Thanks for posting. Hallbach arrays are very interesting and his "field cancellation" is exactly the principle I am using. It would have been more interesting if he had gone into more detail and examined some of the other aspects of field cancellation.

Sadly, I need to redesign. At this scale it is very difficult to get the tolerance needed to get this design to work. Also I need to design in adjustment for lining things up. The crank mechanism, whilst beautiful and good in theory is not ideal in practice, so I'll use a much simpler Scotch Yoke system to convert the reciprocal motion to rotary motion.

Off to FabLab. Wish me luck. Here's another render before I go: