Hi everyone it's been a while since I posted!

Julia Rice and I had a chat with Peter F today about motivating the electrostatic violin mode damper we've been building at Syracuse. We came away with a couple of questions to answer:

1. What's the actual displacement of the fibers when the violin modes are excited, and is our shadow sensor noise low enough to measure the modes.

2. What is the origin of the two violin modes per fiber? Is it fiber ellipticity built into the fibers themselves or asymmetry their attachment points AKA the ear/horn. If it's the former, we'd expect the main oscillation axis of the modes to be randomly oriented. If it's the latter we'd expect 4 modes per test mass to be dominant in DARM, and 4 modes to be much weaker, we'd also perhaps expect to see a regular spacing between pairs of modes.

I still have remote access (thanks Dave for helping me get my ports sorted with NoMachine), so I logged in to grab some plots of violin modes excited vs damped times.

I found an excited time and a quiet time by plotting the ISC_LOCK guardian state and the monitor for ITMY MODE5 which I know is a problematic mode (first plot). I used 5 November 2025 17:59:09 UTC, gps 1447264767 for the modes excited time since this was after many locking attempts with the mode5 monitor was ~20k counts. I used 17 November 2025 20:57:12 UTC, GPS 1447448250 as my quiet time AKA t=0 on the first plot.

Second plot is the money plot! When the modes are excited their amplitudes in DARM range between 1e-16m/rt(Hz) and 1e-14m/rt(Hz), we can roughly convert this to fiber displacement by multiplying by the fiber to test mass mass ratio. We did some back of the envelope calculations with Stefan and there's a factor of 2pi in there somewhere too.

I labelled the plot referencing this handy violin mode spreadsheet from Rahul et al, which set me thinking about the violin mode origin in question 2 above...

I can see the regular spacing with one large mode and one smaller mode in the ETMs for the excited spectrum. BUT it's not like there is 4 modes visible in DARM and 4 that are totally anti-aligned: the smaller modes are still very much visible in the DARM spectrum.

ETMX9 (516.781 Hz) has a small buddy ETMX7 (516.678)
ETMX1 (511.2993 Hz) had a small buddy ETMX6 (511.18)
ETMX3 (508.938 Hz) has a small buddy ETMX4 (508.844)
ETMX8 (507.36) has a small buddy ETMX2 (507.493)

For the first 3 pairs, the bigmode is higher in frequency, not sure if significant. Also, in the damped spectrum the larger modes can become sufficiently damped that the smaller mode is now louder in DARM, I wonder if that might be because we have spent many more hours tuning and damping the settings for the large modes?

There are also some modes that break the trend: ETMY1 and ETMY6 are very close together.

For the ITMs, things are less clear. ITMY1 and ITMY2 are similar height in DARM, same with ITMX8 and ITMX7, and ITMX3 and ITMX5. Things are confusing around the clusters of modes 504 Hz and 508 Hz.

I had previously always assumed that ITMY mode5 and mode6, the extremely close modes, are on the same fiber. But perhaps mode5 is on the same fiber as mode7, and mode6 is on the same fiber as mode8!

To better visualise the data, I made a table of the mode "buddies" and their frequency separations here, check it out if your interested. There is a goofy plot in there with no x-axis and mode spacing on the y-axis. The frequency spacing of pairs of modes is centered around 0.1Hz.

I do think there is scientific value to having a shadow sensor at the fiber: it's clear that the modes are not perfectly aligned or anti-aligned with the DARM axis, I was able to identify all 32 modes. In the egregious case of ITMY, at least 7 of the 8 modes are large in DARM. That being said, the regularity of the spacing means I am quite confident (for most modes!) which two modes are on the same fiber.

Please let me know if I've missed something or if you have any thoughts!

TITLE: 04/03 Day Shift: 1430-2330 UTC (0730-1630 PST), all times posted in UTC
STATE of H1: Planned Engineering
INCOMING OPERATOR: None
SHIFT SUMMARY:

Final HAM1 checks were made by Keita & Sheila in the morning.  ISCT1 was pulled away from HAM1.  

