Hugh noticed that some HEPI ITMY slow channels were not correct in the DAQ. I did some investigation and found:

Of the slow channels (datarate=16) in H1ISIITMY.ini :

channels associated with the FEC were correct

channels associated with Filter Modules were correct

other slow channels were incorrect, either showing zero or other data (like channel hopping)

Fast channels appear to be OK (within the limited sample I used)

The problem is not seen with HEPI BS or ETMX. Looking at the models they all look identical.

Reboot/Restarts:

I first restarted h1hpiitmy. Then tried restarting all models on h1seib1. Then tried a restart of the mx-streamers. Then recompiled and reinstalled the h1hpiitmy model and restarted it, followed by a DAQ restart (DAQ status was good, so this should have been unnecessary). All to no avail.

The only thing I haven't done is reboot the computer.

It appears this problem appeared when we upgraded to RCG2.8 on Tue 12 November.

CP3 alarmed, appears to have been filled
Cleaning at end Y
Apollo torqueing bolts at end X
Corner Station chiller 2 tripped, restarted by Ski

08:52 Joe and Gerardo swapping PSL light pipe shutter
09:05 Filiberto to look at end X transmon cabling
09:39 Joe and Gerardo done swapping PSL light pipe shutter
10:02 Sheila removing beam block from PSL table
10:34 Hugh reports that the corner station HEPI pumps tripped off early this morning
10:51 Dave, Kiwamu restarting H1 LSC model, DAQ restart
10:56 DAQ restart
11:16 Hugh modifying corner station HEPI pump controller (WP 4325)
11:19 Dave restarting all models on H1 ISC EY
13:02 Thomas V. working on ITMX optical lever enclosure/bellows
13:08 Safety inspection walking through the LVEA
13:28 Gerardo to LVEA to find Thomas V.
14:16 Arnaud running transfer functions on TMSX
14:17 Apollo removed dome and north door from BSC 10 (WP 4326)
15:22 Jeff K., Sheila and Arnaud purposely tripping chamber watchdogs
15:23 Hugh done modifying corner station HEPI pump controller (WP 4325)
15:36 Gerardo and Thomas V. done with optical lever enclosure/bellows

Last night the power watchdog on the PSL tripped again.  I turned the last back on this morning after the shutter was installed. 

Note:  this morning I transitioned to science mode, then stefan and I reentered to make a measurement of the modulation depth with the OSA.  When I transitioned back to commisioning mode I forgot to turn the AC on, so there was probably quite a temperature fluctuation.  Now the psl is in commisioning mode.

Since the PSL is at least working reasonably now I updated safe.snaps for the iss, fss, and pmc. 

Kyle, Gerardo 

Replaced (2) each 1 1/2" check valves, (2) each  3/4" check valves, (2) each exhaust air valves + actuators, left and right tower desiccant (175lbs/tower 1/8" alumina) and the two nearest downstream filter elements.  The (2) large air valves which supply the towers during the "drying" cycle are also nearly wore out and will need replacing at the next service interval. 

Kyle, John 

As alumina dust had been generated and had been visible in the 1st downstream filter element, we made rudimentary particulate measurements from the 80 psi air (before 1 psi regulator) sampled between filter stages 3 and 4.  

For comparison, we routed nitrogen from a UHP N2 bottle+regulator via 8' of 1/2" poly tube through a 0-5 GPM (liquid) rotameter which exausted into one end of a flow box.  A particle counter was placed inside of the flow box near the opposite end.  

