B. Weaver, J. Kissel
WP 12109

Betsy, myself, and several others are looking to propose redlines to the WHAM3 flange cable layout (D1002874) in prep for O5 (see T2400150 for DCN and discussion there-of). 

However, in doing so, we discovered that the flange layout may have double-counted the shared cable for the MC2/PR2 M1 (top) RTSD / T1T2 OSEMs. Other drawings (e.g. the Cable Routing D1101463 and the wiring diagrams D1000599) indicate "yes, there's an extra 'SUS-TRIPLE' entry somewhere between the D6 and D3 allocation," but we wanted to be sure.

As such, Betsy went out to HAM3 today and confirmed that YES the MC2/PR2 M1 (top) RTSD / T1T2 cable, labeled "SUS_HAM3_002" in the wiring diagram or "SUS-HAM3-2" in real life, does come out of D3-1C1 and *not* out of any D6 port, and thus we validate that D6's list of 4x DB25s to support the 'SUS-TRIPLE' (i.e. MC2) in D1002874 is incorrect.

Pictures attached show the D3 flange, and highlight the SUS-HAM3-2 cable at the flange going to D3-1C1, and then the other end of that cable, clearly going into the MC2-TOP/PR2-TOP satamp box in the SUS-R2 field rack (S1301887).

J. Kissel, S. Koehlenbeck, M. Robinson, J. Warner

The LHO install team (Jim, Mitch) -- who has experience installing in-chamber fiber optic systems -- have reviewed the two options put forth by the SPI team for optical fiber routing from some feedthrus to the future location of the SPI follower breadboard on the -X side wall of the HAM3 ISI using Eddie's mock-ups in D2400103. Both options (WHAM-D8 or WHAM-D5) are "evil" in several (but different) ways but we think the lesser of the two is running the fibers from D5 (the currently blank flange underneath the input arm beam tube on the -X side of HAM3; see D1002874).

In support of this path forward, one of the primary evils with D5 is that access to it is *very* crowded with HEPI hydraulics piping, cable trays, and various other stuff. 

Here I post some pictures of the situation.

Images 5645, 5646, 5647, 5648, 5649, 5650, 5651 show various views looking at HAM3 D5.

Images 5652, 5653, 5654 show the HAM2 optical lever (oplev) transceiver, which is a part of the officially defunct HAM Table Oplev system which -- if removed -- would help clear a major access interference point.

• Grounding check for ISC. Problem found and fixed.
• Used the viewport simulator on +X door to see if the beams (ASAIR and OMC TRANS) will come out of the viewports. They'll be fine.
• Took pictures of the table layout.
• Jim put an wafer on the table.
• I tried to take good picture of the OMC DCPD surface, but it's impossible without removing the OMC shroud, so we gave up.
• Rahul took SUS TFs and they're fine.
  • HAM5 gatevalve was closed, foil removed, and the plastic cover was put on.
• Jim unlocked ISI.
• Cleanup work by everybody.
• HAM6 ready for the door.

Some details.

Grounding.

I and Jim checked the ground loop for everything on A6 flange (ISC) again. I disconnected in-air cables from the feedthrough, put a DB25 breakout board on the feedthrough and tested the connection between pin13 (in-chamber shield) and the chamber ground.

The only genuine problem we found and fixed was a ground loop for ASC-AS_C QPD cable. Jim rerouted that cable that was bunched up under the ISI and it was fixed.

Also, when testing ground check for F1(OMC DCPD)/F2(OMC PZT)/F3(OMC QPD), disconnecting these three isn't enough. Don't forget to disconnect 3MHz cables (D4-2B1 and D4-2B2), otherwise there's this path which gives you anything from a few hundred kOhm to 4MOhm. We wasted some time because of this.

F1 in-chamber shield <-> DCPD preamp signal ground <-> RF cable SMA shield <-> SMA Rack ground -> chamber ground.

After disconnecting RF cables, don't forget to torque them using a dedicated torque wrench for SMA.

As of now, in-air cables on D6 are as follows.

AS_AIR beam position check

TJ and Tony installed viewport simulator on the +X door after adjusting the hoop positions on the real door.

From the picture I previously took (ASAIR_beampos.jpg), I know that the ASAIR beam position is about the radius of the BDV optic toward -Y direction at the BDV when it's open.

