[Enomoto, Yuhang, Matteo, Eleonora]

In the past days we observed an oscillation in the SHG lock at about 20 kHz.  This is likely to be due to the change of the loop shape as we removed both a low pass fiter and a high pass filter in the control servo.

Before finding out that the LP filter in the High Voltage Piezo Driver could be disabled (see entry #1016), Matteo had designed a HP filter with the goal of componsating its effect and allow for a larger loop bandwidth.

The TFs of this two filters and their combination are shown in the attached picture (taken from entry #585).

As suggested also by Raffaele, the LP and the HP where non exactly compensating above 10 kHz and this could avoid the oscillation at 20 kHz (likely due to a Piezo resonance) that we see when we remove both of them. 

As a temporary solution we decided to keep both the filters on. We will study a better filter to be implement in a new version of the control board.

[Enomoto, Yuhang, Eleonora]

Yesteday, just after openening the gate valve between the pipe and the end chamber we were able to lock the filter cavity again (after more than 2 months!) 

The alignment was recovered in the past days and we were able to see some dim flashes even with the closed gate valve.

The lock acqusition was smooth with the 1/f filter shape but we got an oscillation at 155 kHz when swhiching to the 1/f^4.

The gain setting was:  PIEZO GAIN = 5.   INPUT ATTENUATION = 9.5 (almost maximum)

The oscillation is likely to be caused by a piezo resonance. (See main laser piezo resonance characterization in entry #859). We couldn't get rid of the oscillation by changing the piezo gain but we noticed that for some gain values, the oscillation is at lower frequancy, about 75 kHz (probably another piezo resonance.)

In the end we found a good setting of the gains (we had to change the input amplification also): PIEZO GAIN = 4.   INPUT ATTENUATION = 1.6 (much lower than before)

In this configuration the UGF is about 16 kHz with phase margin 62 deg (see pic 1). By eye, the spectrum of the error signal seems even lower than before but we need to check again the calibration to confirm this.

Note that:

- the locking photodiode is now the qubig one without DC (as the one with DC is used for OPO) but the gain of the RF channel shoud be the same.

- the change of the PIEZO GAIN affects the piezo dynamics, as reported by Pierre in entry #747. 

Participaint: Enomoto, Eleonora and Yuhang

Today we did the filter cavity green reflection characterization again after achieved its lock. Here I want to put some information we found for the green reflection from filter cavity.

Firstly let's review the set up. The configuration is we put a BS for filter cavity green reflection extracted from Faraday isolator. Small part of green goes to FC locking. Another part is used for our characterization and it is sent to a good height by using a periscope. Then let's look at some information:

1. The reflection seems to be cutted by something if you look at our green at a decent distance. See attached figure 1. It is taken several months ago by me and Marc. As you can see, there is a very clear boundary around the green light. Although the brightest part is smaller than this boundary, it is essential to know where it is cutted. And we confirmed that it is cutted by one side of Faraday isolator whose cover is not dismounted.(See attached figure 2)

2. The reflected beam shape is quite bad. I think you have already noticed that in the attached figure 1, the beam seems to be flatted by astigmatism. This effect becomes quite obvious if you look at the beam detected by the beam profiler. See attached figure 3, the beam shape is really horizental oriented ellipse. However, the axises of our beam profiler detection is accidently aligned to two direction that have the same dimension. That means we cannot have a numerical estimation of this astigmatism by chance. But this brings also an advantage, it is roughly the average of the long axis and short axis of this ellipse. So it is reasonable to continue the measurement even in this case.

3. We did the measurement and fit of this beam directly although it is quite collimated. Besides, this beam is quite unstable. So you can see the points we took are quite scattered. Then we did the fit and the result is shown in attached figure 4. As you can see, the beam waist size is quite different in these two cases(900 and 500 um with an error of roughly 10 percent). Also the waist position has a quite large difference(-300 and -260 cm with an error of roughly 10 percent). 

4. The last method we tried is to put a lens and do beam characterization after lens. By this characterization result, we propagate back to the beam before lens by using JaMmt and ABCD matix. However, this time we set up the average inside beam profiler as 20 while last time it is 5. Now the number we can read becomes more stable. Then we took this more stable data and did the fit. The result is shown in attached figure 5. 

