LIGO Weekly Report

ALL REPORTS ARE TWO WEEKS: 9.27 to 10.11
WBS 1.2 LIGO Operations--Administration

There was a teleconference on Thursday, October 4. A number of items were discussed:

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. . .

**NOTE**Linda has returned from her medical leave.Welcome Back!!
>From: Cleveland Mak <mak_c@ligo.caltech.edu>

PLEASE NOTE: This report covers three (3) weeks due to illness on reporting day last week.

Advanced LIGO (Frey)

A draft copy of the quarterly progress report for the end of August 2001 has been distributed for comment.

LIGO Change Control Board met on October 9, 2001. The following Change Requests were discussed:.

These minutes are sans signatures. Copies of the signed Change Requests will be distributed via hard copy.

Quality/Safety (Tyler)

LIGO Hanford Observatory (LHO) Operations (Raab)

LIGO Livingston Observatory (LLO) Operations (Coles)

Interferometer: The new periscope was installed on the PSL table. Its main resonance is at 205 Hz instead of 100 HZ as in the old one. This will improve the frequency noise of the laser below 100 Hz. Before and after noise spectra are posted on the LLO i-log 10/9/01, and spectra with damping material can be found on 10/10.

A decision has been reached to provide an acoustic enclosure surrounding the PSL table, and a recommended solution has been offered to Stan & Dennis. Laser Safety Interlock installation is moving along well. All ISC table interlocks should be completed today. Beginning revision of Laser SOPs to reflect the forthcoming changes. (Jonathan Kern)

GC: Win2kserver has crashed. All of the domain data is gone. I have tried restoring it by several methods, but have not been successful so far. Print services on this machine are no longer available. The stored files on this machine are intact.
LDAS: Ordered the QFS licenses for LDAS and CDS. (Shannon Roddy)

CDS: Installed latest version of Perl and Tk on CDS network. Checked all processors to boot from the target area. Added 4 new channels of the PST readbacks to the master.config list. I am planing on updating the operating system on all CDS
work stations to solaris 8. cleaning up mass storage room rack layout. (Chethan)

Other: LIGO Livingston has hosted tours for a wide range of visitors in these last two weeks - Gulf Coast Travel Club, Jones Creek Academy's 2nd Graders, Doyle High School's Physics Class Seniors, and Istrouma Area Girl Scouts.

Detector/Technical Support (Whitcomb, Coyne)

Commissioning:

Hanford
Livingston

Science & Engineering Support

See also the daily electronic logs for the installation and commissioning activities:

Hanford Detector Log

Livingston Detector Log

LHO COMMISSIONING

Much training of operators in aligning the interferometer, locking the various pieces, etc. Pretty confusing when you have two people trying to align the Y arms of the 2 km and 4 km interferometers simultaneously. Apparently the 2 km cameras have greater sensitivity to the 4 km light, and vice versa.

2km Commissioning

Stan Whitcomb reporting

The mystery of the frequency jumps in the 2 km laser is solved, sort of. We found that the pump diode current control signal from EPICS, violated the recommended grounding scheme from Lightwave. When we removed this signal, the frequency jumps went away. We also saw that the actuation coefficient for the slow actuator went up by a factor of 3. No one has a good model for this, but if you say "grounding problems" forcefully enough, most people will nod and go along with it.

Aligned and locked the 2 km interferometer again.

Measurements of the noise in the coil drivers of the large optics suspensions controllers suggested that this might be contributing to the observed interferometer noise. This was confirmed by turning off the local damping after the interferometer had locked and seeing the noise level drop in the region below about 200 Hz.

As a consequence of this finding all the large optics suspension controllers on the 2 km interferometer were modified to incorporate the most recent DCN's and other work.

*Jumpered out the first two stages of the Chebychev filter. Also removed the coil driver compensation.

*Grounded the unused pins of the Max 333 switches.

*Removed the input Max 509 chips and replaced them with sockets and 3k Ohm resistors to eliminate input noise in POS, Pitch and Yaw input matrixes.

The controllers have been reinstalled (except for the BS which remains on a small optics controller) and realignment has begun.

4-k Commissioning

Stan Whitcomb reporting

We finally succeeded in locking the X arm of the 4 km interferometer. The key steps were sorting out which control signals went to which end, how to get the digital suspension controllers into acquisition mode, and reducing the frequency noise of the light entering the interferometer below about 10 Hz. With these changes, the X arm locked for periods up to 10 minutes. It was even possible to shift the suspension controller into Run mode, though the arm typically fell out of lock in 10 seconds or so.

The MMT was realigned to better center the beam in the recycling cavity, but this did not significantly improve the clipping on the dark port. After a careful assay of the clear aperture of the APS telescope by tilting the BS, we attempted to clear the clipping by adjusting the HAM4 table. To my slight surprise, we were largely successful. The beam is pretty clean coming out of the vacuum system, maybe a little flattened along one side, but not bad enough to stop further alignment and locking.

Commissioned the EO shutter on ISCT1, and got the first demodulated rf signals from the reflected port photodiode. Working on the WFS for the modecleaner.

We then succeeded in obtaining short locks for the Y arm on the 4 km interferometer and have light on the Y end QPD. Resosnces are clearly seen in the DAQ signal.

LLO COMMISSIONING

Rai Weiss reporting

The new periscope was installed on the laser table. The major mechanical resonance, associated with the periscope and seen in the laser frequency spectrum, was moved from 100 Hz to 200 Hz.

Bids have been collected for the construction and installation of the acoustic isolation enclosure to be erected around the laser table.

Work continues on the wavefront sensors for the mode cleaner and the differential motions sensed at the antisymmetric port of the main interferometer.

Displacement data along one arm was taken to establish the coefficients in the model for the seismic feed-forward control. The wiring for the fine actuators was completed two weeks ago during Jay's visit to the site. Another run using data from both cavities is scheduled for the weekend of October 13.

