|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
The LIGO Executive Committee Agenda for Monday September 15, 2003 will be:
(Meeting time: 10:30 am Pacific Time)
Open meeting 10:30 - 11:30
Special Items: S3
readiness
no report
LIGO Weekly Site Telecon (Jasow)
A site teleconference was held on Thursday, September 11, with the following subjects discussed:
PROPERTY MANAGEMENT (Chargois)
From: Ed Chargois <chargois_e@ligo.caltech.edu>
DOCUMENT CONTROL CENTER (Turner, Mak)
>From: Linda Turner - turner@ligo.caltech.edu>
Web pages for the DCC give simple how-to's for document numbering, easy access to the latest on-line documents, and search capabilities for the DCC database. Take a look. . .
ACCOMPLISHMENTS
Organized, sorted and filed documentation.
|
09/11/03
|
Packages
|
Faxes
|
|
In
|
24
|
15
|
|
Out
|
18
|
35
|
Press
here to access the DOCUMENT
CONTROL CENTER WEB PAGE.
COST SCHEDULE CONTROL SYSTEMS (Cunningham, Brambila, Kaufman, Salone)
From:
Esther Cunningham <esther@ligo.caltech.edu>
Press
here for ACCOUNTS
PAYABLE HISTORY DATA .
From: "Brambila, Ruth" <Ruth.Brambila@caltech.edu>
Financial reports can be found at: http://docuserv.ligo.caltech.edu/~fireport
SUBCONTRACTS MANAGEMENT (Petrac, Jasnow)
From: irena@ligo.caltech.edu (Irena Petrac)
SUPPORT (Baldon, Lloyd, Tischler)
>Irene Baldon
ADVANCED LIGO (Cost Schedule Control Systems) T. Frey
From:
Thomas Frey <tfrey@ligo.caltech.edu>
Progress
Period from 09.05 to 09.11
Accomplishments:
For list
of documents that are being used to develop Adv. LIGO Cost and Schedule,
see
http://www.ligo.caltech.edu/~tfrey/Cost_MTG_082002/
Continued development of a proposal for implementing Primavera Project Planner Enterprise.
Continue testing of Primavera Enterprise version 3.5.1. See the links below for information / technology demonstrations:
Primavision - http://primavision.drmcnatty.com:8080/phoenixwl/login_p3ec Login using Username: admin and Password: admin.
Progress Reporter - http://www.projwebsite.com/GroupServerP3e/en/ Login using Username: admin and Password: admin.
Static HTML Example - http://ligo.caltech.edu/~tfrey/SUS_WebProj/ (AdL Suspensions Only) NOTE: This is after importing and before any data corrections.
Static HTML Example - http://ligo.caltech.edu/~tfrey/DAQ_WebProj/(AdL DAQ Only) NOTE: This is after making adjustments to the WBS so that it displays as expected.
Started research regarding how MS Project 2002 compares to Primavera Project Planner, as requested by Gary Sanders and Richard Fischer.
Met with Carol and discussed the project control system for Adv. LIGO.
Attended Carol's presentation of the DARHT project.
Started work inputting additional coding as requested by Carol to identify deliverables, milestones, and interface activities.
Attended meeting with Dennis C., Carol W., and Janeen R. regarding AdL SUS.
AOS - Still need completed WBS dictionary and BOE.
IO - Still need completed WBS dictionary and BOE.
SUS - No action items pending. I will make a first cut of schedule changes for Janeen to review.
SEI - No action items pending.
ISC - Still need completed WBS dictionary and BOE.
LDAS - No action required at this point in time.
INSTALL - No action items pending.
DAQ - No action required at this point in time.
PM - No action required at this point in time.
FAC - No action required at this point in time.
SUP - No action required at this point in time.
Reports (Lindquist)
We are working on two proposals and work plans:
Change Control/Contingency (Lindquist)
The following change requests are open.
|
CR-030015
|
FY 2003
Livingston Observatory Detector Maintenance Expenses (Increment)
|
R. Wooley
|
July 14,
2003
|
|
CR-030016
|
Hanford
Facilities 2.2--Divide the Large Equipment Access to Facilitate Movement
of Large Items (currently assigned to FY 2004 liens list)
|
J. Worden
|
July 31,
2003
|
Human Resources (Akutagawa)
From: Cindy Akutagawa <cindy@ligo.caltech.edu>
Summary of Commissioning Activities at LIGO Hanford Observatory (compiled by M. Landry)
Studies performed on the output table old periscope design indicate
considerable
supression of internal modes with the addition of a damper
. Foam and steel plates
were affixed to the back surface of an old periscope and internal modes
were
measured with an accelerometer. Accelerations at resonant frequencies
were
reduced with dampers by factors of a few to ~35.
2K IFO
As was recently done for the 4k, the 2k acoustic shield (meatlocker)
was
placed over the AS output port ISCT10. Old periscopes (without
dampers)
are currently being used on a new
optical layout .
A malfunctioning fan allowed the X mid-station to heat to nearly 80
degrees .
