Considerations on Image Input System for Printed Wiring Board Pattern Inspection

Abstract

This paper investigates the relationship between an automatic inspection system's capability to detect small defects and its image input system characteristics. Assuming the use of a semiconductor line scanner, three types of image degradation in the image input system are considered : defocusing in the optics system, averaging in the line scanner cell, and blurring due to mechanical translation during scanning. Corresponding image degradation models are developed and simulated for model defects. The calculated capability to detect small defects shows that defects twice the size of pixels can be completely detected at the presence of defocusing where the confusion circle diameter equals the pixel size and blurring due to mechanical translation during scanning. A method is shown that achieves the above focus level without visualizing the scanned image. Finally, measurement of the small defect detection capability of a prototype inspection system for printed wiring board masks is described.