Authored By:
David Geiger, Vincent Nguyen, Hung Le, Stephen Chen, Robert Pennings, Christian Biederman, Zhen (Jane) Feng, Ph. D., Alan Chau, Weifeng Liu, Ph. D., William Uy, Anwar Mohammed, Mike Doiron
Flex International, Milpitas, CA,USA
Summary
Automated Optical Inspection (AOI) is advantageous in that it enables defects to be detected early in the manufacturing process, reducing the Cost of Repair as the AOI systems identify the specific components that are failing removing the need for any additional test troubleshooting1-3. Because of this, more Electronic Contract Manufacturing Services (EMS) companies are implementing AOI into their SMT lines to minimize repair costs and maintain good process and product quality, especially for new component types. This project focuses on the testing of component package 03015 which is challenging for AOI.
Highly-automated and effective test methods are becoming a more and more important topic in our industry today. Advances in modern manufacturing technologies have been making factories smarter, safer, and also more environmentally sustainable. Finding and implementing smart machines which provide real time information is critical to success. Currently we have been successful in using 2D/3D AOI for production; however not for the upcoming 03015 components. Therefore, we are working with AOI vendors to ensure successful testing of this component type, with a special emphasis on optimizing algorithm threshold settings to detect defects.
We have been working with five AOI vendors with 5 test vehicles (PCBAs). Each PCBA board has 246 components with three different pitch sizes (100microns, 150microns, 200microns). The results of Attribute GR&R, Defect escapes, and False Call PPM (parts per million) will be presented.
Based on the data which we received up to now, every set of data (5 sets - still waiting for results of AOI system 3) is from the algorithms of 2D AOI although some machines have the3D AOI capability. These machines have shown different levels of performance. AOI system 5's results have an excellent acceptable level for Attribute GR&R; both AOI system 5 and AOI system 6 have only several percentage points of a Defect Escape rate. However, this study is just in its infancy; more improvement and testing will be performed. We will continue to provide new test results from all suppliers.
Conclusions
Current AOI machines have different levels to test 03015 components; however, all AOI machines involved for this project used 2D AOI function;3D algorithms were not usable due to component reflection.
Attribute Gage R&R results are acceptable for these five machines, AOI System 5 had excellent results.
Based on the data which we have now, AOI System 5 has the best performance for Defect Escape %; however, no machine had False Call PPM (parts per million) < 5000.
This study is just the beginning. More boards (with no reflection) are needed to test with the AOI machines, especially when using 3D Algorithms.
More improvements to the machines are coming from several of the AOI System R&D teams.
|