We had an accident with the new high current coil controllers due to a cable that had been moved and/or a failure mode that had not been considered in the original design of the over current safety circuit. The design was to turn off the coil driver if an average current of 150ma lasting for 200 seconds was applied. In the accident the ETMX survived a current of 320 ma in each of the five control coils. The coils were held at approximately 85C for 8 hours. In the early part of the accidental heating, the pressure in the test mass chamber rose from 1.4 to 7 x 10^-8 torr. RGA scans made after the accident showed no hydrocarbons. The increase in the pressure was most likely due to water. Measurements of optical scattering by the HR surface of the etmx while the X arm cavity was locked showed no increase in scattering. We survived the accident. Several actions are being taken to avoid another incident:

cable connectors are being changed so that it is not possible to misconnect the cables,
additional sensors for failure of any of the power supplies attached to the coil controller are being incorporated.

Peter King
I measured most of the transfer functions of the common mode servo card and found them to behave as documented in the as-built schematics provided by Rich and Flavio. In some cases the switches were jumpered out to permit the measurement - as it was measured in the lab and not out on the floor -these jumpers will be set back to their installed positions. The electrical switches will then be tested out when the servo is plugged into a Eurocard cage

SCIENCE & ENGINEERING SUPPORT

Electronics

Dennis Coyne

A couple of weeks ago we instituted a weekly, project-wide CDS meeting to communicate electronics problems, issues, failure reports, reliability concerns, etc. The meeting is 11:00 PT Wednesdays. Representatives of the commissioning efforts, observatories, 40m, LASTI and Lab engineering attend the meeting.
A problem reporting and tracking system for all detector hardware (not just electronics) is being implemented by Bruce Sears (another instance of Gnats)
An electronics traveler (E010180) has been implemented for conveying the initial acceptance testing to the end users and to capture the revision update and repair history for all of the electronics modules

Suspensions

Janeen Romie
Received an e-mail from Bill Tyler outlining his concepts for preventing a wire mixup like the molybdenum incident. I will implement them as soon as possible. Working with Myron, Todd and others on accounting for all new osems and providing more spares for the sites.

IO Periscope

Mike Smith, Ken Mailand

The fabricated parts for the next 3 IO periscopes were completed. These units are for Hanford (2km & 4km IFOs) and LASTI. The periscopes are being welded. We are on schedule to have at least 2 periscopes for Hanford completed by 10/19.

*** Note: See Livingston elog (username: reader, password: readonly) for frequency noise spectrum reduction with the new IO periscope.

LLO Seismic Feed Forward

Rolf Bork

Finished up code for LLO that reloads filter coeffs on the fly from the reflected memory. Ed Dawe has begun testing. Filters seem to work as designed. Note that I am temporarily using WFS2 thru WFS4 GDS excitation points as inputs to microseismic filter banks to facilitate testing. If either WFS2-4 start to be used and/or this design is finalized, permanent GDS excitation points will need to be added.

Pentek DAC Noise Testing

Jay Heefner, Lori Robison

We have performed some tests at CIT and it appears as though we have two problems with the pentek DACs:

The output noise of the device goes up by several orders of magnitude when the output clock rate is set at 2^20 Hz. This high output clock was selected in order to minimize the delay in the system. It appears as though the penteks need to have the output clock rate set to something less than 131072Hz in order to meet their noise specs.
The code is not deterministic to a level better than 1-2usec. This causes us to jump back and forth across clock boundaries and introduce noise into the system. This level of uncertainty pushes for a clock frequency less that 131072Hz.Once the tests are completed we will summarize the results and call a meeting to discuss what we should do to correct the problem. Right now it looks as though an output clock rate of 65536 would fix both problems, but this could introduce unacceptable delay for some of the higher bandwidth servos in the LSC.

In our testing we have also discovered several penteks that have inherently noiser DAC channels. We are in the process of developing a test procedure so that all modules can be tested and the suspect units returned to manufacturer for repair. The noise tests performed at CIT used GPS derived timing signals, the same as

at the observatories. By doing this and confirming that the DACs meet their noise specs (with lower output clock rates) we have indirectly confirmed that the timing system is not introducing phase jitter and noise into the system, at least at a level that can be determined by the testing. At some point we will need to measure the jitter on the GPS clocks and determine at what level it will become the limiting factor in our input and output noise specs.

DAQ RAID

Rolf Bork

Received new RAID enclosures (rack mount replacements for initial incorrect shipment of tower units). In process of configuring units (takes 5 hours), then will test on development DAQ system.

PSL

Peter King, Lee Cardenas

Examination of the spectrophotometer data provided by REO on the new pre-modecleaner mirrors suggested that the coating was not as specified. Lee is currently in the process of measuring the mirror transmission. If found to be inconsistent REO have said that they would re-work the parts.
Measurements (Beam waist, Rayleigh length and M square etc..) in progress for the new NPRO S/N 259
The LLO PSL data for e2e has been gathered. Currently I'm in the process of format conversion.

Rick Karwoski

We have been working to put hardened QC measures into the board build-test cycle. A formal procedure is now in place. We will continue board testing once test procedures for each of the board types are completed.

Sander Lui
Finished writing the preliminary draft on the FSS Servo card Acceptance Test Procedure. Rita is helping me put it into a better format before I circulate it for review. In the process of writing the Acceptance Test Procedure for the PSL Frequency Reference Card.

LSC

Rich Abbott, Mohana Megeswaran, Flavio Nocera , Szabi Marka

RFPD production continues. Some problems have delayed shipments. Mike Zucker has repaired two modules which will be shipped to the observatories.
In-situ testing of the modified timing module (by Rolf Bork and Szabi Marka) has been successful. The prototype of the new timing board works well when installed in the excitation engine rack. The extended ramp results in a more accurate fit. Production will begin (order of 60 units).
Parts are on order for the new 2 omega box.

Luca Matone

In the month of September, a few closed loop measurements have been taken on the 2k LHO simple Michelson, X and Y arms. The goal was to first address the frequency response of the mirror actuators and then account for all the phase in the X and Y arm closed loop measurements.

