|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
The LIGO Executive Committee Agenda for Monday August 14, 2000 will be:
CANCELLED DUE TO LSC MEETING
The LSC meeting is next week.
WBS 1.2 LIGO Operations--Administration
The
P-card is now a travel card for sixteen members of the LIGO staff, effective
Monday, August 7.This is a pilot
program, designed to determine a set of requirements, which will lead to
the development of a new software module to be used by the whole campus.LIGO
is the only participant in the pilot program.
From:
Ed Chargois <chargois_e@ligo.caltech.edu>
From:
Cleveland Mak <mak_c@ligo.caltech.edu>
|
|
Packages |
Faxes |
| In |
|
|
| Out |
|
|
Press here to access
the DOCUMENT
CONTROL CENTER WEB PAGE.
From:
Esther Cunningham <esther@ligo.caltech.edu>
Press here for ACCOUNTS
PAYABLE HISTORY DATA.
From:
"Brambila, Ruth" <Ruth.Brambila@caltech.edu>
Nothing to report.
From: Florence Kaufman
<fkaufman@ligo.caltech.edu>
The
monthly reports (Allocation of Actual Costs for Construction and Advanced
R&D) for period ending July 2000 have been put on the network. E-mail
has been sent to the Task Managers noting the location of the files.
I am working on gathering the details for the open encumbrances for
the above accounts. I will email them to the Task managers as soon as I
have completed them.
Continue to review the open encumbrances on the Construction and Advanced
R&D accounts to see if the requested removals have been made. I will
continue to review these accounts weekly until all encumbrances have been
removed.
The financial reports on the web provide supporting detail.
http://docuserv.ligo.caltech.edu/~fireport
http://docuserv.ligo.caltech.edu/~finance
Irene
Baldon
Progress
Period from 8.4 to 8.10
Accomplishments:
From:
Kris Duncan <kris@ligo.caltech.edu>
Completed
the June 2000 CSSR and ran draft performance reports for review.
LIGO construction is currently reporting 98.2% of scheduled work complete.
Initiated
work on the July 2000 CSSR with the calculation of earned value and input
of the actual cost and commitment data.
Will
continue with the reconciliation of CIT and LIGO cost data, it appears
all retention issues have been resolved.
Initiated
the incorporation of the change requests from the August 1, 2000 CCB.
Rebuilt indexes in COBRA, revised invalid estimate dates and advanced the
calendars.
No report this week.
General Items:
--------------
(F. Raab)
We are moving hard on prepping to begin installation
of 4K next week. Rai is leading an effort
to characterize the alignment before moving forward on attempts to lock
the power recycled Michelson plus 2-K arm configuration. Today,
the SST teams have their poster session presentations at Battelle Auditorium.
Caltech SURFers are up here on a road trip and we will have a miniSURF
Seminar Day Friday. Last minute details are falling into place for the
LIGO Public Lecture and LSC meeting next week.
OPTICS/COC/SEI INSTALLATION: After reinstallation of ITM-y last week, we used the COS autocollimator to backproject from the IAS position along the Y-manifold to MMT-3. We discovered that the projected reticle pattern is displaced ~1.5cm to the South on the MMT-3 target. The projected reticle is also not centered in the autocollimator mount on HAM-2. We replaced the autocollimator in HAM-2 and checked the centering of the projected reticle at ITM-x. It was perfectly centered in the elliptical baffle ~1mm. We then looked at the projected reticle at ITM-y. It is displaced East by ~1cm. We then measured the ray projected from HAM-2 as it intercepts the BS and were able to accurately locate its intercept by placing a machinist's square on the LOS baseplate, and marking where it entered the face of the tower with respect to the scribe marks we used for centering. We discovered that the ray is intercepting the BS is ~17mm North of center! Joe Hanson has discovered the source of these errors. He setup a corner cube on a tripod and physically checked the laser-ranged path length of the IAS theodolite with a tape-measure. It was wrong. We are using 40mm corner cubes, and our theodolite's prism constant has been set to 30mm for every optic we have positioned. This means that the range length displayed by the theodolite has been in error by 10mm (short) during the LLO installation. Simply, whenever we have laser ranged we have placed optics 10mm closer to the theodolite than intended.
