To verify that the QOSEM readout has the expected sign convention, as detailed in D2500284, I injected a longtidual offset into the BBSS M1, via the QOSEM coils, and monitored M1, M2, M3 OSEMs. The QOSEM readout sign should be consistent with the BOSEMs and AOSEMs.

Figure 1 plots M1 F1_x, M2 LL, M3 UR and the test L injection into M1. Due to the location of these OSEMs they should all readout with the same sign. Here we see that the sign is consistent between the QOSEM, BOSEM and AOSEM for this given longitudal offset, verifying that the QOSEM sign convention is indeed as documented.

[Arnaud, Michael, Jim, Fil, Shoshana]

As mentioned in 90702, one of the readouts for the CRS HoQI amplifier wasn't working, after investigating we figured out that it was an issue with a broken op-amp in a transimpedance amplifier (HoQI B, channel E/COS). Fil switched out the op-amp so we now have a working chassis. Our best guess is that the op-amp failed after plugging it into the rack.

The HoQI on the left side of the CRS (SN01-005, currently HoQI2) had much lower fringe visibility than the other one (SN02-007) since yesterday. We figured out that the laser collimator was positioned in a way where the beam wasn't hitting the center of the photodiodes. The laser collimator was re-aligned and we were able to greatly increase the fringe visibility so the HoQIs now match.

HoQI 1 (SN02-007), SIN: 86.7%, COS: 79.4%, MCOS: 84.9%

HoQI 2 (SN01-005), SIN: 81.4%, COS: 82.7%, MCOS: 80.1%

CRS was balanced, currently has a estimated resonance frequency of 15.7mHz, but we will re-check it in the morning. We are leaving the laser on over night hoping to get some noise measurements, both due to the air currents in the clean room, and the fact that we didn't ground the damping capacitors, it looks unlikely that it will damp down overnight.

Arnaud got the ellipse fitting working, the python script is in /opt/rtcds/userapps/trunk/isi/h1/scripts/write_ellipse_values3.py

We added a gain 0.0847 um/cnt (1064 nm/4pi) to the HoQI Dist filter banks to calibrate the HoQIs into distance. We also changed the CRSRY_PHI_SUM matrix elements to 2.75 to put the RY channels into urad.

An unrelated mystery: HoQI1 and HoQI2 are switched, we aren't sure where in the chain they get flipped, but right now we are guessing it's at one of the cables to the chassis

Previously (90441), Tom, Ibrahim, and I had found out that the QOSEM coil drivers were wired opposite as compared to other OSEMs. To compensate for this at the time (since we were attempting to use damping), we had swapped the signs of the damping gains. But we had only swapped those gains, meaning that other actuation coming through the TEST or OPTICALIGN filter banks was still pushing the coil drivers the opposite way from expected. Today we finally decided to fix it to reduce confusion wrt the direction the BS was moving when offsets were put in (offsets were moving the BS in the opposite direction than expected). 

To do this all I did was put the M1 DAMP filter gains back (before, after) to being negative and then swapped the signs of the gains in M1 COILOUTF (before, after). To put the beamsplitter back with the OPTICALIGN offsets that it had before, I then swapped the sign of each offset. 

I will update the SUS Sign Conventions document to mention that the QOSEMs require the opposite COILOUTF signs from nominal.

So now EVERY BBSS OPTICALIGN slider offset from before 2026/06/24 00:50 UTC requires the opposite sign to be recreated.

Ndscope has a new system for picking trace colors that tries to maintain a high contrast between the trace and the background.

You can test the new system by running "ndscope-test" instead of ndscope on any workstation, laptop, or login portal.  Give it a try and let me know what you think.

Arnaud, Oli

As part of our BBSS alignment and investigation, I used /ligo/svncommon/SusSVN/sus/trunk/BBSScomparetripleparams.m to look at the DC values of our current model when actuating at M1 and reading out at M3. 

I made a few edits to comparetripleparams.m in order to allow me to do this easily, and I've committed those changes as r13039.

The plots (P, Y) show that the DC value for M1 P to M3 P is 0.029 rad/Nm and M1 Y to M3 Y is 0.129 rad/Nm. These plots can be found in /ligo/svncommon/SusSVN/sus/trunk/BBSS/Common/Results/comparetripleparams/comparetripleparams/2026-06-23_M1toM3Model/triplemodelcomp_2026-06-23_M1toM3Model_M1toM3.pdf r13040.

