The LIGO
Executive Committee meeting for March 14, 2005 is cancelled because of the
Staffing Committee meeting.
S4 continues, with good duty cycles over the last week - H1
(84.3%), and H2 (75.1%), L1 achieved a very high duty factor over the weekend,
including one day where science mode fraction was 98%, L1 currently running at
~7 Mpc inspiral range
No report.
LSC MOUs and Research Plans and Progress Reports
For a web page summary showing the status of LSC MOUs and associated
Attachment updates see:
http://www.ligo.caltech.edu/~irena/Revstatus/Reviewstat.doc
Non-LSC MOUs
A site teleconference was held on Thursday, March 3, 2005. The following
items were among those discussed:
>From: Ed Chargois <chargois_e@ligo.caltech.edu>
>From: Linda Turner - turner@ligo.caltech.edu>
>From: Cleveland Mak <mak_c@ligo.caltech.edu>
>From: Esther Cunningham <esther@ligo.caltech.edu>
>From: "Brambila, Ruth" <Ruth.Brambila@caltech.edu>
>From: Gina Salone <gsalone@ligo.caltech.edu>
>From: Florence Kaufman <fkaufman>
>From: irena@ligo.caltech.edu (Irena Petrac)
>From: Ed Jasnow <jasnow@ligo.caltech.edu>
>Irene Baldon
>Dorothy Lloyd
Nothing new to report.
The DCC Steering Committee met the morning of March 9, 2005 pacific time. Discussion centered on the recently released DRAFT requirements document and potential vendors. A pointer to the requirements document has been provided to members of the Executive Committee. Several interesting vendors are being explored.
No open change requests.
>From: Cindy Akutagawa <cindy@ligo.caltech.edu>
>From: Bill Tyler tyler@ligo.caltech.edu
No report.
S4 continues apace, with a higher H1 duty cycles over the last week - H1 (84.3%), and H2 (75.1%). H2's was impacted by an FMY optical lever problem (see below). As usual, some lock losses correlate with earthquake activity. A typical suite of injections was introduced on the interferometers; burst injections are here reviewed. Injections of fake pulsars injections were started (continuously). This elog underscores however that the injection process is not yet bulletproof. Fast channel calibration was investigated.
Roughly 50 Pasco High Geometry students visited on 3/2 under the auspices of MESA. Mike Landry presented a World Year Public Lecture at Triumf in Vancouver on 2/26. Fred Raab spoke to ~90 members of the Richland Rotary on 3/3 and to a meeting of the Majorana Collaboration on 2/26
We achieved a very high duty factor over the weekend, including one day where
our science mode fraction was 98%. The only interruptions were due to
large earthquakes and planned commissioning activities.
On the evening of Wed. 9th we finally got back to high power running. Diagnostic measurements seemed to indicate that the problem was
in the common mode board. A switch to a spare showed this not to be the case. The
problem was eventually traced to a bad slider on the MC board (#MZ note: a
subtle fault which did not appear in standalone testing, but only manifested
itself in the fully integrated interferometer!). This was fixed by inserting a
voltage divider and using a different slider setting.
On Thur. 9th we had a downtime to install a new Wenzel crystal oscillator.
This had a dramatic effect on our phase noise bump at ~ 2 kHz. We also replaced
ASPD2 with a modified version, the old
photodiode was twice as susceptible to AS_I saturation.
From Fri. 4th until Wed. 9th we had a wandering line feature in our noise
spectrum. This was traced to the lsc front-end crate. Swapping the pentek ADC
did not fix the problem, which turned out to also be visible in the dark noise.
We noticed that it was acoustically sensitive, and sensitive to blocking fans
at the back of the crate. The problem disappeared while we were diagnosing it,
apparently due to cleaning the crate fans of lint buildup (or perhaps because
our activities fixed some other intermittent acoustic coupling).
Currently we are running at ~7 Mpc inspiral range and we anticipate
no further interruptions for commissioning activities.
