The planned TCS activity on Calibration of actuators carried out by Nardecchia, Taranto, Lumaca started at 15:00 UTC and went on without major problems for the whole shift. Under request of Lumaca, connected from remote, I relocked ITF on IR at 21:45 UTC in order to keep ITF well aligned for tonight.
On-line Computing & Storage
Control Room phone line (2600) always busy for the incoming calls. Service team informed. To contact the Control Room dial 2601.
SBE
At 16:15 UTC the LC and SBE of SQB1 opened. With the help of Gouaty (contacted) we closed them properly and restored the vertical position of the bench.
Today afternoon, after ISC finished the shift, I proceeded with the CHs calibration, taking the first measurement for the west arm.
The actuators' powers at the beginning of the shift are shown in the table
CH (W)
DAS IN (W)
DAS OUT (W)
WI
0.08
0.23
0.57
NI
0.16
0.33
1.97
I switched off the WI DAS and CH at 16.10 UTC to let the CP cool down.
At 18.45 UTC the WI-CH was switched ON @82 mW.
At 21.19 UTC the WI-CH was switched OFF.
Since this work will continue during the weekend, all the WI actuators have been left OFF and, with the aim to repeat a longer measurement also with NI-CH @115 mW, at 21.29 UTC also all the NI actuators were switched OFF.
I gave a closer look at the data reported in this entry. In particular to the first 10 minutes since the lock is acquired, because there is where a lot of the thermal lensing occurs (see figure 1): after 10 minutes, the curvature is already 55% of the steady state.
Checking the values of B7 for the four locks under investigation (again after 10 minutes), the intracavity power during the December 1st long lock is lower with respect to the other locks by less than 2% (figure 2).
Furthermore, I made a fit of the B4_112MHz_mag data (figure 3) with an exponential function to extract the decay time during these four locks. Figure 4 shows an example of such fits, where the outliers have been removed and only the decading transient has been considered. The results are reported in the table below.
Lock date
Decay time [s]
21-12-01 13:04 UTC
2.8e3
21-12-02 09:14 UTC
4.3e3
21-12-02 11:26 UTC
1.7e3
21-12-02 12:44 UTC
2.6e3
So, the conclusion is that, at least during the transient (first 10 minutes of the locks), there is no such big difference in the intracavity power during these locks, and that the decay times of the 56 MHz sidebands are not correlated with the power in the arms.
If you want to see the "dance" of the sidebands and carrier of B4 beam, here is a video of the B4 phase camera images around the moment where SR_Ty was moved horizontally during the lock of the 30/11.
SR_Ty was moved at 16:40:37 UTC (1min2sec in the video), it has hard to see if the beams moved with SR.
The effects of SR TY alignment are still visible on B5 QD2 56MHz signal during the Dec 2 locks, but less clearly. One needs to look at the trend data to remove some of the high frequency noise.
Figure 1 shows the lock around 9:15 UTC. The 0.4 urad step on SR TY makes a clear step on B5_QD2_H_56MHz, and also on SDB1 TX that follows the B5 DC beam. The latter is surprising as this is the vertical direction while we move SR horizontally.
Figure 2 shows the lock around 11.25 UTC. The 0.2 urad steps on SR TX do not show any clear signal on the B5 quadrant signals.
Figure shows the lock around 12.45:UTC. The series of 0.2urad steps on SR TY seems to increase the error on B5 QD2 H 56MHz this is less clear than in figure 1, but that could be because we are moving away from the zero of the signal. The starting point was already at 0.02V on the quadrant, which in Nov 30 data corresponded to the end of the linear range of the quadrant error signal sensitivity to SR (and to a ~1urad misalignment).
In conclusion B5_QD2_H_56MHz still seems to be a good candidate for SR TY control. For SR TX the vertical signal of the same quadrant doesn't seem to have visible information.
Quick Summary: IMC and RFC locked; all Suspensions loops closed, all SBE loops closed, NE Pcal laser off; NEB Eurotherm not reachable; TelescreenArt process crashed.
No activities ongoing.
Images attached to this report
AdV-ISC (Commissioning up to first full interferometer lock)flaminio - 0:40 Friday 03 December 2021 (54086)
Print this reportA few observations about two recent locks (30/11 and 01/12)The two plots attached to this elog entry are the ones that I showed at the meeting earlier today.
