A new approach to network design to obtain optimal parameter values simultaneously with an optimal set of component tolerances is proposed. An automated scheme could start from an arbitrary initial acceptable or unacceptable design and under appropriate restrictions stop at an acceptable design which is optimum in the worst case for the obtained tolerances.
Chawla, B. R., Gummel, H. K. and Kozak, P. 76.64 'An MOS circuit simulator for integrated circuit design' Proc. 1975 ISCAS pp 2 2 8 230 A M O S circuit simulator is described which bridges the gap between logic simulators and circuit analysis programs. Circuits of up to 1000 gates can be simulated with operational speeds of approximately 2 ms of realtime per gate per nanosecond of circuit time. Daae, H. 76.65 'Computer-aided optimal synthesis of bandpass filters' IEEE Trans. Vol CAS-22 No 6 (June 1975) pp 4 9 0 - 4 9 5 A method is developed for the computeraided synthesis of a particular case of linear lumped, passive, bilateral, lossless bandpass ladder filters. It involves the numerical values of the open circuit immittance function in three sets of equations referred to as the terminal, the remainder and the junction algorithms. The process is applied to the example case design of a 70-MHz 14thorder filter. Dutta, S. R. K. and Vidyasagar, M. 76.66 'Optimal design of nonlinear DC transient circuits without solving network equations' IEEE Trans. Vol CAS-22 No 8 (August 1975) pp 661--665 A method is described for the optimization of nonlinear DC circuits. A performance index is defined to measure the difference between the desired and the actual specifications. The novel approach taken here is to treat the network equations as equality constraints. The problem is then optimized using penalty function techniques. The algorithm is easily amenable to package programming.
Ho, C. W., Ruehli, A. E. and Brennan, P.A.
76.67 'The modified nodal approach to network analysis' IEEE Vol CAS-22 No 6 (June 1975) pp 5 0 4 - 5 0 9 The nodal method has been widely used for formulating circuit equations in computeraided network analysis and design programs. However, several limitations exist in this method~ including the inability to process voltage sources and the current dependent circuit elements in a simple and efficient
manner. A modified nodal analysis (MNA) method is proposed here which retains the simplicity and other advantages of the nodal analysis while removing its limitations. A simple and effective pivoting scheme is also given. Ho, C.W. 76.68 'Modified nodal approach to DC network sensitivity computation' IBM J. Res. Develop. Vol 19 No 6 (November 1975) pp 565--574 Programming techniques are presented for computing DC sensitivity vectors of nonlinear electronic circuits. The modified nodal approach is used as the method of formulation for the circuit equations, in which multiple performance objectives can be accommodated. Numerical examples illustrate some of the techniqdes discussed.
Ho, C. W., Ruehli, A. E. Brenna, P. A. and Zein, D . A . 76.69 'Interactive circuit analysis and design using APL' Proc. IEEE 1975 Int. Symp. Circ. and Syst. (ISCAS) pp 2 1 6 - 2 1 9 The modified model approach to network analysis is implemented in APL. The vectormatrix nature of the network equations is exploited by APL programming. As an example, circuits of 300 elements are analysed with 64k-bytes of storage. Kaplan, G. 76.70 'Computer-aided design' IEEE Spectrum Vol 12 No 10 (October 1975) pp 40--47 This paper presents a general overview of CAD applications in electrical circuits. Problems associated with circuit design, optimization, tolerance assignment, circuit, yield and sensitivity calculations are discussed. A table listing some general information about the various currently available major computer-aided circuit design programs is also presented. Lee, A . Y . 76.71 'Signal flow graphs - computer-aided system analysis and system calculations' IEEE Trans. Vol CAS-21 No 2 (March 1974) pp 209--216 Concepts which promise to extend many fundamental results of network theory to the general systems are introduced. Thus, system gain as well as sensitivity calculations can be formulated for those systems which can be represented by signal flow graphs.
are examined and an estimate of computational costs for each set is made. Liebert, T. A. and O'Niell, L . A . 76.73 'STATE - a program for the statistical evaluation of transmission systems' Proc. IEEE 1975 ISCAS pp 224--227 STATE is a Monte Carlo analysis program used to evaluate and optimize performance of 2-wire and 4-wire transmission systems whose requirements are stated statistically. In addition to the built-in capabilities, users can, if necessary, supply FORTRAN subroutines for equipment characterization, performance measure calculations and control of program flow.
Masden, K., Schjaer-Jacobsen, H. and Voldby, J. 76.74 'Automated minimax design of networks' IEEE Trans. Vol CAS-22 No 10 (October 1975) pp 7 9 1 - 7 9 6 A gradient algorithm for the solution of nonlinear minimax problems has been developed. This simple-to-use algorithm is well-suited for automated minimax design of networks and compares favourably with recent minimax and least pth algorithms. Proposed algorithm is used to design an equalization network for a reflection type microwave amplifier. Resai-Fakhr, M. G. and Temes, G.C. 76.75 'Node elimination in linear active circuits' Electron. Lett. Vol 11 N o 6 (March t975) pp 1 2 1 - 1 2 2 A method is described for the elimination of nodes in a linear circuit containing dependent and independent sources. The connection between this operation and a cycle of Gaussian elimination in the nodal analysis of the circuit is also established. Sobhy, M. I. and Dief, A . S . 76.76 'Application of the perturbation matrix in network design problem' Electron. Lett. Vol 11 No 14 (July 1975) pp 2 9 4 - 2 9 6 The concept of the perturbation matrix is used to develop a set of simultaneous equations relating the shifts in eigenvalues of the state matrix to variations in either the resistive or reactive elements of a network, or to the loss factor of reactive elements. A powerful and versatile computer program based on these equations has been developed and applied to a wide range of network design problems.
Wing, O. and Huang, J. Leung, KoH. and Spence, R. 76.72 'Multiparameter large-change sensitivity analysis and systematic exploration' IEEE Trans. Vol CAS-22 No 10 (October 1975) pp 796--804 The effect on linear circuit response of simultaneous changes in a number of components must often be computed. Two efficient methods for prediciting this effect
76.77 'SCAP - a sparse matrix circuit analysis program' Proc. 1975 ISCAS pp 2 1 3 - 2 1 5 SCAP is a circuit analysis program based on a sparse matrix technique to compute the frequency response of a circuit consisting of lumped and transmission line elements. Special data-structure techniques are used to minimize both the arithmetic and nonnumeric operations in the solution process.
COMPUTER AIDED DESIGN