There is new anti-slip (yellow) paint on part of the ramp at the OSB door enttrance ramp near the Control Room.  This paint should be dry, so it some signs have been removed.  (Waiting on more paint to finish job.)
LOG:

• 1445 optics lab technical cleaning (kim)
• 1448 X-arm tumbleweed check (randy)
• 1510-1657 alum-A-lift moved from staging to hi-bay to West Buy (randy)
• 1541 technical cleaning lvea (kim)
• 1545-1555 checking mega cleanroom (sheila)
• 1603 -1828 Opening HAM1 covers + light pipe + moving ISCT1 out of the way  (keita.sheila.randy)
  • They had to put back up the laser safety barriers for ther South Bay & reconnect isct1
  • 1738:  Light pipes closed, HAM1 covers closed, taking down laser barriers.
  • 1758:  Sheila moving over to SQZT7 & going to local laser hazard
    • 1825 SQZT7 work is done & Sheila is out of the LVEA. Laser is still ON, so Sheila can continue SQZ work from Control Room.  Sheila will turn this laser off at the end of the day--Will want to turn this laser OFF before venting so we can return to full Laser Safe (Operations Laser Safe?).
  • 1801-1828:  Moving ISCT1 away from HAM1 (first have to crane a door out of the way and then they can forklift) [randy, keita]
• 1625-1739 Taking photos of BSC3 for Neil (carlos)
• 1632-1749 Recon of racks:  susR2@ham3, seiC2@cer (jeff)
• 1634 Anti-Slip coating appliction to OSB door entry near Control Room (tyler)
  • Able to paint the sides, but need more coating material for the center portion.  This material takes ~4hrs to cure.
• 2001-2115 Survey work on back-side of the Y-arm (bubba)
• 2030-2046 Checking susR1 rack check near HAM2 (jeff)
  • 2052-2057 returned to rack (jeff)
  • 2204-2208 back to susR2
• 2224 -2233 H2 Bldg walkabout (Josh F)

J. Kissel, D. Barker
WP 13140

Executive Summary: Dave and I created, compiled, installed, and started running the SPI pathfinder's front-end software model -- h1spih23.mdl -- on the h1seih23 front-end computer.
ALL DAC OUTPUTS are 0.0, and saved in the SDF system as such.

Details:

With all the work done ahead of time in 2024 G2402138, subsequent work on the LHO test stand in 2025, and now armed with the real, final ADC/DAC channel usage (LHO:89775), I've *easily* 
    (1) copied over ${userapps}/spi/x1/models/x1spih23.mdl to a new ${userapps}/spi/h1/models/h1spih23.mdl
    (2) made sure Matlab had access to the library part directory under ${userapps}/spi/common/models/
    (3) updated the ADC inputs and DAC outputs to match the production channels / analog electronics,
    (4) commited to svn, making sure to turn on the 'Id' and 'HeadURL' properties, and
    (5) successfull compiled the top level model

The model rate is "32k" (2^15 = 32768 [Hz]), in  order to comfortably support modulation / demodulation at 4096 [Hz].
We're storing most of data processed from the ADC to frames at 2048 [Hz], tho. 
Only 8x channels at the full 32768 [Hz] rate -- 60x 2k channels in total.
    - 4x related to the DAC output, so we can keep track of the digitally requested 4096 [Hz] LO signal (original digital clock, the phase shifted COS and SIN and clock to be sent to the DAC), and
    - 4x related to the raw demodulated output of the IFO PDs.

Dave assigned me DCUID = 187

The model has 5x shared memory (SHMEM) IPC senders:
    - H1:SPI-H23_DIFF_L :: differential length change between ISIH2 to ISIH3 in IFO X coordinates (in the XY plane, at ISI's Z = 0 coordinates).
    - H1:SPI-H23_TO_ISIH2_RY :: The rotation of ISIH2 in pitch (+RY follows the right-hand rule; positive RY rotation in the direction of one's fingers with one's right thumb pointed in +Y; the +X side of ISIH2 goes more -Z and the -X side goes more +X)
    - H1:SPI-H23_TO_ISIH2_RZ :: The rotation of ISIH2 in yaw (+RZ follows the right-hand rule; positive RZ rotation in the direction of one's fingers with one's thumb pointed in +Z; CCW if looking down on the chamber from above.) 
    - H1:SPI-H23_TO_ISIH3_RY :: The rotation of ISIH3 in pitch
    - H1:SPI-H23_TO_ISIH3_RZ :: The rotation of ISIH3 in yaw

