Ring-core fluxgate magnetometers for use as observatory variometers

Abstract Ring-core magnetometers, based on the published design for the MAGSAT fluxgate magnetometer, are or will soon be in use in observatories in the U.S.A. and in Canada. Their low noise level and good stability with both temperature and time make them well suited for this purpose. We describe the critical performance features and design aspects for these instruments. Noise in all ring-core magnetometers has approximately a 1/frequency response in power spectral density. The principal source of the noise is from the ring-cores themselves. In the best instruments, which use a d.c. bias supply to enhance dynamic range, noise from the voltage reference used in the bias supply can approach significant levels at large offset values (∼ 60 000 nT). In these instruments noise in the 0.001–100 Hz band is typically less than 0.1 nT amplitude, and reduces accordingly with reduced bandwidth. Good thermal stability for the sensor is achieved by matching thermal expansion coefficients of all the sensor materials, and by compensating thermal expansion of the sensor with an increase in feedback current. Changes in the sensor-winding resistances provide the cues for the compensation circuitry. In settings where long cables separate the sensor from its electronics, platinum resistance elements included with the sensor winding can ensure that variations in cable resistance will not significantly affect the temperature compensation. Thermal stability of 0.1 nT deg −1 are typically achievable for the sensor. Thermal stability for the electronic components depends primarily on the stability of the d.c. bias supply. These can be designed for a stability of 3 ppm deg −1 , which translates to 0.2 nT deg −1 for a field of 60 000 nT. Long-term drift results primarily from the aging of the sensor winding and the d.c. bias supply components Engineering data for the best available components suggest that drift of less than 50 ppm yr −1 is possible. This translates to 3 nT yr −1 in a 60 000 nT field. Our results from experimental data indicate that drift rates of 1 nT yr 1 are achievable. More experience will be required to confirm these values. The long-term and thermal stabilities achievable with ring-core magnetometers raises an additional issue. In many settings it is difficult to ensure that sensors will not tilt. However, as little as 3 arc-seconds of tilt can cause a 1 nT error in a variometer measurement. In order to take maximum advantage of the capabilities of the ring-core magnetometer it is necessary to either stabilize the sensor orientation to a few arc-seconds, or to measure sensor tilt and rotation to the same accuracy.