The LIGO Executive Committee Agenda for Monday, April 24, 2006 will be:
(Meeting time: 10:30 am Pacific Time)
1. Announcements
2. Comments on Weekly Report
3. LSC Issues (Saulson)
4. LIGO Lab Operations
5. R&D and Advanced LIGO (Shoemaker)
6. CHANGE CONTROL BOARD/TECHNICAL REVIEW BOARD SESSION AS NEEDED
Special Items:
The APS Meeting is this week. Perhaps because of that, this report is somewhat abridged.
Negotiations continue to finalize the new MOU with Virgo; the most recent discussions were in a teleconference between the LIGO Directorate and Virgo Spokesperson on Thursday 20 April.
At last Friday's LSC Executive Committee meeting (mainly devoted to approving APS talks), we also approved the choice of LSU as the host of the August LSC meeting, and chose the dates of 14-17 Aug 2006 (Mon - Thurs.)
Good luck to all of the speakers at the APS meeting!
Received the August 2005 MOUs listed below approved and signed-off by Peter Saulson. These will be forwarded on to Jay Marx for his approval and signature and then submitted to DCC for posting.
|
MOU |
Attachments |
|
GEO |
ACF, DAT, LAS, OPT, SUS, Z |
|
Loyola |
DAT, Z |
|
Trinity |
OPS, OPT, OUT, Z |
The Sannio MOU received last week was submitted to Marx for approval and signature.
Carleton, IAP and
Working on formatting Embry-Riddle's MOU into the LSC pdf format.
Received and submitted signed-off original to DCC of "Attachment Number 7" of the MOU between Virgo and LIGO.
>From: Rod Luna <rluna@ligo.caltech.edu>
>From: Linda Turner - turner@ligo.caltech.edu>
>From: Cleveland Mak <mak_c@ligo.caltech.edu>
|
Week Ending 04/20/2006 |
In |
Out |
|
Packages |
23 |
13 |
|
Faxes |
25 |
17 |
>From: "Cronin, Holly" <Holly.Cronin@caltech.edu>
>From: "Brambila, Ruth" <Ruth.Brambila@caltech.edu>
>From: Florence Kaufman <fkaufman>
>From: Gina Salone <gsalone@ligo.caltech.edu>
>From: Ed Jasnow <jasnow@ligo.caltech.edu>
See Advanced LIGO below.
>From: Cindy Akutagawa <cindy@ligo.caltech.edu>
>From: Bill Tyler tyler@ligo.caltech.edu
A two-day site safety audit of the Livingston Observatory was successfully completed. The audit team found that the LLO generally exhibits a safe operation as evidenced by a good safety record. A few, mostly minor safety related items were identified by the audit team which need to be addressed. Despite recent hurricanes and on-going building construction, the team noted that the site "housekeeping" was quite good. A draft audit report will be distributed for comment during the next week.
No report.
No report.
See also the CDS weekly meeting minutes in the commissioning archives
Rolf Bork
4/13: Began looking into data corruption problems detected on fb1 at LHO. It look like about 90% of the errors are in data from the fast ADCU on H2. In fact, when the fast ADCUs were turned off over the weekend, no data corruption errors were detected. Alex enabled some diagnostics in fb1 and the fast ADCUs were turned back on during maintenance on Tuesday. Of course, no errors have been detected since (at least as of yesterday). The other change that was made Tuesday was that the Framebroadcaster net was moved to fb0 from fb1. Both Framebuilders run this task all the time, but only one is physically connected to the DMT network at a time. The machine that is connected is subject to retransmission requests from the DMTs, so it is still possible that those requests overloaded fb1 at times. Anyway, we will continue to watch. Barker is also making a list of control computer reboot times to see if the other 10% of errors found (those which include more than just fast ADCU channels) are correlated to that.
4/20: Continued to try and
diagnose data problems with fb1 at LHO.
Though we haven't done much except to add diagnostics, there have been
no problems noted this week. Dave Barker is writing up a summary of what has
been done and will continue to be done to monitor the data.
No report
Helena Armandula
2ITM04 arrived at LMA. Scatter map measurements of the central 80 mm area will start promptly according to Laurent Pinnard.
