先进通信技术

Search documents
PCB层数,创新高
半导体行业观察· 2025-05-10 02:53
Core Viewpoint - OKI Circuit Technology has achieved a significant milestone by introducing a 124-layer printed circuit board (PCB), surpassing the previous industry limit of 108 layers without increasing the standard thickness of 7.6 mm, which could revolutionize substrate design in fields such as artificial intelligence, defense, aerospace, and advanced communication technologies [1][4][10] Group 1: Technical Advancements - The increase from 108 layers to 124 layers represents a fundamental shift in PCB manufacturing capabilities, achieving a 15% increase in signal layers without compromising the standard thickness [4][10] - Traditional PCB designs face mechanical and thermal performance limits before reaching 100 layers, and exceeding 108 layers typically requires thicker boards or reduced reliability, which OKI has successfully avoided [4][5] - The use of ultra-thin dielectric materials with a thickness of only 25 µm enables strict impedance control and supports high-power AI chips, crucial for next-generation high-bandwidth memory (HBM) [5][6] Group 2: Applications and Benefits - The 124-layer configuration is expected to open new avenues for semiconductor testing in AI, as it enhances wiring capacity and shielding potential, which is vital for maintaining signal integrity [9] - The PCB's design supports integrated grounding layers and microvia arrays, minimizing crosstalk and signal loss while improving thermal performance, making it ideal for aerospace and defense applications [9] - The PCB meets MIL-STD-883G reliability standards after over 1,000 thermal cycles, ensuring electrical integrity in extreme environments [9] Group 3: Cost and Production Challenges - The material cost for the 124-layer PCB can reach up to $4,800 per square meter, with a production time of 16 weeks and a yield rate around 65%, which is lower than the typical yield of 85% for 108-layer PCBs [10] - Mechanical stress from thermal cycling can exceed 80 MPa, potentially leading to issues in fine-pitch BGA packaging, complicating the diagnostic process [10] - While current applications are limited to niche, high-performance areas, underlying innovations may gradually permeate broader markets as advancements in additive manufacturing and AI-driven EDA tools emerge [10]