|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
The LIGO Executive Committee Agenda for Monday July 15, 2002 will be:
(Meeting time: 10:30 am Pacific Time)
Open meeting 10:30 - 11:30
Special Items: Status of LHO 2K recovery, Preparations for NSF Annual Review, Preparations for S1, Schedule for S2
Rescheduling the S1 Run - Given the
status of repairs at LHO, and the promise of several commissioning advances,
the S1 run will be started at 8:00 am Pacific time on August 23 (the day
after the LSC meeting) and be completed at 8 am Pacific time on September
9.
no report
LIGO Operations--Administration
There was no site teleconference held on Thursday, July 4 or Thursday, July 11, 2002.
The list of current actions revised to reflect
the status of open actions assigned through June 20, 2001 may be found
at
ACTION
LIST.
From: Ed Chargois <chargois_e@ligo.caltech.edu>
>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
ACTIVITY
| For two weeks ending 07/11/02 | Packages | Faxes |
| In | 57 | 48 |
| Out | 13 | 42 |
From: Esther Cunningham <esther@ligo.caltech.edu>
Press here for ACCOUNTS PAYABLE HISTORY DATA .
From: "Brambila, Ruth" <Ruth.Brambila@caltech.edu>
From: irena@ligo.caltech.edu (Irena Petrac)
CONSTRUCTION:
We are evaluating the impact of Caltech's policy regarding premium pay for working shifts.
SUPPORT (Baldon, Torres, Lloyd, Tischler)
>Rita Torres
>From: Ryan Tischler <rtischle@ligo.caltech.edu>
Progress Period from 06.28 to 07.11
Accomplishments:
The following change requests have been submitted:
| CR-010012
Revision B |
WBS 1.4.4.1 | Closeout Construction Budgets for Initial Computer Equipment Complement at the Sites | P. Lindquist |
| CR-020007 | WBS 1.1.4 | Finishing and Furnishing of Livingston Staging Building | G. Stapfer |
| CR-020008 | WBS 1.1.4 | Purchase and Install Audio Visual Equipment in the New Hanford Auditorium | O. Matherny |
| CR-020009 | OPs | Digital Common Mode Servo to Mode Cleaner Path | D. Coyne |
A LIGO Change Control Board (CCB) was held Tuesday, July 2, 2002. Change Request CR-020007 was discussed. Items totaling $157K were approved. Minutes have been prepared and will be distributed pending Gary's review and approval (LIGO-M020281).
From: Cindy Akutagawa <cindy@ligo.caltech.edu>
Summary of Commissioning Activities at LIGO Hanford Observatory (compiled
by F. Raab)
-------------------------------------------------------------------------------------
(see ilog for details - if it's not in the ilog, it didn't happen...)
2K IFO Repair
-------------
On Friday, June 28, laser light from an errant SM2 in the input optics
chain cut the suspension wire on MMT2, ending any hopes of a science run
this July. (The SM2 control difficulties were initiated by moderate ground
vibrations from an earthquake on the Russian/Chinese border. The ground
shaking was typical of what we get about twice a month from earthquakes
around the world.) We entered the diagonal vacuum section on Tuesday, July
2 to diagnose the extent of damage and to re-hang MMT2 if possible. Unfortunately,
the earthquake stops did not save the fallen optic due to a design flaw.
Magnets had been sheared off the optic. We removed the MMT2 suspension
module, leaving stops behind on the table to ensure we could get accurate
replacement of the module after repairing the mirror. We then pumped the
diagonal section.
We decided we could replace MMT2 with a newly outfitted spare faster
than we could replace the magnet assemblies on the damaged optic. The spare
was outfitted, balanced and then vacuum baked over the July 4 weekend.
We also needed to outfit new baffles for the MMT2 and MC2 mirrors in HAM8
to shield their wires from future laser damage. On Tuesday morning, July
9 we re-entered the diagonal section to replace MMT2 and insert the additional
baffling. The total vented time for this vacuum incursion was approximately
5 hours. We relied on our alignment stops for the module and our optic
balancing procedure and chose not to attempt an in-air alignment of the
input optics. This done to minimize contamination and avoid additional
risks entailed by such an extra procedure. After allowing several hours
of pumping to let the floor and chambers relax, we were able to align the
input optics up to the recycling mirror and back to the REFL photodiode
from the control room, using camera views of the MMT2 and MMT3 to thread
the beam through the small aperture of the Faraday isolator. MMT2 biases
are now typical of other input optics, so our procedures worked.
It should be noted that another earthquake
occurred as we closing up the vacuum system, giving a similar level of
ground shaking as a week earlier and once again causing the SM2 control
system to become unstable. At least we have repeatability! This time the
wires were protected by the additional baffles. Clearly we need to find
the cause of this instability.
The wire-cutting incident and its aftermath
raise a number of concerns. Both 4K interferometers are susceptible to
the same wire cutting incident that has occurred in the Hanford 2k. Furthermore,
the spring-tipped motion limitiers are way too soft to limit motion enough
to prevent magnets from being broken off mirrors. This means that a large
number of our optics are susceptible to being damaged in future. Obviously
we should be prepared to fix these problems whenever we make new entries
into any of the vacuum systems.
