On friday 29/06/23 the NE was opened and three Si wafers placed inside for dust contamination monitoring were retrieved (named NE_290623_N, W, S). Those wafers were placed on the 11/06/23, right before closing and pumping of the tower. Hence they experienced 1 pumping/venting cycle and a total of 18 days in the closed tower.

Attached there are the histograms of the particles found on the surface of each wafer. Our camera apparatus is sensitive to particles larger than 8um and counts of particles with diameter D<25um are understimated because of low sensitivity. On the samples there were found:

• N: n=31 particles over (particle count density of  )
• N: n=13 particles over (particle count density of  )
• N: n=6 particles over (particle count density of  )

The largest object found is on the NE_290623_N wafer (see in fig.largest_obj_NE_290623_N.png). It is approximalely 0.5mm in length. In the histogram is the bin at D=150um diameter. This is because the particle detection algorithm reconstruct the diameter of a dust particle from its area assuming it is round in shape.

This is a summary of the intervention performed on the SPRB bench during thursday afternoon and friday morning shifts.

The goal of the activity was to investigate the astigmatism of the B4 beam, in particular by reducing the tilt applied on the L2 lens. The short and preliminary summary is that reducing the tilt of the lens did not have any significant effect on the astigmatism of the B4 beam observed when the bench was suspended back under vacuum. However we noticed that the B4 beam image on the camera was quite round when the bench was suspended in air, thus the astigmatism seems to appear during the transtion from air to vacuum.

The venting ot the SPRB minitower started around 13h15 utc on 29/09. The minitower was opened around 14h40 utc.

First we blocked the bench and tried to measure the initial tilt of the L2 lens (see pictures 1, 2, 3). We estimated this tilt to be of the order of 10 deg.

Then we replaced the L2 lens by a temporary HR mirror and we adjusted the tilt of the mount to try to target an angle of 3 deg (picture 4).

We displaced the glass beam dump which was initially masking the edge of the mirror M1 for the ghost beam reflected by the L2 lens. We installed the beam dump near the optical path between the mirrors MMT_M3 and M1, as shown in the attached Optocad drawing. Actually the beam dump was advanced by ~10 mm closer to the M1 mirror with respect to the drawing.

We then rebalanced the bench and Alessandro adjusted the bench position using the stepping motors, until he succeeded to close the position loops.

At this point we checked the position of the beams reflected by the temporary HR mirror, using the ITF in NI single bounce. We adjusted the tilt of the mount in order to have the B4 beam reaching the beam dump at the expected position (about 10 mm from the edge of the beam dump). However we noticed that the second "B4 beam" (the one reflected by the other side of the POP and lower than the main B4 beam) was close to be clipped on the mount of M1. Thus we installed another beam dumps to mask the bottom edge of the M1 mirror and dump the second "B4 beam". The added beam dump had a weight of 177 g (including support, column and clamp) that we compensated by removing the same weight of ballasts mass.

The Picture 5 shows the situation on the bench before the modifications done during the shift. Pictures 6 and 7 shows the situation after the reshuffling of the beam dumps.

Finally we installed back the L2 lens in place of the temporary HR mirror and rebalanced the bench.

With the single bounce beam and the bench suspended and controlled in air, we started the realignment of the beam on the SPRB sensors. Noticing that the beam was partially clipped on the mount of Mmot1, we decide to act manually on the M1 mirror to realign the beam horizontally and, to a less extend, vertically, first trying to center the beam in the hole of the SPRB diaphragm, then trying to center the beam on the B4 cameras. This allowed us to have the beam also reaching the B4 photodiodes.

We then adjusted with picomotors the position of Mmot2 and Mmot1 in order to center the beam on the quadrants.

Finally we adjusted the centering of the beam on the B4 photodiodes using the picomotors of Mmot3 and we checked that the beam was still aligned on the B4 phase camera: it was still passing through the two diaphragmes on EPRB set as reference yesterday (https://logbook.virgo-gw.eu/virgo/?r=57238) and the images of the PC seemed to be centered. We then closed the minitower.

We started the pumping of the SPRB minitower around 11h05 utc on 30/09. The evacuation lasted for about 1h.

Then the position of the suspension under vacuum was recovered by Alessandro, and it was possible to close all the controls of the bench.

Once the bench controlled we adjusted again the angular position of SPRB in order to have the beam centered on the photodiodes, the phase camera, the quadrants and the cameras.

An interesting observation is that the beam shape has changed between the time when the bench was in air and the time when the bench was back under vacuum.

Picture 8 shows the image of the B4 beam with the bench suspended in air. And Picture 9 shows the image of the B4 beam with the bench suspended in vacuum. In both case we are looking at the NI single bounce. This needs further investigations.

