Session H20 - Gravitational Radiation - Experiment.
MIXED session, Sunday morning, April 30
102A, LBCC

[H20.001] End-to-End simulation program for interferometric gravitational wave detectors

A time-domain simulation program has been developed to provide an accurate description of interferometric gravitational wave detectors. The program simulates the time-evolution of fields, optics, mechanical structures and electronic and control systems. It is written in C++ and its modular design makes it possible to simulate wide variety of experimental configurations and processes using the same software without modifying the program. The flexibility of the simulation environment makes it easy to add new physics or functionalities. This is being utilized to build a model of LIGO (Laser Interferometer Gravitational-wave Observatory) with the aim of aiding in the shakedown and integration of the interferometer subsystems, and ultimately the optimization of detector sensitivity.

[H20.002] Advanced Seismic Attenuation System for the LIGO II Gravitational Wave Interferometric Detector

The advanced Seismic Attenuation System (SAS) is developed to isolate the LIGO II test masses from all external mechanical disturbances. The SAS performance depresses the external noise levels well below the mirror internal thermal noise, for all frequencies above 6-8 Hz. SAS is conceptually similar to the Virgo superattenuator chains but employs more advanced attenuation units. It is mainly a passive system, with the exception of inertial damping to neutralise internal resonances. The first elements of SAS are, an Ultra Low Frequency Inverted Pendulum and a Geometrical Anti Spring Filter (GASF). These oscillators absorb the microseismic peak perturbations. They are instrumented with an inertial damping system that couples the signal of higly sensitive position and acceleration sensors to actuators in the frequency range between 10 mHz and 4 Hz. This pre-attenuator is followed by a chain of 3 or 4 passive GASF filters, operating above 0.4 Hz, which isolates a multiple pendulum mirror suspension from the seismic noise. We expect to achieve even smaller residual, integrated r.m.s. payload motion than the 50 nm already achieved by Virgo. The measured performance and the passive nature of the GASF units ensure that the required attenuation is reliably achieved, free of external coupling and electronics excess noise and with a large safety margin. The extreme softness of the pre-attenuator allows for precision positioning with negligible external forces. Multiple independent SAS chains may be mounted in each LIGO vacuum chamber to handle multiple optical elements.

[H20.003] Characterization of LIGO II/SAS Inverted Pendulum as Low Frequency Pre-Isolation

We have developed an advanced seismic attenuation system for the future LIGO II detector. Our design consists of an Inverted Pendulum (IP) holding stages of Geometrical Anti Springs Filters (GASF), which isolate the test mass suspension from ground noise. The ultra-low frequency IP suppresses the horizontal microseismic peak. The three legs of the IP flex at Maraging steel joints, which have structural damping. Tunable counterweights allow for precise center of percussion tuning to optimize good attenuation up to the first leg internal resonance (\sim60Hz). The IP can be tuned to very low frequencies, by carefully adjusting its load. We achieved \sim12mHz pendulum frequency for one of the radial pendulum modes. The quality factor (Q) of the IP is compatible with structural damping. Q is proportional to the square of the pendulum frequency. Q was measured from \sim2500 (at 0.6Hz) to \sim2 (at 12mHz). All measured functions are in very good agreement with our models. We therefore expect excellent attenuation in the low frequency region, from \sim0.1Hz to \sim50Hz, which is necessary to obtain small off band residual motions of the payload. The extremely soft IP requires minimal control force, which simplifies any needed actuation (e.g. at \sim10mHz a load of 500Kg requires control forces of \sim2mN for a 1mm excursion).

[H20.004] Novel Design and Preliminary Testing of Linkless Geometric Anti Spring Filter Pre-Isolation

We have developed an advanced and simplified filter for attenuation of vertical seismic noise. This new design is based on the non-linear behavior of curved and mechanically constrained blades instead of the geometric anti-spring effect of the original design. Some of the internal filter resonances are eliminated or increased in frequency. The main internal blade mode frequency, however, is lowered by 30The wireless concept permits more compact assemblies, with the penalty of a smaller payload rating of a few hundred Kgs (instead of almost a ton). This makes the new units suitable for lower level filters in a SAS chain and for their implementation in multiple seismic attenuation chains in a vacuum chamber. These wireless filters will also be beneficial for use in smaller interferometers like TAMA.

[H20.005] Controls of Seismic Attenuation System (SAS) for the LIGO II Gravitational Wave Detector

The Seismic Attenuation System (SAS) has to be actively controlled over a frequency band of up to several Hz in order to damp its own rigid body modes (inertial damping), to generate DC local and global positioning, and to reduce residual rms motion to acquire the locking of the interferometer. The control system incorporates signals from local sensors (for displacement and acceleration) and the interferometer and generates adequate feedback signals for various actuators on different levels of the SAS chain. The control system is organized in a hierarchical scheme. With a large dynamic range at higher stages of the SAS, it damps internal modes of the system which minimizes requirements for the suspension control. The control system is a Multiple Input and Multiple Output (MIMO) that can be separated to simple Single Input and Single Output (SISO) feedback loops by using fast DSP boards. SAS controls are limited to a frequency band well below 10 Hz, to avoid noise injection in the gravitational wave band. Above this frequency, the SAS behaves as a completely passive seismic attenuator. According to simulated SAS performance based on measured seismic noise, achievable residual r.m.s. motion of SAS is a few tens of nm above 100 mHz. A similar system for VIRGO has already achieved 50 nm r.m.s. displacement.

[H20.006] Performance of Geometric Anti-Spring Filter (GASF) for Seismic Attenuation in Advanced Gravitational Wave Detectors

The frequency band of some interesting and possibly frequent gravitational wave events spans a low frequency range of up to 10 Hz. The main limit in this band is seismic noise, which requires a good isolation system to supress it down to the thermal noise level. We are developing a Seismic Attenuation System (SAS) for LIGOII to realize this low frequency isolation. The SAS, essentially a passive mechanical filter, is designed to provide low frequency isolation starting at 10-100 mHz for all degrees of freedom. The main obstacle for the isolator is to support heavy test masses, while retaining the softness of the system for low frequency attenuation, which is achieved by a Geometric Anti-Spring Filter (GASF) technique. The GASF consists of cantilever blades linked to a load using a geometric anti-spring effect. It can achieve an ultra low resonant frequency (\sim100 mHz), supporting a mass of a few hundred kg. The mechanical setup and materials used are specially selected to reduce creep, excess noise, and high vacuum compatibility. The prototype filters have been fabricated already, and their isolation performance measured. These results would be presented in this talk.

[H20.007] A very low noise monolithic Horizontal accelerometer.

We present a new low noise, low frequency, horizontal accelerometer. The mechanical design and the machining process aim to improve the sensitivity in the frequency region between 0.01 and 1 Hz, where metal internal friction and thermal elastic effects become critical. The accelerometer mechanics is shaped as a small folded pendulum in order to obtain a very low resonant frequency and low mechanical losses. A folded pendulum is essentially a mass suspended on one side by a simple pendulum and on the other by an inverted pendulum working antagonistically. The straight pendulum positive gravitational spring constant is balanced by the inverted pendulum’s negative one; by changing the center of mass position one can lower arbitrarily the resonant frequency. The only dissipation is in the anelasticity of the mechanical flex joint and in the readout/actuation system. If the spring constant is minimised, the mechanical losses are minimal. The monolithic design of the accelerometer eliminates the stick-and-slip friction localised in the flexure clamps. Low stiffness, 10 micron thick flex joints are achieved by EDM and electropolishing. The instrument is equipped with a low capacitance position sensor; the signal from the sensor is filtered by a PID controller and fed back to the mass through capacitive force actuator for feedback closed-loop operation. The sensor noise matches the expected thermal noise performances, 10^-12 m/\sqrtHz , with measuring range of a few microns. The expected sensitivity, less than 10^-11 m/ s^2 / \sqrtHz around 150 mHz, is a factor 30 below the state of the art limit. This accelerometer was designed to be integrated in the active control of the LIGO II mirror seismic isolators.

[H20.008] MSE: a mechanical simulation engine for the LIGO end to end model.

MSE is a fully tridimensional simulation. This code is based on an object oriented design. It provides a set of fundamental mechanical objects (masses, beams etc.) that can be combined to represent a complete mechanical system. Once the model is assembled, an equilibrium point is searched . Next the linearized dynamics of the system is evaluated around the working point. The model of each mechanical object can be refined according to the precision requirements of the effect under study. The code provides methods to automatically improve the internal representation of the system. The linearized dynamics is used to calculate the progress of the change of positions and forces, with given initial conditions. The matrix representation become larger when the actual system is complex, or very finely described. Basic rules to construct the matrix are well formulated. Models validated for simpler configurations, automatically work for more complex cases

[H20.009] Characterization of a Low Frequency Power Spectral Density f^-\gamma in a Threshold, Multi-stable Model

This study investigates the influence on the low frequency thermal spectrum of selective cooling of a number of normal modes in a mirror at room temperature. The aim of this study is the cooling of the mirrors for the measure of distances between masses in gravitational interferometers. The interest of this problem is that many systems are characterized by a typical f^-\gamma spectral tail at low frequency. The model used shows an interesting threshold type behaviour. When bi-stabe or multi-stable potentials are considered, a 1/f low frequency tail naturally appears. Both analytical and numerical methods have been used. The results show that no reduction of the power spectral density is appreciable outside the resonances of the system while an external force is introduced in order to produce in-resonance damping.

