LIGO Weekly Report

There was a site teleconference held on Thursday, May 16, 2002. The following items were discussed:

Action Item #104: Determine Weber Bar Display Locations. Closed. At Hanford the bar itself will be displayed inside the new building while the enclosure will be outside.

Action Item #106: Prepare punch list for building at Livingston. We are currently several punch lists beyond that referenced by this action assignment. For status, see below. This action is closed.

Action Item #109: Provide estimates for furnishing new buildings for contingency list. This action is closed. Estimates provided by Otto and Gerry. Ed Chargois is looking for excess property furniture.

Action Item #110: Provide pointer to Operating and Control Room Procedures on the Web Site. These procedures are actually a part of the Visitor's Site. The site has been checked and includes the appropriate material. This action is closed.

Action Item #111: Review Visitor's Program Proposal for constraints regarding High School Teachers. Response provided. There appear to be no constraints. Still governed by PMA rules. This action is closed.

Construction Livingston: Work on punch list items continues. Gerry has forwarded eight page copy of new punch list to Ed Jasnow and the contractor is working some of the issues. Biggest item is floor, which has not been started. HVAC contractor being pressured (pun intended) to address problems.

Construction Hanford: There will be a claim (arguably without foundation) regarding stainless steel radius wall siding. The siding was fabricated approximately 20 inches too short. In any case will result in six week delay. The audio/visual equipment is due to be delivered next week.

The list of current actions revised to reflect the status of open actions assigned through May 16, 2001 may be found at ACTION LIST.

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. . .

Advanced LIGO (Frey)

Over the next few months we must prepare and submit the following proposals, work plans, and reports:

We will initiate the preparation of written material for this (FastLane) submittal. I conceive of something relatively short, like the current quarterly report, but emphasizing Operations. Sections will be required for Hanford, Livingston, Caltech (Science Runs, milestones, Data Analysis, 40-Meter Upgrade), MIT, Seismic Prototype, LSC, and Cost vs. FY 2002 Budget Status. We will also require a proposal for FY 2003 from MIT.

Quality/Safety (Tyler)

LIGO Hanford Observatory (LHO) Operations (Raab)

Better locking behavior was obtained on both interferometers in the last week. Rolf, Alex I. and Ashfaq K. are working with David and Richard to install new hardware and software on 4K. Matt, Rana and Stan are up working on commissioning.

Mike Landry was looking at free-swinging 2k IFO fringes and Stan's sharp eye picked out odd-looking sideband fringes, indicating that 2-omega sidebands might be too close to antiresonance in the arms, which could be expected to cause trouble. The RF frequency had been re-tuned a few weeks ago to match better the mode-cleaner passband and it was feared this may have precipitated our locking problems on the 2k IFO. RF was tuned part-way off (300 Hz out of 29.5 MHz) causing a dramatic change in the sideband fringes as they returned to a more normal appearance. Unfortunately, this did not help the locking behavior.

Mark Guenther and Landry searched the e-log and con-log to document all changes between when we last had good locking behavior and the more recent weeks of poor behavior. They identified not a sharp step, but an inexorable slide into bad behavior. In the midst of this slide we had increasing problems with "ratty-looking" SPOB signals, reflecting the sideband power resonating in the PRM. The large and sudden SPOB fluctuations cause large fluctuations in the input matrix optimization used by the lock-acquisition code during "acquire" mode and may cause bad coefficients to be "locked-in" to "detect" mode. This could cause problems in acquisition, hold time in "acquire" mode and variable results from lock to lock in "detect" mode unless a known good set of matrix coefficients is reloaded consistently. Stan and Fred investigated the SPOB signals and decided to test whether the rattiness might be a saturation phenomenon. Not wishing to beat about the bush and taking advantage of a good signal to noise ratio, the light intensity on the SPOB photodiode was reduced dramatically by 10x. This appears to have cured the rattiness problem. At some later time we may want to optimize the attenuation, but this should do well for the near future.

Another impediment to stable locking has been the occasional large and frequent bursting in in the MICH loop. Occasionally the bursting has been severe enough to throw the IFO out of lock and, even when smaller, the bursts contaminate data runs with frequent glitches. Episodes of good and bad burst behavior were often demarcated by incredibly slight (mV) touch-ups of alignment, but lately the problem has become more dramatic and stubborn to reduce. Rana analyzed the 2k MICH loop and found that its phase margin was quite poor near the unity gain point. Theorizing that fluctuating gain due to angular fluctuations could destabilize the loop and cause bursting, he reconfigured the loop gain profile to improve phase near the unity gain frequency. This dramatically improved the bursting behavior and MICH_CTRL looks about the best it has ever looked. There is still significant bursting in the PRC_CTRL in detect mode, but this should be improved by running with common-mode servo engaged.

