Checking Runout on a CMM 1

[Belanger]

I've been around CMMs quite a bit, but it was never involving runout. Can runout be checked on a CMM? We always talk about TIR/FIM and explain runout in terms of the dial indicator...blah blah... but I'm curious how it's actually done on a CMM if the part itself isn't set up to rotate.


John-Paul Belanger
Certified Sr. GD&T Professional
Geometric Learning Systems

http://www.gdtseminars.com

 

[dingy2]

J-P:

A CMM is not the best approach for the 2 runouts. I would suggest using a chuck (divider head) to contact the datum and then use a dial indicator (not digital) on a stand to confirm the circular or total runout on 100% of the circumference. I would say that this approach would give a 95% confidence in the result depending how one contacted the top of the feature.

I know that one could use a CMM to create an axis on the datum and take numerous points around the circumference on the feature. If the points were take in one slice, then the difference in the point readings would reflect the circular runout. If we took numerous points around the circumference in full length, the difference in readings would give one the total runout. I would say that this approach is about a 70% or lower confidence in the result since we are not take 100% of the circumference. We are comparing taking 6 - 12 points rather than 100% of the circumference using a chuck.

Dave D.

http://www.qmsi.ca

 

[Belanger]

Thanks Dave. You know how it is... we are so ingrained with explaining runout in terms of the dial indicator method, so I had to stop and think about how do-able it is on a CMM.

My hunch was that it would be hard with a CMM. I figured that it's because it is too tedious. But I didn't realize that the readings would actually be of less confidence.

John-Paul Belanger
Certified Sr. GD&T Professional
Geometric Learning Systems

http://www.gdtseminars.com

 

[dingy2]

J-P:

The readings are fine but the method using A CMM is suspect at best and that is why, in my opinion, the results would have 70% or less confidence.

Dave D.

http://www.qmsi.ca

 

[PaulJackson]

Every inspection is an estimation... there are no perfect gages!
The jaws of the chuck that you use to check runout with are never exactly coaxial with the axis of the spindle bearing that the chuck spins on.

The bearings themselves always have error... but most of all no one can ever inspect the infinite set of circular cross-sections perpendicular to the axis to reveal the one with the greatest measure of runout error.

Does this mean that no one method can be labeled "best" or "most accurate"? you bet!

So can one estimate circular runout on a CMM? Of course they can! The quality of the estimation depends upon the accuracy and precision of the tool and the skill of the operator using it.
DCC CMM's are typically easier to use to accomplish the analysis because the software can drive the probe tip to touch (or continuously scan) the surface of revolution perpendicular to the resolved datum axis orientation no matter what its rotation from the machine's axis is. With a manual CMM one has to set up the datum feature(s) axis reasonably perpendicular to the machine's axis to get the best estimation of the circular cross-section's runout.

For both types of CMM each point is figured for its radial displacement to the established datum feature(s) axis and the difference of the MAX and MIN values can be reported for circular runout. Naturally a reasonable number of individual cross-sections would be examined according to the expected or observed variability in the cross-section results. If one wanted to estimate the total runout the MAX and MIN radius values of all cross-sections sampled would provide the total runout.

Sometimes when there is an accurate spindle, with an equivalently accurate chuck and a sufficiently long and stable registry of the datum feature that may be the best tool to use for the inspection and sometimes other tools or even a CMM is better suited to perform the inspection.
Paul