OPTIMIZATION OF DISTRIBUTED-LUMPED-ACTIVE NETWORKS,
Abstract
There are many engineering systems which cannot be reasonably modeled by lumped elements, but must be modeled by distributed elements. This is especially true in electrical networks at high frequencies and in integrated circuits. Herein are discussed some computer optimization methods for the design of lumped-distributed networks in which the parameters of the network are adjusted until certain engineering requirements are met. An efficient way to adjust the parameters is to utilize the gradient of the performance index. Two different algorithms are developed to generate the gradient of the performance index with respect to the adjustable parameters. The first method assumes that the network has a cascaded structure, and the gradient is obtained by calculus of variation techniques. The second method which proved to be more efficient is the adjoint network method. The adjoint network method developed previously, using Tellegen's Theorem, for lumped-active networks in branch formulation is extended to distributed-lumped-active networks. For this, the port parameter formulation is used. These methods are applied to several practical examples. The first examples are on the design of UHF broad-band amplifiers with both real and complex loads (a slot antenna). The third example is the low noise matching of a slot antenna to a transistor amplifier. The last example is a design of a low-pass distributed-lumped-active filter. Some of the designs have been built and the measurements show close agreement with the calculated results. A survey of stability criteria which were used in the designs is included. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 01, 1970
- Accession Number
- AD0711058
Entities
People
- Mohammad-ebrahim Mokari-bolhassan
Organizations
- University of Illinois Urbana–Champaign