Rolf, Jim and Dave. We are declaring a success on the RCG2.8 upgrade. All user models are running under 2.8.

The DAQ was recompiled against 2.8. It was reconfigured to acquire FE slow data on the DAQ network. It still writes 32 second full frames, one minute second trends and one hour minute trends. These will be extended in the future when more memory is acquired.

The commission frames were renamed from H-H1_R-GPS to H-H1_C-GPS so science frames (R) and commissioning frames (C) now have different names.

LDAS will start permantly archiving the science frames from now onwards.

Following LHO's upgrade to MATLAB 2012whatever, I was disproportionately dismayed to see that the nice system shortcuts for loading working folders did not make the transition. Hugo was able to help me figure this out by:

1. First we tried just copying the shortcuts.xml from the 2010 (/ligo/home/controls/.matlab/R2010b) folder into the 2012 (/ligo/home/controls/.matlab/R2012b) folder. This didn't work.

2. Matlab seemed to make a shortcuts2.xml at this point. We opened the old shortcuts.xml in a text editor and copied the contents into the new shortcuts2.xml. After another MATLAB restart, this fixed it. Shortcuts restored.

I'm attaching the old shortcuts.xml file as a txt.

Arnaud and Kissel took the opportunity of today's DAQ and internet outages to re-do the B&K Hammer testing of the ITMx (since we had moved the suspension on the ISI last week by 6mm, thus compromising the mating stiffness between the 2).  They discovered that the corner cube was still mounted to the ITMx, which unfortunately surprised Jim.  They removed the corner cube and are leaving it off.  Jim will recheck his table balancing.

We have recompiled all user models (saved with matlab2012b) with RCG2.8. The exception is the HEPI models which are currently compiled with a 2010a version of hepitemplate.mdl (we ran into the feCode error today when recompiling).

All front ends are now running 2.8 code. The DAQ is recompiled, and is reconfigured to get slow data from the front ends directly. The EDCU only acquires non-FE slow data, its channel rate is reduced to 3656.

Next we will work on reconfiguring the DAQ to rename the commissioning frames from H-H1_R-GPS to H-H1_C-GPS, renaming existing files in /frames/full

We shutdown the PSL front end last and started it first. The PSL laser was down for about 30 minutes due to the upgrade. This is representative of the PSL recovery time if the experts are on hand at the time the front end is taken down.

These are now running the same code as the dust monitors in the optics labs (alog 8294).

The EDCU channels will be updated with the next DAQ restart.

If there are no problems after a week, I will create a tagged release and update the target directories.

We are upgrading H1 to RCG2.8 this morning, all front ends will be restarted and the DAQ will need to be shutdown for extended periods.

Alastair H., Thomas V.

We took the MEDM screens that were recently developed at LLO and added&committed them into SVN and then checked them out to the LHO userapps/release/tcs/common/ area. Then we changed the macro substitutions to conform with LHO's sitemap, the channels showed up nicely so it was a good success.

Jim and Dave. We upgraded all mdl files used in the userapps area from 2010a to 2012b. To verify that we did not break anything, all the models were recompiled against RCG2.7.2.

The only problems we encountered were with the files hepitemplate, isihammaster and isi2stagemaster. The isihammaster was resolved by saving with 2012b, re-opening and resaving. The other two files required model changes. I'll leave the details to Hugo's alog.

Some IOP models (SUS watchdogs) reference mdl files in the advLigoRTS area. Some of these were created against very old matlab versions (2008a, 2008b for example). We have left these alone on our upgrade, but perhaps all rtscore mdl files should be upgraded at some point.

We only worked with RCG2.7.2 today, and only compiled (did not install). All 2010a versions of the mdl files were saved at .mdl.2010a files so we will easily be able to revert to 2010a  if we encounter problems during the 2.8 rebuild.

I adapted the previously existing code for bi-directional RTS-to-RTS and RTS-to-Beckhoff communication to work with an EPICS link. The code uses a keep-alive bit on the channel to verify the link is still up. Its current time-out is set to 10sec (same as DAC-ADC-based version).

New files:
 - isc/common/src/RtCommEPICS.c    : the c-code for receiving (in svn) - this is an adapted version of RtCommunication.c
 - isc/common/models/STATECOMM.mdl : Added a new function block for the EPICS-based link

Currently I am running the link in loopback test mode between h1odcmaster and h1odcx.