Side note: while swapping between the L2 lens and the temporary HR mirror, we noticed that the side screw of the L2 mount is misbehaving. The white part at the edge of the screw (Picture 10) seems to be deformed and not rigidly attached to the screw any more.

After having revover a normal behavior for EIB, we checked the signals of the injection. It appeared that the beam was quite low on the IBJM quadrant, also low on B1p. This makes sense since we had to lower it to recover the alignment.
We brought the EIB Tx set point back to -600 urad as it was before the noise could appear (it had been set to -525 urad on tuesday).

We slightly adjusted the setpoint from -600 urad to -596 urad to recover the same position on the IBJM_v quadrant. By taking a closer look now it seems that we can still improve it (figure 1). We may want to do it in the coming days.

The IMC followed and went back to the position it had before the noise could appear.
The excess of noise that could be seen in the last two days disappeared. See RFC signals before the noise on EIB and now (figure 2)

With the CITF locked and the SPRB under drift control (using the B4_QD2 112 MHz signals, see Fig.1), we checked the centering of the EPRB diaphragms. We adjusted very finely their position and close the diaphragms at minimum aperture to have a good reference for the PC alignment during the afternoon shift.

When this check was done, the SPRB angular position was the following : TX = 152 urad, TY = -655.5 urad.

Then we put the ITF in NI single bounce, and we checked that for the same angular position given by the SPRB LVDT, the B4 beam was still passing through the EPRB diaphragms.

In addition we recorded a picture of the two B4 cameras in single bounce: see attached file.

For B4_Cam1, the gaussian fit parameters are: wx = 334 um and wy = 705 um.

The goal of the shift was to check the WI DAS setting for the compensation of the CH @ 93 mW (needed to mimic the YAG at CARM NULL) with the HWS-INJ.

Before the shift, the HWS-INJ illumination was checked but no beam could be seen, as happened last time (entry 56150). Camilla and Piernicola went to EIB to fix the butterfly basis of the SLED beam and the beam reappeared on the sensor.

At 15.12 UTC a west arm scan with the green laser to measure the cold matching (used as a reference) was performed. The arms were locked on IR and drift control on the arms was activated a little bit later with respect to the measurement beginning (15.42.25 UTC) and no effect was visible on the HWS wavefronts and curvature.

At 15.20 UTC HWS-INJ acquisition started and at 15.24 UTC WI CH was switched on @ 93 mW. When the curvature reached the steady state, at 16.41 UTC I switched on the WI DAS IN @ 400 mW and OUT @ 1.1 mW, the curvature wasn’t reaching the zero, as already seen in the entry 56150.

The entire trend of the curvature and the matching measurement are shown in figure 1 and in figure 2, respectively.

At 21.20 the initial TCS powers CH @ 120 mW, DAS IN @ 460 mW and OUT @ 1015 mW were set.

The aim of this shift was to recover the centering of the HD PD1 and PD2 wrt the beam, after the shift of yesterday (entry 54809). The idea is to scan the PD sensitive area by deflecting the SC beam with M4 and then move the PDs with their motor. 

SHIFT summary: 

• We confirmed that the lateral peaks observed with the LO when moving the motors of the PD2 along the x-direction are also visible by scanning with the SC through M4. The peaks are visible only on PD2 (ie not if we scan PD1) and only along x. Furthermore we do not see the peaks with the cameras. This suggests that the peaks are due to the lens in front of PD2. The beam width is different at the two PDs. 

• The M4 actuators seem to work reliably with the ramp (just one issue with fixed bias). We observe that M4 ramp amplitude can have sign and adds to the bias (ie it is not centered on the bias)

WARNING: we left PD2 not well centered in x wrt the beam.

2022-02-10 09:26 UTC
First of all we aligned the SC returning back to EQB1 with the LO by overlapping their spots on the homodyne cameras using mainly the pico motors of the SQB1 mirror M13, and minor adjustments to the bias of the M4 mirror For the first alignment we moved the pico motors by:

V: 1650 steps back
H: 1350 steps back

2022-02-10 10:05 UTC
We removed the HD_MIR and we shuttered the SC. 
We checked that the LO is aligned with the CC by looking at the magnitude of HD_DIFF_RF_4MHz channel: the maximum was 2.5mV. So we can consider SC, LO and CC to be all aligned together
Before starting the scan we changed the bias of M4_Y from 0 to 0.7 and we came back in order to check the reproducibility.
In this operation we lost the alignment. 

2022-02-10 10:50 UTC
We inserted the HD_MIR and we recovered a first alignment between the LO and the SC using the camera. We set the M4_Y bias=-3.7.

2022-02-10 10:54 UTC
Removed the HD_MIR and shuttered the LO we checked the HD_DIFF_RF_4MHz: maximum at 0.5 mV.
We fine aligned with different M4 bias values:

X=1
Y=-4.3

We recovered the maximum value of 2.5mV.

