Since the two magnets of end mirror fell down, this measurement does not make sense. I brought back the rotation angle for END oplev to 0 deg.

[Takahashi, Aritomi]

Since there is a new 6.35Hz peak in END oplev spectra as reported in elog2808, we opened the end chamber and checked the suspension. We found that two magnets of end mirror fell down! This is the reason why the Z correction does not work recently. Takahashi-san will glue the magnets next Monday.

Takahashi-san glued the magnet to the input mirror and closed the input chamber. Since it takes one day for the glue to be fixed, we will evacuate the input chamber after tomorrow.

The attached figure shows the CCFC error signal on 20210622 and 20211227. The elogs on 20210622 and 20211227 are elog2597 and elog2770, respectively.

I checked the first measurement of the shinkosha 7 with s polarization at the input,

As shown in figure 1 there are some vertical stripes present in the measurement.

I tuned both the hwp and qwp in the injection and found out that when injecting s polarization, the p polarization readout is minimized for hwp = 0.1 deg instead of previously 1.3 deg.

Then, I also added beam dump to catch the PSD reflections and a small wall to isolate the imaging unit from the injection part.

The beam dump were removed because they are supposed to be used for visible light...

In any case the good news is that we can recognize similar patterns to the previous measurement with tilted beam.

Finally I restarted the s polarization measurement.

Recently, a 6.35Hz peak is visible in END pitch and yaw oplev spectra (Fig 1). For END oplev diagonalization, I changed the rotation angle for END oplev so that the 6.35Hz peak disappears in oplev yaw spectrum, assuming that the 6.35Hz peak is a natural pitch peak. I changed the oplev rotation angle from 0 deg to 1 deg. The END oplev spectra before/after this rotation are shown in Fig 2 (red, blue: after rotation, green, brown: before rotation).

To check END actuation, I put an excitation (12 Hz, amplitude 1000, sine wave) in END length (K1:FDS-END_LEN_ex2). The 12 Hz peak is visible in both pitch and yaw oplev spectra (Fig 3). The 12 Hz peak height was 13.5 and 20.8 for pitch and yaw, respectively. The current EUL2COIL matrix is shown in Fig 4. This matrix might not be optimized for L2P, L2Y coupling.

In the case of a rigid cavity, we can estimate internal losses without locking. We just need to scan cavity and look for the reflection signal as shows in Fig.1.

1. We scan cavity and know that about gamma = 5% power is not coupled to cavity for TEM00 resonance.

2. We take the reflected power when cavity is on resonance for TEM00 (33.2mV) and off resonance (36.6mV)

3. We get Rcav = ((33.2/36.6)-gamma)/(1-gamma)

4. RTL = T1*(1-Rcav)/2/(1+Rcav) = 0.206%, here T1 = 0.08. Note: the polishing/coating company (LASEROPTIK) specifies AR coating reflectivity is less than 0.1%.

5. escape efficiency is T1/(T1+L)

In this way, we get escape efficiency of 97.5% for the new OPO.

Katsuki, Marc

We measured the distance between the blade at z = 25 mm and the last steering mirror.

From this distance and knowing the 2.5deg angle of the beam we found that the beam is hitting this last steering mirror at y = 134.4 mm (in the frame of the translation stage).

We placed the blade at this position and z = 110 mm and tweaked the steering mirror to get half of the power transmitted.

After little back and forth, we measured the horizontal angle of incidence of 0.024 deg and the vertical one to be 0.015 deg.

Because this is more than 10 times smaller than KAGRA requirement on the C-axis angle wrt to the surface normal we can consider that we are now in normal incidence.

We realigned the imaging unit , did the polarization calibration and started the measurement with shinkosha 7 (new center at x = 398 mm and y = 162 mm).

However we are not sure if this substrate has a wedge or not which means that we might see the fringes that Manuel observed...

Yuhang and Michael

We rearranged the OPO optical layout to use proper photodetectors as mentioned by Yuhang. The layout with breadboard spacing (not to scale) is shown in figure 1. A photo of the setup past L1 is shown in figure 2.

