Runout issue

[[Ji...]]

We have ran into a part that we run at least 3 parts weekly that is giving us a bad runout on one part. We have straightened the probes to where styli are within .008" total to one another and sigma is less than 20 millionths. We know the part is straight and square. Our Z value is .0004" between the 2 bores and our Y value is .0002" and are .0003" on the form. The runout is saying it is .004". We looked at the angle in Features and it was .01" of a degree. We are looking for a way to explain how the runout is showing that bad on the one part. The other two parts show runout to be about .001" on the same set up. They also have worse numbers in ZY by a few tenths. Any ideas what to look for to proof out the runout calculation on the Zeiss would be appreciated. We have ran the part several times and keep getting the .004" with 4 different people running it. I think we have ruled out operator error and shanking error. We have the Datum bore as a cylinder and have checked the other bore as a circle and cylinder and get real close to the same number. We are running slow and have 2500 points on the cylinders with filters and outliers set as we always run them.

[[De...]]

The most likely cause is form error in the datum feature causing the outer tangential algorithm to return a datum axis which is skewed so your result is shown with excessive runout. In order to check this make a second characteristic and set the evaluation of the datum to least squares.