We reconstructed the history of the cavity transmission. 

At the beginning of June we locked again the FC, after more than 6 months (entry #1385), at that time the injected power was 34 mW and the PD in transmission read a bit more than 4 V (which corresponds to ~ 7000 counts). (entry #1404)

After then we decided to reduce the injected power to 12 mW and the transmitted power was reduced accordingly ( ~2300 counts). (Not reported on logbook!)

Recently Yuhang found that by reducing the gain of the locking servo the transmission could reach 5000 counts. (entry #1644) We didn't remeasure the power at that time. 

Few days ago, after some realignment work on the bench (entry #1647), we found out the transmission was back to 2200 counts. (entry #1655). In that occasion (even if it is not reported on the logbook)  the AOM was also realigned and this brought to a large increasing of the power injected into FC (30 mW). The power was reduced by tweaking the half waveplate after the main laser.

Last Thursday (26/09) we spent half a day to carefully check and asses the alignment level (which is reported in entry #1674). The conclusion of that work is that the cavity seems well aligned. 

It seems to me that this mysterious change in the cavity transmission is connected to the change in the injected power, which is largely affected by AOM alignment condition. Also, the only straightforward reason why a reduction in the locking servo gain should bring an increasing of the transmitted power is that the loop was oscillating, and this is likely to be due to an increased power.

Anyway, I think that from now on we should better monitor (and report on the logbook) the level of power injected into FC and the cavity transmission.

Yuefan, Eleonora and Yuhang

For DC normalization, we found out why we couldn't connect. The reason is there is a range of output for photodiode DC sum, it was not able to fit in this range. But now it is fine after some adjustment done by Eleonora(mainly just adding a -1 factor to the DC sum).

For RF signal, in the beginning, we just checked the PDH signal from each quarter of quadrant. But we couldn't find any signal. Then we did some further check, including:

1. quadrant power supply(power and bias were switched on)

2. quadrant DC signal(to make sure the beam is hitting on PD, we checked this signal and it was fine)

3. LO sent to demodulator(it was fine)

4. RF signal from quadrant(before demodulation, checked directly from spectrum analyzer) we found this RF signal was only around -40dBm

5. We amplified this RF signal by 18dBm then we could see a small signal. 

6. Check demodulator for AA system. we took the RF signal from Qubig PD(used to lock filter cavity). The green power on Qubig PD is around 200uW for now. Then we just replace of demodulator for this signal. And then compare the demodulated signal(PDH signal). We got PDH signal pk-pk as 50mV. While the signal we got from another demodulator is ~500mV. So it seems this demodulator is not as good as what we were using.

7.  By using the same demodulator, we took RF signal from Qubig PD and quadrant and got PDH signal(see attached figure 1, blue curve is signal from Qubig, yellow curve is signal from quadrant). Note that green power on quadrant is ~1mW while green power on Qubig PD is ~200uW. So is seems quadrant is not so sensitive to green light or it doesn't have enough bandwidth. 

8. We also tried to move the beam totally on one quarter of quadrant. See attached figure 2. Roughly, the PDH signal was increased by a factor of 4.

9. We also tried to check the other quadrant and the other demodulation board. I shows similar result. (Attached figure 3 shows PDH signal when beam is centered on quadrant, attached figure 4 shows PDH signal when the beam is mainly located in the first quarter of quadrant)

Above is the check we did on last Friday. Since quadrant worked well in NIKHEF, we will check again these days.

Yuhang and Yuefan

As we discussed in the last filter cavity meeting, we may have a beam cut issue in AOM/EOM/iris. We checked again these optics by putting a steering mirror/lens after it and look at beam shape far.

For the cut issue of the iris, we found a better point where the higher-order diffraction beams are more separated. Attached pictures 1 and 2 show the position where we were putting the iris and we are putting it now. This more separated point is located just after the first lens after AOM. The new set-up is shown in the attached picture 3.

We also checked the beam cut issue of Faraday isolator and AOM. The checking point is just before the injection to PR chamber. We found out the main cut comes from AOM. We could see from the attached picture 4 that the most powerful center part could go through AOM/FI but not centered. After twinking a bit the position of AOM, we made the beam go through AOM from the center part(as shown in the attached picture 5). 