Subsequent alignment drifts in the ETM hampered locking efforts for
a day.
SURF student Evan Goetz's project analyzing the carrier and sideband
content
of the 2k POX beam is summarized here
.
4K IFO
Odd behaviour noted on the 4k this week includes the secular drifting
of RF phase
as a function of thermal heating, as seen
on WFS2 . Once the ifo locks and SPOB
begins to grow as the thermal lens forms, the ASC then knocks the 4k
out of
lock at a SPOB of ~280. In the linked study, WFS2 phase is observed
to change
108 degrees as the thermal lens is established. The state of
the WFS in high
power mode is summarized here
.
Another observation that may be cause for concern is the reduction
in sideband
buildup
produced by the thermal lens. Daniel finds the effect of the thermal
lens appears to have decayed from a factor of 2.1 to 1.6 in the last
14 weeks.
The new MC board was installed and made operational, however, it has
the side
effect of unlocking the reference cavity and PMC (most of the time)
when
the full ifo lets go.
Switching in the analog dewhites has induced transients in ETMY, such
that
for the last few days, we've had some difficulty getting into full
CM mode.
We want to sort this rapidly as we have to re-establish robust locking:
the
mini run (M3) will occur this Saturday Sep 13th, 6pm-midnight.
The run
is predominantly to test DMT monitors and the state of DMT infrastructure.
DAQ
Framebuilders fb0 and fb3 received sundry upgrades
: new operating system,
software, location. An fb3 T3 disk space reduction was postponed until
later this month.
IFO Commissioning:
------------------------------------------------------------------------
After a week struggling with bizarre laser behavior, we replaced our
pre-mode cleaner (PMC) cavity on Saturday. The fault observed with
the
previous unit seemed to defy explanation (it was somehow generating
a
counter-propagating mode internally) but everything seems normal after
the swap. We now believe with some tweaking we can get over 6 watts
into
the mode cleaner, pretty near the design value (though it will take
some
time to learn how to use that much power). Lee Cardenas has graciously
sent a new spare PMC from Caltech to cover us in case of any further
mishap.
Flavio Nocera and Ben Abbott are here this week installing updates to
our intensity stabilizer (ISS), including the new lower noise sensing
photodiodes. The updated system is operating robustly and they are
now
cabling up new DAQ connections and testing performance.
We completed installation of internal absorptive baffles and then
hoisted the new acoustic enclosure into place over ISCT4. We are hoping
it will provide a spectacular noise improvement like that posted for
H1
last week. The interior is eerily quiet.
We've had continuing difficulties with our common-mode servo. We found
and fixed a sign error and a wiring error, but there was an electronic
board fault that took a long time to track down. The component failure
probably occurred during the power outage some weeks ago and was latent
until a more recent power down, at which point the board was unable
to
reset. Ben Abbott isolated and rectified the fault, and we finally
acquired common-mode operation in the wee hours of Thursday morning
after over a week of false starts.
Safety and security (Rich Riesen)
------------------------------------------------------------------------
The site security gates are now fully operational.
$MZ$ NOTE TO LLO VISITORS ARRIVING AFTER HOURS
OR ON WEEKENDS:
You
must now contact the control room via the intercom
at
the LLO site entry gate to gain initial entry. For an
extended
visit, please contact Rich Riesen by email
(riesen@ligo-la.caltech.edu)
to arrange for a personal
access
card.
The Laser Warning Signs on all 4 ISC tables are now automatically
controlled by card swiping in and out.
The PSL table swipe units have been moved from the PSL table
to the
outside wall of the acoustic enclose. They function the same
as before.
The LLO laser personnel status sheet has been up-dated and is
posted
outside the change room door.
I only have 2 operators left that need to be trained for registered
laser user status.
Performed weekly IR scan on the PSL table and the 4 ISC tables.
Found no errant beams.
There have been no LLO incident reports for this reporting period.
Completed weekly site safety tour and found no apparent safety
concerns.
I have initiated CPR training for LLO staff.
Continuing work on updating and modifying LLO safety documentation.
Interferometer commissioning and S2 analysis (Ken Yoshiki Franzen)
------------------------------------------------------------------------
1) installation and test of the f2p filters
2) learning how to run WaveMon
LDAS (Igor Yakushin)
------------------------------------------------------------------------
LDAS admin:
1) Replace non-Seagate drives in T3s;
2) Tested that Shannon's temporary fix to VPN solves DB2 connectivity
problem at CIT;
3) Wrote LDAS shutdown and powerup procedure.
LDAS data analysis:
1) Finished parallelizing waveburst. The job throughput seems to be
now
only constraint by the amount of available nodes. If I use all the
nodes
at CIT, I can get 30 times better throughput than before. Doing output
directly from nodes to files reduced the output and MPI communication
overhead practically to zero (without direct output to files it was
impractical to parallelize waveburst due to the overhead in
ligolwAPI/metadataAPI, eventmonitorAPI and MPI communication). A single
job can run now as much faster than before as the number of nodes
available. For example, my typical job is using 64 nodes and does 64
times more work than before without increasing the running time.