Addressing the mirror actuators:

In order to estimate the actuator's frequency response, we locked the simple Michelson to the laser by driving FMX with a bandwidth of ~10Hz. We then began to drive FMY from 10 Hz to 1 kHz, taking the transfer function between the ASY error signal and the excitation. To first order, this transfer function is the actuator's tf. The magnitude of the two measurements show, as expected, a 1/f^2 dependence from ~10 Hz to 1 kHz. The phase, however, differs from the expected -180 deg, phase leading by 30deg at 20 Hz and phase lagging by 50 deg at 1 kHz.

It turns out that even though the first order approximation in use explains the frequency dependence of the magnitude, it does not predict the phase frequency dependence.

The phase measurements were then compared to the output of a model that included the electronics recently characterized (Anti-Aliasing, Anti-Imaging, delays, and the digital filters) and a simple pendulum tf for the actuators. An overall gain was adjusted in the model to make it agree, in magnitude and phase, with the data. The phase agreement, to within 10deg, was achieved without adjusting (or adding) any parameters in the model, just adjusting the overall gain.

Addressing the closed loop tf measurements on the X and Y arms:

We measured the closed loop [1/(1+OL)] response of the two arms, for different gains, different frequency ranges and at different times. We compared these measurements to the output of the model described above, assuming a cavity pole at 180Hz. Once again, to achieve agreement, only the common gain was adjusted to fit the data (no artificial parameter added and parameters not adjusted). The result is a semi-quantitave (certainly qualitatively) agreement with the measurements: <5dB off in mag and ~10deg off in phase.

Accounting for the phase:

The digital filtering enabled to lock the cavities consist of the usual combination .1.1:1,10 & 1k:10,100 & 2kBW. To account for the phase, we also have to add the non-negligible filtering properties of all the electronic boards in the system (this is true for frequencies above 100Hz). A table is given below giving the phase contribution phi (in deg) of each board:

	phi(@100Hz)	phi(500Hz)	phi(1kHz)
1.1:1,10 & 1k:10,100 & 2kBW (analog version)	0	0	0
the above is sampled at 16 kHz à to the z-domain	-2	-10	-15
the above + Anti-Imaging board Rev A	-4	-18	-35
the above + Anti-Alising	-7	-33	-65
the above + delays (computation and travel time to/from mid)	-10	-48	-100

GDS

Daniel Sigg

Added the dmtviewer code into the new gds tree of the cvs archive. This incorporates Masahiro's histogram extensions. The event analysis code has also moved into the new cvs archive. Compiled the whole thing on a solaris/CC and a linux/gcc machine. The solaris/CC compilation also required a fair amount of patches in the standard DMT distribution.

DMT

John Zweizig

This week I contined to make the new version of the DMT software ready for prime time. I also worked on installing the current version of the DMT software at the 40m and on organizing the PEM Deep Mine effort of the Burst analysis group.

40 Meter Interferometer (Weinstein)

Thermal Noise Interferometer (Libbrecht)

Our main efforts at the TNI are involved with hanging our new set of test
mass mirrors, which is relatively straightforward but time consuming.
Since the beginning of classes is soaking up quite a bit of Eric's time,
this task is currently moving slowly. We expect the pace will quicken
again very shortly.

        On another front, Shanti has been making progress on the photothermal
experiment, which seeks to probe thermal conductivity and thermal expansion
in a mirror surface. To do this, we have built a low-finesse IFO cavity,
with a flat "test" mirror at one end. Resonating in the cavity are two
cross-polarized laser beams: a pump and a probe.
        The pump beam, about 350 mW and horizontally polarized, is chopped at a
frequency between 100 Hz and 4 kHz. This drives a fluctuation in the light
power stored in the cavity. A fraction of this power is absorbed at the
IFO's end mirror, which expands and contracts in response to the pump beam.
        The probe beam, about 80 mW and vertically polarized, has 14.75 MHz
sidebands, and is used to measure the length of the cavity. This length
signal is also fed back to the laser, to keep it resonant with the cavity.
        To prevent stray light from the pump beam being reflected along the probe
path (and acquiring sidebands), we have been testing various polarizing
schemes, using a combination of polarizing beamsplitters, Faraday
isolators, and plane-polarizers. One of the challenges is that nearly every
transmissive optic is slightly birefringent, especially when stressed.
Through careful adjustment, this type of systematic noise is approaching a
level comparable to the background electronic noise.
        Unfortunately, we are finding that the polarization contamination is also
non-stationary. We suspect thermal effects in the transmissive optics, so
the noise may diminish once the entire experiment has been allowed to
settle a bit. There are also slight alignment errors in the polarizing
beamsplitters that we are working on correcting.

LASTI (Zucker)

LASTI (Bayer, Fritschel, Harry, MacInnis, Mason, Miller, Mittleman,
Ottaway, Phinney, Rollins, Shoemaker, Zucker)
=======================================================================

Grout under the BSC piers continues to grow stronger. A cleaning crew has
cleaned and rewaxed the high bay floor and we're now moving equipment
back in.

Hydraulic actuators:

Hydraulic pre-isolator design
A decision has been made to use counterwound springs to reduce the
weight on the hydraulic actuators. Springs for use on the BSC chamber
have been designed and sent to several spring manufacturers for quoting
and any helpfull suggestions on their manufacturability.

Our GDS RAID disk tower was swapped and a rack-mount version has
now come back; will install and test end of this week.

PSL:

Progress on the PSl is now back in full swing, the stand in reference
cavity has been aligned and we are now in the process of trouble shooting
the servo. The laser arrived and will be commissioned as soon as the
enclosure arrives (~2 weeks).
Dave

Data Analysis and Computing (Lazzarini)

E2E Physics meeting
-------------------
Recent activities at sites (Matt), LSC characterization (Luca), Mechanical
simulation (Virginio), inclusion of seismic noise correlation data in Han2k
(Bill Butler) and the mode mismatch problem (Hiro and Biplab) were discussed.

