Mark,
I personally would prefer to use actual data when setting probe tension. Probe tension is not the only factor involved. There is also probe integrity, measuring speed, cosine error and probably a few more that I don't know about. We had four Mitutoyo CMM's in a facility that I worked at and each one was set up different. You can also try swapping probe tips to see if that improves your standard deviation. One thing that I would suggest is to set up your CS at the center of your reference sphere by measuring it at least 2 or three times and making the average of those coords the origin. Verify in your program that your measurement vector is correct and I prefer to measure at 45 degree angles to your CMM "home" coords.


Duane Wade
CMM/Gage Technician
TeleFlex Automotive - Pedals Division
301 W Ohio St.
Kendallville, In 46755
219-349-1985
dwade@tfxauto.com