This morning the AA script was running but the structures in the sensitivity, related to the DIFFp alignment were present. The AA script was correctly aligning the iTF, see figure 1, but slowly. I have boosted the alignment of DIFFP by hand at 7:30 UTC. As it is visible in Figure 2 the structures are reduced in the new working point.
Fabio has updated in the ITF.ini file the DIFFp set points
We replaced the B1 photodiode boxes with new ones. The only difference being new pre-amplifiers with same design low noise resistors:
- bulk metal foil for the 500Ohm load resistor (Y1624500R000T9R - Résistance à puce CMS, Couche métallique, 0805 [2012 Metric], 500 ohm)
- metal film for the 1.5kOhm load resistor (SMA-A0102BTNL1K5 - Résistance CMS MELF, 1.5 kohm, SMA-A Series)
- thin film for the other resistors used in the analog filters and gain stages
Figure 1 shows the analysis of last nights data after the installation. In red is the Audio channel difference, which is flat down 50Hz, and in cyan is the fitted flicker noise from last December. The flicker noise is completely gone.
Figure 2 shows for comparison the data from the previous night, were the flicker noise was actually slightly higher than what was fitted in December.
In both cases there is a steep rise in noise below 30Hz, which is partially coherent between the DC and Audio channels. This would need to be investigated further but is not a concern as these frequencies the intereferometer is limited by other noises.
Figure 3 shows in addition to the DC and Audio channel differences, the spectra of the Audio channels for each individual photodiode. Were the noise on each Audio channel is over a factor 100 higher than on the difference in between 10Hz and 20Hz, so imperfect subtraction between the two is a likely explanation for this steep rise in noise.
Upon arrival, ITF was relocking with ISC and SBE crews checking and tuning ITF.
Science Mode set again at 01:03 UTC with BNS Range 50 Mpc.
Guard Tour (UTC)
23:40 - 00:20
01:40 - 02:10
03:10 - 03:40
05:00 - 05:30
ISC
02:49 UTC - Diffp_AA red flag on DMS (DIFFp_TX_SET out of threshold). PyAATuning launched with 20 iterations.
ISC
(29-01-2020 02:45 - ) update values in ITF_LOCK.ini at the end of the iterations.
This is the report of the overall activities carried on during the day from the ISC side:
- after the work on SWEB: the goal was just to relock once, without expecting any particular change aside, perhaps, the demodulation phase of B8_6MHz, but this was not actually the case and the phase had not changed;
- we added to the initial loop filter for NArm the same simple boost (Arm_boost_aq) which was already there (undocumented, at least up to my knowledge) for WArm, in order to increase the initial lock robustness;
- we had instead the usual troubles in the transition to dark fringe, which the actions we performed yesterday did not cure completely; what we did was:
- we slightly changed the PRCL loop filter, in order to have a little more gain at low frequency without impacting basically anything else;
- we reshuffled a bit the commands sent (mainly to LSC_Acl) in that time frame when the unlock occurs (between the hand-off of MICH and the one of DARM), in order to possibly avoid timing conflicts and ramp-less executions, especially for the loop gains; the MICH one in particular has quite a change between MICH_OFFSET_0_1 and Dark Fringe, and the former needed to be increased slightly;
- after these actions we did not suffer from other unlocks, to be verified with some proper statistics;
- after the work on SDB2: here we expected the need to tune all the demodulation phases from that bench (B1p_56MHz for DARM and [B1p, B5]_QD2_56MHz for DIFFp/PR/BS), and this was carried out in the usual way without any trouble;
- as usual, after tuning the DARM phase we went in "fake" LOW_NOISE_1 (with the alignment loops open) and tuned the QPD phases there, then we closed the alignment loop by hand; the point is that, despite the QPDs were seeing the motion of the optics (either manually induced or by injecting some line), closing the loops basically did not have any effect whatsoever; we tried to understand the situation for a while, then we tried to lock OMC1 anyway while carrying on the investigation; eventually the ITF unlocked;
- we decided to acquire the following lock in a completely automatic way, and in this case all the loops were properly closed without any trouble (in the same way as we did by hand, obviously); we still have no idea about what happened bespite (despite ruling out already the most trivial possibilities); logfiles and data will be analyzed tomorrow;
- after that we went to LOW_NOISE_3_SQZ without issues and we set Science Mode at 01:03 UTC.
