It is necessary to keep at the disposition test signals with a high spectral purity for series measuring e.g. in the area of metrology or electronics. Commercial signal generators do not satisfy higher requirements in this area. Our contribution describes chances of the improvement of their quality while using of suitable filters. There is described the construction of a special generator with a high spectral purity of a signal.

The paper presents an improved implementation of a method pointed out for the static characterization of the new generation of high resolution Digital to Analogue Converters (DACs). Interesting aspect of this method is that the problem of the signal acquisition with high resolution is shifted to the simpler problem of the signal acquisition with high speed. In particular, the method is based on the comparison of the DAC output voltage with the analogue reference sinusoidal voltage and the detection of the Zero Crossing Sequence (ZCS) of the resulting signal by high speed ADC. The variation of the ZCS respect to that occurring in the sinusoidal reference signal is used to evaluate the DAC transfer characteristic. The improved implementation concerns the proper design characteristics that can be used as selecting criteria of the test equipment. In particular, the substitution of the ADC for the detection of the ZCS by the comparator device is investigated without affecting the accurate detection of the zero crossing. Moreover, the effects of the noise parameters affecting the generator feeding the reference analogue signal, as amplitude and phase noise, on the evaluation of the DAC static characteristic is analysed by considering test equipment using the comparator device and the high speed ADC, alternatively.

The main point of new approach to the theory of sigma-delta modulation is application of a discrete two-values distribution law for a quantization noise instead of a uniform distribution law accepted before. Due to the new approach, the variance (standard deviation in the square) of the quantization noise becomes dependent on input signal by parabolic function. The noise vs. frequency is found for different input signals taking into account all frequency range from zero to half of sampling frequency. Using dependence of standard deviation on input signal and frequency, different characteristics of sigma-delta modulation can be predicted, including SNR. Difference between analytical and simulation results for new theory can be driven to any small value.

In this paper the idea and the selected results of numerical simulation of an uncertainty designation procedure in the measurement channel equipped with ADC is presented. The basic principles of that procedure implementation are shown in literature. The assessment of correctness of the mentioned procedures is the main aim of the carried out simulations.

In this paper a dynamic behavior model of analog to digital converters is proposed. The model is aimed for post correction in wideband applications. The suggested post correction method is a combination of look up tables and model based correction.
The model consists of three components. The first is a component represented by a Hammerstein model; that is a static nonlinearity followed by a time invariant linear filter. The second component is a nonparametric model caused by significant deviation from the characterized integral nonlinearity and the output from the Hammerstein model. The third component contains of the remaining deviation and is considered as a random model error.
Results from simulations verify that the examined ADC can be described by an ordinary Hammerstein model and a static look-up table.