• My only chance of hand aligning my way through SRCL to get to NLN was thwarted by a pesky lockloss.
• While trying to relock, MICH alignment got ugly
• While trying to get MICH re-aligned, there was a pretty large EQ in Central America that tripped the arm ISIs
• Untripping ISIs at 1um/s and making preperation to try re-locking. Kiwamu is here addressing the recycling gain/SRCL ugliness.
• winds aren't an issue so far
• PSL dust monitor sounding regularly

≈19:02UTC the wave arrived. All arm ISIs tripped at 19:07.

TITLE: 11/24 Day Shift: 16:00-00:00 UTC (08:00-16:00 PST), all times posted in UTC
STATE of H1: Locking
OUTGOING OPERATOR: Nutsinee
CURRENT ENVIRONMENT:
    Wind: ~17mph Gusts, 20mph 5min avg
    Primary useism: 0.15 μm/s
    Secondary useism: ~0.71 μm/s
QUICK SUMMARY:

TITLE: 11/24 Owl Shift: 08:00-16:00 UTC (00:00-08:00 PST), all times posted in UTC

STATE of H1: Meh

INCOMING OPERATOR: Ed

SHIFT SUMMARY: Anamaria left couple of suggestions of things that I can do to help improve the range. I was going to try for an hour but caused a lockloss in the process. Have been having trouble locking all my shift. There're only two ISI configurations I could try without using BRSX (which is still rung up). USEISM_NOBRSXY took me the furthest towards NLN (I was in NLN once with this configuration) but it seems unstable at ~37 mHz.

LOG:

9:08 Lockloss as I tried to engage SRC1 P. Not sure how. I think that's because I forgot to turn on the -20dB filter and turn off the integrator. Sorry. I was going to try moving SR2 and hope that SRM follows.

10:12 Locklosses couple of times during PRMI to DRMI transition. Turned out SRM pitch got stuck in some bad alignment. I moved SRM during PRMI locked then transitioned. ASC was able to engage without issue.

10:18 Lockloss 1 minute into NLN.

10:48 Lockloss the twice at SWITCH_TO_QPDs.

11:13 Lockloss at CARM_ON_TR. I wonder if the useism and wind is becoming a problem.

11:25 Another lockloss at SWITCH_TO_QPDS.

~11:50 Tried to engage BRSX after another lockloss at CARM on TR but it's still rung up. Reverted the configuration to no BRX.

12:02 Tried switching to USEISM_NOBRSXY. It's not in the reccomend list. But with BRSX rung up I kinda run out of options.

          Made through SWITH_TO_QPDS with USEISM_NOBRSXY configuration

12:30 Arrived at NOISE_TUNINGS. Staying here for half an hour.

         While staying here I noticed AS90/POP90 dropped suddenly. I tried several things to improve AS90/POP90. Moving SR2 seems to work. See alog31813

13:08 Run a2l

13:14 lockloss.

13:27 Went to LVEA to toggle noise eater. Checked with LLO ops (Jeramy)

13:50 Kyle to BSC8 to continue his work from yesterday.

14:32 Kyle out

14:41 Lost lock a few times at PRMI_LOCKED and DRMI_LOCKED. Tidal seems unstable. Recently lost lock at SWITCH_TO_QPDS. I'm switching the seismic config back to Windy no BRSX

15:12 Switching ISI config back to USEISM_NOBRSXY after 3 locklosses at FIND_IR. Wind reached 20 mph.

15:39 Lockloss at LOWNOSIE_ASC (IMC-F ran away at 37 Hz). PRMI striptool screenshot attached.

(see WP #6332)

~0605 - 0630 hrs. local -> Kyle in and out of LVEA 

Attempted to de-energize PT180 by removing the fuse at terminal #111 in the Y-arm vacuum rack.  Noted that the old PT180A signal wiring had been lifted from the terminal strip and was deliberately zip-tied out of the way.  The cabling to PT180B was still landed having only the black conductor fused at terminal #111.  Curiously, the indicated pressure value on the MEDM screens did not change as the result of removing this fuse.  Next, I removed the external gauge electronics, isolated PT180 from the site vacuum volume via closing the 1 1/2" metal-sealed isolation valve and then opened the pump port valve so as to pump the PT180 volume with the locally mounted turbo.  Heating was started with the initial variac settings of 20%.

Unfortunately, we are using the BPG402-SE version of this gauge - where the "S" indicates "short" nipple.  This limits the gauge bake temperature to a modest 80C as measured at its CF flange as the internal PCB can't be thermally radiated at greater temps.  This may be a significant limitation as the current theory is that water vapor is the source of the undesired gauge behavior and a relatively short duration bake at relatively low temperatures won't likely remove very much water.  (I can cheat a little and adjust the non line-of-site surfaces to much higher temps.)  

SUMMARY:  

Disregard the indicated pressure value shown for PT180 as it is not valid (the value at the time of gauge de-energization is latched in Beckhoff?) 
Also, I will need access to the LVEA later this evening and twice tomorrow, Friday, to make measurements and adjustments. 

