Surface Mount International Conference Proceedings


CONCURRENT ENGINEERING IN HIGH 1/0 BGA IMPLEMENTATION

Author: I.A. Sterian
Company: Celestica
Date Published: 4/28/1997   Conference: Surface Mount International


Abstract: The need for faster product introduction has increased the risk of introducing products and processes that are not fully developed. Design for manufacturability can help with processes that are currently in practice; areas where there are no design rules or prior experience remain risk areas. One way to reduce the risk is to have a product Spetilc test vehicle. Generic test vehicles are designed to test one component, such as one BGA fully wired and daisy chained, or one process, such as paste-in-hole for PIH components. Electrical product test vehicles are designed to test the function of the product to make sure the electrical properties are met. Product specific test vehicles are meant to mimick the electrical properties of the end product, and incorporate the variety of components that will be used in the end-product. This paper addresses the lessons learned in introducing a high i/o BGA family: 272, 313, 503, and 596 i/o, into the manufacturing process. A test vehicle was used to mimick the actual end-product, and thereby reduce risk of process issues arising during critical prototype stages. There were lessons learned in board design assembly, rework heatsink attach and reliability testing. The net result of the test vehicle effort was to implement a high i/o BGA process at low cost, and without impact to a tight prototype schedule. The current ongoing work involves implementation of these high i/o packages in volume production, while maintaining the O-5 PPM/join defect rate seen in prototypes[l][2]. From a design point of view, the test vehicle had to be the same size and thickness as the ultimate product, and had to include peripherals for the BGA such as SMT connectors for logic analyzer access, and fiducials for x-ray inspection of the BGA. Also, the spacing of the components was varied to push the limit of BGA to BGA spacing and BGA to passive spacing. From an assembly and rework point of view, the test components included cavity up overmolded packages up to 40 x 40 mm in size and a SMT EMI fence that had to be attached around a BGA component. Rework included water soluble flux, no-clean flux, and no-clean paste dispense. Rework was forced at predetermined sites up to two times at the same location. From a reliability point of view, the test plan that was followed included accelerated thermal cycling, torque testing, shear testing, cross-sectioning, and thermal ship/shock. Based on the test results, all BGA components, except one supplier were qualified, and a preferred rework method and heatsink attach method were selected. The final assembled test vehicle is shown in Appendix 1. The Cisco/Celestica team identified what new technologies were to be used in a new generation of products, and decided which of those technologies need to be incorporated in a test vehicle prior to actual product prototype builds. A test plan was written and approved prior to any process development or design work starting. The test plan identities the tests to be done and what constitutes a pass. The design, assembly, and test of this vehicle was completed in 6 months, with follow-on work continuing in the next 6 months. This paper will discuss the decisions made in technology assessment test plan selection, design features, and results of the testing.



Members download articles for free:

Not a member yet?

What else do you get when you join SMTA? Read about all of the benefits that go along with membership.

Notice: Sharing of articles is restricted to just your immediate work group. Downloaded papers should not be stored on an external network or shared on the internet.


Back


SMTA Headquarters
6600 City West Parkway, Suite 300
Eden Prairie, MN 55344 USA

Phone 952.920.7682
Fax 952.926.1819