On Wednesday we unlocked H2 SUS ETMY for the first time after cartridge install, and last night we took a set of transfer functions to diagnose its health. Attached are the results; they look excellent! 

Among the other usual attachments,
2012-04-12_H2SUSETMY_M0_ALL_TFs.pdf -- Main Chain, individual comparison with Model, shows cross-coupling and OSEM basis response
2012-04-13_H2SUSETMY_R0_ALL_TFs.pdf -- Reaction Chain, individual comparison with Model, shows cross-coupling and OSEM basis response
allquads_120413_H2SUSETMY_ALLM0_TFs.pdf -- Main chain, zoomed out version of the below described plots
allquads_120413_H2SUSETMY_ALLR0_TFs.pdf -- Reaction chain, zoomed out version of the below described plots

In the main chain (M0) chain's allquads_120413_ALLM0_*.pdf, I compare 4 measurements, all monolithic lower halves,
2012-01-10 H2 SUS ITMY, Phase 2b approved measurement (On test stand, mounted to BSC8-ISI, Chamber-side Test in LVEA)
2012-02-16 H2 SUS ITMY, Phase 3a approved measurement (In-chamber, partially at vacuum)
2012-03-14 H2 SUS ETMY, Phase 2b approved measurement (On test stand, mounted to BSC6-ISI, Chamber-side Test at End Station)
2012-04-12 H2 SUS ETMY Current measurement (In-chamber, mounted to BSC6-ISI, glass ERM, vibration absorbers OFF)

and in the reaction (R0) chain's allquads_120413_ALLR0_*.pdf, I compare
2011-11-22 X1 SUS QUAD04 Phase 1b approved measurement (On test stand, mounted to Solid Stack, fully laced, metal dummy mass)
2012-01-24 X1 SUS QUAD03 Phase 1b approved measurement (On test stand, mounted to Solid Stack, fully laced, metal dummy mass)
2012-03-15 H2 SUS ETMY, Phase 2b approved measurement (On test stand, mounted to BSC6-ISI, Chamber-side Test at End Station, glass ERM)
2012-04-13 H2 SUS ETMY Current measurement (In-chamber, mounted to BSC6-ISI, glass ERM, vibration absorbers OFF)

At first glance, the TFs look great. 
- No major rubbing, the dynamics, overall, appear, as-usual, strikingly similar to suspensions of similar type. 
- On the reaction chain, Pitch looks fine, implying the lacing cables haven't moved or at least changed the dynamics for the worse. 
- The main chain looks just down right great and consistent. A couple of the degrees of freedom even cleaned up since Phase 2b (see L, T, and R). Nice!

Upon further inspection, I see a couple of things that raise some yellowy green flags on the reaction chain:
    - The second R0 longitudinal mode @ 0.85 is bifurcated. What you can't see is that the magenta curve (the previous ETMY measurement), is also bifurcated, but the lower frequency of the two split modes is lower in magnitude in the magneta curve, and is therefore hidden behind the cyan. Because this mode is *not* bifurcated in the two metal mass measurements, I'm curious if this is something peculiar (I dare say "wrong"), with ETMY. The problem is that we have statistics of one on a glass ERM. All metal mass measurements don't show this feature. This good very well be just how the plant changes between a metal and glass mass.
    - Similarly with the second R0 transverse mode. Though the metal mass measurements show two modes around that frequency (one much more pronounced than the other) the two measurements of the one glass ERM show that this mode shape has changed toward shall we say "more complicated." Of course, transverse is always a noisy measurement since only the SD OSEM is driving it (as opposed to the other DOFs which have at least 2).

Finally, zooming out and looking at, for example allquads_120413_H2SUSETMY_ALLM0_TFs.pdf, I notice that our favorite, high-frequency "what looks like a giant mechanical zero, but we've proven to be just electronics cross coupling some where in the chain, which we've convinced ourselves was the temporary ribbon cables we use in our mock-feedthroughs" feature is STILL PRESENT after the install into chamber, where we're using nothing but production cables and real-deal feed-throughs. We convinced ourselves with FMY that this was true, but perhaps, we were too hasty with our conclusions. Some nice additional proof will be a comparison between vibration absorbers OFF vs. vibration absorbers ON, which I will get tonight.

GregG, JimW

Following SUS's adding vibration absorbers, SEI went in and rebalanced. This included re-adjusting the gaps on 3 of the CPS's as fallout from yesterdays replacement of allegedly defective sensors. We also did some refining of the cable routing on corner 3, as it is a very busy area (as shown in the attached photo) and was showing evidence of rubbing from TF's run earlier in the week. BSC6 should now be in pretty good shape for testing over the weekend.

BSC8--Saw activity around this chamber, so I'm assuming they were able to remove the Arm off the door flange.

