PItch BS channel shows an excess of noise. Damp loops are open and even if I remove DC on coils for the aligment, noise is still there.

This was observed serveral times recently and after a while it goes back by itself. I think it is something electronic, but what?

Yaochin and Yuhang

As we reported in elog1867 and elog1857, we improved the matching between LO/BAB into AMC. However, we didn't care about the residual p-pol component after the optimization. And actually, it is really necessary to optimize them and remove the p-pol component.

LO part (The AMC spectrum before the change is attached as picture 2) (s-pol 11V, p-pol 14mV)

     Actually, I am confused that how can we have a p-pol component from LO. Because LO is provided by IRMC TEM00, TEM00 should provide a quite clean s-pol light. 

     1. We replaced the first mirror after IRMC with a PBS plate.

             After this, the p-pol peak is reduced from 14mV to 9mV(attached picture 3). (We also double-checked that total power doesn't change after the replacement of PBS, so this p-pol reduction is a real      reduction effect from PBS)

     2. We put an HWP just before homodyne BS.

            By rotating HWP, we could almost remove the p-pol component totally(attached picture 4).

BAB part (The AMC spectrum before the change is attached as picture 5) (s-pol 4V, p-pol 43.2mV)

     For the p-pol component inside BAB, it comes from the not optimized HWP. So we just rotate HWP. And then the p-pol is basically removed totally. (attached picture 6)

We measured the squeezing spectrum before and after this change, but there is no obvious change.

Yaochin and Yuhang

We found we could balance homodyne by aligning the flipping mirror and the lens before the far 'eye' of homodyne. (As shown in the attached figure)

By aligning pitch/yaw of flipping mirror and pitch of lens, we could also recover the balance of homodyne.

Today Miyakawa-san managed to solve the long-lasting problem we had to run python script called by medm button (see entry #1656).  

He modified the command used by the medm desktop icon to call medm to: bash -c 'source /kagra/apps/etc/client-user-env_tama.sh && cd /opt/rtcds/kamioka/k1/medm && /opt/rtapps/epics/extensions/bin/linux-x86_64/medm -x sitemap.adl'

Now the command is also calling a script he made (/kagra/apps/etc/client-user-env_tama.sh) to include all the needed evironmental variables.

Somehow when opening MEDM with this command the PV info on the channel cannot be seen anymore. The icon MEDM (edit) was not changed so, by using that one, we can still read PV info ( but not run python scripts) 

Conlusion: we can now run python screep from MEDM screen open by desktop icon, and this should help to better manage control loops, while we wait for guardians.

Yuhang and Yaochin

We reported sometimes the CC1 loop cannot be locked.

Today we found the cable which is sending error signal for the loop has some connection problem. When we shake the cable, the error signal sent to the loop got a lot of noise.

We have already replaced it with a new cable. We checked the new cable doesn't introduce noise when we shake it.

[Optics]
I adjusted the lenses' positions for mode matching.
I tried to measure the beam profile inside the cryostat chamber, but it was too difficult to measure the distance.
Fortunately, some amount of the beam was reflected by a PBS in front of the chamber due to the imcomplete alignment of FI and it propagets on the optical table.
The profile of this kicked beam is almost same as the beam which incoming to the chamber.
Therfore I will use this beam to measure the beam profile.

[Cryostat]
I assembled another window for 80K shield as attached picture.
I am planning to use two 80K shield windows for my experiment.

[Next Step]
Measure the beam profile.
Install a FP cavity to characterize fused silica mirrors.

Simon

I have set up a measurement bench in TAMA to easily and fastly take reflection data of a 1064nm laser on samples (mostly optics, I guess) as a function of the angle of incidence (from 0 to 90 degrees) and input polarization.
The system works fine but it still not entirely finished. For example, we plan to install a rotation-gear on the angular rotator to compensate the 2*AoI effect, so that a manual change of the PD won't be necessary anymore.

Especially today, I was trying to further understand the laser. I recognized that the used PD cannot measure powers more than 8mW (which is strange as there is a threshold of 50mW written on it...). But putting the laser at such low power-gains, the stability of the laser is not good at all. Therefoe, I was inserting a ND-filter in front of the laser and increased the gain so that the laser-power is stable (change of power after 10 minutes <1%).

