TITLE: 09/15 Day Shift: 1430-2330 UTC (0730-1630 PST), all times posted in UTC
STATE of H1: Corrective Maintenance
INCOMING OPERATOR: None
SHIFT SUMMARY:

LOG:

• 2000 Covering Afternoon DAY shift (from TJ)
• 1945 JAC modematching post eom-install (masayuki)

• 2005-2050 Checking on VAC crew @ EX (gerardo)
  • 2015 Travis joining
  • 2025 Tony & Randy out
• 2042-2223 TCS plumbing clean-up (removing old piping and cutting pipe...soldering possibly tomorrow [tj]
  • opened vestibule Exit near the Y-Beam Manifold for old tcs pipes (no access alarm yet)
  • 2107 oli joining
  • 2227 clean-up & trash removal (tj)
• 2106-2109 returning equip to Cleaning Room area (travis)
• 2119-2251 PSL lens adjustment from masayuki's mode matching (jason.ryanS)
• 2149-2305 leak detector to EX & continue pumping annulus (jordan,gerardo.travis)
• 2150-2228 checking vea camera install options at EY (fil)
• 2221-2244 LVEA & optics lab trip to gather viewport installation parts (camilla)
• 2252 jac mode matching mesurements at 62W post-lens-moves from jason/ryanS (masayuki)
• 2300-2326 Local Laser Hazard for Squeezer (sheila)
• 2326 LVEA to Laser HAZARD (camilla)

R. Short, J. Oberling, M. Nakano (remote)

This morning Masayuki ran an assessment of the JAC mode matching after our EOM swap yesterday and came up with new lens positions for the mobile JAC mode matching lenses (IO_MB_L1 and IO_MB_L2).  The result we were given:

SITE INSTRUCTION -- two-lens solution (IO_MB_L1 + IO_MB_L2):

move IO_MB_L1 by 2.0 mm DOWNSTREAM (toward JAC) (dz_L1 = +2.0 mm)

move IO_MB_L2 by 34.6 mm UPSTREAM (toward the PMC) (dz_L2 = -34.6 mm)

(decomposition: common -16.3 mm = waist position, differential +18.3 mm = waist size)

L1-L2 gap: 60 mm -> 23 mm

Moving IO_MB_L1 went smoothly with no change to the beam alignment.  However, we were not able to move IO_MB_L2 the full requested amount due to interference between the lens mounts.  Instead of the requested 23mm separation between the lenses, the best we could get was ~26.5 mm.  To get any better than this we would have to completely change the mounting setup for the lenses.  After moving IO_MB_L2 we had to yaw it a little to recover the beam alignment onto our irises.  While the irises looked good in the PSL enclosure, when we tried to lock JAC it wouldn't lock; Masayuki had to do some alignment tweaks to get the JAC to lock.  This isn't entirely surprising given that our alignment irises are several meters away from the JAC in HAM1, so what looks good on the irises isn't necessarily good enough to lock the cavity.  Once the cavity was locked, Masayuki began tuning things up and we left the enclosure.  He's running his mode matching assessment code now, so we'll see what the results show (preliminary report: "The mode matching seems better." -M. Nakano).

Randy, Tony, Travis, Gerardo, Jordan

Once we got the go ahead for re-installing the BSC door, Randy, Tony and I re-installed the +Y door without any issues. There was some particulate/viton pieces at 12 and 3 o'clock on the flange, but we were able to remove them using IPA wipes.

Bolts were torqued, then Travis and Gerardo set up the turbo and aux cart to pump on the annulus volume. Annulus pumping in progress.

While looking at the ETMX for it's rubbing issue of late, we have been doing some data mining to understand why the ETMx Main and Reaction chain vertical BOSEMs (M0 and R0 LF and RT) are not where I would have assumed to leave them when we installed the "new" ETM test mass and rebuilt the suspension in 2018.  Today, these vertical sensors are all ~6-8k counts when they should be closer to 11k per the open light voltage counts we determined in 2018 alog 41937.

I worried about temp and slow drift which could be also contributing to the rubbing, but in fact, it seems like something happened immediately at closeout in 2018 that put these vertical sensors off the nominal centered 11k value.  (sigh)

In the end, hunting for an event or evolution which could have lead to the L1 LL rubbing of late has not been fruitful.

Anywho, since I waded through the deep water of everything-ETMX-SUS via temp change hunts, etc, I will post it here for the next time we wonder about what the longer history of the EX VEA temp has been doing:

Dec 2014 - CR excursions turned back down from 68 to 65

Jan 6, 2015 - 15883 - John turns temp “back to 65F (down from 67F)” - likely after a vent?