Modal Model
-----------
(Hiro) When the field in a cavity is calculated using up to n+m = 2 modes,
then the correction to the TEM00 field due to the mismatch comes out
to be proportional to eps^4/(1-R) after locking, where eps is the
order of the mode mismatching. For the high finess cavity, this
correction is not negligible. The steady state field calculation in a
cavity using mathematica has been extended to include up to m+n=4. If
the correction term, eps^4/(1-R), is proved to be cancelled out by
including the 4th order mode, we need to expand e2e to include the
mode up to n+m=4.

e2e installation
----------------
(Ed Maros)
- Started work on compiling system for MIT
- Worked on updating installation documentation

Alfi
----
(Bruce)
- Working on event processing for port and connection widgets for the Alfi5
display.
(Melody)
- Continued working on displaying member nodes in alfi5. Worked on displaying
the nodes' icon images, colors, location, and some mouse actions.

Spent the week making the last development changes for LDAS
in preparation for a code freeze on Friday. We will test the
pre-release over the weekend and early next week with plans
to have the next release of LDAS available middle of next
week.

Identified a problem with the new libtools and automake for
builds of LDAS on systems that have never had LDAS installed
before. This is being looked into and should be fixed for the
release.

Continued working on documentation for LDAS systems running
on a single computer. Added to the LDAS documentation "How-To's"
instructions for installing and configuring the LIGO database.

Worked with the Universiy of Wisconsin Milwaukee folks to
configure and debug LDAS for use in a Griphyn demonstration.

Added the ability to run the frameAPI without a frame file
cache for systems that don't conform to the normal LDAS
directory tree. This will increase the frame I/O access times
but allows people with LDAS on laptops to configure the location
of frames arbitrarily.

Added and tested new support to insert database query results
into the dataConditionAPI and also to pass them along to the
wrapperAPI.

Implemented 1-sided PSD and CSD of real sequences, 2-sided for complex
sequences to the dataConditionAPI. This will be checked in prior to
code freeze. Also verified scaling factors and units are consistent
with the new SCCB requirements.

Modified unit tests for new PSD/CSD functionality and added some more
checks and tests eg. check for satisfying Parseval's thm as specified
by SCCB to the dataConditionAPI.

Finished the support for a resource file in the wrapperAPI and tested
in standalone wrapperAPI. Will be making changes to the "How-To's" to
reflect these changes.

Made all the changes to how the eventMonitorAPI and frameAPI construct
data for writing out process frames per the University of Texas
Brownsville specifications.

Did some HPSS stuff including adding tapes to storage classes, labeling
and importing 9940 tapes into the new silo and addressing some user HSI
problems.

Tested dual port Fibre Channel card in various E450 slots: the best
performance (as expected) was with the card in a 66MHz 64-bit slot and
(somewhat unexpectedly) with the QFS DAU at 128K. A single port of the
dual port card in this configuration was a little less than that of a
single port card, ~70Mb/sec writes, ~53Mb/sec reads for single threads,
~77Mb/sec writes, ~92Mb/sec reads for 8 threads (the equivalent single
port card numbers were ~80, ~54, ~80, ~92). But using both ports of
the dual port card didn't come close to doubling that, coming in at
~90Mb/sec writes, ~78Mb/sec reads for single threads, ~100Mb/sec
writes, ~170Mb/sec reads for 8 threads.

Spent much time trying to get interswitch links to work on the FC
switch, so far unsuccessfully. A call is in to Sun.

Found out he GC reassigns IP's based on user not machine. I had a
conflict with my laptop and a new user that was brought
on-line. Installed SCSI pci cards in both Caltech admin machines, also
put two 18g drives on the above machines. The added space will be used for
Greg's backup script.

Set up web based tool for updating test shared objects into the
LDAS system at MIT to support a rapid software development cycle there.

Finalized the configuration for the next Sun equipment order and
am awaiting confirmation on one last technical question before placing
the order.

Developed a migration path that provides additional support
to commissioning efforts via LDAS storage resources while allowing
LDAS to continue to develop over the next 6months to get ready for
the science run.

Worked on rebuilding of /ldcg.
       1. ccmalloc-0.3.4
       2. Updated libpng for solaris to 1.2.0
       3. Updated to lesstif-0.93.12 for Linux
       4. Updated to patch-2.5.4 for Solaris
       5. TCL and TK installations now follow regular naming
          conventions.
       6. Updated to wxGTK-2.2.7 for Solaris and Linux
       7. Updated to xemacs-21.1.14 for Solaris
       8. Standardized xerces package name

1) Worked on knownpulsardemod DSO code. Committed latest changes to cvs
lalwrapper archive on 10/08/01. Started writing script to submit LDAS
jobs to run the DSO on a LDAS beowulf cluster.

2) Dealt with follow-up issues with fssnap and backups. (The Solaris 8
fssnap command takes a snapshot of a file system.) I've learned from
Stuart that the following error message

> The problem here is that xntpd runs in the realtime scheduling class
> with a directory on / as its current directory. The OS won't let you
> lock a filesystem if something in the RT class (be it xntpd or anything
> else) is resident on it.
...
>Already logged as bug #4458695
...
[Solution:]
>
> You must stop all processes that runs in the RT-class that is started
> from /.
>
> The most common culprit is xntpd. Stop xntpd during backup.
...

3) Moved a fiber channel card from the ldas dataserver at LHO to the
framebuilder fb0.
The T3s /cache/t3-13, /cache/t3-14, /cache/t3-19, and /cache/t3-20 are
no longer
mounted on ldas dataserver. These t3s will be on loan to the CDS
framebuilder, fb0.
Currently this gives the LHO control room a 23 hours data cache; this
could be increased to 69 hours once some bugs are sorted out.

4) Placing an order with Sun for 4 workgroup server QFS licenses. QFS is
part of the longer term solution to manage data services to CDS and LDAS
from the framebuilder.
Data Analysis Activities

* met with Massimo et al. several times regarding the FCT grid size vs.
grid size obtained from fitting factor in templates. We seem to have
reached some conclusions that we can make the FCT match the
minimum required template fitting factor.