The following components are used. Some of them I'm not sure though:

Laser: Lightwave laser
WP1: ?
WP2: ?
KPX1: KPX094AR.33
BS1: ?
KPX2: KPX094AR.33
KPC: KPC034AR.33
L1: ?
---
Equipment placed for OPO setup
FI: Thorlabs IO-5-1064-VLP
HWP_FI: Half wave plate
EOM: New Focus 4003 Resonant EOM
AOM: AA Optoelectronic MT110 IR 27
L2: f = 40mm
SM1: Steering mirror 1
BS2: 50/50
SM2: Steering mirror 2
L3: f = 75mm
SM3: Steering mirror 3
BS3: ?
L4: f = 30mm
PD_R1: Thorlabs PDA05CF2
L5: f = 50mm
PD_R2: Thorlabs DET10N/M InGaAs biased
IN: OPO input mirror
OPO: OPO assembly block
BS4: ?
CCD: Camera
L6: ?
PD_T: Thorlabs PDA36A
 

In order to avoid polarization coupling when measuring birefringence with a tilted IR beam (the pump beam) we plan to take future measurement with the pump beam at normal incidence.

Yesterday I removed the last 2 steering optics on the 1310 nm beam path to have enough space to add 2 steering mirrors on the pump beam path.

I also installed a new power meter head (S146C) that has enough range to be used for both absorption and birefringence measurements.

I started by aligning the vertical direction using both a marker on a IR viewing card and iris.

Note that the injection table height is 28 cm with beam height of 5.5 cm and imaging table height is 24.5 cm (so beam height there should be 9cm).

Then I started to align the horizontal direction but it is slightly more tricky as the injection and imaging tables are not aligned.

I installed the razor blade and measured 1.5 deg incidence angle in this plane.

Today we will check the relative position of the blade with respect to the last steering mirror on the pump beam path to fine tune this incidence angle.

Michael and Yuhang

Last week, we replaced two QUBIG PDs for FC GR lock. An RF output only PD is currently used for FC GR lock, which was found to have flat spectrum across tens of MHz as Fig.2. The QUBIG PD with DC and RF outputs is taken out and used for new OPO reflection, which has spectrum as Fig.1. Comparing Fig.1 and Fig.2, we can also find that the RF only PD has a better SNR. However, after a test of DC RF QUBIG PD for new OPO, we found it cannot read phase modulation taken from the reflection of OPO. Although from its specification, PD bandwidth (1-100MHz) is enough to detect 40MHz signal, we decide to not use it since we couldn't find by checking either spectrum directly or demodulated PDH signal. On the other hand, we will use it for monitoring DC response. Considering a BW of 100MHz, which has a very small rise time of 3.5ns.

A PDA05CF2 will be used for acquiring PDH signal for locking. This PD will be placed in the reflection of OPO. It has been proved to be able to provide large enough error signal.

We will use PDA36A-EC in the OPO transmission. If 0dB is enough to measure transmission signal, we can have 10MHz bandwidth, which gives 35ns rise time. However, NIR may have larger rise time since this PD is made of Si. According to Isogai paper, transmission doesn't need fast response. So maybe PDA36A-EC is enough. We will see this later. On the other hand, I took some dark noise measurement, which is summarized in Fig.3. The dB value in the legend shows different chosen gain of this PD. This measurement shows much less dark noise indicated in the specification of this PD. 0dB should be 300uV, 20dB should be 250uV, 40dB should be 340uV, 60dB should be 800uV. One issue is that the RMS I got is only from 10Hz to 3.2kHz. Higher frequency noise must be integrated but I didn't measure. But anyway, we can see that with a factor of 10 gain increase, the noise is not increased by the same factor.

[Takahashi, Aritomi, Yuhang]

Today we opened the input chamber and checked the suspension of input mirror. Because of misunderstanding, we vented not only the input chamber, but also the PR/BS chambers.

Takahashi-san found that one magnet of the input mirror fell down. According to Takahashi-san, the magnet could touch the coil holder and the input mirror could be tilted because of it. In this situation, we tried to recover the tilt by picomotor and then the magnet could fall down. Takahashi-san will fix the magnet and close the input chamber next Monday.

After Takahashi-san adjusted the suspension, we could move the picomotor for input mirror and the green reflection overlaps with the injection now.

Finally, we started to evacuate the PR/BS chambers.

Since the oplev beam reflected by the input mirror was very high as reported in elog2798, I tweaked a steering mirror just after the oplev laser source for input mirror (Fig. 1). The oplev sum for input gets around 5000.

