|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
The LIGO Executive Committee Agenda for Monday March 26, 2001 will be:
(Meeting time: 10:30 am Pacific Time)
Open meeting 10:30 - 11:30
Special Items:
No report.
WBS 1.2 LIGO Operations--Administration
A site teleconference was held on Thursday, March 22, 2001. Discussion included the end of February costs and action #85 concerning VISA issues. Tuesday Gary, Elizabeth and Ed will meet with the immigration specialists recently hired onto the Caltech staff (E. Sommers and H. Willis).
The next site telecon is scheduled for Thursday,
March 29, 2001. The list of current actions revised to reflect open
actions assigned through March 22, 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
LSC MEETING, LIVINGSTON
| Packages | Faxes | |
| In | 27 | 37 |
| Out | 9 | 39 |
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>
From: irena@ligo.caltech.edu (Irena Petrac)
Support (Wood)
Progress Period from 03.16 to 03.22
Accomplishments:
WBS 1.4.1.2 Project Controls (LIGO Construction)
A reminder that I require contributions for an end-of-February Quarterly Progress Report. I have requested contributions this week. Thanks!
Also working with David Shoemaker to set up assignments for Advanced R&D Annual Report. We should be sumbitting this report (necessary to get FY 2001 funding) over the next month or so.
The following Change Requests have
been submitted:
| CR-000018 | WBS 1.1.4 | Curbing for Service Roads at Livingston | G. Stapfer |
| CR-000019 | WBS 1.2 | Additional Lab Equipment | D. Coyne |
| CR-000020 | WBS 1.1.4 | Staging Building and Renovations to Existing Building--Livigston | F. Asiri |
| CR-010001 | WBS 1.1.4 | Return of Unused Construction Budget To Contingency | F. Asiri |
Press for the latest Contingency Needs Projection.
From: Kris Duncan <kris@ligo.caltech.edu>
The financial reports on the web provide supporting
detail.
http://docuserv.ligo.caltech.edu/~fireport
http://docuserv.ligo.caltech.edu/~finance
From: Ed Jasnow <jasnow@ligo.caltech.edu>
(D.Cook)
4k build up ... It is our hopes to start closing up the 4k mode cleaner
HAM chambers near the end of this week if the IOO alignment goes as
planned. We have started machining the viton safety stops and will
replace
the ones on the large optic suspensions that are currently installed.
The
gaps will be set to a minimum distance (~.5mm) allowing maximum seismic
protection, but not to inhibit the several degrees of freedoms needed
to
steer optics and allow for shifts in table tilts etc (there are no
guarantees that these improvements will prevent damage from future
shakers). The ETMy-4k is the last of the core optics waiting
for
installation, but the schedule is not set yet. We will re-vent the
X
manifold on Monday. During this time we will align the ITMx-4k optical
levers through the arm cavity baffle, mark the baffle position and
remove
the baffle so not to clip the 4k beam when we first transmit the beam
down
the tube to the End station.
2k
recovery and retrofit ... If all goes according to plans, we will begin
the realignment of the 2k core optics starting
with ITMy by Monday. The
alignment configuration will use a different
method as purge air capacity
and access to the beam tubes and chambers is
limited along with some of the
optics still installed. We will be replacing
the balance of OSEMs and new
safety stops at this time as well. As we have
found three loose magnets on
the MMT3 after it was removed we'll be checking
the balance of the core
optics in situ for similar problems. These are
optics that were thought to
be ok and not re-suspended. The RM and MMT3 were
installed on Wednesday and
await alignment. The ETMy will come out of the
vacuum bake load early next
week and will be re-suspended. The Mid station
will be staged next week for
installation. The IOO optics and 2k mode cleaner
rebuild is still on going,
but is parallel 4k activities.
We continue to have a lot
of support through out these activities.
Several folks are gaining valuable experience and added insight to
the
interferometer workings.
Hopefully we can wind up
these activities in about the next 3 weeks. (any
bets?)
OPTICS: This week has been spent locating and positioning ISCT-3,
and capturing the pickoff beams from the BS, ITM-x and y on
ISCT-4. (Kovalik/Overmeyer/Traylor/Kern)
Operations Support: Now that the LSC meeting is over, the control
room staff has resumed shifts which cover extended working hours to support
commissioning activities. This week, the Operation Specialists devoted
about 50% of their time to control room support and the remainder to various
tasks which included: Tune-up of the video titler system, LOS circuit board
repairs, Vacuum Bake support for LHO and the 40 meter
LHO optics repair support at LHO, ISCT-4 table alignment, shipping
and receiving, Particle Counter repair, and CDS Software upgrades and refinements.
