An important performance metric for any system is reliability — does the system do what it’s supposed to do, when it’s supposed to do it, time after time, without fault or failure. Obviously, system reliability doesn’t just happen; it’s not just dumb luck when a system delivers flawless performance, whether it’s your car, your cell phone, your garage door opener, or your home computer.
System reliability is designed-in. It’s not a function of test, nor is it a result of customer “burn-in” (failures found and reported at the user level, then fixed in the next generation of the system). Reliability must be a key consideration from the beginning of the design process. And designing-in reliability, particularly for systems of even moderate complexity, means simulation.
With the right simulation and modeling choices, you can analyze your system’s performance metrics as you change design parameters and operating conditions. Simulation, though once an optional design flow activity, is now required for all but the most basic multi-domain designs. In addition to helping you refine your nominal design, simulation lets you run system experiments without the expense and overhead of building and testing hardware prototypes. With its flexibility and efficiency, this virtual prototyping methodology trumps hardware prototyping in most system design flows.
Defining and executing simulation experiments is often the most direct way to verify design performance under a variety of system conditions. The more experiments you run, the better your design coverage and reliability; and the more you use automated methods to setup your experiments, the more experiments you can run. Simple math; no rocket science required.
SystemVision Experiment Manager helps you define system experiments, and then automatically executes your experiments via a direct link to SystemVision. Awhile back I posted a three-part blog (Part 1, Part 2, Part 3) introducing Experiment Manager basics using a simple motor drive circuit. What I didn’t mention, however, was using MiniTab® Statistical Software to help automate experiment definition and subsequent data analysis.
Minitab is a comprehensive statistical and graphical analysis program often used by companies in their Six Sigma and related quality improvement programs. We recently posted a new video illustrating how to use Minitab with Experiment Manager to define and run a series of experiments on a basic RLC circuit, and then analyze the simulation results. The video is just five-minutes long and you’ll have an opportunity to download an information packed white paper: Model-Driven Design for Six Sigma.