Matt Evans and Rana have been attempting to reconfigure gains so that we can better suppress the total RMS fluctuations with the servoes and buy enough head room to increase the light levels on the antisymmetric port (AS) photodiode. (Our best sensitivity above 500 Hz with good stability has been dark-noise limited.) So far, they have not succeeded to get a good transfer to stable operation in common-mode configuration at higher dark-port power. Single-arm 2k frequency noise measurements made at 8x higher power on the AS photodiode confirm that frequency noise on the light is limited by radiation pressure fluctuations that shake the mode cleaner mirrors from 40-200 Hz. The frequency noise at 100 Hz is approximately 10x spec due to 200 ppb/rHz fluctuations in laser intensity. Nonetheless, this leaves room for significant improvements or identification of other noise sources in this frequency regime. Work will continue on this effort.

Summary of Commissioning Activities at LIGO Hanford Observatory May 16, 2002 Report

Work focused on recovery from software upgrades on 4k, and work on 4k PSL and input optics. Repair of a "glitchy" coil driver seems to have cured stability problems on the 2K. Construction work has shifted to extreme noise generation this week as asphalting is being done. Expect this will terminate on Monday, May 20.

As we recover functionality of the core optics, work has continued to improve the frequency noise on light delivered by the input optics. Bill Kells has been working to level the mode cleaner, so that some of the electronic notch filtering of MC bounce modes can be removed. This would simplify the configuration to allow more flexibility in moving down crossovers in the frequency control system. Another major contributor to RMS frequency noise has been line harmonics on the PSL output light. Rick has been leading an effort on grounding of equipment to mitigate this and it appears that this noise contribution will be drastically reduced. Rana tested the "Super Optical Levers" on the 4k end mirrors, which worked extremely well. We will be implementing these as we get the 4k IFO back on its feet. Rana also installed a digital comb filter that can deal with residual line harmonic noise.

Hugh is continuing to make progress understanding the common mode tidal actuator.

LIGO Livingston Observatory (LLO) Operations (Coles)

(Jonathan) Much of the past week I've spent working with Larry Jones and Hugh Radkins devising a procedure that will facilitate the installation of fine actuators on our ITMs. Had several telecons with the Stanford group discussing fabrication issues of the new welded actuator. Began a much closer collaboration with Marcel Hammond and Ken Mason in order to tie loose ends up and move toward completing the drawing details and getting work into the shop. Met with a few local fabricators to discuss the practicability of various assembly schemes.

(Marcel) Efforts to fabricate a "first article" pre-isolator springs are complete. Fabrication began 5/15, est. complete-5/22

Work continues for post-fab i.e. procuring vendors for heat treat, surface coating, joining etc

Work on stiffness calculations continues. i.e. forming the stiffness matrix. The matrix is formed. I am in a model refining mode

I am integrating the pre-isolator springs into the Ken mason's pre-isolator model

(Jonathan) So far, so good. We appear to have uncovered all of the obvious bugs, and the only alarms we've received all week were caused by people inadvertently opening an exit door before swiping their cards. This is an easy mistake to happen, but people are developing the habit and it's occurring much less frequently.

(Shannon) Finished building a workstation for a software package that will help with the computer management here. installed VNC on a couple more computers. Updated the access rules on the firewall. Looking into some classes on the PIX offered by Cisco. Loaded some more CDs into the "online" software storage. Installed P-card software on a couple of new machines.

(Chethan) Working on clearing out some disk space on LLO1 to have sufficient disk space for S1. Wrote a script to clear all core files. Investigating file transfer speed between CDS and the PC machine in control room. Preparing to setup Sun Blade 1000's for DMT. Setup ip addresses for these machines in host table on LLO1.