Bill Kells
Yet again we have new data
on the investigation of our contaminated ITM07 HR mirror. We (again, all the
hard work and implementation by L. Zhang) have re instituted (this had been
done previously with a less refined and less extensive scanning last summer)
simultaneously measuring absorption and scatter during scanning. Thus the
correlation between scatter and absorption may be studied. The importance of this has been brought out
from analysis of the recent absorption [alone] scans of ITM07. They indicate a very irregular,
"blotchy" absorption distribution. Unfortunately the density of these absorption "centers" coupled with the
resolution of our scans (~150 micron) is just not sufficient to conclude
whether there is any uniform film component to the contaminant. On the other hand, if all the absorption is
strongly correlated to anomalous scatter then we may conclude that the
contaminant is essentially "granular" in nature.
The correlation scans show
nicely that the absorption "centers" which dominate the total mean
absorption are significantly correlated with high scatter. This leaves the
puzzle of why no background pixels are found with absorption comparable to that
of cleaned surface scanned regions (and the fact that SEM scans show only
isolated grains of contamination).
One explanation is that the contamination was sub-micron "powder" (smoke?) or aerosol droplets. Deposited, these would have correlated scatter and absorption, yet individually be to small to "see" (at the ~150 micron resolution). The observed scan "blotchiness" would be the result of the "powder" deposition statistics.
No report.
This week we were still vented to air for the completion of the installation of the quad
I installed the classic damping feedback, I also improved it a bit to increase the phase margin while keeping the noise transmission low. I also installed the modal damping. Both controls match a 10sec settling time requirement and work fine. I then took some measurement you can see on the ILOG. I am going to spend the rest of the week working on the estimator and the modal control to see how to improve it or at least understand it better
Mark, Brett and Janeen, with support from Predeep, Ken, Rich and Myron are working on finishing up all in-vacuum tasks for the quad prior to a Friday deadline for closing up the BSC chamber. Earlier in the week, we spent a couple of days trying to determine if the optical table was level and after trying about 6 different methods, we determined that the data collected and the risks addressed pointed toward a level-enough table. We've replaced the magnets that broke off. The electrostatic drive feed through is installed and the wires installed in the connectors. They'll be hooked up tomorrow, as they'll be in the way today. This morning, we worked on aligning the global control osems on the penultimate mass, in X and Y. We'll spend the afternoon aligning the other osems. We will also get the eddy current dampers ready today. The electronics and cables are all hooked up and checked out. The osems are all reading correctly from the right places. An LED failed on the face 3 OSEM of the main chain, which still has to be replaced.
The initial results with the CEO laser amplifier were very disappointing. Significant beam jitter was observed on the transmitted output. It appears that this likely as a result of a mechanical instability of the rod in the laser head as opposed to fluctuations in the laser pump diodes. We next intend to determine whether this is caused by turbulence in the laser head or from pulsations from the chiller unit of the laser.
We are starting to obtain quotes for the clean room for the Y-end HAM, which is the HAM that the HAM SAS will be installed in.
Continued working on e2e modeling of Advanced LIGO Mode Cleaner triple suspension. By including the state space matrix generated by Mark Bartton's Mathematica code into the previously built e2e triple- suspension code and injecting white noise to the input port corresponding to suspension point yaw degree of freedom (DOF), computed transfer function from the suspension point yaw to all DOFs (x, y, z yaw, pitch and roll) at all three stages. The results are currently being analyzed.
The investigation of the in-lock state showed that the XARM is resonant not only on the carrier (as it should be and as it is the case for the YARM) but also on sidebands. This could cause a leak of few percent in the arm power: under studying. Implementing new LSC code for the in-lock state.
Melody, Hiro, Tom and Shannon worked together to resolve the problem in installing e2e at LLO. Some gnu software was too old, and software installation was placed in an unusual path. Now the latest simulation engine and GUI are available at LLO.
One new addition is a sticky note, which corresponds to comments in text based source codes.
Brown:
Chatterji:
Mendell:
I will be traveling to the APS meeting, April 22-25, to present preliminary Stacklide S4 results. I will be working to finish those results by June. I am also working to generate version 1 (v1) SFTs for the Einstein@Home S5 search using data from the beginning of the run up to the middle of March. I am also working to phase out v1 SFT generation, and get version 2 (v2) SFT generation running again at the sites. I plan to start publishing v2 SFTs by the end of next week, for transfer to CIT and tier II centers, after I return from the APS meeting.
