This Is For Matt D.

you said always use perpen to center line in TP check this out.

 DIM LOC3= TRUE POSITION OF CIRCLE F5  UNITS=IN ,$
GRAPH=OFF  TEXT=OFF  MULT=10.00  OUTPUT=BOTH  FIT TO DATUMS=OFF  DEV PERPEN CENTERLINE=ON  DISPLAY=DIAMETER
AX    NOMINAL       +TOL       -TOL      BONUS       MEAS        DEV     OUTTOL
X      0.8770                                      0.8770     0.0000           
Y      0.0000                                     -0.0008    -0.0008           
Z      0.8320                                      0.8325     0.0005           
DF     0.5000     0.0100     0.0100     0.0102     0.4998    -0.0002     0.0000 ---#----
TP        MMC     0.0200                0.0102                [COLOR="Red"]2.5766 [/COLOR]    2.5464 ------->
END OF DIMENSION LOC3   

DIM LOC3= TRUE POSITION OF CIRCLE F5  UNITS=IN ,$
GRAPH=OFF  TEXT=OFF  MULT=10.00  OUTPUT=BOTH  FIT TO DATUMS=OFF  DEV PERPEN CENTERLINE=OFF  DISPLAY=DIAMETER
AX    NOMINAL       +TOL       -TOL      BONUS       MEAS        DEV     OUTTOL
X      0.8770                                      0.8770     0.0000           
Y      0.0000                                     -0.0008    -0.0008           
Z      0.8320                                      0.8325     0.0005           
DF     0.5000     0.0100     0.0100     0.0102     0.4998    -0.0002     0.0000 ---#----
TP        MMC     0.0200                0.0102                [COLOR="red"]0.0019 [/COLOR]    0.0000 #-------
END OF DIMENSION LOC3

[AJHVW93]

Are you talking about the head heathen? sometimes they will match

[MIKEY]

i think I know why: I think it's because I am using a generic cir. where I only use the extracted values so when it calculates perpen. to center line it's uses the theoretical values which in this case some number because i never modified the theo. values of my generic feature. Could that be it? I just did something: I placed v1, v2 and v3 in the theo. of the generic circle. I think it fixed it.

[AJHVW93]

sounds right!!!

[MIKEY]

it didn't fix the problem, the value is better now yet very different

[MIKEY]

 Dim Loc9= True Position Of Circle F1  Units=in ,$
Graph=off  Text=off  Mult=10.00  Output=both  Fit To Datums=off  Dev Perpen Centerline=off  Display=diameter
Ax    Nominal       +tol       -tol      Bonus       Meas        Dev     Outtol
X     -0.8770                                     -0.8770     0.0000           
Y      0.0000                                      0.0007     0.0007           
Z      0.8320                                      0.8325     0.0005           
Df     0.5000     0.0100     0.0100                0.4999    -0.0001     0.0000 ---#----
Tp        MMC     0.0200                0.0000                [color="red"]0.0018[/color]     0.0000 #-------
End Of Dimension Loc9
Dim Loc1= True Position Of Circle F1  Units=in ,$
Graph=off  Text=off  Mult=10.00  Output=both  Fit To Datums=off  Dev Perpen Centerline=on  Display=diameter
Ax    Nominal       +tol       -tol      Bonus       Meas        Dev     Outtol
X     -0.8770                                     -0.8770     0.0000           
Y      0.0000                                      0.0007     0.0007           
Z      0.8320                                      0.8325     0.0005           
Df     0.5000     0.0100     0.0100                0.4999    -0.0001     0.0000 ---#----
Tp        MMC     0.0200                0.0000                [color="red"]0.0056 [/color]    0.0000 --#-----
End Of Dimension Loc1

[MIKEY]

makes no f ing sence

[MIKEY]

here is my code (partial) Oh and it's a round part with 8 windows cut out on the sides of the part. There is no rotation at this point, just Leveled to A and X Y zeroed to the circle.

