[Yamamoto, Yuzurihara]
I updated the summary page of Omicron trigger. Especially, I adjusted the limit of each axis.
On the weekend, the Omicron trigger generation was stopped accidentaly. The cause was related to cache file. In this afternoon, Yamamoto-san fixed this trouble and he restarted the job.
[Tomaru, Ushiba]
Today, we checked the cables of MNH2 photosensors and found that one cable doesn't seem to be connected well (When we touch that cable, the output fluctuates a lot).
Other than that cable, we found no strange behaviour, so we decided to repair that cable and see noise level is recoverred.
We thought we repaired it inside the cryostat but found that it is difficult due to the accessibility, darkness, and so on.
So, we decided to take out the photosensor and repaired the cable.
Cable repairement was finished but not yet reinstalled.
We will do it tomorrow morning.
Note: Since the repairement was not et finished, we locked the suspension to the secrity flame.
So, please do not shake and control the suspension tonight.
With Hirata-san and Ikeda-san.
See pictures in the album BS Remedying work after O3.
We added some weight by replacing some ballast masses:
In the transfer functions the resonant frequencies are 63 and (around) 92 mHz, measured with a resolution of 5 mHz. As before, L and T look strongly coupled.
With Takahashi-san and Hirata-san.
See pictures in the album BS Remedying work after O3.
They redisrtributed the ballast mass on top of the IP uniformly. In order to do this we had to remove the outer end of the magic wand, which was on the way of the arch weights. We need to shorten such a piece by a small amount in order to put it back. Afrer redistribution, some studs, nuts and washers remained unused. The mass that was removed was:
In the afternoon we measured transfer functions.
We'll rearrange the masses to try to reduce the difference between those frequencies.
Rotating the PR3 Traverser Motor in the Yaw Direction
At the request of miyo-san, I rotated the Traverser motor of PR3 in the Yaw direction.
I used cont_serial_py3.py to run the Traverser motor.
First, to check the direction of rotation, we moved the motor drive to +1[mdeg] in the indicated value, and then moved it to -168[mdeg] in the negative direction y.
Please refer to the image for details. As a result, the count of OPLEV moved from 40000 to 0.
Akutsu, Uchiba, Ikeda, Hirata, Fabian, Miyo
After the alignment work with Akutsu-san and Ushiba-san so that the green beam hits on the center of NAB in IXA from PR2 by aligning the PR3 mirror, we offload the actuator output in BF yaw and IM_PIT by using the traverser and picomotor. After that we confirmed that PR3 can be requested to DAMPED as usually. Now I'm measuring the transferfunctions in all DOFs for PR3 to check its health. I'll post it later.
I think report about the alignment work will be posted by Akutsu-san later.
Rotating the PR3 Traverser Motor in the Yaw Direction
At the request of miyo-san, I rotated the Traverser motor of PR3 in the Yaw direction.
I used cont_serial_py3.py to run the Traverser motor.
First, to check the direction of rotation, we moved the motor drive to +1[mdeg] in the indicated value, and then moved it to -168[mdeg] in the negative direction y.
Please refer to the image for details. As a result, the count of OPLEV moved from 40000 to 0.
This is the remnant of the file that was deleted when I tried to investigate the VIS DAQ error by copying the model file to the local directory.
The source of the deletion was the deletion of the files and directories copied locally, so they are unnecessary files.
Continue work in 17137.
It wasn't easy to obtain a reasonable L-T coupling ratios at the resonances since the modes are L-T coupled.
I reverted the LVDT2EUL and EUL2COIL matrices back to the geometric one derived by Mark a long time ago and the results are pretty much the same.
But, plotting L/T and T/L "transfer functions" yielded interesting properties, i.e. the two transfer functions are mostly flat above 0.1 Hz, (roughly) equal in magnitude (~0.23) but with opposite sign (0 and 180 degrees). See Fig. 1 and 2.
These symptoms are consistent with a rotational transformation. And it makes sense as sensor cross-coupling would show the same spectrum shape to another channel.
Measuring the coupling ratios between Y and L/T are relatively straightforward but required several iterations, as L and T resonances overlap.
At the end, I arrived at an LVDTALIGN matrix as shown in Fig. 3.
I then compared the transfer functions before and after applying the LVDTALIGN matrix. These are shown in Fig. 4, 5, and 6. References plots are those with the new LVDTALIGN matrix.
As can be seen, L/T coupling receives a significant amount of broadband reduction while Y to L/T and L/T to Y has minimal coupling at resonances.
There's a caveat. There's a chance that by applying the new LVDTALIGN matrix, we rotated the sensor frame to the actuation frame, but not necessarily the true L/T.
However, this cannot be confirmed, and I am satisfied with the results so I will just leave it this way.
