Accuracy and resolution of direct resistive sensor-to-microcontroller interfaces

Abstract

This paper analyses the accuracy and resolution of direct resistive sensor-to-microcontroller interfaces using theoretical and experimental methods. Three calibration techniques are evaluated: single-point, two-point and three-signal. This last method is a two-point calibration technique that needs a single calibration resistor. For each calibration formula, we analyse both the effects of the internal resistances of the microcontroller pins on the accuracy and the resolution, which is evaluated by the combined standard uncertainty of the calculated resistance. The experimental analysis was performed by measuring resistors in the range of Pt1000-type temperature sensors with two commercial microcontrollers (AVR AT90S2313 and PIC16F873). The experimental results were similar for both microcontrollers and agreed with theoretical predictions. For the AVR, the three-signal measurement method yielded a 0.01% relative systematic error and a 0.10 Ω resolution when averaging 10 calculated resistances.