The Predictive Power Of Engineering Models For Printed Circuit Boards

Printed circuit board (PCB) manufacturers are boosting their investment in inspection, test and analytics to meet the increasingly stringent demands for reliability. The result is a growing demand for predictive design techniques that enable engineers to hone manufacturing performance and reconfigure operations to be more efficient and reduce waste.

Predictive design is about analyzing current and historical data and generating a model to help predict future outcomes.

Now, a new predictive design tool, developed by Draper, has been tested and validated to show it can predict material failure rates and help determine design requirements for printed circuit boards and similar products. In tests, engineers found that the tool accurately predicted methods to reduce and eliminate laminate crack initiation and propagation on a printed circuit board.

The study, prepared for IPC APEX EXPO 2022, was conducted by designing printed circuit boards, using different configurations and materials, and subjecting the boards to environmental stresses and other tests. Using principles of predictive design, engineers developed data sets to guide them. One data set included tests of the PCB materials for such attributes as fracture toughness and thermal conductivity. Another set included detailed design characteristics of the PCB itself.

With the data, the engineers developed a virtual test bed, called a finite element analysis (FEA) computer-aided model, and used it to run scenarios of various PCBs. The FEA model successfully predicted a 50 percent decrease in internal stress that would reduce or completely eliminate PCB laminate cracks.

Just by tweaking the kind of materials and design layouts of a printed circuit board, the team was able to reduce the number, kinds and severity of defects in the board when very minor design changes are made. In their quest to develop a new model, the team, led by Wade Goldman of Draper, capitalized on material characterization testing data and detailed design characteristics of the PCB itself to build a robust model.

“The FEA model allows us to move attributes around to reduce the number and likelihood of cracks in the PCB,” Goldman says. “If each design works as expected, PCB designers and manufacturers are no longer limited to manufacturing by trial and error. Instead, they can spend their time evaluating design changes that might be useful in order to reduce defects and not change their processes.”

The model arrives at an opportune time for the technology industry, Goldman adds. “These days it’s all about designing higher density PCBs. As a result, interconnects are becoming smaller, which introduces fragility, and the industry is making more material choices to make higher density interconnects work. You need a predictive design tool to support that effort.”

The engineers stated the predictive design tool they developed and implemented in this study has been validated and is expected to provide PCB designers with a model for determining design rules for future products. Goldman says, “With this new capability, we have taken a big step in being able to manufacture higher quality printed circuit boards.”

Goldman and his Draper colleagues Hailey Jordan and Curtis Leonard received an honorable mention at the IPC APEX EXPO 2021 for a paper on a similar topic, “Analyzing Printed Circuit Board Voiding and other Anomalies when Requirements Covering the Anomalies are Vague.”