For a sanity check, I measured the L/T free swing spectra and compare them with the seismometer x/y spectra. They should give evidence of whether or not the sensors are well aligned to the interferometer X/Y.
The coherence between the four channels are shown in Fig. 7. As expected, around the microseism frequencies (very large microseism today, peaked at 1 um/sqrt(Hz) during measurement, good for this measurement), L is correlated to seismometer y, but not x, and T is correlated to seismometer x, but not y.
This is very good evidence that L and T are readouts are truely L and T.
Nevertheless, the resonance frequencies as measured in diaggui for the L/T modes are roughly 59.8 mHz and 66.7mHz
All measurements are custom files that are located in /kagra/Dropbox/Personal/Terrence/vis_commissioning/srm/. I put them there so I have convenient remote access.
Next steps would be to align and intercalibrate the geophones and seismometers for blending and sensor correction. I might do actuation diagonalization, but I don't think it's that important.
BTW, the OpLev signals are zeroed out. Is there an optical lever installed for SRM? This will become very important very soon as I go into control tuning, even for the IP.
Additionally, can we have a sensor correction path for the optical levers? With this path, we'd be able to measure the actual control performance of the suspension as an out-of-loop sensor.
Without that, the optical levers will be covered seismic noise and it'd be very suboptimal. It'd be difficult to utilize the optical levers for local controls when the seismic noise level is high as well.
This is the remnant of the file that was deleted when I tried to investigate the VIS DAQ error by copying the model file to the local directory.
The source of the deletion was the deletion of the files and directories copied locally, so they are unnecessary files.
Date: 2021/6/18
Members: Dan Chen, Darkhan Tuyenbayev
We added DAQ channels in calcya model:
Also, we made holes on the front panel of a circuit box in YPcal for GSK readout port.
[2021.06.11]
Sorry, very late report.
I covered the FFU part on the top of the MCe cover with aluminum foil (Fig. 1), hoping that it would lower the noise of the air turbulence. In conclusion, unfortunately, there was no particular improvement...(Fig.2)
The other thing we can do is to turn off the FFU in the MCE booth.
This is related to [log 17175].
I added the feature on Pastavi Now you can use Q-transform to visualize the trigger, which was specified by Omicron. The detailed options are under construction.
Glitch trigger is regularly running and generating trigger lists for oplevs of IMC. I updated the configuration for summary page to summarize these trigger lists. I attach an example for MCI-Yaw.(Fig.1)
At the bottom of summary page, there is a table of glitch triggers.(Fig.2) Usually there is button to link ligoDV. But, the KAGRA full data is unavailable on ligoDV. Instead, I set a link for Pastavi to see the several plots at the trigger time. The modification of gwsumm library is described in wiki page.
After that, I turned off the picomotor driver and detached the relevant cables to avoid dangers, because I had no idea about the nice proper way to lay the cables...
With these measurements and their interpretation, I think the following can be also granteed; the green beam would not hit at the inner wall of the vacuum tube between PR2 and PR3.
Cont'd from 17154.
Relocated the periscope on the POP table so that PR2_TRANS_FORWARD will be picked off with another periscope in the future. Send the green beam from the relocated persicope via POP-POM through the PR2 center to the PR3 center.
Of the POP cover, there was yet another black wall remaining at the side of the PR2 chamber. I firstly dismantled this wall to widen the workspace (Figs. 1, 2, and 3). To access the periscope or relocate it, this wall must be detached definitely.
About the motivation for the periscope relocation, read the previous report (17154). The first and the most important step is to preserve an optical axis that I want to recover after the relocation. Indeed, I want to recover the optical axis that connects the upper periscope mirror via the viewport window to POP-POM (and to the center of the PR2 mirror as well). In the real work, I tried to preserve the optical axis from the upper periscope mirror to the viewport window. Let me say that differently; I want to realize this optical axis from the upstream optical axis; "the upstream optical axis" means the one starts from HBS and go in the direction that is determined by the incident angle to the HBS and the HBS's tilt angle. These two parameters, i.e. the incident angle to the HBS and the HBS's tilt angle must not be touched this time, as I'd like to keep this optical axis while relocating the periscope. In the first look, it seemed the two optical axes that I wanted to connect with the persicope to be relocated were twisted each other (nejire no ichi), so one could have estimated this work would be hard.
As considered in 17154, I put an iris diaphragm in front of the upper periscope mirror and match the hole to the green beam (Figs. 4 and 5). Also, I put a cover on the viewport window and marked a point at which the green beam penetrated (Fig. 6). The photos taken without any ND filters (actually I used laser safety goggles OD2 or OD4 or their combination as such filters) show saturated green spots. With appropriate filters in front of my smartphone camera, the saturation or halation seems reduced.
