Which association criterion is used by GOM to find a datum line for a cylinder which covers less than 180°?

[[Er...]]

Hello all, 

ISO 5459 states that 'Maximum inscribed' is the default association criterion used to find the datum center line of a cylindrical hole.

ISO 5459 currently addresses complete cylinders only. However, the 'Maximum inscribed' criterion is applicable to partial cylinders as well, provided that they cover at least 180°, in which case the feature in question is a Feature of Size and opposite point pairs are present.

It turns out GOM is (fortunately!) capable to establish a datum line even if the cylinder covers less than 180°! It cannot use the Maximum inscribed criterion, since inscribed cylinders will keep growing and eventually 'pop out'.

Now my question is:

Which association criterion ('fitting method') does GOM Inspect apply?

• Technically, a partial cylinder which lacks opposite point pairs would fall in the 'Complex surface' category. This would imply that a cylinder with fixed diameter would have to be associated in such a way that the maximum deviation would be minimized.
• Alternatively, Gaussian fit might be used.

Thank you very much for your help!

Best regards,

Erik de Lange

 

 

 

[[Ch...]]

Hi,

 

due to missing an exact specification we have defined the maximum inscribed cylinder as cylinder with the biggest radius that axis lies in the convex hull of the point data. For each zylinder that cover at least 180 degrees this is no constrait at all (because it always fulfill), but for smaller coverage it forbids "pop out" effect. Of course be free to use the gaussain fit in combination with geometry computation method if this is a better solution for you.

 

Best regards

Christoph Schult 

[[Er...]]

Hi Christoph,

 

Thank you for your quick reply.

This sounds as either a Non-negative Least Squares or a MiniMax algorithm.

Could you perhaps shed a little more light on the math that you apply?

(Background of this question: We use GOM scanners for product release and CMMs for day-to-day process control, and would like to know how these compare.)

Best regards,

Erik

 

 

 

 

 

[[Ch...]]

Hi Erik,

 

sorry but I cannot explain the algorithm more in detail. I can only tell that he will find the biggest cylinder inside the convex hull of all points so that all point lies outside of the cylinder, but I cannot tell you anything how we achieve this results because this algorithm is intellectual property of Zeiss. 

I hope you can understood that I cannot tell you anything more.

 

Best regards

Christoph Schult