Remedial plan is to move the BS CAS, leaving the 50/50 surface of the BS where it currently is but putting the spot where we want it on the BS. We also need to move the ITMy to center the projected reticle and then realign using IAS. (Kovalik / Riesen / Kern)
PEM:Program to collect microseismic
peak data is in running on delaronde. It is collecting average power between
.1 and .2 Hz every 30min and creating a power spectrum every 12 hours.
It is doing this for all nine seismic channels (3 channels on each of the
three seismometer - X, Y and corner stations). Program has been checked
with both software injected sine as well as signal generator injected sinewave
into a DAQ channel. Thank you to Fitra Kahn and John Zweizig for their
help. (Anthony Rizzi)
FACILITY: A meeting was held with the architect and the engineers
working on the design of the Staging and Auditorium building at the site.
We reviewed the progress of the Construction Document. N. Traylor, the
audio/visual specialist for the auditorium, presented the single line diagram
for the auditorium audio/visual system.
Erosion control, contractor is working on slope stabilization along
the borrow pit along the south arm.
LIGO II System meeting held at MIT this week.
All activities reported elsewhere.
| Installation
& Commissioning:
Livingston |
Other Science/Engineering
Activities:
Issues/Concerns |
See also the Installation web page
Wed. night we opened the Y arm gate valves. Replicating prior results we locked the Y arm individually (no PRM), aligned for best throughput, then aligned the power-recycled Michelson to agree with these incident beam and ITM alignment settings. We were then able to lock the PRM on the sideband with low gain, and found that arm cavity fringes did not disturb the lock in this condition. However we were still unable to coax the arm cavity into controlled resonance (at least not convincingly). The prime objective for this run was to record the transient signals (reflected, transmitted and internally sampled fields and feedback voltages) and compare with E2E results to see how well we understand the optical plant; unfortunately a bug in the NDS software deleted several critical channels from the reduced data set, and we didn't end up with useful data. Rai Weiss is going to replicate the run this week. Dave and Daniel are working on a software patch to get all the needed channels on disk.
Rolf Bork
Tracking 1 sample (60usec)
delay in LSC input to output. Ran Jay's old single channel
test software, and this delay does not appear. Working thru the code
differences in Pentek module setup from that code to the present LSC code,
finally found what is causing that 1 sample delay. The LSC code requests
that data be "packed" ie two channels of 16bit data packed into one 32
bit word, to reduce the number of VME read cycles. Test code sets
up for one channel of data per 32 bit word. Appears when set to packed
mode, Pentek actually waits for second sample before providing data. When
the LSC code is modified to req one channel per 32 bit read, the 1 sample
delay goes away as well. Delay is now 35usec, signal in to signal out with
all digital processing, instead of the 95usec seen with the packed mode
of operation. I will probably need to contact manufacturer again on why
data packing causes this delay.
IO Baffle: The aluminum frame for the glass will be reworked for the proper 4K beam hole locations. New glass panels were received, but were found to be manufactured incorrectly and will be shipped back to the manufacturer for rework.
Beam Dumps: The layout
errors for the 2K beam dumps in BSC4,7,8 were discovered. A new part for
a vertical mounting tube was designed, and will be made from one
of the existing vertical
mounting tubes to correct the installation problem in BSC8; the existing
parts can be used to complete the installation in BSC7.
Haisheng Rong
In an effort to understand
the balancing anomalies with the small optics, Haisheng measured the static
charge on the SOS mirrors with an electro-static fieldmeter. The field
strength ranged from 0.02 - 0.80 kV/inch for the 4 mirrors I measured.
The results are qualitatively consistent with the "wire test" in which
a wire is used to attract the mirror. The one producing higher field strength
attracts the wire stronger. At the level of 0.8 kV/inch, the mirror balance
angle is noticeably affected by objects in the close vicinity of the mirror
surfaces.
