Kiwamu, Sheila

After the transition to science mode on friday the temperature in the PSL enclosure has stabilized, now the daily fluctuations are about 1 degree F and the 20 min oscillations are completely gone (bottom right of attached screen shot).  The laser power (top right) is slightly more stable, and 20 minute oscilations in the ref cav trans are completely gone (lower left). 

We will stay in science mode unless someone is actively inside the PSL.  

Kiwamu and I went in this morning (we transitioned to commisioning mode first) with three goals: Kiwamu wanted to investigate the status of the rotation stage, to try to reduce the comination of the PMC trans camera with light from the reference cavity, and to search for reasons for the ref cav misalignment/ fix it.  

The image on the PMC trans camera is clearly different when the ref cav is locked and when it is not.  We confirmed that this is not due to a stray beam coming at a large angle but something that travels almost along the same direction as the PMC trans beam.  There is also a large "smudge" in the image.  We found that by blocking the power meter on the PMC refl port we got rid of the smudge. TIlting the power meter slightly did not move the smudge position on the trans camer at all, although it did seem to reduce the amount of power in it.  It seems as though this is light that is scattered into the PMC reverse propagating mode from the power meter, and then clips somewhere on the way to the trans camera.   

We also seem to have light from the reference cavity in this "smudge".  This could be because the reference cavity circulator is slightly off, and some of the reflected light gets back into the PMC in the counter propagating mode.  We did not try to adjust the quater wave plate, but it is possible that could help.  

In the end we added an iris before the PMC trans camera, there is still a difference between the on screen image with and without the reference cavity locked, but it is much better and the centroid position calculated by Chris's script is almost the same when the ref cav is locked and unlocked.  A picture of the camera before and after the iris was added is attached.  For future reference, on the PMC trans camera up on the table looks like down on the camera, and vice versa.  

We then moved on to looking into the ref cav alignment.  We tried taping and pushing gently on the optics in the path from the PMC to the ref cav, nothing was obviously loose and the alignment always returned to where it was when we took our hands off.  We noticed that the beam shape after the EOM was bad, and when we looked at the aperatures on the EOM the beam seemed high.  We brought the EOM up, which improved the beam quality.  When we were done adjusting the EOM we had (with the ISS disabled) 12.9mW incident on the EOM and 12.6mW after.   This clearly misaligned the reference cavity, we recovered the alignment using the camera and the iris at the front of the cavity as references, and adjusting the top periscope mirror and M27 in pitch.  We were able to get back to about 0.6V on the transmitted PD by walking the beam once the cavity was locked.  We expected to be able to get more than 0.9V, so something is still amiss.  Also, the spot position on the reflected camera has moved.  This does not make sense because there is no reason the reference cavity position should have moved (we think) and the beam is aligned to the cavity.  When the cavity is locked we see bassically nothing in reflection right now, and the spot position on the trans camera down't appear to have moved much.

We then realigned the ref cav reflection PD, using the DC power on a scope.  Kiwamu noticed that the value on the MEDM screens for the refl DC was not changing while I did this.  It seems as though the input is grounded in the model, he is looking into how we can get this recorded in the digital system.  

The EOM misalignment could have been a symptom or a cause of the problems we have been having.  We decided to leave things be for now, and have a look at trends over thanksgiving when no one should be going into the PSL and the PSL is in science mode.

Last week there were issues noted with the Picomotor cable for EX TMS, and it was traced to mislabeling of the (4) Mighty Mouse connectors on a D1000238 cable (basically you want the Mighty Mouse closest to pin-1 of the 25-pin connector to be labeled "D" (@ EX it was labeled "A"). 

I went ahead and confirmed that our (6) remaining D1000238's were labeled correctly (they are), and then also marked the shell of the DB25 with a "1" on its outer surface close to where pin #1 is (see attached photo).  This is because it's impossible to confirm the orientation of the DB25 (or where pin#1) is when it's connected to another cable and screwed down to a cable bracket.  So, if one marks the outside of the connector it's easy to determine the orientation of the connector without unplugging everything.

guardian/cdsutils deployment

We [Jamie, Dave and Jim] recently deployed initial releases of the guardian and cdsutils software packages at LHO:

• guardian - the main guardian codebase
• cdsutils - python library and cli for CDS interaction, including ezca channelaccess library. the cdsutils python library is a depndency of

