Abstract: Presently, in the application of Taguchi method, the quality loss function contains the quadratic term only, but the linear and other high order terms are neglected. This is unadvisable for the smaller the better situations, and the linear term should be included in the quality loss function. In this paper, the design issue of the quality loss function for the smaller the better situations is studied by determining the loss coefficients of the linear and quadratic terms. Algorithm is proposed to calculate the loss coefficients.  Comparative analysis is carried out and results show that the loss function with the quadratic term only is just a special form of the quality loss function under the smaller the better characteristic. The results are verified by a practical case problem.

Key words: taguchi method, the smaller the better characteristic, quality control, quality loss function