Libary parts (and revs)
    - spi/common/models/SPI_LIBRARY.mdl  rev 29747
    - isc/common/models/QPD.mdl rev 18653
    - cds/common/models/rtdemod.mdl rev 31460
    - cal/common/models/CAL_LINE_MONITOR_MASTER.mdl rev 28567

Once Dave installed and turned on the model, I made SDF files for the model softlinks the userapps repo:
    /opt/rtcds/lho/h1/target/h1spih23/h1spih23epics/burt$ ls -l
    lrwxrwxrwx 1 controls   controls      66 Apr  3 16:08 OBSERVE.snap -> /opt/rtcds/userapps/release/spi/h1/burtfiles/h1spih23_OBSERVE.snap (rev 34930)
    lrwxrwxrwx 1 controls   controls      63 Apr  3 16:08 safe.snap -> /opt/rtcds/userapps/release/spi/h1/burtfiles/h1spih23_safe.snap (rev 34929)
and did the same for the filter file:
    /opt/rtcds/lho/h1/chans$ ls -l H1SPIH23.txt 
    lrwxrwxrwx 1 controls controls 59 Apr  3 16:12 H1SPIH23.txt -> /opt/rtcds/userapps/release/spi/h1/filterfiles/H1SPIH23.txt (rev 34931)

Next up MEDM screens!

Jeff, Josh, Erik, Dave:

We started the h1spih23 model for the first time on h1seih23 at 15:51 Fri 03apr2026. It has been allocated dcuid=187.

Erik added this model to CDS puppet. I did a build, rev-update, build-rev-locked and install. Erik then started the model on h1seih23 with no problems.

h1spih23 has not been added to the DAQ, therefore it is missing its DAQ channels and has a bad DAQ status.

h1spih23 runs at 32K. It has 5 SHMEM IPC senders running at that rate. SEI receivers will be added at a later date.

I have added h1spih23 to the CDS Overview (some hand editing was needed since it is not in the DAQ running config yet).

Next week we will add this to the DAQ.

J. Kissel

Now that we have a better handle on how much rack space the SPI pathfinder will take from making it fully functional in the optics lab (see LHO:89269) I've uploaded a new version of the rack layout design to G2401479-v3 -- a copy of which is attached here for convenience.

Then, with this in hand, I went out to SUS-R2 and:
    - Finally removed all evidence of the iLIGO style U-height labeling, 
    - Re-labeled the rack in the standard 42U rack convention with super large font for ease of reading, and
    - installed rack screws in all locations where the SPI chassis will need them.

The first two pics show BEFORE (labeled with counting starting at the top) vs. AFTER (labeled with counting starting at the bottom). In this modern convention, forced on this old rack, u-height "1" is missing; true for all of these racks (e.g. true SUS-R1 as well; see LHO:83156).
The last pic shows U18 thru U2 with all the rack screws and sockets prepped.

I'm in the process of feeding this info back to Dean so he can update the Altium version of the SPI wiring diagram D2400111.

J. Kissel

More experience with the SPI in the optics lab yields an explicit ADC / DAC channel usage list for the SPI pathfinder. I've uploaded it as a part of G2401479-v2, but attached here for convenience. This kind of critical info / easy look-up table should be a part of the Altium wiring diagram, D2400111, but sometimes given there saturation of information it's nice to just have a look-up table especially when building front-end models.

Enjoy, future us!

