5 m-measurement system for traceable measurements of tapes and rules

Transcription

1 5 m-measurement system for traceable measurements of tapes and rules Tanfer Yandayan*, Bulent Ozgur Tubitak Ulusal Metroloji Enstitusu (UME) PK54, 4147 Gebze-KOCAELI / TURKEY ABSTRACT Line standards such as measuring tapes and rules still form a major component in the traceability chain for dimensional metrology. The calibration or verification of these standards is carried out in order to check whether they comply with the classes given in OIML standards or with the user specifications. 5 m-measurement system was constructed by UME for calibration of workshop rules and tapes up to 5 m long. The system is mainly composed of 6 m rail system, mechanical parts and optical units. The rails are kinematically located on a heavy marble construction and a carriage, which employs a camera for probing of the scales on the tapes, is moved a long the rails during the measurement. The image of the scale taken by the camera is viewed on the monitor screen together with software. The operator can perform the probing process by simply placing the measured scales on the viewed target with the help of an X-Y table located on the carriage. The X-Y table and the carriage movement are measured by a 6 m incremental linear encoder integrated to the system or optionally by a laser interferometer. The measurement values are transferred to the computer for evaluation according to OIML standards. The developed software running in windows environment also performs temperature measurement for required corrections. The software automatically takes the measurement results from either incremental linear encoder or laser interferometer, depending on the option chosen in the software) and processes it for representation to the user. The overall uncertainty for the system with the linear encoder is U = [(.5) 2 + (.5L) 2 ] 1/2 mm, where L is measured length of the tape or rule in meters. Keywords: Measuring tapes, legal metrology, length 1. INTRODUCTION The calibration of measurement tapes and rules are generally carried out by a comparison method. The reference and test tapes are stretched out side by side and length of the test is determined in terms of the reference. This method is applied to the workshop rules (no stretching) and tapes of OIML R35 Class I (trade measures) up to Class III 1. The accuracies for OIML standards start from ( L) mm to ( L) mm (L is the length in meters). The tapes are normally quoted as standard at a specified temperature and tension. European standards and directives of 73/362/EEC also specify the accuracy classes of the tapes for European market 2. The comparison method requires reference tapes or rules, which are traceable to SI unit of length. Therefore, absolute length measurement of the reference standards is necessary. Most of the National Metrology Institutes (NMIs) designed and manufactured their own equipment. The detailed information and the uncertainties ranging from about.1 mm to.1 mm (for the length independent part) can be obtained from BIPM (Bureau International des Poids et Mesures) web site 3. 5 m bench of National Weights and Measures Laboratory (NWML) in UK 4 and most NMIs systems utilize laser interferometers for measurement of the length. The maximum length of the bench in the first place depends on available laboratory size. Smaller size benches can also used. Larger tapes can be measured in several intervals and the measurement results are connected and reported with higher uncertainties. EUROMET (European collaboration in measurement standards) length project of 677 has been agreed for calibration of tape standards in order to support confidence for such services of European NMIs 5. *tanfer.yandayan@ume.tubitak.gov.tr: phone : fax : Proceedings of SPIE Vol. 519 Recent Developments in Traceable Dimensional Measurements II, edited by Jennifer E. Decker, Nicholas Brown (23) 23 SPIE X/23/$

2 A bench for measurement of tapes and rules up to 5 m has been designed and manufactured in UME. The measurement system is designed that length measurement can be performed by a linear encoder as well as by a laser interferometer. The purpose of using the linear encoder with.1 µm resolution is to reduce the cost for length measurement system. This was achieved by 5m-measurement system of UME and 3 USD cost of laser interferometer is reduced to 5 USD by using the linear encoder as a measurement system. The linear encoder is calibrated by the laser interferometer and the traceability of the measurement is assured. The laser interferometer can also be used for tape and rule calibration optionally with the system. However, it is usually reserved in the laboratory for other calibration purposes M-MEASUREMENT SYSTEM The 5 m-measurement system is located in the dimensional laboratory, the temperature specification of which is 2 ±.5 C. It is 6.6 m long,.6 m wide and 1.5 m high. General view of the facility is illustrated in Fig. 1. The measurement system mainly consists of base construction, 6 m rail system, mechanical parts and optical units. Figure 1: General view of the 5 m-measurement system 2.1. Base construction and the rail system The base is made of 3 marble blocks each 2.2 m long and of cross section 6 mm wide and 2 mm high. These are supported on 3 steel frames interconnected to each other. The marble blocks are rest on the steel spheres located on the adjustable screw mechanism. The heavy marbles can be leveled precisely with help of the adjustable screw mechanism. The steel plates supporting the adjustment mechanism of rails are screwed into the marbles. They are positioned every 6 mm along the 6.6 m marble construction. There is only one rail for each side. Each is made of 6 meters length of centreless ground steel rod about 25 mm in diameter. Two rails are spaced 36 mm apart and are supported in adjustment mechanisms positioned every 6 mm which enables vertical and lateral adjustment of the rails. During the construction process, the rails have been leveled using precise optical level systems. The straightness of the system is that a fixed point on the carriage stays within the area of a.5 mm diameter cylinder over the 5.5 m measurement range. 42 Proc. of SPIE Vol. 519

3 2.2. The Carriage The carriage is moved along the rails with help of the operator. It employs a tilting mechanism for further adjustment and carries a precise X-Y table for fine adjustment of the optical units (Fig. 2). The carriage can be clamped to the one of the rails and the moving parts are translated by micrometer of the X-Y table along the direction of the measurement axis. As illustrated in Fig. 2, the incremental linear encoder measurement head is connected to the X-Y table and the micrometer movement along the measurement axis is measured with this incremental linear encoder. It is also possible to connect the laser optics to the carriage and to perform measurement using a laser interferometer. The X-Y table carrying the optical probing unit enables fine localization of the scale marks by translating the camera on the scale marks. It is only used in X direction. The X-Y table is commercially available, Mahr X-Y table PKT. Optical Unit (CCD and Mag. Lens) HP Laser X-Y Table Laser corner cube interferometer measurement axis Micrometer Carriage Line scale of Linear encoder X-Y table HP Laser corner cube Tape Tape Rails Linear encoder Tape support Front View Marble Figure 2: The carriage and main parts Left View 2.3. Tape support, and tape clamping-tensioning system A 6 m aluminum bar with a cross section of 1 1 mm is used for tape support. It is located between the rails and level adjusted with respect to the carriage movement during construction in order to maintain the camera and its optics focused on the tape. The tape is stretched out on the flat surface of the 6 m aluminum bar. The tape clamping-tensioning system is shown in Fig. 3. One end of the tape is securely anchored while the other is connected to a hanging weight by a wire. The wire is thin and sufficient strength to withstand tension. The wire is passed over 1 mm diameter pulley wheel, which has low friction bearings. The pulley wheel is adjustable in lateral and vertical position. The connections to the tape at both ends incorporate swivel joints in order to minimize the torque transmitted to the tape via wires. The both ends have lateral and vertical adjustment. One end where the tape is anchored also has a longitudinal adjustment. Proc. of SPIE Vol

4 (a) (b) Figure 3: Tape clamping-tensioning system (a) beginning and (b) end point 2.4. Localization of scale marks (optical probing) Localization of scale marks is carried out with help of a camera. An analogue black and white camera with a magnifying lens is used to transfer the image of the scale on the monitor screen. The system magnification is (5 ) The operator can perform the probing process by simply translating the camera over the scale marks using the X-Y table micrometer. At the same time, the scale marks are observed on the monitor and an attempt is made in order to place the marks on the cross target (Fig. 4). The cross target is made by the software and can be adjusted according to scale mark widths by the user (Fig. 4). The depth of focus for the camera optics is 1.6 mm and the field of view is about 8 mm. (a) (b) Figure 4: Localization of scale marks (optical probing) (a) Adjustment of the cross target (b) Measurement position 2.5. Length measurement system A 6 m incremental linear encoder of Heidenhain is utilized for measurement of X-Y table micrometer and the carriage movement. The linear encoder employs a steel line scale with expansion coefficient of 1 ppm and has a resolution of.1 µm. The manufacturer specification for the accuracy of the linear encoder is ±5 µm. The linear encoder is fitted to the tape support as shown in Fig. 2 in order to reduce the Abbe offset. The attempt is made to be as close as practicable to the tape surface. The steel line scale is 18 mm in vertical direction and 75 mm in horizontal direction away from the tape surface with this position. Hewlett-Packard (HP) model 5528A laser measurement system is also used for length measurements. The HP interferometer optical axis lies centrally between the carriage rails, and is positioned as close as practicable to the tape surface to reduce the Abbe offset (Fig. 2). 422 Proc. of SPIE Vol. 519