(Rich
Riesen)
GC: Configured the Fore 2810 100 mbps switch to connect servers
and heavily used machines to it. This will allow ~10 times the bandwidth
the
servers have now. Put together an "audio visual" cart which has the
Dell with the LCD monitor on it. There is a projector attached to the Dell
along with a wireless networking card so that any time anyone needs
a projector, all they have to do is roll it into place and provide power.
Testing
PHP and MySQL on the new web server (after a moderately difficult compile).
Once we have put it through its paces and made sure that everything
is functioning properly we will move the web pages, ilog, etc to the new
server. We will also consolodate other things that are
running under linux/apache at the present time. This includes the "Operator
Checklist" and the associated database. Installed a network monitor to
observe overall traffic etc over our T1. It can be found at http://fulton.ligo-la.caltech.edu:3000/.
CDS: Installed FrameMaker and Adobe Distiller on the CDS machines.
There are several ways to make a PDF on CDS now. Had the local Sun
representative come in and repair an Ultra 10.\
LDAS: Working with Joe Hanson who is coordinating the electricians
to provide the necessary power in the mass storage room for the LDAS equipment.
Cosmic Ray Counter: The scintillator counter was delivered to LIGO last week. On Saturday, Ray Frey of U of Oregon, along with Joe Langdale, Mike Stewart, and Mark Coles, unpacked and inspected the device and it was placed under BSC2 with Mike Fyffe's help on Monday. It is offset from the chamber center to avoid a dripping chilled water line that also runs other BSC2.
Outreach: Rai Weiss, Bonnie Wascom and Rusyl Wooley participated in the St. Bernard School Annual Science Symposium in Lafayette, Louisiana. This was an excellent opportunity to reach a large number of enthusiastic students. A detailed article will be sent to Dave Beckett, along with a link to photos taken at the symposium, for the LIGO newsletter.
Thanks to everyone that assisted with setup, support, and take down for the LSC meeting!
| Installation&
Commissioning:
Livingston |
Other Science/EngineeringActivities:
Issues/Concerns |
Seealso the Installation web page
New magnets have been glued to the standoffs. In the case of MMT1, a
magnet
fell off MMT1 during the bake and we decided to strip this optic and
start
again, this was done and currently the guide rods have been reglued
on it.
We will glue the magnets on this and balance this optic tomorrow. All
other
repaired 2K optics passed the magnet test and we will soon start
re-stringing them.
Current status is:
1. Mode cleaner aligned with flashes observed, maximum yaw setting on
any MC
optics is less than 0.5, maximum pitch setting is 0.6
2. Mode matching telescope is aligned, with the spots centered on all
MMT
mirrors including MMT3. Maximum yaw bias is 0.5, max pitch bias is
~3
As a precaution, we decided to check all the OSEM voltage settings and
found some disturbing findings. Some particularly low (~0) and high
voltages
(~1.6) were found. So after carefully re-checking the levelness of
both HAM 1
and 2 we set about re-setting all of the OSEMs. This was finished late
Wed
afternoon and all OSEM voltages were between 0.85-1.15V. In addition
a blown
OSEM was discovered and replaced. Fortunately, given the location of
the
problematic OSEMs we should not have to do much major re-alignment.
We will do
a final check tomorrow.
Richard McCarthy is working on getting the electronics and field wiring in plac to attempt to lock the mode cleaner in air, in part by borrowing components from the 2 km IFO. An attempt may be made late this week.
The amplitude stabilized oscillator AC OSEM circuits
were tested on the
mode cleaner and MMT3. The circuits were found
to damp the masses.
Preliminary measurements indicate that the circuits
are 3 times noisier
than the DC OSEM. Further work is needed to establish
the cause of the
excess noise and hopefully to determine a fix.
We collected 3(+1) days worth of reduced data set (RDS) on disks
(~0.5TB) at the observatories and ~1TB full data on tapes at LDAS.
To
access the data at the observatories you have to have access to FORTRESS
and/or DECATUR. On DECATUR the E3 data is stored at /export/raid1/E3/.
