At Kiwamu's request, I have updated the script which runs from the 'Turn WFS ON/OFF' button on the IMC_WFS_MASTER.adl medm screen.  It appears this script was ported from the 40-meter lab (a few channel were still listed as C1:) in addition to touching a number of other switches and filters in an undesirable way.  I commented out everything except the gain setting for the slider, so it should function solely as an ON/OFF switch as advertised.  Changes have been committed to SVN.

I will take the LLO aLOG offline to migrate it to a new server during the maintenance window on Tuesday, May 12, 2015.

Work Permit:
https://workpermit.ligo-la.caltech.edu/view.php?permit_id=2544

 Re Yesterday log on the STS2 work.

Last night I swapped the field cables for HAM2 and ITMY (A & B) STS2s at the interface chassis.  This morning the traces from all three STS2 look good.  See attached, the pink traces are references and notice the noise in the .5 to 5Hz area.  All the other traces are consistent--the coherence below is exactly the same story.  Since the signal from the HAM2 STS2 looked bad with two different cables, inferring the contacts have been worked a bit, this suggests the problem is more likely on the interface side of the connection.  I'm going to do some cable wiggling and switch them back this morning.

Thanks much to Robert Schofield for assistance & teaching.

Sift wiki page

Lock times: (ODC-MASTER_OBS_INTENT & DMT-DC_READOUT_LOCKED)

1114582937 1114624965 (42028s)
1114629970 1114639925 (9955s)
1114643198 1114646416 (3218s)

There were three locks on 2nd May for ~10.5hrs , ~2 hrs and 3hrs (continued to 3rd May) respectively.

The first lock loss happen after turning on 504.8 Hz Violin mode damping loop (alog). Not sure about the second one (Nutsinee was guessing that the initial alignment was not done for the full day so may be the optics started drifting), but the third lock was lost because PSL tripped (alog).

The first lock had stable inspiral range of ~8MPC while the second had slightly better range ~10MPC.

Three dips in the inspiral range of the first lock were caused by three loud glitches. (Glitch follow up).

OAF and GDS h(t) calibration still differ between 20Hz and 110Hz (and a bit above 2kHz).

There were few interesting glitch bands: 10 to 40Hz starting from 12:30UTC. , 80-300Hz through out the locks. Most of the loud glitches were seen around this frequency band.

Both UPV and Hveto have found H1:ASC-AS_A_RF45_Q_PIT_OUT_DQ to be the most effective veto channel.

10-40Hz glitches might be related to angular sensing and control channels and output mode cleaner angular channels.  STAMP-PEM has shown that OMC-ASC_QPD_B_PIT_OUT, OMC-ASC_ANG_Y_OUT, ASC-AS_{A,B}_RF45_Q_PIT_OUT channels have high coherence with DARM in 10-40Hz frequency band which started after 12p.m. Might be the cause of excess low frequency glitches starting from 12:30p.m.

Daily CBC result has an interesting vertical band of glitches around 12 to 12:30UTC. It looks like a messy PSD estimation, but the timing is not  convincing that it's involved with the noise floor jump  which seems to be closer to 12:50 UTC.

J. Oberling, R. Savage, P. King, J. Bartlett

The goal for today was to replace the flow sensors for both PSL chillers with new ones with no moving parts; we have to bring both chillers down to replace one, so might as well replace both.  Once we swapped out the flow sensors both PSL chillers quit working, giving the same error; we are in touch with the service dept to figure out what happened.  In order to get the PSL back up and running we installed the spare PSL chillers, which we had also installed new flow sensors in.  As it turns out, these new flow sensors don't play nicely with the chillers for some reason not known to us.  We therefore re-swapped the old flow sensors that were originally in the spare chillers.  As our luck would have it, the diode chiller flow sensor was reporting no flow when we could clearly see flow; we replaced this flow sensor with one from the original chillers.  By this time we had lost track of which flow sensor was from the crystal chiller and which was from the diode chiller (the one we originally suspected as beginning to fail, thereby causing the PSL trips we have been seeing).  In other words, we traveled in one huge circle today, and the original problem with the PSL diode chiller flow interlock tripping may not be fixed.  Luckily, we are all now experienced in swapping these flow sensors out, so if the PSL does happen to trip again because of the diode chiller flow interlock, we will swap it again.

As it happens, the PSL tripped as I was writing this alog.  Jeff and Peter had already made it out, so Rick and I swapped the flow sensor and got everything to work after some issues.  More detail on that to come tomorrow, as now I'm tired and hungry.

