| Budget Amount *help |
¥7,800,000 (Direct Cost: ¥7,800,000)
Fiscal Year 1999: ¥3,600,000 (Direct Cost: ¥3,600,000)
Fiscal Year 1998: ¥4,200,000 (Direct Cost: ¥4,200,000)
|
|---|
| Research Abstract |
Magnetostrictive effect inherent in the steel shaft provides noncontacting means of detecting torque applied to the shaft. A problem encountered in this scheme is low accuracy due to magnetic inhomogeneities being inevitable for the case-hardened steel shaft. Our torque sensor consists of induction hardened SNCM steel shaft (25 mmφ) and a magnetic head with a pair of planar figure-of-eight pickup coils embedded in a ferrite core. The objective of this research project is to study factors affecting the accuracy of the torque sensor and to find methods of reducing influence of the factors or to improve accuracy of the torque sensor. Results are summarized below :
1. Reduction of zero-level-fluctuation
Zero-level-fluctuation of the torque sensor is strongly dependent on the balance of the inductance-resistance bridge, which is a main building block of the torque sensor. When the bridge is unbalanced, zero-level-fluctuation is enhanced. A bridge was tuned with a variable resistor connected p
… More
arallel to one of the pickup coils. With numerical simulation, this tuning condition is found to be unique, that is, global minimum in terms of the root mean square error. However, the tuning condition for a minimum zero-level-fluctuation does not meet, in general, the condition for a minimum offset. We indicated a method to avoid this problems, in which dummy impedance components having averaged values of those of active heads are to be used in the counter branch of the bridge circuit.
2. Drift due to ambient temperature
Observing temperature variation of the permeability of each component used in the torque sensor, the most responsible component for the temperature stability is found to be casehardened steel shaft. This is due to the relaxation of compressive stresses in the case region of the shaft due to thermal expansion. Drifts of the torque sensor are observed in a temperature range from 40℃ to 140℃ for various initial offsets. It is found that the drift becomes smaller for smaller initial offset. As an example, temperature drift was -22% for an initial offset of 0.047% normalized to the excitation voltage and was reduced to -7.5% for an initial offset of 0.013%.
3. Eddy currents effect on the zero-level stability
Sensing coils induce eddy currents on the surface of the steel shaft. We investigated the influence of the eddy currents on the zero-level stability taking the excitation frequency as a parameter. We found that zero-level perturbs over several period of revolution when the excitation frequency is low and that this phenomenon disappears when the excitation frequency is higher than ten times of the rate of revolution per second. Less
|
