Hi, pretty interesting....the object will always be to accurately copy the target to reproduce it with the mill........with whatever means so far devised if the outcome is as good as the target then the solution is already solved.

With that thought in mind, how much more accurate for $2,350 is the Heidenhain model compared to the $90 dollar Renishaw device, both on EBAY, and I assume they both work on the same principle.

If there is a whole field of difference in the accuracy of the plot in a side by side comparison, then the problem of DIY making is the key......where does the problem lie if it only needs 3 sets of balls and 3 round rods to do the sensing.

In the magnet design I'm working on there would be 3 sets of magnets top and bottom at 120 deg that opposed the central shaft, so holding it in suspension in a magnetic field......this already exists in a magnetic bearing design and I intend to look closely at that principle to see if it has merit on a smaller scale.

The shaft could be mounted in a tube with silicon rubber, but it has no adjustment for sensitivity, but it still has possibilities if used in only small doses.

Silicon rubber, if used as small bushes inside a tube and the OD of a shaft, would work in all directions once cured, so there are endless simple possibilities........the shaft only has to move .02mm at most.

I can quite well imagine that the probing cycle would take quite a few hours probing at the rate of one step forward each time and one step sideways for the next row of hits.

I think that I would have mostly an all round type probing need as opposed to a purely 2D or flat plane one, and that would mean the 4th axis I've got planned is going to get moved up the ladder of priority a notch or two.
Ian.