An Improved Characterization for Predicting a Capability index with Dependence on Manufacturing Target Bias

A target bias dependent capability index predictor is proposed and compared to two other commonly used paradigms. The formalism for the proposed approach assumes a normal (Gaussian) distribution in the values for a specified product parameter. A probabilistic description of the manufacturing process is used to predict the dependence for the fraction of rejected components on a short-term bias- independent capability index and a normalized target bias. Comparisons of paradigms for predicting both the fraction rejected and target bias dependent capability indices were tested for the special case of a three sigma process. The proposed capability index predictor can be implemented with either "canned" routines or with reasonably accurate analytic versions of the error function and its inverse. The demonstration indicates the proposed more accurate approach, is a less pessimistic predictor than the commonly used industry standards. Applicability of approach to formally include the impact of target bias on long-term capability index is discussed.