As competition among firms becomes increasingly intense, there is growing interest in simulation as a means of speeding up product development, but before designers can put simulation to practical use, there is an urgent need to improve the automation environment. In this study, quality engineering was combined with an in-house analytic model of a photosensitive drum drive unit, a direct product experiment was performed with both control factors and noise factors assigned to an orthogonal L108 array, and a nominal-is-best evaluation was carried out. In selecting the optimal conditions, both the variability and the mean value of the maximum amplitude on the photosensitive drum, which is a smaller-is-better characteristic, were reduced by reducing the sensitivity, and the reduction was almost perfectly reproduced in a confirmatioll experiment. When components were actually manufactured under the optimal conditions and tested in a working copy machine, the measured mean value was nearly identical to the mean value predicted by the simulation.
