How to map a ball screw?

I'm looking fro help as to the best way to map a ball screw? I understand the comp file structure of commanded vs actual and backlash but want to check what the best way to do the whole process of checking was?

I have some varying errors in my mill, which may be binding in the jibs etc so I need to test the whole system and start by setting the scaling. My suspicion is that it is just the center part that has worn as the ends seem to ok when checked over a small distance using a 10mm travel DTI.

I have now borrowed a 5mic glass scale and DRO as a standalone system for checking. Assuming it is the ballscrew I was going to do an initial sweep in one direction and check the full travel for the scale check and then check again in small increments and log the variations for the mapping file

My question is , Do i need to do it in both directions as full sweeps or move back and forth?

Thanks Andy
Averaging lots of samples seems like the best option, I think that's were I'm going wrong buy using too small a sample and introducing more error

Can i log it with linux cnc or would I need a a separate pc and program? any pointers as I've not done any data logging yet.

cheers

Seems easy enough. famous last words :oD

I'll have a play and see how I get on.

Initial questions are where does "print" print to?

and "The current value of the input is stored in parameter #5399" where is that

I think you can only have 256 points per axis.

This seems to be right. The question is, how does linuxcnc handle the positions between these 256 values?
Is there an algorithm that avarages the space between or is it stepwise (linear compensation before/behind the values) until the next value is reached?

Does anybody know something about it? (have not found anything)

I have a lathe with both rotary encoders and linear glass scales. Both are physically connected to encoder pins and both are set to home together at the same instant (wired connection of index pulse). Only rotaries are used for position feedback. (Scales are too coarse at 0.005um).

I would like to map the ballscrews with reference to the glass scales.

Can anyone share the G Code which has automated the axis mapping process and writing to a file as described above?

Thanks.