4k IFO investigations
---------------------
Emphasis has shifted to getting ready for implementation of the common-mode
servo. A quick investigation of locking behavior uncovered good and bad
news. The good news is that the interferometer in a well aligned state
would typically lock up in about 5 minutes. The bad news is that there
would typically be half a dozen to a dozen "pops" when the control system
had gotten through the dicey part of the lock cycle, with buildups above
400, for each time that the control would actually complete the lock acquisition.
Once we get such a build up, most of the pops should be leading to a stable
lock. We suspect some saturation problem is the culprit, since we continue
to see large fluctuations in angle in ETMx. This remains under investigation.
We have installed new LSC code to improve the loop-shaping capability
and we have been improving routing of some signals in the hope of reducing
pick up on DSCs. There is some preliminary evidence that some transients
seen previously in AS_Q may have disappeared with the new code. We are
preparing to make larger changes to hardware and software in the upcoming
week.
FACILITIES
------------------
The auditorium is all painted and the contractor is getting ready to
put
the seats in. The office wing is missing the floor covering and the
drop in
ceiling. The atrium and the laboratory wing is being taped and spackled.
The electrician is far enough along that we think the power will be
turned
on next week. The biggest concern at the moment is the HVAC system.
To this
late date we are still finding numerous problems in equipment that
was
specified and equipment that was inadvertently omitted. In addition
the
sequence of operation is not possible with the equipment specified.
Everybody is cooperating (more or less) and hopefully the problems
will be
resolved shortly.
Detector Commissioning: We looked at coherence between OSEM sensor signals on HAM1 and MMT3's optical lever's signal. The idea is to detect table motions through OSEM sensors signals and see its correlation with MMT3's optic's motion. On HAM1 table, different OSEMs have different orientations, so by looking at coherence of MMT3's optical lever's signal to various OSEM sensor signals, it will be possible to infer what table motion is most effectively coupled into MMT3's pitch or yaw motion. Preliminary measurements indicate that MMT3's pitch signal has the highest coherence with OSEMs parallel to beam line (such as MMT1's rear OSEM) and MMT3's yaw signal has the highest coherence with OSEMs in a diagonal orientation to bean line (in the direction of MC1's side OSEM). (Raghu, Yamamoto, Yoshida)
WFS2,3,4 whitening board transfer functions were measured. Some variations in gains with respect to each other and respect to design specifications were found and fixed. (Gaby, Andri). The new CM topology has been implemented. This resulted in a slight noise reduction across the band; from 60-200 there is at best a factor of 2 reduction in spots. Mostly it just reveals the patchy character of the noise in this band. Above 300 Hz, the slight variation is just due to a slightly higher AS port power. Maybe ~50%. We are now essentially at the controller noise floor, in the broadband. There are other patches of noise which exceed the "at rest" voltage noise, but the broadband noise isn't going anywhere from 60-200 Hz. (Rana)
LDAS:
* Fixed a communication timeout problem with the slope code: previously
LALApplySearch() would only return once per run which caused the code to
timeout on some segments (which had the most amount of peaks), now LALApplySearch()
returns after each N peaks are found. As a result the code now runs several
times faster on segments with high number of peaks.
* Made slope code produce a record in a search_summary table. Previously
if slope did not find any event on a segment, no entries would be inserted
into the database which was very inconvenient for accounting
purposes.
* Sent an old Foundary switch to Caltech.
* Received and built the new rack.
* Made an LDAS@LLO hardware list for Albert.
* Learning signal processing.
(Igor Yakushin)
GC: Upgraded the modems for the dialup lines at LLO. Installed some "help desk" software for tracking GC issues. This is available at http://bermuda.ligo-la.caltech.edu/helpdesk/. changed the configuration of the firewall for the LLO webcams. Still working some issues on computers for the students. Working with Mathworks for Matlab licenses. Working on consolodating some of the web based databases and looking at the effect this will have on the web server. Looking at Apache version 2 and seeing what differences there are. This would be to fix a vulnerability for apache. Trying to track down a couple of virus issues. This involves the daily reinfection of several of the machines here with the klez virus. Have not been able to track down the source yet. Have determined that some software I was evaluating is causing random lockups on Win2k. Unfortunately the uninstall feature of the software does not work. I will have to work with the vendor on this. Getting pricing on PCs due to the need for a couple additional machines for the SURF students. LDAS: Installed the two VPN test boxes in the control room. I will work with Igor on this later this week to get the database replication set up through two U10 clients (Shannon Roddy)
CDS:
Worked on an epics interface for the True Time server. We now have
an medm screen with an idiot light on it to show if the true time server
is good or bad. Working on getting big brother for CDS network. We will
have a medm screen for the big brother. Installed new Dataviewer on CDS
side. DMT trends on CDS is now working. (Chethan)
Seismic Amelioration This week in seismic amelioration I, along with
the help of George Noid, Rob Berry, and Jonathan Kern, tested the machined
spring for compliance. I have every indication that the results agree with
the FEA analysis, but I am holding all results until the models can be
properly compared. I worked with Joe Lacour to finalize plans on the for
the HEPI actuation system with the intent that we start cutting metal this
week The Spring Tester had a faulty signal conditioner and a replacement
has been ordered and should be reinstalled very soon. (Marcel Hammond)
no report
no report
LASTI (Goldman, Harry, MacInnis, Mason, Mittleman, Ottaway,
Rollins, Shoemaker, Zucker)
=============================================================
VACUUM ENVELOPE: Pressure is in the mid 10^-7 range and falling with
time as expected. We had a couple blackouts during the recent
heat wave which exercised the emergency shutdown provisions for us;
all
protections worked correctly and the system recovered without
incident.
