PR 261316 Graphical Display of Profile is located at center of ball.

huangyhg

Graphical Display of Profile is located at center of ball.  
<< Doug Schueneman  --  04/22/09  13:49:30>>
There is a Profile in the attached part, made from a body axis scan using the 38.1mm reflector.
If you look at the graphical display of the Profile dimension on the screen it is drawn around a line made at the center of the ball.
This looks terrible, and begs all sorts of questions. questions like what is the software using for a nominal?  Doesn't look like it is using CAD.  What is going on here...  etc, etc, etc.
Expected Results:
The graphics should be shown using the CAD model, or a section fromt the CAD as a nominal with the tolerance shown around it.
<<END>>


<< Changes made by Don Turcotte (Field Changes) -- 04/23/09  11:23:38>>
Action:  Don Turcotte to Ryan Gundry, Status:  OPEN to REVIEW
<<END>>
<< Changes made by Don Turcotte (Resolution Data) -- 04/23/09  11:23:06>>
** Bug Fix - New In Current Version -  -  -  **
<<END>>
<< Don Turcotte (Change Sets)  --  04/23/09  11:22:51>>
Fixed in V44B.
Files inserted to server
------------------------
V44B\BASESCAN\BASICSCN.CPP
<<END>>
<< Don Turcotte (Development Notes)  --  04/23/09  11:22:33>>
I have fixed this with help from Ryan.  This is a problem only in V44B since earlier versions don't have the ProbeComp and CadComp checkboxes.
Files inserted to server
------------------------
V44B\BASESCAN\BASICSCN.CPP
<<END>>
<< Bret Naylor (Management Notes)  --  04/22/09  16:41:15>>
Don and Ryan, let's talk about this one.
<<END>>
<< Don Turcotte (Development Notes)  --  04/22/09  16:58:09>>
Sent the following e-mail to Ryan:
"Ryan,
The program attached to this PR has a Body Axis Scan with CadComp checked.  The graphical analysis of the profile dimension on this scan is drawing away from the cad surface by the probe radius.  I have tracked this to CPCDBasicScan::get_points(…) in BASESCAN\BASICSCN.CPP.  This method is called by get_points_for_profile(…) from the graphical_analysis(…) method of the profile dimension to get the THEO points.  The following "if" forces the "type_of_point" to BALLCENTER since no PROBECOMP was checked for this scan.
  switch (type_of_point)
  {
  case SCANROWCOUNT:
    *out_num_points = 1;
    break;
  case SCAN_HITDATAFORROW + TIPRADIUS:
  case SCAN_HITDATAFORROW + CENTROID:
  case SCAN_HITDATAFORROW + BALLCENTER:     // same as OFFSET
  case SCAN_HITDATAFORROW + VECTOR:         // same as ANGLE2
    type_of_point -= SCAN_HITDATAFORROW;
  default:
    CHitsContainer **hits = bGetPath ? &m_pPath : &m_pHitContainer;   
    if(*hits)
      nHitCount=(*hits)->Count();
    if(!GetStatus(BASICSCANSTATUS_PROBECOMP) && type_of_point == CENTROID)
      type_of_point = BALLCENTER;
This results in the THEO points being probe comp'd away from the cad surface.
If I change the body axis scan to do both PROBE COMP and CAD COMP or only PROBE COMP, then the graphical analysis works correctly (the THEO points are on the surface of the cad).  It seems that the profile dimension is working correctly.  The issue is the design intent of PROBE COMP and/or CAD COMP.  I'm not clear what the intent is here.  Do we need a status bit for BASICSCANSTATUS_CADCOMP?  No such bit exists.
"
<<END>>
<< Changes made by Tim Wernicke (Field Changes) -- 04/22/09  13:57:54>>
Category:  Dimension Info to profile, Action:  Tim Wernicke to Don Turcotte, Assigned:   to Don Turcotte, Priority:  Critical to Stop Rel.
<<END>>
<< Changes made by Doug Schueneman (Field Changes) -- 04/22/09  13:57:36>>
Priority:  to Critical
<<END>>

huangyhg

// PR261316 - CADCOMP is not ballcenter
    if(!GetStatus(BASICSCANSTATUS_PROBECOMP) && (GetParentScan() && ((CPCDscan*)GetParentScan())->GetCADCompensation() == false) && type_of_point == CENTROID)
      type_of_point = BALLCENTER;