The fine tuning of the working point and the analysis of the new configuration will be done in the next shifts.
Before installation the B1 photodiode frequency response has been characterized at LAPP by shining light on it with an LED. (The LED had been characterized to have a flat response to the DAC actuation using a photon calibrator photodiode).
Figure 1 and 2 are for SN38, in blue is the raw transfer function between driving and the PD, in red after correctiong for the delay between the samples of the photodiode and the samples of the channel driving the LED, in yellow after applying a correction filter. Zero at 1.46Hz and pole at 1.6Hz. Figure 1 is for a white noise driving, which is loud above a few Hz, figure 2 is for a comb of lines, giving a measurement between 0.1Hz and 10Hz.
Figure 3 and 4 show the same for SN42, the same fit works well.
Previous photodiode, installed in March 2019, had a fit with a zero at 1.4Hz and a pole at 1.6Hz.
The correction filter in SDB2_Photodiodes has beent updated today (Jan 28) at 22:11:30 UTC.
ITF found unlocked with the crew that were still working on WEB trying to solve a problem before start the recovery, therefore I set the ITF in TROUBLESHOOTING MODE from 14:40 UTC to 15:17 UTC. At 16:53 UTC we were able to lock the ITF in the last state, however the weather condition were very bad and moreover a big earthquake of magnitude 7.7 happened at 19:10 UTC in Cuba shakered the suspensions for a while, therefore we took the opportunity to work on SDB2 and replace the B1 photodiode.
Air Conditioning
from 17:40 UTC to 21:38 UTC - UTA DET in "portata nominale".
SUSP
from 17:49 UTC to 21:15 UTC - following the request of the DET crew all the mirrors has been misaligned except the NI.
As a general remark, please consider that blocking the YAG beam on the laser bench implies a number of operations with possible non-negligible impact on the recovery time of the ITF, including but not limited to thermal load variation on the ISYS benches, and the risk of having the control loops not immediately in the optimal working point afterward. It should be considered as an option when other operations are discarded for some reasons, i.e. when safety of humans or equipments cannot be guaranteed in a different way, and should be planned in advance when possible.
This morning, on request, we blocked the beam on the laser bench in order to let people work properly on SWEB.
When restoring the system we had some troubles to close the BPC loop because the corrections were too high.
We brought them back to values close to 0 and restore the system in its norminal value.
A few more details about the ISC activity:
- the unplanned ISC support to the shift started at 21:24 UTC, when I was called to restart the Lock Acquisition with no beam hitting the B7 photodiode;
- with the West Arm locked and since the NI seemed aligned enough (looking at the B1p camera), I started to move the NE marionetta in order to get some flashes on the B7 PD;
- after a little less than an hour, at 22:17 UTC, I noticed that someone else was moving the suspension at the same time; I called the Control Room and the Operator told me that he was restoring the last saved position (which is in principle correct) but that:
- he was not told that I had been called before and I was working on the Lock Acquisition;
- the work on SNEB was still ongoing and it made no sense to try to lock before such work completed (and rightfully so);
- after the work on SNEB finished, around 22:48 UTC, I was called again in order to resume the recovery;
- I could only get flashes on B7_DC up to 0.2 mW, so I had to 'hack' the lock of the arm in order to succeed;
- once the North Arm locked, though, the angular dithering control could increase the arm cavity power up to 0.46 mW (little more than 50% of the standard), meaning that SNEB was misaligned;
- since nobody was present anymore, I waited until the Operator was given instructions about how to change the bench position;
- after that, Lock Acquisition could proceed and eventually succeeded, after a couple of trials and a couple of tunings:
- Marco tuned the SSFS demodulation phase;
- I tuned the PRCL demodulation phase, which was off by quite a bit (Delta = -0.35 rad); this will hopefully help in the transition to Dark Fringe (PRCL had a very low UGF during the transition), but most probably there is something else which needs some adjustment.
