HAM3 unlocked, ISI tfs look good.

Hugh unlocked HAM3 this morning after SUS people finished in there this morning. I've taken dtt tfs and they look ok. Hugh reported purge is high, so tfs look a little ratty, but resonances seem clear. 

Weekly PSL Chiller Reservoir Top-Off (FAMIS#6554)

• Crystal Chiller topped off with 250mL of water.
• Diode Chiller OK
• Checked Filters and they were OK.
• Closing Out FAMIS # 6554

PSL FE Power Watchdog Trip

At approximately 12:33 UTC (4:33 PST), the 35W FE power watchdog tripped, causing the HPO power watchdog to trip ~6 seconds later (see 1st attachment).  Looking at the laser power for the 35W FE shows the MOPA and the NPRO lose power at the same time, thereby causing the watchdog trip (2nd attachment; the little bump in H1:PSL-OSC_PD_AMP_DC_OUTPUT is simply residual light from the still operating HPO hitting the PD, it goes away once the HPO power watchdog trips).  It is unclear what caused the FE to lose output power in this way.  Putting all the signals from the 2nd attachment on one plot and zooming in, it appears the watchdog tripped before the FE lost output power (3rd attachment).  Possibly a glitch in the PSL Beckhoff software?

The PSL was restarted without issue.

Squeezer coax cable installation

Nutsinee, Terry, Daniel

We installed all the in-air coax cables for the squeezer that were on hand. Six weeks after the promised delivery date, it still seems ~20% are missing due to being back ordered...

Fil et al. started laying down the long D-sub cables.

HAM2 and HAM3 closeout status

Today, in HAM2/3 we:

• Painted FC on MC1, MC2, MC3, PRM, PR2, and PR3
• Pulled FC on PRM, PR2, and PR3
• Unlocked PRM, PR2, PR3
• TBC

BSC 1 North Door

While installing the north door on BSC 1 this morning, as we were inspecting the flange,  I saw what appeared to be a rust spot on the flange in the sealing area of the inner O-ring at roughly the 5 o clock position (facing the BSC from inside the biergarten). 

I wiped, rubbed, and wiped again. I repeated this several times.  There was no drag felt on the alcohol wipe, or on my clean gloves, no "rust residue" or any other foreign material came off on the wipe or glove. 

I put a piece of tape on the bolt head located near this position and showed Kyle the picture. Kyle said he would make a comment on this post.

The area of interest bridges the inner O-ring contact patch and might result in an inner O-ring leak.  We began pumping the annulus space soon after this door was installed and will be able to get useful information via the pressure data as we will have atmosphere on the VE side of this inner O-ring for the next several days.  Though undesirable, an inner O-ring leak wouldn't require immediate attention as would a leak across an outer O-ring.  The main issue resulting from an inner O-ring leak is the scenario where relatively large amounts of air get introduced into the annulus volume as part of an annulus ion pump replacement.  We have work-arounds for this case and, as a matter of policy, always assume that an inner O-ring leak exists when replacing ion pumps or otherwise have a need to introduce air into an annulus space.  

John,
I have the stones

4.4 x 10-6 Torr this morning at the cart backing the locally mounted turbo ("hung" turbo).  This is excellent for the vent/pump duration.

Ops Day Shift Summary

h1asc model make-install, will restart model tomorrow

WP7230

Sheila, Dave:

I have modified h1asc.mdl to reassign the ADC inputs. This model was compiled and installed, and is waiting for a restart tomorrow morning.

WP7230 also covers fixing a PSL bulls-eye channel issue. We believe this was resolved in hardware. Sheila will check if a software change is still needed. If it is, she will add it to her model changes (WP7227).

 

IMs Clear of Rubbing -- OK for Doors When Ready

J. Kissel

I've taken standard transfer functions of the IMs as we begin close-out of the HAM2 chamber. 
All IMs -- IM1, IM2, IM3, and IM4 -- are free of rubbing, and their dynamics look virtually identical to just-prior-to-vent measurements.
We've still got plenty of other things to do, but the IMs are ready for HAM2 door closure.

Huzzah!

Attached are the results and comparisons.

