J. Kissel, S. Koehlenbeck (in person!)
D2400107 for the annotated layout of SPI Pathfinder ISIK's Transceiver Assembly.

After an easy recovery from other-the-weekend, site-wide power outages (LHO:90413), we resumed last week's investigations of what might be causing the MEAS IFO to drift (LHO:90382).

%%%%%%%%%%%%
Executive Summary: No magic bullet found. Today we ruled out the collective idea we all had that the picomotor-actuated mounts were the problem (LHO:90384) by replacing every pico-actuated mount with manually driven mounts. We also replaced the last optic we hadn't swapped in the MEAS path to the MEAS IFO -- M_B1. No impact -- we still see the MEAS IFO alone drift.
%%%%%%%%%%%%

It was only a half-day of forward investigative progress given the power outage, but having Sina on-site really helped confirm my (in)sanity and accelerate investigatory iterations.

Upon arrival in the lab Sina re-confirmed my construction and alignment of the D2400107-v5 board and the quality of things:
    - Centering on optics with IR card :: good.
    - Centering on photodiodes with IR camera :: good.
    - No stray beams (even from what would be AR coatings of beam splitters) :: good.
    - The as-we-left-off Friday assessment that with M_M2 and M_B4 down-graded to manually driven mounts, we still see drift.
    - Having one or the other REF or MEAS beam into the MEAS IFO blocked, that the power on each beam is stable without drift.
    - After an alignment drift, quickly touching the alignment of all three M_M2, M_B4, and M_M1 cannot recover the MEAS IFO efficiency.

As we continue to grasp at straws, we replaced the last optic in the MEAS path into the MEAS IFO -- M_B1. This optic splits the MEAS beam on the board feeding it to the MEAS and REF IFOs, so we were hesitant to swap it. But now that we're getting so good at re-aligning the board after changing out optics (once optic is placed, re-alignment can be done within 2-5 minutes), we figured we must. 

So, all at the same time, we
    - unlocked all manual drive set-screws on all optics (recall the REF path had been "locked down" the day I'd gotten ~95% efficiency),
    - loosened and re-tightened R_B1's 1/4-20 bolt that secured the mount to the breadboard. It wasn't "loose" but it wasn't "cranked on" either.
    - Replaced M_B1 optic 
    - Replaced the M_B1 manual-drive pitch actuator's 1/4-100 alignment screw and bushing because it was "crunchy" and "stiff" to use. This had been this way, but when I was driving solo I'd suffered thru it and locked it down the day I got 85% efficiency on the MEAS IFO and didn't look back. Now what we're touching the mount anyways, we figured we might replace it in case the issue was some sort of stick-slip thing happening.

Upon doing all these things ... things looked solid for ~10 minutes... and then the efficiency started to drift again. Especially after turning on the Clean Room fans.

This again resurrects the theory we thought debunked: "it must be the environmental conditions..."
Why did we rule this out before?
    (1) The REF IFO does not see any drift. Ever. The board is so small and so local, and both IFOs are sampling the same beams. *Very* hard to believe that the few differences in paths and path lengths would cause only ONE of the IFOs to drift. The environment should impact both IFOs equally.
    (2) We *do* see phase noise when we make environmental changes -- as expected. Path length differences caused by
         - physically manipulating the board, its mounts, or optical fibers,
         - waving hands over the board
         - talking, even at a medium volume
    (3) Any environmental change in the efficiency we can cause -- be it: 
         - clean-room fans ON/OFF, 
         - clean-room or room lights ON/OFF, 
         - people coming in-and-out of the clean room, 
         - standing in different positions around the optical table
        -- they all only some-times seem correlated, and never is it reproducible.

But we've run out of things to change, so today we're going to try to adjust the environment and/or other things that *should* be common to the IFOs:
    (a) We're gunna being in an incandescent lamp to try to create thermal gradients
    (b) We'll put on our tin-foil hats and enclose the breadboard in a temporary foil enclosure
    (c) We'll try adjusting the NPRO seed laser controller's parameters (like diode temperature)
...and keep trying all the other wild theories we can concoct.