Science Mode was eventually achieved at 00:38 UTC.
This will hopefully also solve the frequent crashes of the WARM_LOCK node, which is the manager of SUS_BS and it drives itself the BS DSP for the angular dithering control.
As part of further investigation chopper mode is turned ON at 12.12 UTC.
If our preliminary conclusion is right the noise should appear again. This seems to be the case in the current live plots on VIM.
ITF found in Science mode.
At 7:06 UTC ITF in Maintenance mode; below a list of the activities comunicated to the control room:
- vacuum and cleaning operations;
- cryotrap refill;
- check of the SWEB mulstisas (see next dedicated entry); to perform this activity the beam was blocked on the bench.
At 11:55 UTC all the maintenance activities were concluded and we started the recovery of the SWEB/ISYS; recovery still in progress...
Air Conditioning
UTA DET in "portata nominale" from 9:50 UTC to 10:20 UTC.
TCS
standard checks/refill of the TCS chillers.
This morning, on request, we blocked the beam on the laser bench in order to let people work properly on SWEB.
When restoring the system we had some troubles to close the BPC loop because the corrections were too high.
We brought them back to values close to 0 and restore the system in its norminal value.
As a general remark, please consider that blocking the YAG beam on the laser bench implies a number of operations with possible non-negligible impact on the recovery time of the ITF, including but not limited to thermal load variation on the ISYS benches, and the risk of having the control loops not immediately in the optimal working point afterward. It should be considered as an option when other operations are discarded for some reasons, i.e. when safety of humans or equipments cannot be guaranteed in a different way, and should be planned in advance when possible.
Today at 11h UTC, I have enabled the injection of a permanent line at 2012.5 Hz on both NI and WI marionettes. Their goal is double:
- monitor the data transmission stability and delay from the RTPC (CALnoise) to the suspension DSP
- remove an artifact in h(t) when switching from earthquake (EQ) mode to standard mode of the ITF. In last week tests, h(t) = 0 for 16 s at the end of EQ mode around the time when Sc_NI,WI_MAR_Z_CORR reaches exactly 0. Having this line will prevent to reach this condition.
ENV_EER_MAG_2 was moved on Monday 27/01 between 16.02 UTC and 16.09 UTC. I moved it from below the MC tube to on top of the MC tube to be closer to the tube and have a look at possible magnetic noise carried by the tube from the MSB to the CEB. Also added a figure containing the RMS of the probe and one of the current chanels of the MCB. There is clearly so comherence. Further investigation will follow.
ENV_EER_MAG_1 was removed on Tuesday 28/01 at 09.07 UTC. This descision was made since we have sufficient evidence of the splitter in the EER room and the outside part being the source of the noise being investigated. The channel has been deactivated from colleting data.
In the last weeks the SUS_BS automation node was reporting continuosly this error:
TangoError: DevFailed[ DevError[ desc = Failed to connect to device lcuserver/bs/sa...]]
This timeout error was slowing down the node implying being late in providing data to the Fbm_Alp frame buider and impacting DAQ latency.
Profiting of a DOWN moment yesterday evening, the LcuServer/BS and DspServer/BS Tango SAT servers have been stopped/restarted.
After that the BS_SUS node has not complained anymore about "Tango Error" and the situation in term of DAQ latency seems back to normal (Fig. 1)
This will hopefully also solve the frequent crashes of the WARM_LOCK node, which is the manager of SUS_BS and it drives itself the BS DSP for the angular dithering control.