The first shows some of the interesting signal during the 25' long lock on 30/11. - In the middle of the lock the SR mirror alignment in TY has been changed. As already noticed this reduces the B1p_DC and increases B7_DC/B8_DC. Moreover the increasing trend in B1p_DC is stopped. At the same time there is a decrease in the overlap of carrier and LSB as measured by the phase camera on B4. - During this lock SPRB is kept aligned using the centering of the 112 MHz signal on QD2 (top right plot). In order to do so the bench is moved in TX and the same motion is visible on the QD2 DC V i.e. the carrier beam in vertical. This confirms that the carrier and the sidebands are moving differently and that the sidebands are moving more than the carrier (at least vertically). The relation between the motion in the bench in TY and the QD2 in the horizontal direction is less obvious and it seems to change slope when the SR mirror is moved (also in TY).
The second plot shows the sidebands (56 and 6 MHz) and the carrier during the long lock on 01/12. - While all four sidebands look pretty much superposed, the carrier is slightly shifted in position (and also rotated). If one looks at their motion as a function of time the sidebands seems to "dance" coherently while the carrier is on a different rhythm (a slower one).
Images attached to this report
Comments to this report:
bonnand - 10:01 Friday 03 December 2021 (54090)
Print this report
If you want to see the "dance" of the sidebands and carrier of B4 beam, here is a video of the B4 phase camera images around the moment where SR_Ty was moved horizontally during the lock of the 30/11.
SR_Ty was moved at 16:40:37 UTC (1min2sec in the video), it has hard to see if the beams moved with SR.
Today afternoon, we continued the NI CH calibration (First part: entry 54005, second part: entry 54071). The actuators' powers at the beginning of the shift are shown in Tab.1.
Tab.1
CH (W)
DAS IN (W)
DAS OUT (W)
WI
0.08
0.24
0.54
NI
0.16
0.39
1.95
We switched off the NI DAS and CH at 15.16 UTC to let the CP cool down.
At 17.43 UTC, we switched ON the NI CH at 160 mW. We decided to increase the length of the measurement, in order to better approach the steady state (wrt 54005) and to have the same time scale of the measurement done at 220 mW (54071).
At 21.10 UTC the NI DAS actuators were switched ON.
We left the arms locked on the IR.
Detector Operation (Operations Report)gherardini - 23:01 Thursday 02 December 2021 (54083)
Print this reportOperator Report - Afternoon shiftThis afternoon shift was dedicate to NI TCS actuator calibration, the work was carried out with the ITF not locked, concluded at 21:30UTC; the SQZ activity is still in progress... I left the two cavities locked on the infrared.
other activities:
-SQZ: installation of green PSB and system alignment recovery, activity carried out inside the detection lab from 16:30UTC;
As noticed by Bonnand (in 54075) and Ruggi, yesterday's arm transmitted power at CARM offset zero was lower than usual (actually it was the lowest since a few weeks), most likely due to a misalignment of the test masses (see figure). In today's attempts, the cavity power was higher, together with the sideband decay rate. Figure 1 shows a comparison between yesterday and today's locks, including the (currently out of loop) test-mass centering dithering error signals.
The possible (positive) effect of yesterday's lower arm circulating power (due to the misalignment) on the sideband decay rate and lock longevity should be further evaluated.
Today' shift focused on checking the reproducibility of yesterday's long lock (54067) and to further check the influence of SR misalignment on the SDB1 and SDB2 quadrant signals.
Three locks at CARM offset zero were achieved:
9.15-9.40 UTC, ~25 minutes long, a small tuning of SR TY (0.6 rad) was performed and triggered a large improvement in the dark fringe stability and level at 9.18 utc (see figure 1 and 2). A sudden ~85 Hz oscillation noticeable on DARM and DIFFp TY led to the unlock (figure 3).
11.26-11.38 : the engagement of SSFS and the 1f handoffs of PRCL and MICH were tested successfully during this lock. In the second part of the lock the alignment of SR TX was moved in steps of 0.2 rad each (see figure 4).
12.45-12.58 : further misalignment steps of SR were tested in this lock, both in TX and TY (see figure 5 and 6)
Further attempts failed due to difficulties in engaging the lock of the CITF in the afternoon.
Some notes on the SR misalignment steps and CARM offset zero behaviour:
In today's locks, no clear effect of the SR misalignment was noticed neither on the SDB1 LC (closed on the SDB1_B5_QD2 DC quadrant) nor on the SDB2 56MHz quadrant signals (see figure 2-6). This is in opposition to what observed in yesterday's long lock (see figure 7) and in the 30th november locks (54061 ). The reason of this change of behaviour is still under investigation.