Shawhan:
Sutton:
Most of my time this week has been spent on the xpipeline network analysis package. I implemented in matlab a function for producing Lazarus BH merger waveforms based on an analytic approximation in the original Lazarus paper (Baker et al PRD 65 124012). This form avoids the discontinuities that are present in the discretely sampled waveform supplied by Baker that has been used in bursts simulations to date. I finished the main improvements in power spectrum estimation in xpipeline. The new method is based on the LPEF (whitening filter) coefficients, and handles lines much better than the old version. I'm now testing cross-correlation timing estimates in xpipeline for narrowing down the region of the sky that needs to be tested for consistency with a GWB.
Yakushin:
LDAS
Work has started on timing out certain system calls to prevent deadlock situation. The first system function to have this timeout will be lstat in the diskcacheAPI.
Fixed and cvs changes for Utilities page rebuild frame cache function reporting error with missing variable ::cntlpasswd if globus connection is used.
Corrected the issue where cntmon was unable to reboot a utility if the utility's log file was missing (PR#3010). In addition, more debug purple logs to proc readsshPipe to detect why a script invoked by cntmon quit.
System tests were done using version 1.8.154. The tests depended upon seeing /archive. With its removal from the development system this week, the necessary frame files were moved to the /scratch file system.
The putStandAlone test script would sometimes fail to insert the correct lstring length in the ilwd for database insertion (PR#3012). This has now been corrected.
Reviewed the significant changes in the frameCPP code between the current version and the version being used in the framebuilder. These were reported to Alex and the LDAS group.
TCLGLOBUS
All unit tests for the GSS-API have been created and verified to work.
Unit tests for the XIO attributes have been completed (PR#6, PR#9 and PR#10).
An issue with the XIO Tcl Channel was discovered which prevented the cancellation of the XIO timeout. Correctly using globus_xio_attr_cntl() instead of globus_xio_handle_cntl() resolved the problem.
GRID COMPUTING
No production activity. Waiting on hierarchical inspiral pipe code from inspiral analysis team.
Reported problem with Condor-G/DAGMan/glibc2.3.6 to Condor team that causes DAGMan to abort. No response has been received from the Condor administrators in four business days. Will review costs and benefits of obtaining Condor source code until Condor team is willing to support Fedora Core 4. In parallel, a delegation from LIGO will address the support issues with Condor management at a meeting with the Condor PI on April 27-28.
Attended burst analysis meeting to explore feasability of supporting the burst analysis on OSG. There workflow is significantly less complex than the binary inspira and should be easily ported.
Worked with SRM team lead to clean up drm.rc configuration file for supporting SRM in OSG 0.4.1 release.
Attended OSG telecom on OSG 0.3.7 validation results.
Reviewed draft document outlining the validation of OSG software stacks for pre-release candidates. Gave feedback to authors addressing the phase of development for which this document would be most meaningful and reached agreement that this is a provisioning phase document as opposed to an integration phase document.
Lazzarini: A number of LIGO people will be
in
Caltech
(Dan Kozak)
(Phil Ehrens)
(Erik Espinoza)
4/13 node120 - Kernel Panic
4/13 node122 - Kernel Panic
4/14 node255 - Kernel Panic
4/16 node54 - Disk 2 died, replaced
4/17 node283 - System Powered off without warning.
4/18 node227 - Kernel Panic
4/19 node31 - Kernel Panic
4/19 node148 - Kernel Panic
4/19 node37 - Kernel Panic
4/19 node75 - Kernel Panic
4/19 node159 - Kernel Panic
(Stuart Anderson)
MIT
(Keith Bayer)
(Igor Yakushin)
(Dwayne Giardina)
(Greg Mendell)
I am generating version 1 (v1) SFTs from h(t) data from the beginning of S5 up to Dec. 17, 2005 at CIT. After Dec. 17, 2005, h(t) and v1 and version 2 (v2) SFTs have been generated at the sites. The v1 SFTs up to middle of March will be used in by Einstein@Home. I am also working to restart v2 SFTs generation at the sites, now that h(t) generation is beginning again, and phase out v1 SFT generation. The v2 SFTs following a specification that will allow them to be published and transferred from the sites to CIT and tier II centers using LDR. I should have this running by the end of next week (after returning from the APS meeting). Also, occational RLS errors during the LDR transfer of data causes some gaps the RDS data downstream. I am working to fill in these gaps, and ways to prevent them in the future.