A more complete description and usage of these packages is described (or will be) on the awiki links above, but very briefly:

cdsutils is ageneral purpose tool for interacting with the CDS system that is meant to include much of the commonly needed functionality that used to be available in the old ezca/tds name space (read/write/switch/avg/etc). It is both a python library, and a CLI that wraps some of these functions for use at the command line:

controls@operator0:~ 0$ python -c "import cdsutils; print cdsutils.avg(2, 'H1:LSC-DARM_OUT')"
{'H1:LSC-DARM_OUT': 0.0}
controls@operator0:~ 0$ cdsutils avg 2 H1:LSC-DARM_OUT
H1:LSC-DARM_OUT 0.0
controls@operator0:~ 0$ 

See 'cdsutils --help' for a brief usage.

guardian is the in-development aLIGO automation system. The 'guardian' program itself serves two purposes:

• interface to execute a single state from a guardian system state machine description module.
• launch the guardian state machine engine daemon for the state machine described in a guardian system module

See also 'guardian --help' for command line usage.

installation

Both cdsutils and guardian were installed in /ligo/apps, and should be available for all users:

controls@operator0:~ 0$ which guardian
/ligo/apps/linux-x86_64/guardian/bin/guardian
controls@operator0:~ 0$ guardian --version
guardian svnr308
controls@operator0:~ 0$ which cdsutils
/ligo/apps/linux-x86_64/cdsutils/bin/cdsutils
controls@operator0:~ 0$ cdsutils --version
cdsutils svnr95
controls@operator0:~ 0$ python -c 'import ezca; print ezca.__file__'
/ligo/apps/linux-x86_64/cdsutils/lib/ezca.pyc
controls@operator0:~ 0$ python -c 'import cdsutils; print cdsutils.__file__'
/ligo/apps/linux-x86_64/cdsutils/lib/cdsutils/__init__.pyc
controls@operator0:~ 0$ 

For the moment they are installed directly from the SVN, but as they stabilize they will be installed from properly versioned releases.  The svn checkouts for both packages are in "/ligo/svncommon/CdsSVN/":

• /ligo/svncommon/CdsSVN/cdsutils -> https://redoubt.ligo-wa.caltech.edu/svn/cdsutils
• /ligo/svncommon/CdsSVN/guardian -> https://redoubt.ligo-wa.caltech.edu/svn/guardian

h1guardian0 setup

Infrastructure for management of guardian daemon processes was setup on h1guardian0.  This system manages the starting and stoping of guardian daemon processes, and logging their output.

The guardctrl tool, included with the guardian installation, is for interacting with the above described daemon supervision system.  It includes commands for creating, starting, stopping, restarting, etc. the daemons, and for accessing their logs.  It can be executed from any CDS workstation, and communicates with the setup on h1guardian0 via ssh.

the control room workstation opsws7 (behind the operator station) appears to have a bad fan. It is turned off and will go out for repairs soon.

Since the TMSX TF was OK according to JeffK, I and Corey went into the chamber and removed the temporary alignment stabilization tooling.

I made sure that nothing is left on the TMS ISC table.

Transitioned the ETMy HEPI Actuators to Run mode.  We are ready.

Mike V, Gerardo M, and Mitchell R

Saturday we were able to complete the primary baffle assembly and transferred to the balancing fixture. Current steps now are to complete the balancing process and transfer assembly back to the installation fixture. This should be completed by mid-late afternoon today. This would ideally mean that the Cryobaffle assembly will be ready for install Tuesday. 

J. Kissel

I'm slowly working my way through the 14 suspensions that need chamber close-out testing after last week's furiously awesome effort to get doors on before the Thanksgiving week. It's going slower than expected because, for some of the SUS, I'm taking "Official SUS-style" TFs for the first time at LHO, so I've gotta tweak the measurements, make sure the SUS are set up correctly, and also because -- even on a Saturday! -- I'm interrupted with people calling me ask work questions ( THANKS JANEEN, GOSH (-; ). Anyways, enough whining. So far, I've seen no signs of badness -- excellent! Details below, and plots proving the goodness to come later.

----

I've now got and exported the (DTT) data for all degrees of freedom of M1 to M1 TFs
TMSX, RM1, and RM2
and they've been committed to the SVN. I'll process the data later when I find out from my colleagues in what the state of the matlab analysis software lies for these orphaned SUS types that we're now re-adopting into the family.

