Just before we lost lock due to either making power up and down or injecting anti-squeezing, I went to the floor to phase the new LSC-REFL_B.

Unfortunately I needed to advance the LO phase by about 30deg or about 9.3ns, which I couldn't do with the delay line, so as a bandaid solution I delayed it. I fine tuned it down to (16+1+1/2+1/4)ns=57.5deg.

360deg*9E6Hz*17.75s=57.51deg

This means that right now H1:LSC-REFL_B_RF9_[QI] is equivalent of H1:LSC-REFL_A_[IQ] except for the sign of B_RF9_Q (no screenshot of fine tuning measurement as the lock was lost during the measurement).

We need a longer cable for the patch panel to the demod than the current one to delay the photodetector side. For safety margin we'd like to add more than 30 deg, so let's say 40 deg ~ 12 ns ~ 2.4m longer than the current cable, then use delay line to delay the LO side a bit.

TITLE: 11/30 Day Shift: 16:00-00:00 UTC (08:00-16:00 PST), all times posted in UTC
STATE of H1: Commissioning
INCOMING OPERATOR: None
SHIFT SUMMARY: 7.0 Magnitude earthquake from Alaska put us out for a bit, but commissioning is now back in full swing. The "Large Earthquake" button was pressed just before the EQ arrive, thanks to a SEISMON alert. The "Recover EQ" button was pressed to bring it all back. The useism is still around the 0.5um/s range and wind 0-15mph. To deal with the elevated useism, Jim is trying out a some new blend filters that are in the USEISM state of SEI_CONF, this is the current state.
LOG:

1556 Karen to MY to clean
1626 Karen leaving
1637 Vanessa to LVEA to start cleaning
1803 Kyle to LVEA to check on pumps
1818 Kyle out of LVEA and heading to MY
1833 Keita to LVEA to check on Relf DC signal
1914 Keita out
1952 Kyle back
2058 Fil, Peter to MX
2525 Fil, Peter back
2151 Dave to end stations to get card info
2202 Kyle, Gerardo, to ends, mids, and LVEA to drop off equipment
2233 Dave back

2325 Sheila, Haocun to LVEA to SQZ bay

I've been asked about about the 1.1hz peak on HAM3 a couple times today, so I should post about what I have done. To start, the peak is still there. A couple weeks ago, I tried swapping around the ISI interface chassis and AA chassis to see if this changed anything. There are 2 of each of these chassises, all of the corner 1 & 2 sensors are plugged into the first interface chassis, which is plugged in the the AA chassis, while corner 3 is plugged into the second interface chassis, then the second AA. I had hoped that this would show one of the channels on the interface chassis was bad, but no luck. I could try swapping other rack components, but it quickly starts getting much harder to do the swaps.

Separately, this morning during the Alaska earthquake, I did a repeat of Keita's measurement from alog 45182. This was a much bigger earthquake than the one from Keita's alog, and after talking with him, it seems like this matters. In the bottom left of his first plot , the peak frequency and Q are a bit different for the v2 cps to v2 gs13 transfer function than for the other 2 corners, I've blown this tf up in my first atttached plot. In my second image, the same transfer functions larger Alaska earthquake this morning don't show the same difference. The Q is a little lower for the blue V2-V2 trace, but the frequency of the peak and Q are more similar to the other 2 corners. Keita suggests that this means the difference depends on the amplitude of the drive, which is why my earlier driven transfer functions didn't show anything, too get good coherence I had to drive hard enough that it buried whatever is allegedly causing the V2 GS13 to behave differently.

I didn't catch it from Keita's plot before, but I think the fact that the resonance frequency and Q change with input motion makes me kind of suspicious that this is a bad GS13, either sticky or a weak flexure. I don't really know how to confirm this, yet. 