2015-2016 - No one writes actual temp values just things like “seasonal change so I turned a heater on”

May 2017 - 36271 - Upgrade to HVAC controls everywhere, EX week of ~May 15-18th, Kissel alog on calibration discrepancy 64 is actually 68…

May 22, 2017 - 36331 - Jeff/Bubba request to increase from 64 to 66 degC

May 30, 2017 - 36536 - Bubba changes which sensors are in-loop at EX

June 14, 2017 - 36894 - Bubba changes which sensors are in-loop at EX

Jan 23, 2018 - 40233 - Bubba/Hugh banter about where sensors in EX are mounted

Feb-Jun 2018 - Vented, warm with Cleanrooms on, new ETM/AERM installed

________________________________________

June 1, 2018 - 42283 - Chamber closeout TFs ETMx

Vertical sensors get “bumped” or set from nom 11k to ~8k somehow.

June 6, 2018 - 42351 - Bubba puts temp back down to 65degF after chamber closeout

June 26, 2018 - 42664, 42686 - Chandra asks to put temp up to 68defF to aid in outgassing/pumpdown, Kissel/Chandra banter about temp increase. Chandra states that 65degF is more energy efficient than 68degF

June 29, 2018 - 42722 - Bubba puts temp back down to 65degF

Jan 27, 2022 - 61472 - Bubba/Fil move some sensors/cables, “The temperature at the end stations by request of Jeff Kissel several years ago was maintained at 65 degrees F.” - Probably need to go back to fact check this since originally Kissel suggested 66 per logs found above.  0.5degF change but Vert’s see -2k OLV change, weird.

Mar16 ’23-Jun 22 - EX "Rubbing" and 11Hz issues, offsets put on, then later turned off because not an issue “anymore” (68034, 68121, 63766) - So, in hind sight, could have had the L1 LL flag close and maybe brushing in 2022 then it kinda alleviated itself.

Aug 28, 2023 - 72497 - Kissel nice blog analysis of temp on EY, FWIW

Sept 14-15, 2026 - Rubbing issues on ETMX - L1 LL flag brushing.  Vented, moved LL BOSEM over away from flag.  See that ETM M0/R0 V BOSEMs have been at ~6-8k instead of nominal ~15k for many many many years (SUS is “UP” which correlates to “COLD”).

Sept 15, 2026 - 91912 - Richard raising EX from 65F to 67F as read out by HVAC chans. (This should drop the suspension and more center the V BOSEMs, but likely not an actual problem.)

After various ETMX transfer functions and offset tests, it looks like ETMX isn't rubbing anymore!

I did these tests this morning and wasn't able to get any rubbing. For everthing I did, L1 LL (and everyone else) moved as expected and no one got stuck anywhere. The offset tests that I did involved putting offsets in the L1 COILOUTF banks, M0 TEST banks, and R0 TEST banks and making sure L1 osems were all responding how we expected them to.

We noticed that L1 UL has been a little more jaggedly compared to the other L1 osems since we went to air, but the osem spectra looked good (comparing to its brethren) and putting a drive in M0 to move L1 around shows that UL is moving and behaving normally. We think it's probably just related to its current flag vs osem position as we put air in the chamber.

Transfer Functions
Settings:
- HEALTH_CHECK
- OPTICALIGN OFFSETS ON
- DAMP OFF
- ISI in ISI_DAMPED_HEPI_OFFLINE

M0
Data:
/ligo/svncommon/SusSVN/sus/trunk/QUAD/H1/ETMX/SAGM0/Data/2026-09-15_1540_H1SUSETMX_M0_WhiteNoise_{L,P,Y}_0p02to50Hz.xml
/ligo/svncommon/SusSVN/sus/trunk/QUAD/H1/ETMX/SAGM0/Data/2026-09-15_1540_H1SUSETMX_M0_WhiteNoise_{T,V,R}_0p03to50Hz.xml
r13152
Results:
/ligo/svncommon/SusSVN/sus/trunk/QUAD/H1/ETMX/SAGM0/Results/SAGM0/Results/2026-09-15_1540_H1SUSETMX_M0_ALL_TFs.pdf
r13153

R0
Data:
/ligo/svncommon/SusSVN/sus/trunk/QUAD/H1/ETMX/SAGR0/Data/2026-09-15_1615_H1SUSETMX_R0_WhiteNoise_{L,T,V,R,P,Y}_0p03to50Hz.xml
r13148
Results:
/ligo/svncommon/SusSVN/sus/trunk/QUAD/H1/ETMX/SAGR0/Results/2026-09-15_1615_H1SUSETMX_R0_ALL_TFs.pdf
r13150

Spectra:
L1
Data:
/ligo/svncommon/SusSVN/sus/trunk/QUAD/H1/ETMX/SAGL1/Data/2026-09-15_H1SUSETMX_L1_OSEMINF_spectra_invac_vs_inair.xml
r13149

Please do not make any changes to real time models with simulink or foton until after the RCG upgrade. 