Using the tester that came with the fieldmeter, he also tried to verify the performance of the ionizing gun we use to blow dry the mirrors after cleaning. It turned out that the ionizing gun did not seem to remove any charge. Either it's broken or it has never worked. Further investigation is underway.
[Dennis Coyne] The effects of this error are as follows:
HAM3 Installation: The ITMy telescope is installed and is awaiting re-alignment of the BS and ITMy.
Arm Cavity Baffle: The ITMx-arm cavity baffle, and ITMy arm cavity baffle are installed and aligned. The Glass baffle was removed and is placed against the wall of the manifold. The ETMx-arm cavity baffle is being assembled.
PO Mirror: The ITMy PO mirror alignment is awaiting re-alignment of the BS and ITMy mirror.
Beam Dumps: The beam dumps are installed in BSC2 and are awaiting final positioning after the BS and ITMy mirrors are re-aligned.
IO Baffle: The aluminum frame for the glass has been reworked for the proper hole locations and is being baked. New glass panels were received, but were found to be manufactured incorrectly and will be shipped back to the manufacturer for rework.
APS beam dump: The APS beam dump frame was shortened to accomodate the shorter clearance of the 4K APS telescope from the HAM table in the 4K design.
During the last trip to LLO, some discoloration of the cooling water had occurred (something which was hard to pick up with the green laser safety goggles). Lee has replaced the water in the chiller and in the process had to flush the cooling hoses a number of times to remove some buildup of material. Hopefully this would have improved the cooling to the laser. The performance of the laser will be monitored to see if this improved anything.
At present the EPICS PID record is being looked into for controlling the SLOW actuator. Hopefully when implemented, the software PID loop will perform a little better to keep the FAST actuator voltage centered than the current open loop software control.
We have received the parts for the new osem heads at MIT. Myron & Janeen have begun to define and test the assembly steps involved in bonding the devices (LEDs and PDs) and the optical filter into the head. Once the procedure has been refined, Myron will assemble 19 heads for the LHO 4k installation; they should be ready for final cleaning/baking by the middle of next week.
No report received.
LASTI (Zucker, Mason, Kruzel, Smith, Shoemaker)
------------------------------------------------
MZ: writing up experiment plan and working on control topology for
review at LSC meeting next week.
Ken Mason: Created a preliminary optomechanical layout of the LASTI
system. Matt Smith has been given the task of marking and drilling
the
pier locations. Ed Kruzel is ordering hardware and locating
miscellaneous parts.
Matt Smith: I started to help in getting ready for the first HAM
installation at MIT. We are going through the drawings and ordering
hardware and fasteners. We have ordered the anchor bolts and adhesive
along with the drill bits. Hanford is sending MIT their core
drill. I
have been reading the anchor installation instructions and am planning
on marking the holes in preparation for drilling next week.
I have
been talking with Rich Riesen on the tools and fixtures that he is
sending to MIT. We will start gathering all the material and kit what
we
can and verify that we have everything we think we need. We need
to get
some in vacuum hardware cleaned and baked at one of the sites.
We will
also vacuum bake the spring seats and shims at MIT.
Simulations and Modeling
Lock acquisition
* Biplab worked on fixing bugs of the modal model based
power recycled
michelson. There still is some minor bug related
to the coordiniate
change upon the reflection of the field, and a
work-around patch is
used for now.
Biplab ran Luca's lock acquisition model (Matt's
model + mirror
misalignment) for various cases of misalignments
to observe locking
and determine validity of recent fixes in code.
* Matt is working on implementing codes for the new lock
acquisition
controller.
At Hanford,
* Biplab is using e2e for the noise study of PSL
* Biplab worked with Brad and Rick on effects of higher
order
modes and motion on error signals from reference
cavity in PSL.
* Biplab ran optics simulation of reference cavity with
its mechanical
simulation data supplied by Cella, after sorting
out some differences
in convention and signs. Got some relatively large
spikes in error
signal. Trying to analyze results.