Raw data templates live in 
/ligo/svncommon/SusSVN/sus/trunk/HAUX/H1/
    IM1/SAGM1/Data/2017-12-20_2122_H1SUSIM1_M1_WhiteNoise_L_0p01to50Hz.xml
    IM1/SAGM1/Data/2017-12-20_2122_H1SUSIM1_M1_WhiteNoise_P_0p01to50Hz.xml
    IM1/SAGM1/Data/2017-12-20_2122_H1SUSIM1_M1_WhiteNoise_Y_0p01to50Hz.xml

    IM2/SAGM1/Data/2017-12-20_2124_H1SUSIM2_M1_WhiteNoise_L_0p01to50Hz.xml
    IM2/SAGM1/Data/2017-12-20_2124_H1SUSIM2_M1_WhiteNoise_P_0p01to50Hz.xml
    IM2/SAGM1/Data/2017-12-20_2124_H1SUSIM2_M1_WhiteNoise_Y_0p01to50Hz.xml

    IM3/SAGM1/Data/2017-12-20_2140_H1SUSIM3_M1_WhiteNoise_L_0p01to50Hz.xml
    IM3/SAGM1/Data/2017-12-20_2140_H1SUSIM3_M1_WhiteNoise_P_0p01to50Hz.xml
    IM3/SAGM1/Data/2017-12-20_2140_H1SUSIM3_M1_WhiteNoise_Y_0p01to50Hz.xml

    IM4/SAGM1/Data/2017-12-20_2146_H1SUSIM4_M1_WhiteNoise_L_0p01to50Hz.xml
    IM4/SAGM1/Data/2017-12-20_2146_H1SUSIM4_M1_WhiteNoise_P_0p01to50Hz.xml
    IM4/SAGM1/Data/2017-12-20_2146_H1SUSIM4_M1_WhiteNoise_Y_0p01to50Hz.xml

ISS 2nd loop array alignment

R. Savage, J. Oberling, J. Driggers, S. Dwyer, P. King (via phone), R. McCarthy, F. Clara

This morning we completed the initial in-air alignment of the ISS 2nd loop PD array.  Recall from yesterday afternoon, we did not make much progress in the alignment.  This was due to the signals out of the PD array dropping for apparently no reason.  After some intial investigation it was decided to re-convene this morning, this time with Rick helping out (as he has worked with the ISS PD arrays extensively).

Some notes from this morning's work:

• The ISS 2nd loop transimpedence amplifier chassis was OFF, thereby preventing us from seeing any signals from the 8 PDs in the array.
  • After talking with Peter, it is unclear if this chassis was ever turned back ON during the vent.
  • This does not explain the issues we were having seeing the beam yesterday afternoon, as the QPD in the ISS 2nd loop array enclosure is read out via a different chassis (per P. King).
  • This chassis, which was a prototype installed before O2, was swapped with a tested and approved production version.  The swap was performed by R. McCarthy and F. Clara.
• We discovered that to really see the beam entering the array, we needed to re-install the temporary HWP after IM1 that Sheila removed yesterday afternoon.
  • We talked to both Ed and Cheryl and were told that this was necessary due to a temporary HWP placed at the base of the PSL periscope.  This HWP was placed in order the purposefully send the wrong polarization into the IMC, I believe to make the IMC flashes easier to see (I may be mis-remembering what I heard, if this is incorrect please correct me in the comments).  This temporary HWP in the PSL enclosure will be removed.
  • This does explain the issues we were having in seeing the beam yesterday afternoon.  As soon as the temp HWP was re-installed in front of IM1 the beam was once again obviously visible on an IR card and the ISS 2nd loop signals were greatly elevated.
• To increase the rate of IMC flashing, we injected a small sine wave (via awggui; f=0.1 Hz, amplitude=1, gain=30) into the L path of MC2.  This made seeing the flashes on an IR card and via the PD array signals much easier.

This taken care of, the beam was then centered on the PD array enclosure input aperture using the first steering mirror (located on the -Y side of the table), and then the signals maximized (to the best we could achieve with a flashing IMC) using the second steering mirror (located on the +Y side of the table).  This was iterated several times until we had a strong signal on all of the PDs and the array enclosure QPD, and the beam was centered on the entrance aperture of the array enclosure.  This is more than sufficient for the in-air alignment; the PD signals will be maximized when the IMC is once again locked (after the corner station is pumped down).

h1susey pci-e cards moved, continuing glitch-blip mitigation

WP7252:

During today's CDS meeting Rolf reported that moving the fast front-end computer's IRIG-B card away from slot-1 resolved timing issues on his test system. To try this out on a production system, this afternoon I moved h1susey's IRIG-B card.