So I placed the camera so that the lens is co-axial with the line connecting the center of the ASAIR beam position by the BDV and the center of the ASAIR-AS_C splitter as good as I can (ASAIR_likely.jpg). Green cross is where I think the beam will be by BDV. Cyan line is the center line for ASAIR-AS_C BS. (Red line is the center line of the lens just so you know how big the parallax could be.)

Anyway, as far as the centering on the BS is not terrible, the beam will come out w/o any problem.

I also repositioned the camera to assess the worst case scenario (the beam will be very close to the left edge of the BS and therefore comes closest to the right edge of the viewport given the fixed position of the ASAIR beam by the BDV), and the beam will still come out of the viewport.

OMC TRANS beam check

We haven't done anything as there's no reason for a big change except that we changed how OMC ISC cables are routed inside OMCS, and that we removed OMCS suspension offsets. Just took one picture through the +Y hoop of the viewport simulator.

Table layout pictures

For posterity.

Couldn't take a good DCPD picture

I tried to take a good picture of at least one of the OMC DCPDs through the space between OMC shroud panels and through the OMC breadboard itself but I couldn't. The view is there but lighting is not (DCPD.jpg).

HAM7:

• HAM7 -> HAM5 -> HAM6 injection, alignment check and power measurement (65110).
• Rebalanced.
• Ground loop check was done, ready for the door. According to Rahul, there's a ground problem for ZM3, needs to be revisited tomorrow.

HAM6:

• New picture of OMCR DCPD shows that it probably has a cracked glass at the edge (65117). No replacement/mitigation at this point.
• Septum viewport adapter was removed (O-ring was stuck to the septum adapter plate side and had to be pulled manually by Jim).
• Septum accelerometer was put back on (see picture), cable was plugged back in, and the digital channel was confirmed to be active.
• Visual check of OSEM blade clearance was done for OM2 (because we rebalanced on Fri.). They were good. OM2 OSEMs were also recentered.
• Preliminary TF for OM1/OM2/OM3 were done while the HAM5-side purge was turned off by Gerardo (HAM6 covers were off). They looked OK but the coherence isn't good for OM1 and OM3. The purge was turned back on after that.
  • Rahul will retake TFs overnight. After consulting with Gerardo, the HAM5 gatevalve was left half-open with aluminum foil cover. That way the corner purge air mostly goes out of that and won't blow into HAM6 so the TFs will be hopefully better. HAM6 purge is on as well as HAM5 purge, so they pressurized relative to the surrounding clean room.
• Jim and Ryan worked on cable routing.
• Jim and Ryan worked on rebalancing.

What's yet to be done.

• If Raful's overnight TFs aren't great, he will talk to Gerardo to turn down the purge.
• Ground loop check for HAM6 needs to be done as some of the cables were rerouted.
• Ground loop fix for HAM7 ZM3.
• Exit photos and sweep.

Following up on Keita and Koji's converation, Sheila and I took a closer look at the OMCR DCPD in HAM6 to see if there are any cracks. We could not positively identify anything resembling a crack on the DCPD (images attached). We could also not see where the DCPD was hole punched. What is the typical size and placement for these holes? More images here. The DCPD is definitely quite dirty. We compared against some DCPDs of a similar design in HAM7 and can confirm that the OMCR DCPD is visibly much dirtier than those.

We also inspected the OMCR shroud input aperture for signs of pitting and other damage. There is no damage to report. We spent some time looking at the output aperture because we saw some potential damage along the aperture edge. After taking photos at various different angles to rule out reflections, we determined that apertures are fine, save for the occasional speckle on the shroud and the rouge edges likely left behind by the machining process. Images attached. More images here. 

We calibrated OMCR DCPD using Pcal integrating sphere. Numbers are to follow. (But note that Koji is worried that maybe the cover glass for that PD was cracked when they punched a vent hole on the diode case.)

We fixed the beam dump that was too high.

After that, we investigated the DCPD signal ground (alog 65081) and we think we know the cause of the problem but we don't have an easy fix. We'll leave it as is.

It seems that one of the cables (either PZT or QPD) that goes to the OMC breadboard was touching the OMC shroud panel. We'll revisit that in the morning, adjust that, then confirm that the laser is still hitting the OMC QPDs. If not we'll realign.

It seems that we have ground loop problem (PZT shield is connected to the chamber). We'll revisit that at some time, maybe in the morning, or maybe after we're done with the injection from HAM7 to HAM6.