In response to Jeff's (re)discovery of low DARM cavity pole (alog 18283), I looked at past data of various DC and RF sideband power signals to see if there was something anomalous. Specifically speaking, I made a comparison between the data from May-1-2015 and the one from May-6-2015.

A first impression I had is that, looking at the AS 90 in-phase signals from the two data, the DRMI alignment looked very good to me and therefore I imagined that losses due to misalignment in SRC have been small. Since the DRMI for the 45 MHz sidebands is (almost) critically coupled, I can do an order estimation for SRC intra cavity losses. First, I simplifed the DRMI transmissivity as

  (amplitude transmissivity of DRMI to AS) = Ts / (1 + Ts + L) ~ 1 - L/Ts

where Ts and L are SRM power transmissivity and intracavity loss (or half of the round trip loss) respectively. AS 90 is made of beatnote of the upper and lower 45 MHz sidebands. So the AS90 should be proportional to the square of the above equation. In order to explain a 2 % reduction in the AS90 as shown in the above table, L needs to be about 350 ppm. This sounds too small in order for loss to reduce the DARM cavity pole as low as 270Hz (see the plot in alog 17889).

Perhaps, I should do a similar analysis for POP90 which showed some increase -- this is usually an indication of misalignment in SRC. However, on the other hand, POP18 behaved anomalously in the sense that it dropped by 40 % while the carrier recycling gain increased slightly. This shounds to me some rf saturation, beam clipping or that sort of things and therefore I am not so keen to analyze the POP signals of the 11 W data.

Link to DQ Report: https://wiki.ligo.org/DetChar/DataQuality/DQShiftLHO20150503

-Four separate locks with stable inspiral ranges between 9 - 12 Mpc. The total lock time for the day was about 8.8 hours.
-The two temporary drops in range are coincident with two of the loudest glitches. Those glitches are at about 1:10:00 and about 9:45:00.
-Three of the four lock losses can be traced to the PSL tripping. The loss of the 3rd lock at about 9:00 is still unknown.
-A line between 10 - 20 Hz (likely due to end test mass bounce modes) is clearly visible in the spectrograms.
-Omicron plots show families of glitches around 10 - 40 Hz (likely related to angular sensing and control), between 100 - 300 Hz (glitches with the highest SNR), at 500 Hz (violin modes with their higher harmonics as well), at ~1600-1700 Hz (possibly due to OMC length dither frequency), at ~1800 Hz (unknown), and finally at ~2100 Hz (also unknown).
-Hveto picked ASC-AS_A_RF45_Q_PIT_OUT_DQ as its winner and vetoed many glitches between 10 - 40 Hz. Three of the other five winning channels were IMC channels and they vetoed many glitches at about ~ 1600-170 Hz.
-STAMP-PEM showed coherence between DARM and LSC-MICH_IN1_DQ from about 44 - 130 Hz during the first 2 hours of the day's locks.  OMC-DPCD_SUM_OUT_DQ also shows coherence at seemingly all frequencies. ISI-BS_ST2_BLND_RX_GS13_CUR_IN1_DQ, and a few other ISI channels, show coherence between the frequencies of about 50 - 160 Hz.  Some of the STAMP results were not trustworthy because it was using Obs Intent instead of DC readout to find data.
-CBC triggers seemed fairly quiet overall but showed some interesting behavior at the beginning of some lock segments. During the first lock of the day there was also evidence of vertical banding (similar to the April 23rd lock).
-Wind speed was relatively low for the day, significantly lower than the previous day. Seismic activity doesn't seem to have caused many issues either.

Summary

Our current quad distribution scheme for DARM should be compatible with the new low-noise ESD driver. Most of our rms drive to the ESD happens below 3 Hz, so the extra compensation required for the new 2 Hz / 50 Hz pole/zero pair shouldn't cost us anything.

Details

In the ETMY ESD drive we currently have passive filters installed (D1500113), with a pole at 1.6 Hz and a zero at 53 Hz. We digitally compensate for these in the L2L drivealign filter for L3.

If we remove these passive filters and install the new low-voltage driver (D1500016), we will instead have two poles at 2.2 Hz and two zeros at 50 Hz, and we'll compensate these digitally in the same way. So we will effectively have an extra pole around 2 Hz and an extra zero around 50 Hz, compared to what we have now.

I've attached a set of spectra of the ETMY ESD drive in full lock. The blue is our current drive. The red is my projection of the drive we would have if we install the new driver and compensate accordingly. It seems that most of the rms in the drive comes from 3 Hz and below, so the total rms (about 3×10³ ct) won't change much when we change the digital compensation.