SCIENCE & ENGINEERING SUPPORT (Coyne)

Seismic Upgrade Project

Dennis Coyne reporting

External Pre-Isolation

The group decided on an approach (designed by Ken Mason) for the coarse alignment capability, which will be added to the ITM chambers. It consists of a set of pneumatically lifted roller plates at the base of the integrated sensor/actuator assembly and X-Y driven translation stages from the initial SEI coarse actuation system. Ken Mason will detail out the changes to accommodate the coarse actuation design and get them to the prototype venders. All of the machining is underway for the spring/actuator assembly.
Ken is building configuration drawings for the different options (hydraulic actuator, electro-magnetic actuator, HAM, BSC, etc.)
Ken and Myron have devised a method of installation at LASTI using a 3 ton, low friction, ball screw actuator which is electric motor driven. This ball screw actuator will be used to lift the crossbeam above a pier slightly to allow the current hardware to be removed and the new assembly to be installed.
Larry Jones has drafted a procedure for the installation of the fine actuation system (FAS) on the LLO ITM chambers for the Piezo-electric External Pre-Isolator (PEP) system. The load on the pier will be relieved by use of the overhead crane and a load cell. Then the scissors table will be lowered to permit removal of the current spacer block.
Marcel has performed finite element analysis on the machined spring design. Stresses appear to be within acceptable limits (though further analysis with reduction due to chamfering is needed). Finite element analysis indicates that a lightened version has a first frequency at 100 Hz, when the boundary condition is assumed to be clamped. More analysis with realistic rotational compliance from the adjustment mechanism is pending. Some further lightening of the spring is possible and may be necessary to have a minimum frequency of ~100 Hz. (For reference the upper unity gain frequency of the system might be ~25Hz in order to enable some suppression of the 12 Hz vertical bounce mode – if possible. Hence the need for a spring resonance of >~ 100 Hz.). The spring stiffness matrix will be defined, from further FEA, by next week for use in the servo analysis.
A first prototype of the machined spring will be made. Heat treatment for the maraging steel has been defined. A test plan is pending.

Electronics status (Rich Abbott, Mohana Mageswaran):

About 75% done on the schematics for the pump station servo. Pump station enclosure into silk screening. Rack mount unit designed; schematics in review. Pump servo controller consists of 2 microprocessors with a 4 line LCD display on the front on the unit. There are currently no communications from this (stand-alone) controller to the outside world. (This may change once requirements for such communication are defined.) However, there is a single composite alarm bit (which is programmable) and can be run anywhere by using a set of closures.

Pump station pressure stabilization servo topology: drives pump speed and by-pass valve; sense low frequency pressure fluctuations at the load; sense high frequency pressure fluctuations at the valve; cross over in a blended actuator; analog servo for the high frequency, digital servo for the low frequency; cross-over re-programmable with coefficients entered into the controller.

Design for the L4C geophone and Kaman inductive position sensor field interface electronics boxes have been completed. Boards due back this week. Electronic components are in-house. Enclosures due soon.

Decided to interface to the L4C geophones (which are integral to the actuator) with a custom designed, low profile, circuit board which converts from the Bendix connector on the L4C to a standard board mounted connector (e.g. DB9) The circuit board will have pins to mate to the L4C connector. The Stanford group will design the circuit board and its strain relief.

BSC Stack Transfer function measurements at LASTI (Rich Mittleman): Measured X-to-X, Y-to-Y, Z-to-Z. Also looked at angular transfer functions as well as off-diagonal terms (e.g. X-to-Y). Signal to noise ratio is good even on the off-diagonal terms. Work continues. Comparison to analysis pending.

Hydraulic External Pre-Isolation (HEPI)

Tests by Jamie at Stanford on the passive RC integral manifold block have indicated a pressure fluctuation attenuation of 5, whereas analysis predicts an attenuation of ~10. Results are being studied. Will also revise pump station pressure fluctuation limits to be clear about requirements at the load and at the pump station.
Pump Station (Ken Mailand)

CONSTRUCTION: Basic station Finished.
MOTOR SPEED CONTROL: Wired and tested 'ready to run'.
PRESSURE LEAK TEST 'air 80/100 psi' Solder joints by CIT plumbing shop. Solder leaks remain in 3 places, 3 fixed, one reworked and still leaking will try to seal all solder joint leaks by 5-16-02.
DESIGN: A future modification to the reservoir filler cap for a float level sensor.
LAYOUT DRAWING: I have finished a scale plan view and partial elevation [acad 2D] layout drawing, of station component assembly.
HARDWARE ASSEMBLY: Resistors are finished and installed.
LONG TUBE RUN: Finished.
TEST LOAD: Finished.
The nozzle design for the hydraulic servo valve has been completed and a drawing exists (though not yet into LIGO template & DCC). Parker no-bid the nozzle but gave Jonathan the names of several companies that produce nozzles for Parker’s products. Parker is willing to assemble and possibly calibrate and adjust our servo-valves for us. Brian Lantz and Jamie will define the adjustment and calibration procedure. Jonathan will pursue quotes for the nozzles and assembled servo-valves.