Recall (alog 39361) that on 09nov2017 I had moved h1susey's OneStop from slot-2 to slot-5 as part of this investigation. Today I returned the OneStop back to slot-2, and moved the IRIG-B card from slot-1 to the newly vacated slot-5.

slot id	7	6	5	4	3	2	1
before	empty	empty	OneStop	5565	Dolphin	empty	IRIG-B
after	empty	empty	IRIG-B	5565	Dolphin	OneStop	empty

It was a messy restart of h1susey. After taking it out of the Dolphin fabric and issuing the poweroff command, the OS got stuck trying to umount NFS after apparently disabling its network ports. I had to power down using the front panel button, knowing that this would most probably cause a Dolphin glitch. Sure enough, on restoration of power the EY dolphin fabric glitched, requiring a restart of models on h1seiey and h1iscey. EY is now back in operation.

Tagging DetChar, so it's on their record books that this change was made.

The upcoming front-end OS upgrade (Gentoo with 3.0.8 kernel) has a much cleaner startup/shutdown process that works with the OS instead of fighting with it.  This should help those issues.

BSC 1 2 3 Chamber closeout work today

Today, Travis and I worked to closeout BSC 3 to finish out the set of chambers needing doors put on.  We:

• Pulled FC from CPX-AR, which needed a second paint job since there was some small bits left in the center of the optic from the first pull.  The second paint-and-dry was successful.  The usual N2ionizing blow was used
• Painted and pulled FC on the ITMX-HR which came off clean.  The surface looked good considering we were inspecting with laser goggles on (same for the CP above).
• Ran a few TFs - discovered major rubbing somewhere.  A few hours of troubleshooting later and we finally found the ring heater structure cable pushing on the main chain UIM.  That alleviated and the TFs all look great.
• For whatever reason, the ITMX needs ~+300 bias to get back on the oplev.  Not sure what that means, but at least we're still on the slider.
• Jim unlocked the ISI - his TFs look good.
• Added witness wafers and 1" optics where they should be: 1 vert on QUAD, 1 horiz under QUAD, 1 horiz under ACB in tube.
• WIped floor and surfaces many times throughout the day.
• Finished full round of high res TFs with Travis and Kissel - will post in coming days.
• Shopped around for all subsystem signoffs to put doors on.  Waiting for Kissel, but all looks good.

 

I attach a screenshot of the pitch transfer function during the Ring Heater cable rubbing. I also include illustrative drawings and pictures of the rubbing site. 

In words -- the ring heater cables (CBL7 and CBL 8 of D1001517), exiting the QUAD lower structure, should be sandwiched between the two inner face plates (aka "figure eights", D060462). On this ITM, it is not, and for the first time, it escaped far enough out of its designated path that it started to lightly interfere with the bottom of the main chain's UIM base plate (D060376).

Attachments:
First -- (from D0901346) A side view of the full solidworks assembly of the quad. The face plates are the long vertical blue things that extend the length of the lower structure. The two in the middle are the inner face plates.

Second -- (from D1001838)An isolated view of the test-mass-side's inner face plate, with the fully assembled ring heater and cabling system.

Third -- (from G1100850) Another view of the test-mass-side's inner face plate, with the fully assembled ring heater and cabling system.

Fourth -- (from D0901346) A screenshot of the QUAD's solidworks eDrawing assembly, where I've hidden enough to show the ring heater cables, and made the face plates transparent. Naturally, the eDrawing is buggy, and the ring heater cables are displaced from where they're designed to be, and are actually interfering with the *reaction* chain's UIM base plate, but you get the point.

Fifth -- a picture Betsy / Travis took of the offending cable rubbing against the *main* chain UIM base plate.

Sixth -- (from D1001517) A screenshot of just the cable assembly to show what "CBL7" and "CBL8" are.

Seventh -- A screenshot of the Pitch transfer function. Blue and Black are nominal, normal transfer functions. Red was while this rubbing was happening. As expected from the modes shapes of the middle-frequency pitch modes nominally at 1.33 Hz, 1.59 Hz, and 1.98 Hz, that had so clearly shifted, we knew that something was (wrong / rubbing) (in / around) the middle masses -- the PUM / UIM.