John inspected the conflat on HAM-8 with the ding in it and pronounced it usable. (The gasket that was removed at the start of ICC showed the imprint of the ding, clearly establishing that the defect had been present prior to the ICC crew removing the conflat.) The Al port covers were located and sent to clean and bake to replace the foil that is currently covering some ports and we will not remove the cleanroom from the HAM-7/8 area until those port covers are in place.

Activities Today:

Ski worked on Mid Station air systems
Cabling work by the SEI crew on the BSC8 ISI
Patrick made some Dust Monitor code reboots
Prep work for HAM in-chamber cleaning

The attached red-lined drawing shows the proposed solution to the Top Mass interference issue, labeled 1) in my previous report. It is a small hog-out of (the adjacent) one of the two Top Mass upper plates. This obviously requires some dis-assembly, but appears to be the simplest and quickest solution. By wrapping the part in foil and milling it in a vise, it can be done with minimal contamination using alcohol as a cutting fluid and careful handling. Then we could possibly get clearance to clean it up in the lab and re-install.

I replaced the klixon switches at X end and mid stations for the drier towers(for the control air compressors) in both chiller yards.  All systems were returned to normal operation following this.

Jeff K., Jeff G.

The BSFM_MASTER simulink model was modified today to accomodate the binary input/output library parts.  

The modified file is in:
 '/opt/rtcds/lho/h2/userapps/release/sus/common/models/BSFM_MASTER.mdl'

The new model "h2susfmy.mdl" was compiled on h2build and "make-installed".  After a DAQ reboot, the new model was confirmed to be functioning.  

Attached are plots of dust counts > .5 microns.

1)
I have found that the interference we've been experiencing between the installed Upper Mass and Tablecloth is an actual design error. There is a slight (.017") nominal interference at the problem location in the assembly model (D1000549 'UPPER INTERMEDIATE MASS TOP ASSEMBLY'), pretty much as we have been seeing. This interference was designed in originally (over a year ago), and was never caught. It is quite buried in the assembly. I've attached two section views of the assembly model with the rear Tablecloth Plate hidden, to give a look at the problem.

2)
Jeff measured the distance from the overhead table to the Top Stage Blade tips (with full suspended load) using a steel ruler with the following results.
To left Blade tip: 112 mm
To right Blade tip: 116 mm
The nominal distance to the blades for that measurement is 103.7 mm.
Was this to the top of the tips, or bottom??? I believe it was the tops.
Also, I would like to repeat this measurement, this time to the blade roots.

3)
We would also like to take another sanity measurement (with full suspended load) from the overhead table, down to the top face of the Suspended Mass's Top Plate D1000393. The nominal dimension measured in Solidworks is 18.98" [482 mm]. See the attached sectioned view (3'rd file).

The crew continued conflat installation on HAM-8 until they ran out of parts. (The only remaining cleaned conflats-feedthroughs are for the one-arm chambers.) Doors were replaced on HAM-7/8 to clear the way for HEPI plumbing work. The west door of HAM-7 had some worrisome-looking spots on it which were scraped off with a razor blade, FTIR-ed, wiped down, and re-FTIR-ed. 

Dave B., Jeffrey G.

After noticing the extra noise from the M1 "F1" OSEM on the FMY, the input filter for the "F1" OSEM was switched on and off to see if it was contributing to the noise.  Switching the filter off settled down the signal and the foton .txt file was investigated for discrepancies amongst the OSEM channels.  The apparent difference was found in the "F1" filter bank which had a different gain.  These input filters should all be identical with a zero at 10 and a pole at 0.4.  The "F1" input filter now matches the other OSEMs' and the noise from this signal has settled to reasonable levels.  

Travis S., Jeffrey G.

On Friday, the diagonalization of the OSEMs on the M1 stage of the FMY was completed for the Yaw DoF.  For Yaw, the drive was a sine wave at 1.4Hz with a 200ct amplitude.  The first attached plot shows the "F1", "LF", "RT", and "SD" OSEMs are isolated nicely at ~30dB from the two Vertical DoF contributing OSEMS - "F2" and "F3".  There was a bit of tweaking involved as far as the amplitude of the drive to see a nice transfer function, but after adjusting the alignment of the OSEMs, an amplitude of 200cts was enough drive to see the isolation.

For the Vertical DoF - "LF" and "RT" contributions - the isolation from the other OSEMs was not as clean. The subsequent two attached plots labelled "*_vert_*" show the first drive at 1.1Hz with 2500cts and the second drive with 3000cts.

