Analysis of accuracy of 5-axis CNC milling machine through standard and modified circular test

In analysis of accuracy of 5-axis CNC milling machine measuring system with telescopic kinematic ball bar was used. Standard measuring test was made. The modified test was proposed with application of investigated inclined basis. The obtained results were compared. The accuracy of machine tool in indirect way was evaluated.

The use of five-axis CNC machine tools for machining objects with complex 3D shapes allows for high geometry accuracy. To avoid machining errors and to obtain high quality of machined parts, an efficient and quick diagnostics of multi-axis machine tools is necessary.
Problems of machine tool diagnostics should be taken into account by their producers and also by users in the current production process. Diagnostic methods use static, dynamic and thermal properties correlated with individual types of machining errors to assess machine health. Errors that should be measured and aimed at their elimination are mainly of a geometrical, kinematic, thermal nature, refer to drive systems and regulators, measurement systems and control [1].
The required accuracy of multi-axis CNC machine tools is provided by diagnostic tests combined with control of servo systems and error compensation implemented by control systems. The methods for testing the accuracy of machine tools were strictly standardized and presented in international standards ISO 230.
The aim of the work was to analyze the accuracy of a five-axis CNC milling machine using a measuring system with a telescopic ball bar type ball bar. A standard measurement test was carried out.
A modified test was proposed with the use of a 45° angled test bed. This innovative approach provided the basis for further research. Measurement results were compared. An indirect attempt was made to assess the accuracy of the machine tool under test.

Kinematic ball bar
The ball bar kinematic ball bar was constructed in 1982 by Bryan [2][3][4]. It is a measuring device made of two balls and a telescopic rod with a built-in displacement sensor. The balls at both ends act like ball joints, and the measuring rod measures the distance between them with great accuracy. The original purpose of this device was the calibration of coordinate measuring machines.
Since its first use in CNC machining centers in 1985, the ball bar has been widely used because [5]:  accuracy of the test is higher than the accuracy of the machines,  device is reliable and easy to use,  measurement results indicate machine tool errors.
Due to this, it is possible to diagnose sources of errors by analyzing and comparing the obtained measurement profiles on polar charts.
With time, other test methods proposed in [6] were proposed.
In 1994, the test of machine tools using a measuring rod was included in the international standard ISO 230-4.
The accuracy of currently manufactured machine tools increases and is comparable to the accuracy of the measuring rod. Therefore, new methods are sought for measuring accuracy in the roundness test [7].

Accuracy check test
The milling machining center DMU 60 monoBLOCK was selected for the tests ( fig. 1). The machine tool, manufactured in 2010, was operated in a modern industrial plant in a three-shift system. The milling machine was equipped with a rotary table and a pivotable spindle and met the criteria of complex multimanic processing in five axes [8].
The accuracy test for the machine tool tested was carried out using the Renishaw wireless ball bar system QC20 -W ( fig. 2).
Measurements were made in a climate-controlled hall at 20°C. The machine achieved thermal stability after two hours of unloaded operation. Accuracy tests using the ball bar system were carried out in accordance with the ISO 230-4 standard as a standard test [10,11].
For comparison, a modified accuracy test was carried out, in which the magnetic base of the ball bar device was inclined to the surface of the machine table by an angle of 45° ( fig. 3).  [8] The software available in the system did not automatically generate a measurement code for the modified setting. Using the CAM module, a test program for the analyzed measuring case was prepared.
Before starting the test, the QC20-W was calibrated on a 150 mm long pattern.
The course of auxiliary operations preceding the measurements included:  fixing the magnetic spindle in the machine spindle and setting the base on the machine's rotating table,  determination of the feed rate 1000 mm/min on the machine tool,  determining the amount of thermal expansion of the machine tool,  loading the test machine of the machine into the memory,  selection of a measuring rod with a length of 150 mm,  connection of the QC20-W device with a computer,  starting the measurement program.

Results of measurements of the standard accuracy test
The results of the measurements were analyzed in detail. The diagnostic test included basic values of errors and deviations as well as their percentages in the total deviation of machine tool roundness.
The graphic runs of the bidirectional circular deviation, the value of which was determined as 22. 8  In the measured deviation of the machine tool roundness, the largest shares showed:  lateral clearance of the X axis (40%),  lag error (14%),  rectangular deviation (9%),  difference in positioning deviation (8%),  backlash of the X axis (5%).
Cross-lateral play (lateral play) of the X-axis is an error that results in uneven movement of the machine tool assemblies along the guides during reversing movements. The main cause of lateral play ("lateral") is the defect of the machine tool guides. This is a different error than the backlash, which is loose axle.
The lag error (servo mismatch) arises as a result of improper selection of the drive parameters of the X and Y axes. The result of this error is that the interpolation path is not a circle. Using small feed values minimizes the lag error.
The lateral clearance of the X axis and the lag error were 54% in the roundness deviation of the standard machine tool test. The machine's roundness deviation was determined by the following components ( fig. 5b):  perpendicular error (48%),  Y-axis straightness deviation (9%),  difference in position deviation (6%),  deviation of the straightness of the X axis (6%),  Y-axis backlash (6%).

Measurement results of the modified accuracy test
The squareness error is the deviation of the direction of the machine axis (e.g. X and Y) from the right angle, which may be local or be the result of incorrectness of the guides along the entire length. The axes of the machine tool may exhibit vertical deformation (i.e. concavity, convexity, inclination) and excessive wear resulting in increased slack during movement. As a result of the perpendicular error on the machine, there is a deviation of the perpendicularity of the milled planes and problems with the shaping of cylindrical surfaces, both external and internal.
The Y axis straightness distortion occurs in the case of local distortion of the machine tool guides. It is probably caused by their wear or damage. The amount of rectilinearity error does not depend on the speed of feed motion.
The perpendicularity error and the Y-axis straightness deviation had a 57% share in the roundness deviation for the machine tool in the modified test.
The proposed modified test measuring the accuracy of a multi-axis CNC machine using a ball bar measuring bar system was an alternative to machining a test item in the form of a truncated cone inclined to the plane of the machine table.
Preliminary results of machine tool accuracy tests by machining test objects and measurements using a modified test showed a good correlation, which may be a prerequisite for further measurement tests and their analysis.
The method of indirect evaluation of the accuracy of machine tools with the use of machining of test objects is presented in detail in [12].

Conclusions
Diagnostic tests of five-axis CNC machine tools with the use of the QC20 -W ball bar measuring system were aimed at improving the technical condition of the used machines.
Based on the results of the standard and modified accuracy test, it can not be unambiguously determined that the machine tool under test meets the requirements. It resulted from a limited number of diagnostic measurements.
The results of roundness deviations measurements for the tested machine (lower than 25 μm) did not exceed the permissible values.
The presented modified measurement test for an indirect assessment of the accuracy of machine tools should be subjected to further testing and careful analysis.