Probably there are still other improvements that can be done to get
another factor 2-3 of performance increase but for now the current
performance is more than adequate.
2) Working on a new feature of the waveburst that would compensate for
non-stationarity of data which greatly affects the analysis results.
CDS Software (Chethan Paramesawariah)
------------------------------------------------------------------------
Installed a few of the new Sun Blade 2000's in the control room. On
Wednesday, the LVEA Y End vacuum crate died. Replaced the crate
and
started the vacuum system within two hours of it going down. Vacuum
remained normal during this time.
Wrote a document "Art of reviving CDS LLO computers/systems after a
power outage" to help starting up the CDS LLO computers.
Working with vijay in getting the bilinearity monitor working.
Writing a robot program "ROBO PLOTTER" to automate the Daily summary
plots during the science run. This program will run once everyday and
elog and save the summary plot automatically with the right date and
time convention.
CDS Software (Ash Khan)
------------------------------------------------------------------------
1) Prepared documentation for the CDS power up procedure,
2) Developed a diagnostic procedure for recognizing epics controller
missing Burt restore after poer-up.
3) Tested LOS coil driver output for the unhandled Fault condition from
RMS mon signal.
General Computing (Shannon Roddy)
------------------------------------------------------------------------
GC: Spent some time troubleshooting what was wrong with a machine that
Jonathan ordered for the Solid Works software. This machine came
from a
vendor in CA and there were several problems with it. In the
end, it
ended up being returned for warranty and is expected back today.
Looking into why several hard drives would not work with a 3ware RAID
controller. I called 3ware and they suggested returning the card
for
warranty. It would not detect any drive plugged into it, so I
was not
sure if this was a drive compatibility problem or if this was a
defective card.
Working on getting quotes for a RAID file server for LLO. Will
probably
end up with the ASA offering for compatibility with the other RAID
boxes
we already have.
Working with RuthAnne at CIT to solve some of the campus VPN problems.
Larry notified me that we may be required to start running the VPN
for
Pcard reconciliation very soon. This has always been a major
headache
and has caused many problems in the past.
I will begin changing UIDs soon for all NIS+ accounts to make it more
compatible cross platform. This is in preparation for the LDAP
move.
Finally getting somewhere with Sun on LDAP support. There will
be a
conference call on Monday between myselft, John Fragalla (Sun), Larry
Wallace, Jeff Weiss (Sun) and Thomas Light (Sun) to discuss problems
and
future plans.
LDAS: Still looking into the network problems with the CIT VPN.
Igor
will continue some testing through Monday. We will then swap
out to
OpenBSD 3.3 to see if this cures the problem.
HEPI and Optics Support (J. Kern)
------------------------------------------------------------------------
LLO Seismic retrofit
Marcel is back at LLO and he, Oddvar and I are completing the
integration of the housing and actuators into the BSC and HAM models.
We've discovering numerous minor conflicts in the mating of parts and
hole locations, all of which have been pretty easily resolved.
The size
of the SolidWorks model, and the number of parts makes the work slower
than we'd like but we're working nonstop and expect to distribute a
final 'clean' assembly containing no legacy parts tomorrow. Many calls
and e-mails are incoming from the manufacturer of the machined parts
of
the actuator, and springs which are questions about the drawings and
tolerances of the parts. Several telecoms held with Dennis & Ed
Jasnow,
and the actuator parts machine shop concerning parts delivery and
software support.
LLO Lab
Ordered additional IR cards from multiple sources per Mike Z's
suggestion, Contacted P King in Sidney to locate a spare PMC
in his lab
and Lee Cardenas found it for us and shipped it to LLO yesterday.
Acoustic upgrade (Gary Traylor)
------------------------------------------------------------------------
The acoustic enclosure is in position over ISCT4. The acoustic panels
and blower have been installed and we are waiting for materials to
finish the light pipes.
Inspiral veto analysis (Yung Yong Wang)
------------------------------------------------------------------------
Still doing the veto studies for the Inspiral data
analysis and finished the re-checking of the 28 SUS channels
with the filters. Only the ETMX and ETMY channels have the
coherent glitches the results are posted in the web page:
http://www.ligo-la.caltech.edu/~wangyy in the title of
"veto studies for S2 Inspiral data Analysis (3)".
I'm also doing the data quality studies with the channels of
WFS1, WFS2, WFS3 and WFS4. Some results are interesting.
Dennis Coyne reporting
·Rich Abbott is revising the electronics requirements document and writing up the electronics design for an informal, EPI-internal, working review meeting on electronics and controls 9/19
Posted views of the LLO pump station design on the
elog for Thursday meeting re. the final approach, some items will need
approval and clarification from the group.
Talked to the chemists re. the LASTI fluid life, and
long term availability, and will be one topic for the pump station meeting.
The fluid biological growth test is showing nothing after 37+3 weeks.[3 weeks dark environment]
- Framebroadcasters at sites are now broadcasting
Frame version 6.
