Event-Based Modeling of Rapid Single-Flux Quantum Basic Cells With Timing Jitter

The realizable integration level of Rapid Single Flux Quantum (RSFQ) circuits has reached the order of magnitude of 20,000 Josephson junctions, which enables the creation of advanced complex circuits like microprocessors or digital signal processors. During design of those complex circuits behavioral arrangement alone is insufficient, instead the inclusion of statistical spread is required for timing and parameter verification. This is usually done by computation intensive transient simulation on the electrical network level. To overcome the resulting time consumption of transient simulations, hardware description languages (HDL) were used to simulate RSFQ circuit behavior. However, HDL tools are most often developed to describe state or voltage level based electronics, and the description of the pulse driven nature of RSFQ-electronics is not straight-forward. In this work, a new approach based on discrete-event simulations is presented. By this method the pulse-driven characteristics of RSFQ circuits can be directly transferred into a model describing the behavior on the transaction level. Furthermore, a tool has been chosen, which supports a hierarchical design approach as used in integrated circuit design. The realized models of basic RSFQ cells include stochastic timing effects. The approach is demonstrated by modeling a non-trivial cell, and compared against the conventional transient simulation concerning the accuracy of the results and the computation time.