XX IMEKO World Congre Metrology for Green Growth September 9 14, 2012, Buan, Republic of Korea A CALIBRATION SYSTEM FOR LASER VIBROMETERS AT NIMT C. Hirunyapruk, P. Rattanangkul, B. Thummawut, V. Plangangma National Intitute of Metrology (Thailand), NIMT, Bangkok, Thailand, chompoonoot@nimt.or.th Abtract: Laer Vibrometer have been ued in many application uch a automotive and hard dik drive indutrie. In metrological field, laer vibrometer are employed a reference tandard for the calibration of accelerometer. To have the traceability chain to the International Sytem of Unit (SI), everal National metrological intitute (NMI) and companie have developed calibration ytem for laer vibrometer. National Intitute of Metrology (Thailand), NIMT, i one of thoe NMI. Thi paper decribe the et-up for the primary vibrometer calibration ytem ued at NIMT and preent the method ued according to method 1 and 2 hown in the international tandard ISO16063-41 [1]. The calibration reult are alo dicued. Keyword: calibration, laer vibrometer, primary calibration 1. INTRODUCTION The development of laer vibrometer calibration ytem ha received increaing attention in many NMI. There have been everal tudie on thi development [2-8]. In mot tudie the ine approximation method wa employed to obtain the calibration reult. Ripper et al. [2] developed a homodyne quadrature interferometric ytem for the calibration of Laer Doppler Vibrometer (LDV). Thi ytem wa capable to calibrate LDV in the frequency range from 10 Hz to 10 khz. The calibration reult were obtained uing the ine-approximation method. Later, the new ytem ha been developed by Ripper et al. [3] to allow calibration of vibrometer with both analog and digital output. Buehn et al. [4] employed ine approximation method and ued the laer vibrometer a a reference tandard. The calibration reult were alo preented. Sine approximation method wa alo employed by Oota et al. [5]. A modified Michelon-type homodyne laer interferometer wa ued. Alo the effect of four parameter on the calibration reult were invetigated. Thoe were acceleration amplitude tability, mechanical ditortion, poition of meauring beam, and pot diameter of meauring beam. In [6], Oota et al. invetigated the effect of demodulator unit characteritic on the calibration of laer vibrometer uing two commercial laer vibrometer with analogue demodulator unit. G.S. Pineda et al. [7] evaluated the accuracy of the different tandard method hown in ISO 16063-41 [1] and concluded that method 1, method 2 and method 3 are applicable up to frequencie of 100 khz or higher if pecial vibration exciter generating ufficient diplacement or velocity are ued. Brun et al. [8] employed a dual frequency excitation for laer vibrometer calibration uing a modified Michelon interferometer a reference to calibrate laer vibrometer up to 90 khz. The aim of thi work i to develop the calibration ytem for laer vibrometer by modifying an exiting primary ytem for accelerometer calibration. The developed ytem employ a fringe-counting method for the calibration frequency between 40 and 800 Hz and ue a minimumpoint method for the frequency range from 1 to 5 khz. Alo the ytem i verified by comparing the calibration reult with thoe reported in SPEKTRA certificate [9]. Following thi introduction, the experimental et-up and the method to calibrate laer vibrometer are decribed. The calibration reult are preented in ection 3. The developed ytem i evaluated by comparing the calibration reult with thoe obtained from the calibration certificate iued by SPEKTRA. Finally, ection 4 contain ome concluion. 2. EXPERIMETAL APPARATUS AND METHODS Experiment were conducted to invetigate the performance of the developed ytem for laer vibrometer calibration. The photograph of thi developed calibration ytem i hown in Fig. 1. Fig. 1. Primary calibration ytem for the calibration of laer vibrometer.
