Torsion Meter
One of the objectives of building model wind turbines is to understand the power available at the rotor shaft. This required the construction of a small dynamometer capable of measuring 0 to 50 watts at 0 to 500 rpm. The simplest method I could devise for measuring torque was to use a shaft with a circular cross section. One end of the shaft was secured to the base of Mecanno framework the other end being supported on a simple bearing and free to rotate. Over a small range, the torque should be proportional to the angular rotation (a twisted version of Hooke's law).
Some initial experimentation suggested that small diameter wood and plastic rods could make suitable torsion bars. A metal shaft with a diameter small enough to provide measurable angular deflections would not have the necessary longitudinal stiffness and might be prone to buckling as it is probable that some compressive forces would be encountered in operation. The first attempt uses 450 mm length of 6 mm diameter wood. This remains straight at twists less than 20 degrees and can support a compressive load of 200 gm.
Initially, flanges were glued to the end of the shaft, the downside of this arrangement was that small misalignments resulted in the shaft bending when installed in the framework. This problem was avoided by using a short length of plastic tube to make the shaft a tight push-fit int the flanges, this tolerated small misalignments and the shaft remains straight after installation. A sketch of the torsion bar is shown below:
The angular deflection of the shaft is measured by IR photo-reflector interacting with a inclined surface as the shaft rotates the distance between the inclined surface and IR photo-reflector changes and the output from the photo-transistor element provides a voltage that is related to the twist caused by the torque. This scheme was adopted to ensure that measurement did not introduce any additional friction into the system.
The relationship between torque and twist is shown in the graph below:
There is a relationship between torque and twist is almost a straight line. Some hysteresious was evident in the results, the source of this is not clear the two likely causes are friction within the wooden torsion bar or the bearings at the free end of the torsion bar.
At the time of writing, only static tests have been conducted. When the torsion bar is subject to fluctations and rotational vibration, the dynamic response might be more complex than suggested by the graph above.
The angular deflection sensor is shown in the the photos below:
Meccano is a great tool for prototyping, the stock items can be extended with custom built items like those shown above. The photo shows the torsion meter during test and assembly, a couple of aluminium plates from earlier projects have been recycled to provide bracing for the framework.
Related Material
The Next Step
Anticipated develops include extending the range of torques by building an 8 mm torsion bar.