Twenty-eight years ago this month, I walked through the doors of my first employer in the Electronic Design Automation (EDA) industry, sat down at my first UNIX-based workstation, and started work as an applications engineer in the simulation and modeling space.
Much has changed in the analog and mixed-signal simulation world since that first day, but it has been a good ride, and I look forward to seeing how the simulation market changes in years to come. During my time in the EDA industry, I have supported engineers in a broad range of applications, from designing ICs on the small end of the product scale, to developing systems for cars, planes, and rocket ships on the much bigger end.
I started out working in the systems end of the simulation pool, where language-based modeling was more the rule of the day. Think physics-based modeling of mixed-technology devices. Recently, I have worked more in the PCB simulation space, which is most commonly the realm of SPICE modeling.
SPICE is a capable model format, but the more interesting and complex technology gets, the harder SPICE has to work to get the modeling job done. Creating complex models with SPICE is truly an art — in my mind almost a genetic talent gifted to a select few.
Language-based modeling, on the other hand, offers flexibility that SPICE cannot and is more straightforward in its use and application. The challenge with modeling languages, however, is they are often proprietary. Using the language from Company A locks you into using Company A’s tools, even if Company B’s tools better support the capabilities you need.
The Institute of Electrical and Electronics Engineers (IEEE) met this challenge with the development of the industry-standard VHDL-AMS hardware description language. Models developed using VHDL-AMS can be used in any supporting simulator.
I recently wrote a white paper titled “VHDL-AMS: What It Is and Why It Matters for PCB Circuit Design” that talks a bit about SPICE and VHDL-AMS, and gives a brief introduction into VHDL-AMS language basics. If your current PCB circuit modeling methods are bouncing up against limits that slow your simulation and analysis work, you might find the paper interesting and language-based modeling worth a look.