* Am in the process of writing code to generate time-domain chirps and
verifying that the FFT of them matches the frequency-domain templates.
Main issue is making sure our conventions match those used by Teviet
Creighton's chirp generating code

* Implemented 1-sided PSD and CSD of real sequences, 2-sided for complex
sequences. Will be checked in prior to code freeze.

* Modified unit tests for new functionality and added some more checks and
tests eg. check for satisfying Parseval's thm as specified by SCCB.

Livingston:
(Tom)
-Still working on the new T1 line. Just working out some of the bugs but
so far it has been working well. Working with LSU in getting the PIX
system installed but that is still some time off.
-Working on getting the VRVS system installed.
-Purchased a P4 Dell workstation for the Electronics
Room to be used for drafting and circuit-board layout.

(Shannon)
-Win2kserver has crashed. All of the domain data is gone. I have
tried restoring it by several methods, but have not been successful so
far. Print services on this machine are no longer available. The
stored files on this machine are intact.

Hanford:
(Christine)
- Installed security patches for MS Office 2K on all guest PCs and
laptops. Provided written procedures for others to update their own
PCs.
- Tried to get NIS+ working on the new server. Discovered that even
though the old NIS+ database files get copied correctly to the new
server, all client computers and user accounts must be reinitialized as
clients on the new server. Thanks to Tom and Larry for helping me. I
will be providing procedures next week for all users to reinitialize
their passwords on the new server.
- Continued work on the Windows 2000 server setup and on setting up the
backup software to backup both the UNIX user accounts and all the office
PCs.

CIT:
(Mike)
-Updated a users Laptop, that included adding a license, for a networking
utility called 'Net switcher' that runs a batch file that stores its license in
the registry.
-Working with Larry, we worked the server room by moving equipment around to
install new switches to upgrade to a GigE network.
- For Luca Matone, loaded new Laptop for him that included all General Computing
software setup up security and networking.
- PC upgrade by adding memory, blowing out the system with
compressed air. The software that was currently on his computer had to be
reinstalled due to registry errors and computer hanging up on the shutdown
process.
-Applying security patches for multiple users in regards the Microsoft
security report that included security holes in Excel & PowerPoint.
-Charlotte Py ran out of disk space instructed to install another hard drive
that ended up being a slaved.
-Moving software & computing components down to sub basement from Barbara's
office and from my office.
-NT administration updating recovery disks and diagnosing errors on servers
plus going over security instructed to tighten up all servers by Larry.
-Onsite & phone user support, that included printing and networking problems.

(Barbara)
- Reworked the LDAS contacts.cgi script for a revised format for the mail
aliases file.
- Refreshed the LSC reports database for August 15, 2001 reports and plans.
- Installed the September newsletter and other quick web site changes.
- Completed documentation and did lots of organizing and housekeeping.

(Lisa)
- Still building vega for web development
- Testing incompatibilities with the tape drives in the cybernetics robot and 2
stand alone drives. We have consistent read errors on tapes using both ufsdump
and tar. Tapes made on the robot system can be read on another robot system but
not on the stand alone. Will be contacting Sony and others to get this one
resolved.
- Resolved a cadence locking problem for some EE folks.
- Burning CDs for those who are leaving us.
- Updated the backup scripts.

(Larry)
-Worked a number of procurement issues, including; procurements for the LDAS
group.
-Finally, took the training course for the new P-Card system. As soon as we get
the s/w from Criselda we will start installing it on the local PC's and ship
copies out to have it loaded on the PC's at the Observatories.
-Setup some new web services for different groups.
-Working on getting things transitioned from the people that have been leaving
the project. A number of WEB items have been turned over to Veronica. Mike has
taken on the responsibility for maintaining the servers. Other duties/jobs will
be redistributed over the next few weeks.
-Running tests on different video cameras. It appears that we have found a
couple that have the resolution and size that would be acceptable for some of
our conferencing. Everyone in the group is investigating different ways to
improve the video conferencing used on the project.
-Worked on a couple of PC's and patched a couple of the SUN servers.
The latest SUN recommended security patch should be installed on all SUN boxes
ASAP.

LIGO II/Advanced R&D (Sanders)

adv. LIGO Optical Modeling
Bill Kells
Working to make the final statement for our paper on MELODY with regard to SB imbalance. That is, is this intrinsic to imbalanced mirror imperfections, or can it be tuned away by longitudinal offsets. Also preparing a final statement for next week's design review on the Carrier matching into a highly (thermal lens) distorted full ifo.

Erika D'Ambrosio
Much time was spent in editing drafts; the one on 1/f noise simulations is now reviewed by Phys. Rev. E and the one on Melody still needs some adjustments. I also finished a note on the parametric instability and Bill is going to check the equations since I had found every solution is damped, basically because of a forgotten minus sign.

adv. LIGO Suspensions
Mark Barton
This week I continued work on my quad suspension model. I did a lot of debugging to check that the wire bending terms I introduced were doing what they were supposed to (they're only small perturbations, so they're unlikely to produce nice easy-to-spot errors).

Silicate Bonding
Bonded two different styles of the GEO ears to a fused silica mass for a static load test.
Phil is welding fibers to them, then we plan to load them and see what happens.
I am making some silica/silica bonds to test their strength after the bonds are exposed to direct heat

From: David Shoemaker <dhs@ligo.mit.edu>
Subject: SWG: NOTES and summaries, telecon 8.30 AM Pacific, Friday 5 Oct

still in costing review process, no bottom line on cost, scope, or exact timing of proposal

2) possible interferometric suspension sensor (David Shoemaker)
(see http://www.ligo.caltech.edu/~dhs/Adv-LIGO/interferometric-gray-paper.pdf)

3) update on the quad characterization and research (Rich Mittleman) (see below)

