|The Web Newsletter||Front Page||Hanford||Livingston||MIT|
LIGO Caltech NewsLow Frequency Seismic Isolation At LIGO: The Vertical Filters
Caltech's Bill Kells Teaches SURFing Students About Waves
A good seismic noise isolation system must progressively attenuate the incoming perturbations in all degrees of freedom. Five of these (the two horizontal translations, the two tilts, and the vertical torsion) are easily dealt with by means of low frequency pendula or oscillators. The tough one is the sixth direction, the vertical motion. In this direction the very structure that supports the optics' weight may represent a noise highway through which seismic noise can channel and from which it will eventually spread out to all the other degrees of freedom.
All gravitational-wave interferometric detectors that fail to expend sufficient effort to deal with this problem will eventually be limited by their seismic attenuation performance in the vertical direction. This is the reason why VIRGO , the French/Italian gravitational-wave detector project, has made such a large investment in this field and why LIGO is following suit.
The secret of vertical attenuation is the use of soft and quiet springs to hold the payload while keeping it at low vertical resonant frequencies.
The catch is that to hold the heavy mirrors and reaction masses it is necessary to use stiff springs which entail high resonant frequencies. Consequently it is necessary to soften the springs at their working point, a problem solved by VIRGO through means of a complex but effective Magnetic Anti Spring (MAS) system.
Tipped off by the Australian Interferometer Gravitational-Wave Observatory (AIGO) scientists, LIGO is researching much simpler and potentially better performing Geometric Anti Springs (GAS).
The concept of GAS involves the use of VIRGO-like cantilever blade springs mounted at an angle and connected to the load by means of a counter inclined link.
Figure 1 at left, shows a simulated profile of LIGO's GAS filter blade under load. A flat blade (dashed) is bent under load (black curved line) and linked by means of a wire (straight line) to a load disk. The vertical force applied to the disk stays almost constant for small vertical movements around the working point.
The horizontal forces of each spring are nullified by a symmetrically mounted spring. The vertical component of the force transmitted by the link is modulated by the changing link angles, decreasing near the chosen working point by almost exactly the amount of the increase of the recall force of the stiff cantilever spring. The result is that as increasing load drags down the load disk, the recall force stays almost constant and the spring acts as an extremely soft one.
The GAS concept has been simulated and tested at Caltech. (The experimental setup, built after the simulation, is shown in Figures 2, 3 and 4 below.) The contraption is built in Unistrut (TM), each blade is clamped between two disk brakes that allow the tuning of its inclination. The counter-inclined link is made with steel wire looped around a nail stuck in the tip of the blade.
Figure 2 (below left) shows the experimenters with their experimental setup: Virginio Sannibale (left) Riccardo DeSalvo, and Sandro Bertolini (right). Absent is Giancarlo Cella. The Unistrut (TM) device is intended not as a real filter but purely to check the GAS principle. Figure 3 (next in line) exhibits details of the experimental setup. Note the big wrench in the upper right corner of the photo which is used to tune the blade's tilt when the disk brakes are released. Note also one of the radial distance tuning screws (top left) acting on the disk brake base. Figure 4 (last in line) is a side view of a blade. Compare with the simulated shape of Figure 1.
Despite the crudeness of the system, this prototype completely validated the simulated model. In Figure 5 at right the comparison between predicted and measured results is made.
Figure 5: The Vertical Resonant Frequency versus payload disk height for different radial positioning of the blades. The left graph is simulated data and the right one is measured data. The three measured data curves correspond to a progressive advancement of the blades by xx and xy mm.
--A resonant frequency minimum of 250 mHz with a load of 150 Kg (330lb) could be achieved.
--The frequency minima with GAS are several times wider than in a corresponding MAS case.
--Additionally a much lower thermal sensitivity is calculated.
--The mass and the perturbations of the magnets are eliminated.
The GAS strength can be simply tuned by changing the radial distance between the blade's base and the link point on the load disk by means of the horizontal set screws behind the support of each disk brake visible in Figure 2. The effect of this tuning is well visible in Figure 6 at left.
In a real filter the vertical resonant frequency would be factory pre-set. A dimensional tolerance of 0.5 mm is required to set the filter at 250 mHz.
The achievement of lower vertical resonant frequencies was impeded in the prototype by the slip and stick friction of the crude links at the tips of the blades.
In the detail of Figure 7 (below left) some counterweights are visible at the tips of the blades; they are intended to nullify the ill effects of the blade's internal resonance. Additionally the resonance would be damped by mechanically resonant eddy current dampers.
The crew is now working on the design of the improved isolation filter shown in Figure 8 (below right).
In August 1995, one of Caltech's own returned to become a member of the school's professional staff as well as a LIGO team member. Even as an undergraduate at Caltech, Bill Kells had connections to LIGO from his association with Detector Group Leader Stan Whitcomb, students together during grad school.
So it's not surprising that Bill was interested in the students participating in LIGO through the Summer Undergraduate Research Fellowship, or "SURF," program. Bill "saw a need for students to have practical application at an elemental level rather than mostly cursory cartoon illustrations in the classroom." To that end he launched a series of informal tutorials for the students, beginning with a visual tour and then using mathematics and equations to show them how to think about LIGO on a physical level.
The Summer group of incoming SURF students was Bill's impetus to begin the tutorials, but the group has attracted other staff persons as well. Meeting every morning for a short half-hour, the group has averaged between 8 and 10 participants each day. The tutorial is confined to a topic that can be captured on one piece of paper, and acquaints the group to the elements and language that are the essence of LIGO. Bill believes that "a small daily dose of exposure to the concepts of LIGO results in better absorption."
For Bill, personal enjoyment comes from facing the challenge of keeping his tutorials at a easy enough level for all group members to understand, but without ever trivializing the science of LIGO. Teaching is a fairly new experience for him, but with a smile he says, "As it goes, 'Where there's a will, there's a way.'"
Bill's tutorials have wound down now with the end of the SURF program in mid-August. But already there seems to be interest from other schools for a similar introduction to the gripping, if sometimes "hidden" excitement of physics.