Abstract
In this paper, we demonstrate an improvement in the accuracy of a low-cost smart temperature sensor, by measurement of the nonlinear curvature correction at multiple temperature references. The sensors were positioned inside a climate chamber and connected outside to a micro-controller via a network cable. The chamber temperature was increased systematically over a wide range from -20 °C to 55 °C. A set of calibration curves was produced from the best fitting second-order polynomial curves for the offset in temperature between the sensor and reference. An improvement in accuracy of ±0.15 °C is with respect to the mentioned temperature range, compared to the significantly higher value reported of ±0.5 °C by the manufacturer for similar conditions. In summary, we demonstrate a significant improvement in the calibration of a low-cost, smart sensor frequently used in research and academic projects over a useful range of temperatures.