Quickly checked the ugf of the JAC servo. At a filter gain of 0.8, the nominal gain was 300Hz. This seems the same as before, indicating the modulation index didn't change much with the EOM swap. The attached plot shows the ugf at 400Hz with a gain of 1.0.

Reminder: we attenuated the power by half to the IMC refl diode and IMC WFS after a power outage during O4, 86884, using a half wave plate along the refl path in IOT2L.

The attenuation was such that we maintained the desired power on IMC refl and the WFS by reducing the refl path power by 2, 86989, in other words, whatever the power outage did to the upstream mode matching, it was such that the power reflected off the IMC when locked had doubled.

To summarize, our nominal power to IMC refl when locked with ~60 W input is 18.5 mW. The WFS nominal is 0.95 and 0.76 mW respectively.

Now we have JAC, so I checked the powers at IMC REFL. Looks like the JAC team turned up the input power to the IFO to 60 W recently. For 60 W input with IMC locked (guardian state 100), we now have 1.57 mW on IMC REFL, 0.079 and 0.064 mW on WFS A and B. I think this indicates we should adjust the half wave plate to increase power to IMC REFL. I'm not sure if we can get back to nominal since the powers are more than an order of magnitude lower than before, but we should be able to at least double it.

Reminder that the "IMC PWR IN" diode is actually reporting the power into JAC. We have a JAC trans diode, but it still has a funny calibration, so I did not include it here because the number it says is totally wrong.

Betsy, Oli, Tony, Camilla

Yesterday when Betsy, Oli and Tony were inside the ETMX BSC (WBSC9) they took some photos of the transmon mirrors,

1.  with the illuminators on
2.  with the illuminators off and an eye-safe Ozoptics green laser injected via the HWS path

Reason for this is that we've had issues with both the ALS 91306 and HWS 91248 beams at EX. 

Full telescope assembly is D1002460. The optics on the transmon table are in D1201457. The mirrors are called from transmon towards ETM: M2 , F2, F1, M1. D1102361 and Madi's cartoon, attached from G2500646 shows this.

Betsy notes that the optics with the side HR facing -X (F2 and M1) appear to have considerably more dust on the those facing +X. Is purge air or similar causing this? The dusty side is direction ETMX faces.  All the mirrors coating looks intact.

• Transmon -  photo a, b, Green QPD A, M12 mirror, again. 
• LPM - photo with beam
• M2 (TMS telescope Secondary mirror - parabolic) - photo, with beam 
• F2 (Folding Mirror 2 - flat) -  photo, with beam, wide
• F1 (Folding Mirror 1 - flat) - photo
• M1 (TMS telescope primary mirror - parabolic) -  photo a, b, c , with beam

The HWS path green beam made it though the transmon and retroreflected off ETMX and back onto the HWS CCD, but the alignment isn't perfect as the return beam looks clipped, photo. 

PicketFence was updated to the latest version and restarted.

Richard, Betsy, Eric.

Looking back many years we can not find a reason for the temperature setting of 65 Deg F at EX.  Today I raised the temperature to 67 Deg F.  Wtih the clean room on we have seen a diurnal swing of 3 Deg F with no effect on the suspension so this rise should not have a negative impact on the suspension.(sagging)  It may actually help.

However,  It will probably change the balance of the BRS.  Jim W is aware

[Jenne, Elenna]

As a reminder, Jenne had engaged a "slow let go" servo for beamsplitter pitch during O4 because of some possible heating transient that affected relocking, see 85002.

Now we have the BBSS, so we don't expect to see this issue. I undid the steps that Jenne did in her alog above to revert the settings:

• I re-engaged the BS M1 ASC lockloss trigger, see attached SDFs in SAFE and OBSERVE
• I removed the separate BS loop on ISC_DRMI line 66 that kept the M1 LOCK pitch gain set to one and disengaged the integrator in FM1, and added BS back into the list on line 63
• I returned the BS to the list on line 133 in ISC_DRMI that clears the history in BS M1 LOCK P and Y, we had separately only cleared history for BS Y so the BS pitch history would slowly bleed off over time

ISC_DRMI is loaded and I checked the code to the svn.

Richard, Jonathan, EJ, Erik, Fil, Dave:

First thing this morning we did a round of h1seih16 power cycling, with Richard power cycling the IO Chassis. The models ran for about 5 minutes this time and then ADC1-timed-out like before.

Then we power cycled the CPU only, and this time the IOP would not run because of a card count issue. And indeed, both showcards and lscpi report that the A2-slot4 is empty, which contains the second 16bit-DAC card.