PSL: Jamie continues to characterize the intensity noise and measure
related transfer functions. Mike fixed the hotrodded FSS board
(which
burned out some weeks ago after a brief hint of spectacular
performance); we haven't yet gotten the reference cavity locking with
the repaired board, but will try to replicate that teaser result when
we do.
CDS: Operating well, except for a blackout-related reboot of the
EPICS and framebuilder node.
EXTERNAL PRE-ISOLATORS FOR LLO SEISMIC RETROFIT:
- MEPI Pre-isolator Design/Fabrication (Ken)
The assembly drawing and assembly procedure for the MEPI actuator has
been completed. Work is underway to complete the overall MEPI assembly
drawing and assembly procedure by monday 7/15. Assembly is expected
to
begin on that date.
Visits were made to Lavallee Machine, Arland Tool, and Southbridge
Sheet Metal. Lavallee Machine had a couple of small parts remaining
to
finish up. 4 skids of parts were taken to SSM for paint. Arland Tool
is
making the large steel foot. They have 2 operations left and will
deliver the parts to SSM for paint the end of this week. I am concerned
about these parts as Arland Tool goes on a shutdown for two weeks
starting 7/12. I have asked the president to call me with a daily
update. SSM is aware they are coming and will paint and deliver them
immediately to us. SSM has subcontracted the machining of the
housings
to APCO machine. They have been slow in starting the machining. John
Colegnesi (VP Engineering) of SSM visited them and demanded they have
them completed by 7/10. SSM has picked up the unfinished housing
previously delivered to MIT for magnetic field testing, and is
finishing it up. So if everything works out SSM will deliver all parts
for MEPI on 7/12 or 7/15.
- EPI Installation fixturing and rigging (Myron, Ken)
A test lift was performed on one corner of HAM20, using a modified
screwjack unit and a cantilever plate designed by Larry Jones to move
the crane hook CG over the gull wing mounting pad. The lift went
very smoothly. There were a few issues, however; there was visible
deflection of the cantilever fixture, which might become serious when
the full weight is involved (HAM20 only has the support structure
and no stack inside). Also there was rotation of the gull wing
flange when the bolts were loosened, indicating some degree of torque
stress is being transmitted to the pier as built. This could complicate
realignment after substitution of the EPI since
the flange torque stress can't be duplicated exactly. We are
studying this
and trying to evaluate how it bears on the site installations,
where spherical bearings could have absorbed some or all of the
flange rotation.
We're instrumenting with some new indicators to do a second lift
test.
- MEPI Actuator Characterization (Gregg, Megan, Rich)
We measured the magnetic field coming from the MEPI actuator when it
is
in place on the large steel preisolator structure. We found significant
reduction in the field at all frequencies when the steel is between
the
actuator and the test point. This is the orientation relevant for
calculating the field at the test mass. We used this value to
calculate
the position noise of the test mass. It appears additional shielding
may
be necessary, as the test mass motion due to the field is similar to
the
science requirement document value between 40 Hz and 1 kHz and is only
a
little below the allowable noise near 1 Hz.
- Interim SHaker Reaction EPI for HAM's (Gregg, Dave, Megan, Rich)
The transfer function from a LIGO 1 PEM shaker mounted on the LASTI
HAM's gullwing to a co-linearly located Wilcoxon accelerometer has
been
measured in the frequency band between 1-100 Hz. A pole zero fit involving
22 pole and 22 zeros has allowed us to accurately fit to these data.
These
data have been inverted to flatten the transfer function around a unity
gain point. At present we are currently attempting to close a transfer
function around this flattened transfer function. A preliminary
calculation suggests that the maxium velocity that this approach can
correct for is 2.5 um/s at 1Hz. This is limited by the allowable power
dissipation in the LIGO 1 PEM shaker.
Simulation and Modeling (Bhawal)
New package beta release
---------------------------
(Hiro) Two new software packages are under beta testing before released
for
public. One is the simulation engine and GUI package, e2e-1.7.7. This
includes
the new alfi5, JAVA based GUI front end, and modeler/modeler_freq ver.020704.
Another package is the new LIGO simulation software, SimLIGO version
020704.
These are being tested, mainly to confirm the installation process,
at LHO,
LLO and by Luca on his laptop.
SimLIGO
---------
Matt previewed SimLIGO at the commissioning meeting when Rana gave
a talk
about his matlab based noise calculation. It is planned that Matt will
talk
about SimLIGO at the commissioning meeting on 22nd. If one wants to
play with
SimLIGO, one can download from the e2e homepage. One can generate noise
curves
under various conditions very easily. A simple documentation is included
in the package.
LSC
-----
(Luca) The LSC digital coefficient files can now be read by a Matlab
script
that converts them to a set of poles, zeros and gain in the s-domain.
The
idea is to keep track of any changes in the LSC control system, implement
the modifications in the recent e2e model SimLIGO and give initial
values
to the Calibration Group fitting routine.
Han2k-multimode version
-------------------------
(Biplab) A standard package for running Han2k model with multimode
settings
is now available for download from E2E homepage. There is not much
difference
in the basic structure and hierarchy of .box files of the original
plane-wave
version and this one. This one is made ready to allow people (who may
not be
too familiar with E2E) to study multimodal issues without spending
much time
in figuring out and changing appropriate settings in proper places
of the
.box files. Of course, one can also use this to run with plane-wave
approximation just like the original version.
