Wire Bonds Defect Inspection
Distinguish defective from tolerable anomalies for improved IC chip yield and performance
Graphical programming environment for deep learning-based industrial image analysis
Wire bonding is the most common method used for interconnection inside many integrated circuits and microchips. It is a delicate process that requires high accuracy. The purpose of wire bonding is to connect the leads on the chip to the packaging material with very thin metal wires. The packaging material transmits signals to other components. Defects like broken or missing wires can disrupt signal transmission. These defects can be varied in type and location, which makes it hard for rule-based machine vision solutions to accurately determine a defective wire bond.
Traditionally, using an Automated Optical Inspection (AOI) system with rule-based vision does not work well. So suspected No Good (NG) cases are inspected by deep learning to enhance the reliability of the inspection process The AOI machine picks the suspected NG cases and feeds the images to a system using Cognex Deep Learning tools. The defect detection tool dynamically extracts the region of interest, and the classification tool categorizes the various defects, distinguishing defective wire bonds from acceptable ones. Sorting the defects helps isolate problems in the process to prevent costly rework further down the line, while successfully identifying micron-level defects improves IC chip yield and longevity of performance.