Surface Profile Tolerance Not Calculated Properly

[[Pa...]]

I have a part we have scanned with an ATOS 200. I am comparing the scanned part to CAD. I am measuring in inches. I have created a nominal surface using patch compound from CAD. I have created a fitting element for this surface. I have then created a surface profile tolerance of this surface. 

 

The profile tolerance is an unequally disposed tolerance of .023U.003 to datum A. I am using ASME Y14.5 for my standard. The reported result looks like this:

 

The statistics look like this:

If I manually calculate the tolerance from the statistics I get an actual tolerance of .071". I have triple checked my manual calculation with respect to ASME Y14.5.1 2019. 

Why do I get a different value from the reported value?

 

-Paul

 

[[Ch...]]

Without a closer look look into your project it is hard to give a good answer. Therefore please contact you local partner to view your project together.

[[Pa...]]

Please sign in to view this username.

 I am working with my local partner. I wanted to post this to the community at large in order to see if anyone else has run into this issue as well. The actual measurement should just be calculated from the Min/max distance. The calculation is straightforward according to ASME Y14.5.1, You need to calculate the growth parameter g.  In the above example g = actual value - U tolerance. In this case g=.024". The reported actual tolerance should be 2g + tolerance or .071". In the above example the actual tolerance is reported as .107". So either I am not calculating this correctly or the software is not calculating this correctly. I have made a spreadsheet calculator and have checked results against zeiss inspect. In most other cases Zeiss inspect calculates the correct result. In the above case it seems like the software is taking the max and min and adding them together. Then it is doubling that value. Here is a screenshot of my calculator:

 

Edited July 7
Put in wrong value. Again.

[[Ti...]]

I don't play games with unequally disposed. I modify the CAD model so nominal is where the part should be made. Tolerancing in unequal directions is ridiculous. Maybe it makes engineers feel important.

I have machinist come in all day long looking at CT scans and many times we don't even look at the data, we look at the mesh vs the CAD model blend. I can tell how much parts are out by just looking at the blend after 3 years of this.

Measured value of surface profile is simply 2x your worst deviation.  When I look at my profile stats the worst deviation is half the measured value. All bets are off when you go unequally disposed. PCDMIS for instance creates a new nominal surface in the middle of the tolerance zone so that it does the math based on equally disposed. If I was forced to use an unequally disposed profile tolerance and saw odds numbers I wouldn't even waste 1 minute on it. I absolutely despise unequal non-sense.

When I make a profile in GOM 2021 unequal the statistics do not match the measured value...ie it is not 2x the worst deviation in the stats.  Also, it is not adding the min max, b/c all those variations don't math up in when I do unequally disposed.  Just be prepared to not get an answer, but if you can redo the CAD model and apply a normal tolerance and the statistics will add up for you.