(Ben Johnson)
(Keith)
(Dwayne)
(
(Christine)
(Veronica)
(Christian)
(Mike)
(Larry)
Mail Statistics for April 13-19, 06
|
Mail Statistics |
April 20, 2006 |
|
Rejected Messages |
17,474 |
|
Virus Messages |
1,814 |
|
Accepted Messages |
7,848 |
|
Total Messages |
51,672 |
Preparing for the baseline review; kicking off the
HAM-SAS fab effort.
From: Ken Mason <kmason@ligo.mit.edu>
Stage 0 assembly is complete and we continue to work on the assembly of stage 2. Progress has been slow due to the installation of the SUS quad prototype into the BSC chamber at LASTI. The test stand is now available to us for the overall assembly.
We are still waiting for several parts which make up the kinematic lock, spring, and actuator assemblies from Limerick Machine. These are the last parts needed and are not holding us up right now.
The Adv LIGO SEI Cost Estimate has been updated for the current baseline of (5) single stage HAM isolation systems. This estimate was based on shop quotes of detail drawings created by Corwin.
Richard Mittleman has measured several of the
stage 1-2 blades using the calibration fixture designed by ASI. We are finding
the springs slightly "softer" than predicted and are planning
From: Dennis Coyne coyne@ligo.caltech.edu
The NSF approved the
contract to Gailli & Morelli
for fabrication of the mechanical/structural components of the HAM-SAS system
which will be tested at LASTI.
From: Janeen Romie romie_j@ligo.caltech.edu
Working at LASTI on
getting all in-vacuum tasks for the quad completed prior to the noon deadline
tomorrow.
From: Ken Mailand kmailand@ligo.caltech.edu
I'm working on a lower
quad installation arm concept, and will send out an email for comments and
input. Its specifications and
limitations some based on earlier requirements in Oddvars
file.
From: Helena Armandula ahelena@ligo.caltech.edu
The next coating run with LMA will try to demonstrate that the Advanced LIGO absorption requirement of 0.5 ppm can be achieved on a doped Ta coating
On Thursday, April 20th we have a telecon
scheduled with CSIRO to discuss an optimized SiO2 doped TiO2 coating.
From: Bill Kells kells@ligo.caltech.edu
Yet again we have new data
on the investigation of our contaminated ITM07 HR mirror. We (again, all the
hard work and implementation by L. Zhang) have re instituted (this had been
done previously with a less refined and less extensive scanning last summer)
simultaneously measuring absorption and scatter during scanning. Thus the
correlation between scatter and absorption may be studied. The importance of this has been brought out
from analysis of the recent absorption [alone] scans of ITM07. They indicate a
very irregular, "blottchy" absorption
distribution. Unfortunately the density of these
absorption "centers" coupled with the resolution of our scans (~150
micron) is just not sufficient to conclude whether there is any uniform film
component to the contaminant. On the other hand, if all the absorption is strongly
correlated to anomalous scatter then we may conclude that the contaminant is
essentially "granular" in nature.
The correlation scans show
nicely that the absorption "centers" which dominate the total mean aborption are significantly correlated with high scatter.
This leaves the puzzle of why no background pixels are found with absorption
comparable to that of cleaned surface scanned regions (and the fact that SEM
scans show only isolated grains of contamination). One explanation is that the contamination was
sub-micron "powder" (smoke ?) or aerosol
droplets. Deposited, these would have correlated scatter and absorption, yet
individually be to small to "see" (at the
~150 micron resolution). The observed scan "blotchyness"
would be the result of the "powder" deposition statistics.
From: Peter King pking@ligo.caltech.edu
I spent some time in trying to get the Xilinx system generator software to talk to Simulink. After some time, I was able to use simulate a filter design in software but encountered some problems in getting the hardware co-simulation completed. The error messages and logfiles from the Xilinx software are not exactly informative. For example: "ERROR: Xst:1817 - Invalid target speed '-10'" was the result of using a pull-down menu to compile code for the DSP. The option to change flags manually is not provided for. After going through a bundle of documentation, the meaning of the error message is that the device files were not installed correctly and that a separate software package should have been installed before the system generator software was installed. Naturally enough, there was no mention of the installation order in the installation documentation.a Not that it made any difference since the problem is still present even when the software is installed in the correct order. From the Xilinx support database, other people have experienced.