BSC6--Vibration Absorbers have been installed

There were tons of Diode Room Access Alarms throughout the day.

* Vibration Absorbers were installed on ETMy structure
* I started pinning up the cables to Stage0 with the PEEK finger brackets DXX00600
* We set the EQ stops to ~1mm from the masses

The SUS is fully suspended and shields are removed.

On Wednesday the code at end X was moved to new code running on h0epics2, a Ubuntu workstation. Yesterday the weather stations at end Y, mid X and mid Y were ported as well. The weather station at the corner station is currently removed.

Find photos of yesterday's installation at BSC8 in ResourceSpace. 

We restarted the internal EPICS Gateway several times yesterday in order to extend the max data size per PV to allow Rolf's new waveform record data access to the control room. The new SUS ini file problem was resolved to double-quotes in comment lines and usage of a non-supported 128Hz acquisition rate. 
The system reservation utility is being upgraded and made ready for export to LLO.

J. Garcia, J. Kissel

Turns out the BSC6 ISI is locked, so we have a perfect opportunity to sneak in a measurement while Vincent isn't looking. Wahoo!
 
We started this measurement at 
1018325503 (Apr 13 2012 04:11:28 UTC; 21:11:28 PDT)
on cdsws1 ( but logged into from the operator station [operator0?] )

We expect it should end in ~7 hours, i.e. roughly around 04:00 PDT. 

Regardless of it's finishing status (or the results from it) -- we are still a go for installing vibration absorbers tomorrow first thing. 

While Betsy and crew were re-installing the lower half of ITMY into BSC8, she was kind enough to shine a flashlight down the beam tube so I could look for signals from the four scattered light PDs associated with the arm cavity baffle, H2 hole.  I clearly saw her flashlight signature as the flashlight was shone on the inner walls of the beam tube.

A first order 70 Hz low pass filter was installed on each of the inputs to the 4 Melles Griot transimpedance amplifiers to cut down on the large high frequency noise.  The filter puts 1k in series with each transimpedance amplifier input (virtual ground node).  This should be fine at the very low light levels associated with the scattered light PDs, but would introduce nonlinearity at photocurrents much higher than 100uA or so.

Each Melles Griot amplifier was left in the 2uA peak transimpedance setting, which corresponds to a transimpedance of 10^6 V/A.  There is also a gain of 10 in the PEM BNC input channels.  The dark noise of each of the four channels as viewed on the data viewer is about 100 counts peak to peak, with small (100s of counts) DC offsets.  The flashlight test caused excursions in the viewed signals of order 1-2 thousand counts.

The PEM channels for each of H2 ITMY scattered light PDs were moved to:
PD1 - software channel 27 (written as "28" on the front of the PEM DAQ chassis)
PD2 - software channel 28 (written as "29" on the front of the PEM DAQ chassis)
PD3 - software channel 29 (written as "30" on the front of the PEM DAQ chassis)
PD4 - software channel 30 (written as "31" on the front of the PEM DAQ chassis)

The last channel of the PEM chassis is dedicated to the duotone input, so it's not

The H2 PSL was shut off today, we will begin moving it to H1 starting the week of the 23rd.

He had a good life.

We started the day with the need to swap out a Probe to match up the electronics board.  We had some intermittent grounding issues and finally got a good ground and things looked better.  I need Vincent to bless the spectra now but I think we are in good shape. We do need to possibly more formalize how the CPS satellite racks/boards are tied to the ground to prevent this problem in the future.

JimW RichA Hugh

Using the BSC repair arm, we reinstalled the ITMy lower QUAD (with wire hung ITMy optic). The arm is now back outside the chamber envelop and ready for removal from the door.  We still need to finish suspension and cabling before we are ready for anyone to look at it.  (But, we need to do work at BSC6 tomorrow too, so will have to interlace activities.)

After a bit of planning, a work permit was generated for the H1 BSC de-install. The tentative plan is to enter BSC3 through the east door and remove optics from BSC3-2-1 and then strip the guts from BSC3 so that we can chamber clean it. The work permit includes staging cleanrooms, etc. and door/dome removal.

Wipe down was completed and the dust barriers were removed. Second vacuum was completed. Tomorrow, we'll take care of close-out and return doors to the chamber. (I hope!)

V. Lhuillier, J. Garcia, J. Kissel

After confirming that all SUS in BSC6 were unlocked, functional, and damped, Vincent fired off an over night transfer function on BSC6 ISI. It has finished (but needless to say has not been processed).


All,

Measurements on LHO_ISI_BSC6:
Started at GPS:1018232359 - 04/11/12 19:19:04
and finished at GPS:1018266991 - 04/12/12 04:56:16

The seismic team

D. Barker, R. Bork, J. Garcia, J. Kissel

We have added two things to the QUAD_MASTER Model:
(1) Added two "TEST" filters which hook up spare ADC channels to Spare DAC channels (i.e. none are hooked up to any real electronics -- so no danger of positive feedback), and
(2) A DACKILL Part, to which the M0R0 watchdog is hooked up as the input flag.