This morning I looked at phasing AS_A_36.  I want to go back and re-do it, but I left it in a situation where the phasing is 15 deg different than it had been, and the AS_A_36_I pitch and yaw signals (which are what MICH ASC transitions over to in OFFLOAD_DRMI_ASC) were around zero, as opposed to the large DC offset I found eith the original phasing.  Phasing to make the 36 signals' zero crossing be at the same place as the AS_45 zero crossing (MICH ASC was locked using AS_45 at the time of the phasing) made the OFFLOAD_DRMI_ASC state much more reliable.  I also copied the gentler turn-on scheme that we use in that state over to the CHECK_AS_SHUTTERS state, which also turns off then back on the MICH ASC.

After this, I was able to get to PREP_DC_READOUT_TRANSITION several times reliably.  I lost lock several times trying to go to DC readout, or trying to go halfway to DC readout.  I measured the DARM loop while I was still RF-only locked, and it looks very strange.  Once we come back from this earthquake, the DARM loop is the first thing we need to look at.  (This is consistent with, for example, the locklosses from last night that Georgia posted in alog 45590, where the first oscillation happens in DARM).

Screenshot shows the weird DARM loop shape.  I didn't save the coherence as a reference, and the 'live' measurement includes a lockloss, so you'll have to trust me that the measurement had good coherence above 10Hz, so including the weird peaks between 10-20Hz, but not necessarily the peak below 10Hz.

Also, I shortened the ramp time of some of the lockloss trigger turn-offs.  I think I'm seeing that the bounce mode of ETMY (that we have the DAMP_BOUNCE state for, once we're on DC readout) is ringing up more often now.  I'll likely change them back to ~2 or 3 seconds, from the 1 second they are at now.

Seismon gave us a heads up of a 6.7 from Alaska about a few minutes before it hit us. We stopped relocking and hit the Large EQ button. Only the BSC ST2's have tripped.

Georgia Craig Dan Brown Hang Sheila Haocun Jenne

Summary: We can get to DC readout, but are not very stable once we transition to DC readout, and we probably need to double check the DARM loop shaping in the transition to DC readout. 

• Today DRMI locking is better than it was yesterday.  PRMI has to be well aligned for DRMI to grab lock, this seems more critical than in the past.  
• Jenne logged a scan of SRCL offsets we did. 45579  We haven't had a chance to double check the calibration of this.  We've removed the SRCL offset from the DRMI locking.  
• We have had difficulty with the state OFFLOAD_DRMI_ASC, similar to what we saw last night.  45565  The problem is that when we turn off the AS45->BS feedback and try to re-engage AS36->BS there is sometimes a large transient.  This might be worse when the alignment of DRMI isn't great.  Jenne and I tried adding the -20dB filter and turning off the integrator and for a gentler engage of the loop.  Later Craig and Georgia edited this again to also turn the inputs off and back on.  This still breaks the lock sometimes.  
• Jenne moved the alignment offsets around and found a set that gave us a recycling gain of 48-49 with 2W input power. 45584
• After that Hang reset the green QPD offsets and the camera offsets, and I went to ISCT1 and realigned the beams onto the cameras and the COMM and DIFF beatnotes.  The COMM beatnote is about 2dBm, DIFF is about -9dBm.  
• the violin mode we are calling ITMY move 7 seems to have had a phase change, we have damped it with a sign flip consistently today so we put this in the guardian. 
• We had several locklosses trying to power up to 10W.  Georgia will add a comment about that.  THe problem was possibly the notch mentioned in Jenne's alog that she says cannot be used to power up. 
• Haocun checked that the squeezer alignment is still good with the change in IFO alignment.  
• We had trouble staying locked long at DC readout, so we check MICH, PRCL, SRCL loop gains just before the transition. They all seemed fine. 
• We also checked the phasing of POP45, Craig changed it from -5 degrees to +7.  We also remeasured the POP9I to SRCL matrix element, which we found to be a factor of 5 smaller than what we had in the past, 0.0047 is the new value we found.  We tried this once but haven't added it to the guardian because in the past we have seen this changing from lock to lock. 
• We weren't able to stay locked for long on DC readout, and have also lost lock in the transition to DC readout.  Craig looked at one of these locklosses and saw that it lost lock while the gain was increasing for DARM
• If the morning team is in the same position we are in tonight, it might be a good idea to check the DARM loop gain during some of the loop reshaping that happens in the transition to DC readout. 