2022-02-10 11:35 UTC
We closed the LO shutter and made use of the SC beam only. At this stage we observe PD1 only: we blind PD2 by inserting HD_MIR so to track the beam deflection with the HD cameras. We use the HD camera fit parameters to track the beam position. We start with SC beam positions in the FF and NF cameras at: FF (x,y)=(-35, +12), NF (x,y)=(+145, -30). 

With the M4 mirror we can change the angle of incidence of the beam on the f=50mm lens before each PDs inside the homodyne box. This produces a shift of the beam along the focal plane (where the PDs active area is located). The relationship between the angle of incidence and the movement is: d~f*theta*dir 
where theta is 100 urad for 1V of input. dir is a factor that takes in account the scan direction: along x dir=2, along y dir=sqrt(2).
We feed a triangular ramp (freq 0.1Hz, ampl 5V, bias 1V) to M4_X to scan the PD active area. 
The beam moved ~50um. Along x the conversion is ~10 um/V. We observe that the position of the beam center in the cameras follows the rampo but that the x and y degrees of freedom are entangled (Fig 1)

2022-02-10 11:47 UTC
We removed the HD_MIR to check the centering of the PD2.
PD2 was centered.

2022-02-10 11:59 UTC
We inserted HD_MIR to double check with the cameras the centering of PD1.
We scanned the entire PD1 area with a triangular ramp to M4_X: freq 0.1 ampl 18, bias -9
In this case the beam moved ~ 180um
Moving PD1_X we centered the PD.

2022-02-10 13:27 UTC
Removed  HD_MIR we repeated the same scan on PD2 in order to check if there are the same two side peaks observed with the LO.
M4_X ramp values: freq 0.1 ampl 18, bias -9
We observe the side peaks also with the SC on PD2 (we checked only one side) but not on PD1 (in none of the two sides as shown in Fig 2). Note that in PD1 the scan can cover the entire plateau while in PD2 we can see only one side of the shoulder.

2022-02-10 13:41 UTC
We also scanned the M4_Y with the same ramp but we didn’t see the side peaks.
M4_Y ramp values: freq 0.1 ampl 18, bias -9
These values correspond to a total displacement of 125um. Along y the conversion is ~10 um/V.
No side peaks (Fig 3)

2022-02-10 14:02 UTC
To be sure we checked the Y direction of PD1 with both the DC signal and the cameras.
We offset the PD1at 4350 to have the scan centered on the shoulder
No side peaks (Fig 4)

The goal of the shift was to investigate if it is possible to reduce the coupling of the stray light by moving the homodyne PDs far from the minimum waist where they are estimated to be located. In particular we aim at reducing the bump below about 10Hz seen in the HD_diff_audio spectrum. We have the LO arriving at the homodyne PDs, EQB1 delay line disengaged, SQB1 back reflector disengaged and SQZ beam reflecting back from the locked filter cavity.

It was not possible to perform a scan along z due to the non-repeatability of the steps taken by the PDs motors. However, a single shift along z by about one Raylight range was attempted. No difference in bump behavior was observed at 10Hz (please note however that we don’t know the starting position in relation to the waist, so the change in beam characteristics may have been small).

Other main results obtained during the shift are: 

• calibration of the motor's steps along x (positive and negative directions) for both PDs. 

• discovery of two side peaks on the beam incident on PD2.

PD motors scans and centering

2022-02-09 15:27 UTC: We start by scanning the x direction of HD PD1: if one could trust the motor step size to be constant, we could hence estimate the waist size and the x-length of the sensor. We did a first scan by hand. 

Starting position:

Manual scan:

We notice that there is a difference in the response of the actuators while moving along the positive and negative x-direction. 

2022-02-09 19:45 UTC
Then we made use of an ad-hoc script that allows for automated scans of the PDs along one direction. 

PD2_X  positive auto scan: 700 steps total, in step of 10 every 2 seconds

PD2_X  negative auto scan: 2300 steps total, in step of 10 every 2 seconds

This way we could estimate that the step size ratio moving in different directions is negative/positive = 2.33. After the scan we recentered the PD wrt the LO beam.
Fig1 shows the different behavior between the two directions. In particular it is also visible the presence of two symmetric side peaks around the central plateau.

2022-02- 09 20:28 UTC
We repeated the same scan along x with PD1. 

PD1_X  positive auto scan: 4000 steps total, in step of 10 every 2 seconds

PD1_X  negative auto scan: 7000 steps total, in step of 10 every 2 seconds

PD1_X  positive auto scan: 1000 steps total, in step of 10 every 2 seconds

We obtained that the step size moving along negative/positive = 2.23. 
Fig 2 shows the behavior in the positive and negative directions.
When moving the PD1 towards negative direction we observe an asymmetry in the PD dc output between when the beam enters the PD wrt when it exits: probably due to non uniformity of PD active area and/or to not symmetric beam. From the positive scan we can measure a waist=15um.