As I was sitting at Noise Tunings I noticed AS90/POP90 suddenly dropped. First I tried adding some offsets to ASC SRC pitch. I couldn't make a different. Moving SRM also seems to have no effect but moving SR2 I was able to improve AS90/POP90. But something kept dragging SR2 back to where it was in both pitch and yaw (bad place). Anamaria mentioned that she turned off SRC during the last lock strech so I tried and SR2 suddenly came back to its happy place. I was able to put AS90/POP90 back to where it was.

Kept losing lock during switch to QPDs. I wonder if useism+wind is a problem now (useism has reached 1 um/s while wind is ~10 mph). It seems to help the detector stay longer in the state if I sit at TR CARM a little. I will try to stay there a little longer.  Switched ISI config to USEISM_NOBRSXY and was able to move on.

John Z, TJ, Andy, David S, Ansel, Young-Min, Jess 

Summary: ETMX optical lever laser glitches are coupling into h(t) and causing high signal-to-noise ratio single-ifo burst and CBC triggers. We suspect the op lev sensor is coupling to h(t) through a damping loop for the L2 stage, as large spikes in the op lev laser power are coincident with spikes in L2 OSEM actuation in addition to h(t) glitches. 

Suggestion: The ETMX op lev laser or power supply should be swapped out ASAP. 

Hour-long clusters of glitches at 10-20 and 30-60 Hz were first noticed on Nov 18 in during a DQ shift earlier this week. The glitches seem to appear sporadically and can be seen in Omicron triggers on Nov 18 at about hours 12:00 and 18:00 UTC. The hveto algorithm, which looks for coincidence between glitches in h(t) and glitches in auxiliary channels, found the ETMX optical lever channels to be significantly correlated with these glitches. (Note the EX SUS rack magnetometer witnesses these glitches as well.)

Attached are two examples from Nov 18 of h(t) glitch correlation with spikes in ETMX optical lever laser power and L2 suspension control.

One of these optical lever glitches was the loudest single-ifo CBC trigger for LHO on Nov 22. They characteristically appear to be a few seconds long in h(t) and broadband, with similar time-frequency content to the optical lever PD signals: [h(t), optical lever (pitch)]. 

TITLE: 11/24  EVE Shift: 0:00-08:00 UTC (16:00-0:00 PST), all times posted in UTC
STATE of H1:  PEM Calibration
INCOMING OPERATOR: Nutsinee
SHIFT SUMMARY:

PEM Injections for the 2nd part of the shift.  Range was not the best (i.e. around 60Mpc)

LOG:

• Kyle in LVEA adjusting BSC8 Pump
• BRSx is in odd state (see Jim's alog & below).
• 0:10 LVEA Dust#10 MAJOR Dust Alarm
• CP4 sporadic alarms
• 4:13 NLN
  • Difracted power drops from 4 down to 1 around DRMI.  So tweaked back up to around 4 via RefSignal slider.
  • Ran A2L
  • Closed REFL & ASAIR Beam Diverters (POP remained open) for PEM Injections
• 5:16 Lockloss
• After a lockloss, tried turning ON the BRSx, but it did not damp out after 30+min, so I returned it to DISABLED damping.  SEI CONFIG remains at WINDY_NOBRSX.
• 6:10  NLN & AS+REFL Beam Diverters Closed.  No A2L run.  Waiting for PEM Injections
  • 6:25 Disabled SRC1 p&y and then tweaked the SRM to try and improve range & POP90. (Anamaria from Sheila guidance earlier)
    • Kept SRC1 p&y off (or "open").  If POP90 strays, tweak SRM to minimize it.

New Operator Sticky Notes:

• Injections:  If one occurs over the weekend, please see Dave's instructions (alog31801).
• BRSx is disabled & SEI_Config is in WINDY_NOBRSX state
• CP4 Alarms

• H1 was hovering with a range around 60Mpc (maybe a little below).  
• Currently, Anamaria & Robert are out on the LVEA floor doing PEM Injection work.
• BRSx continues to be OFF.
• Winds have been around 15-25mph this evening

During the daily run call today, Adam Mullavey mentioned that the hardware injection team is planning to do a series of coherent hardware injections over the weekend.  The software which selects event candidates and generates EM alerts, "Approval Processor", had been configured to treat hardware injections like real triggers, but that would've caused problems for the hardware injection tests.  So this evening (at 21:00 PST) I have changed the configuration of Approval Processor to make it ignore hardware injections.

This configuration change has been documented in the Bursts/EMFollow wiki at https://wiki.ligo.org/Bursts/EMFollow/SoftwareRecord .

This afternoon Annamaria and I looked again at changing alignments.  

We tried again moving the PR3 spot position back towards where I had it last night, and saw that this still improves the noise in DARM and the range although the power recycling gain drops to almost 29.  