I still need to get M1 to M1 TFs of
IM1, IM2, IM3, IM4, MC1, MC2, MC3, PRM, PR2, PR3
and both M0 to M0 and R0 to R0 TFs for
ITMX

I've created new directories in the SUS repo for the HTTS, which hadn't been created yet for LHO because no one had taken any data on them.
The templates for the data I've taken are here:

/ligo/svncommon/SusSVN/sus/trunk/HTTS/H1/RM1/SAGM1/Data/
2013-11-23_2120_H1ASCRM1_M1_WhiteNoise_L_0p1to50Hz.xml
2013-11-23_2120_H1ASCRM1_M1_WhiteNoise_P_0p1to50Hz.xml
2013-11-23_2120_H1ASCRM1_M1_WhiteNoise_Y_0p1to50Hz.xml

/ligo/svncommon/SusSVN/sus/trunk/HTTS/H1/RM1/SAGM1/Data/
2013-11-23_2214_H1ASCRM2_M1_WhiteNoise_L_0p1to50Hz.xml
2013-11-23_2214_H1ASCRM2_M1_WhiteNoise_P_0p1to50Hz.xml
2013-11-23_2214_H1ASCRM2_M1_WhiteNoise_Y_0p1to50Hz.xml

/ligo/svncommon/SusSVN/sus/trunk/TMTS/H1/TMSX/SAGM1/Data/
2013-11-23_2045_H1SUSTMSX_M1_WhiteNoise_L_0p1to50Hz.xml
2013-11-23_2045_H1SUSTMSX_M1_WhiteNoise_P_0p1to50Hz.xml
2013-11-23_2045_H1SUSTMSX_M1_WhiteNoise_R_0p1to50Hz.xml
2013-11-23_2045_H1SUSTMSX_M1_WhiteNoise_T_0p1to50Hz.xml
2013-11-23_2045_H1SUSTMSX_M1_WhiteNoise_V_0p1to50Hz.xml
2013-11-23_2045_H1SUSTMSX_M1_WhiteNoise_Y_0p1to50Hz.xml

For each, I've exported both as .pdf and .txt with the same name (_tf tag for transfer function, _coh for coherence) for use later.

The triples and QUAD should go much quicker since there are already many measurements of them with known good DTT templates, so it's just a matter of opening them up and hitting go. Why DTT, and not the automated Matlab software that you can just set and forget, you ask? Because over the course of 6 hours, there're too many things that can fail with the automated system at this point, and you don't know of a failure until you've wasted the six hours. I'd rather mindlessly click buttons for six hours with DTT, constantly confirming that things are going well, and listen to some good tunes on a Saturday afternoon.

Craig Conley, Mark Lane, Joe DeRenzis, Chris Soike, Rick Savage

The in-vacuum periscope structure for Yend was pre-aligned in the H2 Laser Area Enclosure.  Mirror mounts were removed and it was wrapped and transported to EndY where it was installed inside the A7 adapter.  It was moved to the staging location where it was rotated 90 deg. to allow access for the Manifold/Cryobaffle installation work.

Tools and installation tooling were left inside the periscope frame in anticipation of its move to the final location after the baffle installation.

Photos attached.

Chris, Kiwamu, Sheila,

I just transitioned the PS to science mode, in the hopes that this will at least simplify troubleshooting the aligment problems we are having if we dont have the 20 minute temperature swings anymore. 

Chris has written verry nice scripts that take snapshots of the two transmitted cameras, and calculates the centroid. 

The ref cav trans centroid is temporarily plotted as H1:LSC-DARM_OFFSET and H1:LSC-MICH_OFFSET and the PMC trans centroid is saved as H1:LSC-SCRL_OFFSET and H1:LSC-PRCL_OFFSET. 

we can use this decide if the cavities themselves are moving.

The TCS viewports were installed on BSCs 1 and 3 today. The gap on the optic enclosure for the primary optics was between .030" and .040" on both viewports. So the new Kapton shims created a good gap for pulldown onto the Viton. They were proofed and leak tested Wednesday and Thursday. There was a small issue with the proofing as the Kapton wedge on top of the primary optics doesn't seal (nor is it expected to). So although the pressure for proofing (15.5 psi) was reached it would not hold. This was determined not to be an issue. The leak test went without incident, Helium flow was 1 LPM and each optic was targeted for ~30 seconds and the enclosure was targeted for ~1 - 1.5 minutes. No leaks were seen, calibration using internal leak showed 1.0e-7 torr L/s with a stated value of 9.8e-8 torr L/s.