For the 2500ct amplitude transfer function, the other ("F2","F3", "SD") OSEMs were isolated better at ~20dB, but not quite down to the desired ~30dB. "F1" only displayed a 10-15dB isolation.  The "F1" OSEM response was noticed to be coupled undesirably too much to the Vertical drive.  The filtered signal from "F1" (after the OSEM "input filters" implemented in the medms via foton) had some unexpected noise contribution.  There was some investigations into the signal chain/electronics rack to pinpoint the cause of this added signal noise.  No initial electronics issue was found, so the investigation continued into the software implementation.

For the 3000ct amplitude transfer function, the isolation for "F3" and "F2" was much better at ~30dB.  The "SD" isolation was at about 24-26dB.  "F1" still had the same coupling issues and did not improve isolation from the increased drive amplitude. The "F1" OSEM noise was noticed at the end of the day to settle down when the input filter was turned off on the medm.  Further information to come.

The flags were adjusted with tweezers to align them parallel to the LED light and PDs as well as the OSEM housing themselves.  These adjustments were done in between measurements to improve the isolation.

Apologies for the sideways plots in the pdfs - it was DTT.

After installing the h2sustmsy.mdl last week, I'd finished up the infrastructure by creating a set of generic MEDM screens for H2 SUS TMSY. The overview screen is now linked off of the sitemap. I attach a screenshot of the over view screen (which should look strikingly familiar to other SUS screens). 

The files can be found in 
${RTCDSROOT}/userapps/release/sus/common/medm/tmts/*.adl

Of interesting note -- though the TMTS has its top stage controlled by a stage that is virtually identical to a QUAD TOP stage, the "gotcha" here is that the TMTS TOP stage (M1) is rotated 90 degrees in its "cube" (which on the QUAD is the upper structure cage). As a result, where F1, F2, and F3 control LONG and PITCH on a QUAD (and BSFM), they control TRANS and ROLL on a TMTS. Similarly, where TRANS is the "lonely DOF" having only the SD OSEM to control it on the QUAD, LONG is the "lonely DOF" controlled by the SD OSEM on a TMTS. Hopefully, the cartoon and Cartesian coordinate definition on the overview screens help differentiate this fact.

Late Thursday and Friday morning we worked on OSEM installation for the TMS on the
test stand. We found issues with getting all of the OSEMs aligned with the flags.
We checked the levelness of the test stand overhead table and found adjustment
needed there. As I was leaving the site, Bram and Kieta were heading out to the
End with some jacks and shims for under the test stand feet (at the floor) to
level the overhead table. They expect to make further progress once the table is
level.

Attached are plots of dust counts > .5 microns for today and yesterday.

Filiberto C. fixed the communication wiring to the dust monitor at location 7 in the LVEA (near the previous location of the H2 electronics racks). Michael R. swapped the dust monitor at location 10 in the LVEA (for the H1 PSL enclosure) back to the 227B model.

Second vacuum, inspection, and post-work FTIR samples were completed at HAM-8. One door needs a light vacuum and wipe-down before it is re-installed on the chamber and there are still a number of conflats and feedthroughs to be put on. The FTIR kits will be sent out on Monday.

Mark Lane (Apollo) showed me a damaged conflat knife edge that he had discovered while installing aLIGO feedthrough flanges today.  The flange in question is on the 10" nozzle's 12" CFF located on the bottom center of the north end of aHAM8 -> I inspected it.  It is a deep depression of the knife edge as would be expected from a perpendicular impact and not a scratch as is typically the case. I did not attempt to repair it at this time.  

I'll point this out to John W. when he gets back to give him the option to inspect it himself before proceeding.  

(Corey, Greg, Jim, Vincent)  This work occurred from roughly Tues-Fri of this week

Our Seismometer Sensor Issues:

GS13's:

The small cable feedthru flanges on 5 (of 6) GS-13 Pods for the BSC8-ISI were installed with incorrect hardware.  To not risk possible future vacuum leaks, it's been decided to swap out these (3) Horiz & (2) Vert GS-13's.

Trilliums:

The pressure sensors inside the vacuum pods were not operational for all (3) Trilliums.

Another Galled Bolt From GS-13 Installation

On the first GS13 (Vert @ Corner1), 1 (of 3) 1/4-20x3" bolts galled and broke; see attached photo for the broken stump.  This shouldn't prevent re-installation of the new GS-13, but now this GS-13 will have one less bolt to secure it down (which will make it similar to the Corner3 Vert GS-13 on this same Unit!).  Alas, this is an ailment bound to happen when installing / removing / Installing / etc. etc. Seismometer Pods.