This is for FAMIS #39758.
Laser Status:
    NPRO output power is 1.841W
    AMP1 output power is 70.55W
    AMP2 output power is 138.4W
    NPRO watchdog is GREEN
    AMP1 watchdog is GREEN
    AMP2 watchdog is GREEN
    PDWD watchdog is GREEN

PMC:
    It has been locked 27 days, 19 hr 55 minutes
    Reflected power = 27.29W
    Transmitted power = 103.6W
    PowerSum = 130.9W

FSS:
    It has been locked for 0 days 2 hr and 22 min
    TPD[V] = 0.484V

ISS:
    The diffracted power is around 4.3%
    Last saturation event was 0 days 21 hours and 25 minutes ago
Possible Issues:
    PMC reflected power is high

Yesterday Camilla and I put two irises on the FCGS refl path, 89751, to use as a reference for FC1 alignment, and turned on the cleanroom to test how well this would work.  The temperature swing in zone 4 G was not as large this time as other times that the cleanroom was turned on or off, but it is large enough for us to see.  Around 11:30 this morning I went to the table to check the iris, the beam was off as shown in the photo.  I walked it towards being centered on the iris, and this recovered the green flashes to 40 counts, with CLF locked IR flashes are up to 0.4-0.6, infrequently, and frequently around 0.2 counts.  

While I was doing this, I saw that when FC1 was badly misaligned, looking at the green here is a nice way to see where the filter cavity flashes are, movie attached.  

This screenshot shows what this looked like on the osems and sliders: I moved FC1 +10 urad in pitch on the slider by looking at the iris, which restored the flahes.  The top mass osem suggests that after this adjustment FC1 pitch is undercorrected by -43urad, M2 wit would suggest that I overcorrected by  +4urad, M3 wit says I overcorrected by +10urad.  

Here's a screenshot of all osems earlier in the week with Camilla had nice flashes in IR and green, and here's a screenshot of all the osems now. 

After all the alignment tweaks (from recent Round 2 of HAM1-JAC & SQZ work), this morning Betsy requested we record how our H1 alignment looks. Sheila mentioned noting the SUS sliders is the best way to do this, and using SDF can take care of this for us.  Guardian does not touch SUS slider values and they (for the most part) are Not Monitored by SDF.  So Accepting these SUS slider values is a good for maintaining this prefered alignment.

To capture where we are "slider-wise", went through the exercise of ACCEPTING all SUS Slider Values for H1.  Will go through SDF_OVERVIEW.adl, and open each SUS SDF Table to ACCEPT the slider values. 

The Process:  For each SUS SDF table, I will:

• Open a SUS SDF Table (I started by chamber by chamber starting with HAM1 and went down the line ending with ETMy....pecked away at this from roughly 1600-1745utc/9-1045ampdt)
• Change "TABLE SELECTION" to  "CHANS NOT MON"
• If there is a long list of Not-Monitored channels, switch from "SHOW ALL" to "SORT ON SUBSTRING", and enter part of the slider channel name in the "String Box" to narrow the list down to the slider channels
  • slider channels:  H1:SUS-[SUS]_M[1 or 0]_OPTICALIGN_[P or Y]_OFFSET
• Click ACCEPT for slider channels in the table list
• Click CONFIRM
• Return "TABLE SELECTION" to "SETTINGS DIFFS"
  • Return to "SHOW ALL" & clear substring field, if you did a "SORT ON SUBSTRING"

List of All SUS SDF Tables This Was Done For:

• HAM2:  SUSMC1, SUSMC3, SUSPRM, SUSPR3, SUSIM (for IMs 1,2,3,4), SUSHAM1 (for JMs [1,3], PM1,  and RMs [1,2])
• HAM3:  SUSMC2, SUSPR2
• HAM4:  SUSSR2
• HAM5:  SUSSR3*,  SUSSRM 
  • * SUSSR3:  (looking like sliders are MONITORED (pointed this out to Sheila--noted this optic is rarely moved).  Will keep monitoring, but ACCEPTING the pitch slider which did come up as having a diff)
• HAM6:  SUSIFOOUT (OFI*, OMS [for OMs 1,2,3], OMC), SUSSQZOUT (for ZMs [4,5,6])
  • *OFI only has the one Yaw slider (which is 0).  
• HAM7:  SUSFC1**, SUSSQZIN (for ZMs [1,2,3])
• HAM8:  SUSFC2
• BSC1:  SUSITMY
• BSC2:  SUSBS
• BSC3:  SUSITMX
• BSC9:  SUSETMX, SUSTMSX
• BSC10:  SUSETMY, SUSTMSY

NOTES:

Attaching a screenshot of IFO_ALIGN_COMPACTEST.adl from earlier this morning.  (Note:  MC2, ETMy, & SRM are listed as MISALIGN from Guardian [I pointed this out to Sheila, but this is when Sheila said it's fine since Guardian doesn't change these slider channels]).