All components that are supposed to be installed in HAM6 in this vent were installed. I decided that the centering of the aux laser beam on the 1st iris was not great after all (1mm or so in YAW), so I realigned the laser to two irises carefully. This didn't have a huge impact on the downstream alignment, but it had some impact. Anyway, ASC-AS_C, WFS, OMC itself and OMCR path were readjusted.

See layout1, layout2 and layout3.jpg for the relevant paths including (but not limited to) things that were installed.

There are some caveats, namely,

• DCPD was super filthy and now it's just filthy. Do people clean this at all?
  • Initially it looked like DCPD_initial.jpg. These spots are hard to see by eyes but easy to capture on camera.
  • I used a puffer and it didn't make any difference. I used Vectra wipe with a mild pressure to wipe the glass cover and it became better (DCPD_1st_wipe.jpg) but not good. Second wipe didn't make much difference if at all (DCPD_2nd_wipe.jpg). Maybe some of these are inside surface of the glass? Anyway we gave up and installed this.
• One of the V-dumps we installed with thicker (0.5"?) post is just barely usable for a 4" beam height.
  • See picture, the top of the metal holder for black coated steel is only 1/8" lower than 4". Inside the V, ghost beam flies 3mm above the shiny aluminum, reflected by the coated steel, flies above the shiny thing again, and again and again. This is not how we use V-dumps, you need a much shorter post. If we have a replacement post we'll use that tomorrow.
• We still need to calibrate OMCR DCPD using an integrating sphere.
• Couldn't see reflection of ASC-AS_C, so the beamdump location was just eyeballed.
• Couldn't see AR reflection of the lens for OMCR path, I expect that to fall on the V-damp for the 99:1 reflection.

Also, this is not a caveat, but there are three ghost beams generated by the OMCR DCPD because the glass isn't removed from the PD case.

Tomorrow we'll attack the above caveats in the morning and hopefully move on to OMC DCPD and PZT grounding investigation.

(Weaver, Kissel)

In HAM1, today we:

1. Unlocked HEPI (Jim)
2. Transition to Laser Haz - 2W (int sphere is gone so not limiting pwr)
3. Relock MC at 2W (RyanS, Kissel)
4. In-Air WFS driven pre measurement
5. Take some pics and get a baseline for BOSEM centering and blade height locations
6. Lock RM blades per E2200181 (blades stopped down ~0.5-0.75mm)
7. Checked centering of BOSEMs - still ~centered, maybe a 0.5-1mm low, but TFs and controlability didn't indicate we should adjust
8. RM TFs
9. Realign to WFS
10. WFS driven post measurement - see new higher peak (more in Kissel's log)
11. Pushed blades down the rest of the way to ~1mm down, reiterated 7-10 above - didn't see much change in this new higher f noise.

Next up:

• Realignment of LSC PD path to walk beam (see Keita's previous alog 63544)
• Install baffle on PSL-IN Viewport - out of nozzle baffle mounting rings so having 1 sent overnight from LLO while more are on order
• Ground loop checks  
• Relock HEPI for upcoming door activity

I went to the staging building to fetch a 50:50 P-pol IR splitter and saw Betsy and Srinath baking FC cavity tubes. Here are some pictures.

1st: They can bake two sets at a time and each set comprises two sections connected together. There are 48 sections (i.e. 24 sets), they need to be baked at 110+ Celcius for ~3 days, so this has been a long laborious process which is still ongoing.

2nd: All sets are pumped down before and while baking.

3rd: They already wrapped the heater element around the tube and they're ready to put the insulation material on.

4th: Insulation material is all velcro-ed up (except that it's not, Betsy won't be ready to turn the heater on until the joint in the middle is also covered up by a narrower insulation strip).

J. Kissel

Following the measurement & fitting (LHO:61313), and subsequent modeling of systematic error impact of various levels of compensation accuracy (LHO:61729), I've installed the updated compensation (in the ETMX_L2_COILOUTF filter bank) for the analog frequency response change to the H1SUSETMX PUM (or "L2") driver (described in E2100204, installed in Dec 2021). 

