inspection of run-out (figure 9-6, page 183 of 2009 standard)

[AndrewTT]

How would one physically inspect the total runout of 0.02 mm w.r.t. datums C-D on datum feature C and on datum feature D? If you are supposed to clamp onto datum features C and D to create the axis of rotation how do you now put a dial indicator on these surfaces to measure runout? How do people handle situations like this in practice?

Does anyone know if ASME B89.3.4 (Axes of Rotation: Methods for Specifying & Testing) contain information regarding to my question?

Thank you.

 

[powerhound]

Just toss it up on your CMM or Tallyrond, scan each datum feature, create a datum axis between the two features, then evaluate the datum feature back to that datum axis.


John Acosta, GDTP Senior Level
Manufacturing Engineering Tech

 

[AndrewTT]

I was thinking that you would have to scan or CMM. What did people do before these devices were available?

 

[JNieman]

Designers had to actually think about how to validate parts before assigning requirements

 

[pylfrm]

AndrewTT,

If you physically clamp the datum features, the clamping devices probably won't actually cover the entire datum feature surfaces. There will likely be some room left for an indicator probe.

Alternately, you could rely on the fact that the total runout with respect to datum axis C-D is not likely to be much greater than the total runout with respect to any other axis. If the total runout measures 0.015 when rotating the part on centers in the end holes, then it's probably good.


pylfrm

 

[3DDave]

It's complicated. Like all verifications one needs to determine how much measurement error to subtract from the allowable error.

In this case one can set up a rotation mechanism based on items that clamp on C and D and then lock the part to that mechanism and remove the clamps. The error is the measured runout between the clamps on C and D and the rest of the rotating mechanism. It will also have to include the error of the mechanism that traverses the surfaces - that path needs to be parallel within some amount, which also gets subtracted from the tolerance.

The ugliest part is that it may be that the part is offset and cancels the error in the part, so it is necessary to do this measurement at several orientations to discover that possibility.