The lens we used is 150mm, and the measured result is quite reasonble now. As you can see in figure 5, z0 is both roughly 150mm. Then we used JaMmt propagate the beam back, the result is shown in attached figure 6. From this result, we can see the beam waist size should be 957um, while its position is after the lens about 2m. This means the waist is located not inside the beam going to filter cavity. Besides, there maybe a measurement shift of several millimeters of waist position of the beam after lens. And this can influence quite a lot the waist position. Also the focal lens of our 150mm lens can also be smaller or larger than this nominal 150mm value. Then influence this waist position quite a lot. But anyway, the waist position is not so important for a collamited beam. So it should be fine.

We also verify this result by using ABCD matrix. The method I used is taking the q factor of gaussian beam. Then the free space propagation is just an addation of this distance to this q. The lens is just a modification of the invert of q by 1/f. Since we use a converging lens, the f is positive. We used the detection beam wait size and waist position to reconstruct the original waist size and its position. Then make real and imaginary part equal with each other to have two equations and solve two unkown valuables. The result is shown in attached figure 7. The averange of these two result is 1.2m before the lens and wasit size is 1008um. It complies with the result of JaMmt, and this is reasonable because they are using the same principle. (Actually I want to propagate the error of fit result, but the python code of this error propagation cannot deal with imaginary number and solve equation, so I give up the error estimate in the end.)

I used OD filters to reduce the probe power.
I avoided using polarizing optics to limit the polarization fluctuations effects on the intensity.
I made the beam pass through the pinhole in 2 positions.

I used the surface reference to maximize the signal and made a scan. See screenshot 1.

The beam was not exactly passing through the center of the lens, so I think it made it astigmatic on the PD. 

I started over the alignment of the imaging unit. I moved the sphere with respect to the lens to find the sharp image of the blade.

In order to not saturate the PD, I changed the laser current from 300mA to  200mA, and the OD 2 to OD3. 

I found a larger signal and I maximized it moving the whole imaging unit, I made a scan for 3 positions of the imaging unit 20,25,30mm, and the maximum is at 30mm. See last 3 screenshots.

Participiant: Enomoto, Yuhang

Since the mirror to do the coherent control is necessary for the measurement of power threshold of OPO gain amplification. We did the mirror replacement the day before yesterday. Also recovered the alignment of green mode cleaner.

Since I found we used to use high voltage deriver not in a proper way. The reason is demonstrated in the e-log entry before. Since this information is crucial for the filter design, I measured these transfer functions for which the high voltage drivers we didn't use properly. The result is uploaded to our wiki page with the name as listed below. However, I don't know if this coherence level is enough for our user.

For the data, please go to the following page to download. https://gwpo.nao.ac.jp/wiki/FilterCavity/Transfer%20Function

Participiant: Enomoto, Eleonora and Yuhang

The week before last week we found the problem of high voltage driver, and this is also the reason why we found there is a low pass filter inside our high voltage driver. So we decided to use it in a proper way, which means switch off the sensor-in switch. However, we found the lock of SHG becomes unstable after did that. So we did the characterization of open loop transfer function again.

We found the unity gain frequency of previous set-up is 795Hz while the phase margin is 89 degree which is fine.

However, the unity gain frequency of revised set-up is 376Hz while the phase margin is 66 degree which is worse. Besides, there is also a peak going above 20k Hz and brings oscillation.

Since Pier is developing servo in APC and we also want to use high voltage driver in a proper way, so we need to measure a list of open-loop-transfer-function again with the correct set-up of high voltage driver.

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Participaint: Eleonora and Yuhang

Besides, the remote control of 290m targrt is recovered based on yuefan's elog. However, its control requires the change of IP address and netmask. For the convience of remote control, we kept that computer inside our clean booth and we used another ether cable to achieve remote control.

[Takahashi, Matteo, Eleonora, Yuhang, Enomoto]

With the help of Takahashi-san, we have opened the vacuum chamber in the end room (only the top part).

[Note that even if the venting was started last Friday the vacuum level was still not low enough so we had to inject some air before opening the chamber.]

The goal was to recover the alignment of the end mirror and solve the issue with the yaw picomotor reported in the entry #954.

By visual ispection, we could confirmed that the picomotor was stuck, as it reached the end of its range, and the intermediate mass was touching the damping magnets.

We have manually put back the yaw picomotor at half range and moved the suspension with the help of a traslation stage (pic1) and also by hand, to make the reflected beam superpose on the incoiming one. The superposition was checked by letting the incoming beam pass through the hole of the second target and check the reflection from the end mirror on the rear side of the second target.