FRINGE-COUNTING METHOD The experimental et-up for the fringe-counting method i hown in Fig. 2. In the experiment, a reflecting mirror wa connected to the top of a B&K exciter head model 4811. A B&K ine generator model 1051 generated an input ignal to drive a B&K exciter model 4805 through a B&K power amplifier model 2707. Michelon Interferometer, coniting of a beamplitter, a reference mirror, a photodetector and a Helium-Neon laer with wavelength of 632.8 nm, wa ued to meaure reference diplacement amplitude. The laer pot wa pointed to the middle of the reflecting mirror through a beamplitter 2. The photodetector output wa amplified by a B&K meauring amplifier model 2636. The amplifier output wa ent to ocillocope to be monitored during the laer beam adjutment procedure. For fringe counting method, the output from the amplifier model 2636 wa compared in frequency with the excitation frequency from the ine generator by uing an Agilent univeral counter model 53131A. The diplacement amplitude,, and the acceleration amplitude, a, generated by reference tandard can be calculated by uing equation (1) and (2) repectively [1]. = R (1) a = (2πf) (2) where λ i the wavelength which i 632.8 nm. Here R i the ratio of the fringe frequency f come from the photodiode ignal to the vibration frequency (f) generated by the ine generator. A Polytec laer vibrometer model CLV-1000 with a decoder model CLV-M050 wa ued a a device under tet (DUT). A ame a the beam generated by the reference homodyne laer, the laer pot from the DUT wa pointed to the reference mirror through the beamplitter 2 a hown in Fig 2. Both laer beam, which were generated from DUT and reference laer, were pointed to the ame poition on the mirror. The meaured analogue voltage from the DUT wa ent to an Agilent digital multimeter model 3458A and the ocillocope. The output from the digital multimeter wa ent to PC for analyi. The acceleration generated by DUT can be determined from the meauring amplifier wa analyzed to find the minimum point during Beel meaurement uing a B&K ignal analyzer model 2035. A the helium neon laer with λ = 632.8 nm wa ued, the reference diplacement amplitude for the minimum point were elected from Table 9 in [1] and the reference acceleration wa calculated uing equation (2). Alo the acceleration of the DUT wa determined from equation (3). 3. CALIBRATION RESULTS In thi ection, the calibration reult uing ytem hown in Fig. 1 are preented. In order to verify thi developed ytem, the reulted are compared with thoe obtained from the certificate iued by SPEKTRA. The calibration method pecified in SPEKTRA certificate i the primary calibration method according to ISO 16063-11. The SPEKTRA calibration wa performed uing a SPEKTRA vibration exciter model SE-09, an APS vibration exciter model 113AB, a Polytec tandard laer vibrometer model CLV1000 and a SPEKTRA calibration ytem model CS18 HF-DKD. Both tandard laer vibrometer and vibrometer under tet meaured velocity at the ame point on the haker urface. The calibration wa done by comparing the acceleration of the vibrometer under tet with that of the tandard vibrometer [9]. Fig. 2. Experimental et-up for laer vibrometer calibration uing fringe-counting method. a = (2πf)V R (3) where V and R are the analogue voltage output from the DUT given in V and the meaurement range in mm /V repectively. The DUT ued here ha three meaurement range, which are 2 mm /V, 10 mm /V and 50 mm /V. The experiment hown in thi paper were conducted for all three range. MINIMUM POINT METHOD The et-up for the minimum point method i hown in Fig. 3. Thi et-up i imilar to that for the fringe-counting method. But for the minimum point method, the output from Fig. 3. Experimental et-up for laer vibrometer calibration uing minimum point method.
Fig. 4. Deviation of acceleration amplitude from the reference tandard for the range of 2 mm /V: - - meaured by NIMT and obtained from SPEKTRA certificate. Fig. 7. Difference between acceleration deviation meaured the range of 2 mm /V. Fig. 5. Deviation of acceleration amplitude from the reference tandard for the range of 10 mm /V: - - meaured by NIMT and obtained from SPEKTRA certificate. Fig. 8. Difference between acceleration deviation meaured the range of 10 mm /V. Fig. 6. Deviation of acceleration amplitude from the reference tandard for the range of 50 mm /V: - - meaured by NIMT and obtained from SPEKTRA certificate. Fig. 9. Difference between acceleration deviation meaured the range of 50 mm /V.
Fig. 4, 5 and 6 preent the comparion reult between NIMT and SPEKTRA certificate for the meaurement range 2 mm mm /V, 10 /V and 50 mm /V repectively. The vertical value hown in thoe figure are the deviation of acceleration amplitude meaured by reference laer and laer under tet. Thi deviation i given in % and calculated from 100. (4) Fig. 10. EN Ratio between the reult obtained from NIMT and SPEKTRA certificate for the meaurement range of 2 mm /V. In Fig. 4-6, (- -) repreent the deviation of acceleration amplitude meaured by NIMT, while ( ) how the deviation obtained from SPEKTRA certificate. The reult obtained by NIMT were calculated a the arithmetic mean of three repeated calibration performed on different day. It can be een that the deviation of acceleration amplitude obtained by NIMT are higher than thoe reported by SPEKTRA. Thi occur for the whole frequency range and for all three meaurement range. Fig. 7-8 how the difference between acceleration deviation meaured by NIMT and that obtained from SPEKTRA certificate. Thi difference i determined from D D (5) Fig. 11. EN Ratio between the reult obtained from NIMT and SPEKTRA certificate for the meaurement range of 10 mm /V. where D and D are the deviation of acceleration amplitude meaured by NIMT and SPEKTRA repectively. It can be een from Fig. 7 that the maximum difference between NIMT reult and SPEKTRA reult for the meaurement range of 2 mm /V occur at frequency 80 Hz. However, it i found in Fig. 8-9 that the highet deviation for the frequency range of 10 mm /V and 50 mm /V preent at 125 Hz. Thee maximum deviation are -0.15%, -0.13% and -0.13% for the meaurement range of 2 mm /V, 10 mm /V, and 50 mm /V repectively. In order to illutrate the agreement between the calibration reult from NIMT and SPEKTRA, EN ratio wa calculated uing the following equation. E = (6) Fig. 12. EN Ratio between the reult obtained from NIMT and SPEKTRA certificate for the meaurement range of 50 mm /V. where U and U are the expanded uncertainty of meaurement declared by NIMT and SPEKTRA repectively with the coverage factor of 2. Here U i 0.25% for frequencie up to 5 khz and U i 0.3 % for the frequency range between 40 and 800 Hz and i 0.7 % for the frequency range from 1 to 5 khz. Fig. 10-12 preent the EN ratio of the reult between NIMT and SPEKTRA certificate for the meaurement range of 2 mm mm /V, 10 /V, and 50 mm /V repectively. It i found that the EN ratio i between 0.08 and 0.38 approximately. Again, the maximum EN Ratio occur at frequency 80 Hz for the meaurement range 2 mm /V and at 125 Hz for the range of 10 mm /V, and 50 mm /V.