After setting up the laser, I tested the system with a Brewster-polarizer (TFP-1064-pw-2037-uv) which we use for the PCI. The input power for the measurements was set for both S and P polarization to 6.53mW.
Please see the attached graphic which illustrates the results. For P-pol, a minimum at 63 degrees was found (R = 3.5%), while for S-pol, the reflectance did not change much at all over the entire measured range (which is somehow expected, I guess).

Please note that the data taken at 80 degrees AoI are probably not very accurate, since I believe that the beam was clipped (I need to check this further, maybe installing a lens is better).

Today I assembled a window for 80K shield with 1~1.5 um broadband AR coated glass.
After that I found that a washer for M4 screw cannot go through the though hole, and it is a little bit difficult to install on 80K shield.
I will ask Ueda-san (KEK) about this.

Yuhang and Yaochin

Firstly, we checked the amplification of BAB with 60mW. The factor is around 30. While the characterization we did last year shows amplification of roughly 60.

Then we measured SQZ and ASQZ at different green power

From the above data, loss is 25.1% and phase noise is 19.4mrad. The loss is consistent with Aritomi-san's result. While the phase noise is a bit reduced.

We also measured the CC2 phase noise. The measurement is done with a 40mW green power and 110deg CC2 demodulation phase. Then we got CC2 pk-pk 62.4mV. We use this for calibration and got a phase noise of 4.25mrad.

[Camilla, Irene Federico]

I updated the control filter for the PR PIT.

The old filter is in ASC2G and notches.

The new filter works with the filters, test and boost2, with the gain of 1.0. The residual decreases more than 10 times 

Yuhang and Yaochin

We should put AMC offset always at 150V, because different offset will result in different alignment condition.

Attached three pictures show the situation of AMC's offset as 40V, 150V and 240V. We could see quite different alignment situation.

[Yuhang, Matteo, Eleonora]

On Tue 3/12 morning we found the FC has difficulties to acquire the lock and to keep it, while in the past months lock was very robust and easy to get.

After some investigation (and having tweeked the gain of ramepeato) we found the input power was reduced from 12 mW to 9 mW, due to temperature change of SHG.

The day before the air conditiong was switched from summer to winter mode by Yuhang.

The injected power was put back to 12mW and the gains were tuned to get UGF ~ 15 kHz.  (See pic 1 and 2). Anyway it seems a loop oscillation happened from time to time and error signal looks noisier than usual. It was quite late and we coudn't keep investingating any longer. We haven't been in TAMA yesterday and today because of F2F.  Will check the situation tomorrow.

The main effect causing degradation of the overlap between LO and reflected beam form FC when it is ON and OFF resonance is the relative phase change between TEM 00 (resonating the cavity) and HOM (not resoanting). Raffaele did a quick calculation to show this and the numbers he found are in rough agreement with those we observed. I attached a pdf with a similar calulation and a plot that shows the overlap degradation as a function of the not coupled power into FC. 

[Eleonora, Matteo, Raffaele]

On 29/11 dIthering was implemented also in Yaw, with the same scheme used for pitch.

Dithering lines are injected on INPUT and END mirrors at a frequency of 16.5 Hz and 17.5 Hz respectively. Each error signal obtained from transmitted power demodulation is filtered with a simple integrator and fed back to its own mirror.

From the time taken by the loop to recover the alignemnt after misalignig a mirror on purpose, the UGF seems smaller 50 mHz. I tried to measure openloop TF with a swep sine injection at very low frequency but coherence was not good, no matter the excitation amplitude.

Note that with the current filter configuration (simple integrator) the UGF of dithering is limited by the interaction with the damping loops. See attached simulation. We discussed the possibiliy to shape the filter to avod the instabilty. Anyway since damping UGF is quite high ~1 Hz. in the present configuration the sum of the two loops is stable. Another possibility would be to go back to DC controlled optical lever and add the dithering DC correction to the set point of this loop. I think this is what is done in Virgo.

I put fused silica mirrors inside the cryostat chamber in order to construct a FP cavity with cavity length od 9.9cm.
Then I injected a laser into the cavity and tried to see the flash with PD which is located at transmission port.
So far I could not see any transmitted beam.

I will improve the alignment and mode matching with two lenses.

