PR 261642 Cannot output Flatness of any constructed set of (2) Measured Plane

huangyhg

Cannot output Flatness of any constructed set of (2) Measured Planes...
<< Tim Wernicke  --  05/06/09  08:52:51>>
Cannot output Flatness of any constructed set of (2) Measured Planes.
Steps to reproduce:
Open the attached part program and create any type of Constructed Plane out of FEET and FEET2.  Create a Legacy Flatness dimension for example.
Results:
Measured results are 0.000.  Why?
<<END>>

<< Don Turcotte (Development Notes)  --  06/09/09  13:53:51>>
Sent e-mail to Andy:
"Andy,
In #261642, I made a change to allow the user to create an FCF flatness on a constructed set of points.  The example in the user's program was two plane features that they wanted to evaluate as a single set of points.  I implemented a flatnessDSFitCompensate(? method in CPCDcons_set(? to support this.   This method uses the DSFitFlatness algorithm in PCDSFit.dll passing in the array of tip radii.  This is the same algorithm used by all the plane features except BF plane and CPCDAFPlaneLaser which use the flatnessPrecompensate(? method which ignores the tip radii.  
I have found in tests with the program attached to #261642 and also the program attached to #262430 that the flatnessDSFitCompensate(? method calculates a slightly larger flatness value (about 30 microns) larger than the flatnessPrecompensate(? method.  So for example, the FCF flatness on a set constructed from the scan in #262430 returns the same result as the flatness on a BFRE plane constructed from the scan.  This flatness is about 30 microns larger than the FCF flatness on a BF plane constructed from the same scan.  
Is it correct for the FCF flatness on a constructed set to use flatnessDSFitCompensate(? (as opposed to flatnessPrecompensate)?
"
<<END>>
<< Changes made by Don Turcotte (Field Changes) -- 05/26/09  09:23:54>>
Action:  Don Turcotte to Yanhua Huang, Status:  OPEN to REVIEW
<<END>>
<< Changes made by Don Turcotte (Resolution Data) -- 05/26/09  09:23:40>>
** Bug Fix - Likely Exists in Previous Versions -  -  -  **
<<END>>
<< Don Turcotte (Development Notes)  --  05/26/09  09:21:53>>
I have implemented this for FCF flatness and cylindricity using the underlying points of the features of a constructed set.  In addition, PC-DMIS now displays an error message for all cases where there are not enough points to calculate a form deviation (or not any points, for example, as in the case of a constructed plane).
Files inserted to server
------------------------
V44B\CONSFEAT\CONS_SET.CPP
V44B\DIMENS\DIM_FLAT.CPP
V44B\DIMENS\FEATCTRLFRMUTIL.CPP
V44B\INCLUDE\CONS_SET.H
V44B\INCLUDE\FEATCTRLFRMUTIL.H
V44B\MENU\BASEDLG.CPP
V44B\MENU\FEATCTRLFRMDLG.CPP
V44B\DIMENS\DIM_CYLY.CPP
<<END>>
<< Brian Gudauskas (Customer Notes)  --  05/20/09  15:56:02>>
The issue here is that sometimes you might want to use the centers of the input features in a set (bolt hole circle), or the points from those features.  I always thought PC-DMIS should have an elegant way of specifying whether you used the centroid of the hits of the features to construct a set, and when multiple selection is possible.
<<END>>
<< Don Turcotte (Development Notes)  --  05/21/09  12:22:48>>
I have uploaded the changes to allow FCF flatness on a constructed set.  I am looking at implementing this also for cylindricity.  Also needs more testing...
<<END>>
<< Changes made by Bret Naylor (Field Changes) -- 05/20/09  16:38:52>>
Priority:  Critical to Stop Rel.
<<END>>
<< Don Turcotte (Development Notes)  --  05/08/09  10:57:48>>
I have added the error message "Feature XXX does not have enough point data for analysis." when trying to create any form dimension (flatness, circularity, cylindricity, straightness) on any feature that has either no point data or insufficent point data.  I have added this to both the legacy and FCF dialogs.  This is an existing message string resource.
<<END>>
<< Don Turcotte (Development Notes)  --  05/07/09  17:14:27>>
I have this working with a constructed set using FCF's (for flatness) since the FCF uses a minmax algorithm and does not care what the underlying feature is (just needs points).  
I will look at putting up a message when there is no data to create a form dimension.  This will also apply to circularity, cylindricity, and straightness for both legacy and FCF form dimensions.
<<END>>
<< Anders Isaksson (Customer Notes)  --  05/07/09  11:20:13>>
Shouldn't PC-DMIS then say "Impossible to calculate flatnes without hits!" instead of giving the false result 0.0?
<<END>>
<< Don Turcotte (Development Notes)  --  05/06/09  17:00:23>>
This does not work on a constructed plane because flatness requires probe hits.  The flatness dimension could be modified to get the probe hits from the features that make up the constructed plane.  This same problem exists with other types of form as well (circularity, cylindricity, straightness).
Another solution is to allow FCF flatness to be performed on any constructed set of features.  Since FCF flatness is a minmax flatness and does not need a least squares solution plane, FCF flatness could be allowed on any set of points.  This is also true of FCF circularity, cylindricity, and straightness.  I am looking at modifying CPCDcons_set to support this.
Alternatively, we could implement a construction for planes (also circles, cylinders, lines) that collects the measure points (THEO, ACTL) from the selected features.  This way the dimension would not have to be modified since this new type of constructed plane would have measure hits to use for any form calculation (legacy or FCF).
<<END>>