Signal Integrity Analysis Based On Model Order Reduction

Deep submicron IC design will bring many new challenges. Under such background, this paper first introduces the principles of model order reduction and the significance of signal integrity analysis. Several classical model order reducion algorithms are introduced, and their advantages and disadvantages are compared. In this paper, a new algorithm based on observability canonical form is proposed. It has a simple structure and is easy to check the stability of reduced model. In the third part of this paper, model order reductions are performed on several interconnect models using various kinds of algorithms. The results prove that model order reduction techniques can obtain a much smaller model than the original circuit without losing much accuracy. Signal integrity analysis based on this technique is feasible and effective.