Generalizations About Component Flatness

Bev Christian, Linda Galvis, Rick Shelley and Matthew Anthony
BlackBerry
Cambridge, Ontario, CANADA

An examination was made of a database of collected component warpage information in the hopes of finding a simple correlation between a physical attribute of the components and the warpage measured at reflow temperatures. This would allow companies to by-pass actual experimentation, at least for a first pass approximation of the expected warpage.

The measured flatness information was compared to various component outer and inner dimensions. In specific instances some correlations have been discerned, but no over-arching generalization can be made.

In this paper an examination has been made of over 250 different components that have been examined using shadow moie. There is no one attribute that provides a perfectly linear, exponential or polynomial correlation with the warpage observed for all plastic encapsulated components examined.

For various subsets of the components tested, correlations have been possible with respect to component thickness, substrate thickness, total volume of the ICs in the component or IC die thickness when there was only one die. There were factors that could not be taken into consideration because of lack of data.

Some of these were: packaging house or houses used by each component manufacturer, particular epoxy used for each component and curing conditions for each epoxy. It is expected that the type of epoxy would be the most important of these three mentioned.

Certainly component thickness and even more so IC volume are important considerations when examining for component flatness. But, as the scrutiny given to the data presented here shows, neither dimension tells the whole story regarding the warpage of components.