- DAQ now allows the retrieval of "obsolete" data channels (signals which have been renamed over time or are no longer being acquired by the system) via the Network Data Service (NDS) to dataviewer.
Timing System and ICS110B channel hopping:
- After many hours of staring at the system I am convinced
that the channel hopping is caused by timing system glitches in either
the clock driver module, the GPS receiver, or the clock distribution. Testing
to date points more towards clock distribution and the clock driver module.
Hopefully we can find the source in the next several days.
- In the process of looking for glitches we discovered
a design flaw in the receiver circuit of the clock driver module. We have
modified two boards here at CIT and so far it seems that the channel hopping
problem has been reduced, but due to the random nature of the jumps we
won't be certain until we try it on several systems for many days. Richard
has also modified a clock driver in the 4K at LHO. This system seems to
be particularly prone to channel hops. We will let it run for a while and
see if it helps.
Adv LIGO DAC Module: We are ordering 14 modules for
use in LIGO.
Richard installed the UPS with the extra batteries
in the Mass Storage Room.
We received all Agilent 53132A timer units. LIGO got
quite a deal as the vendor upgraded the originally requested 53131A ($2,048)
units to the significantly more accurate and expensive 53132A ($3,178)
model without increasing the originally price ($1,200) I negotiated out.
This can be considered as a ~40% (~$7,600) or ~60% ($17,800) saving over
MSR while having fully warrantied and calibrated instruments.
We are installing the Agilent units in the VEAs in Hanford and setting up a test-stand to speed up the EPICS interface development.
The PMC was sent to LLOattn: J. Kern to replace the unit which failed.
I ran FFT simulations for the cold interferometer
with and without ETM: there is no boost in the sideband power according
to the model.
In order to investigate further why this happens, I modified the FFT code to extract informations from the electromagnetic field in front of the ETM and evaluated from the maps the diffraction loss: for the carrier I found 1e-6 which is exactly what we expect for its spot size at that location while for the sidebands I found 0.352 (that cannot be confirmed by analytical estimations, since g1.g2>1 for the recycling cavity in that state).
Contamination Cavity # 1
The new sample "PEEK" wire spool four of them for
the OSEM is under test. Cavity is locked and we are
taking ring down and beat frequency measurements everyday.
So far we can tell from the measurements that the sample is clean. Does
not show signs of contamination.
Absorption Test Measurement prototype in standby
( New laser coming the first week of September 2003)
Scatterometer system
New and complete scan for the HR surface Reflectance
for the 2ITM04-C (Inner test mass mirror)
is underway. Complete plot and results will be released
by Dr. Zhang.
OTF Lab at Lauritsen ROOM 38
Cavity #3
Still pumping with two new mirrors from REO of ~72ppm.
We have made a complete new modematch both ways calculated
and measured.
Cavity is locked BUT the power output and cavity visibility
still low same as before with the other two
mirrors from REO.This
time we concluded it is due to the mirrors scattering and coating surface
.
We attribute this because we took ring down measurements
to be as 18.25microsec. for this set
and the previous
set measured 18.73microsec.BUT back
on 7/22/03this cavity had a good
pair
oforiginal
mirrors of ~70ppm and the ring down was then 24.12 microsec.Big
difference!
Therefore, our next step is to replace this cavity
with the original ~70ppm mirrors which they have
been cleaned by Helena.
Cavity #2 Test cavity in STANDBY.
Akira has been looking into getting the TNI's discriminant up, and hence the electronic noise down. He measured the electronic noise coming out of the photodetector and calculated that, as expected, it is consistent with the current TNI noise floor. He is now in the process of redistributing the local oscillator power between the Pockels cell and the phase shifter. So far, we have been splitting the power 50/50 between these two devices, but the phase shifter has its own circuit to set the voltage and could take a lot less on the input, leaving more to go to the Pockels cell and thus increase the modulation depth. We calculate that 90/10, with 90% going to the Pockels cell and 10% to the phase shifter, would provide enough voltage to operate the phase shifter, while increasing the modulation depth by approximately root-2. If we see such an increase, and if the noise floor goes down by a corresponding amount, then we will work on boosting the voltage further.
On the coating measurements for advanced LIGO, I have been working on data analysis for our photothermal experiment. Specifically, I have been looking at how to extract the thermal conductivity and the thermal expansion coefficient of coatings from a photothermal measurement. I've been doing both calculations and library research.
It turns out there is a large body of literature on using photothermal displacement spectroscopy to measure thermal conductivity and thermal expansion in both isotropic materials and thin films. Several groups have solved for the photothermal response of multi-layer, coated samples analytically. The usual approach is to solve the thermal diffusion equation with appropriate boundary conditions (which differ from paper to paper) then solve the navier-stokes equations for the displacement that results from the nonuniform heating. These solutions are somewhat involved, and in at least one case the solution to the diffusion equation merits a paper by itself, with the navier-stokes solution in a follow-up paper. In every case, the solutions are quite complicated, and the different groups go to a lot of effort (i.e. numerical analysis) to fit their theory to data to extract material properties.
