Calibration of digital 3-axis MEMS accelerometers: A double-blind «multi-bilateral» comparison

Nowadays, digital sensors based on MEMS/NEMS technology, are produced in very huge quantities by manufacturers. Beyond typical applications, e.g., in the predominant market of smartphones, these sensors began to be largely used in a wide range of monitoring systems and survey, in applied engineering and in advanced application, such as Internet of Things (IoT) and Industry 4.0. In these technological applications, accuracy and reliability of digital MEMS sensors, are emerging quality attributes of interest, allowing to improve, in general terms, the management of processes under investigation, on the basis of high data quality. On the other hand, quality of measurements data depends on a specific metrological characterization, in terms of traceability and accuracy. Although some methods and procedures for the digital MEMS sensors calibration, as well as proper sensitivity parameters for digital output, are nowadays available, it is not possible to calibrate every single device, but it is necessary to define suitable protocols of sampling, in order to deliver statistically acceptable levels of performance and reliability while reducing manufacturing costs. In this paper a doubleblind «multi-bilateral» comparison on large-sample scale of 25 digital MEMS accelerometers calibration data, is carried out and results are discussed.