5 Use of the linear encoder or the laser interferometer is optional in the software. They can both be used for calibration of tapes and rules. The HP laser interferometer is also used in order to calibrate the incremental linear encoder Temperature measurement system Platinum resistor thermometers (1 ohm: Pt1) are used to determine the temperature of the tape and the linear encoder. The sensors are placed on the tape support surface. Three temperature sensors are used along the 5 m-measurement system. Temperature corrections can be applied using the readings associated with each interval being calibrated Data collection and the software Comprehensive software has been written in Visual Basic in order to provide prompts to the operator. It controls flow of the data and provides the operator with a program presentation in a windows format whereby the operator can click through the measurement process. It also controls the flow of the data from the ancillary equipment for temperature measurement. The software reads the data from either linear encoder display or laser interferometer display. An option is available to select the source of measurements. General view of the program is shown in Fig. 5. After transfer of the measurement values, required correction is made (temperature corrections and error correction for linear encoder) and the results can be evaluated according to OIML standards. The software also maintains the error correction files of the linear encoder. This file is created during the calibration of the linear encoder with the laser interferometer. The readings taken from the display of the linear encoder are corrected with respect to the carriage position along the measurement range. The operator can see the corrected measurement values of the linear encoder on the computer screen simultaneously. Figure 5: General view of the software 3. TAPE AND RULE MEASUREMENTS Measurement of tapes and rules can be performed either using the laser interferometer or the linear encoder. While the rules are placed on the flat tape support without stretching, the tapes are required to be stretched out using the weights Proc. of SPIE Vol

6 according to manufacturer specifications or international standards. The tapes are usually loaded by forces of 1 to 5 N Alignment and tensioning process The lines are localized at the nominally the border of the scale. This is performed by usually taking into account the first approx. 3 mm of the lines. The scale marks are observed on the monitor screen along the measurement axis and required adjustments in the lateral plane are made using the tape clamping-tensioning system. Eventually it is ensured that the graduation points lie on a line closely parallel to the axis of the measurement system. The weights are chosen according to specified loaded force. It is hang to the wire that is anchored to the end of the tape Measurement process and use of the software The tapes and the rules are calibrated in 1 regularly spaced intervals unless there are some specific costumer demands. The start position, end position, number of cycles and the number of intervals are entered to the software. The program then calculates the target values and represents them to the operator on the screen. The program also allows the user to enter non-regular intervals. In addition to that, required identification such as serial number, costumer name, material and thermal expansion coefficient of the rule or tape can also be entered to the program. Temperature measurement is performed before the measurement and the temperature values are transferred to the software automatically. The operator then follows commands on the screen and positions the carriage to the target points shown on the screen. After completing the measurements, another temperature measurement is performed. The measurements results are corrected to the reference temperature of 2 C and given to the operator starting from the origin (zero). A graphic form of the results is also given by the program (Fig. 6). During the measurement process, the carriage is first set over the zero graduation line at the beginning, then sequentially at each line up to the maximum. The sequence is then repeated in reverse order ending on the zero mark. Finally the carriage is displaced and returned to the maximum graduation again. This is called bi-directional mode in the software and chosen by the operator before the calibration. The software allows the user to perform measurement in unidirectional mode as well. Figure 6: Outputs of the software 424 Proc. of SPIE Vol. 519

7 4. CALIBRATION OF THE 5 M-MEASUREMENT SYSTEM Calibration of the 5 m-measurement system has been performed in different steps. First the tilting errors i.e. angular errors of the carriage have been investigated using a laser interferometer angular optics and an electronic level meter. Then, further tests including length measurements capability and optical probing performance, have been performed in order to determine accuracy of the system Angular errors of the carriage Pitch and yaw error of the carriage along movement range of 55 mm have been determined using angular optics and roll error has been determined by the level meter. Fig. 7 illustrates the angular errors of the carriage during travel of 55 mm in 1 mm intervals. Arc secs PITCH Arc secs YAW Arc secs ROLL Figure 7: Angular errors of the carriage When the angular errors are studied, the reason for the large deviations at the end of the measurement range is found due to bending of the rails on that particular region. The attempt is made to reduce the influence of these deviations by using the compensation files for the line scale. And also the errors, which will be caused by these deviations, are estimated and included in the uncertainty budget. Angular errors of the X-Y table micrometer have also been investigated. The tests have been carried out when the carriage is positioned in 6 different locations along the rails. The angular error values of the X-Y micrometer have been found less than 4" over the ±2 mm measurement range of which is used for measurement. Proc. of SPIE Vol