I think we can do better, or at least the same thing much easier. In all cases, the formula for the photothermal response is only complicated because it must be valid over all frequencies. In the asymptotic regimes, where the thermal diffusion length is either much greater than or much less than all other length scales (e.g. coating thickness, spot size), the formulas are very simple. Right now I am looking at interpolating between these regimes to get a formula that is easy to fit to the data, and from which it is easy to extract the thermal conductivity and expansion coefficient. Getting the formula is easy, but checking its accuracy against the full analytical result takes some time. That's where I am right now, but even if an interpolation is not accurate in the intermediate ranges it's not a show-stopper. Using an AOM to chop the beam, we can just measure the response in the asymptotic regimes and fit different formulas in different frequency ranges.
Ivan is also working on calibrating the instrument. We want to have
the calibration ultimately traceable to some standard when we present numbers
for coatings, and that's going well.
HEPI (Mittleman, McInnes,Hardham and Mason)
We have started testing the pump station to see how many chambers it
can
drive. Using resistors supplied by Ken Maliand, which simulate a chamber,
we have looked at the power spectrum of the hydraulic pressure at various
points along the line as a function of the number of chambers being
driven. The preliminary conclusion is that the existing pump station
should be able to drive three chambers. When a fourth chamber is added
in
the pressure noise increases considerable and the pump starts to get
hot.
We still need to drive three chambers for a number of hours to check
for
temperature effects.
Thermal Compensation for Inital LIGO (Ottaway)
Work continues on the optical design of the for the CO2 laser delivery
system
From: Jonathan Allen <jallen@ligo.mit.edu>
This week I received and built working PCBs for the mechanical shutter
controller.
I am creating a drawing for the front panel to accompany this PCB.
I am adding offset adjustments to the demodulator for the phase camera.
I have provided completed AC wiring for the High Bay and 069 lasers.
Simulation and Modeling (Bhawal)
E2E Weekly Physics Meeting
----------------------------
Luca gave a summary of activities at LHO. Luca, Matt, Virginio, Hiro,
Biplab
discussed about noise curves and WFS loops.
FFT
------
(Hiro) We discussed about how to go for the FFT study with the real
phasemap
data included.
One is to revisit the simulation inputs to make sure all input
information
are properly set. This is to make sure that the result we obtained
by
placing the ITM/ETM phasemaps are correct. Also needed is to
understand the
large difference of the performance due to different ETM losses.
The
effect
of the super polishing will need to be estimated -
(1) psd is used to estimate the loss due to the scattering and
(2) by putting the loss in FFT run to see the effect on the LIGO
performance.
Second is a bigger task to estimate the effect of the mirror aberration
by placing phase maps of all 6 mirrors. The hot state LIGO will
be
simulated
by the thermal model by Bill kells.
WFS loops
------------
(Biplab)
Feedback to Luca who is currently trying to gain-up the WFS loops
with
the
final
aim of increasing the input power to Hanford 4 km interferometer
:
All single loops implemented by Luca are stable. Especially WFS1-DETM
loop
is stable even at 10Hz Unity Gain Frequency (UGF).
However when all WFSs are turned on together, the IFO goes out-of-lock
easily.
The best among "together" cases that I tried so far is like this:
WFS1(DETM), 2A(CETM): UGF 5.0 hz
[Optical lever off in ETMs but ON in other mirrors]
Other WFSs: UGF 0.5 Hz.
It remains locked until about 10 sec with seismic noise level
of only
0.1.
It does not seem to be robust enough to tolerate the full seismic
level.
Mechanical simulation
------------------------
(Virginio) Held a teleconf. for the e2e mechanical simulation
on
Wednesday 10th
1PM. Participants Giancarlo Cella, Mark Barton, Virginio Sannibale,
and Hiro
Yamamoto.
Conference Minute:
Discussed the MSE Status and future development
Modules available and working after some basic test in the frequency
domain
- Wire
- Beam
- Rectangular Blade
- Rigid Body
- Position Sensor
- Force Actuator
Module to be implemented and tested and tasks
- Triangular blade (Virginio)
- Position Actuator (Giancarlo)
- Stiffness and mass matrix extraction algorithm. ("repechage"
of the
old
algorithm MSE (Giancarlo)
- System identification algorithm to generate the state space
representation
from the frequency domain representation (Giancarlo)
- State space Integrator (Virginio)
- Study of the feasibility of interfacing different state space
representations with back reaction (Virginio)
- Compare Mark's simulation/measuremets result with the
quad sim.
(Mark, and
Virginio)
Radiation Pressure
-------------------------
(Hiro) Juri Agrest, a graduate student of Univ. of Pisa, is visiting
LIGO to
identify his thesis topic. As one candidate, he is looking into
simulation
for advanced LIGO. The specific topic that he was working on
during his
short visit is the radiation pressure effect.
