Macromodel construction and verification

Macromodeling is the abstraction of information from a detailed description at a low level to a less detailed description at a higher level, with the premise that during this process the quantities of interest are preserved and accurately represented. The goal of macromodeling is to predict the value of these quantities of interest without having to perform a complete and detailed analysis, especially for large-scale systems. Macromodeling relies on the presence of repetitive structures, such as logic gates, because it assumes that circuits may be partitioned into a relatively small number of similar sub-circuits, which can be macromodeled separately. In this article we discuss circuit macromodeling and describe a procedure for the generation and verification of behavioral macromodels. The methodology is based on the application of dimensional analysis for the simplification of the forms of the macromodel equations. The aspects of the construction of macromodels, such as the setup of the experiments, the implementation, and the verification of the macromodel, are also discussed. We will show dimensional analysis, circuit simulation, and the Monte Carlo method to be indispensable tools in the generation and the validation of behavioral macromodels.