8 4.2. Investigations for accuracy of length measurement with carriage movement It is practically impossible to locate the laser interferometer optical axis to the tape or rule measurement axis. But the attempt can be made to be as close as possible. With the arrangement shown in Fig. 2, the laser optical axis is subjected to errors caused by pitch angular movements. The influence of other angular movements is reduced to minimum with this arrangement. The Abbe offset for pitch direction is about 18 mm. This results in 13 µm (with 15 pitch error) position error for first 3.5 m measurement range of the system. After 3.5 m measurement range, maximum position error is estimated as 31 µm with 35 pitch error. To evaluate the estimated Abbe error values, some tests have been performed by varying the position of laser optics and keeping the linear encoder place constant. The graphs were plotted by taking the difference between the laser and the linear encoder readings. The difference between the results taking from two positions of laser optical axis in vertical direction indicates the influence of about 18 mm Abbe offset. These are shown in Fig. 8 and Fig. 9. Encoder-Laser (µm) a b c (a) (b) (c) Figure 8: Tests for Abbe offset values in pitch direction while small offset used for yaw direction Encoder-Laser (µm) a b c (a) (b) (c) Figure 9: Tests for Abbe offset values in pitch direction while large offset used for yaw direction The influence of the yaw motion can be seen from Fig. 1. After examining the test results with angular errors in Fig. 7, it can be said that the errors are repeatable and correlated with the angular errors and the Abbe offset values. 426 Proc. of SPIE Vol. 519

9 Encoder-Laser (µm) a b Figure 1: The influence of the yaw motion (a) (b) 4.3. Error compensation for linear encoder measurement The linear encoder is calibrated using the laser interferometer with intervals of 1 mm over 55 mm. The error correction file is determined using mean values of 3 sets measurements. The software uses this file to correct the readings taking from the linear encoder display. The corrected values of the encoder display are simultaneously shown on the monitor screen. For any position in the measurement range (-55 mm), the actual value used in the error correction is a linear interpolation between the two nearest points. The separation between these two nearest points is the calibration interval of 1 mm. The corrected readings are also checked with the laser interferometer. The results taken before and after calibration are shown in Fig. 11. Error (µm) 2 15 Before After Figure 11: Calibration of the linear encoder 4.4. Optical probing check The repeatability of setting the coincidence between crossed target and the defining line is checked by moving the X-Y table micrometer from different directions and distances. The standard deviation of less than 2 µm has been easily been achieved when the test have been performed while the carriage being on various places of the rails. 5. MEASUREMENT UNCERTAINTY The uncertainty of measurement has been estimated according to the ISO Guide for the expression of uncertainty in measurement 6 and has been expressed in a length dependent form of [(a mm) 2 + (b L) 2 ] 1/2 using a coverage factor of k=2. Here, a is the constant value, b is the length dependent value and L is the measured length. Analysis of the uncertainty contributions has been investigated in detail and has been combined in the above form. While it is not intended to reproduce all the details here, it is worth recording the components in short. Proc. of SPIE Vol

