Efficient and practical experimental designs for high volume manufacturing

An experimental design technique that permits process optimization studies to extend beyond process average response curves and that optimizes with respect to primary (main) variables and secondary (tolerated) variables is described. The method modifies W.J. Diamond's (1981) approach for the inner array and modifies G. Taguchi's (1986) approach for a high-volume manufacturing environment and is experimentally efficient. It is capable of estimating average effects of primary and secondary variables, effects of primary variable process variances, and the process average/variance interrelationship. In using advanced experimental design techniques, three major goals have been achieved; (1) increased manufacturability, (2) improved quality, as shown by a reduction in outgoing DPPM levels, and (3) streamlined assembly processes and decreased cycle time.<
>