The distribution manifolds at the load (chamber end) are yet to be designed.

Hydraulic actuator (Jonathan, Corwin, Brian): Working on issues of accessibility and repairability for the proposed welded design (vs bolted). Issues are related to the fluid manifold which is integral to the actuator structure and the integral bellows, bellows shield and limit stop. Resolution expected this week.

Modeling (Brian): “real world” model with 6 dof and compliance of the real structural plant is in process. Ken Mason will do an FEA on the stiffness of the V-block weldment (spring/actuator assembly structure) to determine the actuator support compliance.

Electro-Magnetic External Pre-Isolator (MEPI)

Greg Harry measured the temperature rise of the coil to be 11C for 3A drive current. With suitable heat sinking and free convection, this is reduced to 1C. This seems adequate.

Started to measure the EM coupling from the coil to the L4C geophone.

CDS Software

Rolf Bork reporting

Alex, Khan and I completed installation of new Digital Suspension software at LHO. Still remaining is addition of new timing modules for LVEA large optics (Khan).
Alex is preparing the digital suspension software for the 40m lab. The software and VME system for the small optics will be installed Thursday or Friday. We will still need to modify the AWG and TP manager to use the 5579 (PCI) reflected memory network instead of the VME based reflected memory used at the sites for DAQ.
I'm at LLO remainder of this week and thru the weekend. The latest LSC code and new timing modules are being installed/tested.
Lori has modified the ICS110B ADC module initialization software to reduce the sigma delta pipeline delay from ~38 samples to 10 samples. This sofware will be incorporated into systems using the 110B after some further testing, specifically noise characterization.

CDS Hardware

Jay Heefner

Differential Driver/Receiver:

Two of each of the 4 board types have been stuffed and are being tested. A design ommission that set the gain of the DAC to AI path to 1/2 has been corrected. The correction will be incorporated to the design prior to shipment to the sites. The A-A to ADC path boards will be tested next.

Optical Lever Whitening:

A schematic for the upgraded board has been circulated for comments. It should be complete and ready for layout next week.
A new differential driver/receiver chip has been found and tested. It is the THS4150. Preliminary tests show that it may work very well in applications where we need differential in and/or out. The noise is ~20nV/rtHz at 100Hz and the common mode rejection should be better than 60dB. The measured CMRR is ~54 db but the number is not final. If this part works out, it will be incorporated into the design for the board. This will make all off board connection completely differential.

Rich Abbott, Falvio Nocera, Ben Abbott

Testing is underway of the new ISS at the 40m lab. Preliminary results indicate that the auto-zero function is behaving well and a first look at the residual noise outside the loop shows that the servo is performing according to the design. The variable gain amplifiers seem to function well. Much more to follow before Friday when we ship the first unit to LLO for installation next week.

Sander Liu

LEMO Anti-alias Filter chassis - Working on three new units for the 40M

Micro-Seismic Signal Processor-

Remote Interface Box: Testing of the first article is complete. It works well. We will proceed to design the metal chassis so that the rest of the units can be built.

Post Processor: In the process of testing the first article.

PSL

PeterKing

The optically contacted pre-modecleaner fabricated in the PSL Lab is completed. Only two small defects in the optical contact were noted and recorded (see image).They are sufficiently small enough that they should not affect the performance.

NPRO #393, returned from Lightwave after a warranty repair, was tested. The output power remained stable during an overnight test.In addition the relative intensity noise seemed to have improved.

Optics Analysis

Erika D’Ambrosio

I took a lot of data regarding the thermal lensing effect, taking into account only the phenomenological change in the refractive index. The simulations were aimed to see the level of power increase when from the "cold" configuration the interferometer attain the "hot" values that are the ones the radii of curvature have been designed to match. I used both the FFT-code and Melody and looked at both the sidebands and the carrier. The sidebands start from a power gain that is ~25% of the "maximum hot value" according to the FFT-code and ~50% according to Melody. There is no discrepancy between the carrier results obtained by the two codes and both show the carrier is almost not affected as theoretically predicted.