Nutsdinee, Elli

The power on the EY HWS was too high and was saturating the HWS camera.  We replaced the mirror HWS-M4 on ISCTEY with a Newport 10B20NC.1 beam sampler, and dumped the transmitted beam. 

Robert, Dave. WP5191

I modified all PEM models to use spare ADC channels for Robert's testing. At the end stations spare channels on the first ADC (ADC0) were fed into the PEM model. Since this is a lower ADC card than the one currently used, I replaced the current ADC0 part with ADC1, then added a new ADC0. The final configuration is ADC0=card0(the PCAL card), ADC1=card3(the PEM card). On the first ADC, PCAL is using channels 0-7 and 28-31. PEM is using channels 8-13

At the corner station things are easilier as PEM has three ADC cards and the last card has 11 spare channels (channels 21-30 were used)

All the new channels are acquired by the DAQ in both science and commissining frames at 2048Hz. The addition to the science frame is temporary pending a decision to permanently add them.

I have named the new channels H1:PEM-{loc}_ADC_{card}_{chan}_OUT_DQ where:

loc=EX,EY,CS 

card= physical card number (starting at zero)

chan = channel number (starting at zero)

h1pemex, h1pemey, h1pemcs and the DAQ were restarted.

7:25 Richard and Peter to LVEA working on PSL UPS

7:48 Richard and Peter done

8:30 Fil and Andres to LVEA STS cabling

8:45 Craftsmen Cabinets on site

9:04 Joe D to LVEA for maintenance

9:06 Jim loading new isolation filters on ETMx ISI

9:14 Christina to both end stations

9:25 Fil and Andres out of LVEA

9:36 PSL team swapping chiller

9:47 Cris and Karen done at EY, going to EX

10:12 Joe D out

10:18 Cris and Karen done at EX

10:34 Nutsinee to LVEA near HAM 4 for HWS work

10:39 Elli and Nutsinee to EY for HWS work

10:50 Nutsinee out of LVEA

11:56 Hugh out of LVEA

12:33 Nutsinee to HAM4 for more HWS work

13:54 Elli to HAM4 HWS work

14:31 Hugh to beer garden for STS work

14:57 Hugh done

15:48 Greg and Elli to LVEA hunting parts

Elli, Nutsinee

Today we fine tuned the sled beam position on all optics, then get the green beam through both irises and hit the HWS. We moved the bottom periscope mirror back a little as the pitch knob ran out of room. We still don't see a decent return sled beam. 

The Sensor was locked, moved, oriented, leveled, iglooized, on concrete, repowered and masses pushed.  Looks like it needs another push (manual says many pushes are sometimes required.)

There was some evidence that part of the cabling could be the problem we are seeing with the HAM2 STS.  This move will allow us to test the main house field cable.  We can still check the Sat Box and Cable by exchanging them for the B unit which appears good with the newly overhauled PEM Vault unit.  The Chassis can be check too, one step at a time.

At Kiwamu's request, I have updated his PD Null script, pdOffsetNull_ver2.py, located in opt/rtcds/userapps/release/lsc/h1/scripts, to include the balance of the LSC PDs.  The list of PDs zeroed by this script is now:

'LSC-POPAIR_B_RF18',
'LSC-POP_A_RF9',
'LSC-POP_A_RF45',
'LSC-POPAIR_A_RF9',
'LSC-POPAIR_A_RF45',
'LSC-POPAIR_B_RF90',
'LSC-ASAIR_B_RF18',
'LSC-ASAIR_B_RF90',
'LSC-REFLAIR_A_RF9',
'LSC-REFLAIR_A_RF45',
'LSC-REFLAIR_B_RF27',
'LSC-REFLAIR_B_RF135',
'LSC-REFL_A_RF9',
'LSC-REFL_A_RF45',
'LSC-ASAIR_A_RF45'
'LSC-X_TR_A_LF',
'LSC-Y_TR_A_LF',
'LSC-TR_X_QPD_B_SUM',
'LSC-TR_Y_QPD_B_SUM',
'LSC-POP_A_LF',
'LSC-REFL_A_LF',
'LSC-POPAIR_A_LF',
'LSC-REFLAIR_A_LF',
'LSC-ASAIR_A_LF',
'LSC-POPAIR_B_LF',
'LSC-REFLAIR_B_LF',
'LSC-ASAIR_B_LF'

The changes have been committed to the SVN.