Metrological qualification of data acquisition systems

Abstract

The paper treats the metrological qualification of data acquisition systems illustrating general ideas and a practical test method. The necessity of an improved industrial standard for users and manufacturers is discussed, remarking the convenience of a model identification approach. Following this line of reasoning, a new test method (modified statistical domain analysis) for simultaneous measurement of nonlinearity, amplitude noise, time noise and global error of a digitizer is presented, putting in evidence that it provides identification and validation of a first-order error model of the system. This is obtained via a single acquisition of a reference sinusoidal signal and a unique consistent scheme of computations. Simulation and experimental results illustrate the practical advantages and limits of the test, the latter emerging at the higher frequencies, when the adopted memory-less model of the ADC is no more valid.