I also plotted the adjustments of the lengths performed by the two programs and I observed similarities and differences. Both show that the length of the arms does not need to be readjusted and shorten the length of the recycling cavity. However the decrease of the recycling cavity length is much larger for Melody. For the FFT-code that "retuning" is only "quadratic" in the distortion while it is "linear" for Melody.

Optical Metrology

GariLynn Billingsley

A discrepancy between CSIRO and CIT Radius of curvature results has been resolved.There was a nearly constant offset of ~ 6nm when measuring the Sag of the PRMs for the 40M.This turned out to be the result of a re-calibration of the CSIRO WYKO frame grabber card.Measurements are back to the traditional ~< 2nm level of agreement.

Optical Contamination Cavities

Lee Cardenas, Liyuan Zhang

OTF Lab.

Contamination Cavity #2 Continue taking measurements for ring down and beat frequency.

Contamination cavity # 1It breaks lock easily due to the electronics (oscillator and RF supplier malfunction)

We looking into fixing the electronics.

40 Meter Interferometer (Weinstein)

PSL: Rich Abbott, Flavio Nocera, and Ben Abbott have been testing the PSL intensity stabilization servo (ISS), and integrating the DCPD with the servo. Testing was hampered by the discovery of a 7kHz oscillation on the light intensity, which has been isolated to the frequency stability servo. Therefore, the FSS has been disabled in order to test the ISS. Preliminary results look good... a 20 dB reduction in intensity noise at ~ 1kHz. We will investigate the FSS oscillation next week.

Thermal Noise Interferometer (Libbrecht)

We have spent our time since the last report measuring various noise sources, including electronic noise and laser frequency noise. In the North Cavity, we find that electronic noise at its present levels will prevent us from reaching the same sensitivity level achieved in the South Cavity in December. Because of its higher optical gain, we do not expect this limitation to apply to the South Cavity.

One thing we discovered while hunting down the source of our excess electronic noise was that our gain-boost stage was no longer operating according to spec. If no boost stages were active, or if only one stage was switched in, the unit worked as designed. When both boost stages were engaged, however, the output developed a large dc offset, nearly railing. We adjusted an internal trimpot to remove this dc offset, and the unit is now performing well with any number of boost stages active.

LASTI (Zucker)

We resurrected the electron multiplier on our Balzers RGA and can now in principle detect much lower partial pressures of hydrocarbons or other high-mass contaminants (as little as one part in 10⁷ of the base pressure, which is predominantly water vapor at 1-2 x 10^-6 torr or so). At this increased sensitivity, we're finally able to see some putative HC peaks. However, they behave as if they could be generated on the RGA ionizer itself, which is well known to generate CO and CO₂ on its own. We thus may only have an upper limit (in the 10^-13 torr range) to the HC levels inside the machine. Lesson of the week: if you look hard enough you will ALWAYS find something.

We did some minor modifications to the turbopump controller settings to force tripping and backfill on forepump faults (e.g., empty or soft seal gas dewar).

Speaking of which, the membrane air dryer components have arrived from Ballston; Myron is installing them now. Next week we will commission the dryer and wean ourselves off N₂ seal gas dewars for good.

Ken Mason reports-
Completed the coarse actuation system design for both the hydraulic and electromechanical isolators. The system will allow +/- .2 inch coarse adjustment in two horizontal axes and in the vertical axis using many of the existing parts. The drawings for parts already being built have been updated and sent to the suppliers. Changes to the housing caused the delivery to be changed from May 17th to June 3rd.

A weldment has been added which will allow the isolator to be preassembled in a bench area. The parts are held in position with locking screws, which are backed off when the assembly is on the pier in its final position.

Myron has begun design and specs on a precision HAM installation fixture based on commercial precision ballscrew jacks. A manufacturing rep will visit Monday to detail the requirements.

Data Analysis and Computing (Lazzarini)

Software Systems (Blackburn)
The migrations from TCL eventloop driven data sockets to TCL channel based data sockets finally payed off this week. We've been working with the new TCL channel interface for less than a day so results are preliminary, but indications are that we are looking at roughly a 40% improvement in the system performance and about a factor of 3 increase in reliability. Many thanks to everyone in the LDAS software development group for the effort to get this ready for the next release.

The frameAPI resampling code is now directly returning the appropriate data structure for storing resampled channel data into processed frames.

The controlMonitorAPI's client now has to ability to point and click on a graphical representation of jobs issued to the system and pull up the associated log files.

