Polarization Maps on Sapphire Samples from SHINKOSHA #S1, #S2, #S3, #S5, #S6

[Simon, Pengbo]

Attach to this report are the polarization maps and polarization angle distribution maps.

As can be seen, the polarization maps of these five samples both very homogeneous.

Recent setting of OPO temperature and p pol PLL

green power (mW)	OPO temperature (kOhm)	p pol PLL (MHz)
0		305
40	7.19	165
60	7.2	150

OSTM coating absorption

Pengbo, Simon

Attached are the figures summarizing the results on the absorption measurment of OSTM's coating (-> Sigma Koki).

As can be seen, there are many spots with absorption excesses. Most likely, they are due to defects of the coating. Other than that, the absoprtion is quite homogenous with a mean value of 16 ppm.

Measurement of IR power loss from two PBS before and after the in-vacuum FI

We have two reference points for IR injection to filter cavity(FC). They are from the two PBS before and after the in-vacuum FI. And this will be also a loss for squeezing in the future. Also, if we want to monitor the power we are injecting to FC, this can be a choice of checking point.

So we measured the power arriving at these two points. 

power loss from PBS before in-vacuum FI (first PBS)	0.59uW	attached photo 1	so the polarization before PR chamber is optimized better
power loss from PBS after in-vacuum FI (second PBS)	2.78uW	attached photo 2	so the polarization before the second PBS is not optimized very well

We could optimize the HWF after the in-vacuum rotator, but we will lose some isolation factor.

Actually, Matteo said maybe we can have the choice to optimize the Faraday rotator, if so, our maximization of IR reflection should be done by optimizing the Faraday rotator.

PDH signal comparison from FC reflection(qubig PD and NIKHEF PD)

To compare the signal from qubig PD and NIKHEF PD, I did this check. (There was already a check reported in the entry 1670)

The situation of green power is

green power on qubig PD	210uW
green power on NIKHEF PD	1.2mW

The electronics after PD is

1. qubig PD	then goes to RFamplifier(14dB)	then goes to demodulator (used for FC locking, from the mini-circuit company)	then detect on the oscilloscope	----------
2. NIKHEF PD (I put all light on the first quarter of this PD on purpose)	then I take the RF signal from the first quarter	then goes to RF amplifier(14dB)	then goes to demodulator (used for FC locking, from the mini-circuit company)	then detect on the oscilloscope

The result of PDH signal is

qubig PD pk-pk	164mV	attached Fig.1
NIKHEF PD pk-pk	14.8mV	attached Fig.2

Note: If I put these signal to NIKHEF demodulator is qubig PD pk-pk 35mV(reduce by a factor of ~5), NIKHEF PD pk-pk 10mV(reduce by a factor ~1.5).

Frequency tuning of coherent control sidebands (CCFC)

[Aritomi, Yuhang, Yaochin]

We could see transmission of both coherent control sidebands (CCSB) from filter cavity at CCD camera with 40mW green. Pic.1 shows upper CCSB transmission at camera. We found that when CC PLL frequency is 7MHz, AOM frequency of upper CCSB is 109.04179MHz and lower CCSB is 109.02990MHz while AOM frequency of carrier is around 109.036 MHz. In CC locking of filter cavity, either of CCSB is on resonance and other CCSB should be separated by ~100Hz. So we tuned CC PLL frequency in order to have ~100Hz of CCSB separation frequency. Note that AOM frequency separation corresponds to twice of CC frequency separation.

CC PLL frequency (MHz)	AOM frequency of upper sideband (MHz)	AOM frequency of lower sideband (MHz)	CC frequency separation (Hz)
7	109.04179	109.02990	5940
6.997	109.03608	109.03599	45
6.9967	109.03665	109.03528	685
6.9963	109.03728	109.03461	1335
6.995	109.03991	109.03189	4010 4010
6.993	109.04322	109.02855	7335

Frequency separation with 6.997MHz of CC PLL is close to frequency separation that we want, but AOM frequency with 6.997MHz of CC PLL is not so precise since CC transmission seems not changing at camera around this region and we couldn't distinguish upper and lower sideband.

Pic. 3 shows plot of CC PLL frequency and CC frequency separation. The fitting result is

CC frequency separation (Hz) = -1907605 * CC PLL frequency (MHz) + 13347486

If you want to have 108 Hz of CC frequency separation, CC PLL frequency should be 6.996930 MHz or 6.997043 MHz.

After we set CC PLL 6.997MHz, we found that CC1,2 error signal oscillated since demodulation frequency of CC1,2 was not tuned. After we set demodulation frequency of CC1 6.997*2=13.994MHz and demodulation frequency of CC2 6.997MHz, the oscillation dissappeared and we could lock CC1,2. However, when we open the CC2 loop, CC2 error signal has low frequency oscillation (Pic. 2) due to residual frequency difference between CC PLL and demodulation phase and it is difficult to remove.

