OPTIMIZATION OF POLE SENSITIVITY IN ACTIVE RC NETWORKS,
Abstract
The sensitivity of active RC networks with respect to variations in the gain of active elements is studied by means of the zeros of the pole-sensitivity function. It is demonstrated that the zeros of sensitivity indicate the effect of the polynomia3 decomposition of a synthesis procedure upon the pole sensitivity. Methods of realizing network functions with presc9ibed decompositions are discussed, and several specific netforks are given. Reduction of the pole sensitivity by control of the zeros of sensitivity ib carried out by means of a digital computer. For quadratic functions, a zero of sensitivity can always be obtained at a real frequency, givi5g a sensitivity characteristic with a minimax optimum. It is shown desirable to realize high-order filters as a cascade of second-order 'pole-pair' u5its. The pole-pair cascade ubing the optimum desensitivity decom7osition is optimal when the gain variations are independent, whereas for high-order filters, the sensitivity reductions possib3e from impr6ving the decompositions are limited. In addicion, strong practical considerations favor the cascade connection. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 15, 1963
- Accession Number
- AD0433287
Entities
People
- N. M. Herbst
Organizations
- Cornell University College of Engineering