TITLE: 06/02 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: 8mph Gusts, 4mph 3min avg
    Primary useism: 0.02 μm/s
    Secondary useism: 0.14 μm/s 
QUICK SUMMARY:

• FAROing on HAM3, HAM2 work, and HAM7 continues
• DAQ_KCO channels are disconnected

Yesterday (June 1) I replaced a failed power supply on the Comtrol communication device at EY that bridges a dust monitor and a weather station to the network.

Workstations were updated and rebooted. This was an OS packages update.  Conda packages were not updated.

Today started with removal of the HAM7 +X and -Y doors, WP 13283.  These were already on 4 bolts/piece, so removal was relatively quick.  Nothing to note other than the usual O-ring sticking business. The 2x 12" CF blanks on the +Y door were also removed for ISI locker access.  Since the final cleaning happened ~1 hour from door removal, particle counts were very minimally elevated, 40 @ 0.3um and 20 @ 0.5um (see pic).

We then moved on to the X-beamtube manifold vent work, WP 13284.  The X-manifold turbo pump was valved out.  We then cycled GV7 open, then hard closed again.  The GV opened at 10 psi, and hard closed at 20 psi, so we believe the piston O-ring replacement is working well.  Then, the X-manifold vent was started.  After burping the system a few times and confirming there was no communication past GV7, we started venting the manifold section.  We noted that the Kobelco's duty cycle seemed to increase audibly as the pressure in-vacuum rose, which isn't surprising given that we have 7 doors and 2x 12" ports open on the corner volume, so we elected to keep the vent slow, maybe 5 torr/minute.  We stopped the vent for the night at around 400 torr.  Tomorrow, we'll finish venting and open GV2.  

After JAC was recovered (alog 90421) we moved IM4 by negative 500urad YAW (from -387.1urad to -887.1urad) and adjusted PRM so that it retroreflects (without bothering to adjust IM2, 3 and 4 in the interest of time) and Jenne started centering the REFL A and B using RMs.

However, RM2 still railed.

Turns out that the accuracy of retroreflection matters because 100urad rotation of PRM causes the equivalent of 280urad rotation of IM1 as far as the beam angle leaving RH9 (IFO REFL pickoff) is concerned.

The retroreflection technique was nothing sophisticated, we used an IR card with a hole, put the card between the IFI output baffle and DKDP baffle, put the forward going beam in the hole in the card so it goes through. Rahul would make a huge (like 1.5mrad) YAW rotation for PRM, first positive YAW to find the reflected beam on one side of the hole in the card, and then negative YAW to find it in the other side of the hole, and find a good PRM YAW offset where the reflected beam for +1.5mrad positive and negative offsets are balanced on the card. The retroreflection accuracy is no better than +-500urad.

Anyway, using this technique we moved the PRM slider offset to [P, Y]=[-1165, +1000] (from [-1165, 20.5]).

Out of curiousity I changed the YAW offset of PRM from +1000 to +700 so [P,Y]=[-1165, +700], used RM1 to center REFL_B and RM2 to center REFL_A and it was easy without railing RM2 nor RM1. See attached.

I'll  see if there's a space for an iris on IFI where both faces are clearly visible. If so we can probably use an IR sensitive camera as a better resolution viewer, center the iris for the forward going beam well, and observe the centering of the return beam on the iris.

(However, it's still a good idea to reduce the output impedance of the RM OSEM coil drivers. Small actuation range is fine if things are working well, but they're not.) 

Jim, Shoshana, Alexandra, Arnaud

Today we continued setting up the H2-PSL lab to start the testing the HOQIs.

* Received and moved the laser system (RIO+Laser Driver+TEC), and readout (HOQI TIA) in the H2-PSL (see picture 1) 
* Gathered needed electronics from the EE shop (See list below) and plugged them in the lab
* Bagged and tagged the peek HOQI components from VPW (waveplate holders, photodiode holders, in-vac cables)
* Visually inspected the new batch of 10 flexures after c&b (see picture 2), and brought them to the H2-PSL (no issues found so far)
* Started setting up the Panasonic optical readout for CRS balancing. We will need some basic components from the optics lab to hold this readout.