In the first lock, the SR alignment change 'triggered' a (lately recurrent) sudden change in the ITF behaviour. This effect has been observed several times in the past days too, and often happens a couple of minutes after reaching carm offset zero without any particular operation triggering it (e.g. today at 12.46.20 UTC, see figure 8). The usual features of this behaviour change on many signals, including the following:
on B1p, B4, B7, B8 DC signals
on the B1p and B5 galvo and quadrant signals,
beam shape observed on cameras and phase cameras
DARM error signal and correction spectra (this also leads to an offset change on the test mass AA centering signals obtained from demodulating the dithering lines from the DARM correction sent to the input mirrors)
DIFFp error signals
a reduction of slope of the B4_12 and B4_112 MHz signals
sidebands balance
Additional operations performed:
The engagement of the SDB1 drift control using the LC_B5_QD2_enbl relay has been added in the automation in DRMI_LOCK.py. It is now automatically engaged once the 3f handoff of the CITF is performed, and it is disabled in the DOWN state.
The demodulation phase of B2_169MHz (used for MICH and SRCL) has been reduced by ~0.5 rad in the 160mW and 480mW locking steps, in order to avoid unlocks during the CARM offset reduction procedure.
The ITF was left in DOWN state in order to proceed with the TCS calibration activities.
Images attached to this report
Comments to this report:
valentin - 20:45 Thursday 02 December 2021 (54084)
Print this report
An additional note:
As noticed by Bonnand (in 54075) and Ruggi, yesterday's arm transmitted power at CARM offset zero was lower than usual (actually it was the lowest since a few weeks), most likely due to a misalignment of the test masses (see figure). In today's attempts, the cavity power was higher, together with the sideband decay rate. Figure 1 shows a comparison between yesterday and today's locks, including the (currently out of loop) test-mass centering dithering error signals.
The possible (positive) effect of yesterday's lower arm circulating power (due to the misalignment) on the sideband decay rate and lock longevity should be further evaluated.
Images attached to this comment
mwas - 8:37 Friday 03 December 2021 (54089)
Print this report
The effects of SR TY alignment are still visible on B5 QD2 56MHz signal during the Dec 2 locks, but less clearly. One needs to look at the trend data to remove some of the high frequency noise.
Figure 1 shows the lock around 9:15 UTC. The 0.4 urad step on SR TY makes a clear step on B5_QD2_H_56MHz, and also on SDB1 TX that follows the B5 DC beam. The latter is surprising as this is the vertical direction while we move SR horizontally.
Figure 2 shows the lock around 11.25 UTC. The 0.2 urad steps on SR TX do not show any clear signal on the B5 quadrant signals.
Figure shows the lock around 12.45:UTC. The series of 0.2urad steps on SR TY seems to increase the error on B5 QD2 H 56MHz this is less clear than in figure 1, but that could be because we are moving away from the zero of the signal. The starting point was already at 0.02V on the quadrant, which in Nov 30 data corresponded to the end of the linear range of the quadrant error signal sensitivity to SR (and to a ~1urad misalignment).
In conclusion B5_QD2_H_56MHz still seems to be a good candidate for SR TY control. For SR TX the vertical signal of the same quadrant doesn't seem to have visible information.
Images attached to this comment
rocchi - 15:23 Friday 03 December 2021 (54091)
Print this report
I gave a closer look at the data reported in this entry. In particular to the first 10 minutes since the lock is acquired, because there is where a lot of the thermal lensing occurs (see figure 1): after 10 minutes, the curvature is already 55% of the steady state.
Checking the values of B7 for the four locks under investigation (again after 10 minutes), the intracavity power during the December 1st long lock is lower with respect to the other locks by less than 2% (figure 2).
Furthermore, I made a fit of the B4_112MHz_mag data (figure 3) with an exponential function to extract the decay time during these four locks. Figure 4 shows an example of such fits, where the outliers have been removed and only the decading transient has been considered. The results are reported in the table below.
Lock date
Decay time [s]
21-12-01 13:04 UTC
2.8e3
21-12-02 09:14 UTC
4.3e3
21-12-02 11:26 UTC
1.7e3
21-12-02 12:44 UTC
2.6e3
So, the conclusion is that, at least during the transient (first 10 minutes of the locks), there is no such big difference in the intracavity power during these locks, and that the decay times of the 56 MHz sidebands are not correlated with the power in the arms.
First of all we have replaced the cable of RF_PD in order to reduce power losses (photo attached).