**Sheila will probably touch some Squeezer-related sliders this afternoon---This would be for FC1 (see her alog 89771).

SPI has been added added as a subsytem in the userapps directory tree.  MATLAB can open SPI models by name and h1build should build spi models (not tested yet as there are no models to build).

You'll need to wait an hour after this post for puppet changes to propagate and log out and back into a workstation for the MATLAB path update to take effect.

Fri Apr 03 10:12:27 2026 INFO: Fill completed in 12min 23secs

TITLE: 04/03 Day Shift: 1430-2330 UTC (0730-1630 PST), all times posted in UTC
STATE of H1: Planned Engineering
OUTGOING OPERATOR: None
CURRENT ENVIRONMENT:
    SEI_ENV state: MAINTENANCE
    Wind: 5mph Gusts, 3mph 3min avg
    Primary useism: 0.02 μm/s
    Secondary useism: 0.17 μm/s 
QUICK SUMMARY:

GV7 was closed last night (~530pm)--still only soft closed to the X-arm (Y-arm has been [hard] closed ~2weeks).

Closeout work in HAM1 will need to continue today and then move tables away from HAM1 in prep for a Monday vent.

Betsy informed me that the commissioners were done with the peek along the X-arm and I closed GV7, once again it does not want to hard close, currently sitting at "soft close" with 50 psi of instrument air in an attempt to close it.
I tried increments of 10 psi with the new regulator, I took the pressure up to 90 psi, and it sat there for about 30 minutes, no luck closing.  Probably the gate valve is not "hard closing" due to the leaks within the gate valve's piston system.
I've set the pressure down to 50 psi and left it and maybe it will hard close, with the leaks present lots of air is wasted.  Also, this configuration allows for crane usage along the XBM from now on.

Following yesterday's work (alog 89745), ISCT1 was moved in place (but not landed, so the table is much higher than usual for now) to check the in-air beams.

ALS beam from the PSL table as well as green ALSX beam:

Was already good. See ALS_beampos.jpg. 

Note that ALSY beam is supposed to be at the same height as ALSX but shifted horizontally by 10mm or so in -X direction on the periscope mirror. According to Mike Smith's L1200282 Table 1 (which is the table we should use though there are other similar tables in that document) ALSY is shifted by 8.5mm in +Y direction relative to ALSX on the septum window (and the two beams are diverging after 3" mirror in HAM3).

REFL air and POP air:

REFL was already hitting the ISCT1 top peri mirror but was close to the +X edge (REFL_before.jpg). POP air beam was coming to the same mirror mount as was described back in May/25/2025 in  alog 84558 (especially this picture in that alog) though I didn't know/remember. (Turns out that I was away for a week when that was done, my recollection only included my May/09/2025 alog 84334.)

I didn't like that POP and REFL beam are both at the edge of the mirror and "fixed" it, i.e. I undid the work people did on May/25/2025 (sorry). I adjusted 90:10 air-vac splitter (M12 in layout.png) to make sure that the beam is not close to the edge of the last steering mirror for the POP air path (M16), and used M16 to steer the POP beam back to the top POP periscope mirror in ISCT1.

I found that the beam was too much in +X direction at the bottom and didn't hit the bottom periscope mirror even when I maxed out the YAW adjustment range of the top peri mirror. (My guess was that people found that back then and decided to shift the POP beam onto the REFL top peri mirror, but Sheila didn't think so even though neither she nor Camilla remembered exactly why.)

Anyway, since I still didn't like that the beams are both close to the edge of the mirror, I decided to keep the "new" configuration where REFL hits REFL periscope and POP hits POP periscope.  

I shifted the REFL top periscope mirror by ~1" in +X direction so the REFL beam is not close to the edge of the mirror. Betsy took the picture of the REFL beam on the top periscope mirror.

The REFL beam hits the bottom periscope mirror. Further adjustment is necessary downstream.