By the way, the green spot is at 10~11 mm off-centered in the left (-X direction) as roughly analyzed in Fig. 7. Let us compare this number with the "design" described in 17154, which says the green spot would come 14 mm off-centered left from the center point (26+24+30=80mm diameter effective area is assumed, so the center is at 40 mm, and 40-26=14 mm). So will the 4-mm difference be something??? Note that the designed number can vary if I choose a different steering angle for POP-POM in LightTools.... in other words, I may be able to prepare another "design" that would have this 10-11 mm off-centered version. Anyway, I can say the beam spot is close to the designed position.
Then, I detached the periscope, and modified the periscope system like a mirror-symmetric one. And somehow relocated the persicope (Figs. 8 and 9) so that
I do not want to repeat this again... a nightmare like jambow-manbow, but there is still a theory of simple geometric optics behind. Use your brain!
Then, thanks to my luck and efforts, the green beam came back to very close to the one before relocating the periscope. The green beam went thourgh the viewport window and illuminated POP-POM, and the following beam reflected at POP-POM shoot close at the PR2 HR target center. So I tweaked HBS in pitch and yaw to let the beam spot come at the PR2 HR target center (Fig. 10).
Firsly at PR2, I actuated the picomotors on POP-POM with a joystick to guide the green beam to the PR3 center; there has been a target plate as well (17140) by Hirata-san and Yano-san. It is about 11 m away from PR2 to PR3 (see here), so I went to PR3 to see the acutual situation. Then I learned the spot tweaking needed to be more precisely; it was displaced particularly in yaw (again, note that, for these precise beam spot centering, the ND filters were essential; saturated beam spot would be mis-understanding). So I contacted Ikeda-san and asked him to tweak the picomotors from MEDM (17165). See Fig. 11.
Coming back to the PR2 area, I checked several other points. For example, the locations of the IR (PR2_TRANS_FORWARD) and the green beam in the PR2 chamber (Fig. 12). Again, due to high power of green, it was difficult to check the situation on POP-POM. Later check it. The beam separation just out of the viewport window (i.e. in air) is shown in Fig. 13 ;99-71 = 28 mm separation, while the "design" in 17154 says 24 mm. Hmm... 28/24 = 1.16 times larger angle?? Let's see. When we will take the full alignment of the interferometer, these difference might change... (Revisting this: in the "design" I forgot some thinckness of the K400 flange (25 mm), the viewport assembly (36 mm; see here), plus the distance from the viewport window to the sensor card (say, 40 mm). Then at this point (25+36+40 ~ 100 mm away), the separation is calculated as 26 mm (I think this 2 mm increase is consistent with the additonal distance of 100 mm, as the incident angle of the main beam to PR2 is about 0.68 deg according here, so the two beams would separate twice of this, meaning 0.68*2*pi/180 = 0.02 rad, or 2mm/100mm; thanks you for suggesting this rough estimation last night, Enomoto-kun), which might be within error of 28 mm measurment; is that a little arbitrary??)
Finally, got back the PR2 target, and checked again the green beam centered (Figs. 14 and 15).
Tweak PR3 to send the green light to BS, IXA, IXC, and...
The periscope mirrors would be still PYD-20. There were bright green transmission beams from behind these mirrors. These mirrors will/must be replaced with more appropriate ones in the future...
With these measurements and their interpretation, I think the following can be also granteed; the green beam would not hit at the inner wall of the vacuum tube between PR2 and PR3.
Current laser shutter was replaced to the new one.
It works only local mode now.
Once we connect a cable from BO, it can be controlled remotely with some setting for RT model and medm etc.
It has also readback monitor, and alarm monitor for mulfunction.
To open/close the shutter, you can press the orange toggle button at center.
This is still the first trial version and it will be replaced to the second versoin soon with some feedbacks.
To share the output of Omicron between workstations, I run rsync between k1sum1 and k1detfs0. After that, I turned on crontab for same rsync on k1sum1.
I added the description about this on wiki page.
OMG... Nice carefulness to find this, but can imagine the toughness from now on...
When I and Hirata-san looked at around PR2, I was attacked by the ondotori dropping down from somewhere. I somehow returned it to the chamber, but I am not sure I returned it to the exact previous position. Please ckeck it later.
I checked current photosensor spectrum and found that MN H2 photosensor has a strange noise spectrum after yesterday work (see fig1).
Red and blue represent the spectra before (9am, 17 of June) and after the work (5pm today), respectively.
As you can see, MN H2 noise becomes roughly 10 times larger.
I also checked the time series data around the work and found the strange jump of DC output from photosensor on H2 at around 3pm (see fig2).
I will investigate photosensor H2 in the next week.
Performed Picomotor drive at the request of Akutsu-san.
POP_POM VAC_Y [ch.4] was moved from -2655 to -2275 and then to -2975.
After that, I turned off the picomotor driver and detached the relevant cables to avoid dangers, because I had no idea about the nice proper way to lay the cables...