The chamber has been closed and the punping down has been restarted. The gate valve between the chamber and the pipe is still close. According to Takahashi-san, the vacuum level to be reached before opening it is 2e-7 Torr.

Even if the alignment has been recovered we were not able to see any flashes, probably beacuse the beam distortion induced by the gate valve window is too large.

Some comments:

1) The good direction for recovering the alignment was not the one in which the picomotors reach the end of the range, meaning that the suspension position and the overall alignment hadn't moved too much. In any case we prefered to move the picomotor to exactly mid range (and have an optimally centerd intermediate mass), and then move the whole suspension to recover the alignement

2) We confirm that the yaw picomotor is quite critical as by moving it it's easy to make the intermidiate mass touch the magnets. For this reason also in KAGRA during the in-air alignment procedure, the whole suspension is usually moved with the help of some pusher (see Fig. 2)

3) We checked that the beam was well centered on the end mirror and slightly adjusted the camera in the end bench in order to have the beam also centered on it.

4) Some pictures of the current position of picomotors and intemidiate mass are shown in Figs 3, 4, 5.

[Eleonora, Matteo, Yuhang]

On Mon 22 we found out that the supervisor PC used for Labview control system was not able to connect to the CPU (NM2 for TM: 133.40.121.78), which manages the ADC and DAC for the telescope suspended mirrors (PR and BS).

We checked this CPU (which is in the central area) and find out that it was not possible to swich it on anymore. Futher investigations suggested that the problem was in  the power supply unit.

I noticed that slightly touching the mount of the first mirror on the HeNe probe path makes the beam deflect a lot. So I changed it with a more rigid mirror mount.
I also wanted to clean the mirror (Thorlabs 1/2" Protected silver mirror PF05-03-P01), but I broke it, so I replaced it with a Newport 5103 General Purpose Silver Coated Mirror that I found in Tama.

Then I aligned the HeNe probe using the pinhole. The beam passes through the following positions of the pinhole:

I noticed that the reflection of the first lens on the IU was going up to the chopper. This is very likely to generate noise in phase, so I tilted the lens of the IU.
I also noticed another reflection from the sample that is going to the chopper. I can't tilt the sample, so I solved it putting a simple aperture near the chopper. See picture.

I used the surface reference to align the pump and maximize the AC signal.

I used the new power meter for the pump:
pump power: 74mW , chopper on: 37mW, with surf ref sample: 27mW

I moved the imaging unit to find the maximum of the signal on the surf ref.

The maximum is not in the range of the IU stage. Probably because tilting the lens to move the probe reflection from the chopper, I missed  the focus of the IU telescope.

Therefore I started again the IU alignment. I put the IU stage at 60mm. I found the sharp image of the blade. I looked for the maximum of absorption signal again:

So I put the IU at 70mm.

I made a scan of the surf reference.

I noticed that the DC is not linear with the power of the probe. I measured 6.4V for 1.71mW, and 6.5V for 2.58mW. Then I replaced the LDS9 power supply with the 12V battery and the 2.58mW gave 9V of DC. So I decided to keep the DC below 6V with a OD wheel, and continue using  the LDS9 power supply.

After putting the wheel, I missed the alignment pump/probe, so I used again the  pinhole to center the probe.

Then I made a scan of the surf reference again.

And a scan of the bulk reference. The transmitted power is 20mW (of 37mW incident)

[Yuhang, Eleonora, Matteo]

In order to match the green beam transmitted from the green MC into the OPO we have characterized the green beam produced by OPO when it is locked with s-pol beam. Its paramenters will be the target ones for the mode matching telescope we have to design.

Since the beam is very weak (about 0.2 uW) we couldn't use the beam profile and we used a blade to do the characterization, instead.

Procedure: we attached a cutter blade to a traslation stage (See Fig.1) in order to be able to progressively cut the beam, by small and known steps.  For each step we recorded the power of the beam with a PD placed after the blade. Since the beam in gaussian the shape of this function is expected to be a error function (erf):

P (x) =  a*erf(sqrt(2)*(x-x0)/w)+o 

By fitting the data with this function we could extract the beam radius w for each position z at which the measurement was performed.

The beam size and the corresponding waist (in meter) are reported below and plotted in Fig 2.  The zero of the z-axis is taken at the incoupling mirror side of the OPO.

