Extended Computation Techniques
Extended Computation Techniques
In 2024.2 a new option has been added to change the computation technique used for a particular geometry type:
When fitting geometry to a measured data set there are always errors to some degree. This can be more significant in portable metrology. The traditional High Point solution specified by the standards are true to true to the constraints of high points controlling the assembly of parts but this does not consider measurement error. On the other extreme, a least squares or RMS fit is the most reboust solution but may ignore actual part deformities. These additional fit solutions offer intermediate options that may better reflect a specific user's needs.
A very brief explanation for each of the feature evaluations are as follows:
-
LeastSquares. L2 is the default method for our GD&T features. This is due to it being the most robust (least failure to evaluate) of the evaluation techniques, and historically was this way from the beginning.
-
MinimumSeparation. L1 normal evaluation. It is particularly useful in the design of thin-walled parts. For cylindrical surfaces, it represents the radial separation between concentric cylinders, for example.
-
HighPoint. L (Infinity) normal evaluation.
-
LeastSquaresHighPoint. Least Squares and then find the highest point and translate the resultant plane to this location.
-
MinimumSeparationHighPoint. Minimum separation and then find the highest point and translated the resultant plane to this location.
-
EqualizedHighPoint. Solving the plane location using ISO5459 where the maximum deviations are equalized.
-
EqualizedLsqHighPoint. Solving the plane location using ISO5459 where the square root of the sum of the squares of the deviations are minimized.
-
LeastSquaresHighPoint1Stddev. Least Squares and then translate the resultant plane 1 standard deviation along the normal.
-
LeastSquaresHighPoint2Stddev. Least Squares and then translate the resultant plane 2 standard deviation along the normal.
-
LeastSquaresHighPointHalfway. Least Squares and then find the highest point and translate the resultant plane 1 half the distance to this location.