Words such as positioning accuracy are widely used, and there is no clear regulation on how to define them. Generally the AIMA proposal is the most widely used. This method of consideration is to repeat the positioning in the same direction 7 times from any point, calculate the average value X and the degree of dispersion σ from the obtained data, and define the deviation ±3σ centered on X, that is, the maximum / The minimum value is the system accuracy, and ±3σ is defined as the repeatability accuracy. This method is reasonable at first glance, but it is only suitable when the actual error conforms to a normal distribution, and it is wrong when the error has a special law. Furthermore, it is difficult to analyze the causes of errors when there is only this evaluation method. The author here introduces another method to evaluate the cause of the error from the periodic regularity that appears with the continuous movement of the platform. 5.1. Measurement of the positioning accuracy of the machine tool A semi-closed-loop controlled machine tool is measured using a grating ruler, and a large error (30-40μm/240mm) is obtained. In order to find the reason, the positioning accuracy and posture accuracy were measured with a laser length measuring instrument. The measurement points are shown in Figure 9. Figure 10 shows the vertical steering accuracy and horizontal steering accuracy of the platform, especially the horizontal steering accuracy has a large error. Figure 11 shows the results of measuring positioning accuracy with grating ruler and laser length measuring instrument in each measuring point, which is represented by a solid line. The results for the individual measurement points vary considerably. Calculate the value of all position errors at each measurement point including all position errors from the attitude accuracy shown in Fig. 10, plus the lead accuracy of the ball screw, indicated by the dotted line in Fig. 11. It can be seen that this is basically the same as the positioning accuracy, but the lead error of the ball screw shown by the dotted line is very small, and the positioning accuracy is basically determined by the posture accuracy. Generally, flatness is often measured as the attitude accuracy of the platform. In the case of relatively small travel, the measurement result is only 4μm. Although it is often misunderstood as high accuracy, it can also be understood that this has a great impact on positioning accuracy. The attitude accuracy of the platform is greatly affected by the positioning accuracy, and the temperature rise also has a great influence on the attitude accuracy in a wide range of fluctuation errors. In the case of full-closed-loop control, the influence of attitude accuracy will also appear. For example, when using a grating ruler to control the position, there will be a large positioning accuracy error in the center of the platform. 5.2. Evaluation of backlash Fig. 12 shows the result of NC mechanical step feed, resulting in a backlash of 20 μm. Using the same machine, continuously increase the conveying speed to measure, as shown in Figure 13, the empty distance is negative (moved). This is because there is a delay in starting relative to the actual movement commanded, so there is an overtravel when stopping. Figure 14 is a mechanical model with delay and overtravel states established. As shown in the model, delay and overtravel have the same relationship in mechanics, so the reciprocating air travel is twice the delay at start or overtravel at stop. Now, if there is only a delay but no overtravel, it will become the same as in Figure 12. As shown in Figure 13, the error of the delay overtravel is larger. Because the overtravel will change according to the mechanical state at the time of stop, the platform model is assumed to be a spring mass model with 1 degree of freedom, and the relationship between the conveying speed and the empty travel is obtained, and compared with the measured value, as shown in Figure 15. It can be seen that the calculated values ​​are in good agreement with the measured values. 5.3. Measurement of small platforms Figure 16 is an example of positioning accuracy measurement under semi-closed-loop control and full-closed-loop control in precision positioning platforms. This platform is composed of a ball screw (lead 3mm) and a crossed roller guide. The semi-closed-loop encoder is 1000 divisions/1 revolution, and the full-closed loop uses an optical grating ruler with a resolution of 0.1 μm. The measurement results show a high accuracy even when the measurement error is included. Although this evaluation method evaluates what level of accuracy the platform has, it is impossible to know what error elements are included in this accuracy. In order to explore the cause of the error, as shown in Figure 17, it is the result of measuring the progress of continuous feeding. The measurement method is to trigger the built-in encoder of the motor, collect 100 points of data in one revolution, and find periodic error components. In order to analyze this narrow-range fluctuation component, the data is subjected to slope correction, smoothing (moving average), and the vertical axis is expanded to obtain the graph in Figure 18. Although a laser length measuring instrument was used for this measurement, there are fluctuations of ±0.1 to 0.2 μm due to environmental factors such as minute vibrations and air conditions. The data were smoothed to remove this component.