Here is a comparison of POP build ups in counts, un-normalized now and before the last few weeks of work.  

TEAM SQZ

--- Pump laser multimode fixed and beat note optimized (Terry, Sheila, Nutsinee) ---

The setting that Daniel set on Monday (current 2.122, temperature 27.63C, see alog45518) no longer worked the next day. As Terry and I tried to optimize the beat note it keeps disappearing. Finally we could find the beat note on the wrong side of the PSL (above) but not below which seems to suggest that we were close to mode hopping region. Sheila and I then turned down the current to 2.116 A and turned up temperature to 29.57C. Once we were operating at a good region the beatnote strength measured -15.5dBm out of the detector (the best we've ever had), -19.2dBm out of the pre-amp monitor, -2dBm on the medm screen with ~183uW from each arm. BBPD volt was 1.58 V. The polarity has to be flipped when use beat note error to feed to PSL PZT (was negative, now positive). 

We checked the PLL lock point for both ALS. One of them still locks at the same place as it did before the vent which indicated that the PSL frequency hasn't changed.

--- OPO locking ---

Self locking is now much worse than before now that green is better mode matched. To get any reasonable scan without asymmetry I'd have to shoot in .4mW into the OPO. PDH Vpk = 75.2 mV. 80MHz phase still okay. Attached a screenshot of a setting that works at 7mW input to the OPO (according to a model, we should get the same cross-over and UGF as before). I haven't had time to characterize anything before I left on a vacation. I suspect that the better mode match could cause more problem locking at high input power if the intensity isn't stabilize. Note that if the OPO is too noisy, you will have a hard time engaging the EOM. Without EOM engage it would be almost impossible (probably impossible) to lock LO.

--- LO locking with Mephisto LO (Daniel, Nutsinee) --- 

Step 0) swap the PZT output back from OPO pzt to LO pzt. Actuating on OPO PZT with error signal from pump laser LO would be shooting yourself in the foot, as someone might say...

Attached the setting that works as of yesterday. Compensation and boost1 can be turned on. That would have a UGF of 4.6 kHz. The UGF can be pushed further if we install the LO notch filter box (which we had and was removed...)

--- Homodyne Recovery (Haocun, Nutsinee) ---

I tried to leveled both LO and seed beam a bit better to 4 inch height before I realized that the Homodyne diodes are slightly higher than 4 inch. Both LO and seed should be somewhat leveled but shoot up to HD in a small angle. The power somehow became imbalanced while I was trying to optimize the fringe visibility and got stuck at 97% (PD A) and 94% (PD B). Haocun rescued us and the best we had was 98% at least on one of the diode (I think this was true for both diodes).

More squeezing measurements were taken today, as shown in the following table, and plot attached.

• The last measurement was from yesterday, and is probably biased because the SEED beam was clipped on the beam splitter, so I didn't include it into the plot.
• The EOM was noisy and hard to lock with high green power (> 17mW input)
• The blue stars are the NLG I measured with SEED=1.2mW input, and the red circles are the theoretical NLG values calculated from the OPO threshold from alog 45528. I used SEED=1.1mW that day, and the green beam transmission shouldn't change. It seems there is a systematic difference of the NLG measurements, and we should look into this. (Use a scope to measure. Also, from the past we found that the NLG was related to the different dual resonance peaks overlap.)   
• The results are confusing that with such many dB's of squeezing, the level of anti-squeezing of several measurements are low. I cannot find an optical efficiency and phase noise value combination to match the squeezing and anti-squeezing. One probable reason I can came up is that there were not enough phase to cover the most anti-squeezing.
• If we trust the squeezing levels, the optical efficiency should be around 80%. The loss budget gives ~85%.

As outlined in LHO alog 44418,  front end CAL model changes included change in Pcal model to include the in vacuum optical loss  measurements carried out during the vent (LHO alog 40894). 