So although the aperature of AOM and FI are not ideal, we confirmed that the most powerful center part is not cut.

Matteo, Simon

attached are the first results of our experimental try to map also the polarization homogeneity of fused-silica substrates.

In total, I think that we can say that the homogeneity is almost perfect. The fluctuations in the map are probably due to fluctuations risen from the input polarization which cannot be filtered out. That would explain at least the stripe-like pattern in the map along the way of the measurement.

Anyway, the next steps will be to convert the bench back to a PCI for taking an absorption map of the AQ2 sample.

Matteo, Simon

attached, please find the maps and the distribution analysis of the polarization measurements on the Tama-sized Sapphire samples from Shinkosha.
The maps are taken with different polaization angles indicated in each figure, with 0 degrees being pure P-polarization.

Obviously, the birefringence effect is relatively homogeneous and divided into three parts with different offsets. The offsets itself may have their origin basically in the non-zero incident angle of the pump-beam.

with help of Tomaru-san, Namai-san, Ueda-san from KEK, and Sato-san

We did renovation work of cryostat.
First, we detached the optical breadboard from cold head.
Then we moved cryostat chamber about 80cm to extract cold head, and now that we can implement viewport windows.

Next step is to implement viewports to 80K shield, and install modified 4K shield to the chamber.

[note]
The cryostat is located different position temporally.
Cold head was attached to optical table using apiezon grease and indium.

Yuhang and Yuefan

While designing the telescope of the automatic alignment, we had one question is that how close we could  put the galvo from the quadrant to have decent range. To answer this question, we should at least have the idea how much the reflection beam could move.

So in order to do this, we tried to do a long term monitor of the DC signal of the quadrant without closing the galvo loop.

The problem is that the filter cavity could not stay lock during the time we record the DC signal, so we only could have data of less than one hour.

• In the first try(fig7), during around 1h of locking, the recombined DC pitch signal has a maximum movement of 0.1, first we tried if we manually moved the beam in pitch to have 0.1 movement, the galvo was totally able to bring the beam back to the center.(fig6) But then we thought maybe it is better to do a rough calibaration to have an idea about how much the beam moved on the quadrant, with the chaning we saw on the dataviewer.

1. Record the DC pitch and yaw signal with Dataviewer (fig 1), we could see that for pitch and yaw, the reading are both around 0.2.
2. For the pitch, we record the beam height at that moment by put a ruler in front of the quadrant. The position of the ruler also recorded. (fig 2)
3. Move the screw of the galvo to move the beam around 1mm in pitch. Actually the movement done by the screw has a coupling between pitch and yaw, but because we are only checking pitch this time, we didn't care about the yaw value changing.
4. Checking the difference in the signal in Dataviewer again. (fig 3) Pitch moved to 0.1
5. Starting from the position of last step, where yaw reading is -0.14, Repeat the steps above on yaw (fig 4), then it changes -0.22. (fig 5)

         So the calibration results are like below,        

• For the second(fig8) and third try(fig9), we compared the DC signals with the cavity green power tranmission. There were several short time unlock and relock, so we could see that everytime the cavity relock, it seems lock again at different position. And while the alignment of the cavity is drifting away, the tranmission power is lower and lower, and finally cavity unlock, we could see the DC signals also goes to zero slowly, from this two fact we had some guesses.

1. Since everytime we tried to overlap the input and reflection green by only checking the viewport, the beam may not always go to the same direction, which results in the reflection beam height changing we obsered yesterday and also in the past weeks. But of course closing the galvo loop could make the situation better, but the galvo mirror is quite small, if the beam pointing change is too much, the beam could totally miss the galvo. So probably we should do another long term monitor with galvo loop close.
2. What we expected the DC signal should be like when the cavity slowly misalign is that in one or both direction the signal should firstly increase and then a sudden drop to zero when the beam is totally off the quadrant. But what we saw is quite different. After discussing with Matteo and Eleonora, we found out this could  be caused by the fact we didn't normalise the pitch and yaw signal with the total power, so the slowly drop we saw is because the power reduction when the beam is moving out of the quadrant.

