.7)

Michael and Yuhang

After pointing out the place where issue appears in elog2916, we followed the suggestion from Pierre and checked the resistor connect through J10.

1. We disconnected three pins of J10.

2. We put resistor meter between these pins. It was found that the resistor between blue and black wires becomes infinity after touching wires.

3. We peeled off the heat shrinkable tubing for black wire at the place of connecting to potentiometer. But no issue was found.

4. We peeled off the heat shrinkable tubing for blue wire at the place of connecting to potentiometer. We found the inner steel part of the wire has broken.

5. We soldered the broken blue wire.

6. We tested OPO locking servo and found it worked again. But we will see if it always work in the next week or so to confirm its recovery.