LIN1       =FEAT/LINE,CARTESIAN,UNBOUNDED
            THEO/<-0.877,0.2625,-0.5806>,<0,-0.0050286,0.9999874>
            ACTL/<-0.877,0.252,-0.5808>,<0,-0.0008748,0.9999996>
            MEAS/LINE,2,WORKPLANE
            HIT/BASIC,NORMAL,<-0.877,0.2625,-0.5806>,<-0.0000004,-0.9999874,-0.0050254>,<-0.877,0.252,-0.5808>,USE THEO = YES
            HIT/BASIC,NORMAL,<-0.877,0.2613,-0.3417>,<-0.0000004,-0.9999874,-0.0050254>,<-0.877,0.2518,-0.3419>,USE THEO = YES
            ENDMEAS/
LIN2       =FEAT/LINE,CARTESIAN,UNBOUNDED
            THEO/<-0.877,-0.2454,-0.3415>,<0,0.0004678,-0.9999999>
            ACTL/<-0.877,-0.248,-0.3415>,<0,0.0003269,-0.9999999>
            MEAS/LINE,2,WORKPLANE
            HIT/BASIC,NORMAL,<-0.877,-0.2454,-0.3415>,<0.0000004,0.9999999,0.0004677>,<-0.877,-0.2478,-0.5204>,USE THEO = YES
            HIT/BASIC,NORMAL,<-0.877,-0.2453,-0.5204>,<0.0000004,0.9999999,0.0004677>,<-0.877,-0.2478,-0.5204>,USE THEO = YES
            ENDMEAS/
DIM DIST1= 2D DISTANCE FROM LINE LIN1 TO LINE LIN2 PAR TO   YAXIS, NO_RADIUS  UNITS=IN ,$
GRAPH=OFF  TEXT=OFF  MULT=10.00  OUTPUT=NONE
AX    NOMINAL       +TOL       -TOL       MEAS        DEV     OUTTOL
M      0.5000     0.0050     0.0050     0.4999    -0.0001     0.0000 ---#----
            ASSIGN/V4 = DIST1.MEAS
PNT18      =FEAT/POINT,CARTESIAN,NO
            THEO/<-0.877,0,-0.4461>,<0,0,1>
            ACTL/<-0.877,0.002,-0.4461>,<0.0000836,0.0000341,1>
            CONSTR/POINT,MID,LIN1,LIN2
            WORKPLANE/YMINUS
            MOVE/POINT,NORMAL,<-1.093,0.0293,-0.6393>
            MOVE/POINT,NORMAL,<-1.0926,0.0294,0.2124>
PNT17      =FEAT/POINT,CARTESIAN
            THEO/<-0.9165,0.0294,-0.0034>,<0,0,1>
            ACTL/<-0.9165,0.0294,-0.0001>,<0,0,1>
            MEAS/POINT,1
            HIT/BASIC,NORMAL,<-0.9165,0.0294,-0.0034>,<0,0,1>,<-0.9165,0.0294,-0.0001>,USE THEO = YES
            ENDMEAS/
            MOVE/POINT,NORMAL,<-0.9165,0.0298,2.4575>
            TIP/T1A90B90, SHANKIJK=-1, 0, 0, ANGLE=90.008
            MOVE/POINT,NORMAL,<-2.3093,0.0571,6.9771>
            MOVE/POINT,NORMAL,<-1.2376,-0.01,-0.4276>
PNT32      =FEAT/POINT,CARTESIAN
            THEO/<-0.8949,-0.01,-0.8169>,<0,0,1>
            ACTL/<-0.8949,-0.0101,-0.8139>,<0,0,1>
            MEAS/POINT,1
            HIT/BASIC,NORMAL,<-0.8949,-0.01,-0.8169>,<0,0,1>,<-0.8949,-0.0101,-0.8139>,USE THEO = YES
            ENDMEAS/
PNT33      =FEAT/POINT,CARTESIAN
            THEO/<-0.8949,-0.0095,-0.1009>,<0,0,-1>
            ACTL/<-0.8949,-0.0095,-0.1037>,<0,0,-1>
            MEAS/POINT,1
            HIT/BASIC,NORMAL,<-0.8949,-0.0095,-0.1009>,<0,0,-1>,<-0.8949,-0.0095,-0.1037>,USE THEO = YES
            ENDMEAS/
PNT49      =FEAT/POINT,CARTESIAN,NO
            THEO/<-0.8949,-0.0097,-0.4589>,<0,0,1>
            ACTL/<-0.8949,-0.0098,-0.4588>,<0,0,1>
            CONSTR/POINT,MID,PNT32,PNT33
            ASSIGN/V1 = PNT18.X
            ASSIGN/V2 = PNT18.Y
            ASSIGN/V3 = PNT49.Z
F1         =GENERIC/CIRCLE,DEPENDENT,CARTESIAN,OUT,$
            NOM/XYZ,<V1,V2,V3>,$
            MEAS/XYZ,<V1,V2,V3>,$
            NOM/IJK,<1,0,0>,$
            MEAS/IJK,<1,0,0>,$
            DIAMETER/0.5,V4

[John Kingston]

You are creating a generic circle from the X and Y location of one point and the Z of another point?
All you are doing is creating a location for it. Not a size and not true vectors.
If the vectors of those points varies, but you use 1,0,0 for the generic, the TP will vary.
What exactly are you doing? There has to be a better way.

[MIKEY]

yeah, hmm I am gonna try to explain: ok I have a TP callout on the square looking slot, so I am measuring the width---> creating a mid point ---> measuring the length--> creating a midpoint----> creating a generic circle with a diameter or the width and X Y location of the midpoint of the width and with a Z location of the Lenght.

Basicly I need Y(width) and Z(comes fron the leength) to calculate the TP.

[vpt.se]

Deviation perpendicular to centerline also takes the nominal and measured vectors into account.

[MIKEY]

well, Matt told me that he always uses perpen. to center line option under the TP.

Now I trust Matt, he is really good, but in this case when I use perep. to center line as you can see in my previous post the number do not calculates properly unless I unchech perpend. to center line.

[John Kingston]

Just for shiits and giggles, try this.

1) Take a line on each of the sides. (2 hits per line)
2) Construct an intersect point at each corner.
3) Construct 2 lines, using the corner points opposite each other.
4) Construct an intersect point from those two lines.

(note: you could use midlines instead of the corner points. I am at home and thinking this up on the fly.)

Calculate the TP using math, not the software.

Another question came to mind as I typed this.
Can't you use the slot feature? That should do what you are trying to accomplish.

[MIKEY]

I have never used a slot feature, but s h it thats pretty goodd, why didn't I think of that.

I am at home now, but I will try that thanks.

[MIKEY]

although the result should be the same