Ops Day Shift Summary

Fast shutter and PZT2 shutter woes

Sheila, Keita, TVo

Last night I ran into a funny situation while trying to scan the OMC with the IFO beam in single bounce off ITMX.  In order to do this scan, we need to have the fast shutter down and the PZT2 shutter down and the outputs enabled so that you can actually do the sweep.  Effectively, there are two shutters we want to have open FASTSHUTTER_A and PZTSHUTTER_A.

However,  when we tried the logic here to open the fast shutter by setting the threshold to negative (H1:SYS-MOTION_C_SHUTTER_G_THRESHOLD), this would shut off the PZT2 output. Then setting the same threshold to zero or positive, the PZT2 output would be enabled and the fast shutter would close.  SO, there must be some logic that I am missing.  Keita suggested  power cycling the Fast Shutter Driver which Sheila noted having some funky readings (pictures attached).  However, this doesn't seem to have fixed the problem.

Double-beam hunting

Robert, Sheila, Nutsinee

Today we removed a view port cover on HAM5 north door (faraday side) to look for the IFO beam reflected off AR side of the thin film polarizer. Attached a couple photos we took with our IR camera (Nikon D7100 with IR filter removed)-- We found the beam hitting the HAM6 side of the ZM2 cage. The beam moved towards HAM4 when we did positive yaw on SR2 and moved up when we did negative pitch. This is consistent with how the rejected polarization beam moved in HAM6.

According to Koji's calculation and Corey Austin's response to our alog last night (alog41403) we suspect that our wedge is in the opposite direction.

mixing tabs and spaces for indentation not just a bad idea for python code

A word of caution that indenting code using a mixture of spaces and tabs will result in code which looks very different depending upon which editor is being used to view it. In the two attachments a snippet of the same C code is shown being viewed using the editors vi and gedit. The vi example shows consistent indentation, the gedit version shows erroneous indentation. This is because the code mixes tabs and spaces for indentation, and vi defaults to a tab size of 4 spaces, and gedit 8 spaces. This can be further compounded in that the code itself can redefine the tab size with the directive "-*- tab-width: 6;" which is honored by some editors (e.g. gedit) and not others.

Take home message: it is best to adopt Python's PEP8 standard of always using 4 spaces for indentation for all LIGO code to avoid this confusion.

WHAM6 ISI Balanced, yes, again. Pulled Cables 4 & 5 for ASC_OMCR_(QPD A/B & PICOMOTORS)

Refer to D1300122, sheet 8.  Pulled cables 4 & 5 completely from chamber along with their Cable Table Bracket (CB-5).  Cables disconnected from D6-F4 & F6.

With that out of the way and extraneous spode jettisoned, the ISI was unlocked and balanced to the locked Z & tilts.  Got within several um/urads, relocked, updated the target position, and unlocked.  Balancing then required just a 50g adjustment and locking/unlocking is again just a few micros.

Left the ISI locked

Cables removed:

D1000225-v2 180" S1105030--This is in ICS but was not in an any assembly.

D1000223-v8 216" S1202611--This cable is not in ICS & there is no D1000223 in the WHAM6 ICS Assembly D0901822.

Will stash these cables in the cable totes in the LVEA west bay. And uploaded to ICS.

New EOM ready to install

[Calum, Rich, Koji]

The crystal of the new EOM was cleaned and re-mounted in the unit. The impedance of the unit after the cleaning was checked. After a few tweaks of the variable components in the matching circuit, decent 50Ohm matching for all the ports have been recovered.

---

Crystal Cleaning

- All the jobs were done in the optics lab.
- Made a crystal holding fixture using opto-mechanical components (Attachment 1). Note that a piece of Kapton tape is used to make sure the crystal does not drop.
- Checked the particulates on the optical surfaces with a microscope. The surfaces were named left and right (Attachment 2). There were large number of particulates on the surfaces. (Attachment 3/4)
- Painted FirstContact on the both optical surfaces. We intensionally painted the side of the crystal too so that we can attach a peek tub there for peeling the paint.

