probing profile on the lathe and then...

bogus105 wrote: 2. i can then open log file in excel-like spreadsheet program and use linear interpolation (somehow) to create new, smooth profile.

I doubt that linear interpolation is the best solution. putting the points into a NURBS block is likely to work better. Look at the docs for G5.

You can offset the profile simply by having the tool and the probe different lengths. (They will be different lengths, obviously, you just need to lie about by how much they are different).

Point-cloud to surface is a hard problem. Your problem is only slightly easier.

NURBS! this is interesting. i looked at docs for G5 and even tried to write small G5.2 block and my linuxcnc suppory NURBS.
So i know how to digitize the profile, i know that NURBS may be the solution, probe-tool offset should not be a problem but how to connect the mentioned elements? The problem is how to automate probing and creating NURBS block based on digitized data?

Also i do not fully understand L and P parameters in NURBS block. Have you guys have some good web address explaining these parameters? linuxcnc docs do not cover this subject in details.

Rick G: do you know forum name of Mr. Dewey Garrett? i used forum search for his name but with no results

bogus105 wrote: So i know how to digitize the profile, i know that NURBS may be the solution, probe-tool offset should not be a problem but how to connect the mentioned elements? The problem is how to automate probing and creating NURBS block based on digitized data?

I think I would start by trying to use a Python script. You can do a lot with LinuxCNC from Python. And you can make a GladeVCP or PyVCP button run Python code. (I would use Glade, embedded code is easier there).

The Python probably needs to output the profile to a file, then tell LinuxCNC to load and run the file.

The actual probing can probably be done by sending MDI commands through the Python interface.

linuxcnc.org/docs/2.7/html/config/python-interface.html