According to the measurement result of entry about green mode cleaner output beam characterization and OPO green output beam characterization, we can design the telescope to match them. From the above entries, we know the source waist is 287um while the target waist is 24um. Based on the revised version of optical layout which I just modified today(the position of green mode cleaner and distance between bench and chamber are modified)(https://gwpo.nao.ac.jp/wiki/FilterCavity/OpticalLayout). I also attached the part of bench we care, see attached figure 1. In this picture we know the distance between waists should be 82.5cm. Besides, we should avoid putting lens around 10cm, 27.5cm, 55cm and 75cm. And putting lens between 5cm and 72cm. 

By using all the information listed above, I used JaMmt did the simulation. See attached figure 2, which shows the initial condition I gave. Besides, I upload the green lens we have in our clean booth on our wiki page(https://gwpo.nao.ac.jp/wiki/FilterCavity/Optics). See attached figure 3, it shows the lenses I used. Actually I didn't put lens with very small focal lens, although we have lots of small focal length lenses. The fitting shoes many good results. I attached three here. Some of them even tell me the mode matching is 100%.

---------------------------------------------------------------------------------------

Since we found the mirror to reflect green after dichroic (outside OPO's incoupling mirror) should be farther than last version of optical layout. I changed it and did the simulation again. The result is attached in 7th and 8th figure.

During the last few days, we took some data which contains more than one TEM00 and also 87.6MHz sideband. We also extrapulate some more information from the data we have already have. From the analysis we did for them, I found

1. Extrapulate more information.

For the fit of only one peak, we found the fit can give more than one result. This is reasonable, since there is couple between finesse and fsr. For example, we can have two totally different result for the same measurement we did for TEM00(as in attached figure 1, you need to zoom in to see the detail). In this picture, the finesse 75, fsr 4.54GHz can give a perfect fit while the finesse 57, fsr 3.44GHz can give as well. The difference of these two fit is only that I give two different initial range of Finesse. For the first one, I give the Finesse range around 50. However, for the second one, I give the Finesse range around 70. However, the good news is that we can have FSR we expect if we fix the finesse around 70.

For the measurement of TEM00 and sideband. I also found the fit can give more than one set of result. If I give the original Finesse around 70, means around our expected value. We can get a reasonable result of cavity length. This agrees with the fit of bandwidth. See attached figure 2. The finesse now is 72 while cavity length is 39mm. Besides, this fit result give the similar calibration factor with the former fit.

However, all the fit with a finesse value around 70 give FSR around 4GHz, while the direct measurement of FSR gives 2.8GHz. The reason can be PZT cannot response linearly with our driving HV signal.

2. Measurement with TEM00(two) and sideband together

The whole measurement and fit is attached in figure 3. You can see even visually that the distance between twoTEM00 and sidebands are totally different. That is the reason why you can see the fit cannot give a result. But anyway, I tried to fit these two peaks seperately, the result is in the attached figure 4 and 5.  This time, the calibration factor becomes ridiculously different while the fit result of FSR and cavity length becomes also quite different. You can see from the seperate fit, the time difference in the first peak is 0.000501 while 0.000340 in the second peak.

3. Measurement with TEM00(three) and sideband together

The whole measurement is shown in attached figure 6. In this figure, you can see the difference of 0.2GHz in FSR causes the displacement of the peak away from the standard position. Also this causes the fit failed. I also choosed these peaks and fit them seperately. Firstly, the calibration factor is fitted around 1200, 1300, 1100 (MHz/V) seperately.  From this point of view, we can deduce the PZT scanning velocity firstly increase and then decrease. And the fit of the finesse and FSR of the first and third peak give a similar result(first: FSR = 4.9GHz, Finesse = 75  third: FSR = 3.9GHz, Finesse = 61 ). However, the fit of the second has a very large error.

The laser current is set at 200mA.
I sent the laser output on a PBS, to check how the jumps show up in the two polarizations. 
I collected the reflection with the PM100D-S145C. This is most of the power: 58mW.
The transmission (few mW) goes on a 50mm lens and a DET10N with load resistance of 500Ohm.

I watched together the analog output of the PM100D and the DET10N signal at the oscilloscope in AC coupling.
DC of PM100D is 335mV; DC of DET10N is 900mV.
I tried a similar measurement some time ago, but ince the PM100D signal is covered by high frequency noise, I couldn't see anything.
Now I filtered the signal with a SR560 set at AC coupling and low pass first order at 3kHz, gain 5.

The result is in the attached video.
Channel1 yellow is the PM100D filtered analog output. Channel2 blue is the output of the DET10N
The signal are completely anticorrelated. Apart from high frequencies of the PM100D that are filtered out by the SR560.

