Temperature Drift Performance of Fluxgate Sensor with Open-loop Measurement

The magnetic fluxgate sensor exhibits a significant temperature effect in its practical applications. This paper presents an in-depth analysis of the working principles of both open-loop and closed-loop measurement circuits commonly used with magnetic fluxgate sensors. Based on this theoretical foundation, the study focuses on how the sensor's intrinsic characteristics vary with temperature under open-loop conditions. The objective is to provide experimental evidence that can guide the design and optimization of temperature drift suppression techniques in closed-loop configurations. By building upon the fundamental operational principles of the magnetic fluxgate sensor, the paper derives and compares the circuit transfer functions for both open-loop and closed-loop measurement systems. It is demonstrated that the error sources present in open-loop measurements are more directly reflective of the sensor's own performance characteristics, as they are not masked by feedback mechanisms inherent in closed-loop designs. To achieve accurate and reliable open-loop signal detection, a dual-operational amplifier (dual-op-amp) bandpass filter was employed to isolate the second harmonic signal, followed by phase-locked amplification to precisely measure both the amplitude and phase of the open-loop output. Performance temperature tests were designed based on the distinct behaviors of different sensor parameters under thermal variation. Experimental results obtained over a temperature range from –40℃ to +80℃ show that the zero-point drift of the magnetic fluxgate sensor remains within ±0.5 nT, while the phase shift reaches up to 60°. Additionally, the open-loop gain varies by approximately ±5%, and the noise level fluctuates between 4～7 pT·Hz^1/2 at 1 Hz. Although the signal phase is the only parameter that undergoes a substantial change in open-loop measurements, the phase-sensitive demodulation mechanism in closed-loop systems is highly responsive to such variations. Consequently, the observed phase drift has been experimentally verified to result in significant zero-point drift in closed-loop measurements.