* Ed built GCC-2.95.2 @ Hanford (/home/e2e/local/bin/gcc)
in order to
build the latest e2e software usable. The existing
software is too old.
In order to avoid the conflict with the LHO software
maintenace, necessary
unix tools were built in e2e directory.
Adlib
* Hiro completed and committed to cvs the equation parser
package and
related files.
Alfi
* After repeated discussions, it has been decided that
the main focus of
the alfi development is to make alfi rock solid.
The new feature addition
will be limited to those which are really crucial.
This week, Ed and Bruce
are working hard to make one stable version available
by the end of this
week.
* Several bugs were fixed, including the core dump related to renaming.
* Bruce worked on the interface to define settings in a tree view.
* Alfi is compatible with multi line strings, but is not
friendly.
LIGO Data Analysis System
Software Systems (Blackburn)
Last weeks Report
The LDAS software group has been
divided roughly into two teams this week
with the "A Team" fully dedicated
to the dataConditionAPI Mock Data Challenge.
Though the MDC shouldn't be described
as going great, it can be described as
going well. There were several
days of code development and testing at the
unit level that spilled onto the
MDC and has left us with a later start than
anticipated for the beginning of
the integration tests. We will bring in the
new LDAS integration tester to
try and make up some of the lost time during
the full API integration testing.
There are a number of lessons learned by
this MDC which we will document
and share to the benefit of future MDCs. A
few changes to the internal culture
and policies for LDAS development will
also benefit from lessons learned.
At this time, it looks like the final
suite of tests will not be completed
by Friday, but with a bit of work it
should be possible to complete
the MDC remotely given the current health of
the system.
The work to improve the communications
overhead in the wrapperAPI has already
nearly double the performance of
the code on an 8 node calculation. We have
identified several other areas
for improving performance and expect to get
another factor of 2 to 4 squeezed
out of the performance in the next week.
UWM has been attempting to compile
LDAS on their Alpha system to continue
using the wrapperAPI and continue
code development there on dynamically loaded
shared objects. Unfortunately,
LDAS compiles are producing an internal gcc
compiler error on the Alphas which
we don't see with gcc on Pentiums. There
was a bug that was identified on
a UWM LINUX Intel platform which we were
not immediately able to reproduce
on LDAS hardware but thanks to the its
turning up there we did manage
to find a way to reproduce it and fix the
bug on both platforms.
The support to the frameCPP library
for appending frames to existing frame
files has now been implemented
and tested.
The controlMonitorAPI now has a
strawman interface for the LDAS beowulf and
will soon be made part of the distribution
and build process.
This weeks Report
The first MDC involving the dataCondtioningAPI
has formal ended its testing
phase this week. Even though one
or two outstanding issues remain, Sam Finn
called the MDC a tremendous success.
The MDC left us with an LINUX 2.2.x
kernel bug which caused us to be
unable to use more than 2 CPUs in threaded
concurrency tests. It was discovered
during the MDC that this is a known bug
in the 2.2.x kernel scheduler and
will not be fixed until the 2.4.x kernel.
We also are left with strange failure
for the fft and mixer actions when very
large data sets are involved (2^23
samples). This bug has not yet been tracked
down but may be related to a memeory
allocation that fails and is not caught.
The MDC team was able to send over
2000 LDAS user commands to perform linear
filtering, mixing, ffts, resampling
and other basic math operations on data
during the final days of the challenge.
Much thanks to all that put in the
long hours getting the last minute
features and fixes into the code.
The requirements for the wrapperAPI
and mpiAPI were slightly modified this
week to include options to allow
for more efficient usage of communication
bandwidth at the discretion of
the search algorithm implementation. A strong
effort was made to improve the
communications cost even in the worst of
circumstances where all information
had to be communicated between the slaves
and the master. This resulted in
a factor of nearly 10 improvement in the
communications cost for very small
data sets involving a minimum number of
filters. Our beowolf cluster was
taken down last week to upgrade its os to
something in this century. We will
be retesting the wrapperAPI under the
new configuration soon.