Phase noise with/without feedback to INPUT mirror

Aritomi, Yaochin and Yuhang

After we feedback CC2 correction to INPUT mirror, we calculated the phase noise. The result is shown in entry 1719. But when we compare this result to entry 1614, we found after feed it back to INPUT, the acuumulated RMS is even higher.

Actually, if we look into the detail, the main difference comes from basically high frequency region (kHz to 100kHz). Personally, I have the impression the phase noise is different sometimes. For example,  in entry 1522, we reported phase noise reduce after long time laser on. But for example, today(we kept laser on also for several days), the phase noise is a bit higher and we couldn't even lock CC2 loop(we could lock one month ago).

So it is important to compare the high frequency region of phase noise with/without feedback to INPUT mirror. We did this measurement with filter cavity unlock. Because we have some issue with the CC2 loop error signal(Aritomi-san may report later), so the comparison is not calibrated to the unit of [rad].

The comparison is only in high frequency region because we could lock CC2 only for very short time when there is only feedback to CC2 phase shifter PZT. But at least from this measurement, high frequency is not affected whether we feedback to INPUT mirror or not.

Results Of SHINKOSHA#S1, SHINKOSHA#S2, SHINKOSHA#S3 Absorption Maps

[Simon, Pengbo]

Attached to this report are the XY maps and distribution histograms of the absorption coefficient from the Shinkosha #S1, #S2, #S3.

CC2 error signal with/without test mass feedback

We took a movie of CC2 error signal when test mass feedback is engaged (blue line in attached movie). Test mass feedback is engaged around 12s in the movie. CC2 seems better with test mass feedback. Attached picture shows CC2 correction signal with gain 0 (red line) and gain -1 (blue line). 

CC2 free running phase noise when turbo pump is ON/OFF

During this measurement, filter cavity is locked.

Alignment of IR with green and IR overlapping

[Aritomi, Yuhang, Yaochin]

First, we found that IR transmission was only 450 and it was due to optimal p pol PLL frequency changed. Current optimal PLL frequency with 60mW green is 150MHz. After this change, IR transmission became 2500.

Then we aligned IR with two steering mirrors on the bench to make green and IR overlap at first and second target.

Current mode matching is as follows. Mode matching is improved a bit and 93.5% now.

Mode	IR transmission
TEM00	2000
HG10	120
HG01	180
IG20	120
offset	94

Although green and IR should overlap, IR transmission is still fluctuating. Attached picture shows spectrum of IR transmission.

Off resonance reflectivity is 79-81%. On resonance reflectivity is 44-54%. Note that p pol PLL is 305MHz without green and current BAB power before OPO is 186mW.

Check the overlap of GR and IR

[Aritomi, Yuhang, Yaochin]

We could find green and infrared beam from the first and second targets. We draw a circle of beam edge (as shown in attached pictures) on the screen and compared the profile of green beam and IR beam.

In the end, we confirmed both IR and GR overlapped very well.

Attached picture 1 and 2: IR and GR on the first target (before overlap)

Attached picture 3 and 4: IR and GR on the second target (after overlap)

Note: In entry 1709, we couldn't see IR on the first target. This time we could see IR on the first target, the reason is figured out that the iris inside the camera was closed too much. After the fully open of it, we solved the problem.

First trial of input test mass feedback of CC2 correction signal

[Aritomi, Yuhang, Yaochin]

First we checked CC2 correction signal. Correction signal is 8.2 Vpp (Pic. 1). As you can see from the picture, it is saturating. This may be saturation of servo.

Then we injected this correction signal to ADC CH13 directly and tried to feedback to input test mass. We just set the gain -0.5 for filter. Surprisingly, the test mass feedback worked well without any filter. Pic. 2 shows CC2 correction signal when gain is 0 (blue), -0.5 (green), -1 (red). CC2 correction signal is suppressed when the loop is closed. CC2 control is more or less stable with test mass feedback now. We will improve CC2 with proper filter. Pic. 3 is CC2 in loop phase noise when CC2 is stable with test mass feedback (CC2 error signal is 56.4 mVpp). Note that free running phase noise and in loop phase noise are measured at different day. Turbo pump was ON.

Comparison of the absorption map between SHINKOSHA#5 and SHINKOSHA# 6

[Simon, Pengbo]

Attached to this report are the XY maps and distribution histograms of the absorption coefficient from the Shinkosha #5((shown in the attached figure 1 and 2), and Shinkosha #6(shown in the attached figure 3 and 4), the color bar scale for the two maps is the same.

As can be seen, sample #6 has better absorption than sample #5, and we think this is due to the inhomogeneous of the absorption along the z-axis of the samples.

For these two samples, we choose the center of the z-axis. We can check our assumption by doing another XZ maps.