The data products constructed by the eventMonitorAPI are now collected into a single container before transmission to the designated API. This decreased the time spent in the eventMonitorAPI during a dataPipeline command by a factor of as much as 5 in typical upper limit search codes.

The lightWeightAPI was modified to correctly work with the new Xerces-1.7 XML C++ I/O library. We will be migrating to this version in the next
release.

A set of cgi based webpages are being designed in PERL for a secure seb server and will provide an automated means to request an LDAS user name and password. The next release of LDAS will require users to apply for passwords through this webserver.

The GCC 3.1 compiler is now available. We have begun the first phase of testing with this compiler and hope to be building and testing LDAS using it in the next week.

The ILWD socket test has been updated to use the new TCL channel calls to the data sockets. Test results from using this new script indicate that data transmission is about 20% slower do to the overhead of the TCL channel. A small price to pay in exchange for the overall 40% speed up in LDAS jobs and increased reliability.

The narrow band features correspond to the periodic frequency corrections imposed on the 60Hz mains in order to maintain frequency stability in the presence of varying loads.

It is surprising that the East-West cross-correlations of the frequency errors exhibit coherence over periods of order 1 yr. It is not clear whether this is going to be a problem for, e.g., long-term integrations of cross-correlations between sites. The degree to which these grid correlations couples into the detector(s) needs to be understood first.

The U.Fl. group has been looking at mains peaks in power monitors and AS_Q signals as part fo the Stochastic Working Group activities. Long term correlations we observed and were also seen to decay; however the time scale provided by these grid authority datasets is much longer than anything we would have seen to ate during engineering runs.

LIGO II/Advanced R&D (Sanders)

The Burst UL working group met at MIT from May 5-8 to push E7 data through the burst analysis pipeline. Attending were: Peter Saulson, Sam Finn, David Shoemaker, Alan Weinstein, Laura Cadonati, Stefan Ballmer, Ed Daw, Igor Yakushin, Julien Sylvestre, John Zweizig, Daniel Sigg, Rauha Rahkola, Szabi Marka, Erik Katsavounidis (host).

It was a very productive meeting. Over 1000 LDAS jobs were submitted to ldas-mit for generating burst triggers using three different burst filters, and most of them succeeded. Veto triggers were generated by DMT monitors absGlitch and Gide. The EventTool/ROOT was used to study and optimize the vetos. Then EventTool was used to cluster burst triggers; apply vetos, plot veto efficiency, livetime, and trigger rates; form coincidences and delayed coincidences between H2 and L1. We studied the efficiency for a variety of burst signals injected into the data (as far away as 10 pc!). We got almost all the way through the pipeline, on triple-coincidence playground data (3.2 hours). Significant progress was also made on the external (grb) trigger analysis pipeline. There was much discussion of the final statistical analysis and the form that final results might take. Much progress was made, and much more work is planned for refining and finalizing the procedure and then running on the full E7 data.

Lee Cardenas and I traveled to Stanford to check the performance of the 10-W laser loaned to the laser group up there. Shally Saraf had reported that the laser output power was only 4-4.5 W. On arrival a quick check of the state of the laser revealed:

It was decided to increase the NPRO diode current to bring back the NPRO power close to the same level at which the laser was shipped, so that now the NPRO current was 2.21 A. This brought the NPRO power to 415 mW, c.f. 405 mW when shipped. The optical train to the power amplifier was re-aligned and adjustments were made to the thin-film polarizers. Afterwards the state of the laser was:

Stanford University Laser Lab. (Shally)
126 MOPA S/N 104 with NPRO # 248
Completely re-aligned and power output obtained to be 10.5 watts.

Symptoms found:
water level very low in the cooler. Water is very dirty - we added distilled water. Shally will replace water later.
NPRO output power 262 mw after EOM (original power 380mw) with 1.89 Amps of current. Current did not change.
Both Thin Film Polarizer on the optical train was reflecting ~100 mw of power (major offset)
NPRO power going into the Amplifier only 85mw (original power 290mw)
MOPA output power ~ 6.00 watts after laser window with a 33.0 amps of current with a 25.5C diode temperature.
The potentiometer to adjust laser parameters on the power supply not functioning properly.