#MZ Note: This recovery was technically challenging (you have to admit,
those hardware problems turned out to be pretty obscure!!), but the method also
presents a good model for performing measured and coordinated repairs during
science operations. Each step was thoroughly prepared in advance by the team
and downtime was coordinated with LHO, giving a minimal sacrifice of triple
coincidence... and substantially improved performance for the remainder of the
run. Great team effort!
In the process of digging through logs and connection status of the firewall
I found a windows machine that was infected with some form of windows virus or
worm. This machine does not belong
to LIGO. The owner of the machine
has been notified and is taking care of the situation. The Norton install was expired and had
not been updated in a while. It was
scanning the net for port 445 and had > 1700 open connections when I found
it.
Working with Syskonnect on a PCI express card that is not working. The engineers in Germany were able to
duplicate the problem and are working on it.
Made several more changes to the firewall. It has turned out to be quite useful to have a linux
firewall instead of Cisco. I have
been able to do many things with it that I was unable to do with the Cisco hardware. Performance is also better.
Still trying to find a pgp/gpg package for windows that will work reasonably
well for "non-technically inclined" users. I have evaluated several and have not found one that works
easily and is compatible with existing tools in use yet.
Purchased the media kit from Sun for the EduSoft packages. It comes with Sun Studio 9. Unfortunately, Daniel Sigg needs Sun
Studio 10. I am working with Sun
on licensing for it. A site
license is only $100.
Looking into several Sun software packages that may be useful.
Started working on a way of scripting Solaris patch checks under Solaris 9.
no report
1) Last week we had a telecon with IPG. They reported that the system is cleaned
up, new diodes are installed and a fiber block has been installed. The system
now gives 80 W. They still need some time to install new and hopefully better
diodes and making some other optimizations before going up to 100 W. We
requested that they test the system for seven days instead of only 24 hours
before shipping which they agreed to do. They again mentioned that if we change
wavelength from 1064 nm to 1070 nm or even better 1075 nm the system would be
much more robust. We asked them to give the 1064 nm another shot since it is much
better for our purpose and what we have paid for. They will keep us updated on
the progress. We are expecting to get news tomorrow Friday March 11.
2) The Melody Advanced LIGO mode cleaner model with beam injection through
the curved mirror clearly exhibit a great deal less of thermal distortion in
the reflected beam even at high power. I need to verify the validity of this
before going on but it looks very promising. Please find some results at http://www.ligo-la.caltech.edu/~franzen/melody/AdvLIGO_MC_m3_injection.html.
The UF group is discussing if and
how we would change the current MC design in accordance with these results.
1) Waveburst online pipeline including trigger production with 100+1 time
shifts, triple coincidence, plotting figures of merit, publishing triggers in
ROOT and ASCII formats (ROOT files are made browsable under Carrot so that one
can quickly plot various pictures using the web browser) and comparing loud
events with hardware injections is fully automated and running great at CIT: http://ldas-pcdev1.ligo.caltech.edu/~igor
Two styles of analysis are performed online:
a) High threshold analysis in which the threshold
on geometric significance is set so high that only a handful of the most loud
triple coincidence events survive (half of S4 generated only 110 such events so
far and most of them have already been explained out by hardware injections).
For such an analysis we look both at zero lag and time shifted triggers.
b) Low threshold analysis in which we do not look
at zero time lag but only use time shifted events to study the properties of
the background. Made a presentation about the waveburst online status at the
burst group telecon this week.
2) Finished production of S3 triggers to be used for the final analysis (contrary
to the ones with the different time shifts used for tuning the thresholds);
3) Working on incorporating software simulations into the online waveburst
infrastructure;
4) Communicating with the potential SURF student concerning the possible research
projects for summer;
5) Redoing 'time between consecutive events' picture for S2 paper in black
and white on Erik's request.
6) Currently online waveburst infrastructure provides triggers and plots either per job or for the whole run up to now (actually waveburst is configured to keep about 6-8 hours behind the real time). Working on providing daily figures of merit and statistics.
see also the CDS weekly meeting minutes in the commissioning archives
Alex is working on the PC104-based epics system for monitoring the vacuum signals from LLO mid-stations. This should be done in the next couple of days.
