Łukasz Małkiewicz

Reduction of influence of gain errors on performance of adaptive sub-ranging A/D converters with simplified architecture

The paper presents the results of studies pertaining to the influence of gain errors of inter-stage amplifiers on performance of adaptive sub-ranging analog-to-digital converters (ADCs). It focuses on adaptive sub-ranging ADCs with simplified architecture of the analog part – using only one amplifier and a low resolution digital-to-analog converter, that is identical to that of known conventional sub-ranging ADCs. The only difference between adaptive subranging ADCs with simplified architecture and conventional sub-ranging ADCs is the process of determination of output codes of converted samples. The adaptive sub-ranging ADCs calculate the output codes on the basis of sub-codes obtained in particular stages of conversion using an adaptive algorithm. Thanks to application of the optimal adaptive algorithm, adjusted to the parameters of possible components imperfections and internal noises, the adaptive ADCs outperform, in terms of effective resolution per cycle, conventional sub-ranging ADCs forming the output codes using simple lower-level bit operations. Optimization of the conversion algorithm used in adaptive ADCs leads however to high sensitivity of adaptive ADCs performance to the inter-stage gain error. An effective method for reduction of this sensitivity in adaptive sub-ranging ADCs with simplified architecture is proposed and discussed in the paper.

Adaptive sub-ranging ADC with autocalibration of internal components offsets

The paper presents a new method for autocalibration of offsets of internal components in adaptive sub-ranging analog-todigital converters (ADCs). The adaptive sub-ranging ADCs, like other iterative ADCs, are very sensitive to offsets caused by technological imperfections. To achieve higher effective resolution in a sub-ranging ADC, offsets should be accordingly diminished, which is achieved using sophisticated techniques usually causing increase of power consumption as well as complexity and size of ADC. But the structure of adaptive sub-ranging ADC and the way of calculation of output codes create a possibility to compensate offsets during conversion of input samples. Moreover, using the internal digital-to-analog converter (DAC) as a source of calibration signal in the procedure of offsets estimation and addition of a calibration program in the digital part of the converter, enable automatization and autonomization (no additional instruments are needed) of the calibration process. Usefulness and limitations of the proposed solution were confirmed in computer simulations whose main results are presented and discussed in the paper. Implementation of the proposed solution enables to increase effective resolution of ADCs and simultaneously to weaken the requirements on acceptable level of offsets of ADC components.

Difficulties with assessment of adaptive cyclic ADC nonlinearity

The paper discusses the possibility of application of nonlinearity measures – INL and DNL, as recommended in IEEE Std 1241, to adaptive cyclic ADC. The first difficulty confronted was not knowing the nominal transfer function of adaptive ADC. To overcome this difficulty, there was proposed an algorithm for its analytical determination on the basis of the models and parameters of the components of the analog part of the converter and of the codes computing algorithm. Using this tool, validated in simulation experiments, there was established a non-uniformity of the adaptive ADC nominal transfer function, concerning both thresholds and quantization levels. The latter is the second, unremovable obstacle to application of INL and DNL measures to characterization of adaptive ADC nonlinearity. In the paper, there are explained the causes of the non-uniformity and there is shown a possibility and conditions of realization of adaptive ADCs with uniform nominal transfer function. Finally, there are presented the results of simulation analysis of the influence of the established non-uniformity of nominal transfer function of adaptive ADC on THD - another popular measure of ADC nonlinearity recommended in IEEE Std 1241 - which show that the non-uniformity does not worsen THD compared with THD obtained for adaptive ADC with similar but uniform transfer function.

Analytic and simulation-aided methods for improvement of intelligent cyclic ADCs performance

The paper presents a brief overview of main results obtained by the author in a new direction of researches in the field of analog-to-digital (A/D) conversion theory and applications. The object of analysis in the work is a new type of converter - intelligent cyclic A/D converter (ICADC), whose particularity is transition to computing codes of samples using efficient iterative algorithms. This creates the possibility of adjustment of the parameters of the analogue and digital part (codes computing algorithm) of ICADC, which allows to improve the quality of conversion. In the paper there are discussed the developed methods of assessment and improvement of ICADC performance.