owrap>Yahoo! is not affiliated with the authors of this page or responsible for its content.
Broadband Magnetostrictive Shaker 1 David L. Hall and Alison B. Flatau Broadband Performance of a Magnetostrictive Shaker Broadband Magnetostrictive Shaker 1 David L. Hall and Alison B. Flatau
Broadband Performance of a Magnetostrictive Shaker
David L. Hall* and Alison B. Flatau
Department of Aerospace Engineering and Engineering Mechanics
Iowa State University
Ames, Iowa 50011, USA
*Author to whom correspondence should be sent, 2019 Black Engineering Bldg, (515) 294-0088
Resubmitted (18 Aug., 1994) to: Journal of Intelligent Material Systems and Structures
ABSTRACT: Performance data from a magnetostrictive broadband (100 Hz-10 kHz) vibration source
is presented. An original design for a magnetostrictive transducer was built and tested. The transducer
used a rod of ETREMA Terfenol-D (EDGE Technologies, Ames, Iowa) as the motion source (rod
dimensions: 51 mm long x 6.35 mm diameter). This communication will convey the characteristics of
magnetostrictive transducer behavior. The shaker is considered to be a single input (electric current)
single output (force) system. For low signal operation it behaves in a linear fashion; output force is
proportional to input electric current and harmonic content is low. Comparisons with a commercially
available permanent magnet shaker show that the magnetostrictive shaker's linear region extends to
high enough forces (or displacements) to be of use as a vibration excitation source. Higher input
currents eventually lead to degradation of the magnetostrictive shaker's linear behavior; output
harmonics become appreciable, force levels increase disproportionately and, of course, linear systems
analysis techniques become inappropriate. The magnetostrictive shaker discussed is capable of
producing peak acceleration amplitudes in excess of 1960 m/s
2
(200 g's, at frequencies greater than 4
kHz with a 24 gram load using sinusoidal excitation). Experimental results are presented for: 1)
output force as a function of frequency, load, and current amplitude (30 Hz-10 kHz with sinusoidal
excitation); 2) typical electrical impedance as a function of frequency; 3) frequency response
functions as Newtons per ampere, Newtons per volt, and meter per ampere; and, 4) performance
comparisons made with a commercially available permanent magnet shaker. Experimental results 2, Broadband Magnetostrictive Shaker 2 David L. Hall and Alison B. Flatau
3, and 4 are for operation of the magnetostrictive shaker in its linear range (linear in a least squares
sense).
INTRODUCTION TO MAGNETOSTRICTIVE MATERIALS AND ACTUATORS
Magnetostrictive materials strain in response to magnetic fields. Terfenol-D is a
magnetostrictive material composed of rare-earths, terbium and dysprosium, alloyed with iron (Ter =
terbium, fe = iron, nol = Naval Ordinance Laboratory, where the material was first developed, and D =
dysprosium). Terfenol-D's utility arises from its giant strains, 1000-2000 x 10
-6
m/m, due to
readily attainable magnetic field amplitudes, i.e., typically less than 1000 Oersted 79.6 x 10
3
amp/meter.(Butler, 1988) Terfenol-D was the magnetostrictive material selected for use in the
transducer design to be discussed.
When Terfenol-D is produced in rod form, its small "oblong" magnetic domains are
intentionally oriented with the longer axis perpendicular to the rod's longitudinal axis. Application of a
magnetic field along the longitudinal axis of the rod,
in either direction , causes the magnetic domains to
rotate so that their longer axis is parallel with the rod's. As this occurs, the rod gets longer (and
smaller in diameter).
A few items of general interest about Terfenol-D warrant discussion. First, it tends to deliver
larger net strains if it is compressively prestressed. When the material is manufactured in rod form
this prestress acts to rotate more magnetic domains further away from an axial orientation. The net
result is the possibility of increased strain with the application of a magnetic field. Second, its
compressive and tensile strengths are approximately 700 and 28 MPa, respectively.(Butler, 1988)
The material should not be drilled or threaded since it can only be described as "brittle." Third,
Terfenol-D filings are flammable.
