Do You Trust Your Simulator?

Admittedly, I’m a bit biased. After 20+ years in the EDA industry, I’ve developed increasing confidence in modern simulation technology. In short, I generally believe what SystemVision tells me about the systems I analyze. With the right models and simulation settings, I am confident my simulation results will closely match lab measurements on prototype hardware.

Part of building confidence is simply practice – just using the simulator and its models to get to know its strengths, weaknesses, and yes, workarounds. All simulators have several of each. I have unlimited access to any Mentor Graphics tool I want to use, so practice sessions can start any time. Not bragging, just stating a nice perk of working in the EDA industry. I regularly sit down at the keyboard to put SystemVision through its paces. We have a large number of demonstration and example circuits designed to illustrate many of SystemVision’s features and benefits. But what about engineers who do not have the luxury of unlimited tool access and may only use a simulator once in awhile? Confidence in simulation is often inversely proportional to industry tenure.

Some mechatronic engineers, particularly those with more than a handful of years in their industry, tend to question what a simulator and its models tell them about their system. More often than not, they would rather take a prototype to the lab and make measurements. New engineers fresh out of school, on the other hand, where simulation and modeling are often an integral part of education, are more inclined to trust simulation. Much of what new engineers learn in school is centered around using a variety of EDA tools to complete their assignments. So which approach, hardware testing or simulation, makes the most sense? The answer isn’t as clear-cut as it may seem. Even with the best simulation tools and detailed, accurate models, the most successful design flows use a combination of simulation and hardware test. But with better design tools and more accurate models, the transition point can shift towards doing more simulation and less bench testing. The key can be summed up in one simple concept: do your due diligence. Yep, validate both your models and your simulator. You need to get comfortable with what simulation and modeling can do for you and your design. And the more comfortable you are with simulation and modeling, the more you will trust it to do the bulk of the heavy lifting in your design process. Build confidence in your tools, use them to speed your design process, then verify design operation in the test lab.

I won’t try to tell you how to qualify either your simulator or your models. Qualification processes are as unique as engineers and the companies they work for. And what your company uses to qualify a simulator may not be good enough for you – you might add your own tests to the qualification flow. Since the success of your project can depend on the tools and methods you use, design tool qualification can be a bit of a personal process. Just like any craftsman (you do consider yourself a craftsman, don’t you?), you have to trust your tools.