Calibration of piezoelectric accelerometers at INTI

Transcription

1 PROCEEDINGS of the 22 nd International Congress on Acoustics Acoustical Measurements and Instrumentation: Paper ICA Calibration of piezoelectric accelerometers at INTI Alexis Zapata (a), Ramiro Benevenia (b), Lucía Taibo (c) (a) INTI - Instituto Nacional de Tecnología Industrial, Argentina, gzapata@inti.gob.ar (b) INTI - Instituto Nacional de Tecnología Industrial, Argentina, ramirob@inti.gob.ar (c) INTI - Instituto Nacional de Tecnología Industrial, Argentina, luciat@inti.gob.ar Abstract The National Institute of Industrial Technology, INTI, is the technical referring of the argentinian state in the assistance for the industrial development in the country. Besides, is the National Institute of Metrology, (NMI) due to its legal responsibility on the national measurement standards. The SI units of physical quantities, such as meter per squared second (m/s²) for the quantity of acceleration, is realized in the Vibrations Laboratory, UT Acoustics of INTI,and disseminated to external customers through calibrations of transducers and equipment. In former times, INTI participated in the first intercomparison in acceleration that took place in America within the SIM (Interamerican System of Metrology) with the participation of the NMIs of Canadá, EEUU and México (NORAMET) and Brasil and Argentina (SURAMET). The frequency range comprised 50 Hz to 5000 Hz. The system for the primary calibration of accelerometers is a simple Michelson laser interferometer with a single detector and an electrodynamic shaker according to the guidelines given in ISO In the present work, recent improvements in the method are described, which allowed to extend the frequency range from 10 Hz to Hz. The charge sensitivity results of a laboratory standard accelerometer B&K 8305 calibration are given, and a good correspondence with the values stated by INMETRO in a previous report was observed. Accordingly, a bilateral SIM comparison will be encouraged in order to validate the current CMCs of INTI. Keywords: vibration, vibration transducer, accelerometer, primary calibration, laser interferometry.

2 Calibration of piezoelectric accelerometers at INTI Introduction Primary calibration of piezoelectric accelerometers by laser interferometry began at INTI during the past 80`s decade before this method became an ISO standard, with the development of a low-cost arrangement that underwent continual improvements up to present times. In 1999, the first comparison on the calibration of vibration standards in America, was officially carried out under the framework of the Interamerican Metrology System (SIM), joining two subregions: NORAMET (Canada, Mexico, USA) and SURAMET (Argentina, Brazil). The participating laboratories were the National Metrology Institutes of five american countries, i.e., National Institute of Standards and Technology (NIST - U.S.A.), National Research Council (NRC - Canada), Centro Nacional de Metrología (CENAM - Mexico), Instituto Nacional de Metrologia, Normalização e Qualidade Industrial (INMETRO - Brazil) and Instituto Nacional de Tecnología Industrial (INTI - Argentina). Interferometric calibrations of three standard accelerometers and charge calibration of a signal conditioner were performed by each of the NMI s. The overall results obtained were highly satisfactory [1], [2]. INTI has declared CMCs on acoustics and vibration at the BIPM data base and holds ISO/IEC accreditation, sustained by peer review from INMETRO (Brazil) and PTB (Germany). 2. Primary interferometry calibration of vibration transducers 2.1 Evolution of the primary interferometry calibration system The primary calibration system of accelerometers developed at INTI is a homodyne Michelson interferometer with a single detector and an electrodynamic shaker. In the initial arrangement [3], both the optical paths and the motion of vibrator were in the horizontal direction. The optical system and the shaker were mechanically decoupled from each other, resting on two independent inertia blocks supported by vibration isolators. In 2002, a thorough revision of the system was undertaken on account of the new international standardized requirements [4], and the need to overcome the drawbacks in the alignment of the optical elements in the initial set-up. The horizontal interferometer was replaced by the vertical design that is currently in use, being a comparatively simpler arrangement, with reduced dimensions and lower installation costs, inspired in a version developed at the Kobayashi Institute in Japan [5]. The new set-up comprised two independent mechanical mountings to support: i. The optical components: an optical table was self-constructed according to the restricted budget available. It was made up with a heavy rectified granite plate, supported by four active commercial selflevelling vibration isolators (tolerance: +/-0,25 mm, vertical resonance frequency <1,35 Hz, vertical transmissibility at 5 Hz <0,1). The optical supports were directly glued to the granite plate. ii. The vibration shaker: it was mounted on a steel block, supported by four springs (natural 2