Simon

I finished the spectroscopic measurements of the colored Sapphire samples from 1600 to 300 nm wavelengths (respective figures are attached).

I measured the transmittance (T) and the reflectance (R) for both blue and green sapphire samples, and calculated the absorbance by 1-R-T = A.
In short, the green samples have the lowest absorbance at 1064 nm wavelength (A = 0.05), while the absorbance of the blue ones ranges from 0.143 to 0.158.

This is the scheme of the AA telescope now

According to the measurement yesterday, we made the new design for the telescope, this time seems we could use one telescope to get gouy phase around 90 degree difference in two axis of the beam.

Firstly, we calculated the reflection beam waist position according to the measuremnt. (colors refer to different curves in last entry)

The size of the waist in two direction seems much closer than all the measurement we did before.

Then because the space limitation, we fixed the two quadrant position as before, which is around 1.9m and 2.3m from the FI, and both the positions have some margin of 5cm.

The beam we used to first determine the lens is the orange one, because it has smaller divergence than the other one. So with the orange as the initial beam, we tried to put different focal length, and move the lens position to have the beam waist around 2.1m which is in the middle of the two quadrants. By checking the gouy phase difference, we could know that we need longer focal length or shorter.

The final results we got is to put a lens with focal length of 1m, and 0.82m from the FI. 

Then the quadrants information are as below

Then use the distance to calculate the gouy phase of the other beam axis (blue). By slightly adjusting the lens position and the quadrants position, we could get the gouy phase difference around 90 degree for both axises. In this case the lens position is 0.75m.

Yuefan and Yuhang

We did beam height measurement on the bench and the result is reported here. However, we could see that the beam is still not flat at the height of 76mm. So we adjusted beam height again. After the adjustment, the beam height after AOM and until the last steering mirror is maintained at the level of 76mm.

After this activity, we aligned the filter cavity again. And the reflected green beam seems to be better. We think at this point, we should turn the work to design and implement the final telescope for AA telescope. We also measured the reflected beam. The result is attached and Yuefan will use this result to design the new AA telescope.

Note that if reflectivity of HR coating of PPKTP is 99.995%, we can explain 10% loss from OPO in loss and phase noise measurement.

I discussed with Shoda-san and we looked better into the issue about python scritp launching from MEDM (see entry #1651). We couldn' t solve the problem.

So I decided that the best solution is to use only a bash script, as for the moment we simply need to change loops gains.

I prepared two bash scripts to open and close the loops repectively ( open_loops.sh, close_loops.sh), just using "caput" command. The scripts are located in /home/controls/FDSscripts.

I modified the shell command MEDM to launch them and now we can close and open the loops by pressing the button, even when we open MEDM from desktop icon.

We found the filter cavity green transmission DC is changed to ~2400 at maximum after last week's work. Although we tried to change servo gain and aligned filter cavity as well as possible, we couldn't bring it back to 5000.

Yuefan and Yuhang

We found the beam height of the green beam between SHG and filter cavity has very large beam height change after the work of beam center on each lens.

We measured the beam height at several points. (The beam height at Faraday isolator is a very important parameter. We may send a tilted beam if the beam height is wrong. Also, Astigmatism beam could be cut if the beam is not center in this FI.)

Note: Lens1~5 is the lens in the sequance from SHG to PR chamber. (Details are also shown in the attached figure 1)

We also checked the beam shape. Attached picture 2: beam shape after AOM. Attached picture 3: beam shape before injection to PR chamber (magnified by a 50mm lens). Attached picture 4: beam shape seen at the end camera. Attached picture 5: reflected beam shape seen before FI.

We measured the beam parameter and fit with basic Gaussian beam function before and after today's work. The comparison is attached in the last two figures.

In the past days I spent some time to modify again the damping loops.

The main reason was that the loops that I have implemented before (see entry #1641) had the first UFG at too low frequency (<100 mHz) and it took to much time to change DC offset to align the cavity. Also I'm not sure if the signal in the 10-100 mHz corresponds to a real motion of the suspensions or is oplev noise.