The first prototype all metal quadruple suspension and quadruple
reaction has been operating at MIT for about 2 months now. It is mounted
on an isolated platform in a vacuum can, although currently there are no
plans to evacuate the can.
We are currently working on characterizing the system to see if it
agrees with the existing model. This involves measuring the normal mode
frequencies, eigen vectors, Q's and coupling to the actuators. In the near
future we will start investigating the control available when actuating
between the third mass in the reaction and main chain.
Quite a bit of data has been gathered in the form of transfer functions
from some combination of actuators driving the upper mass to position
sensors mounted on the frame, observing the lower masses.
For both calculation and measurement purposes the modes can be divided
into four groups, Vertical (4 modes expected), Rotations (4 modes
expected), Longitudinal-Pitch (8 modes expected), Transverse-Roll (8 modes
expected). The data taking/analysis has progressed the furtherest for the
vertical modes, some of which can be viewed in
http://tintagel.mit.edu/~richard/Telecon/
(more data can be found with a bit of exploring)
ModelVerticalDisplacement, "Model Vertical Eigen Vectors" shows the
predicted mode structure.
Vertical Power and Transfer, "Vertical Power Spectrum of Mass 2,3,4
(9/11-12/01)" and "Transferfunction Magnitude from input to Position
Sensors" show the power spectra and transfer function from a vertical
drive on Mass #1 (uppermost mass) using the actuators with no vertical
damping. There are four labeled peaks which seem to be vertical modes.
The first three labeled peaks agree very well with the model in frequency
and qualitatively in amplitudes. The fourth predicted mode does not agree
at all in either frequency or amplitude with any feature that I have been
able to identify, admitly this measurement was taken without the enclosure
and could be improved by repeating it with the top on.
Vertical Damping, "Vertical Damping Ratio", is the above transferfunction
divided into a similar run with the damping on the vertical actuators
turned on to between 1/4 and 1/3 of full scale. The four label peaks show
up as dips, while other features show a dispersion character.

Topics for discussion and thought
- what constitutes "verifying the model"? and how far should we go?
- identifying all 24 modes
amplitudes, frequencies, undamped/damped Q's
- measuring isolation, high/low frequency
- complete set of transfer functions
- Should we replace the second mass?
Both of the top two masses (which are in the form of sandwichs to
constrain their blade springs) had to be reinforced to reduce the
flexure of the plates, resulting in a change in position of the
center of mass. The fine leveling adjustment on the top mass
worked very well and we will probably want to incorporate this
into the second mass, which is quite difficult to level as is.
- Any further thoughts on the global control experiment
Some electronics problems encountered
- oscillation (~30Hz,62Hz) in local damping, which aren't always
repeatable.
- very high frequency (~1MHz) oscillation in the sense monitor (I
think)
- the local damping electronics are very compact, which makes then
extremely difficult to modify.
We are currently using two DSPT Siglab I/O devices. These each provide
4 A-to-D inputs and 2 D-to-A outputs. In the near future we will be
switching to a DSpace control unit which not only has more I/O channels
but also has the ability to implement real time digital filters for
adjusting the local and global controls.

SWG progress report of the Syracuse group
submitted 3 October 2001 by Peter Saulson

In the anelastic experiment, Scott Kittelberger is working on
getting data from the silica block now hung from its all-silica
welded suspension. Two problems he is dealing with are the high
Q of the suspension modes, and what appears to be excess fixed
birefringence from the welding operation that attached the
stubs to which the silica fibers were welded.
Scott and Sasha Ageev are also checking the overall
calibration of our birefringence apparatus, against calculations
that Sasha has done with the Algor FEA program.
Sasha and Steve Penn have measured Q values of a new
9 mm thick rod of Suprasil 2. It was disappointingly low,
below 20 million, as have all of the samples we have measured
from our last shipment of 20 mm thick stock of Suprasil 2.
We are trying to figure out how these samples might differ
from our previous supply. We would have expected Q of 50 or
60 million at least, or even higher if the trend of higher Q with
greater diameter had continued.
Andri Gretarsson has built a capacitive displacement
sensor to use for absolute calibration of his violin mode
sensor.
We had a very productive visit from Gregg Harry a couple of
weeks ago. He and Steve measured Q's for one superpolished
silica disk to be sent to Lyon for the coating experiment. Steve
has measured Q's of a second disk.
Josh Smith has measured Q's in a sodium-silicate bonded
sample supplied by the Glasgow group. The Q values were comparable
to those in the KOH-bonded samples, but were definitely lower.

He is consulting with Jim Hough about possible causes. One
feature of the sample is that there are some visibly poor regions
in the bond.
Two graduate students have started 1-semester Graduate
Laboratory projects with our group. Both will be looking at
possible Q improvements from annealing. Belkis Cabrera will
anneal some fine glass fibers to look at the question of
surface loss in pristine samples, and Jose Luis Lopez will study
superpolished disks, to see if annealing can repair the surface damage from polishing.

With Alastair Heptsonstall in Caltech and Peter Sneddon writing his thesis progress in Glasgow has been somewhat slower of late. We have been working on understanding the intrinsic losses in the sodium silicate bonds. Assuming that the modulus of the bond material is essentially the same as that of fused silica and that the bond thickness is around 80 to 100 nm as measured by an atomic force microscope we obtain an unphysically high loss for the bond material (phi~4). This is consistent over a number of experiments and work continues to determine directly the elastic modulus of the bond material and to determine the bond thickness in a different way.

Looking at Bench we find that the formulae for ribbons and fibres does not seem to have been implemented in the form in which it was submitted by us and we are chasing down the errors resulting from this.

We have also been working recently on putting together the Conceptual Design Document for Advanced LIGO suspensions, and this was presented at the suspensions design review on 20th September 2001

SWG related activities this month mostly finalizing the content of the joint Syracuse/Glasgow/Stanford papers on dielectric coating loss. These have now been submitted to the LSC for review.

4 samples for the systematic coating elastic loss study are about to be sent to Glasgow where Sheila is headed for the next month, so these substrates can be measured pre-coating.