New Simulink model and MEDM screens were generated  (shown in LHO alog 45559) based on the equations outlined in document T1800046.

Upon review, we found some mistakes in the new changes. The attached screenshots show the correction needed to reflect the scheme described in T1800046.

The first attachment is the corrections required in Simulink model and those corrections need to be propagated to the MEDM screen as well. The second attachment highlights the naming inconsistencies in the MEDM screen. 

Chandra requested that all 8 beam-tube vacuum alarm levels be reduced from 8.0e-08 Torr to 5.0e-09 Torr. Details in attachment. System was restarted.

We've walked the SOFT offsets, as well as PRM and SRM pointing to get ourselves back to a PRG of ~48 (we were closer to ~39 the last few days since the vent).  It's not totally clear why the QPD setpoints needed to be made different, but we haven't checked how different (if at all) our input pointing is.

Hang reset the green initial alignment setpoints, and we set the IR trans QPD offsets, and the POP_A QPD offset, and now have the full ASC locked at this alignment.  Hang also rephased the AS72 while we were locked with this new alignment before we closed the loops.

We're not at the 0,0 point of AS_C right now (I've put an offset in that QPD as well), but we actually should move SR2 and SRM such that we're back to centered on AS_C so that we're going through the OFI nicely.

While sitting at 2W full lock for several tens of minutes, the 3.4Hz instability that comes from PRC2 Pit started to ring up.  For right now, we've engaged the lowpass that has the notch for this mode in it (it's an ELP6 that has a notch at 3.4Hz in PRC2_P) and that fixed things, but we'll likely want to remove this lowpass for the power-up.  Probably this isn't usually a problem for us since we don't spend large amounts of time here, and just get to LOWNOISE_ASC where that lowpass/notch is engaged in the guardian.

   Posted are the monthly BRS Drift Mon and BRS Aux Channel plots. Drift mons look OK, as neither is trending towards it's boundaries. 

We had a problem with the ISS tonight, which was probably caused by the second loop.  We haven't taken the time to look into it but have disabled the second loop by turning off the input H1:PSL-ISS_SECONDLOOP_ENABLE, which made first loop stop railing.  Attached is a screenshot. 

J. Kissel
ECR E1800246
IIET 11305

I've (nearly*) completed several MEDM screens that are reflecting the new upgraded time-dependent correction factor calculation in the fron-end calibration pipeline, in support of splitting the calculation of kappa_pu into kappa_p and kappa_u, plus the new supporting calibration line, as described in T1700106. The changes were installed into the CAL-CS front-end model moons ago (see LHO aLOG 44459, and G1801594), and thus I'm just completing the effort started in LHO aLOG 44752 to develop a user interface.

This system doesn't work yet (we also don't have an IFO back yet to test it), but now that the use interface is complete enough to debug the parameters and settings and we can finally move forward.

*There're a few bug fixes and a bit of aesthetic cleanup let to be done, but we'll get to that in due time. 

These new screens live in the following corner userapps repo:
    /opt/rtcds/userapps/release/cal/common/medm/
with names akin to the file names of the respective screenshots attached below.

Also in due of time, I'll update G1801594 to reflect the new changes.

WP 7968
Over the course of developing these screens, I'd identified a few tweaks, extra EPICs records, and bug-fixes to the front code along the way, so Dave graciously agreed to support a couple of h1calcs model, and subsequent DAQ, restarts yesterday evening, a. la. LHO aLOGs 45539 and 45543.
These model changes have been committed to 
    /opt/rtcds/userapps/release/cal/common/models/
as a part of the clean-up.

Relevant FRS ticket: 4526

Incursion activity to fix the problem: alog 45391

I checked if the ASC REFL_A DC segment 4 problem described above was fixed for good by walking the bean on the WFS using RM1 from one segment to another so the beam is contained by only one of the four segments. At first I was glad to find that the problem was gone (SEG1, 2, 3 and 4 showed about 8200, 8300, 8100 and 9000) but later found that the digital gain of 2 was somehow put back in after our work in the above alog.

This means that the problem came back after we thought that we fixed the problem.