State of thermal lensing in Han2k
-----------------------------------
Biplab and Hiro prepared a plan to study transition of Han2k IFO from
cold to
final hot state. This final state seems to be different from what was
predicted
earlier and this difference might originate from a lower value of absorption
loss.
The first step would be to determine how this state differs from a
optimally
mode-matched state. Then the transition of locked state from cold to
this
hot state and behaviour of IFO in this final state will be studied.
Code development and maintenance
------------------------------------
(Hiro) g++ 3.1 and 64bit option :
On alterf, gcc 3.1 is available. modeler package is compiled using
this new
version of gcc with 64 bit option on. There are few changes needed
due to the
change (better standard compliance, and 64/32 issue), and the code
is still
being debugged. Hopefully this increase speed because most of the simulation
is done using double as the real number.
(Hiro) Profiling :
The SimLIGO profiler was reviewed and identified offending primitives.
Now much of the CPU time is spent in FUNC, expression parser, which
is used to build the LIGO module mainly expressing complex mathematical
expression. A few possible improvements are thought out, and is under
implementation.
(Ed Maros)
- Worked on installation document
- Worked on build-e2e script
Alfi
-----
(Bruce)
- Working on positionable ports in Alfi 5.
(Melody)
- Continued working on the automated tester for alfi5.
LIGO Data Analysis System
Software Systems (Blackburn)
This week has been dominated by two tasks which address weaknesses in
the
version of LDAS released for the S1 Run back in June. Both of these
primarily
deal with the frameAPI, through benefits may be apparent in other APIs
if
the first of these can be resolved.
The first task is to improve the threadedness of the frameAPI while
handling
multiple user requests. The code base for 0.3.0 primarily used the
TCL event
loop to schedule multiple resquests. But the event loop has an inherent
FILO
(first in, last out -- "...like sitting in the tail section of a Boeing
747...") behavior which under heavy load causes the system to
stall. We are attempting to make as minor an alteration in the frameAPI
code
as possible before the true start of the Science Run so that we do
not
introduce any unwanted bugs. This is being accomplished by fixing up
the
C++ thread class to allow smarting backgrounding of I/O tasks. Our
first
couple of attempts at this have shown only minimal improvements. However,
we have discovered that the default attributes for the threads in this
class
are not properly configured for an SMP box and we are now in the middle
of
several experiments to determine the best options for Suns and Linux
kernels
with SMP hardware.
The second task is to introduce truely modular code for creating the
so
called RDS (reduced data set) frames. This will be a fully threaded
function
which handles all I/O, cuts and linear filtering (resampling) of the
original
raw data in the background. This task is being broken up into two parts:
one
to deal with specific cases where no resampling is being requested;
and the
second to deal with mixed cases where some channels are resampled and
some
are not. The C++ code for the first case is already implemented and
unit
tested, but system tests are only now beginning with no results to
report.
The controlMonitorAPI is also being enhanced to include support for
viewing
the drifts in the time load on each API for given job classes. A new
interface
is also being added to monitor the size of the queue. Several other
minor
changes are also being made to simplify user interactions where time
permits.
A code review of the frameCPP library source has begun by a new software
developer. This is in preparation for the implementation of a version
5
Frame Spec. This will be targeted for inclusion in LDAS after the Science
Run #1 and before S2.
Hardware Systems (Anderson)
Caltech
-------
(Dan Kozak)
* HPSS migration continues. I'm now improving the scripts so that
migration can be ongoing in multiple classes of service
simultaneously. I also spent time working with STK on
failed Redwood
drives. Installed hsi on wulf (the Datawulf machine).
* Installed Fibre Channel card and software in saiph and made the fibre
connections necessary to give saiph access to the T3 disks in
Booth.
Worked on getting the FC switch configured correctly.
I now (as of
this morning) have it working and believe that the underlying
problem
was a simple one: the Java switch GUI doesn't work reliably
when
displayed on my Linux desktop machine. When displaying
on a Windows
2000 laptop and on a Sun it worked fine. This answers
many
questions/problems I've had with switch configuration.
* Put in order for 1050 ea. of Type 9940 tapes.
* Have sent a query to Linda Feather re: a proper SAM-QFS license.
* Installed latest OpenSSH on acsls machine.
(Al Wilson)
* More attempts on getting Apache web server working on a Sun box. I
will get
with GC for helpful hints.
* Testing RH7.3 kickstart file using ldas-sw for distribution.
* Nodes, nodes, and more nodes. We set up 16 beowulf nodes, on the
LDAS-CIT system.
(Stuart Anderson)
* Reorganized 215 Synchrotron to physically split the LDAS-CIT and LDAS-TEST
systems to different parts of the room and prepare for the remaining
construction work to install the fire suppression equipment.
* Initial planning for purchasing ~20TB of disk to hold reduced data
sets at the Observatories and the full S1 data set at Caltech.
* Tested the new RedHat Beta release (limbo) and found it wanting for
automount and laptop suspend-to-disk.
MIT
---
(Keith Bayer)
* PC Raid bid from pcsforeveryone.com is proceeding
(they ate the cost for the RAM).
* a/c install prep work.
* Built second rack for cluster machines.
* Ordered several more 20 amp circuits for future ldas lab expansion.
* Made adjustments to BBrother at MIT.
* Removed some old frame data to reduce space problem on gateway /export.
