Pengbo, Simon

After having found a good calibration for the upgraded system, Shinkosha#7 has been characterized again in terms of absorption (input P-polarization). The results can be seen in the attached figures (Z = 76), compared to the ones before the upgrade (Z = 74.1).
The basic shape as well as the mean aborption coefficient are almost the same, which is a very good achievement.
However, there are also small differences. It seems that the center absorption, which is at maximum in both measurements, is higher than before the upgrade. Also, there are basically two peaks in the absorption-coefficient distribution for yesterday's measurement. The reason is not clear. Maybe the change in polarization has some effect on the results. We are going to further analyze this phenomenon by using a polarization control on the pump-beam.

Today BS was less noisy, however, as reported in elog 2046, filter cavity was unlocked by the filter cavity length correction signal drift.

Actually I was thinking maybe it is related to I just turned on laser. But even after more than one hour of laser-on, the length correction signal is still drifting. Anyway, in the end, the signals including FC_GR_tra, FC_GR_correction, FC_IR_tra and PDH_detuning were monitored for one hour.

We could see that during this one hour, without any re-alignment, the detuning never jumps. The detuning changing range is roughly from 40Hz to 80Hz.

There was one period of time, green is quite stable. This is different from elog 2047, which has green misalignment. We could see during this green stable time, IR detuning drift is still happening.

Simon, Pengbo

Today we started the IU adjustment with the bulk reference trying to find the maximum of AC/DC value.

We start setting the IU back to initial position ( -> 70 mm) , decreaing the distance then find the maximum at 54 mm with a value of 0.0158. We stopped here since the maximum position is too far away from the achievable distance towards measuring KAGRA-size samples.

Then we moved the first lens of the telescope 2 mm away the second lens and adjust the crossing-point with pin-hole and surface reference sample(IU back to 70 mm). We fastly found it and adjust the AC to be maximized with second lens -> R_surf = 16 1/W.

After that we did another adjustment of IU to find the AC/DC maximum, and found it  was at 69 mm. We make a bulk reference scan and get R_bluk = 0.713 cm/W. Then we start the absorpiton measurement on KAGRA-size #7 sample.

(This is the report of yesterday20200217) Aritomi, Eleonora and Yuhang

We found BS moved quite a lot, this was figured out later that it is because of the saturating of BS coil drivers. Then we tried to remove the DC offset of BS, which didn't work very well since we found BS moved more than before. In the end, we open the oplev local control for BS and did the experiment.

We locked filter cavity with GR and closed the pitch dithering. The BAB is sent to filter cavity and we demodulated the filter cavity reflection to have PDH signal. We detuned IR by 60Hz and leave it for one hour. During this time, the filter cavity was unlocked several times. We relocked three times, once we just relock and twice we re-aligned filter cavity.

The filter cavity GR/IR transmission and GR filter cavity length correction and IR detuning was monitored.

The result is attached. We can see different alignment causes different detuning value.

There was a zoom-in part, which we can see the correction singal doesn't change but alignment and detuning keep changing.

Eleonora and Yuhang

Today we tried to monitor detuning drift but filter cavity always got unlock(almost every 5 minutes). And we noticed that the length correction signal was always drifting. Considering the change of 3000counts, we can convert it to the voltage we send to laser PZT is drifting by about 3.66V(3000count/1000count*0.61V*(100/50)).  Maybe it drifts out of range and made filter cavity unlock. However, from the manual of laser, the PZT driving coefficient is 1MHz/V. And the PZT driving range is +/-65MHz. It seems they are not consistent.

Besides, the reason for this correction drift is till under investigation.

Since the parts for fixing the cavity have been delivered, I started the installation of a folded cavity.

1. Installed a stage and put a spacer on it.
2. Fix the position of the spacer.
3. Install a mirror and injected a beam.
4. Adjusted the alignment roughly.

Now the "transmitted" beam can be picked off with two steering mirrors.

Pengbo, Simon

after the successful polarization test, we started to do also the absorption tests in order to see whether we can reproduce our results from before the upgrade.