We then tried exploring alingment of IM3, and moved it by 2000urad in pitch.  We saw the power recycling gain increase from about 32 to 32.3 at best, this alingment was with the PR3 spot in the bad jitter noise location and the IM3 pitch alignment slider at 27940.  We were able to make the jitter noise worse by moving IM3 in yaw, but not to make it better. The attached screenshot shows the extremes of the noise in the bucket while we were moving things around.  

We reverted the IM3 alignment and I was going back to the lower noise PR3 spot position when I moved to quickly and broke the lock.  

Keita, Sheila

At 11 am pacific time (19:00 UTC Nov 22nd), there was a sudden shift in alignment in the SRC that caused bad noise in DARM. The first attached screenshot shows the noise in DARM changing.

At this time, there is a shift in the alingment of SRM and SR2 in both pit and yaw, and other interferometer optics don't move significantly.  The OMs and OMC suspensions move as well, but that appears to be a result of the feedback. There is also an increase in POP90, a small drop in AS_A DC, and a small increase in AS90.  The signals on all 4 quadrants of ASA36 I change (the in loop signal, where we are using a funny matrix to take care of an offset).  On AS36 B, we see pitch signals, but not much signal in yaw. Also, the RF levels on all quadrants of AS36 WFS change by 5 to 8dB, which could be the result of the alignment shift.  

We looked at conlog for any changes at the time, there were only PI settings changing, we don't think there was anyone in the LVEA or CER at the time.  

The most obvious sign for an operator that this has happened would be that POP90 jumps suddenly in the middle of a lock stretch.  If this happens again, people could (go out of observe) open the SRC1 ASC loops by ramping the gains to zero, then try moving SRM to bring POP90 back to a normal level.  

We have known for some time that the indicated level of CP4's LN2 fluctuates too rapidly to be physically likely and is more probably due to some transient GN2 phenomenon which effects the validity of the differential transducer.  Occasionally this indicated level will deviate enough as to enter into the alarm region and will then get the attention of the CR operator (Corey just asked about this).  I guess we'll need to investigate this more seriously.  Chandra is aware of this issue but is gone until next week.  

The Vacuum Team gets notified via automated text messages of these alarms.  

Until this gets resolved, CR operators only need contact the Vacuum Team Members if CP4 stays in alarm for a prolonged period (say hours). 

(see WP #6332)

Did a few "on/off" states for Robert S. so as to compare spectra while IFO in low noise.  He OK'd the continued running of these pumps during the intended bake-out of pressure gauge PT180 located on BSC8's dome.  As is, the apparatus is ready to go with the exception of the gauge electronics which remain connected and energized.  I am ignorant as to the impact, if any, of removing this while it is energized.  I've been told that this gauge has an in-vacuum PCB that limits the bake temperature to 80C at the CFF and I can just see myself going through the bake exercise only to have to vent and replace the gauge because I had "popped" the internal electronics by removing the external electronics while energized!  Spoke with Patrick T. who is only familiar with the Beckhoff end of things, left msg with Richard M. who might know more pertaining to the hardware.  Mfgr. tech support is not available via phone during off hours etc....  

As such, I am leaving PT180 exposed to the site vacuum volume (YBM) and the pumps etc. running.  The locally mounted turbo pump is isolated from PT180 via a closed metal valve and its exhaust valve is open to the supporting pumps located on the LVEA floor.  I will attempt to start the actual bake-out on Friday - hopefully by then I will have determined how to de-energize the gauge or discounted the concern of removing the external electronics while energized.  

Until we put in a more permanent solution next week, over the holiday weekend here are the instructions to allow the operator to disallow hardware injections from running, say in the event of a GRB.

From the SITEMAP MEDM, on the X-Arm line, select the CAL pull down and press the PINJX button

You will see the CAL_INJ_CONTROL MEDM for the x-arm.

The part we are interested in is the Transient signals portion in the centre of the screen. Open the  filter modlue screen by pressing the TRANSIENT related display button.

The filtermodule can be in one of two states depending upon whether a transient hardware injection is currently runng (not a high probability occurance).

If an injection is not running (the most probable scenario), the GAIN will be 0.000. In this case, turn OFF the output ON/OFF button to the right of the LIMiT switch.

If an injection is running the GAIN will be 1.000, set the GAIN to 0.000 (it will ramp down over 2 seconds) and then turn OFF the ON/OFF button.

At this point wait for the ring-down period of time (nominally one hour).

To permit hardware injections, turn ON the ON/OFF button (in the unlikely event an injection is running, either wait for it to finish or set the GAIN to 0.000 again)

Frequently I find that SRC will not lock properly. I then misalign SRM and look at the spot centering on te ASC_QPD to check the alignment of SR2. Usually I find that it's been driven off diagonally in one direction or the other making it difficult for the loops to align and converge SRM into the cavity. After re-centering I will re-align SRM and re-request SRC_ALIGN and it should lock and converge properly. Maybe some hand alignment will be necessary.