Tweet
I often have to measure short cones / large diameter.
I noticed than measuring or constructing a cone with a length of 1 mm gives strange results, and if I change from 1 to 2 mm, results arebetter.
I can't post a real case here, but I created a case which shows the problem.
I constructed a BF cone from 2 ellipes (circles with a little defect !). I used BF because I can't easily construct a circle from hits in this case.
CE1 is the circle constructed from cone at 0.5 mm of origin, and with ref vector as ZPLUS. Why does the vector of the circle isn't <0,0,1> ?
CE2 is constructed from intersection of a line between hits of each ellipse and a plane F2 at Z0.5 mm.
The center seems to be right (half of the Y defect of Ell2), and the center of CE1 is far from it...
Could it be a bug ? If I change the Z of ELL2 to 10, both centers are close together.
(I know that CE2 doesn't represent exactly the intersection of a cone and a plane, so it's logical that results can be different, but not like this !!!!!)
I noticed than measuring or constructing a cone with a length of 1 mm gives strange results, and if I change from 1 to 2 mm, results arebetter.
I can't post a real case here, but I created a case which shows the problem.
I constructed a BF cone from 2 ellipes (circles with a little defect !). I used BF because I can't easily construct a circle from hits in this case.
Code:
ELL1 =FEAT/CONTACT/ELLIPSE/DEFAULT,CARTESIAN,IN
THEO/<0,0,0>,<0,0,1>,300.05,300,<1,0,0>
ACTL/<0,0,0>,<0,0,1>,300.05,300,<1,0,0>
TARG/<0,0,0>,<0,0,1><1,0,0>
START ANG=0,END ANG=0
SHOW FEATURE PARAMETERS=NO
SHOW CONTACT PARAMETERS=NO
ELL2 =FEAT/CONTACT/ELLIPSE/DEFAULT,CARTESIAN,IN
THEO/<0,0.005,1>,<0,0,1>,300.5,300.45,<1,0,0>
ACTL/<0,0.005,1>,<0,0,1>,300.5,300.45,<1,0,0>
TARG/<0,0.005,1>,<0,0,1><1,0,0>
START ANG=0,END ANG=0
SHOW FEATURE PARAMETERS=NO
SHOW CONTACT PARAMETERS=NO
SCN3 =FEAT/SET,CARTESIAN
THEO/<0,0,0.5>,<0,0,1>
ACTL/<0,0.0025,0.5>,<0,0,1>
CONSTR/SET,BASIC,ELL1.HIT[1..20],ELL2.HIT[1..20],,
CON1 =FEAT/CONE,CARTESIAN,OUT,LENG,NO
THEO/<0,-0.7246,-666.5177>,<0,0.0010375,0.9999995>,1.3115,299.9538,300.5441
ACTL/<0,-0.7253,-666.5189>,<0,0.0010385,0.9999995>,1.3118,299.9537,300.5442
CONSTR/CONE,BF,SCN3,,
CE1 =FEAT/CIRCLE,CARTESIAN,IN,NO
THEO/<0,-0.0326,0.5>,<0,0.0010375,0.9999995>,300.2499
ACTL/<0,-0.0326,0.5>,<0,0.0010385,0.9999995>,300.2499
CONSTR/CIRCLE,CONE,CON1,HEIGHT,0.5,REF_POINT = ORIGIN,REF_VECTOR = ZPLUS
F2 =GENERIC/PLANE,DEPENDENT,CARTESIAN,$
NOM/XYZ,<0,0,0.5>,$
MEAS/XYZ,<0,0,0.5>,$
NOM/IJK,<0,0,1>,$
MEAS/IJK,<0,0,1>
V3 =LOOP/START,ID=YES,NUMBER=20,START=1,SKIP=,
OFFSET:XAXIS=0,YAXIS=0,ZAXIS=0,ANGLE=0
ASSIGN/V4=UNIT(ELL1.HIT[V3].XYZ-ELL2.HIT[V3].XYZ)
ASSIGN/V5=DOT(ELL2.HIT[V3].XYZ-F2.XYZ,F2.IJK)
ASSIGN/V6=DEG2RAD(180-ANGLEBETWEEN(V4,F2.IJK))
ASSIGN/V7=ELL2.HIT[V3].XYZ+((V5*V4/COS(V6)))
F3 =GENERIC/POINT,DEPENDENT,CARTESIAN,$
NOM/XYZ,<V7.X,V7.Y,V7.Z>,$
MEAS/XYZ,<V7.X,V7.Y,V7.Z>,$
NOM/IJK,<0,0,1>,$
MEAS/IJK,<0,0,1>
ASSIGN/V9[V3]=V5
LOOP/END
SCN6 =FEAT/SET,CARTESIAN
THEO/<0,0.0025,0.5>,<0,0,1>
ACTL/<0,0.0025,0.5>,<0,0,1>
CONSTR/SET,BASIC,F3[1..20],,
CE2 =FEAT/CIRCLE,CARTESIAN,IN,LEAST_SQR,NO
THEO/<0,0.0025,0.5>,<0,0,1>,300.25
ACTL/<0,0.0025,0.5>,<0,0,1>,300.25
CONSTR/CIRCLE,BF,2D,SCN6,,
OUTLIER_REMOVAL/OFF,3
FILTER/OFF,UPR=0
CE2 is constructed from intersection of a line between hits of each ellipse and a plane F2 at Z0.5 mm.
The center seems to be right (half of the Y defect of Ell2), and the center of CE1 is far from it...
Could it be a bug ? If I change the Z of ELL2 to 10, both centers are close together.
(I know that CE2 doesn't represent exactly the intersection of a cone and a plane, so it's logical that results can be different, but not like this !!!!!)
Comment