We then converted into cts/N using (1e-6)*(1/DACdrive)*(1/coilDriver)*(1/forceCoeff).

For this we used DACDrive = 20/2^16 cts/V, forceCoefficient = 1.694 N/A, and coilDriverTransconductance = 0.011919 A/V. The reason why DACDrive = 20/2^16 cts/V and not 20/2^28 is because we already compensate for the difference between a 16-bit DAC and a 28-bit DAC in the COILOUTF filter bank so we don't need to put that into our calibration.

This gives us:

   Calibration factor [cts/N] = (1e-6)*(1/(20/2^16[cts/V]))*(1/0.011919 [A/V])*(1/1.694 [N/A])

     = 1.623e5 [cts/N]

Then multiplying this calibration factor by our earlier fractional values for P and Y gives us:

   Pitch slider calibration : 4.7e-3 cts[M1]/urad[M3]

   Yaw slider calibration: 20.9e-3 cts [M1]/urad[M3]

B. Weaver, C. Compton, R. Short

Carrying on from alog90708 posted at lunch, Betsy, Camilla, and I offloaded the BS pitch sliders mechanically at the suspension.  This was a tad tricky since we were looking at the beam(s) at HAM3 in front of PR2, so we had to work out blocking 1 ITM beam in order to not be confused by both beams crossing each other after the long PR chain of optics.  In the end, we worked out the process to bring the BS PIT slider from +1000 to 0 which meant pitching the BS optic HR "up" (PUM roll bar "in").  Unblocking the other ITM, we got both beams overlapping onto each other with a little more slider to see how much better we could get them and see flashes in the AS_AIR camera (PIT = -300, YAW +1560).

Once back in the control room, Keita scanned BS alignment to find the SQZ beam reflected by ITMX as well as ITMY in AS camera as well as ISCT1 camera, ending with BS PIT = -520, YAW = 1677. With this alignment, IMC flashes reflected by ITMX as well as ITMY were visible in both of the cameras. Though SQZ beam and IMC flashes are not coaligned,  the corner alignment is already pretty good as of now.

At this point we want to:

1) Recenter the OPLEV

2) Go in and physically calibrate which way the sliders go as seen on the oplev and at the HAM3 PR Beams since there is confusion around the sliders (direction, calibration, cross coupling)

3) Based on 2) Potentially make a request for HEPI YAW correction tomorrow when they resume HEPI "fine" alignment and actuator reattachment (TBD convo with SEI/IAS)

4) Get on with landing HEPI/actuator attachment, etc.

5) Final round of BS PIT SUS mechanical Offloading 

TITLE: 06/23 Day Shift: 1430-2330 UTC (0730-1630 PST), all times posted in UTC
STATE of H1: Planned Engineering
INCOMING OPERATOR: None
SHIFT SUMMARY: SPI in-chamber installation was finished yesterday, CRS work, and alignment work continued ending with Jason recentering the BBS OPLEV as of 23:30 UTC.
LOG:                     

• A wildfire sparked up to our southwest by Benton City around 19:00 UTC, the smoke could easily be seen to the left of Rattlesnake. Multiple helicopters, and a plane were out fighting it. As of the end of shift it looks like they've mostly put it out based on the smoke or the lack of.

J. Kissel, for J. Freed

At the last minute on 2026-06-18, Josh figured out that we had errantly been sending the SPI's double mixer two copies of a COS wave of 4096 Hz rather than a SIN and COS (LHO:90680). This was because someone had errantly put a 90 [deg] rotation in the SIN analog output path. He fixed it before he left, but didn't save it in the SDF system and it got lost during Monday's bootfest of h1seih23 (LHO:90689). I'm not sure if Josh (or anyone) tuned it to any specific value, but I've changed the SIN analog rotation phase to 0.0 and accepted it in SDF so we don't lose it.

Josh may add some more details in the comments below, in the comments to his LHO:90680.

J. Kissel

I'm looking to get some overnight data of the SPI pathfinder so start looking at science signals and debugging. Jennie has been turning OFF the laser via the external-to-PSL Uniblitz shutter controller (see LHO:90491) over this past long weekend, and each night this week because she was worried about the stray beam from the FBR_PWRIN PD, but again -- this is 0.2 [mW] and not an issue for humans or equipment as long as the LVEA is laser hazard. She didn't know the power was so low, and followed her good instincts.

