Was alerted at 9:09UTC

IFO was in NLN but could not move into observing due to 2 (attached screenshot) SDF diffs in ALS Y, which I accepted.

IFO reached OBSERVING at 9:21UTC

TITLE: 11/08 Eve Shift: 00:00-08:00 UTC (16:00-00:00 PST), all times posted in UTC
STATE of H1: Observing at 146Mpc
INCOMING OPERATOR: Ibrahim
SHIFT SUMMARY:

The last few hours of the shift turned out to be eventful with a large earthquake hovering at high seismic levels for 2.5hrs which made locking Green Arms impossible.  In the last 30min of the shift, seismic finally looked like it was sort of coming down, so I attempted an Initial Alignment.  X-arm showed signs of life and eventually was aligned, but the Y-arm gave an error about the Camera Centering being too far off.  I ran a DOWN out of the Initial Alignment due to the Y-arm, to see if I could do anything.  I might have messed up the Xarm because when I tried locking again, X-arm was no longer aligned and had no light in the arm (I returned ETMx & TMSx to where they were when it was aligned).  At this point, I'm letting ISC LOCK take over (both X & Y arms are in INCREASE FLASHES).

Unfortunate timing with the earthquake and frustrating last 30min of the shift.

Addendum:  as I log out, atleast the Y-arm now has life after INCREASE FLASHES and can finally OFFLOAD with a quieter ground.  Xarm....looks a little glitchy, but is trying to offload.....and now it is moving on to FINDING IR.
LOG:

• 0029 Out of PCal Lab (tony)
• 0509 EQ Mode activated (nothing on USGS)
  • 0511 M6.8 Earthquake from Timor/Banda Sea/North of Australia
  • 0512 Squeezer unlocks and drops H1 from Observing
  • 0513 Lockloss
• 0729 Initial Alignment attempted.
  • Unfortunately, something appears amiss with the Y-arm.  It has decent flashes but for the last hour or so it hasn't been able to lock for more than a second or two now.  
  • During the Initial Alignment while ALSY was in the ENABLE WFS state, ALSY gives the note about the "Camera centering being too far off.  Need to turn on slowly." And the node would then be stuck at this point.  I don't see any obvious errors on the ALS screen. 

H1 taken to Earthquake mode at 0509utc; started seeing seismic signal on Picket Fence and seismic blrms.  Squeezer unlocked taking H1 out of Observing and then lockloss 1-min later.  It's been about 15min since the EQ waves hit us and seismic is still increasing (ISC LOCK has x&y arms at INCREASE FLASHES).  Marking time as EARTHQUAKE.

H1's been locked 4.25hrs, but the range is 8-9Mpc lower than 24hrs ago (and this is with quiet breezes now vs high winds last night); it took about 3hrs for the range to slowly ramp up to the range of 151Mpc (where last night we were at about 160Mpc with high winds.  useism has been somewhat high and constant the last 24hrs.

TITLE: 11/08 Day Shift: 16:00-00:00 UTC (08:00-16:00 PST), all times posted in UTC
STATE of H1: Observing at 146Mpc
INCOMING OPERATOR: Corey
SHIFT SUMMARY: Maintenance day with higher winds. There have been a few changes that should help lock acquisition - COMM beatnote improvement and ALS_DIFF IR finding. We are back at observing, though our range is a bit lower than it has been recently.
LOG:

0001 Observing                

I wrote a new search in ALS_DIFF to hopefully help with our issues finding IR resonance during this step in locking. Previously the node would go to 3 or 4 past saved offsets, then if there was a little bit of signal it would try to find tune. Successful offsets seem to always be around 400 and 3250 +/- 100, so there's no need to search the entire range. I have it get near these saved offsets +/- 150, then step 30 at a time to look for for any power. If no power is found around this saved offset, then it will move on to the next saved offset range. As long as I move in the same direction, I didn't run into any odd hysterysis.

This ran successfully a handful of times and seems to be decently fast.