22:00 UTC, ITF found in Commissioning Mode with SNEB LC still opened. The recovery lasted until 23:40 UTC then, with the help of Vardaro and Bersanetti from remote, we started again to relock ITF.
ITF Locked in LN3 SQZ at 00:25 UTC. Science Mode set at 00:38 UTC with BNS Range 49 Mpc.
ITF unlocked (TBC) at 02:11 UTC. Relocked at 02:55 UTC,
GRB
Fermi event created: E361413 at 3:42 UTC. Alert reset at 03:44 UTC.
Guard Tour (UTC)
23:00 - 23:35
01:00 - 01:35
03:05 - 03:40
05:00 - 05:35
ISC
Fine tuning:
- Vardaro tuned the phase of B4 56 Mhz from -1.51 to -1.24. (Adjusting Mode set)
- DIFFpTy tuning from -0.034 to -0.024. After a while, I tuned again to -0.034.
- DIFFpTx tuning from -0.003 to 0.01
SBE
23:00 UTC I went at WE to switch OFF the 4 HEPA filter above SWEB clean area.
The flag SNEB_B7_Cam1_FitPosY became red. To restore the nominal value I manually tuned the value of SNEB_LC_TX_SET. Value changed from -668 to -678
The flag SNEB_B7_Cam1_FitPosX became yellow. To restore the nominal value I manually tuned the value of SNEB_LC_TY_SET. Value changed from -114 to -110.
ISC
(27-01-2020 21:15 - 28-01-2020 01:00) From remote
Status: Ended
Description: Problem with the lock of the North Arm.
A few more details about the ISC activity:
- the unplanned ISC support to the shift started at 21:24 UTC, when I was called to restart the Lock Acquisition with no beam hitting the B7 photodiode;
- with the West Arm locked and since the NI seemed aligned enough (looking at the B1p camera), I started to move the NE marionetta in order to get some flashes on the B7 PD;
- after a little less than an hour, at 22:17 UTC, I noticed that someone else was moving the suspension at the same time; I called the Control Room and the Operator told me that he was restoring the last saved position (which is in principle correct) but that:
- he was not told that I had been called before and I was working on the Lock Acquisition;
- the work on SNEB was still ongoing and it made no sense to try to lock before such work completed (and rightfully so);
- after the work on SNEB finished, around 22:48 UTC, I was called again in order to resume the recovery;
- I could only get flashes on B7_DC up to 0.2 mW, so I had to 'hack' the lock of the arm in order to succeed;
- once the North Arm locked, though, the angular dithering control could increase the arm cavity power up to 0.46 mW (little more than 50% of the standard), meaning that SNEB was misaligned;
- since nobody was present anymore, I waited until the Operator was given instructions about how to change the bench position;
- after that, Lock Acquisition could proceed and eventually succeeded, after a couple of trials and a couple of tunings:
- Marco tuned the SSFS demodulation phase;
- I tuned the PRCL demodulation phase, which was off by quite a bit (Delta = -0.35 rad); this will hopefully help in the transition to Dark Fringe (PRCL had a very low UGF during the transition), but most probably there is something else which needs some adjustment.
Science Mode was eventually achieved at 00:38 UTC.
ITF in Commissioning mode for almost the whole shift (plot attached). Science Mode only from 14:26 to 14:51 UTC, after that the ITF has been manually unlocked and the crew started to work at the installation of beam dumps on SNEB, operation didn't performed due to a "problem" during the installation itself. At 20:45 UTC we tried unsuccessfully to lock the cavities, we weren't able to lock the North arm, therefore I contacted the ISC expert that at this moment (22:00 UTC) is still trying to understand the failure, just before (at 21:00 UTC) I also contacted the vacuum expert in order to confirm us that al the valves were correctly open.
ISC
(27-01-2020 21:15 - ) From remote
Status: On-going
Description: Problem with the lock of the North Arm.