But yeah, I want to get some over night data, so I've turned ON (unshuttered) the SPI laser input via the Uniblitz shutter controller box by the PSL.
I'd like to leave it ON as long as the LVEA is laser hazard.

If we need to go back to laser safe and I'm not around, one can easily shutter the laser by turning OFF the controller box: power it off with the rocker switch on the back, or unplugging the box from the wall, or both. With no power, the shutter defaults to CLOSED, which blocks the SPI pick-off path (only) from inside the PSL, thus making the entire SPI laser system laser safe anywhere outside the PSL.

Ibrahim, Oli, Thomas, Betsy

Before we flew the BBS in, we did a sanity-check experiment to compare slider values with a laser-pointer optical lever. This experiment follows Jeff's G1200698 math for an optical lever (picture also attached).

Setup:

A laser pointer was pointed at the BBS HR side such that the HR and AR reflections could be seen on a paper near the laser pointer. The initial position was recorded on a paper, and then slider values were driven to their maximum range in P and Y. The distance between the laser source and the HR surface was recorded. The experiment was done with BOSEMs and then with QOSEMs.

Method:

When the laser pointer is incident on the BBS, there are two reflected beams (AR and HR). The average distance of these was taken along with the average distance from the BBS. This was done by calculating the distance to the BBS and adding the half-thickness of 30mm. This gives us "one beam" and "one distance' to go off for P and Y. See the attachment below for an image of the paper used during the experiment. This was done once for BOSEMs and then again with QOSEMs.

A few notes:

• The setup was replicated and the distance was remeasured on a different day and this is why the distances are different.
• The QOSEM as-it-is is inversed in its actuator sign convention and you can see this in the attached diagram. One of our action items is to rectify this. So right now, +P = Pitching up instead of down with respect to the HR. In general, we will abide by T1200015, the SUS/Actuator Sign Convention.

Results and Analysis

Pitch is in purple and Yaw is in Orange. The equation used was obtained from "G1200698, the Oplev Infamous Factor of Two", which is essentially the same outline as our experiment, a diagram of this is attached.

This equation is: DØ= DS/2L Where,

Ø is the angle in radians (adjusted to urad below), S is the displacement in the beam spot, L is the distance from the laser source to the optic.

For these data, the total range of the Opticalign sliders was considered (from lowest negative to highest positive). Thus, a half the range would apply if the sliders were zeroed.

Conclusion:

• For a QOSEM Drive:
  • Pitch at M3 read +-475urad (1/2 the total sweep of 949.72 shown above), which is equivalent to 3.16 Opticalign Cts per urad.
  • Yaw at M3 read +-3365urad (1/2 the total sweep of 6730.64 shown above), which is equivalent to 1.19 Opticalign Cts per urad.
• BOSEM and QOSEM ranges are comprable
• No cross coupling was observed. Remember that this was done on the test stand a few weeks ago, before we found P:Y and Y:P coupling in chamber.

Next Steps:

Next, we're going to compare these results with the M3 AOSEM readouts once these are reliable. We will also compare these results with the model (which is used to corroborate slider calibration). Ultimately, this will help with the calibration of our slider values (both within the actuator slider actuation vs. sensing and from counts to real units).

Once we get a working OpLev (it's not on the BBS right now), we'll be able to get a more fine measurement of this kind.

Again, this is just a gross alignment check for information about our pointing and the capacity of our actuation.

J. Kissel, J. Wright, T. Shaffer, J. Warner, J. Freed
(belated aLOG covering 2026-06-22 activity)

After enjoying a lovely long, Juneteeth Holiday weekend, TJ and I wrapped up the few loose ends that were left after the 2026-06-18 super push to get everything installed (see last update; LHO:90676). 
   - As a fall out of LHO:90667, Jim, Arnaud and I discussed the pros and cons of using the current, D1000907-v7, balance mass "Payload & Suspended Mass Assembly" arrangement in regards to the W9 corner. We concluded that leaving the D1000907-v4 configuration for this corner in place -- i.e. having 11.6 [kg] of mass in small, modular, optional components -- was better that having one "giant" 10 [kg] mass in "the same" location (on the table top, but in the same W9 corner). As such, we left the plates as re-installed on LHO:90713, with the acknowledgement that the final configuration to create a balanced ISI may be different even further than D1000907-v7.

