壓縮包內含有CAD圖紙和說明書,均可直接下載獲得文件，所見所得，電腦查看更方便。Q 197216396 或 11970985
The Development and Application of a Planar Encoder Measuring
System for Performance Tests of CNC Machine Tools
W. Jywe
Department of Automation Engineering, National Huwei Institute of Technology, Huwei, Yunlin, Taiwan
In this paper, a measuring device with a planar encoder is developed to test the performance of a CNC machine tool. With the assistance of a PC, this system can be employed for both 2D contouring tests and 3D positioning tests for a CNC machine tool. The structure and the principle of the system, the applications for the general 2D contouring test, the drift test, and the specified geometric part path tests. An actual case study on improving the accuracy of machining a cam are
described. Finally, a new 3D positioning method using the optic encoder is demonstrated.
Keywords: Ball bar system; CNC machine tool; Geometric part path; Planar encoder; Thermal drift test; Three-dimensional positioning; Two-dimensional contouring
1. Introduction
Machine tool performance and consistency is the main determinant of the quality of parts machined by it. It is of importance to check the performance of the machine tool systematically
for direct quality control purposes or to compensate for this uncertainty. Schlesinger, in 1932 [1], first provided a systematic testing method for machine tools. This method was developed as the basis of the ISO standard. Tlusty, in 1959 [2], employed an electric level and sensor to test the spindle accuracy. Tlusty and Koenigsberger [3] and Burdekin [4] indicated new testing rules for machine tools. Burdekin [5] checked the relation between the motion accuracy of machine tools and the machined part. Tlusty [6] proposed a non-cutting testing method. The tests for machine tool performance were then classified into a direct cutting test and an indirect cutting test. Ericson [7] first described the work zone of machine tools. Bryan and Pearson [8] explained the definition and the way to measure the pitch, roll and yaw motion and straightness error. After the commercial laser interferometer [9] was available, the analysis of volumetric errors [10–13] was described. Voutsudopoulos and Burdekin [14] indicated a calibrating model for a coordinate measuring machine. Fan [15] used a laser interferometer and a related device with the assistance of a PC to calibrate different types of NC machine tools. Zhang and Hockey [16] obtained the 21 error components by measuring the position errors. Zhang and Zang [17] designed a 1-D ball array to find the 21 error components, then Zhang [18] described a rapid method to obtain the straightness error. In 2000, Jywe [19] described a method of employing a ball bar system for the verification of the volumetric error of CNC machine tools. Circular tests were developed to check both geometric errors and contouring errors. Burdekin [20] described cutting tests using circular paths for accuracy assessment. Bryan [21] developed the first ball bar system for the contouring test. However, in this system, the uncertainty is high due to the friction between the master balls and the magnetic sockets and no accurate contouring radius was given. Knapp’s system [22,23] used a circular comparison standard disc mounted on the test table of the machine tool and a 2D-probe. The problems for this system are the existence of friction between the 2D probe and the disc, its small bandwidth, which causes the system to be unusable for high-speed contouring tests, and the high cost of the 2D probe. Kakinov [24–27] provided a series of methods using a ball bar system to calibrate a coordinate measuring machine and CNC machine tools. Knapp [28,29] described a rule to reduce the errors due to stick–slip etc. Burdekin and Park [30] modified the original ball bar system by employing a four-rod linkage. Burdekin and Jywe [31] provided a method to diagnose the contouring error and to adjust the parameters of the CNC controller to optimise the performance of the tested CNC machine tool. Ziegert and Mize [32] described a laser ball bar system. All these ball bar systems, including the lateral Renishaw system [33] provide only the radius error during the contouring test. This limits the analysis of contouring error since no individual error in each axis is available. Jywe [34] used two position silicon detectors (PSD) for a contouring test to obtain the contouring error in each axis. One laser source emits a laser beam and the laser beam is split into two vertical lines and projected onto two positioning silicon detectors, which are set vertically to each other on the test machine table. The Heidenhein [35] grid encoder also provides a 2D contouring test, but at very high cost. A planar encoder system was developed [36] for applications such as semiconductor and electronics manufacturing equipment. The system, which has a good dynamic response, can provide up to a 0.1 m resolution in positioning, and it is of importance that it is of low cost. However, the original planar encoder was designed for manual operation. It is not suitable for the contouring test of CNC machine tools due to the following considerations:
1. The original system only included an encoder and a reading head. No related interface and driver are available.
2. Thus there were no related contouring software and testing methods. Thus, in this paper a new computer-aided planar encoder system has been employed and integrated, with the related software, for checking the both the dynamic performance and the geometric error of a CNC machine tool. It is of importance that 90% of the cost of the contouring testing device can be reduced compared to the equivalent Heidenhein grid encoder system. From the previous research, it has been found that the device for the circular test is not always suitable for a 3D geometric error test. Furthermore, these devices are not suitable for a free-form 2D contouring test. In this paper a simple measuring device is designed and developed to check contouring performance with a single axis output. The application for a 3D positioning test is also developed.