So we are operating under the theory that all the recent problems have been caused by a flakey DAC1, manifesting as ADC1 timing errors. We are preparing to replace DAC1.

Card 2
    Extender 1
        slot 1: (10ee d8c6) Xilinx Corporation,LIGO Timing Board
        slot 2: empty
        slot 3: (10b5 3101) PLX Technology, Inc.,General Standards 16AI64
        slot 4: (10b5 3120) PLX Technology, Inc.,General Standards 16AO16
Card 3
    Extender 2
        slot 1: (10b5 3101) PLX Technology, Inc.,General Standards 16AI64
        slot 2: (10b5 3101) PLX Technology, Inc.,General Standards 16AI64
        slot 3: (10b5 3101) PLX Technology, Inc.,General Standards 16AI64
        slot 4: empty <<< Should be 16AO16 DAC1 <<<<
Card 4
    Extender 3
        slot 1: (10b5 3101) PLX Technology, Inc.,General Standards 16AI64
        slot 2: (10b5 3101) PLX Technology, Inc.,General Standards 16AI64
        slot 3: empty
        slot 4: empty
Card 5
    Extender 4
        slot 1: (1221 8682) Contec Co., Ltd,Digital I/O DIO-64
        slot 2: (1221 8682) Contec Co., Ltd,Digital I/O DIO-64
        slot 3: empty
        slot 4: empty
 

• Turned on the sqz laser, Marc Pirello turned on high voltage supplies.  
• with the OPO dither locked on the seed, I adjusted ZM3 to get full power onto SQZT7 IR PD, didn't see light on AS port PDs this way.  
• PMC transmission is normal at 660mW.  SHG green power is 77mW, rather than 100mW normally, SHG launch power is 9mW max, compared to up to 30mW we normally have available. 
• I tried adjusting the SHG alignment picomotors, this didn't improve the SHG transmission much.  
• Filter cavity green flashes cavity modes.  

Jennie W, Jason O, Ryan S,

Summary: PSL EOM swapped, alignment good enough to lock JAC but the mode-matching to JAC looks different as the JAC REFL camera spot has changed.

We entred the PSL enclosure and turned down power through EOM to less than 2W at the manual waveplate.

We then recentred the first iris in the MB path after the EOM in the vertical direction as the beam was slightly too low. The second iris was centred ok.

Then we closed shutter 2 and unplugged the 118 and 9 MHz cables.

We terminated the 45 MHz cable as this is the only one still plugged into an active voltage source.

We had to remove the EOM along with the white spacer it sits on to undo the screws on the bottom of this spacer that go into the EOM unit.

After swapping the new EOM onto the spacer we screwed it back down on the mount.

The beam was now off of the first iris in both pitch and yaw. It was too low and slightly in the +y direction.

To fix this in the y direction Jason adjusted  IO_MB_M2 in yaw. The beam was not making it through the the EOM aperture. We loosened the side screws on white spacer and used slop in the large screws slots which hold the spacer on the bottom mount to yaw the eom and fix this . Then we used the pitch of the IO_MB_M2 mirror to centre the beam in the vertical dirtection on the first iris.

After tightening back the screws on the mount and spacer we took a calibration sweep of the rotation stage to re-calibrate the output power to the interferometer.

Then we took the power guardian to 2W and re-locked the JAC. The relfected beam from the locked JAC has changed shape on the camera so we assume the mode-matching is different.

Masayuki will measure the mode-matching and we will improve this with lens moves on the PSL table tomorrow.

I unplugged some cables to an interface chassis in the CER this afternoon to do some CPS whitening measurements. Not sure what has happened, but when I plugged everything back in, the corner 2 cps didn't come back alive, and are sitting at ~0 cts. I have to leave early today, so I'm out of time to diagnose, but hopefully I can fix this tomorrow morning when I get back on site. This means HAM2 won't be able to isolate, until the corner 2 CPS are working again. 

Sheila, Elenna, Camilla

Sheila was questioning if something is drifting for us to need an initial alignment after the majority of relocks. Elenna and I noticed that BS PIT moves a lot both while powering up /moving spots and while in NLN. Unsure from the BS alignment inputs plot what's causing this.

This was also happening before the break (see below) but the operators were similarly needing more regular initial alignments before the break too.  1 year ago this was not happening, plot.

• Now:
  • While locking (Power up state #500 to NLN #600) BS PIT moves 40urad  plot
  • During a 2hr lock BS PIT moves 13urad plot
• Before Break:
  • While locking (Power up state #500 to NLN #600) BS PIT moves 43urad  plot
  • During a 4hr lock BS PIT moves 20urad plot