To download data from the CalTech archive please follow Peter's
instructions
I also gave talks at the LSC meeting. (E3 summary, E2 Timing
investigation, SNEWS)
Rolf Bork
1) Found that DAQ in EY would periodically shift data from one channel
to
another. This had been seen before but had thought to stop after
changing out
the VME crate about a month or so ago. The only way I can think
of that this
happens is that the ADC module and CPU get out of sync with each other
ie ADC
gets a read command and outputs next value to VME bus but CPU does
not actually
get the data or gets the next data point instead, thereby shifting
all further
data one DAQ channel up. Another problem was seen where a GPS
timing ramp
channel would peridically move three DAQ cycles out in time.
This problem seems
to coincide with the reset of another CPU on the DAQ network.
Dave and I are
discussing the implementation of additional diagnostics within the
DAQ to catch
these errors, such that, at minimum, data is not taken for long stretches
without an indication of a problem and secondly to determine the cause
and
perhaps correct these problems.
2) Keith Riles pointed out 16Hz noise which appeared Saturday afternoon
on coil
driver signals on optics driven by the LSC. This has been seen
before at LHO.
However, when I came back later to try to track the problem, the 16Hz
spikes
could no longer be found. Checking with the operators and the
E log, nothing
had changed in the LSC, ie it had not been reset. Continued to
check these
signals throughout the weekend, but never saw the problem again.
3) We found that if more than a combination of 8 LSC tp/exc signals
were
selected, the LSC controller would lose sync with the DAQ system.
Today I ran
some tests using the LSC system at LHO. While the actual servo
calculation time
does not vary (read ADC, calc filters, write DAC output), use of the
remainder
of the 60usec allotted time varies by cycle (cycle = 1/16384 sec).
This is
because the code uses this remaining time for writing DAQ, writing
TP, reading
EXC signals, and other housekeeping. With no TP/EXC selected, the longest
time
used by the code to do the servo calcs, DAQ, and other housekeeping
is 52usec.
When Tp/Exc signals are selected, it takes approx. 1usec extra time
per Tp/Exc.
So, it looks like it's in keeping with the numbers that 9 LSC Tp/Exc
signals
being selected would drive the LSC CPU beyond the allotted time, causing
it to
go further out of sync the longer these signals are selected, eventually
dropping ADC data. This problem will become worse when the LSC
has to handle
yet more ADC channels, which is being proposed, as then even less time
will be
available for Tp/Exc signals.
If we continue to require more I/O from the LSC for ADC channels and/or
Tp/Exc,
changes will need to be made to improve performance. Some thoughts
are:
a) Spread the housekeeping tasks out over more cycles, thereby reducing
the
maximum cycle time. I will look into this and see what can be
squeezed out
here, but I don't expect a whole lot, as I/O is the key driver.
b) Push ahead more rapidly for the development of an ADC module for
advanced
LIGO and/or get Pentek to make changes to their ADC module. We
initially used
the Pentek module in a "packed" data mode, whereby two ADC channels
could be
read on each VME read cycle. However, we found that in doing this,
the Pentek
module would delay one clock cycle in providing data in this mode,
setting us
60usec behind. (This problem was confirmed by Pentek after we raised
the issue
and they studied their design). We therefore had to abandon this and
only get
one ADC data point per VME read cycle. Each read cycle takes
on the order of
1usec, so this is a costly change. Secondly, the Pentek module
will not read
out in "Block" mode. Block mode transfers on the VME bus are on the
order of ten
times faster than single reads.
c) Change the CPU from a Pentium to a Power PC. There is a vendor
of Power PC
VME processors which uses a proprietary I/O interface to the VME bus
and claims
higher performance in single read/write modes. This would be
a bigger change.
First, we trade an 850MHz PentiumIII for a more expensive (double)
480MHz Power
PC, which may or may not lead to longer servo computation times.
We would also
have to buy a new vxWorks license and learn the nuances of Power PCs.
4) Worked with Russ Wooley on vacuum system gate valve readback problems.
In
the LVEA, the three gate valve on the X arm were indicating both open
and closed
at the same time. It turns out that the Xycom212 VME card which
reads the limit
switches had been replaced at some point, but the traces which select
on board
power had not been removed (required, as we use an external power supply).
Fixing this cleared this problem, and two of the three valves read
correctly.
The third valve then needed its limit switch adjusted.
In the Y end station, we found that the open and closed switch wires
were
twisted together and tied to the same connection point in a PSI interconnect
box. This was corrected and now all LLO gate valves appear to indicate
correctly.
Otherwise I have been trying to take part in the daily commissioning
related work that needs to be done. Including the prep. work
for the E3
run.
Following on from the E3 run, trend data for the last month was
examined in order to set the gains for the PSL generic DAQ interface
card.
