Thank you so much, wolfekeeper and evan_au =)

wolfekeeper, the face of the two cylindrical PMs will be randomly positioned with x,y,z distances relative to the coordinate frame of the primary PM. I should be able to take that into account as magnetic field vectors, but I have yet been able to model this mathematically while considering the rotation (which in this case, I assume will also affect the net magnetic field at different angles)

Are these short cylinders or long cylinders. If they're long compared to the diameter then FEMM will give you a reasonably accurate answer.

These diametrically magnetized cylindrical PMs are of 1cm diameter x 1cm thickness. Sounds like the geomtry of the PMs is the tricky part? I wonder if there are any assumptions I can make to still get a valid result?

evan_au,

Are you trying to transmit torque through a container wall (eg for a stirrer driven by an external motor, without the need for a hole through the container wall)?

yes =) about the same concept. I am transmitting torque on the PMs across the barrier so the PMs would have contactless rotation. I am wondering if anyone has used this method before in any applications. It would be a nice reference.

Since the PMs will be randomly positioned, I would have no guarantee how the actual arrangement will be like. So my model would have to cater for the difference between the two in all the x,y and z axes. The transmitter magnets though, would most probably be pretty close to one another, unfortunately =( as my workspace is pretty limited (to about 30x30cm slightly spherical surface). Any idea how I can find out the optimal distance between the magnets for the highest torque transmitted?

Thanks again! =) =) =)