In the attached PDF I report the new filter performances. I tried to avoid any overshoot but I think they can be further improved.

Some remarks:

- I observed that sometimes an eccess of noise shows up in the whole BS YAW spectrum when closing the BS PITCH loop. I have slightly decreased the BS pitch gain and it has disappeared.

- The pitch peaks just above 10 Hz are quite hard to damp as they are coming from the stack resonance. (See attached a plot of the spectrum of an oplev for a mirror placed directly on the stack.)

Up to now, all the loops are closed one by one by, manually changing the gain from 0 to 1 in the MEDM interface.

Since this is inconvinient especially if we will need to close alignment loops, I prepared a python script to close (and open) the damping loop just by clicking on a button on the MEDM screen. (see attached picture) 

The scripts, which use EZCA module to change values of epics channels, are located in /home/controls/FDSscripts.

A strange issue we found (while checking with Shoda-san) is that the script doesn't work if I open MEDM interface using the desktop icon but only if I open it from terminal.

I contacted Yamamoto-san and he suggested that it can be due python environmental variables and he suggested to try with a bash script that launch a python script. I tried this solution but also in this case it only works if I open MEDM from terminal. I will investigate more.

For the moment, if we want to use the button to open and close all the damping loops in one click please open MEDM from terminal (just open a terminal and type "sitemap").

On Friday afternoon we realizied that we were not able to use diaggui software to perform for fourier analysis anymore. After launching any measurements we got the error message: " test timed out".

I contacted Yamamoto-san in KAGRA and he explained me that this problem is caused by a difference between a system time of the client workstation (desktop1) and a timing signal on the standalone. The system time of the client workstation is synchronized to a correct time via NTP (probably within 1 second accuracy). But the accuracy of a timing signal on the standalone depends on the 65kHz-clock signal from a function generator. So the time on the standalone deviate from the correct time by the long term operation because the function generator is not so accurate for the real-time system.

Yuefan and Yuhang

This work is inspired by astigmatism we had for green reflection from the filter cavity. Astigmatism can be generated by misalignment or refractive error.

If it is because of misalignment, this can be improved by doing the alignment. If it is because of refractive error, we can move a bit beam on the lens to correct the refractive error.

The filter cavity reflection always has the same direction astigmatism no matter what is the alignment condition. So we think most of astigmatism comes from the refractive error.

We always check the beam shape on the first and second iris(along the vacuum tube), but we should notice that the beam shape is difficult to tell. The reason is that we look at the iris while the beam hits on the iris with angle. Also, we checked the beam we see on the filter cavity transmission camera when we misalign the input mirror. Always this filter cavity transmission has astigmatism(like filter cavity reflection). So we think the astigmatism is mainly because of refractive error.

To solve this problem, we tried to center beam on every lens. We have five lenses between SHG and FC. However, we found most of them are not centered in the lens. But this is based on a clear phenomenon that SHG reflection was seen to have some fringes if we reflect it far and large enough to see clearly. The elongation is the diffraction fringes at one side of the SHG hole and caused by the clipping of the SHG hole side. We know that the beam direction of SHG reflection is defined by the axis of SHG crystal and the SHG in-coupling mirror. We need to adjust the crystal and in-coupling mirror to have a better axis.

Although we centered beam on each lens, the filter cavity reflection still has astigmatism. Maybe astigmatism comes from SHG, and we should assemble it again. Maybe astigmatism happens inside the vacuum chamber, and we need to consider how to improve this situation.

Since we decided to have gouy phase 90 degree different on two quadrants instead of exactly 0 and 90 degree, I calculated again the telescope based on the beam waist reported in last entry.

In the case of two different axis has different beam waist size, I made two different telescopes. 

The telscope was designed to have more or less same quadrant position as before, and to have two quadrants one before and one after the beam waist.

1. beam waist of 711um

500mm lens 1.5m from the Faraday Isolator including the height increasing in the periscope.

2. beam waist of 1274um

1m lens 0.975m from the Faraday Isolator including the height increasing in the periscope.

We will check the green injection path these days, if we could have a more round beam in the reflection, we can measure the beam again.