While the modeling effort (LHO:61729) only discussed State 3 for brevity, and "only" the analog low-pass filter in State 3 changed, I also installed updates to the compensation for the Acquire ON and Acquire OFF response as well, because previous measuring / fitting attempts (47166) were flawed: the measurement was done with the AOSEMs still engaged, leaving the AOSEM inductance zero in play which confused the fitting routine. Let's be serious, I really just didn't know what I was doing and I was rushed. Now I'm far more confident in my results because I understand the system much better, and I had the time to calmly measure all the pieces of the puzzle independently. The differences to State 1 and State 2 are small, but now *any* state we choose to run in during observation will be supremely compensated (at the 0.1% / 0.1 deg level, at least below 1 kHz), rather than just "OK" compensated (at the 1% / 1 deg) level.

In addition, all the switchable compensation filters (the "Sim"s in FMs 1-3 and "Anti"s FMs 5-7) have all been converted from "Zero History" input and "Immediately" output to having their input "Always ON" and output set to "Zero Crossing." This feature of the coil driver switching had been lost in the sands of time: having recently re-explained "how the switching and compensation works and why" to Brad, I remembered a slide from a 2009 presentation of mine, G0900112 Slide 9 (be sure to download the .pptx as the animations are critical to the point), and the last bullet on there covers that, under this "SimDW," "AntiDW," and "DW" "dance of the de-whitening plum fairies" model of compensation, the switching should be done on a Zero Crossing in order to minimize glitching that may result from the large, high frequency, gain changes between State 2 and State 3 when we run "LOW NOISE COIL DRIVERS" after achieving full IFO lock/resonance on our way to NOMINAL LOW NOISE.

Finally, for the first time ever, and motivated by the modeling work, I've included an AntiAOSEM filter, which compensates for the non-negligible ~855 Hz zero that arises from the coil's inductance. This filter remains ON at all times, regardless of the state of the coil driver. As such, because there's no intent to ever switch this filter on the fly, I left the input and output filter switch settings as "Zero History" and "Immediately."

Note: I've left the L2_COILOUTF gains as is, bu these will need to be retuned, since (a) the measurement method (LHO:11392 and LHO:9453).  involves driving a line a ~3 or 4 Hz, and there was significant frequency-dependent systematic error in the compensation around that frequency which was unrelated to the actual overall gain imbalance, and (b) there're brand new R and C electrical components in the State 3 path, so I anticipate some ~1% level change in the over all gain (and the coil balancing method claims to achieve 0.1% level accuracy in compensating for gain imbalance).

Attached is a screenshot of how the new filter bank looks when the BIO_REQUEST for L2 is set to State 2.
Also attached is a screenshot of me accepting the turning ON of FM9, which is where I've chosen to house the AntiAOSEM compensation.
Also, also attached is a new version of the State Machine Diagram, conveying which state has which analog response, and corresponding the compensation scheme.
Finally attached is a list comparing old vs. new filter poles and zeros.

I used a flashlight (through the OMC breadboard) to aim at the DCPD in reflection of the 50:50 BS as best as I can , and DCPD_B showed much stronger signal.

This means that:

1. DCPD preamp and whitening chassis work.
2. DCPD_B is the one that sees reflection as of now, DCPD_A is receivning the transmission.

See cartoon, this is a top view looking through the glass OMC breadboard (optical components are under the breadboard).

Maybe this isn't the best place to share this, but life can be frustrating as a deuteranope. Matplotlib has an easy to use colorblind friendly style built in. It can be invoked by adding the line plt.style.use('tableau-colorblind10') before starting to setup your plot (if you've used the usual import matplotlib.pyplot as plt). Maybe this is common knowledge, but I just found it while working on a script.

The first plot is one of the plots produced for the weekly CPS noise monitoring famis for ITMX. The second plt is the same one, but I added the colorblind style to the script. For me, on the first image the H3 and V1 sensors on both subplots are almost the same color. It's especially difficult to go between the legend and the traces on the first plot. Reading these plots on a backlit monitor make this even harder. More than a 4-6 traces, I have a really hard time telling which line is which.

On the second figure, it's much easier. Add in some different linestyles and thicknesses, and everybody's pretty plots can be appreciated by more people.

VOPO relocation was very smooth overall.

VOPO spacer (picture 1) was put in place in HAM7. The spacer didn't have any of the helicoils for 1/4-20, so we had to put them in place.

VOPO platform was moved to a lift cart, wrapped in C3 cover (2nd picture), moved to a big clean roon in the LVEA and unwrapped (3rd pic).