----------------
Silicate Bonding at CIT (Armandula reporting)
Baked sodium silicate silica/sapphire bonds and observed a significant amount of stress.
Run several tests:
a) Baked at 120 degrees C for 24 hrs. freshly made bonds and 60 day old bonds. All showed stress.
b) Baked at 120 degrees C for 24 hrs and at 60 degrees for 24 hrs. All showed stress.
c) Monitored silica/sapphire bonds while heating; observed that stress starts showing at about 30 degrees C.
Bonds baked for 1 hr at 41 degrees C showed irreversible stress.
Tested in tension baked and un-baked silica/sapphire bonds. The baked bonds are stronger even when stressed.
Testing continues in this area.
"Q" measurements
After long delays with substrate manufacturing, the development of coatings aimed at reducing mechanical loss will start soon.
"Q" measurements on the first set of two fused silica 3" dia. x 0.100" thick substrates are completed.
Gregg Harry and Steve Penn performed the measurements.
The parts will be shipped to Lyon shortly.
These parts are going to be cleaned and annealed one at 500; the other at 1000 degrees C.
"Q" measurements will be performed again on them to evaluate mechanical loss induced by the processes, if any.
4 fused silica substrates 3" dia. x 1" thick will be measured at Glasgow by Jim Hough and Sheila Rowan.
Expect to have at least 2 of these parts measured and ready to ship to Lyon in October.
It needs to be decided if it is necessary to coat one of each type substrate per run, (thick and thin), or one of wherever type is available at the time a coating run is scheduled.

Working on the AdLIGO costing effort, specifically, action items from our two costing reviews.
Supported the DRR and CDD last week. We are also working on a lab layout and lab requirements list.

------------------
Mark Barton and Virginio Sannibale have both generated nice numerical models of
quadruple pendulum suspensions. Virginio's is meant to generate mechanical
models for the e2e system, while Mark's is intended for suspension design and
will shortly generate state space matrices for input to Ken Strain's Simulink
model. Both use a variational principle approach very much like Matt Husman's
code, and run very quickly.
Calum Torrie has arrived at Caltech and is settling in. Welcome to West Death
Valley, Calum!
We have been given space in the Synchrotron High Bay for assembling and testing
prototypes.
Fiber/Ribbon Research:
----------------------
Continued study of our fused silica double pendulum shows the dilution factors
as measured by the frequency shift of the violin modes with temperature are
quite close to their predicted values- within 15% or so. The Q's of the first
five violin modes are all in excess of 400,000,000. The fourth mode has a Q of
500,000,000. These high Q's imply that the surface losses of the fiber are
somewhat less than the fitted value from Syracuse measurements, though within
the range they quote. We intend to remeasure the Q's at elevated temperatures
and then with the lower mass removed so as to determine the intrinsic loss.
With Alastair Heptonstall's help we are assembling the vertical bounce Q
experiment and expect the suspension to be in vacuum by week's end.
Silicate bonding Research:
--------------------------
We are bonding two different styles of the GEO ear to a fused silica mass for a
static load test. Once they cure we will weld fibers to them, load the fibers,
and see what happens.
We are continuing our study of how sapphire-silicate bonds cure at high
temperature. Early results: some bonds break upon temperature cycling, but
those that don't are much stronger than unbaked bonds under tension. Quite
Neitzschean.
Finite element modelling of several ear attachment geometries is underway. We
can clearly plot the stress distribution in the bond interface. We see a clear
advantage to moving the fiber attachment point closer to the mirror in that the
maximum tension within the bond is greatly reduced.
Coating loss measurement:
-------------------------
The result of an undergraduate SURF project is that the reflectivity at 632nm of
a coating does vary with strain, more or less reproducibly. We are working to
reproduce this measurement, clean it up, and retest it at 940nm. This
experiment suggests a fairly simple way to measure the Young's modulus of thin
film coatings directly, by stressing single-layer coatings and observing how
their spectral reflectivity changes. This data would be useful input to the
Syracuse coating thermal noise model.

Recent activities of the Advanced LIGO seismic isolation team: (Giaime)
1) Typically LLO has more than a factor of 10 greater noise in the 1 - 3 Hz band than LHO, making daytime operation difficult. In response, our group has bumped up the priority of our planned test of the external hydraulic pre-isolation stage, and have decided to test the BSC configuration first. We hope to carry out a full-scale test early in '02, in LASTI's central BSC, and intend to order enough parts for two complete tanks, so that the second can be used at LLO for "first article" testing if desired. We are also implementing LIGO-I microseismic FF at LLO, and will simultaneously study whether coherent prediction signals are available for higher-frequency noise. Brian is building a dynamic model that will be used to study the problem.
(Brian, Corwin, Ken, Joshua, Dan, Joe, Larry, Rich, everyone else?)

2) The mechanical design for the ETF active isolation platform prototype is complete and HPD in Boulder has begun fabrication. Delivery to Stanford is expected in December.
(entire team, especially Giles, Marcel overlooking HPD and designing key auxiliary parts.)

3) Work on various instrumentation details for ETF prototype continues. These include the high-current voice coil drivers (Rich), possible modifications of the force actuator frame (Gerry), DSP setup, plumbing & wiring (Stanford), etc.

4) Diversion of team effort to 1) above can't help but cause delay in other work. At the moment, the most "neglected" work has been UHV technique research for the LASTI version of the HAM, and the start of work on a BSC version. Also, Hua has been laboring alone on finishing work on the two-stage prototype formerly at MIT.

There are two efforts underway in the high bay. The high-priority effort to demonstrate a pre-isolator solution which can be retrofitted to initial LIGO is dominating. We are working closely with the SEI group to turn the concept and prototypes to date into a system which can be placed under an initial LIGO seismic isolation stack, installed in the LASTI BSC. At MIT, we are preparing the piers (completing the grouting), the materials handling system, and assembling the parts at MIT for the initial LIGO stack.

The seismic support piers have been grouted and the
contractor has cleaned up and cleared off. The grout is expected to
attain sufficient strength by 10/10 to undertake the planned test
of free pier eigenfrequencies, requested by the hydraulic actuator
design team.