2. The 2D Planar Encoder Contouring Measuring System for CNC Machine Tools
2.1 Principle of the Planar Encoder
A planar encoder system, such as the Renishaw RGX grid plate, has been developed for applications such as semiconductor and electronics manufacturing equipment. The system uses a reading head with two orthogonal sensors that read a checkered grid in both the X and Y directions simultaneously. The system has a good dynamic response and can provide up to 0.1 m
resolution in positioning. The software in V-Basic is edited to carry out the measuring procedure.
Figure 1 shows the arrangement of the contouring test using the simple planar encoder. This planar encoder provides positioning information in each axis for 2D contouring. During the test, the planar encoder is set on the CNC machine tool, and the reading head is fixed in the spindle. The computer software can read the sampling data via a counter card.
3. Uncertainty of the Measuring System
3.1 Uncertainty Due to Sampling Procedure
The developed software incorporated the following factors:
1. Sampling must be uniform around the profile and reasonably
independent of the type and speed of the computer. A Planar Encoder Measuring System for CNC Machine Tools 21
2. Sufficient sampling data is required to display and analyse the error at high resolution.
3. Sampling data should be independent of contouring speed, computer speed and contouring radius.
3.2 Uncertainty Due to Thermal Effect
Considering the thermal effect of the system for the tests, if the temperature in the planar encoder is different from that of the machine tool table, the radius error will be affected. If the temperature of the planar encoder itself is not uniform, the out of roundness error will be affected. Although the thermal expansion coefficient of the planar encoder is rather small, to minimise the effects, the encoder should be put on the test machine table for some time to reduce the difference in the temperatures and to let the temperatures of the encoder stabilise.
4. Test Results of a Circular Contouring Path
A simple contouring test is carried out on the XY-plane of a vertical CNC machine tool with a 0M Fanuc controller. The contouring result is shown in Fig. 2. The anticlockwise and clockwise contouring tests at 20 mm radius can meet ISO 230-1 and 230-2 requirements. From the results, the absolute radius error can be found easily. For general contouring systems, only the out of roundness is given. Furthermore, the error for each axis can also be found individually if necessary. This is useful for analytical purposes.
5. Thermal Drift
This contouring system provides a non-contacting contouring test. For general contouring systems such as a ball bar system, only a limited number of runs are executed, due to the problem of winding of the signal cable. In this application, the test run is unlimited. Thus, a thermal drift test can be carried out easily without additional fixtures. For eight-hour continuous clockwise contouring runs, contouring results are shown in Fig. 3 for each two-hour period. The contouring centre for each 30 minutes, is plotted in Fig. 4. The contouring centre drift is significant in the 8 hours. It is important that not only the contouring centre drift is given but also the contouring error form in each run can be obtained. From this test, the performance of continuous runs
can be monitored easily by this system.
Fig. 1. The optical measuring system for contouring performance test on a CNC machine tool.
Fig. 2. The clockwise and anticlockwise contouring test results with
Fig. 3. The thermo drift test results during 8 hours continuous the planar encoder measuring system.
Fig4.The thermal drift test results presented by the drift of the contouring centres.
Fig. 5. The squareness test result on a CNC machine tool with theplanar encoder measuring system.
6. Squareness Error Test by Planar Encoder
The squareness error can be tested easily with the planar encoder. Let the encoder be set on the tested plane. The reading head goes along the square of the encoder. A CNC machine tool was tested and the result is shown in Fig. 5. contouring test.
7. Laser Diode and Quadrant Sensor Contouring System [37]*
Using a laser diode and a quadrant sensor contouring system, the planar encoder contouring system can be verified. Using a 2 mm clockwise contouring radius, Fig. 6 shows the contouring
result using the quadrant sensor, while Fig. 7 gives similar results.
Fig. 6. The contouring test results for square path with the laser diode and quadrant sensor contouring system [37].*
Fig. 7. The contouring results for a square path with the planar encoder.
Fig. 8. The test result for a combination of various geometric shapes provided by the planar encoder measuring system.
Fig. 9. The geometric shape of a specified 2D cam.
Fig. 10. The test results of the path of the CNC machine tool with/without cutter radius compensation.
Fig. 11. The test results with self-calculated cutter radius compensation under different feed rates.