4th pic shows the VOPO attached to the installation arm with the tooling. (Note the aluminum spacer between the VOPO and the lift cart. This is the same spacer that was used when we uninstalled VOPO from HAM6.) We had to remove two balance weight bars at the side to lift the VOPO high enough.

After it was placed on the spacer, three persons needed to push and pull it a bit to set the yaw angle right while the VOPO base plate still contacts with two alignment pins on the spacer.

Georgia went in to HAM7 and used dog clamps for +Y and +X sides (5th pic) , I did -Y and -X sides. Two balance weight bars were reattached at this point.

6th and 7th pic show the newly installed VOPO in HAM7.

We pinched one long in-vac cable between the installation arm and the -Y side door flange while trying to move the installation arm out of the way. Externally, the insulation sheath looks somewhat deformed (8th pic) but it's not clear if there's internal damage. There's another cable in the 8th picture next to this cable, it shows similar deformed-looking sheath at similar location but that one wasn't pinched.

DCC and SN of this are D1000225-230" #1, S1104241 (pic 9 and 10).

Travis also took a lot of pictures.

[JenneD, SudarshanK]

We  lowered the ASC DHARD Pitch Gain to see if we can actually stay locked with smaller gain. We lowered the gain by a factor of 3 (from -30 to -10). The interferometer happily stayed locked but we didnot notice any significant change in the DARM spectrum (Attachment 1). During this time, the purge air/Kobelco was turned on as part of Tuesday maintenance, which might elevate the DARM noise floor but should be similar throughout our study. 

We also made noise injections at these two configurations via DHARD_P_EXC. The sensor corrections was turned off during one of those injections but we didnot see much difference between the two.

At next comissioning opportunity, we will probably try to measure sensing function at the lower gain value because we expect the sensing function to  change (move towards "normal") with lower gain.

Evan G., Keita K., Jamie R., Dave B.,

We tested the transient hardware injection infrastructure to make sure that we could still run transient hardware injections. While the IFO was not in low noise, we successfully made a DetChar set of injections. These were injected starting at 1233537100. We saw the injection on the TRANSIENT filter bank and on the PCALX RX PD. The STATUS did change from 3 to 40 and back to 3. The value was displayed weirdly by DTT, ndscope and dataviewer, possibly because it is a uint32 data type, but we believe it is doing the right thing.

The obs_channel_name variable in INJ_TRANS.py had to be updated to "GRD-IFO_INTENT".

We connected the 5-way coax to TNC adapter that Fil has made to the chamber. 5-way coax cable assignment is in D1300466.

We measured the TF from cable 2 to cable 4 while cable 5 was terminated (attachment 1), then from cable 2 to cable 4 while cable 4 was terminated (attachment 2) and both made sense.

The problem of this configuration is that shells of TNCs are touching with each other (attachment 3, note that cable 2 and cable 5 are terminated).  We put cable 1, 2, 3 and 5 inside different fingers in a glove to insulate things from each other (attachment 4).

For the moment 9MHz connector is exposed but it's not touching any surrounding metals.

For a short while we locked MC with 200mW and measured the coherence between REFL_A and REFL_B 9MHz, and it also made sense (see Georgia's comments).

Keita, Craig

We checked all the LSC PDs while locked in nominal low noise with normal modulation depth settings.  Remember that we multiply the scope output by 14 to get the actual demodulated signal, since the demod board picks off -23 dB for the RF monitor.

REFL9pp ~   560 mV
POP9pp ~    280 mV
REFL45pp ~ 1120 mV (!)
POP45pp ~    70 mV

REFL45 seems to be slewing at ~ 100 V/us, seems pretty fast.

Betsy, Sebastien, Slawek


Work on installation of AMDs on ITMx and ITMy has been completed. 
Each test mass has now 4 AMDs glued to the suspension flats.  All AMDs on ITMy are glued slightly lower than the nominal AMD height position. It appears that quad suspension of ITMy is  ~5 mm longer than the ITMx quad. We have noticed it after AMD installation. This new AMD position should not cause any AMD performance degradation neither increase thermal noise according to our simulation.
Additionally, due to the scratches on the test masses flats we decided to move AMD 4 on ITMx and AMD 4 on ITMy. The exact gluing locations are shown in the attached drawing.
The curing dynamic test of the epoxy EPOTEK 302-3M of all batches was passed, see attached picture.  The bond between AMD and TM looks good for each AMD (see the back face image of the AMDs). 
We can call this installation 100% successful.