I've somewhat strayed from my post S4 list for the past couple of days.
Since our real-time front end code is pretty modular now, I wanted to see if I
could write a perl scipt to generate it. We already do this for generating most
of the EPICS code, which communicates with the front ends. As a test, I am
trying to generate the asc code for the 40m lab. So far, I am able to:
In this fashion, I have been able to generate and run a basic front end system. Something like this could be useful for building LIGO-like test stands. It could also be packaged into a PC running Linux, with some PCI I/O cards, for an inexpensive front end for test systems with LIGO-like capabilities and connections to LIGO Framebuilders, etc. Unless there is some interest in this, I only plan to work on this for another day or two to see if I can fully replicate the 40m ASC code using this method. If nothing else, this exercise has shown where a few more generic code modules would be useful in our front end code.
Along these lines, Alex put together what I call "DAQ in a box" for use at the TNI. This is a Linux PC with an ADC module installed. The PC has a Linux version of Framebuilder installed and code to read out the ADC module. The ADC module is triggered by an external clock. So, it kind of combines our ADCU and Framebuilder into one unit. It has full Framebuilder functionality, so it saves data as frames and provides connections to the standard LIGO tool set, such as DTT. So, again, it could provide a LIGO-like environment for test stands without the expense of VME and high end workstations.
no report
After the recent power outage at LLO, there was a small suspicion that the high voltage power supply for the reference cavity ion pump might be faulty. A check with the new manufacturer suggested that the LED display was faulty but that the power supply itself was okay. During the process I learnt that Physical Electronics sold off their vacuum equipment lines and that Gamma Vacuum now sells them.
OTF Lab. (W. Bridge) Ready
This chamber still has two samples, white Ceramabond, and disks of TRA-BOND #2254 color light brown epoxy. Waiting on the new mirrors to be tested for contamination.
Absorption Test Measurement prototype in standby
Scatterometer system in progress
The Quantronix 60 watt laser shut down
OTF Lab at Lauritsen ROOM 38
Cavity #3
Cavity #2 in progress
No report.
Ken reports:
A design has been completed and all parts ordered for a structure to support
the BSC down tube. This will be used to isolate the down tube from the support
table during pier amplification and other dynamic testing of the BSC structure.
I am currently working on the design of the solid stack for the BSC chamber
for testing the quad prototype at the advanced ligo height. The aim is produce a structure with a
resonance above 50 Hz.
Myron has been Cleaning and staging of equipment (walk platforms, chamber floor,
optic table support, cartridge clean room, etc.) for the BSC chamber work
coming up.
Rich reports:
The BSC is now under HEPI control (again) and is currently taking data to be
used in the design of sensor correction filters. We had more than the usual
amount of cross coupling, also during the day the ground noise at MIT is high
enough that the valves on the actuators are operating at or in their non-linear
range, I don't know how these two things couple.
Using the geophones on the BSC optics table we measured the transfer function
from the support table (HEPI geophones) to the optics table. This should be
useful in the design of HEPI resonant gain.
Laurent is visiting his home institution in
France to develop the necessary modeling tools for the upcoming LASTI work.
Jon is repairing our SOS controllers and
SR560s
Mark Barton talked about
"Suspension Modeling in Mathematica". The physics behind various
tools and two major families of models were discussed. Those are: GEO-style
triple pendulum and AdvLIGO's quad pendulum. These models are used to generate
some numeric state-space matrices which are efficient for time-domain simulation.
These matrices are then exported to modules in e2e simulation set-up.
Viewgraphs available at: http://www.ligo.caltech/~e2e/ME2ET/Minutes05/050310/.
(Biplab) Worked on analysis of
the limitation of phase-camera set-up with its pixels distributed on a spiral coordinate and how
that affects the determination of
modes in a beam.
(Monica) I validate my
analytical calculations with Mathematica and compared them with those done by Hiro and Osamu for the Dual
Recycled Michelson Cavity. The
comparison looked satisfactory. I'm still debugging the e2e code for only Signal recycled interferometer with
Fabry-Perot cavities in arms:
Numerical values of the fields at the symmetric and asymmetric
ports for minor signals will be
checked.