   - Using a beam profiler, we measured the Beam Profile of the beam returning from M_C1 on the HAM23 ISIJ Reflector. This is to, at least roughly, confirm the radii of curvature of the M_C1 mirror. We expected a 2 [mm] diameter beam, and we got a 2 [mm] beam diameter. A more thorough aLOG to come.

   - We did NOT address the stray beam that Jennie mentions in her summary -- yet. I'll also write a separate aLOG on this, but in short -- it's a ~0.2 [mW] beam that's what ~18% reflection there is off of a silicon diode, and it hits the -X/-Z rim of the chamber of the +Y door, and the (splotchy) spot size is ~4-5 [mm] in diameter. The SPI team has been aware of this beam since testing in the optics lab (see mention of it, e.g. in LHO:90455), but hoped that it would land on some part of the SPI Shroud assembly, but it *just* misses it. 


With this last item, we consider the Installation and Integrated Test Plan COMPLETE (T2500024) ... to as good as possible with the HAM3 ISI still locked.

And that ... qualifier is a "just in case" qualifier, as it's a "we'll see what happens" when the ISI gets unlocked, and we've got an excellent amount of remote adjustability to be able to recover the MEAS IFO's alignment if HAM3 moves a lot between 
    - "locked," 
    - "re-balanced and damped" and 
    - "isolated with feedback and DC positioning engaged."
Essentially, we've launched the SPI pathfinder into space, and now its up to out built-in remote controlled actuators and sensors to take us home to achieve our scientific goals (see T2600019).

Super congrats to all, and similarly large thank you. We did a thing!

WP13303 Reimage PWRCS Beckhoff, update Corner HEPI Pump controller INI

Patrick, Erik, Jonathan, Dave:

Patrick created a new H0EPICS_PWRCS INI file. All of the EPICS channel names have changed since the first version of this system, and 40 of the new channels have old-channel equivalents. To preserve recent minute trends, the data acquired so far were transferred to the new names as part of the upgrade.

The sequence followed was:

 - Remove the old names from the edc_green_ioc IOC. 43 disconnects at this point (40 from H0EPICS_PWRCS and 3 from H1EPICS_HEPIPUMPL0)

 - Create the copy scripts to be ran on TW0,TW1 to copy the old files to the new, so that the new channels will "continue from where we left off"

 - Create the new H1EDC.ini and verify its changes.

 - Stop daqd on TW0 and TW1

 - Run the copy scripts on TW0 and TW1

 - Restart the 0-leg, followed quickly by restart of the EDC.

 - Wait till 0-leg was fully operational, including TW0

 - Restart the 1-leg

All went well except for the usual GDS1 needed a second restart, FW1 restarted itself after 5 minutes.

Continued from 90691. Betsy, Keita, Sheila, Oli, Eric, Camilla, Ryan S, Jenne

First thing, we went back to alignment we had ITMX SQZ  the "flashes" and "fly-bys" yesterday, put the same size oscillation on the BS, but saw no flashes of fly bys. Betsy took a video of what the two SQZ beams looked like at PR2 at this alignment, they were crossing each other. 

We then rethought the alignments from yesterday, we are leaving:

• ITMX and ITMY at March DRMI top mass OSEM sensors
• ZM4 and ZM5 to get SQZ beam on SQZT7 irises
• ZM6 to O4 sliders.
• SR3 at the Oplev/slider positions of March DRMI/O4 (all similar)
• SR2 Yaw slider back to O4, Pitch to get SQZ ITMY beam on AS_AIR (+440urad PIT sliders different from O4).
• SRM aligned to improve SRY fringes (+200urad PIT sliders, +400urad YAW sliders different from O4).

Once we did this, we had SRY fringes, Betsy and I then adjusted the BS to bring the beams back on top of each other at PR2. This was a change in the BS sliders from (P 1000, Y 1290) to (P 1350, Y 1180).

Keita then wobbled ITMX and found the SQZ beam off ITMX! Aligned ITMX to get MICH flashes at AS_AIR, AS_A/B and ISCT1 REFL. He had to move ITMX from (P -96, Y 104) to (P -280, Y 94). 