Yaochin and Yuhang

We see anti-squeezing rotates to squeezing @2kHz. But the filter cavity was really unstable.

average: 50 times

Setting of hand-fit parameters:

sqz_dB = 9.5;                      % produced SQZ

L_rt = 105e-6;                    % FC losses

L_inj = 0.17;                     % Injection losses

L_ro = 0.1;                      % Readout losses

A0 = 0.05;                         % Squeezed field/filter cavity mode mismatch losses

C0 = 0.03;                        % Squeezed field/local oscillator mode mismatch losses

ERR_L =   5e-12;                   % Lock accuracy [m]

ERR_csi = 50e-3;                  % Phase noise[rad]

phi_Hom = [-1/180*pi, -11/180*pi, -75/180*pi];         % Homodyne angle [rad] (you can input a vector of values)

det =  2e3;                       % detuning frequency 

int = 0.8e3;                         % frequency range = det+/-int 

Setting of shot noise level/detuning error parameters:

semilogx(fd+34, d_a1+133.5,'k','LineWidth',3) %squeezing to anti-squeezing

hold on

semilogx(fd+32, d_a2+133.5,'LineWidth',3) %anti-squeezing to squeezing

hold on

semilogx(fd+150, d_a3+133.5,'LineWidth',3) %intermediate

[Aritomi, Yuhang, Yaochin, Eleonora]

First we measured frequency independent squeezing with filter cavity (Pic. 1). We had 3.5dB squeezing down to 80Hz with filter cavity.

Then we alined BAB to AMC with 1kHz filter cavity detuning and measured FDS (Pic. 2). Fitting parameters are follows. It seems that detuning changed for squeezing and anti squeezing measurement.

sqz_dB = 13;                      % produced SQZ

L_rt = 100e-6;                    % FC losses

L_inj = 0.20;                     % Injection losses

L_ro = 0.11;                      % Readout losses

A0 = 0.1;                         % Squeezed field/filter cavity mode mismatch losses

C0 = 0.1;                         % Squeezed field/local oscillator mode mismatch losses

ERR_L =   5e-12;                  % Lock accuracy [m]

ERR_csi = 80e-3;                  % Phase noise[rad]

phi_Hom = [-5/180*pi, pi/2];      % Homodyne angle [rad] 

det =  [1074 1004];               % detuning frequency [Hz]

cavity_pole = 59.6;               % cavity pole (Hz)

We measured FDS again with smaller frequency span (Pic. 3). Fitting parameters are follows. This time, it seems that produced SQZ also changed in addition to detuning.

sqz_dB = [13.5 11 13.5];          % produced SQZ

L_rt = 100e-6;                    % FC losses

L_inj = 0.20;                     % Injection losses

L_ro = 0.11;                      % Readout losses

A0 = 0.05;                        % Squeezed field/filter cavity mode mismatch losses

C0 = 0.1;                         % Squeezed field/local oscillator mode mismatch losses

ERR_L =   5e-12;                  % Lock accuracy [m]

ERR_csi = 80e-3;                  % Phase noise[rad]

phi_Hom = [-3/180*pi 70/180*pi 7/180*pi];       % Homodyne angle [rad] 

det =  [970 1005 965];                          % detuning frequency [Hz]

cavity_pole = 59.6;               % cavity pole (Hz)

Aritomi, Eleonora, Yaochin, and Yuhang

For the adjustment of OPO/FC matching and FC/LO matching, we always do the following.

1. Send BAB into the filter cavity. Align IR into the filter cavity while filter cavity alignment is kept in the best condition. Check IR 0/1/2 order modes by changing AOM frequency.

2. Align BAB reflection into AMC while the filter cavity is locked.

However, we found that the matching from BAB_ref into AMC is quite different when the filter cavity is locked and detuned. And we know that we measure FDS while FC is detuned, so we decide to align BAB_ref into AMC while FC is detuned. So now the procedure is as follows.

1. Send BAB into the filter cavity. Align IR into the filter cavity while filter cavity alignment is kept in the best condition. Check IR 0/1/2 order modes by changing AOM frequency.

2. Align BAB reflection into AMC while the filter cavity is detuned.

However, we are not very sure why the matching can be so different when FC is locked and detuned. Because I think the BAB_ref should be quite similar in both cases of lock and detuned since we had 94% of matching from OPO/FC. I put the reason for my thought as following. Please correct me if there is something wrong.