Dynamic Time Step Control Algorithm Enhancements

Analog simulators typically use a dynamic, adaptive time step control strategy in order to minimize the calculations performed during inactive portions of a system’s operation, while maintaining adequate time resolution during active portions. Without this dynamic control of time steps, one could generate too few time steps to accurately render fast-edged events, or spend a large percentage of CPU time calculating redundant values. These Dynamic Time Step (DTS) algorithms are often overzealous in their relaxation of time steps, and thus can render waveforms inadequately or inaccurately. Thus, the user can make poor decisions because of the incorrectly presented waveforms. Some simulators provide useradjustable control parameters for the DTS. The Saber simulator, with four primary DTS control parameters, allows users to “dial in” just about any results they want. Which one is correct is often debatable. The techniques presented here allow the user to override the built-in DTS in order to obtain trustworthy results.