Reaming GS-13 Pod Bases

Just before we were set to install a new Vertical, Jim remembered ALL Pod Base dowel pin holes are undersized for our dowel pins!  So, this meant we had to stop work, get reamers, and do some on-the-fly reaming, before installing Pods.

Drop Tested GS13's

Because the Test Stand and Cleanroom are non-user-friendly, we had to come up with ways to lift the GS13s up to the ISI.  The GS13's were initially put on a pallet and were going to be lifted via the forklift.  Unfortunately, the first attempt at this caused the GS13's (S/N 77 & 103) to be dropped a small distance (~1ft?)---basically we used a flimsy pallet, and when the pallet was lifted, it tipped, and the two GS13's fell to the ground. 

We ended up heaving the Pods up to the ISI by hand.

And Here We Are, Yet Another List of GS13 Locations.

Corner1

• Horiz:  33
• Vert:  79

Corner2

• Horiz:  31
• Vert:  94

Corner3

• Horiz:  16
• Vert:  26 (installed previously)

New Set of Trilliums Installed

Since we received these this week, these monsters were also installed.  Here are the S/N's:

• Corner1:  32
• Corner2:  20
• Corner3:  19

For bookeeping's sake, here are the Seismometers which were removed/dropped that are being sent back to LLO:

GS13:  68, 76, 84, 85, 89, 77(dropped), 103(dropped)
Trillium :  12(dropped) , 26, 36

Note On Connectors

All of the connectors for the GS13's, Trilliums, & L4C's, which were disconnectd for this swap, were re-connected, but we didn't screw them to their Pod plugs this time.  In case any of these Pods fail, we don't want to deal with the tiny screws for these connectors more than we have to (due to cross-threading/galling issues we've seen before).  We must make sure all connectors are securely screwed down before installing an ISI into a BSC/HAM chamber.

Trillium Door Weights

Have noticed a few of these door weights being installed with the helicoil tangs still in place (or the helicoils not installed at all); the helicoils are generally in the D0902614 Trim Masses.  Found that one of our bolts came close to seizing in one of these weights (looks like a tang was mauled in the threads).  I did not reinstall these door weights because we will need new long bolts to clamp them together (the previous bolts are too damaged to re-use).

• Chamber cleaning at HAM8
• TCS alignments
• Wiring, piping at H2 PSL

Maybe the logbook should allow multiple tasks?  Many things could span various "tasks" or subsystems.

OK, we've experienced many failures of the Triax style connectors used for the BSC-ISI CPS (Capacitive Position Sensor) electrical feed-thrus.  These are new to LIGO with aLIGO.  None have been installed on chambers yet.  We thought the eLIGO used BNC connection (HAM6 CPS) was much stronger but not so much it turns out.
We've had many breaks of these Triax connectors before, during, and after Clean & Bake.  The manufacturer measured 6 ft-lbs of torque breaking the triax.  They modified their design and got that up to about 12ft-lbs.  We will install these onto BSC8 & BSC6.  They have further changed their design but have not fabricated this yet so we don't know the strength.

Here are some photos of the breaking I did with the eLIGO BNC feed-thru.  This is the eLIGO triple style--aLIGOs will be duals from a different manufacturer.

*** Bottom line--the three all broke at ~9 ft-lbs. ***

First--Clamped Feed-thru in vice.  A 12" machined rod slides over the connector--12" exactly to the failure point.  I also clamped a round stock above the feed-thru so I could measure deflection while loading-see third image.
I then hung a bucket from 12" groove (first time I used a spring scale too)(image four) and then carefully added door flange bolts to the bucket while measuring deflection of the connector the best I could.  It broke after the addition of the fifth bolt(image 5 & 6.)  It didn't break right away but before I could measure the deflection.  I measured a little increase in deflection after the fourth bolt but not much more than measurement error.

So I repeated with the second connector on the conflat this time eliminating the Spring Scale to make sure the bounce of the scale wasn't impacting the outcome.  This time it broke again after the fifth bolt was added to the bucket but I was able to measure deflection--about .005" compared to near zero for all the previous weight additions.  OK somewhat repeatable.

For the third feed-thru on the conflat (at Vern's suggestion) I turned it around increasing the moment arm to 13-3/4"--again it broke after the fifth bolt-see photo7.  I measured the deflection about six times as far away from the break point as before and so a lot of movement of the metal before it broke was observed--I didn't get an image of the measurement point--as close to the conflat as I could opposite the ceramic side.

I then weighed the bucket and five bolts--8.5lbs +-.25lbs.  The 12" rod measured 1-1/4 lbs.