Search

We have demonstrated a highly sensitive/wideband vertical-component seismometer using an astatic rotational pendulum to obtain a long natural period. This seismometer employs magnetic levitation for removing any parasitic resonances of a spring to support a weight due to gravity and the thermal dependence of the spring constant. The pendulum has a cylindrical plunger-type permanent magnet that has a weight at one side of its end edge. The plunger magnet is inserted into a uniform magnetic field generated by a window-frame-type permanent magnet, and attached to two crossed-leaf spring hinges as a rotational axis outside of the bore of the magnet. Magnetic forces applied to the plunger magnet counterbalance the gravitational force at the weight. To realize stable operation of the rotational pendulum without any unnecessary movements of the plunger magnet, a tilt of lines of the magnetic force in the bore of the window-frame magnet was compensated by a tilted magnetic-pole surface near to its opening. The field uniformity reached 10-4 owing to this compensation. The thermal dependence of a magnetic field strength of about 10-3/K was also compensated by as much as 9×10-5/K by Ni–Fe metal having a negative permeability coefficient. The metal was attached along the sidewalls of the window-frame magnet. To determine the feedback control parameters for a feedback control seismometer, the natural period of a prototype rotational pendulum was measured. It was more than 8 s, and was able to be changed from 5 to 8 s by using an additional magnetic spring, similar to the voice coil actuator of a speaker. This change was in accordance with theoretical calculations, and showed that the pendulum movement did not include a big nonlinearity caused by the tilt of the lines of the magnetic force. No parasitic resonances were found during experiments. A velocity feedback-control circuit and a capacitance position detector to measure the weight position were applied to the rotational pendulum for building a feedback control seismometer. Observations showed that the noise level of the seismometer was less than about 10-8 m/s at 1 Hz. This fruitful value is close to the specifications of the most sensitive seismometer, such as STS-I. However, low-frequency noise of about 10-7 m/s, caused by a buoyancy change at the pendulum weight arising from atmospheric pressure variation, could be recognized. To decrease the noise, a vacuum chamber to isolate the atmospheric pressure variation should be employed in the next step of the study. [ABSTRACT FROM AUTHOR]