One of my favorite customers was having a bad day. "Those CMM operators are disputing my gage results," Sue said. According to the operators, the results from her gages were significantly larger than their CMM readings. The process operators preferred the CMM results because they could use them to help control their process. The gage numbers didn't seem to help, so they decided to do it their own way. "Now my whole inspection process is in doubt."
I told her that this is not an uncommon problem, and I asked for the details. The part in question was a simple shaft with two diameters, and the check was to ensure coaxiality. The gages Sue was using were designed to measure what the print called out, circular runout, which is the old reliable TIR check. This gives operators a number they can compare against the tolerance to determine if the part is good or bad.
The CMM operators decided they would look at the check with a concentricity function. They felt this was more valuable because it gave the operators an X and Y figure that could be used for offsetting the machining center to control the process. While this is an interesting approach, it's also one headed for conflicting results.
Circular runout is a two-dimensional measurement using surfaces to control an axis. The tolerance is applied at any cross-section. When it is used on a surface referenced to a datum axis, as with this part, it will control the total sum of all variations of circularity and coaxiality.
Concentricity is more complex. It can be thought of as three dimensional and uses a series of diameters and their midpoints to generate an axis location. This is why the process control operators like it: It gives an XY location of how the two axes line up.
Why would one be larger than the other? They both start out similarly, with one diameter used as the reference. The part is either held in a Vee type arrangement or a three-point clamping system to set up an axis of rotation. The outside diameter and the holder act like mating parts to create the minimum circumscribed cylinder.
In the case of the TIR check, an indicator is used to watch the test diameter as it is rotated. The indicator will see everything that the diameter is doing, and the movement of the indicator will include: any misalignment or runout between the reference fixture and the test diameter; the angular relationship between the two diameters (coning error); and any form error in the test diameter.
The indicator is placed on the measured diameter, set to zero and the part is rotated through 360 degrees.The measurement must fall within the part diameter tolerance and the runout tolerance. The check should be performed at a number of locations along the axis. This check will not control taper.
However, with the concentricity check, one diameter is held in a fixture to set up a reference axis (or the CMM may create an axis based on a few diameter touches on the reference). The CMM takes a series of diameters on the surface being measured and finds the center of each diameter. Any number of diameters can be used, but four to eight are typical. Then the centers of the diameters are plotted to create a set, or cloud of points that can be used the find the mathematical XY location of the center of the diameter. As long as this cloud of points falls within the specification, the concentricity spec is passed. This check should also be made at a number of different locations (axially) to ensure that the whole length is within tolerance.
Why the difference? Ignoring any error in form or alignment, the TIR check bases its coaxiality reading on diameters, while the concentricity check calculates radii. However, there is another potential reason for different results: The runout check is taking an infinite number of readings, while the CMM is taking four to eight diameters. The indicator is apt to see much more form variation than the CMM.
Which one is right? They both are for their particular call-out. If the print specifies runout, then the part should be inspected for runout. If concentricity can be used to help control the process, that's great for process improvement. Try to keep these apples and oranges separate and use the right check for the application.