He then got the PSL IX and IY reflection in the AS_AIR camera in this alignment. Sliders at this alignment attached. 

However as we are not sure we trust the ITMY alignment form top mass osems (from 90551), we want this Pitch change to be in ITMY, so ITMX is at it's known DRMI pointing. We could try to walk these, maybe along with SRC which might reduce some of our Pitch differences in SR2 and SRM. This change is so small that we will go ahead with mechanically offloading BS Pitch. The BS YAW will not mechanically changed. 

The OMs are also not currently in their O4 sliders/osems position, attached. We could verify the SQZ beam pointing by putting these back if we thought that was needed.

Jennie W, Jeff K, Josh F, TJ S,

Summary: SPI interferometers aligned, shroud installed successfully. One stray beam and possible clipping on shroud to investigate/deal with but otherwise all problems were solved. Obligatory success photo taken by Josh.

Today we:

• reseated the M_M5 mount in its holder. We had troubles with this stick-slipping when we tried to move it. Jeff replaced the 45 deg adaptor part D2400336-v4 S/N 001 type 1 with S/N 002. This seats better in the holder.
• realigned beam to QPD A and B using picomotors M_C1 and M_M1.
• Confirmed other two picomotors M_B4 and M_M2 work and their pitch DOF (in breadboard basis) matches left/right on picomotor controls.
• We had some weirdness in the meas IFO which Josh has solved and will discuss in other alog from Josh (LHO alog #90680).
• During this debugging we used the oscilloscope to plug into analog outputs for meas and ref IFO and a commercial RF source.
• We used the breadboard steering mirrors to get 85.9 % heterodyne efficiency on the reference IFO PD A and 88.9% efficiency on reference IFO B.
• With no adjustment the measurement interferometer PD A had 86% efficiency and PD B had 93.4% efficiency so we will take this (ch4 on scope is PD A, ch2 is PD B).
• This afternoon we installed the shroud around the ISIK breadboard.
• We had to move the fiber spools to avoid the bottom shroud bracket and these are now close to the edge of the beam tube where it meets HAM3 in the -x direction.
• After installing the shroud panels both QPDs had dropped in power by ~2V, will investigate on Monday as the beam does not appear to clip the shroud when we check with the card.
• There is also a stray beam from the FBR_PWRIN_REF PD that hits the upper shroud stand-off on the lower panel on the -y side and then hits the door cover, this will need dumped somehow.
• After moving the fiber spools I checked the FBR PWRIN channel for meas and ref and these both drifted after we rerouted the fibre spools but the level seems to have been dirfting more than this since we got into chamber this afternoon at 14:20. The first cursor in the photo is when we started moving the spools and the second is when I shuttered the laser.

J. Kissel, J. Warner

During SPI install we removed the following side-wall balance mass from the W9 side wall (see first bullet of LHO:90558):
    Plate              Mass [lbs.]   Mass [kg]    QTY    Total Mass [kg]
    D071200 Type 04    7.9           3.583        3      10.750
    D071200 Type 01    1.1           0.499        1       0.499
    D071200 Type 00    0.6           0.272        1       0.272
    D071201            0.1           0.045        3       0.136

    Total Mass                                           11.657

First attachment is a diagram to convey which side wall I'm talking about.

This is inaccurate with the latest version of the ballast / balance mass inventory, D1000907-v7, which states that this side wall has only 1x Type 04 (3.583 [kg]) and a 1x Type 03 (2.041 [kg]), for a total of 5.625 [kg]; much less.

Remember, from LHO:90504 that the total SPI mass is 12.599 [kg]. 
This *excludes* the mass of the lower ISI Shroud baffle (D2400106-v4) and the three upper HAM Table Baffles (D2600007) with all their bracketry and bolts, currently only represented only in e-drawings posted to D2400103-v6.

After talking with Jim, he wants more mass in this -X / +Y corner of the table, because that's where all the new stuff is. So he wants this corner "over" or "heavy" (because the new stuff must be in a fixed position) so that he can adjust the *opposite* corner of ballast mass (which has more open table and side wall access and thus is adjustable).

As such, he says "put all the at 11.657 [kg] back on the corner. 

So we will!

Second attachment is a picture of the wall mass arrangement prior to us removing it taken on 2026-06-09.