Having trouble wraping my little brain around this problem...

I digitizing a part, between centers (headstock is a rotary table) in a mill. I zero out with the digitizer, digitize the patch, then rotate the part 90deg, and digitize again (W/O zeroing again). I rotate every 90deg until I get the 4 sides, only zeroing on the first.

When I cut the part, I to the same...zero on the first side only. However, when I cut, I'm getting minor shifts in the part.(can be in any axis, actually). DO I NEED TO ZERO EVERY TIME I ROTATE? Can anyone comment from experience? The part is NOT symetrical.

Thanks, Jim

I do a lot of this type of work (indexing is what its called) and from you description there is only a couple of possibilites. The most likely is this, your "homing of the axis" is in fact not truely precise. for instance your "a" axis has two set points, the head stock and the tail stock. If one is off you will have undesireable overlap when indexing. The preceding has two possibilities of error, human or machine. Is the machine absolutely precise? That is to say does the headstock and the tail stock (parrallel to the x axis for example) line up perfectly in a 90 deg fashion to the y axis.... is the homing method absolutly precise.
Next check for backlash. Did a bolt or tightening screw come loose somewhere in the machine??
another possiblitity thats is most likely to happen with stepper motors is, do they have enough power? stepper motors since they are not in a closed looped system will not be able to tell if they have fallen behind. If you are really stressing your machine it is possible that the stepper motors are loosing a few steps which would show itself in overlap when indexing the part.
Finally I will add this. If its a home made machine ( for instance a wood router) check and consider every part that will be put under stress as a possible candidate for shifting. Especially the z axis!

Thanks for the good info. I had to figure out the hard way that my zero (y/z)has to be right on the y-z axis. After you think about it for a while, it makes a little sense. I was getting the overlap, but from the zeros being off the axis. Thanks! Jim