Hiro has set up a e2e model to simulate a FP cavity with radiation
pressure
acting and LSC/ASC engaged. Juri has been studying the radiation
pressure
effect by reading appropriate references, both length and angular
issues,
and by running the simulation.
Hiro has worked with him to clarify several issues.
Code development and maintenance
---------------------------------
(Melody)
- Continuing e2e source
code modification to dynamically create
C++ source from FUNC_xxx
equation modules, which would be compiled,
combined to a shared library,
and used during runtime execution.
Worked on getting the macro
values into the source code and resolving
some initialization issues.
Also updated the e2e/lib
and bin directories to the latest version.
Alfi
--------------
(Bruce) Continuing work on bundles and bundlers in Alfi (PR 272.)
LIGO Data Analysis System
Software Systems (Blackburn)
We have begun using a beta release of a public domain version of the
SGI
standard template library. The package goes by STLPort. It is not mature
enough to use in LDAS, but is has very useful debugging capabilities
for
identifying misuse of the standard template library in C++. In order
to
used these debugging features, LDAS must run using it which has been
a
bit slow and painful. So far we have been able to run stochastic jobs
on
a test system configured with it. It has identified several bugs in
the
use of iterators in the frameAPI, dataConditionAPI, lightWeightAPI
and
the ILWD library. These bug fixes are currently under evaluation with
GCC's STL to see if they contribute to the coredumps. We will be testing
all search codes against this STLPort to maximize the debugging benefits
over the next couple of weeks.
We have begun investigating the use of the Intel C++ compiler for Linux
builds of LDAS. As of this week we have been able to compile our object
oriented socket library (ospace) with this compiler but it hasn't as
yet
identified any new bugs in the code.
The dataPipeline command now has the -usertype and -outputdir options
that
have been part of the createRDS command for several releases. This
will
facilitate the dataPipeline jobs that produce large sets of data
products
which would benefit from having localized storage of results. One
example
of this is the SFT frames generated by the periodic group.
At the beginning of the week there were 212 problem reports in the LDAS
bug tracker (and several more were added during the week). We will
be
focusing on these over the next couple of weeks in preparation for
the
release in mid October.
The createRDS Mock Data Challenge (MDC) will take place the week of
October 6th. This doesn't give us much time if a serious issue comes
up
during the MDC before the start of the science run. We will push the
pre-release versions of LDAS to all sites the prior week to facilitate
testing and setup issues. This should give us a bit more schedule for
fixes near the start of the science run.
The memory and cpu load plots in the controlMonitorAPI's GUI have been
enhanced to show the times of the coredumps, as well as the times of
the
restarts (original functionality). This clears up some uncertainties
in
the managerAPI associated with unnecessary restarting of APIs.
Configured a common area for test frames to be used by all LDAS systems
used in development. This allows tests to be easily moved from one
system
to another without having to identify the data and copy it onto the
target
system.
Hardware Systems (Anderson)
Caltech
-------
(Dan Kozak)
* Determined new (2.9a) version of hsi doesn't help get LIGO data from
HPSS more efficiently, so I reverted to 2.6.2 (it works better
for
some things).
* Continued retrieving E7 data from HPSS.
* Migration of SAM-QFS data from 9940A to 9940B continues.
* Continued copying S2 L1 RDS data into SAM-QFS (until IDE-RAID failed).
* Got all of Greg's pemRDS data into SAM-QFS and have exported the tapes
for shipping to LHO.
* Fixed ldas-archive:/archive disk full condition.
* Installed SAN 4.2/SAM-QFS 4.0 on LHO's fb3.
* Fixed staging problem (stuck daemon) at LHO.
* Reconfigured ldas-archive to use multiple 2GB FC HBAs and 3510 RAID
units
as the disk cache for /archive. /archive is now ~10TB
in size
available
at ~500MB/sec.
(Al Wilson)
* Added directories to various user machines.
* Helped Hari add user accounts to the nodes on the bewowulf machine
for CIT.
* Setting a SMP machine for the replacement of ldas-pcdev1. I have run
into
a bit of a snag, it does not work when it has the ip of that
machine.
* Found a bad drive on datacache2 (ide3,4) Replaced drive and another
failed. Will replace when the data is backed up.
MIT
---
(Keith Bayer)
* Brought pcraid4/5 on line.
* Troubleshooting failed raid on DMT machine.
Livingston
----------
(Shannon Roddy)
* Still looking into the network problems with the CIT VPN. Igor
will
continue
some testing through Monday. We will then swap out to
OpenBSD 3.3 to
see if
this cures the problem.
Hanford
-------
(Greg Mendell)
* The main sys admin activity on my part at LHO continues to be working
with LDAS to identify bugs in RDS generation and test the latest
createRDS scripts.
* The disk/tape archiving system at LHO was upgraded this summer to
use
a large tape library and SAM-QFS. To facilitate access to the
data and
standardize the locations for archiving we plan to reconfigure
the T3
disks used by LDAS, the CDS fb3 framebuilder, and DMT for trends.