Fig. 12. The structure of a 3D positioning measuring device.
without sensors, is connected to the spindle of the tested CNC machine tool and to the reading head by two individual balls and magnetic sockets. The centre of the ball on the ball bar on the side of the spindle is the 3D measuring target to be analysed. When the target is reached, the first sample from the planar encoder is then taken at its first position. Without moving the target, the planar encoder is moved to a neighbouring point and a second sample is taken. Finally, the other
neighbouring point is sampled as the third sample. Each of the three samples includes 2D coordinates, the 3D coordinates of the target which can be analysed. Thus each 3D movement
will be obtained by this 1-point and 3-step (1P3S) method. This method can be described as follows. To obtain the 3D positioning coordinates X, Y and Z, a simple model is developed in Fig. 13,
where:
Fig. 13. The model for analysing coordinates of x, y, z.
x, y, z are the coordinates to be analysed.
x1, y1, z1, x2, y2, z2, x3, y3, z3 are the coordinates provided by
the 2D optical scale in the first, second and third step samples.
L1, L2, L3 are the lengths provided by the ball bar. Then,
Solving the equation,
where
Here, two possible solutions can be found. One is on the top of the planar encoder, while the other is below it. Thus, in this application only the coordinates on the top of the ball plate are used. After the coordinate z is found, x and y can also be found. In this application, the ball bar length is fixed, thus L1 L2 L3. To extend the working range a standard or laser ball bar system with a long working range displacement sensor can be employed. In that case, L1, L2, L3 can be obtained by that sensor. To minimise the cost, in this application only one set of planar encoders and a simple ball bar are considered. Thus, A Planar Encoder Measuring System for CNC Machine Tools 27 the coordinates x1, y1, z1, x2, y2, z2, x3, y3, z3 have to be obtained by the planar encoder in three individual samples.
The sampling procedure (1P3S) is:
1. Let the machine tool move to the tested position (one point).
2. Take the sample by the planar encoder (step 1).
3. Move the reading head to a neighbouring position related to the planar encoder; the tested machine is not moved. Take the sample by the planar encoder (step 2).
4. Move the reading head to the next neighbouring position, the tested machine is not moved. Take the sample by the planar encoder (step 3). z1, z2, z3 are affected by the flatness (?Zij) of the linear
XY stage. The flatness of the linear XY stage ?Zij is equal to the ith grid
on the X-axis and the jth grid on the Y-axis.
11. Discussion and Conclusions
In this paper, a planar encoder system was employed for a contouring test of a CNC machine tool. It was proved that this system could be employed successfully for the contouring test. The advantages of this application can be summarised
as follows:
1. During the contouring test, the contouring error for each individual axis can be obtained. This is not possible using a general ball bar system. This function provides more useful information for analysing the contouring error. 2. The system can be employed for a long-period thermal drift test, but the traditional ball bar systems cannot, because in this there is no cable which can be wound up.
2. For contouring a combination of complicated curves such as a cam, the system can be employed while a general ball bar system cannot. Table 1. The verification result of the planar encoder for 3D positioning
With this 2D optical measuring system, the 3D positioning error test can also be performed successfully. Thus this optical encoder can be employed for both dynamic performance and geometric error tests on CNC machine tools. Acknowledgements The work was supported by National Science Council, Taiwan, Republic of China, Grant Number NSC-88-2212-E-150-006.
References
1. G. Schlesinger, Inspection Test on Machine Tools, Machinery
Publishing Co. Ltd, London, 1932.
2. J. Tlusty, System and methods for testing machine tools,
Microtechnic, 13, p.162, 1959.
3. J. Tlusty and F. Koenigsberger, Specifications and Tests of Metal
Cutting Machine Tools, UMIST, Manchester, 1970.
4. G. Schlesinger, F. Koenigsberger and M. Burdekin, Testing
Machine Tools, 8th edn, Machinery Publishing Co. Ltd,
London, 1978.
5. M. Burdekin, “Cutting tests for accuracy assessment”, Technology
of machine tools, 5, pp. 9–11, 1980.
6. J. Tlusty, “Testing of accuracy of NC machine tools”, Technology
of Machine Tools, Supplement 1, 1980.
7. C. Ericson, “Machine alignment – the first step towards product
accuracy”, Tech. Paper MM66-171, ASTME, 1966.
8. J. B. Bryan and J. W. Pearson, “Machine tool metrology”, Lawrence Livermore National Laboratory, Livermore, CA, UCRL71164, 1968.
9. Hewlett-Packard Co, CA, “Calibration of a machine tool”, Laser
Measurement System Application Note 1156–4.
10. W. J. Loves and A. J. Scarr, “The determination of the volumetric
accuracy of multi-axis machines”, Proceedings of the 14th MTDR
Conference, pp. 307–315, 1973.
11. R. Schaltshik, “The components of the volumetric accuracy”,
Annals of the CIRP, 25, pp. 223–228, 1977.
12. R. Schaltshik, “The accuracy of machine tools under load conditions”, Annals of the CIRP, 28, pp. 339–342, 1979.