TITLE: 11/07 Eve Shift: 00:00-08:00 UTC (16:00-00:00 PST), all times posted in UTC
STATE of H1: Preventive Maintenance
OUTGOING OPERATOR: Oli
CURRENT ENVIRONMENT:
    SEI_ENV state: CALM
    Wind: 13mph Gusts, 7mph 5min avg
    Primary useism: 0.03 μm/s
    Secondary useism: 0.38 μm/s
QUICK SUMMARY:

Maintenance & Winds had H1 out of Observing when I arrived, but H1 was ready to return to Observing

• Within a few minutes of start of shift H1 was taken to OBSERVING by TJ after Naoki finished up Squeezer work.
• TCS Powers on diag_main has an error which TJ mentioned to me & he will address this evening.
• TJ also mentioned the ALS_DIFF node was worked on to improve FINDing IR
• COMM Beat note is now better after in-air table alignments by TJ & Sheila.
• Now that H1 is back to OBSERVING, Robert will be making HVAC adjustments once H1 hits around 155Mpc.
• Tony continues work in the PCal Lab.

Struggled a bit with high winds but it calmed down recently and we are back to Observing. There were some SDFs related to Keita's alog74070 and the CO2 rotation stage work.

I disabled 1Hz LPF which was somehow present in the SUM (not NSUM) of the QPDs and WFS DCs that are used for normalization of PIT and YAW. These LPFs ruin the rejection of the intensity noise for QPDs and WFS DCs even at 1Hz.

Here's a list of the filter modules where the 1Hz LPF (all FM1) was disabled: H1:ASC-OMC_[AB]_SUM, H1:ASC-REFL_[AB]_DC_SUM, H1:ASC-AS_[AB]_DC_SUM, H1:ASC-AS_C_SUM, H1:ASC-POP_[AB]_SUM, H1:ASC-[XY]_TR_[AB]_SUM.

Note: Without LPF, when large clipping action is going on and SUM goes very small, I can imagine that PIT and YAW become somehow worse than with LPF. If that ever becomes a problem, enable 1Hz LPF.

Note 2: Green WFS DCs and QPDs don't have the BPF for SUM.

I moved a 48 port fiber switch out of the MSR back into the test stand per WP11515.  The switch had been put in the MSR for a short term host for the file servers, but that did not work out.  The switch has been empty since then.  It is needed in the test stand to help test some high bandwidth optics.

I checked the FC GR SUS/VCO crossover as shown in the attached figure. The crossover is about 60 Hz which is much larger than the blue reference on December 2022. This higher GR SUS gain could cause the instability during the transition from GR to IR. I reduced the GR SUS gain from 1 to 0.5 and the crossover became similar to the blue reference. I updated the SQZ_FC guardian to change this gain. Let's see if this helps the recent failure of FC GR to IR transition. 

Sheila, Vicky, Naoki

Summary:  The sqz blrms at 80 Hz became worse October 27th and we may be able to improve the sensitivity in that band by adjusting the FC detuning.  We do not know exactly why the high frequency squeezing was restored to a better and more stable level last night.  

The first attachment is a set of SQZ BLRMs trends from the last month.  Focusing on the blue line (1.7kHz BLRMS) you can see that the crystal move on Oct 17th increased the level of squeezing available, but also caused the squeezing level to be less stable. This frequency band doesn't contribute much to the range, but you can see that the squeezing level at 1.7kHz seems to correlate with the range, presumably because it tracks the mid frequency squeezing level.  

On October 27th, Naoki centered the beam on FC2 73777, and the 80 Hz BLRMS (red trace) jumped up to +1.5dB, indicating that the squeezer is degrading the low frequency sensitivity of the IFO.  We would like to try adjusting the FC detuning to see if we can get back to the earlier performance. 

Looking at the second attachment, (again focusing on 1.7kHz blue sqz BLRMS), you can see that since the crystal move the squeezing level has drfited as the crystal changes.  The CLF_REFL_RF6 power slowly drops over time, meaning that there is less of the generated 3MHz sideband (and less nonlinear gain).  Every time the OPO temperature is adjusted, the RF6 level steps up as the nonlinear gain is restored.  The bottom left trace shows the RLF transmitted by the filter cavity, which also increases each time the temperature is readjusted, because the generated RLF sideband is increased as the nonlinear gain is restored.  In last night's lock, the FC transmission was larger than it has been in the past month, although the CLF 6MHz wasn't especially large.  We don't know why this is, but it's interesting that it corresponds with the restored (and seemingly stable) 4.5dB of squeezing.  Other things that we've looked at to explain last night's squeezing level were AM4 alignment (shifted a few days ago and doesn't seem directly related to the improved squeezing), and the OPO PZT voltage which jumped up this morning after the IFO was no longer locked to near 100V.  

