HEPI transfer functions
The first set of transfer functions was measured on BSC8 from 2mHz to 500Hz. The 700mHz-10Hz frequency band was measured in 2 configurations:
- ISI undamped (2012_02_15)
- ISI damped (2012_02_16)

The HEPI transfer functions can be seen at:

https://svn.ligo.caltech.edu/svn/seismic/HEPI/H2/ITMY/Data/Figures/Transfer_Functions/Measurements/Undamped/

-          LHO_HEPI_BSC8_TF_L2L_Raw_from_ACT_to_IPS_2012_02_16.fig

-          LHO_HEPI_BSC8_TF_L2L_Raw_from_ACT_to_IPS_2012_02_16.pdf

-          LHO_HEPI_BSC8_TF_L2L_Raw_from_ACT_to_L4C_2012_02_16.fig

-          LHO_HEPI_BSC8_TF_L2L_Raw_from_ACT_to_L4C_2012_02_16.pdf

The ISI damping loops are effective between 1Hz and 8Hz. The effect on HEPI transfer functions is limited but visible on the 1.25Hz and 5 Hz resonances. A comparison between the two configuration can be found at:

https://svn.ligo.caltech.edu/svn/seismic/HEPI/H2/ITMY/Data/Figures/Transfer_Functions/Comparisons/L2L/

-          LHO_ISI_BSC8_Comparison_TF_L2L_ACT_H_to_L4C_H_20120215_vs_20120216.fig

-          LHO_ISI_BSC8_Comparison_TF_L2L_ACT_H_to_L4C_H_20120215_vs_20120216.pdf

-          LHO_ISI_BSC8_Comparison_TF_L2L_ACT_V_to_L4C_V_20120215_vs_20120216.fig

-          LHO_ISI_BSC8_Comparison_TF_L2L_ACT_V_to_L4C_V_20120215_vs_20120216.pdf

If you look carefully, you can see light bumps at the junctions between the different sections (10mHz for instance) of the transfer functions. To evaluate the non linearity, I drove a sine wave at 200mHz and 4Hz using different amplitudes (100cts, 300cts, 1000cts, 3000cts, 5000cts) (attachment).

Low gain on the HEPI pier interface D080521
It is quite surprising to see that the HEPI L4C signals don’t overshoot +/-100 counts on the ADC (attachment 1) when the actuators are not driven.  After checking that all elements of the readout chain were turned on, we measured the L4C gain of one HEPI pier interface D080521. The gain is 22 and a 0.1V input offset gives 3600counts on the ADC as expected.
After comparing calibrated spectra of the HEPI L4Cs (on top of the piers) with spectra of the ISI L4Cs (stage 1), the HEPI L4Cs spectra don’t look wrong.
L4C gain in D080521 seems low.

The default NDS server assignment that is made for new shells has been changed to use h2nds0 instead of h2nds1.  This will affect any program or shell started after 10:38 AM PST on Feb. 17, 2012.

Attached are plots of dust counts > .5 microns for February 15 and 16.

Kyle, Gerardo, John

Had valved-out QDP80 pump from YBM before leaving last night.  YBM pressure was 20 torr at that time.  Pressure rose to 200 torr in ~18 hours, i.e. large unexplained gas load in YBM volume -> Resumed pumping YBM with QDP80 this morning.  Pumped adjacent annulus volumes (BSC8, GV1 and GV3-GV5 already under vacuum).  Connected leak detector in parallel with QDP80 pump path and set detector up for high pressure leak testing (10 torr max inlet pressure).  Once YBM pressure reached 10 torr, Gerardo and I did a "rehearsal" exercise by spraying some feed-throughs on BSC8 -> Noticed a large delayed response after tens of seconds.  Time constant believed to be around 1800 seconds for these conditions.  Signal continued to rise for ~30 minutes after initial (10 sec/flange dwells) until signal "railed" the leak detector at 9.9 x 10-4 torr*L/sec. -> Surprisingly, Gerardo was unable to find this "wopper" leak by utilizing the stethoscope!  We probably could hear this leak with our unaided ear if there wasn't so much background noise in the LVEA. -> For tonight, I am leaving the YBM vent/purge valve "cracked" open a little so as to keep a viscous stream flowing in the correct direction in the event of a QDP80 stoppage during the night.  This way we won't have to valve-out the pump overnight and we can resume leak testing in the morning without having to wait "hours" to get back to 10 torr.

PSL is shut off for the night to allow Robert S. to make accelerometer measurements on the periscope, without interference from the chilled water lines. PMC, ISS, FSS servos are off, everything will be started back up tomorrow morning.

HEPA units are off, make up air is set to 30%, creating a ~0.01" water pressure differential between laser and anteroom. AC units were shut off via touchpad interface, but they may still be on.