2022-02- 09 22:09 UTC
We recentered both the PDs on the beam. We optimized the dc ratio  PD1_HD/PD2_HD=0.992. This was the starting value before the shift.
We took a few minutes in quiet condition as a reference for the stray light bump at 10Hz.

Attempt at reducing stray light bump

2022-02- 09 22:09 UTC
We made an attempt to move PD1 along Z, to see if the stray light effect can be reduced.  We started with PD1 in position z=0 and we made +700 steps, which should correspond to a Raylight range for the 15um beam waist (assuming about 1um/step). After the steps in Z, the power on PD1 was almost lost. DC output went to about 0.3V from 0.77V and the plateau was about 1400 steps along the negative direction. During this scan we arrived at x=-22300 and we noticed that the motor gradually stopped responding as we moved towards the negative direction. The motor responded only in positive.

2022-02- 09 22:43 UTC 
We moved x and y to recenter the PD. 

PD1_X positive auto scan: 40 steps (we were stuck on the waist shoulder moving in negative direction)

PD1_Y positive auto scan: 150 steps

One possible explanation for the larrge correlation between z and y displacement is that the beam goes from the lens to the PD with a slope (assuming 1um/step, theta=15 degrees)

2022-02- 09 23:01 
With PD1 finally centered and the ratio PD1_HD/PD2_HD=0.991, we took data to check if the bump at 10Hz was changed. We did not observe relevant differences with respect to the reference (fig 3).

2022-02- 09 23:13
We moved the PD1 back by -700steps, nominally reaching the original z=0 position. We expect the two directions to have a factor 2 difference as for X and Y so likely the original position is not reached.  We recentered the PD along x and y obtaining again PD1_HD/PD2_HD=0.991. 
Fig 4 is the spectrum of HD_DIFF_AUDIO at the end of the shift

During the shift we moved PD1 and PD2 of the homodyne along x.

We moved PD1 also along z (+700 steps), we recentered in x and y and than moved back z (-700 steps).

We leave PD1 and PD2 centered with PD1_DC/PD2_DC=0.991.

We will report more details in a future entry.

We post the plots relating to the different tests.

In the figure 1 we can see how the air affect the system.
The purple line represent the system in the following configuration: air off, EQB1 box closed, light on.
In the blue line the air was switched on, EQB1 box closed, light on.
We note how the air excites different structures above the 100Hz and how the low frequecies noise (<10Hz) disappeares when the air is switched off.

In the figure 2 the purple line is: air off, EQB1 box closed, light on.
In the blue curve the air was off, the box was opened in the sec 17, 18 and 19. The lights were on only above the sec 17.
Even with the air off, opening some sections, the noise at low frequecies reappears.

1. The figure 3 confirms that tapping horizontally on table side facing SQB1 the peak at 43 Hz gets excited (the purple reference curve is the same of the previous plots: air off, EQB1 box closed, light on).

For the following configurations the reference purple curves represent the system with: air off, open sec 17->19, lights on (only above sec 17)

2. The analysis on the tests with a light dumper confirm that we don't see effect on the 43Hz (no figures are attached) 

3. The analysis of the "mechanical dumping" tests confirm the two structures around 110 and 190 Hz:

• in figure 4 the blue curve shows how removing the HD lid a peak at 110 Hz is excited.
• in figure 5 pressing the HD BS mount there is a peak around 190 Hz.

5. Figure 6 confirms how the 43 Hz noise is clearly excited when tapping on LO_M2. 

The white curtains that separate EQB1 have hairs at their edges, as shown in attached pictures. These hairs are found everywhere in the lab near EQB1.

We continued the work on installation of dumps to block the residual transmission of HR mirrors on EQB1 bench (Logbook Entry 52657). We installed a dump in transmission of each of the last two mirrors on the Green line which were left behind from the previous shift: EQB1_GAOM and EQB1_GM4.

We found that the head of the screw fixing the Polaris mount to the pillar protrudes out of the Polaris mount, hence we could not fix the dump mount as foreseen; we workaround this problem by mounting it rotated by 180° hence with the screw on the bottom part of the mount.

We have not yet installed additional dumps in the IR paths, which were left behind from the previous shift. In addition, we discovered two more mirrors in the Delay Line path wrt the previous shift and also wrt the latest EQB1 design (dated 21 Sept 2021): we need the prepare two more dump mounts to be able to finish this work.

We set out to investigate the origin of the line(s) around 40 Hz that shows up in the homodyne detector differential channel. After a number of tests, we concluded that the element vibrating at that frequency is the LO_M2 actuated mirror (on the LO line just before entering the HD). The two motors sticking out of the back are probably responsible for lowering the resonant frequency, which appears to be instead in the range of 100-200 Hz for the other mounts. We have not yet investigated the complete coupling mechanism (i.e. where the stray light originates and recouples).