Remedies:
Increased NPRO current to 2.22 Amps to bring NPRO power output to be 415 mw after EOM in place.
Re-aligned optical train completely
Fixed and aligned both Thin Film Polarizer to minimum reflectance ~ 0.5 - 1.0 mw.
NPRO power going into the Amplifier is now 340 mw
MOPA output power is now 10.5 watts with 33.0 Amps. of current and 24.8C of diode temperature adjusted and running cooler.
Replaced a new potentiometer.

The vertical bounce experiment is installed in vacuum in the OTF and collecting data. This is an all-fused-silica, triple pendulum suspension, and the vertical bounce motion of the small intermediate mass allows one to cleanly measure the Q of a fiber at high strain (due to a much heavier bottom mass). Initial data are that the Q of this mode is 12 million at the strain intended for advanced LIGO. This compares very favorably with the Q's of bending modes measured previously for an identically prepared, but unstrained fiber, which was 11 million.

This suspension is rich with modes to study, but large seismic excitation of the lowest vertical bounce mode is complicating the measurements. The next step is to float the optical table to reduce the excitation.

Added the corrected thermal noise code to my triple pendulum model, bringing it up to date with the quad model.

This week I extended the validation of my suspension models to asymmetrical cases, to check for errors and to see whether imperfections made a big difference to the thermal noise predictions. In the toy and triple models, I made one of the fibres 1 mm shorter than the others. Initially I found a surprising result: there was a strong cross-coupling of roll to yaw which inflated the yaw thermal noise by a considerable amount at frequencies above the roll mode peak. It turns out that the result is strictly correct for the precise sense of "yaw" implied by the setup of the problem, but that this doesn't correspond to the obvious sense of rotation about a vertical axis for problems with a static pitch displacement, due to issues having to do with non-commutativity of finite angular displacements. It's easy enough to add a correction to make the results more interpretable but this has not yet been done.

SAS suspended Fabry-Perot.
Here is a result of a naive analysis on the 3m FP spectrum. This is a spectrum of the error signal of the 3m cavity, corrected with a single pendulum transfer function (the simplified version of the open loop transfer function). This approximation should be correct up to a few hundred of Hz, as the UGF is around 1 kHz. Of course I have measured the loop trf, and will calibrate the spectrum accurately). The SAS inertial damping was turned off on both the towers. I guess, the measurement was limited by the laser frequency noise above 2 Hz at time. I will take other measurements with laser frequency stabilization on, next week, and of course I will turn on the inertial damping as well. Though there should be various noises, and it will be hard to investigate their sources, the result seems quite good at low frequencies. As far as I experienced, all the peaks below 2 Hz must disappear after installing the damping, without injecting control noises. So we are very close to what we promised (10^-8 m r.m.s. above 100 mHz), if I didn't make stupid mistakes on the calibration.

Michael

Cryo setup still down, Quantum Design spent lots of time replacing parts, both present and replacement cryo pump look bad, they will be back next week with more thoroughly tested components.

Finally found the correct splatter setup and splatted several samples, the first one was polycrystalline (breaks into dust if you touch it), the other samples are glassy (not fragile) on increasing fraction of the sample surface. Tuning for optimal point The setup process ate a lot of samples, now we are short of supplies, need to make more samples. Tried also a MoRuBAlSiP, chickensoup sample, for fun for now.

Prepared fiber vacuum feedthrough in view of Q-factor measurement in cryo setup.
Preparing platform for room temperature Q-factor measurement in large vacuum chamber.

Calculated folded pendulum suspension thermal noise. There is no advantage in using a folded pendulum. The resonant frequency is pushed down and the effective K goes down, but the term 1/2K X² at fixed frequency stays constant. A posteriori, since the mechanical K stays the same, this is not surprising. Now concentrating on calculating thermal noise in thin flex joints.

05/16/02	Packages	Faxes
In	32	26
Out	12	26

master oscillator diode current	1.89 A
master oscillator output power	250-260 mW
power amplifier diode current	23.01 A
laser output power	~5 W
diode temperature set point	25.5 degrees C
operating hours	11,268 hours
master oscillator monitor	89
power amplifier monitor	60

master oscillator diode current	2.21 A
master oscillator output power	415 mW
power amplifier diode current	23.00 A
laser output power	10.6 W
diode temperature set point	24.9 degrees C
master oscillator monitor	214
power amplifier monitor	135

Weekly Report for Week Ending May 16, 2002

Special Announcements:

LSC Issues (Weiss)

LIGO I Construction/LIGO Laboratory Administration (Lindquist)