- The FC cleaning was pretty effective. It, however, turned out that the assembling of the crystal into the housing cause some dusts atached on the surafaces. This particulates come from the boron nitride used to support the crystal. It is a brittle matrial and emits a lot of dusts.

- So the cleaning was started over again.

- The final cleaning showed excellent dust removal (Attachment 5/6).
- Then the crystal was installed in the housing gently such that the creation of the particulates are minimized.
- The surface observation after the assembly showed sufficiently clean surface, particularly the central area.

Impedance measurement

- The impedance of the EOM + matching circuit was measured with a handheld network analyzer.
- The measurements were done before the disassembly (before the cleaning) and after the post-cleaning assembly. (Attachment 7/8)
  The impedance measurement after the cleaning showed a slight devitation from the nominal tuning. This deviation was tuned by adjusting the trimmers and the inductor windings.
- The final measurement showed sufficient matching (VSWR<1.5) at each frequency.

First Pass at Electric Field Meter Install

Luis, Craig, Georgia, Calum, Travis, Betsy, Rich

We had a go at running the field meter in End-X today, but hit many snags.  The spectra we were able to take will be posted soon.  At time of writing, we believe one wire (the Y-minus of the Y-axis differential output) is not connected somewhere in the air-to-vacuum wiring chain.  We will investigate tomorrow.  The good news is that we were able to get a first look at the X-axis spectrum, but frankly we are suspicious of the results because we don't see the expected power line features.

We installed the QPD chassis that supports the electrometer into the ISC rack at U-height 3.  The new unit is S1102829, the old unit is S1102832.

Leak checked GV-to-OMC flange at IP3 isolation Gate Valve

Vertex volume isolated from YBM and XBM and being pumped with Vertex MTP backed by Inficon LD

Sprayed audible flow of helium for 25 second duration at each of the two 180 degree separated leak test slots and also slowly moved around entire joint perimeter.  LD background 9 x 10-9 torr*L/sec. Noted same joint at IP4 had visible gap -> sprayed this joint as well.  Signal crept up to 1.5 x 10-8 torr*L/sec. Signal increase was slow and not typical of metal joint leaks.  I'll notify Chandra R. of gappy joint.  

HAM6 work today (SQZ)

Sheila, Terry, Daniel, Terra, T Vo, Koji, Nutsinee

 

SQZ beam came back!

This morning we checked that the SQZ beam went into HAM5 and still came back out on the other side to the OMC. The SQZ beam was not perfectly overlapped with the PSL beam but we declared that was good enough.

 

Double-beam

Corey Austin told us we should expect two beams reflected back from the .5 deg wedged thin film polarizer, we only saw one beam came into HAM6. Koji calculated the separation angle to be 2.2 deg. ZM2 diameter is 2", facing the beam at >45 deg left us with about ~1" diameter (2.5 cm). The beam separation at ~1m is ~4cm. It is very likely that the second beam doesn't hit ZM2 as the main beam hits the center of the optic.

 

Beam diverter

We moved the beam diverter such that the transmitted red beam from the VOPO shot out of HAM6 to SQZT6 parallel to the green refl and trans. A new back (right) panel for the SQZT6 has to be made, the beam is now 4 and 1/4 inches from the top of the panel and 9 and 1/4 inches from the -y edge of the panel. Here's a photo of the red trans beam position relative to the view port simulator. We also removed the apertures out of the way. Here's a photo of aperture #1 and #2 that shows their positions before we moved it, which was set after the realignment work in HAM5 last week.

 

IFO beam

The IFO beam that reflected to the squeezer side of HAM6 table measured 7.6 uW and was 6mm above aperture. A black glass beam dump was put behind the beam diverter at roughly 33 deg from the beam diverter's nominal. The sketch shows the angle that the beam came in relative to the beam diverter and how the reflected angle was calculated. The IFO trans beam into OMC measured 1.4mW. We also moved the silicon carbide beam dump to capture the OMC refl beam.

 

H1 SUS ETMY Aligned to Optical Lever using OPTICALIGN Slider Offsets; Uses only 35% of DAC Range

J. Kissel,

Now that
    - Hugh unlocked H1 EY's HEPI, and brought the seismic isolation system to nominal performance (LHO aLOG 41392), and
    - The vacuum team's got BSC9 to ~1e-6 Torr
I wanted to confirm that we could align to the optical lever (our best alignment reference until we get the arms back). 

