ASME vs ISO surface profile calculation

[[Ti...]]

Question:

Does anyone know the calculation difference b/w the two?  Not a guess, but really know what is going on?

 

The problem:

I am testing 2025 Zeiss xray from 2021 GOM and some of our big programs went from 2:30min calculation times to 13:00min calculation times. After some analysis, I found it is due to ASME no datum surface profiles taking up the balance of the difference. I ran a 20 mesh study of ASME vs ISO and 95% of the profiles measure almost identical.

The profile that did not measure identical had datums and measured 50% larger in ISO.  I have 15 surface profiles in this program and there are plenty of other profiles with datums that measure the same.

 

I am going to run the part in pcdmis with ASME vs ISO and see what I get on the CMM.  Curious if anyone has definitive knowledge on the math behind the two. 

 

Thanks.

[[Pe...]]

Overall I don't know the math but my experience is that if your CAD has some elements inside the part the calculation period is increasing. It's true that we have only an Atos capsule 12M sensor so not  an x-ray as yours. And if you evaluate a profile surface tolerance with datums in my opinion the calculation will be always slower because of the datums. First we usually delete the inside elements from the CAD that we don't need for the evaluating and after then lauch the profile tol. calculation.

[[Ti...]]

Please sign in to view this username.

  It is just the opposite with the 2025 Zeiss inspect software...when there are no datums the ASME profiles are taking 10x longer to calculate. We have many projects that are going from 3 min to 13 min in the new software. Once I remove the profiles with no datums the report takes less than 3 min. 

 

My opinion is this is b/c it is trying to do a best fit math to minimize the measured value.  With datums the math is easier I would think, it is more black and white.

[[Ti...]]

I tend to think that profile is measured in ISO and ASME the same way. 2x the worst point on the surface.

I say this b/c profile of a line does not have an ISO vs ASME selection for the standard in GOM and the below explanation seems to be discussion how they differ if you select a cylinder for your surface profile...we do not select geometric features for our profiles we create separate surface patches to which you cannot apply a guassian or chebychev filter.  The surface is where it is no matter what. 

 

[[Ch...]]

The main difference between the two standards is what you have to do when your actual data is bigger than the nominal. Typically approaches are

* measuring the distance to the nearest CAD triangle => ISO

* interpolate the CAD so that the size between actual and nominal is comparable and afterwards compute the distance to the interpolated CAD => ASME

 

In 2025 we have to change the computation algorithm due to rare problems with the old approach. We have expecting slightly longer computation times but nothing in the area of 5 times longer. In our internal test we do not found such increase in computation time, but in our internal test we do not found commutation times that are factor 5 or higher. Therefore if possible please provide the project via your local partner to allow us to have an deeper look on the performance issues. Of course we can sign a NDA if needed.

 

Best regards

Christoph Schult

[[DW...]]

Please sign in to view this quote.

Please sign in to view this username.

 The standard for surface profile evaluation is a bilateral zone with one result. That result is the maximum deviation, doubled. This always yields a positive integer, which in my opinion is not helpful at all from a diagnostic stand point. When guys are troubleshooting a parts condition, they usually want a bilateral zone with two results (min/max). Anyway, perhaps you are on to something with the way the software is performing the surface profile with no datum calculation (form only evaluation, the surface with respect to itself). It would be exactly what you described, a best fit alignment (imagine wiggling the actual surface around on the CAD until the difference between them was the smallest). I'm not sure why that would take as long, might be a bug? 

Edited Monday at 02:46 PM