(Hiro) Worked on modeler
programming issues.
(Bruce)
(Melody) Working on fixing existing Problem reports.
Charlton: (A.L. NOTE: Philip Charlton has returned from his teaching stint
in Australia; he will be with us for a quarter, then moves on to the Cardiff
GEO group to work on stochastic background searches with Joe) - started looking
at how to modify the stochastic data analysis pipeline to work on the GRID
Mandic:
I performed several calculations needed for the S3 paper on stochastic analysis.
In particular, I combined the H1L1 and H2L1 results, and I determined the
frequency range containing most of the signal-to-noise for this analysis.
I worked on the code for studying time-shift effects in the stochastic analysis.
With Peter Shawhan, we worked on the hardware injection code, and have successfully
injected a number of different waveforms.
Mendell:
Work continues on making further comparisons between the Hough, StackSlide and PowerFlux codes and
preparing the StackSlide code to generate example preliminary S3 results for
the March LSC meeting.
Shawhan:
Sutton;
This week I finished the LIGO-TAMA untriggered bursts paper, P040050-02,
which has since been distributed to the full LSC for comments.
Weinstein:
Yakushin:
1) Waveburst online pipeline including trigger production with 100+1 time
shifts, triple coincidence, plotting figures of merit, publishing triggers in
ROOT and ASCII formats (ROOT files are made browsable under Carrot so that one
can quickly plot various pictures using the web browser) and comparing loud
events with hardware injections is fully automated and running great at CIT: http://ldas-pcdev1.ligo.caltech.edu/~igor.
Two styles of analysis are performed online:
a) High threshold analysis in which the threshold on geometric significance
is set so high that only a handful of the most loud triple coincidence events
survive (half of S4 generated only 110 such events so far and most of them have
already been explained out by hardware injections). For such an analysis we look
both at zero lag and time shifted triggers.
b) Low threshold analysis in which we do not look at zero time lag but only
use time shifted events to study the properties of the background. Made a presentation about the waveburst
online status at the burst group telecon this week.
2) Finished production of S3 triggers to be used for the final analysis
(contrary to the ones with the different time shifts used for tuning the
thresholds);
3) Working on incorporating software simulations into the online waveburst
infrastructure;
4) Communicating with the potential SURF student concerning the possible
research projects for summer;
5) Redoing 'time between consecutive events' picture for S2 paper in black
and white on Erik's request.
6) Currently online waveburst infrastructure provides triggers and plots
either per job or for the whole run up to now (actually waveburst is configured
to keep about 6-8 hours behind the real time). Working on providing daily
figures of merit and statistics.
The 1.5.1 release of LDAS went out this week. It has a fix to allow merging
non-aligned LLO and LHO frames into a single RDS frame in the createRDS code.
This came about from the expanded feather in the 1.5.0 release of allowing
shorter frames to be automatically generated when data dropouts in the
instruments occur. The 1.5.1 release is now running at MIT, CIT and TEST. LLO
and LHO are free to upgrade when it doesn't interfere with the science run. But
since LLO and LHO do not have each other's datasets at this time, the upgrade is
not significant to their sites. NOTE: LDAS is now running again at MIT.
The stalled LDAS-MIT user accounts that collected during the MIT down time
were updated on the LDAS-MIT system.
Successfully advanced the LDCG and LDAS code to run with the new GCC 3.4.3
compilers on the Tandem 2 LDAS system. Preliminary testing indicated that this compiler
is about 10% better for the overall performance of LDAS on Linux systems. Don't
have a comparison yet for Solaris hardware improvements and expect this will
have to wait until the migration of the code base onto the LDAS-DEV system in a
few days. Also, with this new compiler, we are attempting to go back to the STL
library provide by GCC instead of having to use the STLPORT version of the STL
libraries due to better thread support in STLPORT. Conclusive results with the
GCC STL libraries will have to wait until the code has been challenge tested on
the LDAS-DEV system.