[H20.010] Thermal noise in coupled harmonic oscillators

The current generation of interferometric gravitational-wave detectors typically uses test masses suspended as single-stage, low-loss pendula. More advanced detectors are expected to suspend test masses from compound pendula to improve both seismic and thermal noise. Predicting the thermal noise in a single pendulum is straightforward, but for compound systems the math required can get hairy. Here we present a practical procedure for calculating thermal noise in a system of coupled oscillators, and a set of simple rules for doing ``back of the envelope'' estimates. We also show how the loss angles for each stage can be inferred from the Q's of the normal modes, and vice versa.

[H20.011] Gravitational gradients in gravitational wave detectors: data analysis methods

We present a method of analyzing seismic data at the sites of gravitational wave detectors to determine the possible influence of gravitational gradients as a noise source in the detectors. We use statistical methods to distinguish between local and gobal noise sources, as well as compare our findings to models of gravitational gradients (S. A. Hughes and K. S. Thorne, Physical Review D, Volume 58, 122002). We apply these methods to data taken at the Hanford LIGO site, and present preliminary results.

This work was supported by Pennsylvannia State University and the National Science Foundation. We acknowledge the collaboration of the LIGO project while taking the data presented.

[H20.012] Development of a Double Pendulum for Gravitational Wave Detectors

Seismic noise will be the dominant source of noise at low frequencies for ground based interferometric gravitational wave detectors, such as LIGO now in the final phase construction. Future interferometers installed at LIGO plan to use at least a double pendulum suspension for the test masses to help filter the seismic noise. We have constructed an apparatus to use as a test bed for double pendulum design. Some of the tests we plan to conduct include: measurements of seismic transfer functions of the double pendulum, by using a high precision vibration shaker; comparison of modal damping to point-to-point damping of the pendulum; measurements of actuator and mechanical cross couplings; and measurements of dynamic ranges of actuators, used to control the position of the double pendulum masses and how to split control between the intermediate mass and lower test mass. All these properties will be studied as a function of mechanical design of the double pendulum. Results will be presented from tests of a single pendulum supported by cantilever springs, which can be controlled in all six degrees of freedom.

[H20.013] Separation of LISA Galactic and Extragalactic Signals

Data obtained by the Laser Interferometer Space Antenna (LISA) is expected to include gravitational wave signals from several types of sources involving massive black holes at cosmological distances. In addition, there will be a very large number of signals from short period galactic binaries. For 1 year of observations, the average number of galactic binaries per frequency bin will be large enough at frequencies below 1 mHz so that most of their signals cannot be resolved. At higher frequencies, above roughly 3 mHz, most individual galactic signals can be solved for and removed from the data record. Studies have been started to investigate, for frequencies from roughly 1 to 10 mHz, how much information about the extragalactic black hole sources will be lost because of having to solve for the galactic sources. An axisymmetric but fairly realistic model is being used for the distribution of binaries in the galaxy. It probably will be desirable to subtract out some sources whose reality and signal parameters are only moderately certain, but whose frequency characteristics differ from the massive black hole signals of interest.

[H20.014] Low Frequency Gravitational Waves from White Dwarf MACHO Binaries

The detection of MACHOs in the Galactic halo has led to a great deal of speculation about the nature of the population. We examine the possibility that the MACHOs are white dwarfs of mass \sim 0.5 M_ødot, and calculate the contribution of white dwarf binaries to the gravitational wave background. The low-frequency (10^-5 Hz\, <\, f\, <\, 10^-1 Hz) gravitational wave spectrum from halo white dwarfs would be stronger than the expected Galactic disk contribution, and would dominate the stochastic background in the LISA waveband. These low-frequency gravitational wave detections will yield important clues to the nature of the dark MACHO population.

[H20.015] Starlight Deflection and Parallax Effects in the Gravity Probe B Relativity Experiment Data Reduction

The Gravity Probe B (GP-B) experiment is designed to measure the relativistic drift of a gyroscope in a free fall on a polar orbit around the Earth as predicted by the general relativity (the geodetic and frame-dragging effects).. The GP-B Data Analysis includes processing of telemetry data from several physical instruments placed on the GP-B spacecraft (gyroscope/SQUID readout system, science telescope, GPS receiver, spacecraft's control system).

In this paper we consider the effects the deflection of light from the Guide Star (GS) by the Sun and the GS parallax caused by the Earth's annual motion. To accurately determine the gyroscope's relativistic drift rate, these effects should be included in the data analysis. The purpose of this analysis is threefold:

1) to derive an exact mathematical model of the science signal that would include these effects; 2) to improve the accuracy of the relativistic drift estimation by using the a priori known magnitudes and time signatures of the deflection and parallax signals; 3) to show the feasibility of the independent determination of the starlight deflection and parallax by the GP-B science instrument.

The latter is intended, in particular, to the enhancement of the validity of the GP-B experimental results. The appropriate dynamical model and filtering approach are presented, and the results of simulations exploiting it are given and discussed.

Part H of program listing