OEM and EMS to Combat Head on Pillowing Defects

Research

Latest Industry News

First Major Attempts to Regulate AI Face Headwinds from All Sides

Foxconn confirms rotating CEO system in place since April

Apple planning US$250 million expansion in Singapore to write 'new chapter' in company's history in city state

Samsung & TSMC Bolster U.S. Advanced Chip-Making

Untether Collaborates With Arm For Automotive AI

MORE INDUSTRY NEWS

OEM and EMS to Combat Head on Pillowing Defects

This paper evaluates AXI results from different machine platforms and results from similar platforms operating at different facilities.
Analysis Lab

DOWNLOAD

Authored By:

Alex Chan, Paul Brown
Alcatel-Lucent, Ottawa, ON, Canada

Lars Bruno, Anne-Kathrine Knoph
Ericsson, Katrineholm, Sweden

Thilo Sack
Celestica Inc., Toronto, ON, Canada

David Geiger, David Mendez
Flextronics International, Milpitas, CA, Austin, TX, USA

Mulugeta Abtew, Iulia Muntele
Sanmina Corporation, San Jose, CA, Huntsville, AL, USA

Michael Meilunas
Universal Instruments Corporation, Binghamton, NY, USA

Summary

It is an ongoing battle in the OEM and EMS world to eliminate Head on Pillow (HoP) defects in every day assembly activities. Recent effort in the industry has been primarily focused on defect mitigation, while limited investment has been made to understand the true capability of HoP detection using Automatic X-ray Inspection (AXI).

For this reason, the study presented here focused on evaluating the HoP detection capability of four different AXI platforms using assemblies with known HoP defects.

Conclusions

HoP is not an easy defect to detect in production. The industry's perception of the problem has come a long way from believing that HoP is an isolated phenomenon to now understanding and accepting that HoP is a much more prevalent defect in BGA soldering in the SMT process. In this collaborative study between the OEM and EMS, the overall results are very encouraging. It suggests that with focused engineering effort between the EMS and the AXI suppliers, a reliable HoP detection capability is achievable.

This project was able to quantify the effectiveness of the HoP detection through a lengthy experiment that culminated with a tedious cross-sectioning effort. The results suggested that some machines can potentially catch 100% of the HoP defect while other less effective machines are still capable of catching 90%+ of the HoP defect. Depending on the machine type used, adding HoP inspection to production may potentially result in a small increase in AXI inspection time.

However, with the quantifiable results demonstrated in this study, such an increase is easily justifiable on the grounds that it will help to avoid 90%+ of the HoP defect escapes. By extension, implementation of AXI for HoP detection in volume production on high risk BGAs is now a common practice in some of the OEM and EMS involved in the project.

Despite the very promising results that have come from this project, several other aspects require further study and follow-on investigation to improve HoP detection through the AXI process.

Algorithm fine tuning. Knowing the results of the cards in Group B, the sites involved in the study can further improve the algorithms and threshold settings through re-inspection of the Group A cards that have not been
subjected to destructive analysis.
More work is needed to understand the effect that BGA joint size, pitch and shape may have on different thresholds in the algorithm settings.
Identifying BGAs that are at higher risk of HoP was not part of this study but it is a key to successful implementation of HoP detection with AXI. Component warpage and its relationship to HoP defects will be discussed in Part 2 of this collaboration effort between the OEM and EMS.
Finding a known HoP defect card is not an easy process -especially without destructive analysis. A more reliable laboratory type of tool is needed especially when HoP is not on the outside row of the BGA. Once a known HoP BGA is found, using such a card to fine tune the AXI will lead to a much more successful HoP detection of that same BGA type.
Much of the work that has been done previously has been focused on mitigation steps to minimize HoP, rather then addressing the root cause. The industry will need to focus more efforts on how to address the root cause of HoP (as will be discussed in the Part 2 of this paper).

Initially Published in the SMTA Proceedings