Added the fix to CVS to all creation of 100% of the raw frame channel lists
in RDS frames. The 20,000 or so
channels were causing an FrHistory overflow when this operation was attempted
in the 1.5.0 and 1.5.1 code base. The fix is running on the LDAS-DEV system and
will be in the 1.6.0 release.
Upgraded the LDAS-DEV and LDAS Tandem systems to the new DB2 8.2.1 release.
In our unit tests, this release of DB2 demonstated nearly a 50% improvement in performance.
NOTE: Also upgraded MIT to the
required 8.1.5 release of DB2 since they had not upgraded LDAS since the 1.3.0
release.
With the LDAS CVS repository reopenning this week, several watershed fixes
to the control-MonitorAPI and the genericAPI were checked in and have now been
tested on the LDAS-DEV system.
Development has progressed in the last week to the point that only 4 out of
89 Globus FTP Client functions are remain buggy:
Currently have developed tests for the following six Globus FTP Client functions
to emulate FTP client/server data movements.
Began working on Perl scripts to automate publication of the results of individual
test cases on the web from nightly build.
Now have Ganglia configured and running on the three OSG worker nodes. There
was an issue that was discovered due to the LDAS-CIT and LDAS-TEST cluster being
networked across the switch which was causing the LSC Data Grid to report OSG worker
nodes in its private ganglia. This was resolved by changing the default
broadcast to a specific multicast on the OSG nodes.
Have successfully configured MonALISA on the OSG gatekeeper. Our OSG site
now has a green dot on the MonALISA homepage.
Presently working on bringing up the condor batch system on the three OSG
nodes. There have been several issues associated with sharing these nodes with
LSC Data Grid which are slowing down the completion. Learning how to install
and run condor on the non-standard "condor" unix account being chief
among these.
(Dan Kozak)
(Phil Ehrens)
(Stuart Anderson)
(Keith Bayer)
(Igor Yakushin)
(Greg Mendell)
(Ben Johnson)
(Keith)
(Shannon)
(Christine)
(Mike)
(Veronica)
(Larry)
From: "Thomas Frey" <tfrey@ligo.caltech.edu>
No report this week.
From: Ken Mason <kmason@ligo.mit.edu>
We have received a final DVD containing the final released documentation
from ASI.
Oddvar Spjeld, Dennis Coyne and I have been reviewing the documentation. We
expect to close out the purchase contract with ASI next week.
The BSC seismic isolation design package is complete. All models and
analysis files have been opened and checked. The fabrication and assembly
drawings are ready to send to suppliers for quoting and fabrication.
The tooling and fixturing required to assemble the BSC seismic isolation
system is approximately 60% complete. LIGO engineers will complete this design.
The HAM seismic isolation design is approximately 30% complete. Models and
analysis exist for the stage 1 and stage 2 subassemblies. The pods, actuator
assemblies and spring assemblies are mostly finished. Stage 0 and all
fabrication and assembly drawings have not been started.
All actuators are scheduled to be shipped from PSI today.
From: Janeen Romie <romie_j@ligo.caltech.edu>
Russell Jones joined us for the last 2 weeks to work on the quad. His
participation was greatly appreciated. We look forward to Tim Hayler and Ian
Wilmut from RAL next monday.
I submitted an updated SUS budget to Carol and Dennis.
Got the tablecloth modeled and, with Calum's help, meshed. Worked with Calum
and Russell on the earthquake stop design concept for the quad.
Mike Pedraza had to do lots of work on my desktop but hopes he's found a
solution to the problems I was experiencing with Solidworks, PDMWorks, Symantec
and Algor.
Worked some on the modeling wanted by the LIGO movie folks.
Sent around a preliminary schedule for the SUS breakout session on Monday,
March 21 st at the LSC.