In the following, we report the full list of tests (to be more thouroughly analized later).

We started by closing the shutters of Green, SC, and BAB (only LO to the Homodyne), plus we put the Delay Line on in order to have the LO light circulating only on EQB1 bench.

We first looked at the spectra of EQB1_HD_DIFF_AUDIO with the bench in different environmental conditions (box, HVAC, lights)
One minute stretches of data (first number is GPS start time, data good for at least 60 s):

1321697970 air off, EQB1 box closed, light on
1321698139 air on, EQB1 box closed, light on
1321699508 air off, open (section 18 - above galvo), light on (above sec 17)
1321699625 air off, open (section 18 - above galvo, light off
1321700020 air off, open sec 17->19, lights on (only above sec 17)

We note that the bump below 10 Hz disappears when HVAC is off and the box closed, while it partially survives even with HVAC off if the box is open (plots will follow).

We then investigated the origin of the approx. 43 Hz line in the EQB1_HD_DIFF_AUDIO channel. All GPS start times are followed by  minute of test data.

1. GPS 1321698379: Tapping horizontally on table side facing SQB1, in a point close to the green rotator (and air off, EQB1 box closed, light on): 43 Hz gets excited.

For the subsequent tests, we kept air off, open sec 17->19, lights.

2. Tests with a light dumper:
1321700314 put an rectangular glass IR dumper on HD box out port towards cameras
1321700431 removed the dumper, as a double check
1321700568 put IR dumper on HD box SQZ input port
1321704720 IR dump on HD SQZ input, inside the HD box before the BS (by hand, also touching box and table)

We didn't see any significant effect on the 40 Hz noise.

3. " mechanical dumping" tests (press a finger on top of the optics mount and seeing if the 40 Hz noise is dumped):
1321701487 finger pressing on the waveplate mount at HD box SQZ input 
1321701856 HD lid removed
1321701981 hands on HD box walls (wall at SQZ input and its opposite wall)
1321702097 hands on HD box walls (wall at LO input and its opposite wall)
1321702306 finger pressing on HD BS mount (top)
1321702500 finger light touching on HD BS mount (top)
1321703199 finger pressing on mirror mount on translation stage inside HD box
1321703381 finger light tapping on mirror mount on translation stage inside HD box

Again, we see no clear effect on the 40 Hz noise.

5. Tapping (by using Allen key gently touching the optic mount) tests:
1321703544 on HD BS
1321703656 on HD SQZ input Lambda mount
1321704349 on HD lens in reflection from LO on the BS
1321704463 on HD lens in transmission from LO on the BS
1321705568 tool tapping on actuated mirror LO_M2 before HD

These tests showed that the 40 Hz noise is clearly excited when tapping on LO_M2, while tapping on other optics mount excites structures above 100 Hz.

Triggered by the difficulty of manually setting and blocking the camera focus, I have made the following mechanical upgrades to the imaging facility in the "Perugia Cleanroom":

• Replaced the contraption holding the camera with a custom plate (thanks to Alessio Buggiani for the quick turnouround on the request), which makes the mounting cleaner and a lot more stable.
• Inserted a Thorlabs L200/M lab jack (borrowed by the 1500W lab; thanks Fiodor) under the tip/tilt stage (removing the big washer that was there to prevent the knobs from hitting the breadboard, and is no longer necessary
• Readjusted the position of the two big vertical posts so that the camera is fairly centered on the sample holder (note that the "vertical" of the camera is ajusted manually, as the transverse post that holds it is free to rotate around its axis)
• centered the illuminating ring, and adjusted its height so as it should be quite similar to what it was before wrt the sample holder (I dind't have a precise reference; it can be further adjusted if needed). On the occasion, I also tried to space apart the posts holding the ring as much as possible, so as to make it easier to access the sample holder.

Beside obtaining a general tide-up and increase in stability, the idea now is to keep the camera focus fixed on the closest position, and adjust the lab jack height to move the position of the sample until it is in focus. NOTE: a 1/8" (or 3 mm) allen key (or even beter and hex driver) is needed to adjust the jack height: we should procure one to keep with the facility.

This modifications should allow for a faster and more precise focusing (please note that the jack seems to have a quite obvious backlash, but it is not a big issue once you get used to it).

Picture of the upgraded setup are attached.

This is a summary of the shifts performed on Jul 21 and 22 to install absorbing glasses on the rear part of EQB1 mirrors, in order to block the residual transmitted beams (ghost beams).