Several tools were identified for
monitoring network statistics and job loads
on beowulf clusters this week.
These include bWatch which has now been added
to the controlMonitorAPI (more
free code that we don't have to write in our
quest for design functionality!).
The controlMonitorAPI is now part of the
LDAS build process and will be
available in all future distributions.
Several issues with using threads
came up during the week. As a result we
had a suite of new TCL extensions
written for the frameAPI and the genericAPI
will be rewritten to support threading
at the operator socket level to be able
to start up threads as soon as
possible in the TCL layer.
Hardware Systems (Anderson)
(Hanford)
* Greg Mendell visited the LDAS data-conditioning mock
data challenge at
Caltech from 7/27/00 to 8/7/00 and attended the
ITP Workshop on R-modes
in Neutron Stars, in Santa Barbara 8/02 - 8/03.
* Patches for RedHat 6.2 for linuxbox1 and controlmon
at Hanford were
completed. These boxes are now fully upgraded
to RedHat Linux 6.2.
* Video cards, an 8 port KVM switch, and a UPS have arrived
at Hanford.
The video cards will be installed in the Sun 450
ldas-wa dataserver and
metaserver, and these will be connected to the
UPS this week (on
8/10/00). This will nearly complete the
migration from the Ultra 10's
to the 450's.
* Steps have been taken to get drawings of the Mass Storage
Room at LLO
to plan the beowulf layout there.
* StorageTek has been contacted again about access needed
to the back of
a L700 or 9740 Tape Library.
* sunbox1.ligo-wa will be set up as an ldas machine since
its upgrade to
Solaris 8.
(Caltech)
5 new Sun E450 computers has been delivered, 3 to Caltech,
2 to LLO, and
1 to MIT. There are being configured to run Solaris8
and the LDAS software
as upgrades for the dataserver and metaserver functions,
as has already
been done at LHO.
A fibre optic network connection to the Caltech LDAS production
site in the
Synchrotron building has been made.
A CD recorder has been installed for general use on
ldas-pcdev1.ligo.caltech.edu.
(Engineering Run Support)
LDAS staff met with Stan Whitcomb to discuss tentative
schedules and possible
goals for future engineering runs. The group worked to
identify functionality
in both hardware and software that could support the
detector efforts. With
the LDAS database and DMT working well, these were quickly
identified as tools
for near term engineering runs. The dataConditionAPI
will also be functioning
soon at a level that could support an engineering run,
adding the ability to
process data and record data products in the database.
(Jeff Edlund, CIT Undergrad)
This week I reduced the number of factors that we store
in our pre-array
from M*M to M/2+1. I also identified that it would
be possible to reduce
this array to at least M/4+1 if M is even.
Right now I'm working on figuring out the best way to
use
interdependencies between the constants that we are searching
for to
reduce the search space. Soon I'll be rewriting
the N-Dimensional FCT to
use these new methods.
(Samantha)
- Tested the new CD-R's, they are spiffy.
- Played with w2000 a bit, first order of business was
changing the horrible
blue color of the desktop.
- Did lotsa and lotsa of surfing the web for various
products. May have
found some solutions to things.
- Fixed PC issues, viruses, etc.
- Admired the mountains of sun equipment that arrived.
- Learned how to better restore stuff from tape.
- Learning a few new software programs to simplify various
tasks.
- Looked for various pieces of hardware.
(Suresh)
- On vacation, lucky guy.
(Barbara)
- Due to underwelming feedback on the costbook reports,
I set them aside
and began upgrading the Roster database.
Compiled the code in Office 2000
and began converting objects to the discipline
and organization codes
needed for LIGO II.
- Updated the inauguration image on the LIGO home page
along with a number
of other quick website changes.
(Larry)
- Working a number of procurement issues, presently getting
quotes for memory
upgrades (prices are shooting up), the new plotter
and a couple of network
switches to handle all the new connections taking
place.