From: ctorrie <ctorrie@ligo.caltech.edu>
Russell Jones and Norna Robertson visited Caltech last week. As part of
their visit we, along with Janeen, went through the steps involved with the
"3and1" assembly technique. This involved the upper intermediate,
penultimate and test mass in a mock up of the lower structure. This proved
extremely useful and allow us to test a whole series of assumptions. At the end
of this build we suspended, admittedly briefly, a triple pendulum from a fixed
top mass.
These photos and more are available on my webpage at: photos
Janeen (with some support from me) have been looking at the resonances for
the tablecloth, more to follow.
I have been looking at, with Mike Perreur-Lloyd, extending the use of drum-ended
wires to the upper intermediate and penultimate stages in the suspension.
Ian Wilmut and I have been looking at the blade data and have now selected
blades and associated clamps for the controls prototype.
Valley Precision, a local machine shop, delivered a set of fixtures that
will be used to aid in the assembly of the cantilever blades in their suspended
stages.
CES has now finished all of the suspended parts, we received the penultimate
reaction mass this week, and will now move on to making us 2 upper structures.
One for the controls prototype and the other for coupled tests at Stanford.
The Physics shop are working on the housing for the drum ended wires, clamp wire
clamp assembly jigs and the top stage that interfaces between the suspension
and the structure. This will all be competed in time for next weeks visitors.
Tim Hayler and Ian Wilmut are visiting next week from Rutherford Appleton
Laboratory.
Oddvar, Janeen, Ken Mason and I are working on preparing files from both
initial and advanced LIGO for Tom Lucas who is making the LIGO movie. As part
of this we should get some very useful free renderings.
From: Ken mailand <kmailand@ligo.caltech.edu>
I'm currently learning solidworks, and working on the lower suspension
installation arm to move the lower suspension into the BSC chamber. I’m doing a
more detailed design of the arms, and pivot pin bearing housing. The first
algor stress analysis indicates no problem stresses or deflection issues on the
preliminary design.
Also I have finished designing a platform for the modules for the BSC
chamber, one assembly is in process at the CIT shop, and should be finished by
Friday.
I sent 3D views of the 10x10x3 capacitive position sensor chassis to people
at LHO, and Oddvar at LLO for comment on the clearance to the top flange that
John Worden was concerned about.
The preliminary cost estimate for this bake oven/cleaning station is
~$45,000 +shipping. I will try to
get two other vendor quotes for comparison.
From: Helena Armandula <ahelena@ligo.caltech.edu>
Developing and characterizing a new mirror cleaning process.
Assembling magnets, holders and flags by cold indium welding. Testing and
modifying process.
From: Bill Kells <kells@ligo.caltech.edu>
MOstly ill this period, but we are anticipating a meeting on the AdL
beamsplitter geometry next week, so I am working on considerations for that. So
far I am working through the scaling of what diffractive scatter loss might be
considered acceptable from naïve scaling from LIGO I (where the amount we do
have is presumably now known to be acceptable).
From: Peter King pking@ligo.caltech.edu
The high power photodetector is still going strong after about a month now. No noticeable damage so far, although
the unit is warm.
A board for the next revision of this design was laid out.
From: Michael Smith <smith@ligo.caltech.edu>
No report this week.
From: Riccardo DeSalvo <desalvo@ligo.caltech.edu>
Marco
Mesa Beam Cavity blades were cooked at 480oC with the fixed oven. The
previously tested maraging samples, from Janeen stock, non A&D, now are
cooking at 435oC. The hardening process studies and maraging treatments should
be finished before the LSC meeting.
Creep experiment: we are taking measurements at 40oC. Observed a “fall” of
0.18 mm for the equilibrium point, probably due to the mechanical set up of the
experiment
Debugging piezo driver circuits.
Lab cleared of asbestos, but still to be rehabilitated.
Juri
I found a problem in my simulation routine: the automatic length control
intended to achieve the resonance condition in the cavity wasn’t working as
expected (my results of the previous week were not correct…). I fixed this
problem and now it is working well, I am debugging further and making more
investigations before going to the 2D mirror case and especially to Mexican Hat
mirror shape.
Justin
Measuring modes in the Q setup.
For additional information about this report, contact S. Whitcomb or P. Lindquist