Working with Flavio Nocera on a new theory
regarding the apparent saturation of the AS port PD during a lock
transient. His suspicion is now focused on the bias regulator's
response
during a current limiting event. Also have been lending a hand
towards
getting users familiar with the existing PSL designs, some of which
can be
subtle.
The delivery date of the super polished substrates from Mindrum Precision
has slipped by two weeks to the end of April.
I have been looking in to the use of UV cured epoxies for fabricating
pre-modecleaners. Enquiries have been made to Masterbond and
Norland Optical
about low outgassing epoxies. At present I'm leaning towards
the NOA88
adhesive from Norland. This epoxy is low outgassing, UV curable
and has
a reasonably fast cure time of around a minute or so. A sample of UV
curable
epoxy from Norland Products has arrived. The data
sheets indicate that it would be a good candidate epoxy for use in
fabrication of the PMCs. Low-outgassing, reasonably fast curing
time and
it also has a pre-cure which will come in handy for adjusting the final
gluing position.
I spent some time this week trying to fix the broadband photodiode
for use in measuring the relative intensity noise after the pre-modecleaner
(PMC). For reasons I don't understand, there was a discontinuity
in the
the noise spectrum. A check of the photodetector did not reveal
any
obviously damaged components. However a check of the RF path
showed a large
input offset in the MAX4107, even in the absence of light. The
part was
replaced and the photodetector seems to function okay.
With the forementioned photodetector, I have just completed measuring
the
relative intensity noise before and after the PMC. Analysis of
the results
has not started yet, to see if the PMC provides sufficient filtering
at 25 MHz.
Rick Karwoski, Paul Russell, Ben Abbott, Sander Liu
This week I've been writing up the work I've been doing on the
various uses of magnets in the suspension. An update to T000119 is
nearly
finished.
D. Coyne note: Near term retrofit of the earthquake stops will be done with the reworked chamfer stop tips for the large optic suspensions (conical tip machined cleanly). The small optics suspensions will use reworked, UHV qualified, viton sheet for tips for the face stops (cut from existing cable clamp viton pieces). An improved design for the stops is also being considered on a longer term basis, with the desire to have them available for future opportunities for retrofit as they arise.
After our ISC meeting, I am concerned that
there are bits and pieces, such as microseismic and tidal, that have docs
that I am not aware of or do not exist. My primary concern is that,
now that most controls are digital, control systems are so tightly coupled
that it is difficult to totally isolate LSC from ASC from suspensions from
DAQ, etc. It has all kind of blended together into a MIMO monster.
It is therefore critical that I keep track of all servos and particularly
the I/O involved with them, as in the digital systems, this eats up most
of the allotted time. The microseismic and tidal, for example, could
impact the digital suspension controller in the end station, or LSC processing,
or whatever. I propose that we should develop a new top level diagram
of the "LIGO control system", to incorporate all the ties between systems
and all I/O presently in place and known future; Since I'm most effected,
I will start on this. I think it will be useful in the future as
well in analyzing control performance/architecture as new ideas come along.
</x-html>
Since our last report, we have finished re-installing all of our suspended
optics. All seven optics are in now, with all magnets in place.
The mode cleaner uses old-style fins on its magnets for shadow sensing,
and we are missing a side fin on one of the mode cleaner's mirrors.
This is not a critical item at this time and can be installed at any time
before we close the chamber to pump out. We have rebalanced the stack
and are now in the process of optimizing the positions of the sensor/actuator
heads.
In-vacuo beam dumps arrived from the physics shop and will go in the
chamber soon, and we have begun steering the beam from the new laser into
the vacuum chamber. The beam now impinges on the mode cleaner's suspended
input mirror, and we will begin aligning the mode cleaner next week.
Last week we spoke with Seiji Kawamura about backup plans in case lock
acquisition proved difficult with our current design, and one of the options
we discussed was lowering the finesse of the arm-cavities by replacing
the test masses. With Helena Armandula's help we have begun requesting
quotes on grinding and polishing the spare fused silica blanks we have,
and we expect to hear back from vendors either late this week or early
next week.
We have also been pursuing coatings for our sapphire test masses and
for our backup fused silica test masses, and we have ordered new magnets
and suspension wires for both. (Because they are nearly twice as
heavy as fused silica masses, sapphire masses require stronger suspension
wires.)
In anticipation of our seismic isolation system performance tests in
about three weeks, we ordered a set of accelerometers, controllers, and
mounting hardware. Many thanks to Dennis Ugolini (and indirectly,
Mike Zucker) at the 40-meter for pointing us to the right vendor and models.