We checked Power on SC beam path:
19 mW out of fiber
18 mW after power control PBS
17.5 mW before FI
15 mW first pass after FI (towards OPA)
14 mW second pass on FI (towads HD)
12 mW reflected from IR shutter towards AA PD
0.8 mW after filter before AA PD
In this condition we had a filter in front of RF_PD and we investigated on the SC AA loop with DL on. The strategy was the following:
we moved the SC_PD_MIR in order to center beam on RF_PD with SC AA loop engaged (Err 4MHz I); then, AA loop opened, we opened IR shutter and we looked at PD_MONI scanning OPA and Pump Phase. As you can see in Fig.2 we reached about 97 % of MM. We noticed that M3_X and M3_Y were saturating, so we discharged them and before we adjusted demodulation phase of M3_Y.
After Actuators discharging and realigning by hand, we engaged SC AA loop improving the signal on RF_PD, and then we measured 99% of MM on PD_MONI, as you can see in Fig.3,4. We added the new setpoints of M1 and M3 on the configuration file EQB1_HD_AA.cfg (lines 141-144).
Images attached to this report
Comments to this report:
garaventa - 8:25 Friday 03 December 2021 (54088)
Print this report
The cavity power was lower than usual most likely due to the arm dithering and COMMp alignment loops being disabled, therefore their alignment not being optimal.
The most sensitive lenses are L2 and L3 (excluding LAOM used the last time https://logbook.virgo-gw.eu/virgo/?r=53560). We moved only L2, being after the AOM path and deciding to move L2, if needed when we will replace the GPBS before the AOM.
Moving step by step L4 in direction of the AEI BOX, i.e. by reducing the distance GL3-GL2 to 940 +/-2 mm. Our goal was MM~ 5%, but we decide to stopwhen at 4%
MM = 3.9+/- 0.1 % (see attched pictures)
COMMENTS
At the end of the shift we uload the BPC Mirros GM5 and GM7
I attach the scans we did: only Y_A is not correct and needs to be done again.
We see that TZ and the position dofs are not reconstructed properly. For the positions we may need to acquire slower and larger scans in order to be able to see the displacement of the beam on the PSD instead of a noise. For TZ we will try to adjust the matrices.
Something worth noticing concerning the long lock of yesterday is that the cavity power was much lower than the previous attempts.
Figure shows the cavity power (B7/B8_DC) and the B4_12_MHz and B4_112MHz.
One can see that at the begining of the lock the power in the cavity follows the usual trend but after a few minutes dropped at 13:13 UTC and become noisier and then keep lowering up to the moment where the NE_Ty was moved (not visible in the plot).
I don't know what causes the cavity power to drop but it is not related to the opening of the galvo loop that happened 9 minutes later at around 13:22.
It is also noticeable that the B4 12MHz and 112MHz magnitude is increasing while the cavity power is decreasing but I cannot say if it is related.
The fisrt crash of the FFLMoni52 occurred the 20211130-22h48LT and several times later until we found that there is 2 instances of FFL builder for the rds, trend and trend_ll files.
This was possible only because these servers were removed from the Cascina Cm domain using the cm kill command , here the report of the CmNameServer
NSNov_30_17_47_50.log:NS> Nov 30 17:52:01 Cm>ResquestKill "FmRds" from cm_1
NSNov_30_17_47_50.log:NS> Nov 30 17:52:01 Cm>CmConnectRequestKill> FmRds CmConnect - CmName FmRds, port 28003, Host olserver53.virgo.infn.it, owner virgorun, socket_id 471 (16384,32768), local 0, alive 1
NSNov_30_17_47_50.log:NS> Nov 30 17:52:13 Cm>ResquestKill "FmTrend" from cm_1
NSNov_30_17_47_50.log:NS> Nov 30 17:52:13 Cm>CmConnectRequestKill> FmTrend CmConnect - CmName FmTrend, port 28005, Host olserver53.virgo.infn.it, owner virgorun, socket_id 476 (16384,32768), local 0, alive 1
NSNov_30_17_47_50.log:NS> Nov 30 17:52:31 Cm>ResquestKill "FmTrend_ll" from cm_1
NSNov_30_17_47_50.log:NS> Nov 30 17:52:31 Cm>CmConnectRequestKill> FmTrend_ll CmConnect - CmName FmTrend_ll, port 28007, Host olserver53.virgo.infn.it, owner virgorun, socket_id 477 (16384,32768), local 0, alive 1
Please before using the cm kill command check that the related server is not anymore running
Today afternoon, the CH calibration activity started three hours later (due to the request to give to ISC more time to work and exploit the long lock that eventually ended at 16.42 UTC - entry 54067).
We continued the NI CH calibration started Friday 27 November 2021 (entry 54005).
We switched off the NI DAS and CH at 17.26 UTC to cool down the CP.
At 19.26 UTC, we switched ON the NI CH at 220 mW. Then, the NI CH power has been decreased to the nominal value (160 mW) at 21.52 UTC.
Also the NI DAS actuators have been switched ON at the previous powers (see tab 1).
Tab.1
CH (W)
DAS IN (W)
DAS OUT (W)
WI
0.08
0.24
0.54
NI
0.16
0.39
1.95
We left the arms locked on the IR.
AdV-ISC (Commissioning up to first full interferometer lock)tacca - 23:28 Wednesday 01 December 2021 (54070)
Print this reportComment to ISC shift: SDB1 drift control, long lock (54067)In the attached figures some data acquired by the phase cameras during the long lock (13:04 UTC and 16:42 UTC). - Figure 1 shows the B4 images at the beginning of the lock -> 6 MHz and 56 MHz sidebands are characterized by an almost Gaussian shape; - Figure 2 shows the B4 images just before the end of the lock -> 6 MHz and 56 MHz sidebands are characterized by an almost Gaussian shape. The quality of the sidebands remained almost unchanged after more than 3 and a half hours of lock. - Figure 3 shows some time data acquired by B1p and B4 phase cameras -> both on B1p and B4, after the usual large imbalance at the beginning of the lock, the 56 MHz USB and LSB are better balanced for most of the time of the lock, the imbalance increases after a couple of hours of lock; both on B1p and B4, the 6 MHz USB and LSB are almost always balanced for all the lock. - Figure 4 shows a zoom of the same data at the beginning of the lock -> as already observed yesterday, on B1p USB and LSB cross a first time some seconds after the CARM null condition is reached and a second time after about 1 minute; while on B4 USB and LSB cross a first time at CARM null and a second time after about 1 minute.
The first part of the shift was dedicated to the lock to CARM_NULL started at 13:04 UTC (see report #54067). The lock lasted until 16:42 UTC and, after two following locking attempts, the planned TCS activity (carried out by Nardecchia and Taranto), started at 17:20 UTC. The measurements proceeded for the rest of the shift with the ITF unlocked without major issue. Activity concluded at 22:05 UTC, infrared cavities left locked.
Other activities carried out during the shift:
SGD Green Matching/ Installation of green PBS/ Telescope for AA (Delaurentis, Garaventa, Guo)
No DMS alerts were received during the shift.
Sub-system reports
DAQ
FFLMoni52 crashed at 17:58 UTC and 19:01 UTC. The process was restarted via VPM.
Comments to this report:
masserot, mours - 5:42 Thursday 02 December 2021 (54073)
Print this report
The fisrt crash of the FFLMoni52 occurred the 20211130-22h48LT and several times later until we found that there is 2 instances of FFL builder for the rds, trend and trend_ll files.
This was possible only because these servers were removed from the Cascina Cm domain using the cm kill command , here the report of the CmNameServer
NSNov_30_17_47_50.log:NS> Nov 30 17:52:01 Cm>ResquestKill "FmRds" from cm_1
NSNov_30_17_47_50.log:NS> Nov 30 17:52:01 Cm>CmConnectRequestKill> FmRds CmConnect - CmName FmRds, port 28003, Host olserver53.virgo.infn.it, owner virgorun, socket_id 471 (16384,32768), local 0, alive 1
NSNov_30_17_47_50.log:NS> Nov 30 17:52:13 Cm>ResquestKill "FmTrend" from cm_1
NSNov_30_17_47_50.log:NS> Nov 30 17:52:13 Cm>CmConnectRequestKill> FmTrend CmConnect - CmName FmTrend, port 28005, Host olserver53.virgo.infn.it, owner virgorun, socket_id 476 (16384,32768), local 0, alive 1
NSNov_30_17_47_50.log:NS> Nov 30 17:52:31 Cm>ResquestKill "FmTrend_ll" from cm_1
NSNov_30_17_47_50.log:NS> Nov 30 17:52:31 Cm>CmConnectRequestKill> FmTrend_ll CmConnect - CmName FmTrend_ll, port 28007, Host olserver53.virgo.infn.it, owner virgorun, socket_id 477 (16384,32768), local 0, alive 1
Please before using the cm kill command check that the related server is not anymore running