4. CONCLUSIONS The primary calibration ytem ued for laer vibrometer calibration at NIMT ha been preented in thi paper. The ytem i capable of calibrating laer vibrometer in the frequency range from 40 Hz to 5 khz. The method that thi ytem ue are a fringe-counting method for the calibration up to 800 Hz and a minimum point method for the frequency range between 1 khz and 5 khz. The calibration reult are determined in term of the deviation of acceleration amplitude meaured by the tandard heterodyne reference and by the vibrometer under tet. The calibration of a Polytec laer vibrometer model CLV-1000 with a decorder model CLV-M050 were performed and the calibration reult were compared with thoe reported in SPEKTRA certificate to verify the calibration ytem developed at NIMT. The calibration reult how the reaonable agreement between the reult preented by NIMT and SPEKTRA certificate. Although the deviation of acceleration amplitude reported by NIMT were larger than SPEKTRA, the highet difference between NIMT and SPEKTRA wa atifactory, which wa about -0.15 % at 80 Hz. Alo the EN ratio howed acceptable agreement between the reult from NIMT and SPEKTRA certificate. The maximum EN ratio wa 0.38 approximately. 5. REFERENCES [1] ISO 16063-41:2011, Method for the calibration of vibration and hock tranducer Part 41: Calibration of laer vibrometer (Geneva: International Organization for Standardization). [2] G. P. Ripper, G. A. Garcia, R. S. Dia and M. Zindeluk, Primary calibration of ingle-point laer Doppler vibrometer (LDV), in Environmental Noie Control The 2005 Congre and Expoition on Noie Control Engineering, Rio de Janeiro, Brazil, 2005. [3] G. P. Ripper, G. A. Garcia and R. S. Dia, The development of a new primary calibration ytem for laer vibrometer at inmetro, in IMEKO 20th TC3, 3rd TC16 and 1t TC22 International Conference Cultivating metrological knowledge, Merida, Mexico, 2007. [4] U. Buehn, H. Weiing and G. Siegmund, Calibration of laer vibrometer tandard according to ISO16063-41, in XVIII IMEKO WORLD CONGRESS Metrology for a Sutainable Development, Rio de Janeiro, Brazil, 2006. [5] A. Oota, T. Uuda, T. Ihigami, H. Nozato, H. Aoyama, and S. Sato, Preliminary implementation of primary calibration ytem for laer vibrometer, Proc. SPIE 6345 (Proc. of the 7th International Conference on Vibration Meaurement by Laer Technique: Advance and Application), 634503-1 to 8, 2006. [6] A. Oota, T. Uuda, T. Ihigami, H. Nozato, and Y. Hino, Effect of demodulator unit on laer vibrometer calibration, Proc. of SPIE Vol. 7098, 70981J, 2008. [7] G.S. Pineda, H.-J. von Marten, S. Roja, A. Ruiz and L. Muniz, Calibration of laer vibrometer at frequencie up to 100 k Hz and higher, Proc. of SPIE, vol. 7098, 70981K, 2008. [8] T. Brun, F. Blume and A. Täubner, Laer vibrometer calibration at high frequencie uing conventional calibration equipment, in XIX IMEKO World Congre Fundamental and Applied Metrology, Libon, Portugal, 2009. [9] SPEKTRA Schwingungtechnik und Akutik GmbH Dreden, Certificate no. 0466, dated 04/05/2011.