3 frequency <1,5 Hz, 90 % rejection frequency 4,3 Hz in the vertical axis). The mounting was designed to control the vibrations produced by a set of rotating machinery located at the basement of the Physics building, that affected the calibration laboratory. The Vibrations Laboratory at INTI was peer reviewed by INMETRO, Brazil, in 2004, 2009 and 2015, achieving the accreditation by ISO and supporting the CMCs Appendix C of the BIPM Key comparison database (KCDB) in the primary accelerometer calibration by laser interferometry and the accelerometer calibration by comparison [6]. 2.2 Current primary interferometry set-up and improvement measures A block diagram of the system is shown in Figure 1. The laser emits a single frequency light beam, F1, that hits the beam splitter which sends half of the light to a fixed mirror that reflects the same frequency, F1, << the reference path>>, and the other half, to the moveable element (accelerometer or base plate), which reflects a Doppler shifted frequency F1±Δf, <<the measuring path>>. Both beams are recombined and interfered together at the photodetector to give a beat frequency of zero for the stationary condition, whilst the beat frequency rises as the optics move in either direction. Figure 1: Block diagram of the apparatus A view of the measuring apparatus in the vertical arrangement is shown in Figure 2. 3

4 Figure 2: Primary Vibration Calibration Set-up at INTI The reference light path is horizontal, whilst the measuring path points vertically downward after the beam splitter, passing through a hole drilled in the optics table. Recombined beam Photodetector Reference path Beam splitter LASER He-Ne Measuring path Laser beam Figure 3: Close view of the optical arrangement In order to improve the primary interferometry calibration system a set of measures were implemented over time. The main achievements are summarized as follows: 4

5 Acquisition of specific instruments, such as a new B&K 4808 shaker and an Agilent 33210A sine generator, in order to replace old equipment. More stable signals were generated and a reduction of the systematic errors was achieved. A rectified aluminum plate with threaded holes of regular pattern was fixed to the upper side of the granite table, allowing an easier positioning of the optical parts instead of gluing them to the granite table. Besides, new optical components, such as mirrors, mechanical supports and a beam splitter cube, were acquired. The elastic mounting of the shaker produced high spring deflections and long stabilization times after a change of the accelerometer position. Besides, as the rotating equipment affecting the calibration laboratory was finally dismantled, the vibration isolated mounting was no longer required. The elastic supports of the inertia block were then short-circuited by a rigid connection to the floor, and correspondingly, the optic alignment became easier and less time consuming. A new standard B&K 8305 accelerometer has been acquired. In 2014, its primary calibration according to ISO (Method 3) was performed at INMETRO. This transducer became in the primary standard of the vibrations laboratory at INTI, replacing the 15 years old B&K 8305 accelerometer, which is currently used as a working standard. A new procedure for the calibration of charge preamplifiers was implemented (PEA15), improving the uncertainty budget, and extending as well the frequency range to 10 Hz up to Hz. Most recently, additional measures have been carried out in order to improve the primary calibration set-up: The light beam reflected on the lapped top of the new standard accelerometer is sufficiently intense to be sensed by the photodetector, and consequently, the optical mirror attached to the top of the old standard accelerometer by means of a threaded adaptor, is no longer needed. Accordingly, a significant source of error has been reduced. A more precise torque meter was acquired in order to control the transducers mounting according to the manufacturers optimal values. An overall better repeatability of the method was achieved. In order to rotate the double-ended standard accelerometers under calibration, new mechanical adaptors were constructed. A stainless steel supporting plate was built in order to fix the single-ended transducers, and reflect the light beam on the exposed side. A careful process for the metallic surface lapping was undertaken in order to fulfill a high degree of flatness, face parallelism and surface roughness. 5

