基于GMR传感器的国产伏羲芯片表面缺陷检测系统设计

Design of a Surface Defect Detection System for Domestic Fuxi Chips Based on GMR Sensors

摘要:

国产伏羲芯片组成结构较为复杂，为了加强对国产伏羲芯片的安全保护，确保芯片品质符合设计规格和标准，为此设计并开发了基于GMR传感器的国产伏羲芯片表面缺陷检测系统。加设GMR传感器和检波器，改装核心处理器、信号放大器、数据存储器和图像采集模块，完成硬件部分优化。利用GMR传感器获取伏羲芯片表面磁场反馈信号，拼接国产伏羲芯片表面图像数据，通过信号补偿、标准化以及图像增强等步骤，从磁场信号和图像数据两个方面，提取芯片表面特征，通过特征匹配确定当前芯片表面的缺陷状态，得出缺陷类型、位置以及面积的检测结果。实验结果表明:该系统的缺陷类型误检率降低了2.9%，缺陷位置检测误差减小约2.5mm，同时表面缺陷面积检测误差明显降低。

Abstract:

The composition and structure of domestic Fuxi chips are relatively complex. In order to strengthen the safety protection of domestic Fuxi chips and ensure that the chip quality meets the design specifications and standards, a surface defect detection system for domestic Fuxi chips based on GMR sensors has been designed and developed. Add GMR sensors and detectors, modify the core processor, signal amplifier, data memory, and image acquisition module, and complete hardware optimization. Utilizing GMR sensors to obtain the surface magnetic field feedback signal of Fuxi chips, and collecting surface image data of domestically produced Fuxi chips, the chip surface features are extracted from the magnetic field signal and image data through signal compensation, standardization, and image enhancement steps. The current defect status on the chip surface is determined through feature matching, and the detection results of defect type, location, and area are obtained. The experimental results show that the defect type misdetection rate of the system has been reduced by 2.9%, the defect position detection error has been reduced by about 2.5mm, and the surface defect area detection error has been significantly reduced.