I have created a Geo correction file similar to what I have done in the past. I lined a precision straight edge up in x. It was at zero on the test indicator mounted in the spindle and touching each end of the rail. Along the travel from one end to the other I recorded the deviation at each inch. This revealed a bow in the x that basically created a arc of about .03 inches along its span. I made measurements at each inch of the deviation and created points in cad space. I then mounted a linear scale to a frame and squared it to my straight edge. I took measurements of the travel also every inch and created a offset that was equal but opposite the deviation. This was mapped to my previous map. I squared another precision straight edge to the one already mounted. I the took my y measurements same as x. This went quickly since it was straight and square to x by less then .001 over the whole travel and the screw held less then .002 deviation at any inch move. I decided y was close enough to leave out of my mapping needs. I adjusted the measurements file to create a grid of 2 rows and 49 columns. The spacing was set to 1 and 32. That's to cover the whole travels 0-48 in x and 0 and 32 in y. The origin was set at 0,0.

I took my measurements by running a g-code program that single stepped through the mapped coordinates and using the measure button. The file I created works but has some issues. When I redid my previous tests in x to check my results I found mixed results. The scale measurements of travel distance were reduced from as much as .029 inches down to .0015 or less at any position. That's great and better then expected. It's also repeatable every time. The issue I'm having is with the correction of the bow in x. When I jog or make step moves along the straight edge the indicator shows the bow being corrected and deviation over 48inches is now within .004 of the straight edge. I think I could make this better with a more careful map of the bow. When I make a move along the straight edge in mdi or through a g-code program the error increases drastically. The move now has a error of as much as .015 which is better then .029 but still not good. Oddly it seems the x movement correction is still correct but the y tracking is off whenever I run g-code which is where it really counts.

I'm wondering if the configuration of my map could be causing this even though it works while manual jogging and stepping. The error goes back up to about .015 during rapid as well.

If anyone has any suggestions of where to look please help.

Thanks
Ben

Sent from my E6810 using Tapatalk