As part of the preparation to add a PMC to the squeezer, a new squeezer EtherCAT chassis 5 was installed. As a side effect, the RF channels with prefix H1:SQZ-SHG_TRANS_RF35 have changed to prefix H1:SQZ-SHG_TRANS_RF24, but only in the slow controls.

The slow controls for the squeezer also got a new H1:SQZ-PMC subsystem that is in an error state due to the fact that the the new chassis is not connected to the field chassis yet.

TJ, Camilla. After we found both CO2 output powers have been changing since the timing master swap 74025, we recalibrated the CO2 rotation stages this morning ~ 16:20UTC. Last done in September 72961. 

Followed the README.txt file in /opt/rtcds/userapps/release/tcs/h1/scripts/RS_calibration/. We did this with annular masks in. I'd previously tried this without using the matlab fitting code and hadn't been able to make a good fit. 

J. Kissel (encouraged by S. Dywer, D. Sigg)

Continuing along with the plan of building up a complete picture of the PSL frequency noise as a part of understanding the future *effective* frequency noise delivered via fiber to a seismic platform IFO for longitudinal control (aka SPI L, started elsewhere using the PMC LHO:73905), today I switch tactics and look at the main IFO's frequency stabilization servo. 

Today's primary metric is "IMC_F," i.e. the demodulated output of the PD in reflection of the input of the input mode cleaner, a.k.a the IMC REFL PD, which is a part of the global frequency stabilization servo (FSS) for the detector's pre-stabilized laser (PSL) frequency. For more details on how the global frequency stabilization works, check out the most recent published info on it from Craig Cahillane's paper P2100219 and Chapter 3 section 4 of his thesis P1800022. In short, it's a cascaded set of loops that use 
   - the in-vac, rigid-body ~1 m, reference cavity within the PSL, 
   - the suspended in-vac ~16 m input mode cleaner, then 
   - the common motion of the suspended in-vac 4 km arm cavities themselves 
as frequency reference, with ~300 kHz, ~60 kHz, and ~20 kHz UGF, respectively.

This IMC REFL PD's output is converted into a control signal to be pushed to the PSL laser frequency with the IMC "common mode board" analog servo. The output of that analog control filter is digitized and converted into Hz. Because the FSS + IMC + CARM bandwidths are all above 20 kHz and I'm looking at the PD's output below 1 kHz, we're very much in the "infinite gain" region of the control signal. Thus -- assuming we know the frequency actuator calibration (i.e. the voltage-controlled oscillator drive chain) -- then this control signal can be scaled directly to frequency noise without knowledge of the loop. I didn't do anything here, I'm taking advantage of the filtering that's already in place from the original 2013 characterization / calibration -- see LHO:5945 for details. I'm just reading out the channel H1:IMC-F_OUT_DQ, which is actually calibrated into kHz, and multiplying by 1e3 [Hz/kHz].

I look at this metric in two different states of the detector: 
    2023-11-03 10:53 UTC -- Reference Cavity and IMC are stability locked vs.
    2023-11-03 10:27 UTC -- just prior, when the IFO is in nominal low noise, i.e. FSS and IMC and CARM are locked

As a reference, since folks are not typically used to looking at the frequency noise of the stabilized PSL below 10 Hz, I also show the cartoon trace of the canonical frequency noise of a "free running" NPRO, 100 Hz/rtHz at 100 Hz, and falling proportionally with frequency, i.e.
    df = 100 [Hz/rtHz] * ( 100 [Hz] / freq )

This puts the "free running" PSL NPRO frequency noise, if it weren't stabilized at all, at 1e5 [Hz/rtHz] at 0.1 Hz -- i.e. the microseism.