We can!

This means we can begin on the long list of "Does It Work?" measurements in chamber:
    - Standard suite of spectra and transfer functions of OSEMs (on both QUAD chains and TMSY)
    - Finding the new highest Vertical and Roll modes
    - Measuring their Q to confirm that the new BRDs are functioning as expected (a la LHO aLOG 40098)
    - Turn on high voltage, measure the ESD effective bias voltage

Electric Field Meter Channels Installed in PEM EX Front End Code

D. Barker, P. Fritschel, J. Kissel, D. Sigg

I've modified the 
    /opt/rtcds/userapps/release/pem/h1/models/h1pemex.mdl 
front-end model to include new channels for the production electric field meter (EFM) that's being fit-checked and commissioned now, and will be re-installed permanently later this summer. 

The new channels stored in frames will be: 
    Channel Name                       Stored at      Channel Meaning                                               Units
    H1:PEM-EX_EFM_BSC9_ETMX_X_OUT_DQ   16384 Hz       for the X DOF field, aligned with the IFO's global X          V.m^{-1}
    H1:PEM-EX_EFM_BSC9_ETMX_Y_OUT_DQ   16384 Hz       for the Y DOF field, aligned with the IFO's global Y          V.m^{-1}
where, for now they'll be stored at 16384 Hz while we explore the usefulness of the sensor.

I've also added infrastructure for the band-limited RMS'ing of the signal of up to 5 bands (for now called A, B, C, D, and E), where after looking at the spectra once installed and functional, we'll decide the frequency bands that we'd like to use. Their interesting channels are 
    H1:PEM-EX_EFM_BSC9_ETMX_X_RMSLP_A_OUT16
    H1:PEM-EX_EFM_BSC9_ETMX_X_RMSLP_B_OUT16
    H1:PEM-EX_EFM_BSC9_ETMX_X_RMSLP_C_OUT16
    H1:PEM-EX_EFM_BSC9_ETMX_X_RMSLP_D_OUT16
    H1:PEM-EX_EFM_BSC9_ETMX_X_RMSLP_E_OUT16

    H1:PEM-EX_EFM_BSC9_ETMX_Y_RMSLP_A_OUT16
    H1:PEM-EX_EFM_BSC9_ETMX_Y_RMSLP_B_OUT16
    H1:PEM-EX_EFM_BSC9_ETMX_Y_RMSLP_C_OUT16
    H1:PEM-EX_EFM_BSC9_ETMX_Y_RMSLP_D_OUT16
    H1:PEM-EX_EFM_BSC9_ETMX_Y_RMSLP_E_OUT16

Dave has installed and restarted the model, and restarted the DAQ, so all of the above channels should now be in the frames (LHO aLOG 41390), though there's nothing yet hooked up to it.

MEDM overview screen to come.

Other Details:
    I've wired up the signals to be connected to ADC Card Number 1 (counting from zero), channels 28 and 29 (again, counting from zero). These should correspond to the (dreadfully confusing and out of date) list of ISC channels for the end stations -- Section 4 of T1100472 -- in which these channels are labeled to be on 
    - "ADC 0" (counting from zero, but actually corresponds to physical Card Number 1 in the IO chassis, which has 4 cards) under 
    - "DB9_8" (the 8th DB9 spigot of a standard AA chassis that's hooked up to Card Number 1)
    - channels 29 and 30 (counting from 1)

This is assuming that 
    - the signals get plugged into a "Modified QPD Amplifier" indicated on Rich's new EFM wiring diagram, D1800088, which is a modified ISC Dual QPD Transimpedance Amplifier Chassis (D1002481) in which one half is converted to handle the output of this new field meter, recombined into the DB25 pin output, 
    - sent to an ISC Whitening Chassis (D1002559) where we can gain up the signal before it's finally
    - sent out on a DB9 into Card Number 1's AA chassis, on the 8th port in the channels described above.
but I'll have to confirm all of this with Rich.

