A magnetostrictive motor and its drive circuit and control system have been designed to be especially suitable for robotic applications in which there are requirements for precise, high-force linear actuators. The motor includes a laminated armature made of the magnetostrictive alloy Tb0.27Dy0.73Fe0.2. The armature is sandwiched between two double-layered, three-phase stators, which are energized to make the armature move linearly in “inchworm” fashion. The total range of linear motion is 25 mm. Like other magnetostrictive motors, this motor offers the advantages (relative to geared-down conventional motors) of reduced weight, extreme ruggedness, fewer moving parts, greater reliability, and self braking when power is not applied.
A capacitor is connected in series with the stator windings to correct the power factor. This or almost any other magnetostrictive motor presents a highly inductive load to its drive circuit and therefore operates at a low power factor in the absence of correction. As in other electrical applications, a low power factor is undesirable because it gives rise to the need for a greater drive potential or drive current than would otherwise be needed to deliver a given amount of power. At its resonance frequency of 470 Hz, the motor windings exhibit a power factor of 0.352, but the series combination of the capacitor and the motor windings exhibits a power factor of 0.989 — close to the ideal value of 1.
Because the speed of the inchworm motion depends on both the amplitude and frequency of the drive current, the control system includes one controller that holds the frequency constant and varies the amplitude and another controller that holds the amplitude constant and varies the frequency. Both controllers utilize proportional + integral compensation and implement an integrator-anti-windup scheme to limit accumulation of position-error signals.
The control system includes a position sensor and a 12-bit analog-to-digital (A/D) converter that processes the sensor output. Because the output swing of the position sensor is only one quarter of the input range of the A/D converter, one could utilize only 10 of the 12 bits (corresponding to a position resolution of 49 μm) if one were to feed the raw sensor output to the converter. Therefore, to make use of full 12-bit resolution of the A/D converter, the sensor output is fed to the converter via an amplifier stage gain of 4. Another amplifier stage with a gain of 39 is also included to demonstrate a capability of precise positioning; a position resolution of ≈1.25 μm is achievable when this amplifier is included in the signal path.
This work was done by James H. Goldie, Won-Jong Kim, Andrew E. Barnett, and William R. Snow of SatCon Technology Corp. for Johnson Space Center.