2 GPM indicated flow    Bottled UHP N2      sampled air 
                                     260,000 0.3u        45,000 0.3u 
                                       17,000 0.5u        12,000 0.5u 
 
3 GPM indicated flow    750,000 0.3u        50,000 0.3u 
                                       23,000 0.5u        13,000 0.5u 

30 minutes after routing the purge air supply to the LVEA Class 100 manifolds I moved the flow box apparatus to the XBM purge air connection and measured: 
 
0.8 GPM indicated flow        200 0.3u 
                                             0.0 0.5u

Saw the corner station pump output at zero with max drive from controller.  This condition usually indicates a level trip.  At Pump Stations on mezzanine, indications confirm.  Plus when I attempted to restart, it tripped again from fluid level so I conclude it was a level trip fault.  I lowered the trip point to compensate.  There is only about 3/4" more this can be done.  Why...colder temps? Nah, likely a pretty small effect, air bubbles working there way out?  Probably some of that going on; if it comes out of the LVEA plumbing, it can be released in the reservoir tank.  Leaks?! There are some minor leaks from two of the pumps but these are a drip or two a day kinda leaks, certainly contribute though.  Bottom, line, monitor the levels, clean up the leaks, fix the leaks, add fluid as needed.

I took the opportunity of the shut down to hardwire the FWD switch on the PS controller.  This is already the case at the End Stations.  Consequences: 1) After a shut down, there will be no need to open the panel to restart unless there is a fault to acknowledge on the controller.  This is good because access to the panel requires electrical checkout/approval from McCarthy or having McCarthy do it.  2) Restart will be abrupt if it is not brought up slowly (manually).  This could be bad as the HEPI may shake things around so please understand the process when you restart the system after shutdown.

The dome and door bolts are installed and tightend at BSC 9.
Miscellaneous rigging, tools, o-ring protectors etc. was removed from E X and returned to corner/E Y, as needed.
The north door and dome are removed from BSC 10.
We returned the CPB jigs back to the corner station from E Y. 

In the aLIGO era, folks working at the upper level of BSCs have garbed on the E-module and then entered the Chamber Cleanroom. With the cleanrooms, E-module, etc. in the BSC10 configuration, there is no easy/acceptable way to access the E-module. If you need to do ANY cleanroom work (upper OR lower level),turn right as you come through the Change Room man-door, find the Lower Level Garbing/Staging Room and "bunny-up" there. If you need to work on the upper level or at the spool, pass through the Chamber Cleanroom to access the Spiral Staircase or the CPB Cleanroom. Avoid walking through the Test Stand Cleanroom. If you want to access the far side of the beamtube, there is no need to garb up: just turn left as you come through the man-door and follow the wall around.

I modified the forton files for HAM2 ans HAM3 ISI so the filter locations, and order would match the latest SEI directives. 

I attached a picture of HAM3-ISI blend filters, after updating the foton files. Foton files are attached too.

Kiwamu, Stefan

We put the existing initial alignment system in the h1asc model into its own library part.the files are in /opt/rtcds/userapps/release/asc/common/models :

IAL_LOCKIN.mdl

IAL_MASTER.mdl
 

SEI macro substitution files were created for: HEPI-ETMX and HEPI ITMX - SVN r6548

The SITEMAP.adl was updated accordingly - SVN r6549

I turned ETMX-ISI back on today, to make sure that the recent closeout activity did not impact its functionality.

It is now running under the following configuration:
- Damping
- T100mHz_NO.44 blend on Stage 1
- 250mHz blend on Stage 2
- Level 2 Isolation loops

Apollo asked me to lock HEPI so they could finish torquing the dome on BSC9, so this morning ~8am I went to the end station. Did not use the IPS because DI's are still on the chamber, so I was probably more disruptive than normal, but Apollo said they would probably have to walk on the cross-beams anyway. Apollo is done now, so we will unlock when needed or convenient. This and HAM1 should be the only HEPI's in HIFO-X chain that are locked now.

h1iopiscey has been showing a timing error since 16:15 Wed afternoon (local time). I restarted all the models on this front end and the error has been cleared.

Chirs W, Dave B and Kiwamu I (under WP#4296)

I svn-updated the following directories to reflect the recent changes for the LSC model that Chris has been working/organizing:

• lsc/common/models
• lsc/h1/models
• cds/common/models
  • This contains UGF.mdl
• cds/commonsrc
  • This contains DB2MAG.c

With help from Chris and Dave, I complied, built and restarted the h1lsc model. Now it is up and running. DAQ also restarted.

