Specifications

Solder and Lead Inspection
Introduction
Statistically solder defects are the most numerous amongst all the PCB
manufacturing defects. A few common solder defects are insufficient/no
solder, solder bridge between leads, excessive and improper wetting. Due to
the fact that variation amongst good solder joints are substantial, solder
inspection presents a great challenge to AOI machines. Moreover,
determining a solder defect is somewhat subjective.
YesTek AOI machines use a variety of vision algorithm and illumination
arrangements to perform solder inspections. These include traditional rule
base algorithms, moreover, software architecture is provided to allow easy
extension of the inspection algorithm by linking in extension DLLs (dynamic
link library) at Test Board-time. The DLL can add new vision algorithms and
lighting control to the core software to satisfy special inspection requirements.
Algorithms for Solder Inspection
Pattern Recognition
Pattern Recognition, also known as template matching, is a technique the
software uses to do device marking inspection. This technique can also be
used to a limited extend on solder inspection. Pattern recognition uses a
statistical formula to determine the degree of similarity between the test image
and the template (or templates) trained on known good solder joints. The
comparison produces a score between 0 to 1000 with a score of 1000
meaning a perfect match. For solder inspection a score of 600 is commonly
used.
Histogram Analysis
Histogram analysis compares the ratio of bright/dark pixels to the total number
of pixels. Based on the lighting arrangement, defective solder joints usually
yield a different ratio than the good solder joints. Histogram analysis is a
simple and fast technique. Currently it is the most commonly used algorithm
for solder inspection.
Algorithms for lead (solder bridge) Inspection
Blob Analysis
An adaptive threshold is used to binarize the image, and then the black and
white pattern is checked against the sector’s parameter such as pitch and
lead width etc. Any discrepancy in dimension or regularity will flag an error.