6 Single-ended accelerometer Aluminium base plate Stainless steel plate (lapped) Moving element Figure 4: Single-ended accelerometer mounting for primary calibrations As described in [7], the imperfect behavior of the vibration exciters affects the primary accelerometer calibration, introducing errors due to different factors, such as the rocking and transverse motions of the moving element. In fact, since most accelerometers exhibit a transverse sensitivity due to a misalignment of the maximum sensitivity axis, the coupling of the shaker s transverse motion and the accelerometer s transverse sensitivity creates an error in the sensitivity determination. These effects were experimentally confirmed, observing that the lateral movements of the shaker largely affect the calibration process in certain frequencies. Following the suggestions stated in the cited reference, measurements were taken at four points on the reference surface, i.e. top of the double-ended accelerometer (or on the lapped mirrored plate for the single-ended transducers). The final charge sensitivity curve corresponds to the mean result of the four positions shown in Figure 5, for each measured frequency. This method was introduced in the calibration procedure (PEA12), since averaging allows to cancel the high irregular peaks observed in each individual point. Figure 5: Laser beam at the double-ended standard accelerometer reference surface (top) The same schedule was followed for measurements on the mirrored support plate of the singleended transducers. 6

7 2.3 Primary calibration of the reference standard accelerometer The charge sensitivity of INTI reference standard accelerometer, a new double-ended B&K 8305, has been primary calibrated at INMETRO, Brazil, according to ISO , Method 3. A time after, the same transducer was primary calibrated at INTI according to ISO , Methods 1 and 2, guidelines, using the system described along the present work. Charge sensitivity results are given correspondingly, and as a first attempt of validation INTIs proficiency is assessed based on a widespread standardized criterion [8] At INMETRO The primary calibration was performed according to ISO , Method 3_Sine Approximation, using a primary quadrature interferometric system that covers the frequency range from 10 Hz to Hz (DIMCI 2104/14, INMETRO, ). The declared expanded uncertainty with the coverage factor of k=2 is: ± 0,25% for 10 Hz to 2500 Hz, ± 0.35% for 3000 Hz to 4500 Hz, ± 0.60% for 5000 Hz to 6000 Hz and ± 0.70% for 7000 Hz to Hz At INTI The primary calibration was performed according to ISO , Method 1_ Fringe-Counting Method_ carried out from 10 Hz to 1000 Hz and Method 2_Minimum Point Method_ from 1000 Hz to Hz, covering the whole range from 10 Hz to Hz. At 1000 Hz the stated value is the average of both methods. Mounting conditions Table 1: Measuring conditions at INTI Mechanical adaptor LV120A Tightening torque of adaptor LV120A Tightening torque of the accelerometer Lubrication between mounting surfaces Ambient conditions Aluminium plate 1,50 ± 0,02 Nm 2,00 ± 0,02 Nm Silicon grease Ambient temperature 22,8 ºC to 24,3 ºC Relative humidity 29% to 52% Interferometric measurement of movement On the reference surface of the accelerometer (top), 4 positions according to Figure 5. The declared expanded uncertainty with the coverage factor of k = 2 is: ±1,1% for 10 Hz < 200 Hz, ±1,3% for 200 Hz to 1 khz, ±1,5% for >1 khz to 10 khz. 7

8 [pc/(m/s 2 )] 0,1418 0,1406 0,1394 0,1382 0,1370 0,1358 0,1346 0,1334 0,1322 0,1310 0,1298 0,1286 0,1274 0,1262 Primary Calibration of B&K 8305 Standard Accelerometer Charge Sensitivity according to ISO INMETRO (Method 3, ISO ) INTI (Method 1, ISO ) INTI (Method 2, ISO ) 0, Frecuency [Hz] Figure 6: Charge sensitivity frequency dependence of the standard accelerometer 2.4 Comparison of results between INMETRO and INTI Charge sensitivity results obtained INTI were compared to INMETRO values, considering this last laboratory as the reference one. A widespread method to evaluate the results of a proficiency test, the comparison to the reference value for a measurand of a result reported by a participating laboratory is made by calculating the En value [8]. The En value is calculated as follows: E n = M Inti M Ref 2 2 U Inti + U Ref (1) where: M Inti and M Ref M are the measured charge sensitivity of standard accelerometer done by INTI and INMETRO, respectively, and, U Inti and U Ref are the expanded uncertainty of measurement declared by INTI and INMETRO, respectively, with the coverage factor = 2. The uncertainties are expressed for k=2, which corresponds to a 95 % coverage. A result is considered successful, if the value of the proficiency test is 1 En 1. The En-values obtained 8

