EXPERIMENTAL VERIFICATION OF ADDITIVE ITERATIVE METHOD FOR CORRECTION OF ADC ERRORS IN LSB REGION

Abstract

In application of additive iterative method (AIM) for an analog-to-digital converter (ADC) a combination of AIM with nonsubtractive dithering enables correction of errors in LSB region. In the paper real application of the method is discussed and tested with orientation to measurements based on a single-chip microcontroller. Circuits available on a chip can be employed for the correction and only few additional components are needed. Inverse element (IE), which is a fundamental part of AIM, is created using pulse width modulation output. To achieve precise processing of a signal generated on the IE output technique similar to deterministic dithering is employed. For processing of the measured signal in the initial step of a correction nonsubtractive dithering with a stochastic dither is used. Theoretical curve of root mean square error dependency on standard derivation of added noise is compared with experimental. According to the real dependency a quasi-optimal standard deviation is found and used for final correction. Automatic correction of both linear and nonlinear error of ADC is proven experimentally. Finally a special ending condition of iterative correction is found which is suitable especially for systems with quantizer.