CMM Results double those obtained manually

[[Ja...]]

I have a part that I am measuring true position on. I have made a highly simplified model to ask a fundamental question about the evaluation. I have the part sitting on the table such that the bores are along Z and the shank is along X. The position of the feature is called out to -B-, -C-, and tertiary -F-. I set up my origin to Level Z+ to -B-, Rotate X+ to -F-, Origin X, Y at -C-, Origin Z at -B-. This is what I am assuming:

The X and Y deviations in my True Position report are from the Feature to -C-, where the Y axis is being forced perpendicular to -F-.


When the parts are measured manually, they use a pin to locate on -C-, and they locate on -F- and they square up to -B-. Then they use an indicator to check the Feature. They then rotate the part and check the feature again. Their deviation is coming out very near half of the CMM results. Can anybody see which method is not being performed correctly?

Thank you.

[[Ro...]]

Make a 2d line between C & Feature, then check parallel between your constructed line and F.
If possible probe everything at the same Z height.
Also check the form on everything, especially of you're using cylinders.
Are you using 3 cylinders? if so maybe try 2 circles and a line?

[[An...]]

Don't alter the datums.
See attached.

True_Position_BCF_1.pdf

[[Ja...]]

Please sign in to view this quote.

I tried constructing the 2D line but I receive a parallelism result of 0.8425. This seems absurd. Any clue what may be going on?

[[Ja...]]

Please sign in to view this quote.

Andreas, my program is not altering the datums. I Spatially Rotate to -B-, Planar Rotate to -F-, Origin X & Y to -C-. Origin Z to -B-.
I am measuring the feature, -F-, and -C- as cylinders.

If there was a deviation in perpendicularity as you're implying with the pin (manual) measurement, their results would actually be worse in any given direction.

[[Ja...]]

Please sign in to view this quote.

I found out my issue here, and am now getting about .003 parallelism. This doesn't help explain why my deviation in Y is about .018.

[[De...]]

Usually I find the problems because of incorrect methods of checking the parts on the surface plate.

It sounds like from your description they are aligning to C first, however C is NOT the primary Datum, B is.

The CMM will align to B first then C (and should do C while constrained to B if setup correctly)

If there is error in perpendicularity between B and C the difference in which is used as primary can be surprising.

[[Ja...]]

Please sign in to view this quote.

Thank you for the response Derek,

Can you clarify on what you mean by "and should do C while constrained to B if setup correctly?"

Thank you

[[Me...]]

Jacob,
I have a couple questions about your datums. First, is the B datum only the surface around the C datum or does it also include the surface around the 'feature'? Second, is the F datum the cylinder axis or the surface that the identifier is attached to? Third, did you intersect your C datum cylinder with the B datum to get a single point? I've found that when using cylinders in a base alignment, Calypso sometimes has trouble defining the alignment correctly.

[[An...]]

See attached.

True_Position_BCF_2.pdf

[[De...]]

In the True position characteristic, your secondary and tertiary datum features should have constraints set to constrain the evaluation of those features based on the order of precedence stated in the datum reference frame on the blueprint.

There is a button on the constraint page with the label 'Form Datum per ISO 5459' hitting this should setup your constraints so that subsequent datums are constrained to previous ones.

[[Ne...]]

I've recently been involved in a similar situation where my perpendicularity is reporting .010 out of tolerance but hard checking with a surface plate and indicator we're getting around .004-.005 angle to the inside of the cylinder in my Y direction. I printed out a graph to try and see why its doubling and I discovered its adding the max deviation to the min deviation. So my question is how do I tell calypso to only report the maximum deviation?

[[Ne...]]

I've recently been involved in a similar situation where my perpendicularity is reporting .010 out of tolerance but hard checking with a surface plate and indicator we're getting around .004-.005 angle to the inside of the cylinder in my Y direction. I printed out a graph to try and see why its doubling and I discovered its adding the max deviation to the min deviation. So my question is how do I tell calypso to only report the maximum deviation? And if the datum and feature actually meet each other, why wouldn't the min be .000?