Improving a soldering process applying the dual response approach to a Taguchi's orthogonal array

Abstract

Taguchi methodology for robust parameter design has generated both extraordinary interest due to numerous successful applications and much debate due to some limitations such as shortcomings to measure variability with the signal to noise ratio. In Response Surface Methodology the dual response surface approach has been successfully applied as an alternative to Taguchi's approach. The dual response surface approach uses separate regression models to describe the behavior of the mean and the variance of the response. This article presents an application of Taguchi's robust parameter design to reduce the number of defectives in a solder process. The Taguchi methodology was used to find the appropriate settings for the robust parameters or factors. Then, looking to overcome the possible problems of the signal to noise ratio and trying to obtain a better solution, the dual response surface approach was applied to the same Taguchi's crossed array.