OK, so, to the attachments. I attach four cascading plots of frequency noise where I'm adding more and more traces to the same plot as the plot number increases. Open them all up in tabs, and then walk through them slowly as follows:
(I) 2023-11-03_H1IMC_F_FrequencyNoise_01.png -- this shows IMC F stabilized with just the RevCav+IMC used as reference in blue. This is our starting point for building up an understanding, and I keep the NPRO free-running frequency noise in dashed sea-green. In this attachment you can already kind of guess what's happening.
    (1) Above 10 Hz, there's some noise that falls as (1/f). Littered throughout that noise, are some acoustic peaks scattered throughout. BUT
    (2) Below 10 Hz -- and particularly around 0.1 Hz we see a quite familiar shape -- that of ground motion around the microseism.

(II) 2023-11-03_H1IMC_F_FrequencyNoise_02.png -- this shows IMC F in nominal low noise, stabilized against with RevCav, IMC, and CARM used as reference. The changes actually help further ellucidate and validate the story:
    (3) Above 10 Hz, the (1/f) noise doesn't change. Fine.
    (4) Below 10 Hz, the part of the spectrum we thought was seismic is indeed changed dramatically:
        (a) The 0.5 to 10 Hz region looks to convert 
             (i) from a noisy-ish displacement noise -- that looks a lot like the shape of 0.5 to 10 Hz noise budgets from HAM2-HAM3 optic motion -- dominated by the HSTS MC1, MC2, and MC3 OSEM sensor noise re-injected through the damping loops
             (ii) to better displacement noise of optics quadruply suspended from a collection of BSC-ISIs
            retaining the the shape of first two quad longitudinal resonances at 0.43 Hz and 1.0 Hz, and then falling into the ever present (1/f) noise -- now "earlier" / lower in frequency at around 2 Hz.
        (b) Below the 0.15 Hz microseism peak, we see kinda also what we expect -- that the BSC-ISIs are re-injecting a lot of local tilt into the local longitudinal motion of each of the test masses, and is creating common longitudinal noise in CARM -- increasing the spectral density by a factor 50x, but the RMS only by a factor of 2x because the 0.15 Hz microseism peak dominates the RMS.

To further re-enforce that this is real seismic motion -- I recall that we've done a lot of work converting all ISI's GS13s -- include those of the BSC-ISIs suspended the QUADs -- into local suspension point motion -- and then even further converting those local suspension point motions into the basis of the IFO cavities.
    H1:OAF-SUSPOINT_CARM_OUT_DQ
is that ISI GS13s for the QUADs converted into local Sus. Point and then to CARM all coherently. So, if we convert this to frequency noise, via 
    >> c = 2.9989e8;
    >> L = 3995;
    >> lambda = 1064e-9;
    >> (c / (lambda*L))
        ans =
            7.0551e+10 [Hz/m]

and because the all GS13s are calibrated into nano-meters, we get the final "frequency noise" calibration of these channels as 
    >> 1e-9 * (c / (lambda*L))
        ans =
            70.551 [Hz/nm]

then we land on the black trace in 
(III) 2023-11-03_H1IMC_F_FrequencyNoise_03.png 
    I'm not gunna lie, I'm *delighted* at how well this matches up with what I expect!
    (a) in the 0.5 to 10 Hz region, comparing black to red, we confirm that the QUAD modes are amplifying the ST2 motion on resonance at 0.43 Hz and 1 Hz --and maybe 2.7 and 5 Hz.
    (b) but look at how identically matched the motion is from 0.1 to 0.2 Hz! WOW Wow wow.
    (c) We can tell though, that the GS13s themselves are not reporting real motion below 0.1 Hz -- they're are either "tilt dominated" or "self noise" dominated or something.

But, for me, this confirms without-a-doubt the thing that Daniel and Sheila "knew all along" that "at the microseism, the laser PSL frequency is just following the ARMs."
Maybe someone has plotted this up before for, but I couldn't find it. Anyways, very convincing, and very enlightening.

Finally, 
(IV) 2023-11-03_H1IMC_F_FrequencyNoise_04.png compares this BLUE and RED answer against our bigger picture question -- how does this relate to what we need for SPI L? It looks like it may meet our "bare minimum" needs, but will become interesting if we're trying to push the performance of the SPI.
     Very interesting indeed....

Anyways, just for another point of reference, I also show the ground motion at all stations in X and Y -- since this is another thing that folks are used to looking at (rather than the SUSPOINT CARM). This gives you the feel that this measurement is during a "medium" microseism day, with little-to-no wind.