Abstract
NAS 979 is well–known standard for testing motion accuracy of multi–axis machine tools. One of the testing methods of the standard is to cut a cone frustum and measure its accuracy. Although it is established for long time ago, it is still used since better method not exists. In the previous study, it showed that the moving range of straight axes varies very much if the measuring position varies in case of the machine that has two rotary axes in the table side. In this study, ball bar measurements were done on the real 5–axis machine and try to diagnose the error sources that worsen the motion accuracy.