Measurement of microphone spectrum of CC2 phase shifter

Aritomi, Yaochin, and Yuhang

For CC2 phase shifter, we could always hear the sound around it and we wanted to check what is the frequency of this sound. So we used a sound spectrum analyzer and it shows only one frequency component which is 6.9kHz. Note that we couldn't see this oscillation from the oscilloscope. 

We checked the squeezing spectrum and found that this frequency appears in the spectrum of squeezing. Also in the squeezing spectrum we could find the peak at ~23kHz. These peaks come from the resonance of CC2 phase shifter.

However, we found there was also a clear peak at 16kHz. And we found this peak when we measure OLTF of CC2. So it seems this peak comes from servo? Maybe we can consult with Pierre.

Recovery of IR alignment

[Aritomi, Yuhang, Yaochin]

Last Friday, we lost IR alignment, but the IR alignment could be easily recovered by moving one steering mirror on the bench. I moved both pitch and yaw, but especially pitch was misaligned. Now mode matching went back to ~90% as before.

Measurement of microphone spectrum of CC2 phase shifter

Aritomi, Yaochin, and Yuhang

For CC2 phase shifter, we could always hear the sound around it and we wanted to check what is the frequency of this sound. So we used a sound spectrum analyzer and it shows only one frequency component which is 6.9kHz. Note that we couldn't see this oscillation from the oscilloscope. 

We checked the squeezing spectrum and found that this frequency appears in the spectrum of squeezing. Also in the squeezing spectrum we could find the peak at ~23kHz. These peaks come from the resonance of CC2 phase shifter.

However, we found there was also a clear peak at 16kHz. And we found this peak when we measure OLTF of CC2. So it seems this peak comes from servo? Maybe we can consult with Pierre. (figured out later this 16kHz if from mechanics)

Check of IR and GR overlap on First target/Second target/IR camera

Aritomi, YaoChin and Yuhang

To have a feeling of IR and GR overlap. We checked three points.

1. First target(shown in the attached figure 1 and 2). In the first figure, IR is actually in the center. When I was checking it, I could see the difference if Aritomi-san blocks or unblock the IR beam. In the second figure, there is GR. But it is difficult to tell the overlap level.

2. Second target(shown in the attached figure 3 and 4): It seems IR is higher and a bit left.

3. End camera(shown in the attached figure 5 and 6): It seems IR is also higher and a bit left. 

Diagonalization of length control matrix for Input mirror

Eleonora, Yaochin, and Yuhang

Firstly, we tried to excite FC length by using H1 and H3 coil. But we found H1 is not working(because we sent excitation to H1 only but the mirror couldn't move accordingly). So we checked if we were sending the signal and whether the signal is reaching the coil driver. We found the signal is reaching the coil driver but with some noise. The frequency of this noise is measured as 700kHz.

So we decided to excited FC length by using H2 and H4 coil. The method we used to diagonalize is to measure the response of the mirror optical lever signal. We sent excitation to channel K1:FDS-INPUT_Z_CORR_fil_EXC with a frequency of 4Hz and amplitude of 4000. We set the matrix index for H2 as 1 and measured the spectrum of K1:FDS-INPUT_Y_fil_IN1. There was a peak of 126.9 at 4Hz. Then we set the matrix index for H4 as 1 and measured the spectrum of K1:FDS-INPUT_Y_fil_IN1. There was a peak of 173.4 at 4Hz. So we decide to put index 1 for H2 while index 0.73 for H4 to make the coupling to yaw to be zero.

In the end, we measured the transfer function from the excitation of the INPUT mirror length to FC correction. There should be a response. We also measured the TF from the excitation of INPUT mirror length to INPUT mirror yaw optical lever. There should be not a response. Also the coherence of the above two. The result is shown in the attached figure 1. We have a response from the FC correction signal 10 times larger than the response from the yaw optical lever. Also, the coherence is better for FC correction. This can be a good start for the implementation of the feedback for CC2 phase noise.

Filter cavity GR_tra DC dropped again and its recover

Eleonora, Pengbo, Yaochin, and Yuhang

We found green transmission DC dropped to ~600 counts at the beginning of today's work.

First, we checked the GR higher-order modes we have for the filter cavity.(As shown in the last figure) It was fine.

Then we went to the end room and found the GR_tra PD was tilted almost by 45deg. We think the PD cables have been accidentally pulled during the work to repair air conditioner which took place this morning in the end room. After we correct the angle of this PD, we found the signal went back to ~2700 counts.

We also checked green power at two points(shown in the attached picture 1) in today's situation (the green injected power is 12mW). The first point is just before GR_tra PD and it is 45uW(shown in the attached picture 2 and measured as shown in the attached picture 3). The second point is just after the green BS and it is 1mW(shown in the attached picture 4).