10 5.1. Length independent A major contribution is the error caused by carriage tilt. Considering that the Abbe offset for pitch direction is 18mm, 31 µm position error is estimated with 35 peak-to-peak pitch error (Sin [35/36] 18 mm). Although the laser measurement axis is located in the axis of optical probing unit, possible contribution of yaw error is included in the uncertainty budget. 5 mm misalignment of the laser measurement axis results in 7 µm position error due to 26 yaw error. These are considered as rectangular distributions and combined statistically in the uncertainty budget after being divided by square root of 3. Accuracy for optical probing of the scale marks is investigated by measuring the small line width of the stage micrometer. The deviation of the results has been found less than 1 µm and is considered as rectangular distribution. The repeatability test described in section 4.4. produced standard deviation of less than 2 µm and is included in the uncertainty budget. Repeatability for measurement has been determined by measuring a 5 m tape of OIML class II. Ten measurement results of each point gives maximum standard deviation of 1 µm. This is divided by square root of 3 (as 3 measurements are performed usually) and added to uncertainty budget. Uncertainty contribution for length measurement system of linear encoder consists of several parts. The major part is the contribution of laser interferometer and the calibration of linear encoder. When the linear encoder with error correction file is calibrated using the laser interferometer, the maximum deviation is in the band of 2 µm. The repeatability of this measurement is less than 1 µm. While the first is considered as rectangular distribution the second is taken as normal distribution. Resolution of.1 µm for the linear encoder and of.1 µm for the laser interferometer are also taken into account as rectangular distributions Length dependent Length measurement by laser interferometer results in about 3ppm uncertainty contribution being L. Besides, temperature measurement and expansion coefficient of linear encoder also contributes into uncertainty budget as length dependent part. Temperature measurement of the linear encoder is performed with a maximum uncertainty of.15 C. This gives rise to error of ( L) in terms of length (1/K) is taken as steel linear encoder expansion coefficient. Contribution due to value of expansion coefficient for steel linear encoder is also calculated. This value may vary and can be estimated with an uncertainty of (1/K). As the maximum allowable deviation of temperature from 2 C has been set at ±.5 C, the value in terms of length due to this uncertainty can be calculated by ( L). All parameters explained above are assumed to be rectangular distribution. After uncertainty contributions regarding to the linear encoder calibration, uncertainty analysis of the tape measurement for length dependent part is also performed. This is similar to temperature part explained in above paragraph. Uncertainty contribution of temperature measurement and temperature corrections for the tape and the linear encoder (once more for tape measurement) are calculated by ( L), ( L) and ( L), ( L) respectively Combined and expanded uncertainty The combined standard uncertainty, being the root sum square of the all uncertainty contributions, is calculated. The appropriate coverage factor can be taken as k=2 and the expanded uncertainty is determined by multiplying the combined uncertainty value by 2. The expanded uncertainty for tape measurement with linear encoder, which is expressed for confidence level of 95%, is given by U = [(.5) 2 +(.5L) 2 ] 1/2 mm, where L is measured length of the tape or rule. If this uncertainty is calculated using the laser interferometer as a length measurement unit, it is U=[(.4) 2 +(.4L) 2 ] 1/2 mm, where L is measured length of the tape or rule. 428 Proc. of SPIE Vol. 519

11 6. PERFORMANCE OF THE 5 M-MEASUREMENT SYSTEM AND DISCUSSION The performance of the 5 m-measurement system has been checked by measuring a 1 m gauge block. The gauge block has been placed on the tape support with 46 different interval positions so that the carriage should be positioned to perform measurements in the measurement range of -55 mm. The position of the carriage starts from -1 mm to mm with the intervals of 1 mm such as -1 mm, 1-11 mm, 2-12 mm,., mm, mm. Fig. 12 illustrates the results taken by measuring the 1 m gauge block along the range of -55 mm in 46 intervals. Each measurement result is the mean of 5 individual measurements. The standard deviations are also plotted around the results in order to give information about repeatability. The maximum error for linear encoder is found to be less than 23 µm and the maximum repeatability is 12 µm. It is 18 µm and 8 µm respectively when the laser is used. All results are within the uncertainty specification for 1 m lengths. It is apparent that the laser gives more repeatable results with lower deviations. Further checks are planned to be made using a calibrated good quality steel tape of 6 m. After studying the results, it has been founded that a major contribution is due to carriage tilt occurring on the places where the rails are nor very straight. This is due to use of extruded aluminum holders screwed in to the rails as a holder. The extruded aluminum holder disturbs the straightness of the centreless rail when it is screwed along the rail. The further work will be carried out by replacing current holders with short ones, which will be used only on the supporting points. Error (µm) Linear encoder Intervals (a) Error (µm) Laser Intervals (b) Figure 12: Performance check of the 5 m-measurement system using a 1 m gauge block (a) linear encoder, (b) laser Proc. of SPIE Vol

12 7. CONCLUSIONS 5 m-measurement system presented here enables the calibration of tapes and rules up to 5 m using a linear encoder and optionally using a laser interferometer. The estimated uncertainty for the system when used with the linear encoder is U = [(.5 2 )+(.5L) 2 ] 1/2 mm, where L is measured length of the tape or rule in meters. The performance tests carried out using a 1m gauge block gave promising results. Further tests will be carried out using a calibrated tape of 6 m. It is also planed that further improvement will be performed for tilting error of the carriage. ACKNOWLEDGMENTS The authors would like to thank Okan Ganioglu, Ilker Meral, Orhan Yaman and Nuray Karaböce for their valuable help and support. REFERENCES 1. OIML R 35 EN, Material measures of length for general use, International Organization of Legal Metrology 1985, /362/EEC, Material measures of length., European standards and directives, BIPM web site, 4. Rosenberg C.B., Munteanu C. S. C, and Ferguson R. A. Calibration of flexible tapes to ppm accuracy level, OIML Bulletin, 38, 25-29, 1997, 5. Euromet Length web site, 6. Guide to the expression of uncertainty in measurement, International Organization of Standardization (ISO), Geneva, Proc. of SPIE Vol. 519