Comparison calibration of a heavy tri-axial velocity transducer

The specified quality of a vibration exciter may be strongly affected by parasitic transverse motions occurring as a result of the load exerted by the object to be calibrated. Mass, size, stiffness and mechanical coupling of the sensor have an influence on the quality of the desired - usually uniaxial - vibration motion. Following a request of METAS (Switzerland), a rather large and heavy tri-axial velocity transducer for their local calibration device was calibrated on PTB's multi-component vibration exciter which is designed for higher loads. By laserinterferometric determination of the reference and interference quantities it could be shown that the influence of parasitic rotational motions can, to a large extent, be compensated computationally.

For the secondary calibration of vibration transducers by means of a vibration exciter, the transducer sensitivity is determined from the simultaneously measured signals of the test object and a reference transducer. Thereby, a stiff mechanical coupling with identical motion sequences of the two sensors is required. In this regard, ISO 16063-21 [1] specifies different requirements for the calibration device and the measurement uncertainties that can be achieved with it. Whereas in the case of small light sensors on heavy vibration exciter armatures the disturbing transverse motions remain within acceptable limits, this is no longer the case for comparatively large and heavy transducers.
The transducer investigated was the tri-axial velocity transducer MS2003+ of the Swiss company SYSCOM. This transducer is frequently used for vibration measurements in the building industry. The transducer (mass: 1.55 kg, dimensions: 12x12x8 cm³) is specified for the relatively wide frequency range from 1 Hz to 350 Hz and for a dynamics of more than 130 dB.
At the Swiss metrology institute METAS, the transducer was calibrated on a ball-bearing- mounted low-frequency vibration exciter (Shaker APS-113, SPEKTRA Schwingungstechnik und Akustik GmbH, Dresden) with two different transducer adaptations. The moving armature of the vibration exciter and the test object had almost the same weight. It showed that the first, originally envisaged, transducer adaptation with its displacement between the exciter force vector and the centre of gravity of the moving masses excites pronounced tilting vibrations which considerably affect the quality of the uniaxial vibration excitation. Whereas the parasitic transverse accelerations are still negligibly small in the case of low frequencies - and large vibration displacements - they reach, above approx. 40 Hz, the amount of the nominal component and - partly - even dominate. Such a disturbed vibration excitation is not suitable for a calibration. The then optimized transducer adaptation (see Figure 1), in the case of which the driving force-vector (in x-direction) now acted onto the centre of mass of the moving structure, allowed the transverse accelerations a_y and a_z to be reduced by about one order of magnitude. However, the vertical component a_z still exceeds the permitted relative maximum value of 10 % over wide ranges (see Figure 2). To identify rotational disturbances, the nominal movement of the transducer housing was determined at several points by means of a laser vibrometer. The computational evaluation showed that the x-sensitivity of the tri-axial velocity transducer exhibits an amplitude response which is flat up to 200 Hz if the velocity traced by the vibrometer at the height of the transducer’s measuring axis is used as the reference quantity.

Figure 1: Calibration on the vibration exciter of METAS using the optimized adaptation of the device under test (DUT).

Figure 2: Platform movements on the METAS vibration exciter for the optimized adaptation. Acceleration components a_x, a_y, a_z (measured with the tri-axial acceleration transducer PCB 356B08), nominal displacement d_x (measured with reference acceleration transducer Honeywell QA2000).

To check this calibration result which has been obtained with a non-ideal shaker, a comparison calibration was performed on PTB's multi-component acceleration standard measuring device which is designed for test loads up to 100 kg (see Figure 3). This measuring device uses three orthogonally arranged shakers which drive the vibration armature via a cross-coupling unit with hydrostatic bearings. With the now clearly weaker parasitic movements, the flat amplitude response of the x-sensitivity could be impressively confirmed up to 300 Hz (see Figure 4). The diagram shows a comparison between the amplitude responses measured directly on the different measuring devices and those having been corrected with respect to the tilting movement. The activities described here, which have been carried out in cooperation with METAS, were presented at the IMEKO TC3 Conference 2007 [2].

Figure 3: Calibration on the multi-component acceleration standard measuring device of PTB.

Figure 4: Comparison of the calibration results obtained with the two different measuring devices: Amplitude response of the x-sensitivity; shown are the direct responses and the responses corrected by means of additional laser vibrometer signals.

[1] ISO 16063-21:2003, ”Methods for the calibration of vibration and shock transducers — Part 21: Vibration calibration by comparison to a reference transducer”, International Organization for Standardization, Geneva, 2003.

[2] Chr. Hof, M. Kobusch: Comparison of the calibration of a heavy multi-component vibration transducer on different exciter systems, Proc. of IMEKO TC3 & TC16 & TC22 International Conference, 2007, Merida, Mexico, CD-ROM and www.imeko.org

Contact person:

M. Kobusch, Department 1.3, WG 1.34, E-mail: michael.kobusch@ptb.de