The glass support is half-circle designed to be attached on the mirror mount not to affect the alignment and it perfectly fits on Thorlabs-POLARIS-K1S5 (for 1" mirrors) and Thorlabs-POLARIS-K2 (for 2" mirrors) mounts (Figure 1).
The glass is installed inside its mount using a custom 3D printed support that ensures it is tilted by 2.5° around the vertical axis wrt the mirror (Figure 2); the direction of the tilt is chosen in such a way that subsequent residual reflections between the absorbing glass and the mirror, initially at 45 degrees for most cases, are at ever smaller angles, thus ensuring the ghost beam is estinguished on the absorbing glass instead of quickly escaping out of it and towards the metal mount.

Unfortunately, contrary to what we were told not all the mounts on EQB1 are K1S5 and K2: for this reason, we installed all the glasses on the mirrors with K1S5 and K2 mounts, and left aside for the moment the mirrors whose mounts are different.
In particular, the following mirrors are still missing of the absorbing glass:

EQB1_MT2 (2", Low Wavefront Distortion Gimbal Mirror Mount)
DL_M2 (1", Thorlabs-POLARIS-K1-2AH Mirror Mount)
DL_M5 (1", Thorlabs-POLARIS-K1-2AH Mirror Mount)
DL_M6 (1", Newport Clear Edge Mirror Mount)
EQB1_GAOM (1" Thorlabs-POLARIS-K1-2AH Mirror Mount)
EQB1_GM4 (1" Thorlabs-POLARIS-K1-2AH Mirror Mount)

We will design appropriate fixtures to install the absorbing glasses also on these mounts.

During the shift, we noticed the reflected beam from the Green Rotator was ending up on the EQB1 box, so that we temporary blocked it with a beam block (Figure 3).

Inspired by Michal's idea and after a some additional exchange with him and Marco V, I produced a similar script to Michal's, with some corrections and enhancements:

• the sidebands Michal considered initially actually belong to groups associated to carriers which are far apart in frequency (ITF, subcarrier and green). So beat between sidebands of different carriers can be neglected and excluded from the list.
• the selection criteria for acceptable sidebands should be two:
  • No beat note should be produced in the ITF band (up to some total sideband's order, like in Michal's script). Let's assume f >=2 kHz
  • No beat note shoudl be too close to a possible demodulation frequency (indentified as an harmonic of a sb), not to bother control loops. How much it depends on the loop. but I'm told that 50 kHz minimum distance should be safe.

The attahced python (v3) script:

• only accounts for beatnotes between sidebands referring to the same group
• plots all beatnotes and their order (in a quite messy but hopefully informative plot, if one is patient enough to navigate it)
• lists all beatnotes <2 kHz (none found)
• list all beatnotes closer than 50 kHz from a possible demodulationfrequency (there are a couple, all originating from the ITF carrier sidebands)

Max sideband harmonic to consider for beatnote generation, max total order of a beatnote (sum of the orders of the beating sb), maximum sb harmonic to be cosnidered a possible demodulation and minimum spacing of beatnotes from demodulation frequencies are all configurable. 

This is a summary of the shifts performed on Apr 28 and 29 to inspect and dump ghost beams (GB) on EQB1.

The starting point is VIR-0329B-21, which indicates all the wedges as installed, except for LO_M1 (installed with opposite wedge wrt the document). These are GB resulting from the main beam being transmitted by the HR face, reflected by the AR face and then transmitted again by the HR face: they are expected to be of order 1e-14 of the impinging power, and thus undetectable. We thus place beam dumps based solely on the calculated positions, taking care to leave sufficient clearance for the main beams (which of course are visible).

Additionally, we used beam profiler, power meter and IR-visor to inspect for additional unforeseen GBs. We found two GB exiting from the squeezer, on the BAB and GREEN lines (see details below).

Details follow, divided by beam line (pending actions at the end).

Some general considerations and lesson learned during the shift

• using the power meter we can identify GBs at the level of a few nW
• using the Ophir beam profiler, we can detect GBs of no less than a few hundred nW
• by measuring the BAB+SC transmitted through EQB1_M1 we were able to have an order-of-magnitude confirmation of the performance of the HR of the mirror: we detected a transmitted power of ~3 nW with about ~1 mW of BAB power on the mirror, and 11 nW with the addition of 2.5 mW SC power on the mirror. In both cases T ~ 3e-6 (vs 1e-5 nominal).
• At this time, we were not prepared to dump these transmitted GB. However, we plan to design beam dumps to be placed directly behind the mirrors for this purpose.
• In some situations it is more convenient to install the glass dumpers (which are 45 mm x 30 mm) horizontally on their post rather than vertically. However, in this configuration the beam hits the dumper quite close to the upper edge, and it would be safer to raise those posts by 5 mm.