This proves that the jumps are due to polarization instability.

I also attach a video of the DC of the two detectors. Channel1 is the PM100D not filtered, channel2 (blue) is the DET100D. 

I took a video of the DC signal on the oscilloscope in 2 conditions

video7.avi
HWP at minimum of transmission
range 370uW
display 35uW

video8.avi
HWP at maximum of transmission
range 370mW
display 54mW

[Yuhang, Eleonora]

The green production depends on incident power and on the crystal temperature. We need to find the best temperature to have the best phase matching and maximize the green production. The measurement was performed from 5.794 to 7.787kOm, which corresponds to 312.976K to 304.767K.

The method we used is to change the p-pol beam to s-pol. In this case, we can lock OPO with this beam and at the same time have green production, since only s-pol produces green beam. In transmission of OPO, we put a dichroic mirror which reflects infrared light and transmits green. The reflected beam is focuse with a lens on the  Qubig PD used for the locking,  while the transmitted beam is focused with a lens on a  InGas PD.

The attached picture shows the green production as a function of the temperature. We see two peaks: the largest is at 306.8437 K, the second largest is at 309.3391 K.

 The IR input power was 141 mW.   At the optimal temperature (309.3391 K), we measure a the green power of 0.277 uW  while the IR trasmitted power is 0.189 mW . (Tramsmissivity:  0.13%.  Lower than that measured few days ago #elog 999)

I set the laser current at 200mA.
I used the PM100D power meter with the S145C power meter head to check the laser power fluctuations in several conditions.
I connected the analog output of the PM100D to the oscilloscope, and took some videos
I set the bandwidth of the PM to HI. I swithed OFF the auto range and set the range to 370mW.

PM right after the fiber output
display power: 61.3mW
analog out DC: 365mW
AC: video1 oscilloscope

PM after the first PBS
display power: 58.3mW
DC 350mW
AC: video2 

Then I rotated the half-wave plate to change the  PBS-HWP-PBS system transmitted power 

after PBS-HWP-PBS max transmission
display power: 54.3mW
DC: 327mV
AC: video3 
 

after PBS-HWP-PBS min transmission
display power: 0.0mW
DC: 10mV
AC: video4 

after PBS-HWP-PBS half transmission
display power: 25mW
DC: 150mV
AC:video5 

Changed the PM range to 370uW
after PBS-HWP-PBS min transmission
display power: 27uW
DC: 160mV
AC: video6 

Conclusion: looks like the fluctuations increase when the polarization rotates toward the minimum transmission.
I can't recognize the same jumps I saw in the DET10N

I checked the linearity of the system by changing the incident power of the HeNe laser.

The PD DET10A gives a DC signal proportional to the laser power. And the AC signal obtained after the lock-in demodulation is proportional to the absorption signal and to the DC. So if there is some non-linearity (for example saturation effects) it should cause the AC/DC signal to be not constant.

I put an optical density wheel to reduce HeNe power. I rotated it a bit to change the DC level, acquired some points of AC and DC in that condition, and rotated again the wheel a bit to a different DC value and so on.

The plot shows good linearity.

Temperature change causes the change of cavity length. However, PPKTP has a different refractive index for p and s polarization. So the FSR change for s and p polarization are different. If we want to make both s and p polarization resonant inside the crystal, we need to choose a good temperature or choose a good frequency difference between s and p polarization. As pointed out as the entry about the temperature we should measure, we did the measurement. Besides, we measured the calibration factor before so we can know the s-p frequency difference change according to temperature change.

The measurement was done like this, we firstly check the rotation of HWP so that the transmission of HWP keeps s-pol. At this time(almost all the light is reflected by PBS cube), the angle is 331 degrees. Then put it just in front of OPO housing, rotate HWP to have some p-pol inside OPO. In the end, we make s-pol is higher than p-pol. So we can differentiate them by hight. Then we changed temperature and made measurement.

The measurement result is listed in the last attached figure as a sheet.(detail is in the attached figure 1-17)

Most of the case, the find of frequency different is based on finding the highest and second highest peak. And then use the calibration factor and the slope of ramp signal. However, the find of frequency difference in temperature of 307.4K is performed by fitting the scanning by the addation of two airy functions.

By using these data I plot the birefringence effect in figure 18, the birefringence constant is 397.76 +/- 0.99 MHz/K.