Today we installed the fall protection wire and temporary vertical protection bars for the TMS isc table, and removed the Genie, while the cube is still on EQ stops. It went fine.

Originally the temporary protection bars weren't required at this stage, however the new initial alignment procedure the TMS and the intial alignment people have been working on requires that a large target is mounted on the input aperture of the telescope plate. Since this totally changes the balance of the TMS/tele, we needed temporary EQ stops on the vertical protection bars.

Replacement of FMCS UNC at mid Y.

Jim, Hugo

Issue: 
LLO have a BSC-ISI that is under-loaded by 400lbs. They noticed that its small blades where lifted from their clamp-plates. They measured this gap by inserting shims at the blade/clamp-plate junction. A 1.5 mils-thick shim could be inserted by approximately 18/64” in the whole length of that junction.

Test:
We did the same test here at LHO on BSC6, which is reasonably loaded, in order to help LLO diagnose. The access was limited. Only two blades were tested. They could only be tested along 2” of the junction’s length.

Result:
Observations are the same for both of the blades that were tested. A 1 mil-thick shim could be inserted by approximately 28/64”, and a 2 mils thick shim could be inserted by approximately 19/64”.

Conclusion: 
The gaps measured on LHO-BSC6 are similar to the ones measured on the unit studied at LLO. Hence, the space measured between the blades and their clamp-plates doesn’t seem to be a justification for a BSC-ISI to be under-loaded. 

Suggestion:
Further discussion revealed that our bolts were lubricated before torqueing. Lubrication allows reaching a higher level of compression before tripping the torque wrench. It also revealed that our bolts were torqued to a higher torque value. 

H1 front end racks were installed into the MSR and populated with computers. The first two of the H2 racks were installed and have front ends installed except those currently running the H2OAT.

I remotely logged in from LLO and modified the H2SUSITMY.txt file located in /opt/rtcds/lho/h2/chans/ directory, which contains the ITMY suspension control filters.

I removed 4 AntiAcqL1 filters which shouldn't have been there and which may have been installed awhile back incorrectly (possibly by the prepare.m script I've since modified, possibly by someone by hand).

Testing with the current suspension prepare.m script in /ligo/svncommon/SusSVN/sus/trunk/Common/MatlabTools does not generate this filter, so future use to prepare filters for other QUAD optics should not have this problem.

I copied the new H2SUSITMY.txt to /opt/rtcds/lho/h2/userapps/release/sus/h2/filterfiles and checked it into the svn.

I've take new, pre-pump-down H2SUSFMY data, and have identified that the suspended elements are inhibited by some sort of rubbing. Of course, it's too late fix anything now -- at least for the foreseeable future, and we don't need H2 SUS FMY for the single arm. But, it's sad.

See attached comparison between
H2SUSFMY (111114) -- Phase 2C approved Measurement: pre-cartridge install, metal mass
H2SUSFMY (2012-01-09) -- Un-posted Phase 3A measurement: post-cartridge install, taken to be "final," but before final setting of EQ stops (see below). Taken to be measurement that approved the SUS.
H2SUSFMY (2012-02-15) -- Today's measurement, just before pump-down of BSC8.

We initially approved this suspension, post cartridge install with measurements on Jan 9th (no original aLOG posted, but shown in the attached plots).

We neglected to confirm that the dynamics were OK after EQ stops and BOSEMs were set to the "final" values for both BSC8 Suspensions on Feb 01, assuming that moving the EQ stops closer (but obviously attempting not to touch the suspended elements) would not effect the dynamics.

Note, I'm so confident it's EQ stop rubbing, because we had recently decreed (see M1100256) that we need to get the stops 0.75mm +/- 0.25mm away from the suspended masses. The technique for determining this distance is 

(1) Run in all stops until "just touching" the mass
(2) Given the knowledge of the stop's bolt thread, turn out the bolt the number of turns necessary to create a 0.75mm gap. E.g. a 1/4-20 screw (typical EQ stop size) has 20 threads per inch, or 0.05 inches / thread, or 1.27mm / thread. In practice one can get ~1/4-1/2 turn accuracy, or +/- ~0.0.31mm at best.

However, there are several factors further against you while setting these stops: 
- "just touching" the mass, because running stops in is typically done sequentially is difficult to get: one can believe that running the first stop into "just touching" would move the suspended mass by ~0.5mm or more. 
- For a good fraction of the EQ stops, the point of contact is invisible rendering the check of the gap, one they're believed to be backed off at best difficult (say by using a mirror system), or at worst impossible.

Perhaps we need to revisit this decision/requirement again...

No production GS-13s on this unit but it does have good parts.  No issues getting it in the Storage Container.

See Jim worry it down into position.

We attached the tele/table to the cube under the test end, though the tele/table is still supported by Genie.

Tomorrow we'll fully suspend it.

Kyle, iGerardo