Due to the high bids received on the mezzanine structure for SEI/SUS
staging, we've decided to de-scope the weight capacity and do the design
in-house. The resulting delay will be accommodated by temporarily relocating
other equipment so we can use occupied floor space for pending
assembly work; the mezzanine will still easily be in place well before
it is needed for adLIGO seismic installation (there's sufficient float
in the schedule).

Hydraulic actuators:
Ken held a review with the SEI group of the mechanical design for
the hydraulic actuator supports and articulations. The design was
received well. Two issues raised were the capacity for coarse
adjustment (relatively easy to handle) and a subtle problem of torsion
due to spring unwinding (as of this morning Ken had a novel idea
which should fix this).
Joshua has begun assembling the hydraulic pump/accumulator/regulator
test skid in the machine shop. After much deliberation we elected to
go with stainless Swagelock piping for the prototype for ease of
assembly. We may later transition to VCR or brazed connections for
production.

PSL:
Almost all of the electronics and almost all of the optics for the Pre-Stabilized Laser has arrived at LASTI and is being installed and debugged. The system is effectively identical in hardware and software to the systems at the LIGO Observatories, so familiar to all those who have spent time at the observatories.
The PMC (pre mode cleaner) servo electronics were repaired and the PMC is working.
The 10W MOPA arrived in good order from Caltech. We expect our enhanced
safety enclosure will be in house shortly, at which time we'll
place the PSL table into final position and cut over
from the test NPRO to the MOPA.

A significant effort was made by Betsy Weaver in a visit to MIT, along with LASTI staff, to prepare the initial LIGO Small Optics Suspensions (SOS) for installation in the LASTI vacuum system. The resulting cavity will be used to test the PSL and possibly lead to some exploration of marginally stable cavities. They will be installed in Mid-October.

@ Hongo
Akiteru
I have been having trouble with sensor diagonalization after the JPS
meeting and finally got rid of it. The problem was as follows:
One of two translational modes of the IP has resonance at 65 mHz and it
was almost normal to one of the LVDTs. Unfortunately there is a yaw
resonance of the suspended part at 70 mHz. As a result, the peak of the
translational mode was reduced due to coupling between these two
resonances. It bothers the diagonalization, because we need to know the
exact height of the peak for it. Then, I took a measurement with two
actuators to try to excite just the translational mode, without
injecting force in the yaw direction. This was successful and now we
have a clear peak at 65 mHz and the sensors look diagonalized pretty
well. However, I had been trying to re-orient the translational mode so
that all the sensors can see it well, by changing springs and mass
location... It took so much time, but wasn't very effective. Now I am
reconstructing loops for the inertial damping. (The actuators are
already diagonalized.) I guess we can regain the damping in a few days
and will go ahead to the feed forward test and launch the michelson
measurement in the third week of this month.

Latest news: I have recovered the inertial damping on the y mode (65
mHz) and am workingon the x mode (30 mHz). Everything seems reasonable
eventually.

Szabi, Yuhiko
Szabi has sent us a code to run his stepper motor drivers and Yuhiko has
checked them all working nicely. So, basically, we can close the
chambers.
(Of course I would like to do it after confirmed everything for the
michelson experiment.)

Yukiyoshi, Tatsuo
Yukiyoshi and Tatsuo are working on the suspension. They have measured
transfer functions for actuation on the mirror and the suspension
platform.
The mirror transfer functions looks quite nice. We can see clearly the
advantage of the Virgo recoil mass system. The mirror can be controlled
as a single pendulum at all. I should have described their setup. The
filter one is fixed on the safety shelf of the SAS for the front tank,
and the rest of the suspension is suspended from it. Anyway, the mirror
control looks very easy, however, the behavior of the platform is not
very good. There are so big coupling between different DoFs. I think we
should use the stage just to align the mirror in pitch direction now and
modify it later to be able to control mirror position.

Yukiyoshi
Yukiyoshi is also completing the MSE simulation of the suspension. There
are several parameters which need to be determined experimentally, but
the job is almost done.

@ Caltech
Chenyang
Started his PH 078 with the AC electrostatic drive.
Will be under mixed Alessandro and Riccardo supervision.
Rebuilding a resonant frequency with the inductors characterized by him
last year and the new electrode made by Ale and a Glass mirror in front.

The mirror movements change the resonant frequency (obviously). Will try
to implement a phase lock circuit to automatically lock on the peak of
the resonance or on the side of it.

Alessandro
More on cutting sapphire with water jet and ultrasounds.
Should have water cut of samples early next week, to be then cut with
ultrasounds into flex joints for further tests at Caltech and testing
polishing techniques.

Charlotte, Virginio, Riccardo
Frantic work on several problems on the creep setup, we tried to clear
up our mind and resumed the current situation, for our own sake, in the
report LIGO-T-10112-00-R, The GAS blade creep measurements, problems and
some solutions.
It reads like a thriller, full of surprises and unexpected traps.

Charlotte, Riccardo
Taken delivery of the cryogenic thermal conductance measurement
cryostat, works nice with metal samples down to 2K, will wait for
Alessandro’s early samples for testing sapphire.
Studying how to modify it to measure also the Q-factors of flex joints
as a function of temperature.
Working on design of high performance surface hardened and/or glassy
metal flex joints.

CR-010006	WBS 1.1.4 (OPS)	Roof Shelter for Concrete Pad Adjacent to Erosion Control Pond (Livingston)	M. Coles
CR-010007	WBS 1.2.4	Core Optics Components, Re-polish/Recoat Beam splitter	S. Whitcomb
CR-010008	WBS 1.1.4	Five-Ton Crane for the Staging Building in Hanford	O. Matherny
CR-010009	WBS 1.4.2 (OPS)	Additional 10-W Laser for PSL Lab	S. Whitcomb
CR-010010	Advanced R&D	Seismic Isolation System PD ETF Mechanical Contract (APPROVED)	D. Coyne/ G. Stapfer

Weekly Report for Week Ending October 11, 2001

Special Announcements:

LSC Issues (Weiss)

LIGO I Construction/LIGO Laboratory Administration (Lindquist)