E2E for LIGO I meeting ----------------------- September 25th, 2001 PSL (Peter King) ------------ I haven't been able to complete the measurements I was hoping to do in the last reporting period due to making some alterations with the hardware - both the modecleaner servo and the frequency stabilization servo. The configuration should have settled down now, so the frequency noise measurements should be able to proceed. There was a hardware problem which stopped me from obtaining the latest reflected and transmitted intensity noise spectra. This will be fixed this site visit. However some pre-modified modecleaner servo spectra are available. Input Optics (Dave Reitze) ------------ nothing significant to report Isolation (Mark Barton) ------------ I haven't done anything more on LIGO I E2E stuff, but I now have a Mathematica model of the Advanced LIGO Quad Pendulum which may be of interest. It's based on the same framework as the LIGO I model I did and has the following features: * fully 3D * masses are rigid bodies * wires are stretchable but don't have violin modes or rigidity near attachment points * blade springs have x, y, z DOFs, and the ones with more than one wire attachment point (all but the top two) have a pitch DOF (torsional compliance) * fully analytic up through the point of building expressions for the potential and kinetic energy * solves for the minumum of the potential and finds eigenvalues and eigenvectors numerically * utility functions to plot mode shapes and transfer functions Features to be added soon: * damping * elasticity from bending of wires near attachment points * export of a state-space model to Matlab/Simulink For maximum comparability the initial version aims to be a strict superset of Ken Strain's Matlab model and incorporates his quadopt.m and opt.m as close to verbatim as was possible allowing for differences in syntax. I'll be posting it at some time this afternoon. Optics (GariLynn Billingsley) ------------ I now have a new calibration data file. With this file rms errors are down from ~.5nm to ~.3nm rms over the full data aperture of 150mm. The inner rms (over 80mm) remains nearly the same as before the new file was generated. Complete data are not yet posted to the website. Bill Butler: ------------ I worked on using e2e and reported back to Hiro and Biplab with useability questions regarding the Power Spectral Density funtion. Robert Schofield and I took some measurements of the seismic noise within the corner station. In some cases measurements were limited by the equipment. Measurements are: Dist RM to BS ~13ft BS to ITM ~30ft RM to ITM ~43ft Distance (ft) 0.7 Coherence (Hz) 1 >100 6 ~50 15 ~20 30 ~10 43 ~6-10 (z-z,y-y ~6, x-x ~10) 2 kilometers ~0.4 There did not appear to be any systematic coherence between x-y, y-z or x-z and the seismometers do not have tilt measurements so no data on thetaX, thetaY or thetaZ were taken. Also measurements on relative phase or time delay between impulse responce were not taken. I have not yet incorporated these measurements into the Han2k model. NOTE from Ed Daw (Fri, 21 Sep 2001) ---------------- The other night I took some careful, high resolution spectra using Streckeisen seismometers at Livingston. I thought these might do for input to the e2e. Bear in mind that these spectra were taken at night so that the 1-3Hz region is relatively quiet. (See the comments at the beginning of the attached file for details of the data format, etc.) A summary of what I measured is posted on the Livingston elog from 19th September 2001. Luca Matone ----------- 2k LSC Characterization effort: --Electronics-- The transfer function of the LSC electronic chain, including the Whitening, Anti-Aliasing , Pentek and Anti-Imaging boards have been measured in the past few months. A matlab model was setup to reproduce the measured transfer functions and it was found to agree with the measurement only below 1-2kHz. Now the agreement is within 5dB and 5deg below 8kHz. A matlab function has been written and tested to reproduce the transfer function from the Whitening board to the Anti-Imaging board with any digital filter combination enabled. --Phase accounting-- Once the electronics was determined, we began measuring the Closed Loop tf of one arm and the simple Michelson. For a measurement done on the 2k Y arm, the phase was accounted for to within 10 deg between 100 Hz and 1kHz. We also measured the closed loop tf of the simple Michelson. By factoring out the electronics tf, we were able to get a first estimate of the product sensor*actuator (FMX) tf between 40Hz and 200 Hz. Presently working on a technical note summarizing the work completed so far. << Discussion: Stan suggested the next immediate step should be translating the Matlab code to an equivalent e2e one. >> Virginio Sannibale ------------------ Virginio Sannibale talked about his recent progresses in mechanical simulation and violin modes. << Discussion: When Virginio talked about splitting of violin modes due to difference in suspension wire diameters, Stan pointed out that a similar splitting might also be found for the simple fact of geometry ... for the fact that the wires are not vertical and there's a difference in their oscillation in two lateral directions of the wire. >> Biplab & Hiro ------------- Time domain modal model consistency checks, null tests. Mode mismatched runs using lo-finesse and Hi-finesse FP cavity and the Han2k model. ------- << Discussion at the end: Dennis said he was particularly interested in knowing how the ground motion in Modecleaner (at LLO) might get transferred into the frequency noise and invited others to get into pursuing this problem. On this problem, Stan thinks high frequency isolation for stacks is much better than the LIGO requirement and so probably it's possible to improve the low frequency isolation (and so avoid the above-mentioneed problem) by removing one or two layers of stack, thereby degrading hi freq isolation but taking care that hi-freq isolation does not become worse than the requirement level. >>