O.B. Goncharuk, E.G. Skorobogatov, V.V. Yudin

The work is dedicated to probing of nonlinear effects in amplifiers dynamic characteristics of which (the dependences of an instanta-neous value of output value on an instantaneous value of input value) can have a different form. The dynamic characteristic is nonlinear function that causes a nonlinear conversion of an input value spectrum (signal) ? generation of new spectral components. In a series of cases the non-linearity of a dynamic characteristic will be useful property, however in many other cases it reduces to undesirable effects. In this work the non-linearity of a dynamic characteristic is considered as undesirable property and the problem of completeness of its quantitative characteristics used in practice is considered in common. A harmonic coefficient defined at one-signal impact and a level of beat-moire components defined at bisignal impact fall into those. The analysis conducted displays that both of mentioned parameters, as they are integrated and bound to specific conditions of external disturbance, can provide an insufficient information in some cases. There are good causes to guess, that the degree of distortion (breakup) of an input signal depends not only on its level (power, energy), but also on its spectral "saturation" ? number of spectral components, at that the latter can appear differently under different dynamic characteristics. The procedure of probing is based on computer simulation analysis of a given signal transmission through an amplifier processes with a dynamic characteristic specified with ascending power series. The degree of input signal distortion is characterized by a level of nonlinear distortions - relation of a part of output signal power concerning distortions to a total energy of an output signal (minus constant power component) calculated in decibels. Is taken into account that friendly signal component additively contained in an output signal generally can mismatch an input signal, differing from it by an amplitude factor and temporal shift. The problem of a friendly signal extraction from a complete output signal which arose at it was successfully resolved using functional analysis methods. During research, two qualitatively different dynamic characteristics surveyed, distinguished by "length" of an ascending power series. For the first characteristic the nonlinear terms of a series are rather great, but promptly descend, for second they are rather small, but descend slowly. The calculations were conducted with variation of an input signal level and number of components in its spectrum. The probing conducted fully confirms the made conjectures. The level of nonlinear distortions increases both with magnifying of an input signal level and with magnifying of its spectral components number and both these dependences appear more strongly at a dynamic characteristic with the slowly decreasing nonlinear terms. In the article the general results of research were cited. The main conclusion was substantiated that consists in following: at a spectral distribution complicated enough a harmonic coefficient and beat-moire components level have no sufficient completeness of the information being invariant to an form of an ascending power series approximating dynamic characteristic. In result under unfavorable conditions there can be undesirable effects which can not be predicted resting on these parameters only.