Low-Frequency EMI Prediction by Electromagnetic FEA and Machine Learning in Consumer Electronics Devices with PCBs

Rapid technology advancements in recent years have resulted in increasing functionality, increasing complexity and reduced footprint of consumer electronics devices. Flex Printed Circuit Boards (flex PCBs) have gained popularity as they enable smaller device footprints by creating additional space within the device to accommodate component modules such as permanent magnets (PMs), sensors, cameras, chargers, batteries, etc. With tightly packed flex PCBs and component modules, Electromagnetic Interference (EMI) has become a critical design consideration that must be carefully analyzed and mitigated to prevent significant impacts on device performance and efficiency. This paper presents an electromagnetic FEA solution assisted by machine learning (ML) for low-frequency (LF) EMI prediction in consumer electronics devices incorporating rigid/flex PCBs. Comprehensive analyses of device operating modes, magnetic shielding effects, and component placements are conducted using a wireless charging example with flex PCBs and a general-purpose rigid PCB to study their influence on EMI. Simulation results demonstrate the effectiveness of magnetic shielding and optimized component placement in reducing LF EMI.