We anticipate using these units for vetoing as well once we begin non-Gaussian
noise studies.
The main item on our schedule for this week was to begin repair of our final, broken OSEM controller. We have a deadline in about two weeks to have all seven OSEM controllers working and tuned, so we started work on this early this week. We tested the ones already installed and found one more that was not functioning according to spec. We have diagnosed the problem with this controller, and we believe we have corrected it. We anticipate verifying the repair this Friday (tomorrow, as of this writing). We also discovered that the mode cleaner's shadow sensors were not properly optimized and have corrected this. The optical gains are now much higher than before, and we are now able to achieve over-damping with sufficiently high electronic gains, which we were not able to do before. We are now in the process of lowering the electronic gains to give the suspended mirrors a pendulum Q of 8 to 10, which Peter Fritschel has estimated to be an optimal tradeoff between damping and seismic isolation.
Our final BNC vacuum feedthroughs arrived from Insulator Seal, Inc. Watchful readers will recall that we ordered duplicates from ISI and Kurt J. Lesker in case one vendor suffered delays. Some delays did in fact occur, but since the order from ISI has now arrived we have all the feedthroughs we need.
No report
Hiro attended LSC meeting and gave a talk on current status of E2E.
Suspension system
-----------------
In the LSC meeting at LLO, Virginio gave a talk about his
quad-pendulum model using MSE and it was well accepted. Virginio
got new information from GEO people and is going to improve the
model including blade modules as well.
Lock-acquisition studies
------------------------
(Biplab) work has been affected due to a performance slowdown of the
LIGO
sunbox and sargas computers where the e2e codes were running. Larry
and Suresh are looking into the problem because it appeared only recently.
Code maintenance
----------------
(Ed Maros) Built e2e software at Caltech.
Alfi
----
(Bruce) Continuing work on sanity checks on subnodes, connections,
and ports during loading of Alfi boxes (Problem Reports 199, 205-207).
Vicere':
* Simulation: started work to write modules for linear prediction,
with
the
purpose of exploring a way to reduce the errors
implied by loops of
modules
in a discrete time simulation.
I keep following simulation in Virgo: a comparison
with experimental
data
on transfer function is underway. Great success
of the simulation,
almost
perfect agreement! The Virgo suspensions seem to
behave as expected.
Last Week
Biplab attended Pacific Coast Gravity meeting at UC, Santa Barbara
and gave a talk on the End-to-End modeling and simulation efforts.
Hiro is attending LSC meeting.
Ed Maros installed E2E at LLO.
Physics Studies
---------------
Hiro completed work on generation of noise curves. The algorithm
to calculate the psd of noises with a wide dynamic range (e.g. Seismic)
was tested along with the box files for generating thermal noises (internal,
suspension, violin) made by Biplab.
Biplab continued studying mirror angular fluctuation effects in presence
of offsets.
Code Improvement
----------------
(Tavio) Working on three things: debugging/testing LAPACK & performance
-library changes to matrix class (which is theoretically done except
for
the fact that I can't find the actual .a file for the library anywhere
to
link and test it out).
Finding a faster random # generator (gaussian and flat) than the
linear-congruent one we're using.
Getting profiling to work on the multithreaded version of modeler.
Alfi
----
(Bruce) Working on sanity checks on subnodes, connections, and ports
during
loading of Alfi boxes (Problem Reports 199, 205-207).
Software Systems (Blackburn)
LDAS was demonstrated at the LSC meeting in Livingston this week. The
general turnout for the demonstration was surprisingly large and
generated plenty of questions. One other interesting note related to
this demonstration is that GEO participants went home and quickly
downloaded LDAS on their system for evaluation.The metaDataAPI has now been upgraded to support multiple databases
concurrently. Prior to this upgrade, LDAS had to restart the metaDataAPI
each time we changed over from development to test or production versions
of the database tables. This enhancement will also allow LDAS to make
dramatic changes in table designs which do not easily migrate and still
support access to the older data collected in engineering runs as we
move towards the LIGO I science run. The upgrade was accomplised in just
3 days of coding effort which is a tremendously powerful statement of
value of the combined modular/object oriented structure of LDAS.Support for BLOBs (binary table data) in the wrapperAPI was added this
week and tested using a dynamically loaded shared object developed by
a caltech undergraduate. This will allow search codes running on the
LDAS beowulf cluster to insert things like spectra and waveforms into
the LIGO database tables. Previously, only simple datatypes were allowed.Several bug fixes and enhancements were made to the lightweightAPI
this week. Cheif among these were support for threading and the addition
of complex data support.A modest effort went into cleaning up documentation in the data-
ConditionAPI this week. The latex documentation is being made
standalone and the perceps documentation in the FFT code is being
upgraded to reflect the the new rules.A new set of test scripts which automate the access of all frame
attributes was started this week. This will give LDAS a much clearer
view into the validation of frame software and data stored in frames.Development on the eventMonitorAPI continued this week. It is now
able to receive and register the different data objects sent by the
wrapperAPI. Work will continue on parsing this data and sending it
to the next APIs in the pipeline, e.g., the metaDataAPI.