For POP, since the only thing lacking was the adjustment range of the top peri mirror, at first I wanted to swap the mirror mount (ultima with 2 actuators) with a 3-actuators variation to gain more adjustment range, but it turns out that the adjustment knob on the "third" actuator would interfere with the bracket on which the mirror mount is bolted. As a quick fix, I merely "pushed in" the fixed pivot pin to tilt the mirror in a disired direction, which worked. See top_POP_peri_mechanical_tweak.jpg. 

I also steered down the beam by M16 a bit, shifted the top POP peri mirror in +X direction by ~1" like I did for the top REFL peri mirror, used M16 to steer the beam again, and ended up with an OK position on the top peri mirror. See POP_top_peri_beamPos.jpg.

The beam hit the bottom peri mirror, it's not centered but is OK. Further downstream alignment should be done in the future.

Finally,  POP_REFL_at_VP_position.jpg shows the REFL and POP air beam when the card was held very roughly at the location of the viewport. This more or less agrees with one of the pictures (https://alog.ligo-wa.caltech.edu/aLOG/uploads/84344_20250509211310_PXL_20250509_231320536.jpg) in my May/09/2025 alog and I'm positive that the POP won't be clipped by the VP.

REFL_AIR and POP_AIR Beam dumps:

REFL air goes to beam dump when the REFL beam diverter closes (forgot to take picture but not surprising as REFL air path wasn't touched).
Unfortunately I forgot to check if the POP air beam goes to the beam dump when the POP beam diverter closes. This needs to be checked the first thing tomorrow.

J. Kissel, S. Koehlenbeck
H1 SUS-R1 D0902810, S1301885
TIA D1002481-v3, S2500711

Finished with all required assembly and testing of the SPI pathfinder's in-vac ISIJ Reflector assembly (D2400102), we packed it up such that it's install ready. That means it's supporting QPD photodiode transimpedance amplifier, whose [V/A] impedance has been measured and tailored to suit a 10 [V_p] differential ADC input (see 89739) is no longer needed in the optics lab. As such -- we moved it from the optics lab to its final location in the LVEA -- in the U6 position of SUS-R1 by HAM2.

Pictures:
    BEFORE installation
    AFTER installation
    S-NUMBER and U-Height zoom for confirmation.

We have not cabled up any of its inputs or outputs, but will do -- per D2400111 -- so in the fullness of time.

After opening the X arm,  Jenne used the baffle PD script to roughly align the X arm.

We've found that the green beam was too high, hitting the aluminum part of the top periscope mirror mount. But it was off in YAW, too. We moved the PR3 to see what happens, but we ended up reverting it and moved pico steering mirror for POP path in HAM3.

In PIT, green beam was brought to a place where it's higher than the middle of the top periscope mirror because centering that beam in PIT the IR beam will be clipped.

The IR beam was still very close to the +X-Y edge of the dichroic and also the 90:10 to split in-air and in-vac beams, so pico-ed in YAW too.

At this point there was a HUGE pit offset on POP-X, the beam was hitting the top of the QPD case, and large (but not huge) yaw offset. We couldn't center the beam using PM1, so we used the dichroic.

We had to set the DC gain of the WFS DC interface on the front panel to HIGH to be able to see any meaningful DC signal from POPX, but anyway we managed to roughly center the beam in that H1:ASC-POP_X_DC_PIT_OUT and H1:ASC-POP_X_DC_YAW_OUT go both positive and negative as the beam fluctuates.

We didn't see anything on LSC POP_A at this point. We scanned PM1 and found that the beam is not that far and off mostly in PIT. We used a pico-ed splitter to split ASC and LSC POP to steer the beam on POP_A. We scanned the PICO to roughly put the beam at the center of the plateau. I say "roughly" because we used PRX beam which was freely flashing, so it was hard to go precise, but we can always do it after pumping down.

I say that the in-vac alignment is done (in that we can fine-tune after pumping down using picos and suspensions in HAM1 and also using pico in HAM3 if necessary).

Since we changed the HAM3 pico and made a large adjustment of the dichroic, the in-air POP beam as well as the green beam should have been affected. We'll put the table by HAM1 and see how they land on the periscope, and will adjust the steering mirrors in HAM1 if needed.

Sina, Jim

We tested the functioning of the picomotors in the LIGO-SPI Pathfinder ISIK and ISIJ assemblies. All motors are working (see attached video).