A miniaturized low-power vector impedance analyser for accurate multi-parameter measurement

Abstract Distributed measurements are important in many application fields, from environment to biomedicine. In both cases, the sensor nodes employed in the measurement network have to satisfy many requirements; among them, the most important ones are: i) low power consumption, ii) miniaturization, iii) adequate accuracy, and iv) capability of multi-parameter measurement. This paper presents a Vector Impedance Analyser (VIA) architecture that satisfies these main requirements. The architecture is specifically devised to be interfaced with an array of impedimetric sensors for environmental applications, such as distributed water monitoring, or for mobile-Health/wearable biomedical devices. The proposed architecture is based on delta-sigma digital-to-analogue (D/A) conversion for the generation of the low-noise excitation and on band-pass delta-sigma analogue-to-digital (A/D) conversion for the low-power and high-accuracy acquisition of the impedimetric sensor response. The proposed combination of delta-sigma D/A and A/D conversion allows to i) implement many measurement cores in a single silicon chip with reduced dimensions, ii) achieve a fair accuracy/power trade-off, and iii) tune the operative frequency in real time so as to span the target portion of the frequency domain. A prototype of the VIA is implemented in a 3 × 6-cm PCB board to investigate the potentialities of the architecture. The low-noise analogue circuits of the architecture are implemented in an Application Specific Integrated Circuit (ASIC), while part of the digital circuits are implemented on a commercial microcontroller for better testability purposes. The prototype embeds four independent cores to allow real-time multi-parameter measurement. To prove the performance of the proposed VIA, the prototype is characterized in terms of noise (input-referred noise between 20 mΩ and 70 mΩ in 10-Hz bandwidth, i.e. from 25 to 92 ppm of the full scale), accuracy (average uncertainty of 0.14% of the full scale for the magnitude and 0.72° for the argument, accounting for the limited accuracy of the reference instrument used for calibration), and power consumption (approximately 125 mW per-core including the power consumption of the microcontroller and the ancillary circuits used for power management and communication). The multi-parameter measurement capability is demonstrated by the realistic case-study of estimating the concentration of total dissolved solids in a potassium chloride (KCl) solution by means of direct concurrent measurements of conductivity and temperature.