Hardware Systems (Anderson)
LDAS hardware/sys-admin activities
----------------------------------The purchase orders for the 28TB of LDAS disk storage and 6000-slot tape silo
where signed and placed with the respective vendors. Shipping dates for
all of this equipment are scheduled for the last week in March.Ordered 7 5000VA UPS units to run the large LDAS disk systems at CIT, LHO, and
LLO and coordinated the shipping and installation of the STK 6000-slot tape
silo with the appropriate personnel at CACR.Ordered a small network switch for LHO to support the move of the LDAS servers
to the staging building scheduled for the last week in March.Successfully recorded the E3 data at both LHO and LLO. Ingestion of these
data into the main archive at Caltech have been suspended after 0.5TB to
await the imminent arrival and installation of additional redwood tapes.Installed and begun testing of a new LDAS software server (E220R) which
will free up the current E450 for duty as the database server in the
ldas-dev system.Installed and begun testing of an additional 3GB of memory for a total
of 4GB in the data-conditioning server in the ldas-dev system.Begun plans to specify and order the MIT LDAS system for the duration
of the Engineering runs.
FCT research:Data Analysis Activities
Linqing Wen is at UTB to particpate in a workshop on writing LAL-LDAS code for search algorithms. She will be working on porting FCT to this environment for MPI based analysis.
Charlton:
* Worked on (finally) getting Rick's FCT code to work properly with
the Newtonian chirp function.
In terms of tau0 and frequency, the Newtonian component
of the chirp is
tau0*(6/5)*PI*f0*(f/f0)^(-5/3)
where f and f0 are in Hz. I chose f0 = 175 based on Sathya's paper.
The chirp transform has a factor of 2*PI in front so we
want to divide that out and use
2*PI*tau0*(3/5)*f0*(f/f0)^(-5/3)
We also want to zero out everything below 40 Hz because
seismic noise is dominant:
phi = 0
0 <= f <= 40
= (3/5)*f0*(f/f0)^(-5/3) f > 40
However, putting this function in "as is" doesn't seem to work - all
I get is an image that looks like random junk, instead of a well
defined chirp. I think the dynamic range of phi is too large and the
capacity of float is being exceeded - the FCT contains a division of
two
parts of the phase function, which in this case can give a large number
divided by a very small number, which might cause this.
Things work much better if the phase function is
scaled so that it's min is 0 and max is 1. I tried phi/phi_max
instead. The tau0 becomes tau0*phi_max, so, at the end of the
day, whatever value you get for the estimated chirp coefficient
is divided by phi_max to get the real value. This seems to work out
fine.
Vicere':
* Data analysis: I am getting acquainted with issues related to frame
access and
availability. Had meetings with Alex and Philip
about accessing
frames on
directories: with Kent about needs for the next
MDC (ingestion of
LSC files
and production of frames); with Albert and Roy Williams
about the
frame
exchange with Virgo (and others) using Grid.
Last Week
E3 run (Charlton)
This week and last weekend heve mostly been spent monitoring the
acquisition of, and making plots of, spectra and cross-spectra from
the
E3 run. A selection of current results are viewable at
http://www.ligo.caltech.edu/~charlton/E3Run/
The scripts worked reasonably well, except for a couple of occasions
where
the data conditioning API crashed. In the end I was able to produce
a
set of plots roughly once an hour while the scripts were running, from
about 17 minutes of real data.
MIT:
(Keith)
-Installed new IDE drive and rebuilt ultra-10 with 2.8
tested the apop APOP/POP3 server.
-Requested new MX records for mail server/backup
tested out sendmail masquerading.
-Installed CVS, built CVS repository, and hooked up
cvsweb (interactive cvs web browsing tool).
Livingston:
(Shannon & Tom)
-Configured the Fore 2810 100 mbps switch to
connect servers and heavily used machines to it.
This will allow ~10 times the bandwidth the
servers have now.
