Adjusting control charts for inaccurate measurements

Abstract

The concept of statistics-based control is nearly a century old. The basic assumption is that the measurement is precise and therefore all calculation based on these data is also precise. Typically, this assumption is not far from reality and thus does not affect the proposed control limits. However, in some cases the rounding-off of measurements may lead to erroneous decisions. This study examines the effects of rounding-off on the average (X ?) control lines and suggests a simple method way to design new control lines, based on the conventional average run length (ARL).