Cheryl, Peter, Joe Gleason, Stefan

After making sure the input mode cleaner was flashing with a good alignment, we did the following alignment work for HAM2 and HAM3:
- IM1/IM2 tweak-up to center through the Faraday.
- IM3/IM4 tweak-up to center on irises marking the center of PRM and PR2
- Rotated the PSL polarization by 90deg (45deg on the lambda/2 stage) (required because of the REFL periscopes on HAM2)
- Installed irises for marking the IM transmitted beam, just after MC3 and before PRM
- Installed re-direction mirrors to redirect the IMC REFL into the path of IMC trans (alignment using the just installed irises)
- Joe is still aligning the last steering mirror in the IM4 transmitted path.

Just after lunch h1fw1 started going unstable. We also found that data retreival from h1nds1 was slower then h1nds0. Dan investigated the solaris QFS writer and found it unresponsive. He rebooted the Solaris machine (h1ldasgw1) and performed a file check on the QFS system. No problems were found and h1fw1 was restarted.

We are in the process of upgrading all front end model mdl files from matlab 2010a to 2012b. To do this, we are opening the mdl files with 2012b, making a text change on the top level and then saving the file. The changes are to add the SVN tags $Id$ and $HeadURL$ as strings (or moving them/changing font size if they exist).

If the model is a common library part, it must be unlocked before it can be edited. This is done (after the model has been opened) by the matlab command line:

>> set_param('rtdemod','Lock','off')

where rtdemod is the mdl file name (leave out the .mdl part). You should see the padlock icon on the lower left disappear and you can make changes.

When you save, you have the option of asking matlab to save a 2010a version. Please do this, we can remove these files later when we have tested the system thoroughly with 2012b.

When you close the file the lock is reapplied.

09:11 Mark & Jodi - Installing dust barrier in BSC3
09:50 Koji – OMC work in H2-PSL enclosure
10:38 Mark & Jodi - finished at BSC3
10:45 Justin - transition LVEA south bay to local laser hazard
11:03 Hugo – compiled and restarted the H1ISIITMX model
13:00 Koji – OMC work in H2-PSL enclosure
13:30 Cory - Going to End-Y TMS lab to recover tools/parts
15:10 Koji - OMC work in H2-PSL enclosure 
15:49 Dave – Recycled Framerwriter1  

Since we are in the process of saving all models with Matlab R2012b, the models can no longer be edited with Matlab2010.  So, the default Matlab version has been moved to R2012b.  Note that copies of models that work with Matlab 2010 exist, so it's possible to return if needed.

• Yaw: +23 µrad (CCW)
  • Target: 0 µrad
  • Error: +23 µrad (CCW)
  • Tolerance: ±50 µrad

Alexa, Keita, Stefan and Sheila-

We have the last optic mounted and balanced in a tip tilt (SN 33) for HAM 6 (OM3).  We had difficulty balancing this the optic in this tip tilt,  even with a short bolt, 2 nuts and a washer on the back side we could not balance it.  Keita brought out SN20 and 25 which are intended for India (and do not have cables attached).  We had a similar problem balancing one of these.  Keita then noticed that there is slop in the four scews that attach the piece that clamps the wire to the rest of the mirror holder (picture attached, these are the 2 bolts on top and the 2 bolts on the bottom of the photo.) The tip tilts we were able to balance had those screws centered in their holes while they were off center in the tip tilts we were unable to balance.  We centered them on the india tip tilt and then we were able to balance it using a dummy optic.  We then did the same thing on the TT that was originally intended to be the H1 spare (SN33).

In the process we used a dummy optic from iLIGO that has a scratch in the center to avoid chipping more real optics.  

We took parts from TT SN9:  the peak nudge  (which was missing from SN33), the small screw that tightens the balancing bolt (we dropped this) and one of the OSEM flags (one flag on SN33 was bent).  

In the end, we have not changed anything for the india tip tilts (SN 20 and 25) except correcting the position of the small bolts, we have 3 balanced and ready for HAM6 (SN2,4, and 33) and one H1 spare (SN9) which is missing parts and needs the wires reclamped.  

On monday I will double check all of the serial nubers written in this alog to make sure they are right.  

The attached pdf contains plots of the accelerometer response from B&K impact hammer tests.  There were six locations on the ETMY sleeve structure selected for impacts.  Each plot displays the response in each degree of freedom (X,Y,Z) of the accelerometer.  The orientation of the accelerometer was such that the X direction is vertically up the structure, and the Y,Z directions in the horizontal plane of the optic.  Each measurement has five averages with 0.5Hz resolution.  

Data:
'~/sus/trunk/QUAD/H2/ETMY/Common/VA_Testing/2012_03_26_h2etmy_hammer_test/'

Plots: 
'~/2012_03_26_h2etmy_hammer_test/plots/'