부산대학교 도서관

To eliminate the unwanted spectral components in the processed signals in the automatic control and monitoring systems, most often the low-pass (LPFs) and band-pass filters (BPFs) are used. They are performed on the basis of the analog (continuous) circuits, as well as on the analog-to-discrete (switched capacitors) and digital ones by means of the software or hardware implementation. In such filters the information delay causes a dynamic component of their error. The dynamic error is a part of the total error of the automatic control and monitoring system and it reduces the permissible velocity of its operation. The article considers the questions of the obtaining analytical dependences of the dynamic error values of the elliptic filters (Zolotarev-Cauer filters) to solve the on-line direct and inverse problems of the error distribution of the automatic control and monitoring system for justification of its operation velocity. An important advantage of Zolotarev-Cauer filters is that there are zeros in the transfer function, which can make possible to significantly weaken the parasitic noise with the known frequency components, by locating transfer function zeros at these frequencies when designing the filter and taking a correct solution of the approximation problem. It is shown that for the LPF the maximum dynamic error values can be obtained with high accuracy. For the BPF the approximation errors of dependence of the maximum dynamic errors on the filter parameters are determined by the units of per cent, but in some cases, they can reach 20%.