Livingston
----------
(Igor Yakushin)
* Sent old Foundry Ethernet switch to Caltech.
* Received and built a new rack.
* Made an LDAS@LLO hardware list for Albert.
* Recorded "S1" injection data on two tapes at the request from Szabi.
(Shannon Roddy)
* Installed the two VPN test boxes in the control room. I will
work with Igor on this later this week to get the database replication
set up through two U10 clients.
Hanford
-------
(Greg Mendell)
* LHO beowulf nodes 29 and 32, which spontaneously reboot and hang,
were set to ASA for repair or replacement. The remaining
62 have been
working fine since they were installed at the beginning of June.
* Scripts were run to copy 18 hours of frame data from the earthquake
(on the Russian-Chinese border) on June 28, 2002 at approximately
10:30
am PST, the lock up of H2:IOO-MMT2 on June 28 at approximately
11:57 am
PST, up to the pump action that was started on June 29 at approximately
8:40 UTC. The data was copied to /frame10/brokenMMT2wire062802/Data0
-
/frame10/brokenMMT2wire062802/Data17 and to tape.
The directories are visible from fortress.
This corresponds to GPS times:
709313552 to 709378816,
which correspond to:
Jun 28, 2002 08:32:19 PDT (15:32:19 UTC)
to
Jun 29, 2002 02:40:03 PDT (09:40:03 UTC)
* Pre S1 injection data was copied to disk at LHO and LLO.
The details:
I. At LLO the copied GPS intervals, given as [startTime stopTime), are:
1) 709188000 709201000
2) 709282000 709296000
3) 709372000 709383000
4) 709512000 709527000
This data corresponds to:
1) A 3.61 hour stretch starting from June 26 23:39:47 CDT.
It contains
the interval 709191792 to 709196928 requested
by Isabel.
2) A 3.88 hour stretch starting from June 28 01:46:27 CDT.
It contains
the interval 709282000 to 709296000 requested
by Szabi.
3) A 3.05 hour stretch starting from June 29 02:46:27 CDT.
It contains
the interval 709372000 to 709383000 requested
by Szabi.
4) A 4.16 hour stretch starting from June 30 17:39:47 CDT.
It contains
the interval 709516000 to 709523000 requested
by Szabi.
On decatur (and on the LLO ldas network) the data is found in
1) /frame10/S1Inject/LLO/GPS709188000/Data[0-3]
2) /frame10/S1Inject/LLO/GPS709282000/Data[0-3]
3) /frame10/S1Inject/LLO/GPS709372000/Data[0-3]
4) /frame10/S1Inject/LLO/GPS709512000/Data[0-4]
Symbolic links to all the data can be found in
/frame10/S1Inject/links/LLO/Data[0-16].
NOTE: The directory /frame10/S1Inject/LLO/Injections062902/
is covered
by the above directories, and is now obsolete. I plan
to delete it.
I. At LHO the copied GPS intervals, given as [startTime stopTime), are:
1) 709012000 709024000
2) 709114000 709127000
This data corresponds to:
1) A 3.33 hour stretch starting from June 24, 20:46:27 PDT.
It contains
the interval 709016704 to 709019824 requested
by Isabel.
2) A 3.61 hour stretch starting from Jun3 26, 01:06:27 PDT.
It contains
the interval 709118224 to 709122512 requested
by Isabel.
On fortress (and on the LHO ldas network) the data is found in
1) /frame10/S1Inject/LHO/GPS709012000/Data[0-3]
2) /frame10/S1Inject/LHO/GPS709114000/Data[0-3]
Symbolic links to all the data can be found in
/frame10/S1Inject/links/LHO/Data[0-7].
Data Analysis Activities (Lazzarini)
Mendell:
1) Finished the main tasks that allow the knownpulsardemod DSO to unpack
SFT and ephemeris data read in by LDAS. This is a major milestone
in
that code is now ready to analyze data for periodic signals over long
stretches of time correcting for amplitude and phase modulation, using
the LALDemod and related LAL functions from the periodic package. Some
minor coding tasks remain as to how to store the results in the
database. Currently the signal amplitude and SNR, along with
source
information, are stored in the SIGNAL_DPERIODIC database, but the output
spectrum is not. Code also needs to be added to write to the
job
summary table. However, the effort will hopefully soon turn to testing
and verifying the analysis code.
2) Continued supervising a summer SURF student to test LAL functions
and
to generate synthetic data sets for testing the knownpulsardemod DSO.
3) Wrote and tested LALWRAPPERInitBarycenter, a lalwrapper version of
LALInitBarycenter, to transfer ephemeris data from LDAS structures
(read
in by LDAS from ilwd files) to LAL structures. I've submitted
this
function the the software librarian for inclusion in lalwrapper.
4) Wrote a script, getsftdata.tclsh, that gets SFT data from LDAS and
transfers it to local disk in ilwd format.
Usage: getframedata <site> <ifo> <start_time> <end_time>
<start_freq>
<band_width> <outfile>
Currently the site must be lho or llo; ifo must be H1, H2, or L1.
The good news is that LDAS is very fast at reading SFT files.
For
example, when I ran
$ ./getsftdata.tclsh lho H2 693601392 693687407 59 1 myout.ilwd
to get SFTs in the 59-60 Hz band for 1 day of data, it only took LDAS
60-70 seconds to read this band out of 42 SFTs files. (The returned
file was 2.56 MB in size; I had a fast ethernet connection to LDAS
so
the transfer time was negligible.) That's over 1000 times faster
than
the real time it took to generate the original raw data.
