PR 251525 Line vs Set for Symmetry

huangyhg

Line vs Set for Symmetry
<< Andrew Dear  --  11/08/07  17:29:57>>
The inputs that are supported by Symmetry, as the non datum item(s), are Line-Line or Set.  The Line-Line entry gives incorrect values and the Set entry gives the correct values.  A test prg is attached which shows this issue(s).

Steps to reproduce:
Run, or replicate, the attached .prg in v4.2 (I used 7th Nov Beta).
Results:
For the same numerical inputs, different outputs are calculted by using Line-Line / Set.

Expected Results:
Numerically identical inputs to give equal outputs (results)
<<END>>


<< Changes made by Neil Kay -- 02/19/09  12:51:58>>
Status:  RESOLVED to CLOSED
<<END>>
<< Jerry Naylor  --  08/15/08  09:37:17>>
wasaddedtoreadme43B
<<END>>
<< Changes made by Neil Kay -- 07/23/08  14:04:40>>
Action:  David Petrizze to Andrew Dear, Status:  REVIEW to RESOLVED
<<END>>
<< Don Turcotte  --  12/03/07  11:02:22>>
Reviewed.
OK
<<END>>
<< Changes made by Don Turcotte -- 12/03/07  11:02:22>>
Action:  Don Turcotte to David Petrizze
<<END>>
<< Paola Pallo  --  11/30/07  17:10:22>>
Fixed.
I reworked GetEndPoints() method used for the line-line case in the FCF math: now the results from FCF are coherent with the ASME definitions , and for LIN01 and LIN02 to LIN03 is 2.622 ( not 0  as written by Don's note, where he considered the nominal line without re-executing it), and it is coherent with FCF design.
For review:
------------------------
V43B\DIMENS\DIM_SYMM.CPP
V43B\INCLUDE\DIM_SYMM.H
V44B\DIMENS\DIM_SYMM.CPP
V44B\INCLUDE\DIM_SYMM.H
<<END>>
<< Changes made by Paola Pallo -- 11/30/07  17:10:49>>
Action:  Paola Pallo to Don Turcotte, Status:  OPEN to REVIEW
<<END>>
<< Paola Pallo  --  11/29/07  18:23:24>>
Code modification in progress
<<END>>
<< Paola Pallo  --  11/28/07  18:28:30>>
Phone talk with Don to clarify FCF math. Code changes for FCF math not yet completed.
<<END>>
<< Don Turcotte  --  11/27/07  16:52:28>>
Reviewed.
These calculations still have problems.
Based on the definition of symmetry in "ASME Y14.5M-1994 Dimensioning and Tolerancing" section 5.14 (also the example in Fig. 5-61), I would expect the symmetry of LIN01 and LIN02 to LIN03 to be 0 since LIN01 and LIN02 are perfectly symmetrical about LIN03.  Also the symmetry of LIN01 and LIN02 to PLN01 should also be 0 since LIN02 and LIN02 are perfectly symmetrical about the plane (they are on opposite sides of the plane and the same distance from the plane).  
For the set, PC-DMIS expects points of the set to alternate on different sides of the datum feature.  The centerpoint of each of these pairs of points is compared to the datum feature.  The first centerpoint from (MP1,MP2) is 1.25 from the LIN_1 or PLN_1 datum.  The other center points from  (MP3, MP4) and (MP5,MP6) are on the LIN_1 or PLN_1 datum.  So the symmetry is correctly reported as 2.5.
From "ASME Y14.5.1M-1994 Mathematical Definition of Dimensioning and Tolerancing", section 5.7, the actual value of symmetry deviation is the smallest tolerance value to which the feature will conform.  The tolerance zone is a cylindrical or parallel plane volume centered on the datum feature.  All the midpoints must lie within this tolerance zone.  So the tolerance value is 2 times the max deviation of any center point from the datum feature.
<<END>>
<< Changes made by Don Turcotte -- 11/27/07  16:53:08>>
Action:  David Petrizze to Paola Pallo, Status:  REVIEW to OPEN
<<END>>
<< Changes made by Don Turcotte -- 11/27/07  16:52:28>>
Action:  Don Turcotte to David Petrizze
<<END>>
<< Paola Pallo  --  11/23/07  17:45:09>>
I saw more issues here, related to the math for FCF line-line.
The legacy result is
DIM SYM01= SYMMETRY OF LINE LIN01 AND LINE LIN02 TO LINE LIN03  UNITS=MM ,$
GRAPH=OFF  TEXT=OFF  MULT=10.00  OUTPUT=BOTH
AX    NOMINAL       +TOL       -TOL       MEAS        DEV     OUTTOL
M       0.000      2.000      0.000      5.000      5.000      3.000 ------->
On the contrary FCF result on the same features is 0 (wrong) : I added a fix for this.
In any case there is still  a difference with the legacy result,and I would discuss it with Don , together with the possibility to add a specific label ('legacy_dimension_simmetry.lbl')  to get a report header fully correct . Actually I added a fix to get the datum name, but it is not  correctly used inside the legacy_dimension.lbl.
For review :
V43B\DIMENS\DIM_SYMM.CPP
V44B\DIMENS\DIM_SYMM.CPP
<<END>>
<< Changes made by Paola Pallo -- 11/23/07  17:45:57>>
Action:  Paola Pallo to Don Turcotte, Status:  OPEN to REVIEW
<<END>>
<< Paola Pallo  --  11/22/07  18:03:57>>
Investigating in the different do_math() cases ( FCF math for  line-line case was recently modified) .
I agree that legacy symmetry report should display also the datum feature's name, and I added a fix to get it from the report side.
<<END>>
<< Andrew Dear  --  11/09/07  09:16:49>>
On a related note - when using the legacy dimn, no reference to the datum or the name of the datum feature is present within the report window entry. If this is intentional, my apologies, but it would be better if it were included?  Kind regards, Andrew.
<<END>>
<< Changes made by Tim Wernicke -- 11/08/07  14:55:24>>
Action:  Tim Wernicke to Paola Pallo, Assigned:   to Paola Pallo, Priority:  to Critical
<<END>>