- Finally received the SUN order, 87 boxes, thanks to
the crew from Wilson
House we were able to move things around quickly.
Also, thanks to those
people whose offices we used/are using to store
the equipment as we get the
new units built. The LDAS group has picked up
their units and CDS has picked
up their units. Lisa is working on building the
new units and the 40Meter
has picked up the first one. The rest should be
finished by the end of next
week.
- Fixed a few printer problems. All pretty minor. When
the financial people
print things from the CIT Oracle system we have
no control over those
problems and they need to be fixed by the CIT
group.
- A few virus issues but all were minor since the virus-scan
s/w has been
taking care of most of the problems.
- We had a power hit on Wed., which caused a few machines
to have problems
but most issues were quickly resolved. A few files
were also affected but
so far nothing that has not been recovered. Once
again the UPS units took
care of the servers and network boxes. The PC
servers not attached to the
UPS were put on since they were already down and
Barbara brought the units
back on-line and made sure everything was in working
order.
- Working with Sam and Lisa on a number of installation
items. Some work
with Barbara and Sam on the Office 2000 install.
We have a few problems
to still be worked out.
From: Peter Fritschel <pf@ligo.mit.edu>
o A description of the current LIGO II interferometer design is now
posted
on the website:
http://www.ligo.caltech.edu/~ligo2/scripts/refdes.html
This site will be updated with design additions and changes.
o This week we have been having a meeting at MIT to look at sensing
schemes for the LIGO II interferometers. In attendance: K Strain,
D Sigg,
N Mavalvala, G Mueller, D Coyne, M Zucker, J Mason, J Camp, P Fritschel.
For the readout of the gravitational-wave channel, we are tending toward
adopting the so-called DC readout technique, where the Michelson is
simply
offset a bit from the dark fringe to provide a 'local oscillator' with
which the signal fields are interfered. For sensing of all other
degrees-of-freedom, we are baselining a scheme that involves a single
RF phase modulation of the input carrier light, and a relatively high
frequency subcarrier, on which has a single RF phase modulation is
also
impressed.
From: Mike Zucker <mike@ligo.mit.edu>
LIGO II ISC (Zucker)
---------------------
Participating in interferometer configuration workshop at MIT.
Working
on WBS dictionary for ISC entries and text for proposal description.
From: Janeen Hazel <romie_j@ligo.caltech.edu>
Working on LIGO II Proposal. Updated the Development Plan.
From: Sam Richman <srichman@ligo.mit.edu>
Stiff isolation system (Sam Richman, Wensheng Hua, Joe Giaime)
We have implemented two of the transformation matrices generated by
our collaborators at Stanford in a controller for the stiff prototype.
The first uses just the geometry of the stage to map the position sensors
and broadband seismometers to the locations of the appropriate (nearby)
actuators. The second sends a signal to the vertical forcers when
a horizontal force is applied to correct for the tilt coupling of the springs.
These matrices represent refinements of cross-terms that were previously
put in by inspection. Results are encouraging but have not completely
ameliorated low-frequency tilt effects.
From: Garilynn Billingsley <gari@ligo.caltech.edu>
At CSIRO Chris Walsh reports that he will be
taking a position with JDS
Uniphase as of October 2000. A replacement
has not yet been named.
Measurements of the two 15cm sapphire pieces are proceeding and are
proving very interesting. We have all been a bit perplexed about the
phenomena we see when taking phase maps of sapphire. There are usually
several sets of fringes present. We believe that we now know the source
of these various fringes and can now proceed with evaluating the
suitability of this material for use in transmissive optics.
Of greatest concern are the features which occur at very high spatial
frequencies. The features are linear and regular, with a frequency
of
~.5/mm and a depth of 20 to 80 nm, reminiscent of a diffraction grating.
These appear even when probed with a single polarization. The picture
is
more complex when a circular polarization is used, fortunately we will
not
be contemplating using the material in this manner.
For additional information about this report, contact sanders@ligo.caltech.edu