Data analysis activities -- Shawhan
* Brushed up on my knowledge of Bayesian and Feldman-Cousins statistics,
as
they apply to the burst and inspiral upper limit groups.
* Did some further studies of veto efficiency in the inspiral analysis.
Yakushin:
* Fixed a communication timeout problem with the slope code: previously
LALApplySearch() would only return once per run which caused the code
to
timeout on some segments (which had the most amount of peaks), now
LALApplySearch() returns after each N peaks are found. As a result
the
code now runs several times faster on segments with high number of
peaks.
* Made slope code produce a record in a search_summary table. Previously
if slope did not find any event on a segment, no entries would be
inserted into the database which was very inconvenient for accounting
purposes.
General Computing (Wallace)
MIT:
(Keith)
-Installed software packages on virtual control room machines
-Upgraded all versions of OpenSSH to 3.4 (security patch)
-Troubleshooting wireless cards / hubs
-Moving ilog server over to emvogil-3 (with help from Bruce Sears)
Livingston:
(Shannon)
-Upgraded the modems for the dialup lines at LLO.
-Installed some "help desk" software for tracking GC issues.
This is available
at http://bermuda.ligo-la.caltech.edu/helpdesk/.
-Changed the configuration of the firewall for the LLO webcams.
-Still working some issues on computers for the students.
-Working with Mathworks for Matlab licenses.
-Working on consolidating some of the web based databases and
looking at the effect this will have on the web server.
-Looking at Apache version 2 and seeing what differences there are.
This would be to fix a vulnerability for apache.
-Trying to track down a couple of virus issues. This involves
the daily
re-infection of several of the machines here with the klez virus. Have
not been
able to track down the source yet.
-Have determined that some software I was evaluating is
causing random lockups on Win2k. Unfortunately the uninstall
feature of
the software does not work. I will have to work with the vendor
on
this.
-Getting pricing on PCs due to the need for a couple additional
machines for the SURF students.
Hanford:
(Christine)
- Set up two new PCs for a new staff member and guest.
- Working with Larry on new equipment purchases for the upgrade of
the
LHO bandwidth.
- Planned network and computing needs for the new building.
- Lots of user support for PC network configuration, installing
printers, killing stuck processes, e-mail "how-to", etc.
CIT:
(Mike)
-Loaded a 1.4Ghz computer for Linda that included loading OS and GC
software plus this included backing up data and transferring data to
new
computer.
-Loaded a new computer for Mark that included loading OS and GC software
plus installing additional hardware for DSPACE Simulator unit and loaded
additional engineering software.
-Swapped out a Sun workstation for a SURF student and set him up with
a PC
workstation.
This was required because of some certain software that he has been
assigned to use.
-Set up a safety meeting for Rita Torres to project videotape through
the
projector.
-Worked on Gina Salone's computer fixing some registry errors.
We are currently loading a new Dell PC for her and hope to have her
swapped
out by this week. This should take care of the problems that she is
having.
-Helped Larry install a new, rack mount, SUN server in one the server
racks
and also did some cabling within the server room.
-Added additional software packages for Calum Torrie and Janeen Romie
plus
this included loading service packs 2 & 3 then an additional service
for
edrawings service pack 3.
-Setup the broadcast of two SURF seminars for Alan using VRVS.
(Lisa)
- Worked with Mick and Wendy on getting quotes for parts.
- Began working on troubleshooting some network problems in the 40meter
martian network. It looks like they are having 3 distinct problems.
One of
those is that medm screens crash within an hour if they are run over
a wireless
connection (using wep) and f-secure. Mick and I are doing stability
testing on
different configurations to see which works best. They were also
having huge
collisions on the martian network. Larry, Mick and I swapped
out the old hub
they were using and put in a 16port switch. Mick and I will be
watching it over
the next few days to see if that reduces the instability they've been
seeing.
- Put an interim webmail server into use. Contact me if you need
information
about using it.
- Demo'd EMUmail, another webmail package. This package was developed
on linux
using perl scripts. The perl module dependencies are not well
documented and I
could not get it to successfully work. I am going to look at
a different
package and return to this one later if nothing else looks good.
- More surf/visitor accounts and support.
(Veronica)
- LIGO website: posted PAC12 transparencies. Installed May newsletter.
Posted a few updates to various pages.
- LSC website: reworked the next LSC meeting announcement webpage and
posted it. Updated a list of MOUs. Traced a mix-up of two progress
reports from MSU in the process. Updated the LSC Talks webpage.
- LDAS website: working on updating hardware diagrams, and linking
them to
the LDAS main pages.
(Mick)
-Assisted with installation of faster switch at 40m
-Began diagnosing problem for wireless access (frequent disconnect)
at 40m
-E-mailed companies for product quotes on AIT2 tapes
(Wendy)
-Loaded up a couple of PC units.
-Working on getting quotes for different items.
-Played with scanner
-Installed/updated software on machines
(Larry)
-Lot of things going on the past few weeks in GC. Mick and Wendy have
returned
for the summer so that is a Plus.
-After getting power checked out in the computer server room, after
two breakers
had tripped for no apparent reason, it was discovered that in some
cases there
are two drop lines connected to the same breaker. We are presently
working with
the electronic shop to get a dedicated breaker for each drop line.