(Joe, Gerardo)

PSL light pipe and shutter were removed from the chamber side (HAM01), the main PSL beam vieport was exposed to apply FC on it.  FC application (Joe used a small bottle) was finished by 4:10 PM yesterday, to protect the viewport we attached the old vieport adaptor with 3 screws to the viewport, and covered the 3" center hole with aluminum foil.
Work will continue today.

Stefan, Sheila and Kiwamu with remote assistance from Rich A. and Volker

The mysterious jump in the RF phase (see alog 7941) is now understood and fixed. It was due to a loose connection at DB15 connectors in the EOM box and not due to the SMA connector (see alog 8811 and 8813 for our early detective story). We applied two small in-situ modifications on the EOM box. As a result, now it doesn't show the mysterious RF jump any more.

The box:

I briefly explain the EOM box for those who are not familiar with our custom-made EOM box. The EOM box consists of two boxes -- one contains LC resonant circuits and the other contains the EOM crystal. This two-boxes-design allows one to tune the resonant frequencies by tweaking the LC circuits without messing up the alignment of the EOM crystal because one can simply take out the electronics box and leave the crystal box for solder or tuning some parts in the circuits. To apply voltage across the EOM crystal for normal operation, the LC circuits need to be connected to the EOM crystal. This is done by a DB15 connector attached on each box -- female DB15 on the crystal box and a male DB15 on the electronics box (see pictures shown below). In this way, the two boxes are electrically connected.

A picture of the actual EOM box. The gloved hand is me pressing the SMA downward in order to reproduce the RF jump.

When I pressed the SMA connectors downward in this morning, the RF characteristic of the EOM box changed as if something jumped. This was repeatable, although it seemed that the condition to make it jump was random -- occasionally, pressing the SMA didn't make it jump and sometime pushing the SMA toward the box made it jump. Anyways, at this point, it was clear that the EOM box was the culprit and not the RF cables.

The causes:

• The mechanical connection between the two boxes was not solid.
• The DB15 connectors were not completely all the way in.
• The above two facts resulted in different amount of stray capacitance in the DB15 connectors depending on the relative orientation of the two boxes.
• Though, this doesn't explain the bi-stable impedance modes we have observed. Maybe there was some sort of bi-stable condition in the physical orientations.

At the beginning, we thought the culprit was the SMA connectors (see alog 8811 and 8813). However, this turned out to be wrong as we investigated it further. With a remote assistance from Volker, Stefan and I took the electronics box apart from the crystal box while keeping the crystal alignment. In the process of removal, we found that the electronics box was attached to the crystal box merely by friction of the DB15 connectors and three pieces of adhesive tape. So the orientation of the electronics box was not so solid with respect the EOM box. We then checked the return loss of the LC circuits without connecting the EOM and confirmed that wiggling the SMA connector didn't change its impedance. Instead, we discovered that the DB15 connector can easily change the amount of its stray capacitance -- the frequency of the resonant notch could shift by approximately 1 MHz by very gently touching the DB15 connector with our latex-gloved-hand. We put the electronics box back on the EOM crystal box and wiggled the orientation of them. Indeed, it changed the resonant frequency by about 1 MHz in a discontinuous way. So we determined that the loose connection in the DB15 connectors was the culprit and the mysterious RF jump was induced by some change in the orientation of the two boxes.

It seemed that the DB15 connectors were not all the way in because the two aluminum boxes contacted first.

The repair/modification fixed the issue:

We did the following two repair/modification:

• We put washers in the DB15 connector of the electronics box to make the connector stick out a bit more.
• We introduced four screws to mechanically tighten the connection between two boxes.

The DB15 connector have two screws to support it and we put a washer for each screw. They raise the height of the DB15 connector like shims. Also, we newly installed four screws to make the connection of the two boxes more solid. There were already four screw holes on each box to accommodate them. So we just installed them. After these modifications, we checked the return loss of all three RF ports. We didn't observe the mysterious jump at all, even when the electronics box was wiggled hard. Of course, strongly pressing the electronics box downward shifts the notch position by an order of 10 kHz due to the change in the stray capacitance at the DB15 connectors, but the shift is smooth and not in a discontinuous way any more. So the RF jump issue is now solved.