Actually, the first test was performed with the TAMA#1 sample but we repeated the test with Shinkosha#7 since we don't have any direct reference for TAMA#1.
We can say that the actual results in terms of the absorption coefficient are well reproducible (see attached figures from before [Z = 74.1] and after [Z = 80] the upgrade).
However, there is an issue with the "effectivity of absorption", meaning that we get a factor of 2-3 smaller AC-values than before. The reason is still not quite clear to us. It may be that the size of the beam-waist, which is 10% larger than before, has a bigger influence on the AC-value than anticipated. We will try to further reduce the waist and see what may happen.

At least we can exclude that the polarization of the beam has any significant influence on the absorption. We have tested this yesterday as well.

I installed 2 lenses for mode matching of TEM00 mode beam.
Their positions can be optimized by the stage.

The beam is collimated less than 100 um dia.
The requirement is about 50 um and I need optimization.

Tomorrow, the rest parts for folded cavity installation will be delivered and I will install the cavity with these parts.

This is my fitting result.

Eleonora and Yuhang

We tried to measure filter cavity FIS with scanning LO phase to get squeezing and anti-squeezing level. But since the suspended mirrors are always moving a lot, we couldn't see the expected M shape.

Thank you for the measurement.

Can you add the fitting result of homodyne angle? 

Eleonora and Yuhang

We fixed the CC2 demodulation phase and measure FDS during more than one hour.

During these measurement, filter cavity got unlock twice. Also, for the last measurement, we misaligned filter cavity yaw.

The detuning was stable within ~10min. Even without unlock, detuning changes and unlock almost doesn't change detuning if green is recovered. But detuning will change after a bit misalignment.

Aritomi, Eleonora and Yuhang

Today we monitored BAB PDH signal for more than one hour, which is taken by collecting filter cavity BAB reflection on an RF PD. During this time, filter cavity got unlock several times. 

The orange color line is normalized PDH signal.

The cyan color line is PDH signal before normalization. 

The dark blue color line is filter cavity control loop correction signal.

• We could see basically the PDH signal doesn't drift and it means the BAB is always on resonance. And there is fluctuation, by taking the pk-pk value of this thick line(let's say 8count pk-pk) and take calibration factor of 2Hz/count. We should have RMS locking accuracy of 11Hz. And this is roughly consistent with the measurement of locking accuracy spectrum which give result of 6Hz.
• We could see that in the middle, there is a time filter cavity got unlock, and normalized PDH has an offset from zero.
• There are many times filter cavity got unlock and everytime when it got lock again, the PDH went back zero again.

But we observed at least twice a strange behavior. The PDH becomes smaller and BAB transmission becomes also smaller while GR transmission is kept. After we tune AOM frequency, the BAB transmission was recovered and the detuning was changed a bit. 

Earthquake of magnitude 7 occurred at 19:44 JST in the north of Hokkaido (https://www.jma.go.jp/en/quake/20200213104426395-13193432.html). Mirrors are shaking very much and the lock is difficult to recover.

By monitoring the long term drift of the supesnsions, we observed a  ~24h cycle likey related to the temperature. It is especally visible in pitch. The "inversion points" are at about 5 am and 5 pm. Shorter period oscillations seem to appear in yaw.

Non normalized:  0.011 Hz/count

Normalized:  1.98 Hz/count

[Aritomi, Yuhang]

We locked FC with green and implemented dithering to be sure that alignment of FC is good. We locked IR off resonance (detuning is 1kHz). Unfortunately, today there was a large mode mismatch peak in AMC, which was ~50mV out of 1V. This becomes ~10% readout loss.

Measured loss is 40(1) % and phase noise is 30(5) mrad.

K1:FDS-WFS2_DC_SEG1_OUT    ->  PDH IR 

K1:FDS-WFS2_DC_SEG2_OUT    ->  IR TRA

K1:FDS-WFS2_DC_PIT_IN1         ->  normalized PDH IR  (i.e PDH IR/IR TRA)

Aritomi and Yuhang

The loss measurement doesn't include intra-OPO loss (roughly 8%) in injection.

This measurement also doesn't include homodyne loss (roughly 4%), PD quantum efficiency (roughly 1%) in readout.

total loss: