Yesterday, the h1hpiitmy model was installed on h1seib1. IPC channels (SUS-ISI communication) were added the BSC-ISI and the QUAD models on ITMY and ETMY.
The HEPI at EY is currently being re-commissioned with the Version 4 of the HEPI commissioning scripts (upgrade of the scripts written during the single arm test). The hydraulic pump is running with the servo controller frozen.
The PLC2 process on h1ecatc1 got stopped again [1] and because of it the MC servo board was unable to control, resulting in trouble with the IMC locking.
I did the same recovery procedure as I did before and it came back OK. The settings of the MC servo board was burtrestored to 11pm of yesterday. However we lost the settings of the delay line phase rotator (which is used to adjust the demodulation phase of the IMC PDH signal at 24 MHz) as there is no snapshot file prepared for it yet.
At the moment it is unclear to me when the process stopped and why it happened.
[1] LHO alog #5021
Transfer functions on the fourth BSC-ISI showed some unusual resonances in corner 2 (figures 2013_01_10). I have attached transfer functions from the stage 1 actuators to the L4Cs and the stage 2 actuators to the stage CPSs. After checking/re-torque bolts in this corner and changing the Vertical L4C, the new set of transfer functions look good (figures 2013_01_15). This unit looks healthy. More details will be given in the testing report.
[Patrick, Kiwamu]
Yesterday the IMC commissioning team complained that the EPICS values associated with the corner Beckoff system did't look correctly updated. This was causing confusion --- they tried zeroing the electronics offsets in the MC servo board, but the values in the MEDM screen didn't change at all. It turned out that the slow digital system itself had been running fine and had been under control regardless of whatever the values shown in a screen. Therefore it seemed that something was wrong between the Beckoff and EPICS.
We suspected that this might be due to the fact that we have so many EPICS channels managed in a single Windows machine, namely h1ecatc1. To alleviate the heavy load on the processors we have set the affinity of all the EPICS OPC servers such that they can run any of the CPUs (from CPU0 to CPU9) via the regular task manager application. Note that we have 17 OPC severs running on this machine. This solved the issue and so far it looked running OK.
Though we think we need to come up with a more reliable solution since we maybe having the same issue in some future.
Forgive the omission please. I unlocked HAM2 HEPI right after the 0815 meeting Thursday morning being fully unlocked by maybe 0930pst. Some of the signs may not be right; I'll correct hem asap.
Giacomo, Paul, Keita, Hugo, Matt, Lisa The IMC is locked and happy. The stability problems were due to the ISI, Hugo put it in a configuration which allowed us to recover a stable lock. We measured the open loop TF of the whole system, it was stable up to 30 kHz UGF. We also measured the cross-over VCO/MC2_M3: it is about 15 Hz (it can be stable as high as 100 Hz). The MC2_M3/M2 cross over is well below 1 Hz. A detailed entry with the measured transfer functions and gain settings will follow. We also centered the WFSs on IOT2 and found that a 90/10 beam splitter was used instead of a 50/50 in front of WFS_A, explaining the factor 10 power difference between the two WFSs. We leave the IMC locked for the night, and we should be able to start checking the WFSs signal path tomorrow.
We were trying to get a direct measurement of the M2/M3 cross over..we will try again tomorrow. Clean data starting at 1042513275 .
Paul has implemented a notch filter to push the UGF further up. He has measured the OLTF again and will post the new data in another entry.
The fast/slow path crossover frequency was measured by injecting an excitation signal before the MC2_M3 lock filters (H1:SUS-MC2_M3_LOCK_L_EXCMON) and taking the ratio of the signals immediatly before and after the injection point. With the settings of the relevant parameters specified below, the crossover is between 10 and 20 Hz (15 Hz? the shape of the TF is a bit funny there...). See attached plot.
Fast path gain (H1:IMC-REFL_SERVO_FASTGAIN): -2 dB
MC3 lock filters gain (H1:SUS-MC2_M3_LOCK_L_GAIN): -1000
MC3 lock filters enabled: FM5 (150:4) and FM8 (CLP100)
We used a similar setup to measure the MC3/MC2 crossover frequency, injecting an excitation signal before the MC2_M2 lock filters (H1:SUS-MC2_M2_LOCK_L_EXCMON). With the settings fo the relevant parameters specified below, the TF is well below unity at 1 Hz. See attached plot.
MC2 lock filters gain (H1:SUS-MC2_M2_LOCK_L_GAIN): 0.005
MC2 lock filters enabled: FM3 (0.1:1), FM4 (100:1) FM5 (0.01:0.1) and FM8 (ELP70)
Attached are plots of dust counts > .3 microns and > .5 microns in particles per cubic foot from approximately 6 PM Jan. 16 to 6 PM Jan. 17. Also attached are plots of the modes to show when they were running/acquiring data. Dust monitor 10 in the H1 PSL enclosure is indicating a calibration failure and will probably need to be replaced.
Attached are plots of dust counts > .3 microns and > .5 microns in particles per cubic foot from approximately 6 PM Jan. 15 to 6 PM Jan. 16. Also attached are plots of the modes to show when they were running/acquiring data. I used h1nds1 because h1nds0 was reporting an error. 6300 seconds of data was not available for each plot.
MC locking work 9:19 - 9:52 Hugh to unlock HAM2 HEPI Dust monitor 10 calibration failure 11:00 Richard turning up voltage on ISC DC racks 11:37 LN2 delivery ITMY SUS, ISI models restarted Genie lift delivered 12:46 H1SEIB6, H1SUSB6 models restarted 12:54 Jim shutdown H1SEIH16 frontend, swapped ADC card, restarted frontend Apollo removing items from end Y VEA in preparation for BSC10 de-install 15:31 DAQ restart 15:35 PEM model on H1OAF frontend restarted
h1seih16: was rebooted several times to verify that the ADC originally in the second slot was bad by moving it to the third slot. It was then removed and the spare good ADC installed.
h1hpiitmy model had some mx_stream issues until I realized I have not included it in the rtsystab file. After this was done and the mx streams restarted on h1seib1 its data is now good.
The DAQ was restarted several times because of the ITMY SUS and ISI changes (slow chans renamed) and the H1EDCU_ECATC1.ini file was added to trend the mode cleaner servo board channels. At this point the new h1peml0 ini file was taken by the DAQ, so the h1peml0 model was restarted (FLOOR->VERTEX name change on Seismometer).
Later DAQ restart to update the front end DAQ status channel list in H1EDCU_DAQ.ini.
(Betsy, Gerardo)
Optic is now in the oven.
The heat cycle was initiated at ~10:30am, intending to ramp to 34 degC over 2 hours and then soak for 6 hours before returning to room temp. At 3pm the sensor on the glass probe inside the oven showed 28degC while the controller had reached 34 degC. I attempted to increase the temp of the controller and add more time to the soak, but during the hour of messing with it the temperature started to slowly drop. Apparently the controller does not like ramps less than 2 hours and could not resume heating quickly. Because we were mid cycle, I opted to kill the power since I was leary of adding more time at increased temp during unattended evening hours. A lesson learned for the next optic/glue cure cycles - set the controller temp to ~40 degC and don't mess with it mid-cycle.
The BS cure cycle was 28 degC for 4.5 hours - not quite up to spec, but going to have to be good enough. The optic was left to ramp down to RT naturally overnight.
Today, I added the BS elliptical baffles (both HR and AR) to the suspension in addition to the remaining lower structure support brackets. As of now, the payload of the SUS is as accurate and complete as it can be without having the optic installed. It should not change much once the optic is installed (probably on the order of 10s of grams difference?). Once a decision has been made with regards to the placement of the structure on the ISI table, SEI can rebalance and begin their testing.
Did the same as was done for h1ecatc1. There are now three gateways running on this machine. I set the affinities for these tasks to all of the CPUs. This had to be done earlier for h1ecatc1 to fix problems with the EPICS update rate.
Mark B.
Andres backed out the BOSEMs on PRM and PR3, and the AOSEMs were already out, so we measured the OL counts:
prettyOSEMgains(logOLs('H1','PRM'),'PRM')
M1T1 24867 1.206 -12434
M1T2 27695 1.083 -13847
M1T3 26803 1.119 -13401
M1LF 27765 1.081 -13882
M1RT 29899 1.003 -14949
M1SD 26998 1.111 -13499
M2UL 24409 1.229 -12204
M2LL 25575 1.173 -12787
M2UR 25475 1.178 -12738
M2LR 26427 1.135 -13214
M3UL 24962 1.202 -12481
M3LL 24710 1.214 -12355
M3UR 26711 1.123 -13356
M3LR 25557 1.174 -12778
prettyOSEMgains(logOLs('H1','PR3'),'PR3')
M1T1 25835 1.161 -12917
M1T2 30072 0.998 -15036
M1T3 28666 1.047 -14333
M1LF 25623 1.171 -12812
M1RT 25798 1.163 -12899
M1SD 28238 1.062 -14119
M2UL 17706 1.694 -8853
M2LL 20285 1.479 -10142
M2UR 18746 1.600 -9373
M2LR 17714 1.694 -8857
M3UL 17313 1.733 -8656
M3LL 22891 1.311 -11445
M3UR 24376 1.231 -12188
M3LR 17159 1.748 -8580
Mark B. I entered all the above gains and offsets into the various OSEMINF blocks for PRM and PR3, checked that OSEM2EUL, SENSALIGN, DAMP, DRIVEALIGN, EUL2OSEM and COILOUTF were all correct, and updated the safe.snap files (in the new locations that Jeff K had just moved them to!).
RCG 2.6.1 now authorizes 512 IPC channels (previously limited at 64 channels with RCG 2.5). The communications channels between the BSC-ISIs and the QUADs were added at ITMY (BSC1) and renamed at ETMY (BSC6).
After assessing the location of every subsystem's safe.snaps in the userapps repository, then discussion on the CDS software meeting yesterday, we determined that each subsystem's safe.snap should live here:
/opt/rtcds/userapps/release/${subsystem}/${ifo}/burtfiles/${modelname}_safe.snap
where ${subsystem} = [isi, sus, psl, etc.] , ${ifo} = [l1, h1, i1] , and ${modelname} = name of simulink model (sans the .mdl extension) . ISI and PSL already obey this convention, but SUS had put the .snaps in subfolders of the burtfiles directory named after the optic motivated by still-under-developement Guardian work at LLO.
As of this entry, all SUS (including h1susim, and h1sustmsy) safe.snaps have been moved out of their sub-folder into the "top level" burtfiles folder, and the corresponding soft link in the target directory has been updated:
controls@opsws1:target 0$ ls -l h1sus*/h1sus*/burt/safe.snap
lrwxrwxrwx 1 controls controls 65 2013-01-17 12:46 h1susbstst/h1susbststepics/burt/safe.snap -> /opt/rtcds/userapps/release/sus/h1/burtfiles/h1susbstst_safe.snap
lrwxrwxrwx 1 controls controls 64 2013-01-17 12:46 h1susetmy/h1susetmyepics/burt/safe.snap -> /opt/rtcds/userapps/release/sus/h1/burtfiles/h1susetmy_safe.snap
lrwxrwxrwx 1 controls controls 62 2013-01-17 12:46 h1susim/h1susimepics/burt/safe.snap -> /opt/rtcds/userapps/release/sus/h1/burtfiles/h1susim_safe.snap
lrwxrwxrwx 1 controls controls 64 2013-01-17 12:46 h1susitmy/h1susitmyepics/burt/safe.snap -> /opt/rtcds/userapps/release/sus/h1/burtfiles/h1susitmy_safe.snap
lrwxrwxrwx 1 controls controls 63 2013-01-17 12:46 h1susmc1/h1susmc1epics/burt/safe.snap -> /opt/rtcds/userapps/release/sus/h1/burtfiles/h1susmc1_safe.snap
lrwxrwxrwx 1 controls controls 63 2013-01-17 12:46 h1susmc2/h1susmc2epics/burt/safe.snap -> /opt/rtcds/userapps/release/sus/h1/burtfiles/h1susmc2_safe.snap
lrwxrwxrwx 1 controls controls 63 2013-01-17 12:47 h1susmc3/h1susmc3epics/burt/safe.snap -> /opt/rtcds/userapps/release/sus/h1/burtfiles/h1susmc3_safe.snap
lrwxrwxrwx 1 controls controls 63 2013-01-17 12:47 h1suspr2/h1suspr2epics/burt/safe.snap -> /opt/rtcds/userapps/release/sus/h1/burtfiles/h1suspr2_safe.snap
lrwxrwxrwx 1 controls controls 63 2013-01-17 12:47 h1suspr3/h1suspr3epics/burt/safe.snap -> /opt/rtcds/userapps/release/sus/h1/burtfiles/h1suspr3_safe.snap
lrwxrwxrwx 1 controls controls 63 2013-01-17 12:47 h1susprm/h1susprmepics/burt/safe.snap -> /opt/rtcds/userapps/release/sus/h1/burtfiles/h1susprm_safe.snap
lrwxrwxrwx 1 controls controls 67 2013-01-17 12:47 h1susquadtst/h1susquadtstepics/burt/safe.snap -> /opt/rtcds/userapps/release/sus/h1/burtfiles/h1susquadtst_safe.snap
lrwxrwxrwx 1 controls controls 64 2013-01-17 12:48 h1sustmsy/h1sustmsyepics/burt/safe.snap -> /opt/rtcds/userapps/release/sus/h1/burtfiles/h1sustmsy_safe.snap
controls@opsws1:target 0$