we're facing a problem with a tube part which is approx. 500mm in lenght and consist of multiple diameters.
We have a concentricity on a diameter which is almost in the middle of the tube. Datum is A-C - two diameters - one on each side.
Now the problem is, that with every turn of the tube, concentricity is always different - as told, it shouldnt be.
Example:
Concentricity without moving the tube - 5 measurements - there's no deviation between measurements.
If we turn the tube 4 times for 120° - 1st and 4th measurements are identical.
Tried with different alignments - calculated concentricity is always the same.
Are you able to measure the diameters 360° around? If not and you are only able to measure sections of the circle, this is very likely to happen - especially if they aren't round...
Concentricity is alignment independent.
Also, how are you measuring everything - circles or cylinders? Scanning or touch trigger? How many points? What is the form error of each of the features / datums? Which method of dimensioning are you using (legacy, XactMeasure or Geometric Tolerance command)?
It will if you're only taking a limited number of hits (as opposed to scanning). When you rotate the part, the hits will be contacting a different point on the circle which will mean that the resultant circle calculation will have a slightly different diameter and centre point. If you scan at a fairly high point density, you should see less of a difference when rotating the part and re-measuring.
Don't forget that this applies to the datums as well as the feature being reported. The form error and the number of hits you take on your datum A and C circles will also affect the result. In theory, if all datum and considered features are scanned at a high density and covering a full 360° of each one, then you should see little difference when you rotate the part and re-measure.