After
consulting with representatives of LDAS, CDS, the DMT, and the
commissioning team, we have scheduled downtime for the reconfiguration
to start next Wed, Sep 17 2003, and last 1-3 days.
(Ben Johnson)
* Both Fujitsu T3 disks have been replaced at LHO with Seagate Cheetah4
disks (firmware rev. AA27).
* Continuing work on archiving scripts so that they can be used during
M3.
* While preparing the LDAS section in the Control Room, discovered that
the controlmon box has died (bad CPU fan, bad HDD).
Data Analysis Activities (Lazzarini)
Charlton:
This week I have begun testing the data reduction procedure in earnest,
beginning with a few hours of S2 raw data copied to an LDAS standalone
box. This has uncovered at least one bug in the createRDS user command
(which is probably related to a similar bug reported by Greg Mendell)
whereby the user command doesn't produce RDS data for all the times
requested under some circumstances.
Testing has also revealed a couple of minor bugs in the the createrds
GUI which have been fixed. I've also added an installation script for
the
GUI.
For the RDS MDC, I've revised the planning document again and submitted
it
to interested parties. I have written up a set of requirements for
publication of RDS data to LDR (Lightweight Data Replicator) and Hari
Pulapaka has begun work on a script to implement the publication.
Creighton:
I have finished writing and testing LAL routines to compute matrix
inverses, determinants, eigenvalues, and eigenvectors. I have
started
writing a utility program to place a flat mesh (constant parameter
metric) for a directed pulsar spindown search, using the new routines
to compute the optimal eigendirections of the metric.
Mendell:
1) The stackslide search has reached its first milestone. It
can sum
power from SFTs and output the results. I've posted the link
on the
PULG S2 investigations page for the S2 injections. Since no sliding
was
done; this was just to show that we have preliminary code up and
running. Also, this first version of the code only runs on one
node on
a simple flat parameter space. Future major milestones will be: 1)
adding sliding, 2) running in parallel, and 3) extending the code to
a
nontrivial parameter space.
2) Continued working with PULG on methods for doing parameter estimation
in the frequency domain.
Reilly:
Another week on the DSO. I encountered multiple problems with ldas-cit
being down and ldas documentation. I attempted to do the first
run
through
of S2 data didn't make it. I have been working with Tania to get the
MC
code
in form to add to the DSO and I think we are just about there. I now
know
what DSO changes are necessary to make to do the simulations we want.
Currently these changes are being made. I have re-written most of the
tclsh
scripts necessary for running the new DSO. We now have the ability
to
write
out down sampled AS-Q time series for use at a later time. I
have set
up
the code to generate 4096 down samples and will be running that today.
This allows us to do our downsampling in two steps which should prevent
some of the worries we had about using such a high order filter (i.e.
the
filter required by down sample by 16 in one step).
Shawhan:
Attending the PHYSTAT 2003 conference this week. ("Statistical
Problems
in Particle Physics, Astrophysics, and Cosmology")
Yakushin:
1) Finished parallelizing waveburst. The job throughput seems to be
now
only constraint by the amount of available nodes. If I use all the
nodes
at CIT, I can get 30 times better throughput than before. Doing output
directly from nodes to files reduced the output and MPI communication
overhead practically to zero (without direct output to files it was
impractical to parallelize waveburst due to the overhead in
ligolwAPI/metadataAPI, eventmonitorAPI and MPI communication). A single
job can run now as much faster than before as the number of nodes
available. For example, my typical job is using 64 nodes and does 64
times more work than before without increasing the running time.
Probably there are still other improvements that can be done to get
another factor 2-3 of performance increase but for now the current
performance is more than adequate.
2) Working on a new feature of the waveburst that would compensate
for
non-stationarity of data which greatly affects the analysis results.
General Computing (Wallace)
MIT:
(Keith)
-Patched several more gc machines
-Patched several windows computers
-Troubleshooting backup server scsi raid problems
-Troubleshooting HP4500 printer problems
user issues
Livingston:
(Shannon)
-Spent some time troubleshooting a problem with a machine that
Jonathan ordered for the Solid Works software. This machine came
from a
vendor in CA and there were several problems with it. In the
end, it
ended up being returned for warranty and is expected back today.
-Looking into why several hard drives would not work with a 3ware RAID
controller. I called 3ware and they suggested returning the card
for
warranty. It would not detect any drive plugged into it, so I
was not
sure if this was a drive compatibility problem or if this was a
defective card.
-Working on getting quotes for a RAID file server for LLO. Will
probably
end up with the ASA offering for compatibility with the other RAID
boxes
we already have.
-Working with RuthAnne at CIT to solve some of the campus VPN problems.
Larry notified me that we may be required to start running the VPN
for
Pcard reconciliation very soon. This has always been a major
headache
and has caused many problems in the past.
-I will begin changing UIDs soon for all NIS+ accounts to make it more
compatible cross platform. This is in preparation for the LDAP
move.
-Finally getting somewhere with Sun on LDAP support. There will
be a
conference call on Monday between myselft, John Fragalla (Sun), Larry
Wallace, Jeff Weiss (Sun) and Thomas Light (Sun) to discuss problems
and
future plans.
