« Back to the SERA Actuator home page
Contents |
To simplify the system, the team is currently considering a design in which a larger motor is used and the worm shaft is fixed to the motor output shaft. The axial compliance of the motor is used as the range of motion for the worm shaft and therefore no spur gears are involved. A flexure assembly provides the elastic element and allows for indirect sensing of output torque by sensing the deflection of the flexure. This design has not been built or tested.
The team is considering two major classes of sensing. One would be using a magnetic sensor and a magnet coupled to the flexures; the sensor would sense the angle of the magnetic field and therefore infer the angle of the flexure. The alternative is a strain gauge mounted directly on the flexures which directly senses the strain of the flexures. A third possibility is using load cells instead of flexures.
The benefit of magnetic sensors is their high precision and that they don't require mounting circuits directly to the flexure. The down side is that a circuit board needs to be mounted very close to the worm shaft and a large amount of design work needs to go into the selection of the sensor and ensuring that the magnets' angle changes by enough that large portions of the sensors range go unused, decreasing the precision of the force sensing. Also, the precision of the location of the sensor vs. the position of the magnet is very significant to the results, and since the device relies on very small deflections for the readings (on the order of 1/100,000th of an inch) the team worries that this may be unusable due to our being unable to guarantee that sort of mechanical precision.
Strain gauges give a much more direct sensing mechanism than magnetic sensors. However, they require mounting electronics directly to the flexure and calibration of each individual gauge. Furthermore, they would require a conversion to digital close to the sensor to prevent the electronic noise of the motor affecting the measurement.
The flexure with strain gauge design is essentially what a load cell is, so this is an obvious alternative to the design. The primary difference is that load cell manufacturers calibrate and guarantee the performance of their devices, allowing our design's precision to become another manufacturer's problem. The difficulty is that load cell manufacturers engineer for almost zero compliance and therefore the mechanical compliance of the device is gone. However, the team considers that the very small mechanical compliance was close enough to zero that this was not a serious concern.
The contents of this page are Copyright © 2007 Matthew Aasted, Guilherme Cavalcanti, Jeffrey DeCew, Christopher Dellin, Gill Pratt, Kevin Sihlanick, and Jon Tse.