Items borrowed from the EE shop :

x2 GWINSTEK power supply
x1 Hewlett P power supply
x1 multimeter with probes
x1 Techtonik oscilloscope 
x1 Bnc banana plug
x4 bnc to clips
x4 bnc cables
x1 db9 cable
x3 power cables
x3 breakout boards (x2 db9 and x1 db25)
x3 db9 gender changers

As per WP 13286 I tested the copy of the fmcs epics ioc that I had extracted from fmcs-epics-cds.  Unfortunately the IOC would crash each time I ran it.  There was not time to debug and see if it could be fixed today.  So the mac mini is still running the ioc.

Sheila, Camilla, Disha

After Jim locked the ISI, Marc bypassed the high voltage interlock, and Camilla restored the psams, we reset the alignment sliders following 90483 which was in air.   The dither locking worked without any issue, and we saw that the beam was low on both the irises in the homodyne path on SQZT7.  Camilla adjusted the alignments of ZM4+ZM5 to center the beam on the irses, screenshot attached shows the alignment that she found.  

We placed an iris in front of ZM4 for this alignment, we were able to look from the +X side of the chamber to check the centering.  We also took photos of the IR/green co-alignment, it looks good coming right off the VIP, but looking right after ZM3 you can see that there is a small misalignment in yaw.  We will continue with placing irses tomorow.  

Betsy, Anamaria

With the helpful kits Mitchell had prepared, we assembled the cage baffle brackets and panels on a table by BSC1; this includes CP and HR cage baffles. We chose SN 3's for ITMY and SN 4's for ITMX for both panels and brackets. We also checked that we have the various tools we need chamber side. Still need to make one more set of brackets for ITMX CP baffle, but that can wait. We found and staged two ACB long lockers but only one wedge. 

TITLE: 06/01 Day Shift: 1430-2330 UTC (0730-1630 PST), all times posted in UTC
STATE of H1: Planned Engineering
INCOMING OPERATOR: None
SHIFT SUMMARY: Very busy day with lots of visitors on site! Essentially everything was recovered following the weekend power outage, doors on HAM7 were removed and SQZ work in the chamber started after the ISI was locked, QOSEM work continued on the BBSS, the ISI was balanced on the test stand, prep began for the ITM cage baffle installation, GV7 was cycled and left closed for venting of the X-manifold, more beam alignment went on in HAMs 2 and 3, turbopumps at EY and MY were tested, FARO surveying began in HAM3, CRS building continued, and SPI work in the optics lab is ongoing. The EDC is still showing several disconnected channels, some of which relate to EY dust monitors.
LOG:

[Jonathon, Ollie, Erik v., Tom]

h1susbs front end model was shut down and h1suslo12 was rebuilt and restarted, moving control of mc2 form h1susbs to h1suslo12.

This required DAQ and EDCU restarts.  The DAQ restart ran into problems because some channes were renamed (temporarily, hopefully) from BS to BBSS that were part of the GDS broadcast.  These channels have been renamed in the broadcaster lists for h1daqgds0 and g1daqgds1.

JAC PZT HV was off -> Mark reset the interlock.

JAC heater chassis was off -> I switched it on at the rack.

JAC error signal was not crossing zero -> I updated the dark offset H1:JAC-REFL_A_RF43_I_OFFSET to a new number 2.4.

Beckhoff was reporting "invalid value" error for JAC PZT whitening -> Patrick did something to clear that.

Rahul restored slider numbers for JM1/3, IMC1/2/3, IM1/2/3/4 and PRM to Friday evening numbers as he reported in alog 90419.

JAC guardian didn't want to lock because H1:JAC-TRANS_A_DC_NORMALIZED kept getting below the threshold of 0.5. -> I temporarily changed it to 0.1.

JAC PZT voltage was drifting quickly for 30 minutes or so and JAC kept unlocking (probably because JAC was heating up) -> waited for a while.

JAC still didn't want to keep lock for a long time. -> I put it to DOWN state, followed the resonance manually using the PZT voltage for a minute or so, changed the state to LOCKED and the guardian liked that.

Jenne refined the JAC alignment using PZT and JM1. H1:JAC-TRANS_A_LF_OUT_DQ momentarily went up to 620 and the best number in the past month before today was like 580. It's trending down in the past two hours, now it's ~580 and that's still a very good number.

Temperature drift is still going on, the symptom is that the PZT voltage drifts down to -10 Volts and at that point JAC unlocks, but the guardian relocks to higher voltage fringe position right away, and the drift is slow enough it's bearable. Last lock lasted for an hour.

Now we can set H1:JAC-TRANS_A_DC_NORMALIZED back to 0.5 if we want to.

Aligning beams onto QPDs while the IMC is flashing (eg, during a vent) is very challenging.  We want to know what the pit and yaw values are, but can only trust the values at a moment when there is a flash.

The jupyter notebook that has just been placed into git can (with lots of patient help from Jamie Rollins) read data that has been exported from an ndscope (so, can be data that is not DQ'd), and then will calculate the pitch and yaw values at the time of the largest flash in that time window.  

This is a direct link to the notebook right now for a snapshot of this first iteration, and you can see the plots at the very end and a printout of the pit/yaw values of all of the QPDs that I have collected data for.

Hopefully this will help make things a bit quicker as we work to center on our various ASC PDs and close out this vent.

I have restored the SUS alignment slider values (P/Y) for all the suspensions in HAM1, 2 and 3 chambers (i.e JM1, JM3, MC1, MC2,, IM1, IM2, IM3, IM4) which had changed after the Saturday site power outage. I trended the pitch and yaw slider values for all of the suspensions in these three chambers and restored the ones which I found have changed. I selected Friday evening (29 May 2026, 6.30 PM when Keita et al were done with moving the suspensions).

Marc and I turned on the HAM7 Hi Voltage Bypass in the Mech Rm Mezzanine. We first turned the PSAMS down to zero, they hadn't been zero'ed before Hi-voltage was turned off. WP13282

Belated alog from Friday.

Summary:

After realigning IMC and IM2 on Thursday, we continued aligning IM3 to put the beam in front of the PRM to the nominal position, IM4 to set the beam position in front of PR2, and PRM so the beam retroreflects. IM1 was left untouched, so the only change in IM1-IM2 line is from the change in the IMC alignment.

Clipping on IM4 baffle (HA12) disappeared completely. Beam position on IFI output baffle looked OK. Beam position on IFI input baffle didn't change (as expected) and still looked OK. There are some mystery 

We saw flashes on ASC REFL sensors but they were weak. We tried to center them using RMs but RM2 seems to rail in YAW.

What to do:

Scan PRM to make sure that we're NOT looking at some kind of ghost beam on ASC REFL sensors.

Assuming that we're looking at the real beam on REFL sensors, we can do either (or some of) these:

• Relieve RM2 by rotating RH9 (the fixed IFO refl pick off mirror in HAM2) by positive 500urad or so in YAW. This is the equivalent of the combination of negative 0.53 mrad for RM1 and negative 2mrad for RM2 in YAW. After doing this we can add 530urad to RM1 and 2000urad and continue centering from there. See attached cartoon (doesn't scale but the sign of things is correct) and the script.
• IM1 is almost degenerate as RH9 as far as relieving RM1 and RM2 is concerned, so we can rotate IM1 by negative 500urad or so in YAW. This requires that we tweak IM3, IM4 and PRM again but has an added benefit of moving the beam spot on the IFI input baffle closer to the center. See IM1rotation.png.
  • If you aren't sure how IM1 rotation is translated to the IFO refl beam,  IM1rotation_straightened.png is a simplified and unfolded version. The beam going out of IM1 goes through CWP at the IFI input (deflected but it's not written in the cartoon for simplicity), is retroreflected and comes back to CWP and defflected, the deflection for the REFL beam is depicted here. Without IM1 rotation the beam goes along the black path. With IM1 rotation (red path), the red refl beam converges with the black refl beam where the distance from the conversion point to the CWP is equal to that from the IM1 to CWP. The angle change of REFL is the same with the input and the lateral displacement of red refl relative to black is ~(distance(IM1_CWP)-distance(RH9_CWP))*angle_change.
• Make the OSEM coil driver output impedance smaller for RMs to gain more range. 1.2k to 500 Ohm should be safe. 

Both CO2X and CO2Y Chillers turned back on with no issue. Interestingly as the laser is off they are currently heating the water.