Vacuum
(27-01-2020 21:00 - 27-01-2020 21:15) From remote
Status: Ended
Description: No light on B7 photodiode.
Actions undertaken: Check of the status of the NE valves.
There seems to be a correlation between the fact that the ITF is left not locked and the corrections on one of the piezo acting on the beam-pointing control (see plot). The drift is up to now roughly 1V/5hrs, so it should be fine until tomorrow morning.
ITF found in Science mode.
At 6:25 UTC ITF in Adjusting mode to activate the "earthquake configuration" (see attached picture); at around 6:27 UTC the squeezer unlocked and the ITF switched in Locked mode.
At 6:45 UTC ITF in "Science mode".
At around 7:00 UTC started the planned activity on Wireless (Wifi/Lora) test injections in NEB (IMC instrumented baffle); the should have been carried out in Science mode but at 7:15 UTC I had to switch in Adjusting mode because I was informed that people of the previous activity had been entered the NE building without informing the control room. To be confirmed the exact time in which they entered in the building; according to them it was around 7:10 UTC.
At 7:27 UTC ITF back in Science mode.
At 13:28 UTC ITF again in Adjusting mode to allow Federico to switch off some devices at WE building.
At 13:44 UTC ITF in Commissioning mode for the activity of Federico and to prepare the planned activitivies of DET and SBE.
Preliminary results:
the moving noise lines going up and down between 200Hz and 1000Hz seem to be realted to the Chopper mode ON.
In figure 1 we see a long stretch of data. Yellow is when the copper mode is ON, red is the last signal just before it was switched OFF. Pink is with Chopper mode OFF. In the pink area none of the lines seem to appear, we see in the yellow area.
Figure 2 and 3 are details of the moment of the switch off. There is a hugh spike over all frequencies at the moment of switch off. The noise line (indicated in red) steps perfectly at the spike.
Since the 20th December two streams of magnetometer data (V1:ENV_EXT_MAG_N, V1:ENV_EXT_MAG_W) are sent from Virgo to the working station of Renato Romero. This work started from the signature in the year 2018 of MOU between EGO and Renato Romero: "ACCORDO DI COLLABORAZIONE TRA IL CONSORZIO EGO ED IL SIG.RENATO ROMERO", and its goal is to investigate the possible sources of magnetic noise around Virgo. The software work essentially consisted in finding the best way to send our data from VIRGO to the Spectrum-Lab (https://www.qsl.net/dl4yhf/spectra1.html) application used by Romero for the analysis of magnetic data. After the first attempts with the VORBIS-OGG and FLAC format, we preferred to use the raw-tcp (home-made) format: a Cm server reads data from a frame provider and transforms this data in the raw-tcp format. Although the system is still in the testing phase, at the following link we can find the first more than encouraging results: http://www.vlf.it/virgo/virgo.html .
We started with a wifi test. We performed some injections at transmission powers of 5, 12 and 20W for periods of 150 and 500 seconds (the specific injection times can be seen in the attached log-file). We saw no effect in DARM.
Then, we proceeded with the Lora tests at powers of 5, 12 and 20W for periods of 200 seconds and varying the transmission frequency between 750-900MHz. Again, we saw no effect in DARM.
ITF found in Science Mode, it kept the lock for the whole shift.
Guard Tour
23:37 UTC - 0:09 UTC;
1:35 UTC - 2:04 UTC;
3:37 UTC - 4:08 UTC;
4:56 UTC - 5:30 UTC.
ITF found in Science Mode, it kept the lock for the whole shift. BNS Range ~50 Mpc.
Adjusting Mode break from 20:38 to 20:49 UTC in order to allow Bertolini to restore a good vertical position of SPRB F0. (plot1)
SBE
(26-01-2020 20:38 - 26-01-2020 21:00) From remote
Status: Ended
Description: restore a good vertical position of SPRB F0