-After a number of computer shuffles to get the saiph back on-line
we had to
order a server replacement which is now installed and having the new
OS loaded.
-Worked with Lisa and Mick to resolve a couple of issues in the 40M.
The switch
replacements appear to have eliminated some of the problems. They are
presently
running tests and monitoring things to narrow down the last few problems.
-Worked with Mike on a couple of PC builds and installs. Along that
same line
there are a number of new viruses and hacks going around. So, we encourage
everyone to keep an eye on their system.
-Worked/working with DCC on a couple of issues, recovering old documents
from
tape. Fortunately, the old backup tapes from years ago are still here
and
checking them for the documents being searched for.
-Three of the sandboxes are now in use. There is still some changes
going on as
we learn more about how they are being used. We will need to update
a couple of
the pkgs. on them as well as add some scratch disk space.
-Reworked some of the cabling in the computer room. This will be a
summer
assignment, to get everything relabeled.
-Worked a couple of network issues. In two cases it was a bad cable.
-Worked with a number of SURF students on a number of issues, repairing
their
system, transferring files, e-mail and printing setup and network testing
of
computers they brought with them or have been assigned to from other
areas.
-Worked a number of procurements. Mostly, computer and supply acquisitions.
Still working with Mathworks on getting Matlab licensing setup for
the project.
-Working a number of financial issues concerning the construction account
and GC
systems in the new buildings at the Observatories as well as the new
WAN
connection being worked on at Hanford. Florance is guiding me through
that
paperwork.
-We've started working on some of the logistical issues for the next
LSC meeting
at Hanford. Christine will be gathering more information as time goes
along.
From: Gregg Harry <gharry@ligo.mit.edu>
We have been exploring the possible causes of anomalously low mechanical
Q measurements in disks of fused silica. Originally it appeared
to be an
effect related to electrostatic charge on the surface of the fused
silica
sample, but recent data suggests that the extra damping may be due
to
Coulomb friction with dust in the vacuum chamber. We will soon
be able to
check if the data support this idea, and if so, we will estimate the
coefficient of friction and see if the result is reasonable.
We have also begun an effort to look for different phi values in fused
silica for torsional and flexural motion. A sample is hung in
a vacuum
chamber and we have found the modes in air.
From: Helena Armandula <ahelena@ligo.caltech.edu>
Advanced LIGO Coatings
SMA/LYON
Jean-Marie Mackowski visited Caltech last Friday.
We talked about the next run he'll make for us where the Ta2O5 material
is modified with the addition of a doping material.
He has made several trial runs already; his goal is to make the atomic
structure of Ta2O5 more similar to SiO2; the aim is to reduce mechanical
loss. A detrimental effect of this modification is a 30% increase in absorption
(from 0.4 to 0.6 ppm), however, if "Q" is effectively improved, one can
later work on reducing absorption.
We also discussed some future work. He has the capability of assessing
the mechanical loss of coating materials .
I will set-up a teleconference with J.Marie to get details of these
measurements from him, and see if they could be applicable to our work.
MLD
Finished annealing the Ta2O5/SiO2 coated substrates to our specifications.
Finished the Nb2O5/SiO2 coating run. We should receive it shortly.
These parts have been measured before and after annealing with an interferometer
so we can assess the effects of annealing in coating stress.
Also received two Nb2O5/SiO2 coated 1" samples that showed blistering
after annealing at 500C. We'll ask MLD to anneal 2 more 1"dia. samples
at 450 C for evaluation before the "Q" substrates are annealed.
WAVE PRECISION
They are about to make a Ta2O5/Al2O3 coating run on "Q" substrates.
CSIRO
Roger Netterfield, with CSIRO, also paid a visit last Friday. He announced
that CSIRO just acquired a large Ion Beam Sputtering chamber that they
will commission in August.
Advanced LIGO Suspensions
Received SF4 glass blanks for bonding tests. The parts will be sent
out for polishing shortly.
From: Janeen Romie <romie_j@ligo.caltech.edu>
AdLIGO Suspensions
PRs submitted for two machine shops for MC blades. Delivery due mid-August.
I'll drop off the material to one machine shop, Lobart, this afternoon.
Still working on osem procurement. Alumina heads were due here friday
but the vendor has slipped the delivery date.
Held a meeting with Glasgow, CIT and MIT on the LASTI controls prototype
testing. Quite fruitful - many action items. Jay is working with Harry
Ward on a MC systems wiring diagram.
From: ctorrie <ctorrie@ligo.caltech.edu>
Advanced LIGO Suspensions
Janeen and I continued to extend our search for local companies to supply us some cantilever blades, see Janeen's report.
Larry Jones finished re-checking all of the drawings for the LIGO Mode-Cleaner.
Most of these are now either with Mike or Ric in the Caltech machine shops.
The upper mass, table cloth and blade clamps will be submitted as soon
as some preliminary tests are completed.
Last Tuesday Go Engineer gave us a demonstration on PDM Works and 3D
TeamWorks in the ECR. The demo was relayed to MIT and Glasgow.
TeamWorks allows collaborators at different locations to view and make
comments on a particular assembly or part and each collaborator can take
control of the session.
As a result we are going to buy at least one licence of 3D TeamWorks
and Ken Mason and I are going to try out PDM Works with support from Go
Engineer.
Last Wednesday I spoke with Caroline Cantley, Russell Jones and Mike Plissi in Glasgow regarding the work that Rutherford will contribute over a 2 month period in the next 6months or so, an introductory meeting will be held this Thursday.