The purpose of (1) was to have an independent way of confirming both the already-installed IOP DACKILL, and now the newly installed (2) usermodel DACKILL. For now, we've naively hooked up the M0R0 watchdog stage to it, so the our USER DACKILL merely performs the original intent of the watchdog and prevents all output of the model when the watchdog trips.

HOWEVER, this is too severe of a watchdog tripping, and currently (because we've yet to create an MEDM screen for it) is just a pain in the butt to reset. Eventually, we intend to hook this up to things more similar to the ISI's DACKILL, looking for "rogue" excitations such as coil driver monitor showing signal when "we haven't requested it," master switch being off, etc.

More details / development to come.


Details:
--------------
Since I haven't recompiled the a model recently, I provide a detailed log of what was done in order to actually install and ensure functionality of our changes to the model.

Made above mentioned changes to 
${userapps}/sus/common/models/QUAD_MASTER.mdl
the new test filters added two extra inputs and outputs to the QUAD library part.

THEREFORE made further changes to 
${userapps}/sus/h2/models/h2susetmy.mdl
connecting the TEST filters to ADC0_29 and ADC0_30, and DAC3_7 and DAC3_8. 
NOTE: This connection is unique to the ETMY top level model ... if we intend to keep these test filters in, we'll have to find unique spare channel locations for each chamber's QUAD.


Finally, with changes saved, we ran through the build process:
ssh h2build

cd /opt/rtcds/lho/h2/rtbuild-trunk/

make h2susetmy
(compiled successfully the first time -- we rule!)

make install-h2susetmy

ssh h2susb6

starth2susetmy
.........
Because we added two new filter banks to the model, this affects the .ini framebuilder file., so we restarted restarted framebuilder (the "starth2susetmy" step "automatically does" the DAQ RELOAD button mashing that is normally required when we changed the .ini file without restarting the model.)

telnet h2dc0 8087
shutdown

.........
Saw second light of DACs 1 2 (not 3?!) 5 red on both the H2SUSETMY and H2SUSBSC8 GDS_TP screens.
Learned that that's (another/the) way that you know the DACKILL is working ... it tells you that the DAC card is "no longer in use." (Yep. THAT's what the second bit means.)

SO:
reset ETMY M0R0 watchdogs (because they come up tripped, and that's currently what trips our new USER DACKILL)

reset DACKILL (because we don't have a screen yet, we had to do it by hand...)
    caget H2:SUS-ETMY_DACKILL_STATE
    0 (is BAD -- shows DACKILL is tripped)
    (saw zero)
    caput H2LSUS-ETMY_DACKILL_RESET 1 (triggers "the reset button")
    caget H2:SUS-ETMY_DACKILL_STATE
    (saw 1)

..........
Confirmed ability to drive by taking a quick low-resolution Longitudinal transfer function.

Mechanical brushing was completed using a total of three drills (1 non-starter, 1 short-timer, and 1 marathon-man). Hand detailing with the stainless steel toothbrushes was completed around the fins and under the support tubes.First vacuum was completed including the removal of bellows protection. Isopropanol was decanted in preparation for wipe-down. 

Attached are plots of dust counts > .5 microns.

5 horizontal GS13s and 4 vertical GS13s were received from LLO in shipment load # 3523. Huddle tests were performed.

Horizontal GS13s were initially set with the same orientation: as installed in the ISI. Two of them were then failing. Pods can be installed with 3 orientaions. Each orientation set up is distant of 120˚. We tried rotating the pods by increments of 120˚ to see if the test would be more successful then. GS13s were gently tilted back and forth after each rotation to make sure that the mass was not stuck. For some orientation the mass still felt stuck. The instrument's power spectrum would then reveal a failure of the instrument's response below 20Hz. 

The rotation test was performed two times: 04/10 and 04/10. The working set up found for both times correlated:

• Horizontal Pod# 37 - +240˚ CCW
• Horizontal Pod# 20 - Regular (as in ISI)
• Horizontal Pod #84 - Regular (as in ISI)
• Horizontal Pod #27 - +120˚ CCW 
• (No working orientation was found for Horizontal Pod #18)

However, when tried later this afternoon, this configuration would feature a failure of Pod #20, as reported by Corey.

The combination [leveling of the instruments in their pod + approximate level of the table used for test] could be an explanation for the need to rotate the instruments. Another explanation could be that loose flexures, as some were observed, would impact the symetry of their pull, causing the mass to get stuck in some positions and not in others.

As usual, Vertical GS13s appeared to be robust. Their spectra are all fine.

Test pictures and results are attached.