Design007 Magazine


Issue link:

Contents of this Issue


Page 23 of 73

24 The PCB Design Magazine • May 2016 I-Connect007 Editor Andy Shaughnessy reported that the keynote speaker at the IPC APEX EXPO Design Forum was Dale Parker, a former PCB designer at Shure who is now a CAD manager at Google X. Parker is involved in the development of autonomous vehicles and all sorts of other great ideas at Google X. Ac- cording to Andy, Parker told the crowd, among other things, that EDA tool vendors need to trash all their old 1990s code and start over, this time with artificial intelligence. There has been a lot of activity in the field of AI recently, with such developments as voice recognition, unmanned autonomous vehicles and data mining to list a few. But how could AI possibly influence the PCB design process? This month, I will take a look at the endless possibili- ties. So much time is wasted on reproducing the same thing over and over again on each layout. Current EDA tools, with all their bells and whis- tles, are still very limited in automation process- es and mostly rely on the skills and foresight of the engineer and PCB designer to drive the software through all the hoops. Instead, EDA tools need to predict what the designer is try- ing to do, then look at previous designs to sug- gest alternatives and auto-complete the design where possible. AI is a system that perceives its environment and takes actions to maximize its chances of success. Automating many of the tedious steps in setting up the initial database would be a good start. A standard form factor could be used to establish the initial layout environment ensur- ing that designs are compatible across multiple generations of technology. Although some PCB layout tools allow the designer to load a stan- dard set of predefined startup configuration files, there is still too much manual interven- tion required. The PCB database could predict the fundamental design rules and via stack re- quirements sourced from previous experience. Predictive text, which we all use every day on our cell phones, could provide self-evident nam- ing conventions for supplier part numbers and database fields, greatly speeding up the design definition. Busses and interfaces could be analyzed and categorized with naming conventions inter- preted from the chip pin name assignments, eliminating much of the monotonous schematic cap- ture process. IC power pins could have powers supplies assigned based on datasheet requirements. And a starter set of decoupling ca- pacitors, added to each power pin, could kick off the PDN analysis by Barry Olney IN-CIRCUIT DESIGN PTY LTD / AUSTRALIA The Case for Artificial Intelligence in EDA Tools FEATURE COLUMN: BEYOND DESIGN Figure 1: Artificial intelligence is part of the future of EDA tools.

Articles in this issue

Archives of this issue

view archives of Design007 Magazine - PCBD-May2016