Distribution Optimization of Through-Silicon Via (TSV) Array Based on Genetic Algorithm

• Published in: IEEE transactions on components, packaging, and manufacturing technology (2011) Vol. 15; no. 2; pp. 399 - 409
• Main Authors: Ju, Jia-Yi, Qiang Liu, Qi, Zhang, Peng, Wang, Jing, Zhao, Peng, Lin, Xuan, Yao, Chen-Yang, Zhao, Wen-Sheng
• Format: Journal Article
• Language: English
• Published: Piscataway IEEE 01.02.2025; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Arrays; Co-simulation; Crosstalk; Encoding; Finite element analysis; genetic algorithm (GA); Genetic algorithms; Integrated circuits; Interconnections; Mathematical models; MATLAB; Optimization; Packaging; Scattering parameters; Signal integrity; Simulation; through-silicon via (TSV) array; Through-silicon vias
• ISSN: 2156-3950; 2156-3985
• DOI: 10.1109/TCPMT.2025.3526829

Through-silicon via (TSV) has attracted much attention due to its numerous advantages, including high-density interconnection, low power consumption, and wide bandwidth. When signals are transmitted in TSV array, however, significant crosstalk issues may arise, negatively affecting timing and signal integrity. In this article, the distributions of signal and ground TSVs in an array are optimized by conducting co-simulation of ANSYS Q2D and MATLAB based on genetic algorithm (GA). The optimization aims at reducing the crosstalk in the array. Compared with the reference distribution scheme, the results indicate that the optimized distribution of TSV array can effectively reduce both near-end crosstalk (NEXT) and far-end crosstalk (FEXT). The proposed method is of great significance in solving the crosstalk problem, offering feasible solutions for the design of TSV array.