| The LIGO
|First Page||Third Page||Back to Front Page|
LIGO'S First Science Run - A Special Report
The challenges faced by the Washington observatory occurred mainly before the Science Run, although the resulting delay had a bright side in that it afforded time enough to fine-tune Hanford's 4-km interferometer into a scientifically potent tool ready for use. Meanwhile at Livingston, Louisiana, the difficulties arose during the run itself. At both sites the culprit was the same: ground motion. And for Louisiana this came in the form of seismic noise created by nearby logging activities, as well as meteorological disturbances sweeping in from the active Gulf of Mexico. Such difficulties had always been planned for, and compensations were inherent in LIGO's original blueprint. But acts of nature are notoriously unpredictable, and nothing gives LIGO's long-range planners a wee case of the jitters more than seeing their strategies tested against an often arbitrary opponent. Mark Coles, chief of the Livingston site, tells how the Science Run teamers hunkered down and carried on despite stormy weather and the frequent shouts of "Timber!"
The performance of the LIGO Livingston Observatory's (LLO) 4-km interferometer has been significantly enhanced since the Seventh Engineering Run (E7) conducted late last year. The interferometer now operates robustly in a power recycled state, and the commissioning of the common mode servo has greatly reduced frequency noise and many common mode error contributions. Other improvements were made to the digital and analog filtering, commissioning of the wave front sensing at the antisymmetric port, and in the signal routing to reduce cross talk. See our Livingston column in the May 2002 newsletter for additional details.
The evolution of the noise spectrum of the LLO 4-km interferometer is shown in the figure below. The current state of the interferometer is about as indicated by the 6/13/02 line. Relative to E7, the interferometer operates with more than a factor of ten greater sensitivity everywhere, and nearly a factor of 100 better sensitivity in the critical band from 100 to 400 Hz.
Ground motion at LLO has been the major limiting factor in our ability to lock the interferometer. This aggravating seismic noise occurs almost exclusively during the day and peaks in the 1-3 Hz region. It arises primarily from activities such as the harvesting of trees in the surrounding forests. We had hoped that the piezo-electric actuator system (PEPI) would increase our duty factor by making it possible to lock the interferometer in a low noise mode during the transitional hours at the beginning and end of the work day, when logging and other man-made noises were either starting or ending. (See this document by Joe Giaime (LSU) and Brian Lantz (Stanford) for a recent overview of the PEPI system.) Alas, the tree loggers commenced work immediately adjacent to the site. Harvesting of trees north of the Laser and Vaccum Equipment Area (see map below) began just prior to the start of S1 in the shaded areas to the right of the squiggly red line.
We were thrilled when we found out that the logging company would not work on Labor Day weekend. And we were dumbfounded when another logging company began work on Labor Day adjacent to the south arm. With these activities so close to the site, even the set-up movements at the start of the day cause ground motions large enough to prevent the operation of the apparatus, so there was no transitional time to be had.
Another act of nature that occurred during this time was the presence of two tropical storms in the northern Gulf of Mexico simultaneously. Storms such as this are believed to excite low frequency ground motion through coupling of wave action into the ocean floor, and this pressure wave propagates on through deep geologic structures for very long distances. The micro-seismic feed forward system was designed to handle precisely this kind of problem, and we were very glad to have it available to us. An example of the very large micro-seismic motion encountered during the run is shown in the figure below.
We had anticipated that this run would be dominated during daylight hours by man-made seismic noise sources. Our plan was to operate each night, and possibly over the three-day Labor Day Weekend. Taking into account the number of daylight hours, the twice-per-night seismic disruption caused by passing trains, and a planned calibration, we estimated that the Livingston interferometer would be able to remain locked about 35 percent of the time. Our actual experience was slightly better than that at 43 percent, the increase due primarily to slightly less logging activity than feared. The breakdown of time distribution during the run is seen in the chart below.
Operationally, we learned some important lessons. For example, during this run we subjectively balanced the desire for long, continuously locked sections of data in a stable operating configuration against the observation that the relative alignment of the suspended optics was slowly drifting over time, and that this drift was degrading the noise performance of the interferometer at frequencies above about 150 Hz. This experience reinforced our motivation to commission the remaining angular control servo systems planned for the interferometer.
An important related lesson had to do with the noise stationarity of the interferometer. Many new data monitoring tools were developed and used during this run to investigate and characterize the above-mentioned noise behavior. Various glitch monitors indicated that sudden transitions or discontinuous behavior in the output of the interferometer were rare (although they were common during E7), evidently showing that refinements in the various servo upgrades since E7 have controlled this behavior.