We propose a mathematical model describing quantization frequency selection in the measuring channels of intelligent systems depending on the parameters of noise. We consider internal and external sources of noise, seek the reasons for noise generation, justify the necessity of recording noise in the analog lines of measuring channels of intelligent systems. The parameters of our mathematical model of noise are: stationary white noise with normal distribution, an anti- aliasing bandwidth filter, the variance of noise at the input of an analog- digital converter, the correlation coefficient of noise. We evaluate the dependence of the selection of quantization frequency value on the parameters of noise. We show that if the right sampling frequency is selected in the conditions of broadband noise measuring then the systematic component of error is determined by the bandwidth of the low frequency analog anti- aliasing filter, while the random error is determined by the bandwidth of the low frequency digital averaging filter.