This week I have been working in the laboratory with the prototype blades, these were made in the Physics machine shop. Hopefully, tests will give me useful information that can be used on the next set of blades.
On Monday at the LASTI meeting we discussed in detail the design for
the "hybrid OSEM". Some suggestions will be incorporated and 14 will now
be made in Glasgow for the 2 mode-cleaner suspensions
From: Riccardo DeSalvo <desalvo@ligo.caltech.edu>
Alessandro
Accelerometer paper second draft.
Working on maraging components for negative K GAS.
Testing some different brazes.
Preparing to try to laser implant an evaporated coating of Nickel into
the MoRuB surface oxyde. If successful it could be a good way to
prime the MoRuB surface for low temperature braze.
Akiteru
I have just received small optical benches for the optical levers from
the machine shop this afternoon.
A laser pointer for the new optical lever is dead, after 3 day of run.
This may be due to its life time, or noise of a home-made power supply.
I will drive another pointer with batteries for this weekend and see if
it survives. This will cause some delay to the program, in any case. I
am upset.
I am investigating local control of the IP again. So far, we
could damp the IP resonances very well, down to critical condition,
but
damping with heigher gain has never worked as in servo design computation.
I
am looking into what limits performance and have got few ideas, and
verifying the hypothesis.
I have been testing some alternative servo design for the inertial
damping. In a conventional design, the IP fundamental modes are damped
with the LVDT signals. Gain of the LVDT loop is lower than that of the
accelerometer loop at micro seismic frequency, but still sufficient to
re-inject motion of the reference frame. I am trying to increase accelerometer
loop gain between 0.1 Hz to 0.5 Hz.
In the yaw mode damping, situation is much better, the IP resonance
is successfully damped and the damping has gain even out of the resonant
frequency, of nominally 10 dB, between 100 mHz to 1 Hz.
Eric
Work on braze samples
Trained Brian in various techniques.
Built lead shield and optical density calibration ladder for X-ray
setup.
Tested the setup in the Vet X-ray machine.
Finished the calibration of the X-ray optical density scale for Glassmet
at different thicknesses. Calculated (from particle properties) the
relative penetration ratio of MoRuB and Glassmet; now can read thicknesses
in micron of Glassmet equivalent and convert it in MoRuB thickness.
Lots of data to analyse.
Measuring splat wedge angle and correlate with glassicity.
Preparing a standard calibration and screen format for standardised
x-ray measurements.
The measurement presision is limited by a background and/or a gradient
in film sensitivity and/or x-ray illumination. Working on background and
gradient substraction techniques
Brian Emmerson
New summer student, will focus on X-ray diffraction scattering measurement
to measure residual cristallin structure in glasses.
Stefano and Brian
Got trained in X-ray safety procedures.
Making routine successful splats. Alloyed molded and splat MoRuB17
samples.
Produced samples for Alessandroo’s laser zapping tests.
Work on different brazing techniques. Tried several alloys and
surface preps. Trying more and more complex techniques to find the easiest
possible. AgSn seems to behave less badly. Lots of failures
so far, no success.
Stefano, Ric, Gianni
Designed LVDT position readout for mini stress-strain machine.
The aim is to measure stress strain curves while observing the emergence
of shear bands in a electron microscope. LVDT in production.
Comment
Brazing will be tough, but it is mostly good news it indicates
that the first few atomic layers of oxide are a almost indestructible passivation
of the glass. Per aspera ad astra.
Stefano
Succeeded in good Edm cut of MoRuB.
Designing modifications of a mini stress strain machine to use on small
glassy samples inside electron microscope to detect shear bands during
formation.
Hareem
vacation
Youichi Aso
Working on tentative design of suspension noise interferometer.
Made a program to calculate and display noise as a function of the
relevant parameters. Testing.
Mike
Mounted a second glasmet sample with more conductivity and essentially
concluded a coherent data taking run, hand fitting of the exponential decays
vs. temperature show some systematic difference from the automatic
Quantum Design fitting. Now swamped with data. Manual
fitting successfully measures also failed Q.D points. Found that
the heat conductivity power setting need to be changed much more rapidly
that what made by the Q.D. automatic program because the heat capacitance
plunges at low tempeature.
Kelin Mike
Working on Q-factor measurement hardware, made specialized cable from
cryostat to DSP. Work on photodiode puck.
Youichi and Ric
Went to Livingston for S1 and returned after run abort.
Ric
Work on braze.
Technical note.
We are having plenty of problems in brazing to MoRuB. This is
bad in the short term, good in the long term. The present trouble
in brazing reflects the tenacity of the mono- or oligo-molecular layer
of oxyde that passivates MoRuB. A posteriori it is not so surprising
to find trouble because we are trying to low temperature braze an alloy
made of refractory metals, a refractory alloy oxyde is very stable.
The chemical stability that impedes chemical attack also implies very stable
passivation and will allow very thin flexures without fear of cracks and
corrosion. The mechanical stability of the oxyde layer is also welcome.
We are looking for the easiest way to braze to MoRuB. We have
a virtually assured solution which consists in using back-spattering to
ablate the oxyde layer and Nickel sputtering to prime the native metal
surface for brazing. Of course this is a cumbersome choice
and an easier way would be very welcome.
For additional information about this report, contact sanders@ligo.caltech.edu