Hanford:
(Christine)
- The whole week has been spent working with Criselda, Caltech ITS
and
RuthAnne Bevier to get ITS and VPN accounts for everyone who will be
taking pcard training on 9/16. This included getting 20 people
to fill
out and return to me ITS account request forms with copies of their
Caltech ID cards. The forms had to be faxed to Caltech ITS, where
I
have been working with Erich Snyder to expedite the account setup and
RuthAnne Bevier to get the VPN accounts. I had to contact Ciro
at
Caltech ATC to get updated pcard libraries and the test and development
libraries that will be used in training. I have downloaded, installed
and tested the pcard software and VPN client software on everyone's
computers. I'm also getting some additional computers set up
in the
auditorium for the training session and making sure the network still
works in the auditorium.
- Spent a few minutes purchasing a license for the Calcium calendar
software and getting that upgraded on the web server.
CIT:
(Mike)
-Upgraded Florence's computer and transferred over all her data, plus
loaded
some additional software such as citrix and discoverer for her to
connect
to ATC.
-Upgraded Bill Tyler's computer and transferred over his data from
his
old
computer.
-Worked on Linda Turner's computer troubleshooting some software issues
that
she was having in Adobe Acrobat 5.0. I ended up reinstalling the
software
and adding additional plug ins. I also ran OS & security updates
on her
computer.
-Worked on all NTSRV's running windows update installing current
security
patches and updating firewall security updates.
-Worked with Ed Chargios, showing him how to work with his new palm
pilot
and loaded software to enable him to sync his database from and to
his
laptop & workstation.
-Burned all monthly ghost backups for October to our DVD archive.
-Setup a VRVS meeting that included video and audio
testing.
-Upgraded Jim Covington's computer and transferred all his data over
from
his old computer. Plus I help him setup his desktop settings and
re-mapping
his network drives.
-Loaded a computer for the control room, up on the 3rd floor of W/B,
with
all general computing software.
-Updated our ip database, with all swapped out computers. I have been
doing
a lot of work here.
-Upgraded Calum Torrie's computer with additional memory and clean
up
the
registry on his computer for it was swamped with annoying spyware
software.
(Lisa)
- Setup two new mailman lists.
- Giving cheetah the once over so it can be put into production.
- Swapped in a new server to replace canopus.
- Worked on a dhcp problem in the 40mars network.
(Veronica)
- LIGO website: Set up the videoconference and videotaped a LIGO seminar
by Carol Wilkinson. Scheduled two more videoconferences for upcoming
weeks. Captured and compressed the video of Carol's talk and posted
it
at
the seminars webpage. The audio input by the wireless mic required
some
troubleshooting. Compressed a video clip that she used in her
presentation, for easier and more consistent playback. The clip will
be
posted together with her presentation as soon as it clears through
the
DCC.
Installed Windows security patches on several computers/laptops
currently
in my possession.
Converted an image to resolution high enough for printed media for
a
newspaper publisher, per a request from Caltech PR office.
Posted updates to the roster.
Posted a number of Amaldi 5 presentations as they keep arriving.
Assisted Ryan in preparing a report to NSF on Elba.
- LSC website: Posted belated presentations from the last meeting.
Contacting Peter Saulson regarding a webpage for the next meeting.
Posted
abstracts of the upcoming LSC talks.
- CaJAGWR website: Roster updates and user support.
- Project Science: Updated the program and info webpages for the
upcoming
workshop. Addressed a registration issue. Prepared an updated email
list
for announcement mass mailings.
(Larry)
-Worked a number of procurements. Ordered a disk system to be used
as
scratch
area for the general user community. Wrapping up the SUN maint.
contract. Then
there were a number of misc. items.
-Worked a couple of network issues. We are going to swapout the main
chassis
today, if all goes well.
-Modified a couple of security documents to reflect a number of changes
that
have been made.
-Assisted DCC with a number of issues. Most dealt with document format
translation. Some minor administrative issues.
-Worked with Lisa on the canopus replacement. Still a few things to
work out on
the unit but it is on line for use.
-Spent more time working with the mail system. The mail filters and
mailserver
virus scan has been a big help but going through the logs to check
for
false
positives is taking more time for both Lisa and myself.
-Regular user support and trouble shooting.
Advanced LIGO Seismic Structure
Design/Fab Contract:
The Proposal Evaluation Team completed their evaluation of the Phase II proposals and a report has been sent to Gary Sanders.
Actuator Testing and Redesign:
The two spools of square wire were cleaned by Bob Taylor and returned
to Turbo-Jet. Turbo-Jet will make a test winding of the unanodized
bobbin without using adhesive and will give us a quotation for the re-winding
with adhesive.
Adv LIGO SUS Test stand
===========================
- Wiring diagrams and rack drawings have been generated to convert the
- present setup to one that can handle two triple pendulum suspensions.
Paul and Sander have started the assembly and checkout.