Green line

This line still has some temporary optic (most notably a cube PBS) that substantially increase the scattered light and the number of GB (see IMG_20210429_171850). In addition, we don't have beam dump with the right coating yet. We decided to work on this line after the final optics will be installed and the beam dumps are available. A few notes to ourselves:

• We identified a GB a few mm below the green beam, leaving the squeezer from the same port (see 20210429_224937.jpg). The GB is visible, but we didn't measure its power, nor we attempted to dump it for now (we will do when we work on this line)
• The forward beam reflected by the BS (rather than transmitted towards the AOM) is directed to the side of the squeezer box with no dedicated beam dump. We plan to installa  dedicated beam dump in the future.
• GB in reflection from the wave plates present on the line are clearly visible and will need to be blocked

SC up to the FI

Note that we were unable to increase the power of the subcarrier (SC) to several tens of mW to increase GB visibility. In fact, we could not exceed about 3 mW (see entry 51609). In addition, before inspecting this line, we verified that the SC was well aligned with the BAB using bot the card and the images on the diagnostic cameras on EQB1. 

• based solely on simulations, we placed a beam dump close to the FI to block the GB from SC_M1 (see IMG_20210429_231621.jpg and IMG_20210429_231627.jpg)

SC towards OPO

Simulations indicated that there is no space to safely block GB without risk on clipping for the main beam, so we did not perform any action

SC+BAB from OPO towards SQB1

• We identified a GB of about 10 nW leaving the squeezer BAB port at about 15 degrees from the main beam line. We dumped it in the proximity of EQB1_M1. (see 20210428_205229.png and 20210428_212103.png). Unfortunately, it didn't occur to us to check if this beam originated fromthe BAB or the SC entering the squeezer.
• Based on simulations, we placed beam dumps:
  • just before the FI, right of the beam (see again IMG_20210429_231621.jpg and IMG_20210429_231627.jpg)
  • close to EQB1_M3 (see IMG_20210429_221209.jpg)
• We decided to reconsider placement of a beam dump close to EQB1_MT2 (to dump GB form EQB1_MT1); the distace between GB and main beam is not very big comared to the size of the beams, and placing it without a precise refernece seemed at risk of either clipping the main beam or placing the beam-dump edge clos to the cneter of the GB. We will urther assess this and decide for an action.
• We could not place foreseen beam dumps close to the viewport to SQB1 (left and right of the beam) since the space is all occupied by a temporary delay line and a temporary reference iris that we were asked not to remove at this time.

Local oscillator

We did not place any beam dump on this line for different reasons:

• damping the LO_M1 GB seems a bit risky from the point of view of clipping the main beam
• the GB coming form the LO_PBS has two problems: the temporary delay line restricts the available space, and the two lenses on the LO line are in significantly different positions wrt the simulations.

Pending actions (summary)

• Design and install beam dumps behind each superpolished mirror to intercept the transmitted beam.
• Design a solution to raise by ~5 mm beam dumps whose glass has been installed horizontally
• Work on green line form scratch. In particular, in addition to dumping simulated beams:
  • procure dedicate beam dump for PBS reject beam
  • dump GB out of the squeezer
  • dump GB in reflection of WP
• Reconsider dumper for EQB1_MT1 GB
• Place dumpers close to viewport for BAB line, once temporary iris is removed
• Need to reassess LO line (new lens positions)

During the ghost beam shift, we wanted to increase SC power on the bench to increase the visibility of ghost beams. However, by rotating the waveplate of the power control right after the output collimator of the fiber, we could not raise the power above about 3 mW, which seems to indicate an abnormally high power loss from the laser head to the collimator output (it was expected to be able to reach a few tens of mW). The laser injection current (that we did not change) was set to about 1.3 A, which should be nominal and close to maximum output power (500 mW), although this needs to be confirmed.

As part of the effort to characterize stray light due to dust particle contamination in the SQZ subsystem, last week we investigated the camera system located in ISO 3 Perugia Clean Room. The setup is based on a 10 MP camera with 2/3” sensor and 35 mm focal length, alongside a high-resolution lens at working distance 110 mm, equipped with an illuminating LED ring set above the table (Figure 1). The idea is to use this camera system to acquire images of the wafer witness samples placed in the SQZ environment during past months (Logbook Entries 49008, 49870, 50211).

Time has been dedicated to the re-alignment of the setup, the inspection of the parameters (camera aperture, exposure time above all) and the optimal height of the illuminating ring to get the most uniform background on the acquired image. Tested have been performed on a spare 2" wafer. 

Since the luminosity of pixels display slight random fluctuations, some background pixels may appear particle-like during image analysis. For this reason, we acquired 30 nominally identical images with maximum exposure time and aperture 8, and 30 nominally identical images with maximum exposure time and aperture 16. After opening them as stacks on ImageJ, we combined them by assigning each pixel its minimum value in the stack. A preliminary analysis can be performed on ImageJ, by adjusting the threshold below which pixels are considered part of the background and thus removed from the images. The cleaned images thereby obtained are displayed in Figure 2 (aperture 8) and Figure 3 (aperture 16): one can see that aperture 8 provides more particles than aperture 16.