-Put together an "audio visual"
cart which has the Dell with the LCD monitor on
it. There is a projector attached to the dell
along with a wireless networking card so that any
time anyone needs a projector, all they have to do
is roll it into place and provide power.
-Testing PHP and MySQL on the new web server (after a
moderately difficult compile under Solaris). Once we have put
it through it's paces and made sure that
everything is functioning properly we will move
the web pages, ilog, etc to the new server.
We will also consolidate other things that are
running under linux/apache at the present time.
This includes the "Operator Checklist" and the
associated database.
-Installed a network monitor to observe overall traffic etc
over our T1.
-We have ordered a Sun Blade 100 workstation for testing, and
for use in one of the labs.
Hanford:
(Christine)
-Caltech ITS has finally decided that LHO Sun's can be registered with
the Caltech Sun Scholar Pac to receive relatively free Sun compiler
software. I finally received a license file from Caltech ITS
for the
Sun Forte6 compiler software. I have spent the last three days
downloading the C, C++, Fortran and High Performance Computing compilers
from the Sun web site. Caltech ITS felt this was better than
sending me
the CDROMs. I hope to have the compilers installed by the end
of this
week.
-Installed bison software on the application server. Through
complaints
from users, I continue to identify which software packages are needed
to
be installed on the application server.
-Ordered D1000 replacement disk drives for CDS and computers and PCI
cards for GDS. Purchased software for development of Solaris
code to
run on Windows systems for porting Ligo Apps Tools to Windows.
-Removed security from the TRINET seismic datalogger PC to facilitate
off-site control.
CIT:
(Barbara)
- Continued efforts on the LDAS Equipment web forms. Finished
password-based permissions for deletions, add/update, and view only.
Finished feature to return to calling page after performing an edit.
Now
just have a couple of reporting items. Would be good to clone
this
application for all LIGO equipment.
- Met with Linda Turner about the search tool for the document log.
Have a
minor change to make to the tool. Meanwhile Linda will contact
staff with
many reserved numbers to do some cleanup before the tool goes public.
- Posted the LSC transparencies. As of today we have 82 posted
and about
20 outstanding. Updated the Employment page. Posted some
meetings;
archived 1999 meetings.
- Sent out the revised LDAS home page for review and have a number
of
changes to make.
- Worked on the contacts.cgi script a bit. The script is finished
but I am
adding some error checking and defensive programming.
- Reviewed my router woes with Larry.
(Lisa)
-Configured rana in the 40m to act as a gateway to their cdssol network.
-Installed graffias with samba 2.0.7 and began testing it with windows
2000.
-Configuring an older ultra 10 with sunpci. This required an
update to the OBP
and a second hard disk to be installed.
-Spent some time playing with wlan access points.
- I have 3 old pc's ready to be surplussed.
-The usual round of pc issues.
(Suresh)
-Worked with Larry to resolve Bridge third floor networking problem.
A FORE
3810 ethernet switch has been replaced by working one. We noticed same
resetting
problem of 3810 box as before.
-Worked with HPC technical representative (Mike), Larry and Stuart
to resolve
postscript printing problem with Laserjet 5000N printer in 6th floor
Millikan.
Finally it was found that formatter is the problem. A new formatter
has been
ordered. Hopefully, next week we will get the replacement.
-A new e-mail alias "aligo_sys" has been set up. Also modified some
aliases.
- Working on resolving the running of modeler program rather slowly
in certain
CPUs.
-Resolved some system problem with Sydney Meshkov's Sun workstation.
(Larry)
-Worked a number of procurement issues and maintenance contract setups.
Still have a few more to go.
-Helped DCC move some data around. So far the change for getting the
LSC
conference presentations on to the web has been a big improvement.
-Worked on a couple of SUN boxes. Mainly, just some minor setup and
procedure
problems.
-Trying to work out a number of VPN issues with ITS so we can have
more people
on LIGO using it.
-Finally, was able to check out some of the WAN configurations we were
looking
at for Hanford, with a FORE tech. rep.. Two of the methods we have
been looking
into will work according to FORE.
-Fixed a couple of accounts.
-Spent some time working various issues with Shannon and Tom at Livingston.
(BS) Ilog maintenance, additions, and development:
(2.5 days)
- Major fix to prevent bad
XML characters from being written to the logs
(thus breaking
the log for that day.) This has been an occasional,
but ongoing
problem. With the use of Ilog continuing to grow, it was
a needed fix.
- Setting up more Ilog services
on ldas-sw for various groups.