A picture of the electronics box when apart from the EOM crystal box. This DB15 connector was shimmed by washers.

A picture of the EOM crystal box when the electronics box is taken away. There is a threaded screw hole on each corner and these are the ones we used for installing the new screws.

A top view of the EOM box with the lid off. The green circles indicate the screws that we newly installed.

A top view of the EOM box with the lid off. The green circles indicate the screws that we newly installed.

J. Kissel, C. Vorvick

Calum and Betsy installed the new PR3 baffles (D1300957) yesterday (see LHO aLOG 8619), but only roughed in their alignment with respect to the optic / prisms / wires. Today Cheryl and I went into HAM2, and I aligned the baffles to-the-best-of-my-ability, by-eye, and then tool tightened them to PR3 HLTS structure. My left-right metric was "just covering the edge of the bevel on the optic with the straight portion of the baffles" and my up-down metric was having the "bottom curves of the baffles following the curvature / bevel of the lower half of the optic."  It was certainly a ball-park activity, given that the goodness of alignment depends heavily on one's angle, orientation, and height of viewing. From most face-on views, however, the baffle looks to be baffling the wires and not clipping the optic.

I took many pictures, but did not find out until later that the memory card on the camera was full, so the pictures were not being stored and lost. FFFFUUUUDDDGGEEEE. 
Cheryl, Calum, Betsy, and Kate have graciously volunteered to retake pictures tomorrow before / after they pull of the first contact on all the optics.

We also took B&K hammer transfer functions after securing the baffles (with PR3 freely suspended, and the ISI Locked, and probably with me leaning on the table). I placed the accelerometer in the exact same location and configuration as in LHO aLOG 6014, (4th picture attached) with ACC +X = PR3 -L, ACC +Y = PR3 -T, and ACC +Z = PR3 +V, in the upper left corner of the cage on the HR face, "front" of the cage. The X impacts with the hammer were just below the accelerometer, on the HR face. The Y impacts were on that same strut / corner, at the same height as the accelerometer, on the -T face of the cage, in the +T direction. Comparing these results with the previous results taken before the baffles were installed -- by-eye they look roughly equivalent. The resident LHO expert of post-processing the data is off to LLO for a few days, so we'll post a comparison next week (or I can ask Calum / Stuart tomorrow, we'll see.)

For now, the saved templates live only on the laptop, and live in
C:\Users\ligo\Desktop\SUS Hammer Test\PR3
and are called
SimpleHammerDisplay3-PR3Baffle-ISIlocked-Ximpact.pls
SimpleHammerDisplay3-PR3Baffle-ISIlocked-Yimpact.pls

(Alexa, Sheila)

I am still puzzled by some of the mode matching measurements I took of ISCTEY. I had attemped to compare the accuracy of the Mode Master, Nanoscan, and Knife edge. They seem to approximately agree, but it's hard to say which is most accurate.

Regardless...I examined the beam profile exiting the table enclosure with all the lenses in place using the Mode Master. Refering to D1100607-v13, I found the optimal placement of the telescope to be: ALS-L6 13.5 inches from ALS-M9, ALS-L7 6 inches from ALS-M10. With an "a la mode" script, I determined the mode matching overlap with the TransMon secondary mirror to be 87% for the horizontal profile and 98.6% for the vertical profile. This overlap was computed with a 2.2mm waist at 3.5m from ALS-M11. The waist size and location was determined via T1200200-v1, D0902163, D1201457 as done in the ISCTEX MM alog with a correction of 4ft for the panel location change between EX and EY.

I have attached the scripts and MM snap shot. Note: the first matlab script is just the profile taken by the MM after the telescope, along with the overlap computation. The second script contains the full profile of ISCTEY, which is where some of my confusion persists.