In order to monitor dust particle contamination inside the Clean Room, a new silicon wafer is horizontally placed inside the room (Figure 4). Please don't disturb it.

Last week we removed the silicon witness samples for dust particle contamination monitoring placed on ESQB1 (Figure 1) during early November (Logbook Entry 49870); unfortunately, the horizontal wafer placed in the center of the bench is broken (Figure 2), so there are 4 wafers that will be analyzed with the camera system in ISO 3 Perugia Clean Room.

We placed 5 new samples, approximately as in the same positions as the previous ones (Figure 3):

- On top of AEI squeezer box, horizontal
- On ESB1 bench, near AEI squeezer box, horizontal
- On ESB1 bench, near AEI squeezer box, vertical
- On ESB1 bench, approximately in the middle of the rectangular part of ESQB1, horizontal
- On ESB1 bench, approximately in the middle of the rectangular part of ESQB1, vertical

In addition, we placed a horizontal wafer inside the external Detection Room, on top of threshold of tower structure, behind detection bench minitower (Figure 4)

Horizontal witness samples have been left in their open boxes, vertical witness samples have been mounted on a vertical support with two screws.
Each sample has the clean face exposed and the box lid facing down next to it (ready to be closed and transported away).
Two additional wafers (one horizontal and one vertical) have been left exposed on the bench for 1' and removed, in order to monitor the dust fallout due to handling.

Please do not move the samples and report any activity that may need to be performed in the close proximity of a sample and may compromise its representativity of the average cleanliness condition of its location.

As part of the effort to characterize stray light due to dust particle contamination (especially in the SQZ subsystem), on Thursday, June 11 we placed a series of silicon witness samples (kindly provided by Antonio P) in several positions around the DetLab and the AEI squeezer box. In particular:

• ESQB table, inside plexiglass box, space between the FI train and the telescope: one 2" and one 3" wafers placed horizontally, plus one 2" wafer in vertical
• ESQB table, on top of plexiglass box: one 3" horizontal
• Next to particle counter: one 3" horizontal
• On top of external detection bench: one 3" horizontal
• On top of threshold of tower structure, behind detection bench minitower: one 3" horizontal

Each witness sample, with the exception of the vertical one, has been left in its (open) box, clean face exposed, with its box lid facing down next to it (ready to be closed and transported away). The idea is to remove and analyze them in about 1 month time (with the exeption of one or two control samples to be removed earlier).

Please do not move the samples and report any activity that may need to be performed in the close proximity of a sample and may compromise its representativity of the average cleanliness condition of its location.

Attached PDF shows more precisely the location of the samples.

Considering that the ITF conditions were not stable since some minutes before the start of the test, my impression is that no definitive conclusion can be drawn from this test and it has to be repeated (maybe, as suggested, following a different on/off sequence).

Nevertheless, here some plots showing the behaviour of the sensitivity curve during the test:

• Figure 1 is a spectrogram with indicated the selected times (all ON, all OFF, only HALL on, only DET on)
• Figure 2 is a zoom, showing that around 13:36 a strange spectral noise onset in Hrec (see also Figure ...). Apparently this noise engaged AFTER the DET HVAC was switched OFF
• Figure 3 compares Hrec sensitivity right Before and when ALL CEB devices were Off: there is a clear reduction up to 100Hz  (these are the 3Mpc step)
• Figure 4 is a similar comparison of Hrec Before (purple) and when CEB HALL is ON (blue): no change with respect to the all off (brown)
• Figure 5 is similar, but now the blue curve is with DET airconditioner ON: noise seems to go back to the level BEFORE
• Figure 6 shows the strange noise that onsets right after DET air conditioner was turned off

Indications are that DET air conditioner is the most impacting (scattering?). The Hall air conditioner seems not contributing. The effect of the INJ air conditioner could not be tested.

It seems important to repeat the test, focussing on DET (and INJ air conditioners).

The strange noise onset seems not directly linked to our actions... it has to be further investigated.

A few operations done during this maintenance:

• set up the big shaker on the floor underneath the NI cryotrap (Picture 1) ready for injections (connected to DAC channel ENV_NOISE_CEB_TCSroom).
• verified that the accelerometer (meggit model) named ENV_NI_CT_ACC_Z is attached to the rod connected to the LN2 tank (Picture 2)
• reconnected the (PCB model) accelerometer attached to the NI cryotrap external belly, renamed ENV_NI_CT_EXT_ACC_Z (Picture 3) cabled to ENV ADC in TCS room. See spectra in Figure 4
• at WEB temporary installation of one meggit accelerometer sitting on the  vacuum chamber flange at first floor, named ENV_WE_ACC_CHAMBER_Y (sorry no picture)