Koons, George; Meyer, Steve; Kurowski, Mark (1992, ASQC) Motorola Lighting, Inc, Buffalo Grove, IL 60089
Motorola Lighting, Inc. (MLI) used Design of Experiments (DOE) methods to determine which factors affect customer satisfaction ratings of electronic lighting ballasts.
MLI wanted to achieve Six Sigma quality levels for their electronic lighting ballast by simulating effects of combinations of various performance factors and choosing a combination of factors that promised the highest quality.
They identified eight quantifiable performance parameters that customers look for in lighting ballasts. They established performance criteria for each parameter and identified factors that would affect that performance: variations in component values, fluctuations in incoming voltages, and load induced by different lamp types.
They constructed a five factor, two-level factorial design to test the interaction of five components, two lamp types, and three input voltages. Resulting data was analyzed using regression analysis to identify component effects significantly related to performance characteristics. Output of the regression analysis was imbedded in a Monte Carlo simulation model written with Statistical Analysis System (SAS). The simulation selected component values from probability distributions and predicted how the ballast would actually performed if built with those components. The simulation goes on to predict distribution of target performance criteria over a run of 5000 ballasts.
As a result, MLI was able to select components and other factors that most closely approximated a Six Sigma level.
Customer satisfaction (CS),Design of experiments (DOE),Monte Carlo methods,Motorola Corporation,Performance objectives,Six Sigma