    In addition to creating the new EFM block, I've also taken contamination control to heart, performing a bit of "clean as you go" on the h1pemex model. Attached are screenshots of the before & after. The clean-up includes:
    Top Level:
    - Moving the ADCs off to the left, and connecting them to the central EX block via tags instead of direct wires
    - Reordered in the inputs of the EX block, so that the ADCs come into the block in numerical order
    EX Block:
    - Wrapped the ambiguously named ADC_0 channels (which apparently read out the ESD driver's power monitors) into a subsystem, so they take up as much space as the other sensor blocks
    - Removed all of the unnecessary DAQ channel lists that merely contained commented out copies of the DouTone Channel
    - Completed the labeling scheme for all blocks, aligned them to the sensor blocks, and to themselves.

I've created a rudimentary MEDM screen for the field meter that covers all of the new infrastructure. It can be found under the PEM menu at the bottom. See attached screenshot. No infrastructure has been filled in, but we'll get there.

EndY Yend BSC10 EY HEPI Unlocked

At the request of the controls engineer, the Yend HEPI platform was unlocked.  No issues other than the usual terrible access in unlocking.

Ran Cartesian Range-of-Motion tests of 0.5mm for DOFs X Y Z RX RY & RZ.  Probably sufficient to do X Y & Z but forgot about that before getting started.

The attached plot shows the X Y Z Range of Motion with the local sensors over laying one another.  This shows no bad actors of local sensor clipping, slope differences or changes.  HEPI is clear to +- 0.5mm.  Will check larger ranges most importantly for Y (tidal) another time but this will get us by for now.

The HEPI and ISI platforms are Isolated.

IP11+baffle re-install and EY leak check

Kyle, Gerardo, Mark D. Tyler, Chandra

This morning I vented IP11 (again!) with bottled N2 and then Mark and Tyler bolted it to the new IP11 valve. The chevron baffle housing was already attached to bottom of IP11 from leak testing yesterday (baffle housing SN002 is leak free). I leak checked the bottom flange of new IP11 valve and also the new HV NEG housing/valve on BSC6, with background <1e-9 mbar-L/s. No leaks detected. Later Kyle leak tested the flange connecting bottom of baffle to top of IP11 valve. No leaks detected. EY main turbo is valved back in and good to go for in-vacuum high voltage.

We left a hung turbo pumping on IP11 overnight, backed by leak checker.

BSC10 turbo cart is also valved in with AIP ON.

ETMX Fibers broken during installation

Today, Travis, Mark and I worked on installing the ETMX lower structure into the chamber.  After it was in, while reassembling cables, we noticed that the test mass was really jammed into it's lower stops - upon further inspection, we discovered that both fibers on the left side (viewed from back of the suspension) were gone.  We started looking for the debris and found it just outside of the chamber on the floor and on the trolley that the unit had been sitting in overnight last night and this morning.  We did not find anything like a bug nearby, but investigation ongoing.  Meanwhile, we're revising the schedule to back up a few steps and start the rebuild.  History of lower structure after it was welded:

- Welded main chain unit sat on it's trolly in weld room for ~a week before we were ready to install it.

- Yesterday, we rolled the covered main chain on it's trolly ~60ft from weld room cleanroom to the staging cleanroom.  We then uncovered it, and lifted it with the genie duct jack and placed it on the reaction chain trolley and mated the 2 units together.  The fibers were intact at this point.  We then covered it again and rolled the whole unit ~10ft from the staging cleanroom to the chamber cleanroom and parked it near the door.  We left the cover on it and the genie duct jack parked around it with the forks loose around the structure to aid in protection.

- This morning we uncovered the unit, and then lifted the unit off of the trolley with the genie lift and set it up on the install arm elevator.  We then continued with the installation of the unit into the chamber with the arm.  A few hours later we discovered the broken fibers.

All of the above is standard install procedure, with the same equipment (trolly, cover, genie lift, install arm) used in each of the previous QUAD installations.

Since we found the fiber debris all over the trolley and on the ground outside of the chamber on the floor, it must have happened between yesterday late afternoon when we rolled it there and this mornign after we picked it up and set it into the install arm.