Authored By:
Annaka Balch
Thayer School of Engineering
Hanover, NH, USA
Ronald C. Lasky, Ph.D., P.E.
Indium Corporation
NY, USA
Summary
This project compares past board assembly roadmaps with actual technological outcomes. Its conclusions are mixed: some aspects that the roadmaps covered were very accurate, while others could use improvement. This paper also draws general conclusions on the outline and readability of the board assembly roadmaps. These roadmaps were given to Dr. Lasky and me at no cost from Marc Benowitz, the CEO of iNEMI, for the purpose of this project.
This paper examined the progression of predictions across seven significant aspects of board assembly covered in the 1994, 2002, 2007, 2013 and 2017 roadmaps: 1) Conversion Costs, 2) NPI Cycle Time, 3) Component Trends, 4) Solder Paste, 5) Bar Solder, 6) Wave Solder Flux and 7) Die Attach Adhesives.
Conversion costs were quantified across the 1994, 2002 and 2007 roadmaps and were found to be accurate, if not conservatively estimated (see Figure 5). Even the estimate in the 1994 Roadmap for 15 years out was within 0.05 cents of the actual technological outcome per I/O. NPI predictions were found to be extremely accurate quantitatively as well as qualitatively. The area with the most discrepancy between the roadmaps’ predictions and actual technological outcomes is in component trends. Maximum I/O density, minimum pitch for area array packages and chip speed placement were all overestimated markedly, especially in the earlier roadmaps.
It should be noted that there are discrepancies between these roadmaps, but this project aims to bridge these discrepancies in a comprehensive fashion to better inform iNEMI for future roadmaps.
Conclusions
iNEMI has made many projections over the last century, a majority of them very accurate. With all the advancements in new legislation, many would expect otherwise. Conversion costs and NPI cycle time were all estimated very accurately, even 15 years out in 1995. The area with the most discrepancies was in component trends, where the 1994 and 2002 roadmaps largely overestimated component density capabilities.
While this project makes conclusions about the evolution of predictions since 1994, it should be noted that many of the metrics were complex to compare. The roadmaps contrast greatly in their general outline and what metrics are used. Some roadmaps discuss metrics qualitatively and some roadmaps discuss metrics quantitatively.
This project aimed to bridge these discrepancies into a comprehensive reflection on the roadmap predictions versus actual technological outcomes.
|
