An essential disadvantage of application of the oversapling sigma-delta A/D converters in high-resolution DC measuring systems are local peaks of quantization errors observed for some specific input signal values. The best method of elimination of those peaks is application of additional, deterministic, digitally generated dithering signal to the converter input. When averaging period of the converter digital output filter is a exact multiple of the period of the dither signal, this topology is intrinsically subtractive and elimination of DC residue is achieved without the need of additional digital subtraction. Experimental results of commercially available, nominally 24-bit sigma-delta converter with added simple self-subtractive circuit proved, that full elimination of the most significant peaks of errors was obtained. Increase of effective resolution of this converter (ENOB) was from 1,5 to 4 bits.

The combination of dithering and correcting table is the way of improvement not only of ADC resolution but also of linearity - effective number of bits. Dithering is used to reduce small-scale errors, such as quantization error, while correcting table reduces large-scale errors.
The contribution describes design and practical experience with ADC unit based on single-chip micro-controller 80C552 that uses above described correction procedure. Sub-quantum accuracy has been achieved.

The paper deals with some metrological parameters of the ADC implemented on AT90S8535 and ADuC812 microcontroller chips. As a criterion of ADC precision, which allows to determine the operational condition limits, the number of effective bits has been chosen. More over, the DNL and INL measured characteristics are presented. The results are compared to some of the data brought out in vendors’ data sheets.

The influence of different ADC nonlinearities on the measurement of ADC intermodulation products is discussed. It is shown that the amplitudes of harmonics and intermodulation products can be predicted only if the low order nonlinearity of the transfer curve is dominant and if a sufficient level of input noise or dither is used.

Time-domain testing of A/D converter in single tone mode amounts to the identification of fundamental and harmonic tones in noise. The variable projection method reduces this problem to frequency estimation by minimising a onedimensional function. In this paper it is shown that the bias due to ignoring the harmonic distortion effect is negligible, and this approximation leads to fast frequency estimation algorithms. Three methods are examined (Newton, Gauss-Newton and Golden section search) and compared to the method which takes into account the effect of the distortion through simulations and experimental data. The convergence and implementation issues of the algorithms are also treated.