9 are given in Table 3. In Table 2, the individual charge sensitivity results of the two laboratories are given. Table 2: Charge sensitivity results of INTI and INMETRO Frecuency (Hz) B&K 8305 Standard Accelerometer Charge Sensitivity, Sp (pc/ms -2 ) INTI INMETRO Frecuency (Hz) B&K 8305 Standard Accelerometer Charge Sensitivity, Sp (pc/ms -2 ) INTI INMETRO 10 0, , ,5 0, , , , , , , , ,5 0, , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , ,12980 The declared expanded uncertainties of the laboratories are given in and Table 3: Calculated En-values Frecuency (Hz) En-Values Frecuency (Hz) En-Values 10 0,04 12,5 16-0,03 0, ,02 0,02 31,5 40-0,01 0, ,03 0, ,03-0, ,01 0, ,00-0, , , ,10-0, ,09-0, ,04 0, ,03 0, ,02 0, ,01-0, ,32 0,15 As shown in Table 3, INTI results agree with the reference values within the stated uncertainties of the two laboratories. The largest En-values were obtained in 8000 Hz: -0,32, 6300 Hz: -0,27 and 1000 Hz: 0,27, that are comprised between -1 and +1. Based on a number of assumptions, this is expected to happen at least in the 95 % of the time. 9

10 Conclusions The primary vibration calibration system at INTI based on a low-cost arrangement has underwent successive improvements in order to fulfill ISO guidelines. In the present work, the charge sensitivity of a standard accelerometer was measured at INTI and INMETRO, Brazil by the primary interferometric method. Considering INMETRO as the reference laboratory a proficiency testing of INTI results has been conducted. According to this analysis, the charge sensitivity calibrated by INTI is regarded as quite satisfactory. The obtained results allow an extension of the current frequency range to 10 Hz to Hz, as well as a reduction in the stated uncertainties, with an estimate of <0,8% for the whole frequency up to 5000 Hz, and about 1% in the upper frequency range. Anyway, an official bilateral comparison within the SIM frame will be promoted, in order to validate an upgrade of the CMCs at the BIPM data base, thus extending the vibration calibration capabilities at INTI. Acknowledgments The authors wish to thank to Dr. Gustavo Ripper of INMETRO, Brazil, for his expert advice and encouragement, and Lic. Karina Bastida and the Optics team at the Physics and Metrology Center of INTI, for their guidelines and assistance during the optical system mounting. References [1] Silva Pineda, G.; Payne, B.F.; Ripper, G.P.; Wong, G.S.K.; Taibo, L.N. Acceleration comparison SIM.AUV.V-K1 Final Report, 26, September [2] Taibo, L.; Barceló, L. INTI at the first interamerican comparison on acceleration within the SIM Frame, 4º Jornadas de Desarrollo e Innovación, Buenos Aires, [3] Taibo, L.; et al. Implementación de Técnica Interferométrica para Calibración de Acelerómetros, 2das. Jornadas de Desarrollo e Investigación INTI, Buenos Aires [4] ISO, International Standard : Methods for the calibration of vibration and shock transducers Part 11: primary vibration calibration by laser interferometry, Geneva, [5] Yokota, A.; Komura, H. Comparison calibration of vibration pickup by using transfer function and new calibration methods of multiaxial vibration pickup, Journal Acoustical Soc. Japan 13, 5 (1992). [6] Ripper, G. P. Peer review reports for the Laboratory of Vibrations, INTI, Argentina, INMETRO, October 2004, October 2009, February [7] Ripper, G. P.; Dias, R. S.; Garcia, G. A. Primary accelerometer calibration problems due to vibration exciters. Measurement, Vol 42, 2009, pp [8] International Standard ISO/IEC Guide 43-1, Proficiency testing by interlaboratory comparisons Part 1: Development and operation of proficiency testing schemes,