Fig. 2 shows the oplev spectrum of input mirror (top right). There are large peaks at 2 Hz and 3.4 Hz. We need to open the input chamber and check the suspension.

[Aritomi, Yuhang, Michael]

We found that the green reflection from input mirror is very high and hits on the upper edge of gate valve between BS/input chambers. The oplev light for input mirror is also very high as shown in the attached picture. We tried to move input mirror with picomotor, but the input mirror did not move. We will align oplev and measure the input oplev spectrum tomorrow.

I connected the END YAW to the motor A of driver although I don't know the motor A is working or not.

I took a function generator and an oscilloscope from filter cavity clean room to modulate EOM and acquire data in ATC.

The RF PD is moved to OPO reflection to acquire PDH signal. A minicircuit zx05-1l-s+ is used for RF signal demodulation. 10dBm 40MHz signal is sent to New focus 4003 EOM to get a phase modulation around 0.07~0.21rad. 3dBm signal is used as LO. We got error signal as attached figure. Some glitches are found and we will try to understand it better tomorrow.

One possible reason is that I should put a DC block between RF PD and mixer. Note that demodulation phase is not optimized yet.

To measure effectively the internal optical losses of OPO, we need to inject laser from the in-coupling mirror side (elog2784).

Therefore, I flipped OPO and moved OPO closer to 75mm lens as described in elog2790. However, I found that moving 54mm is far too much, whose reason is not understood yet. After moving OPO farther from 75mm lens, the mode matching situation is better. However, I found that the beam is going into the in-coupling mirror but close to the edge of its hole. So I loosed in-coupling mirror and moved it vertically down. After this, I found it diffcult to recover alignment at the first glance.

However, I realized that it is actually very easy to re-align OPO using transmission camera and injection/reflection beam overlaping (without using periscope/translation stage). The method is bascially the same with the procedure provided in elog2783 apart from different beam parameters and almost not seperating the alignment of crystal and in-coupling mirror but aligning them as almost a whole. And I found it is possible to choose a personal preferred position of in-coupling mirror and just fix it with the crystal side of OPO. Then align them as a whole.

After this easier alignment procudure, checking by eye, I found the laser is far enough from the edge of OPO two sides holes. Then I aligned OPO until mode matching level is around 95%. Now OPO is ready for internal optical losses measurement.

[Aritomi, Yuhang]

After the evacuation of PR/BS chambers, we aligned PR/BS picomotors to center the PR reference and the first target. Then we checked the second target. Although we scanned BS alignment a lot, we could see only scattered light on the second target.

Instead of the nominal PR reference, we used the green beam position in the gate valve between BS/input chambers as PR reference (Fig.1). We centered the green beam in the gate valve between BS/input chambers by PR picomotor and centered the green beam at first target by BS picomotor. Then we found the green beam in the second target. Since the green beam position in the nominal PR reference changed, we made a new PR reference as Fig.2.

We firstly used Rotary Vane Vacuum Pump (Alcatel 2100A) to evacuate. The speed is controlled to be less than 1mbar/s to avoid large pressure applied on in-vacuum components. Its ultimate achievable pressure is 0.03mbar.

Since the turbo pump (STP1003) requires outlet pressure smaller than 0.1torr (0.13mbar), we waited for rotary pump to evacuate PR/BS chambers until pressure reached 0.1mbar.

The small rotary pump and turbo pump were kept on with gate valve closed between turbo pump and PR/BS chambers. So we just opened the gate valve after PR/BS chambers pressure is lower than 0.1mbar.

We will keep turbo pump on until the PR/BS chambers pressure reaches 1e-6 mbar. Then we will open the gate valve between PR/BS chambers and input mirror chamber.

[Takahashi, Aritomi, Yuhang]

Since the BS picomotor did not move, we opened the BS chamber and checked the BS picomotor. The BS picomotors for pitch and yaw were connected to the motors A and B in driver 1, respectively. However, we found that the motors A and B in driver 1 were broken. We decided to use the motor C in driver 1 for both BS pitch and yaw temporarily.

We noticed that the BS picomotor did not move in yaw direction although we heard the sound of the picomotor. Takahashi-san found that it was because the BS was touching the stopper of the mirror. After Takahashi-san fixed it, the BS picomotor could move in yaw.

After we adjusted the BS picomotor, the green beam is hitting on center of the first target. Finally, we started to evacuate the PR/BS chambers.