Signal Integrity Optimization for C-PHY Channel Using Surrogate Model of Tab-Routing Structure

• Published in: IEEE transactions on electromagnetic compatibility 2024-10, p.1-9
• Main Authors: Cheng, Yu-Ying, Wu, Tzong-Lin
• Format: Article
• Language: English
• Subjects: Couplings; Crosstalk; Degradation; Mobile industry processor interface (MIPI) C-PHY; neural network (NN); Optimization; particle swarm optimization (PSO); Predictive models; Printed circuits; Scattering parameters; Signal integrity; signal integrity (SI); surrogate model; tab-routing structure; Voltage; Wires
• ISSN: 0018-9375; 1558-187X
• DOI: 10.1109/TEMC.2024.3476489

This article presents the first comprehensive investigation into the crosstalk mechanism within a three-wire (four-conductor) C-PHY transmission channel based on mixed-mode theory ( X , Y , and C modes). The phase difference between X and Y modes is identified as a primary contributor to crosstalk, leading to signal integrity (SI) degradation. A tab-routing design is first specifically applied to enhance SI in three-wire (four-conductor) C-PHY channels. Additionally, an artificial neural network (ANN) based surrogate model is developed to map tab-routing parameters to eye-opening metrics efficiently. By combining the particle swarm optimization (PSO) algorithm with the ANN-based surrogate model, optimal geometrical parameters for the tab-routing C-PHY channel with enhanced SI performance can be quickly determined. The optimized three-wire tab-routing C-PHY channel, fabricated on a two-layer printed circuit board (PCB), demonstrates a 17.2% improvement in eye-opening and an 8.5% reduction in the occupied area compared to a typical 50 Ω three-wire channel. This article also represents the first application of machine learning (ANN, PSO) to C-PHY SI research, significantly improving design process efficiency. The feasibility and accuracy of the ANN-based surrogate model applied to the tab-routing C-PHY channel are thoroughly validated.