Eddy currents for measurement of twist angles

[suz123]

hey, could someone tell me if it is possible to measure the angle of twist of a rigid shaft subjected to a torsional moment at one end by using eddy current testing?

 

[EWH]

In general, for a circular shaft, a simple, single measurement, and a DC magnetic field, I think no. A rectangular shaft that is thin in one dimension and thick in the other would allow angle sensing within one quadrant with an AC field. You could measure angular velocity with a circular shaft if the shaft is conductive and either not magnetizable or with an AC magnetic excitation, and you could integrate that velocity to get angular position, but the latter measurement would be unreliable. There might be some way to do it with some additional part geometry over a part of a revolution, and to use multiple sensors to get better angular resolution, but it might not have any advantage over regular encoders. (Then again, it might.)

 

[suz123]

im sorry i wasn't clear on the type of shaft.. it is of micrometer dimensions.. hence conventional methods will be ruled out.. optical encoders may be a solution.. but again, size is the constraint here

 

[EWH]

I think a capacitive method would be preferable for MEMS. Magnetism doesn't usually scale down as easily. (Perhaps something adapted from hard drive technology might work, though?)

For instance, I'd imagine attaching to the shaft a pair of conductive sector plates laid out like the blue bits of the BMW logo, with these plates at a small constant distance above similar sectors to which are attached capacitance meters. Or perhaps high-k dielectric sectors attached to the shaft, which pass between pairs of sensor plates. Or a broad, shallow, 1-turn screw thread attached to the shaft with the capacitance varying at a given point according to the changing distance to the sensor plate.

It really depends on the details of what is possible in your fabrication process, how fast and how accurately you need to measure, what you want to measure (position, velocity, acceleration, torque, etc.) and what constraints or freedoms the other parts of the design impose or allow.

 

[suz123]

well the sector plates to measure a change in capacitance is a good idea.. but fabricating such a structure on a micro wire would be an enormously challenging task! I am basically exploring options that can be surface micromachined.. hence 3-d structures would be ruled out.. something in the plane of the wire would be more compatible... in certain instances, a cross beam has been mounted on the wire. and torque applied on the wire would cause the beam to turn and hence actuate some sort of load cell.. but this method cannot be used due to certain constrains..

Basically i want to measure the torsional moment applied. i was looking along the lines of measuring the twist and deducing the torsional moment from that

 

[EWH]

You mean measure the torque? It seems like the torsional moment of the shaft would be easier to just calculate. The measurement of the moment itself would be difficult since any thing you attach to do the measurement would change the moment, and with frictional forces usually being dominant and the moment itself being so tiny...

The torque with one end of the shaft fixed and the other torqued could perhaps be measured if the shaft had one or more flat sides - the flat side would twist relative to the fixed side and that could be sensed using a capacitive, magnetic or best of all optical manner (bouncing a laser off it like an old mirror galvanometer.
Thinking out loud...if the shaft has to be circular, maybe a thin birefringent coating could be applied and a polarizer-analyzer setup could see the stress? Or maybe stress or strain would show up in some kind of Raman change? Or maybe being a crystalline substance, the reflected light polarization will change with twist as the crystal axis changes? Perhaps twist could be somehow converted to a distance measurement and you could use interferometry? Perhaps applied torque could be measured with some kind of watt-balance setup? I'm out of my depth. Surely this is something that someone has worked out how to measure before.

 

[suz123]

well that's it right there.. there isn't much literature on this.. but thanks a lot! I'll definitely explore your suggestions.