Last week
MIT:
(Keith)
-Migrated MIT machines to new printer server.
-Investigated bizarre behavior from Ultra 10
which ended up being an intermittently failing
IDE drive.
-Investigated POP3/APOP server (teapop)
for use with mail server.
-Installed StarOffice on several machines. Patched and installed
maintenance update v3 on several more 2.8 nodes.
Hanford:
Nothing to report.
Livingston:
-Performed a number of fixes on various computers for the E3 run.
Overall things went pretty smooth.
-A great deal of time was spent setting up for the LSC.
-Worked on getting more information on another 3rd party network vendor.
Shannon has made some progress there.
-Albert is now trying to get some information from the LAnet people
concerning
the OC3 connection for the Observatory.
CIT:
(Lisa)
- rebuilt grumium, Ultra 30, with solaris 8. I'm moving it upstairs
today.
- rebuilt Liz's old pc. It's ready to be moved out.
- Got demo licenses for Cadence microwave simulation tools. This
was a courtesy
for EE.
- chased down some email problems between Virgo and Ligo.
- I also discovered an interesting e-mail problem with someone's home
linux box
using imap off of acrux. That problem was resolved by switching
the person over
to pop.
- dealt with some netscape instability under Solaris. If you
are on a sun box
running solaris 8, netscape is part of the operating system.
It appears to be
more stable than running netscape off of an nfs mount. The path
is
/usr/dt/bin/netscape.
- I've been babysitting the new modem pool. I've had a couple
of reports of
problems from the LSC.
(Barbara)
- Finished search tool for the DCC log. It is ready for review
after folks
return from the LSC meeting. Also reworked the forms for reserving
a
document number to eliminate the Author pull-down which takes too long
to
load. Made changes to all DCC web forms and pages to bring them
up to
standards.
- Continued work on LDAS forms for equipment inventory. Have
password-based privileges just about working.
- Set up Access database and web pages for the LSC transparencies.
Created
a new logo for the LSC.
- Installed various web site changes -- LSC, publications, PAC talk,
construction accounts, etc.
(Suresh)
-Moved directories under /export/home3 and /export/home4 in sirius
to 36 GB
drive in Sun Storage D1000. Now directories under these two home directories
are
combined to one single home3 directory. Subsequently, modified auto_home
file in
sirius to reflect thsese changes.
-Received bottom paper tray with housing enclosure from HP and replaced
it in
LaserJet 5000 N printer in the third floor bridge. Sent back the old
broken
module to HP.
-Downloaded FORE ES 3810 and ES 2810 switch firmware's from Marconi
web site
and set up in System Admin Laptop.
-Did some testing on ES 2810 switch with 10 db attenuated single mode
ATM
uplink module. This problem is not resolved yet. Seems like it need
to be sent
to Marconi for further diagnosis.
-Installed application software's, set up printers and e-mail account
in Ken
Mailand's PC after Lisa passed me on with Windows 98 installed.
-Created some e-mail aliases, fixed minor printing problems and help
users in
various ways.
(Larry)
-Spent the last week working with the E3 run at Livingston and assisting
with the setup for the LSC conference.
[Bruce Sears]
* (BS) Ilog maintenance, additions, and development:
(1.0 days)
- Continuing support of
LLO Ilog move to a different machine.
- Setting up new Ilog service
on ldas-sw for various groups.
Peter King
PSL for 40m and LASTI
More components continue to trickle in. Just about all the
mirrors, waveplates and lenses from CVI Laser have arrived. A
fraction
of
the mirror mounts from Newport have arrived.
Jay tells me that apparently the Motorola 162 IOCs ordered will not
work
for
us and will have to be returned for a swap.
From Riccardo DeSalvo
We had no meeting this week only a few remarks.
Finally got hold of some DAC units, unfortunately they will be in Italy
until early next week.
Feed through flanges, water cutting machines down until Monday, we will
get them in Pasadena middle of the week, will mount cables end of week
or weekend trying to ship to Hongo on Monday next week.
Accelerometers, also some delays, all parts should be delivered early
next week to Alessandro. All the rest is ready to go.
Gianni has introduced the modifications and the real vacuum tank
geometry in drawings, after approval need to start the production of
the
two double nail head wire, should do by middle of the week. Also
need
to redesign coil for F0 actuator. Will need force versus position
characteristics from Akiteru for old coil supports.
New hard clamp and wedges being built, soon ready to test for
hysteresis.
For additional information about this report, contact sanders@ligo.caltech.edu