Calibration of five-axis machine tool using R-test procedure

Geometric calibration of five-axis machine tools is presently a very complex procedure, since most measuring devices are restricted to acquiring of individual errors. Industrial experience shows that several days are required for the complete metrological examination of a five-axis machine tool. In this paper, a method based on R-test is presented that allows the complete geometric calibration of compact machining centers with swivelling rotary table within a few hours. This advance in calibration speed is achieved by the combined indirect measurement of all geometric errors with a contactless 3D measurement probe attached to the tool holder of the machine and a reference ball attached to the table. With the newly developed method, almost all errors of linear axes and rotation axes can be calibrated. For the calibration method, a model of the machine errors and the measurement process is required. The position of the sensor at the TCP (Tool Center Point) and the ball on the rotary table are modelled based on the rigid body theory. Each error motion is modelled as a function of their related axis position. A measuring strategy is defined to allow the numerical identification of all relevant errors with minimum measuring effort. For this purpose, measurement data from several ball positions has been used as training and test data. Several error models are defined and tested to compare their performance. For analysis of the calibration quality, the geometric errors of a test machine tool are compensated and two methods are used to measure the residual error: 1) R-test at independent ball position. 2) Circular test in three coordinate planes. Based on the results, the optimal error model for the test machine is decided.