The viewports replaced the blanks on the chambers today. Kyle noted the gap pre-torquing of the bolts was low and the gap was found to be 0.02". Lower than is normally sought (>0.025"). This seems to be from a deeper recessed knife edge on the viewport. John gave the okay to torque the viewports down and continue. They are now metal to metal on the flange.

Picomotor mistery solved: It was an in-chamber cable. Unlike the ones Corey looked at in the lab today, what we have in chamber is "A"=J5, "D"=J2. See Alexa's alog.

In-chamber alignment done: We adjusted the IR QPD sled. Since we already rotated the 50-50 BS 180 degrees yesterday, there was no way the QPD sled could be left untouched, but anyway when we started it was not that bad, and only a small touch up was necessary.

One thing we did NOT understand is why it was easy to do this this time. If you remember the story from EY installation, the reflectivity of 50-50 IR splitter is almost zero for green S (correct) polarization, and it was almost impossible to find the right beam without using a bad polarization. This time, however, it was easy to see the beam with the beam leaving the table as is, and rotating the polarization 90 degrees made it worse.

Maybe the existence of the first contact on the HR side of ETMX is doing something funny for the polarization, though the transmissivity of HR for green is between 3% and 15% according to witness sample measurement by LIGO and 27% according to LMA (https://dcc.ligo.org/C1103232) so it's not that likely that the FC is doing something significant.

Becoming suspicious if we're using correct optics, we checked the DCC number of 1" IR 50-50 splitter and 1" high reflector, but they were both correct.

Anyway, the important thing is that the green leakage light into IR QPD sled is centered on QPDs when the ETM is retro-reflecting the beam, and we (Kiwamu and I) are quite confident that the beam we're looking at is the correct one.

BOSEMs were checked, they 're not touching, and the centering is not terrible either (though certainly it's not perfect). We set the depth of BOSEMs so the output becomes 15000-ish.

Some tools that are not ours are left in chamber.

What's next:

1. Jeff K will take another TF. If it's good, we'll NOT touch BOSEMS.
2. After that, we can give the chamber back to IAS for ACB alignment, and to the SUS.
3. Before closing out, we MIGHT want to push the dust particulates off of the telescope optics using nitrogen gun and wipes (they're dirty and you can see it from the pictures), before pulling the first contact from ETMX.

[Jim, Sebastien]

This week, we've seen the gain of L4C-H2 changing randomly, from nominal gain to a factor of 2 lower.

We tried to chase down this issue without success (see aLOG #8632 and 8666). Everytime we tried something, it magically resolved by itself, and there were no way to reproduced it.

After it happened again today, we might have found the issue though: the in-air cable at the feedthrough felt a little "squishy" and loose. Jim unplugged it and readjusted everything. It feels much better now, and the issue is once again gone.

According to what we saw before, we're pretty confident that was it. If it wasn't, let's keep in mind to start looking at the feedthrough when it happens again.

All the HEPI models have been updated, compiled and installed at LHO on all teh HEPIs. The update went great, no issue so far.

The next action monday will be, on HEPI-HAM2 and HEPI-HAM3, to translate the alignment offsets calculated before when the IMC was locked (see aLOG # 7189 and 7180) from local basis to cartesian basis.

Not a hard task, but I've been having trouble with it. I'll continue Monday.

Tfs were run last night but we have a few empty buckets. Filling those overnight. OPSWS0.

Deed was done ~10 am local.

Day Shift Summary
LVEA Laser Safe

Patrick – Changing dust monitoring sample and hold times on several dust monitors
Gerardo – At End-Y
Filiberto – At End-Y working on the Beckhoff system
Jonathan – Recovering equipment from LVEA
Cory – At End-X gathering TMS parts for shipment to LLO
Sebastian – Restarting HEPI models
Hugh – At End-Y working on HEPI 
Dave/Jim – Restart the DAC
Dale – At End-Y checking on P-Cal Install
Scrap Metal Recyclers on site to change out collection containers
Cyrus – At Mid-X to swap media converters