From: Michael Smith
<smith@ligo.caltech.edu>
Chris Echols, a new
mechanical engineer on the LIGO project, is in the process of updating the BSC
and HAM chamber solidworks models for the ADLIGO
vertex layout assembly.
The Zemax
optical layout has been revised with the output beams exiting from HAM2 (moved
in front of HAM1), and HAM5 (moved behind HAM6).
After discussion with Phil
Willems, I modified the wedge angles on the ITM
compensation plate and the ITM COC as follows: 1) the ITM has a symmetric wedge
of 1.074 degrees on the AR and HR sides, for a total wedge of 2.15 degrees. 2)
The compensation plate (CP) has a 5 min. wedge to eliminate etalon effects. The
ITM PO beam is taken from the AR surface of the ITM. This was done so that the
first surface of the CP and the HR surface of the ITM are essentially
parallel--this allows the use of the Russian interferometric
technique for sensing the thermally induced aberrations in the ITM optical
assembly as an alternative to the Hartman sensor, if desired.
The AOS WBS dictionary is
updated.
From: Rolf Bork rolf@ligo.caltech.edu
Continued work on our Simulink code generator. I've had to add support for more
features in order to be able to produce the code for the Lasti
Hepi system, which is far more complicated than the
quad suspension code. The key items remaining are support for the watchdogs and
polyphase FIR filters. I hope to have the watchdog
ready by Tuesday and Alex is working the new CDS standard FIR filter feature.
The latter may or may not be ready for the planned Lasti
install the week of May 8. If it is not
ready, we will stay with the hardcoded version used
in Ligo for now.
One possible concern is
that there was no watchdog feature defined for the quad suspension controls or
a connection of that signal to the new Hepi system.
I've asked Barker for info on how the suspension watchdogs work on the Ligo aux crates as a starting point. I will still need to
work out with someone how the quad watchdog should be implemented. Communicating that trip signal will then still need to be implemented.
We have done tests communicating between front ends on the new network, but,
other than between front ends and the Framebuilder, have
not yet implemented a standard communication feature in our software (haven't
had time).
The computer and PCI-X I/O
chassis for Lasti Hepi are
due in tomorrow or Monday. The back panel necessary to modify the I/O chassis should
be ready by then as well.
I ordered a GPS network
time server that also has a 1pps output. This unit will replace the existing VME GPS at
Lasti, with the 1pps signal used to drive the new
timing system. The network connection will be used to time synchonize
all of the computers on the CDS network.
Since we plan to remove
the remaining VME equipment on our next visit, we need to replace the existing
VME ADCU. Jay is putting together a new interface chassis w/4pin LEMO and BNC
connectors to interface to our new PCI-X ADC modules. This unit will connect
the existing PSL signals to our new system and provide a number of BNC
connections for general purpose use.
Had a telecon with Joe Giaime and Brian
Lance about the ISI installation at Lasti. Brian is to get me his Simulink files to start working out the controls software.
We are also trying to figure out what the schedule is for ISI controls
installation. I believe that was deferred until our next meeting on Wednesday
while Joe tries to sort that out.
We have not had a lot of
time to devote to the Ponderomotive experiment
controls. Alex had a first cut at the code running. It barely makes the 64KHz minimum cycle requirement (running on one cpu). We are attempting to further split the code over
multiple CPUs, but need to work more on the synchonization
issues (the synchronization semaphores seem to take too long ie 1usec). The 1usec may not seem that long, but when we
want to get to 128KHz (<8usec cycle time), every usec counts. So, we
are continuing to look at ways to save time. One quick way would be to turn off
the 16Hz decimation filters associated with every CDS filter module (for EPICS
output). When we run several hundred filter modules, with each one having this
feature, it consumes several usec (about 4usec in the
quad controls, which has 96 filter modules).
We have been given the go ahead to order the equipment necessary for the DC readout system at the 40m lab. I haven't yet had much time to devote to this. I have ordered a Sun V40, but not much else.
For additional information about this report, contact Stan Whitcomb or Phil Lindquist