Fourth, and most significant from an actuator design standpoint, Terfenol-D strain as a function
of applied magnetic field displays hysteresis, owing to magnetic hysteresis within the material. A
sketch of this hysteresis is shown in Figure 1. Note, both positive and negative magnetic fields result Broadband Magnetostrictive Shaker 3 David L. Hall and Alison B. Flatau
in positive strain (the rod increases in length). To allow for bidirectional motion, a rod is typically
magnetically biased to Ho (via an external permanent magnet or a DC current in the coil). The field
from an electric current in a surrounding wire coil, or solenoid, can then be used to control the net
magnetic field within the Terfenol-D rod. (Recall: H nI, where n is the number of turns per unit
length of the solenoid and "I" is the electrical current passed through the solenoid.)
An oscillating current in the solenoid causes the Terfenol-D rod length to oscillate.
Unfortunately, from a linear systems point of view, while oscillating the material must traverse the
hysteresis loop depicted in Figure 1. Harmonics present within the output acceleration frequency
spectrum of Terfenol-D transducers are due in part to this hysteresis. The relative amplitudes of the
resulting harmonics are primarily a function of how hard the transducer is being driven. More on this
topic is included with the results discussion.
Traditional uses of Terfenol-D transducers include positioners (Miller, 1991) and sonar
projectors (500 - 2000 Hz). Additional work has shown its applicability to use in isolators (5 - 60
Hz, (Hiller et al., 1989)) and mounts (single sine "shock" attenuation (Reed, 1988)). In a recent
project, a Terfenol-D transducer was designed, built, and tested for use in a study of tissue response to
internal vibrations.(Hall, 1991) That transducer was designed to mimic the vibrations of an artificial
heart from 300 Hz to 10 kHz.
THE MAGNETOSTRICTIVE TRANSDUCER
Figure 2 is a section assembly drawing of the Terfenol-D magnetostrictive transducer used in
this investigation. The external geometry of this design was chosen to suit its intended use as a general
laboratory vibration source. It will also serve as a test fixture for measuring the performance of
different Terfenol-D rods (h), springs (e), prestresses (via (i) and (e)), air gaps (between (c) and
(a)), and H
o
values (adjustable using the external solenoid, (g)). Broadband Magnetostrictive Shaker 4 David L. Hall and Alison B. Flatau
A constant magnetic field is provided to the Terfenol-D rod by the cylindrical permanent
magnet, (f), and by a DC current in the external wound wire solenoid, (g). An AC electric current
provided to the internal solenoid results in an oscillating magnetic field along the longitudinal axis of
the transducer. The external solenoid and permanent magnet's fields provide the offset H
o
indicated in
Figure 1, biasing the Terfenol-D to approximately one-half of its maximum strain. The field from the
internal solenoid is used to add to, or subtract from, this "DC" field, the result is bidirectional motion
of (a) relative to the housing. This is the basis for the transducer input-output relationship.
The path for the magnetic flux is (f)-(k)-(i)-(h)-(a)-(c)-(f). The housing components,
(d) and (j), are made of aluminum since it is approximately magnetically "neutral" (and readily
available). The components along the flux path are made of "high" permeability, fully annealed 1020
steel.
The transducer incorporates a simple prestress adjusting scheme. In addition, the design is such
that the cylindrical air gap in the magnetic circuit (between (c) and (a)) remains constant (radial
clearance of approximately 0.010") through the entire strain cycle of the rod.
EXPERIMENTAL DETAILS
A schematic for the experimental set-up is shown in Figure 3. The accelerometer (Kistler
808A) charge amplifier pair, the resistor (used for estimating current values), and the Tektronix
2630 Fourier Analyzer were calibrated before these measurements were performed. For the
accelerometer, the charge sensitivity versus frequency was plotted and a linear least squares curve fit
was performed. Broadband measurements from 0-10 kHz were scaled using the charge sensitivity
implied by the curve fit evaluated at the center frequency of 5000 Hz. The maximum deviation from
this value was eight percent. The standard deviation was 3.5%. Measurements taken at single
frequencies were scaled by the appropriate charge sensitivity values from the calibration data. The
amplifier shown in Figure 3 was simply a standard 50 Watt audio amplifier. One end of the shaker was Broadband Magnetostrictive Shaker 5 David L. Hall and Alison B. Flatau
attached to ground during tests by bolting it to the floor on hard rubber pads. Thus, relative motion
between the ends of the shaker's magnetostrictive rod corresponded to absolute motion of the shaker's
free end.
The influence of structural resonances introduced by the actuator mechanical design and the
shaker mounting configuration during tests were assessed as follows. A modal analysis using an impact
hammer identified a rigid body mode near 1500 Hz due to the shaker-f