Advanced Content Master's Series
Division: Six Sigma Forum
When conducting a design of experiment (DoE), the practitioner is encouraged to run replicates or repeats to increase the power of the experiment and to more accurately characterize the mean of the key process operating variable (KPOV). However, that is not the only leverage repeats and replicates provide. By using Taguchi’s signal-to-noise ratio as the response for the DoE—instead of the KPOV—the practitioner can perform an assessment of both the mean and the variability of the KPOV—optimizing both simultaneously. Based on Taguchi’s loss function, a signal-to-noise optimization would rather have the KPOV slightly off target with less variability rather than exactly on target with large variability—the former is more likely to provide higher customer satisfaction. In this session, the concept of Taguchi’s signal-to-noise ratio will be introduced and explained. In addition, the various signal-to-noise ratio equations will be provided. Finally, two industry examples will be presented that illustrate the power of using Taguchi’s signal-to-noise ratio as the response for the DoE, rather than just the mean of the KPOV. Get the most out of a DoE—leverage those repeats and replicates!