Is the problem me or the print?

[vizionary357]

I have this part on the fixture. I've picked all of the datum features as shown in the image, and I've got a solid alignment. The problem is, when I go to dimension, I get an error that states "secondary is neither parallel nor perpendicular to primary therefore tertiary is not allowed"

Am I doing something wrong? I set datum targets A1-A4 as a plane and made it datum A, the 20.1 mm circle is datum B and the 10.1 mm circle at the top is Datum C. I've never seen this error while programming before which leads me to believe that the datum structure per the print has some sort of error. But I'm not certain. I could be in error. Can anyone chime in on this and shoot me some ideas on how I can fix it so I can proceed with my program?

datum_structure.jpg

[dph51]

Did you rotate to a line between -B- & -C- ?

[vizionary357]

Yep, I still get the same error.

[Matthew D. Hoedeman]

Do an iterative alignment, then mark "USE CURRENT ALIGNMENT" when making your Xact dimensions.

[vizionary357]

I saw your comment and tried that as well. I'm still getting the same error.

[BKulpa]

I think B needs to be measured as a cylinder to restrict all degrees of freedom.

[Schlag]

on a complex shape shouldn't -A- do that ?

[louisd]

I presume your error comes up only when you go to dimension the true position of Datum C Ø10.1? Does it appear when you measure the profile callout?
If so, it's because you can't measure the C datum feature to basic XYZ offsets without a rotational constraint controlled.

I blame this on the drawing. Without a defined datum that controls rotation (aside from the implied orthogonal drawing center-line, which doesn't exist), you cannot locate C datum feature with both an X and a Y offset. You can locate with either ONE axis or a polar radius from B, but not Polar angle or a second axis.

I would measure a cursory C circle. Construct a line, between B and your cursory C diameters, and define that as a 'for now' C datum.
Re-measure C diameter. Report true position to A B and your 'for now' C.
If you absolutely need to output true position relative to the BASIC XYZ coords, then you'll have to trig out your BASIC rotation angle (lengths of all sides are known, so this is easy) and align/rotate to that basic offset angle to square up your rotation to the implied view.

[Matthew D. Hoedeman]

stupid sheet metal stamping stuff.....

Step 1: Align the fixture

Easiest way to check if it is good is a very simple 2D distance between B & C, that is, after all, all that they are looking for, +/-0.025 distance between the holes.

Step 2: Load the part on the fixture, leave the clamps open, but put the 4way & 2way locator pins in. Check the 4 "A" datum targets (circle F = free state check). Clamp the part, pull the pins, measure the 4way, origin on it, check the 2way, then either TP it (to current alignment) or 2D distance it.

[R2ah1ze1l]

Your plane points are at different heights, trap the deviations to create a plane, don't just pop A1-A4 into a constructed plane?

This could be throwing your -B- off because it's tilting in a manor it wasn't meant to.

[vizionary357]

"Trap the deviations to create a plane"

I'm not sure what you mean by this.

[R2ah1ze1l]

Say point A1 is 1.08", point A2 is 1.32", A3 is 1.22", A4 is .5"

They all are coming from some even location. You can use generic features to drive them back to a common location, that would leave the tilt based on deviation from target locations. I will attempt to find an example in my past work.

[R2ah1ze1l]

A1         =FEAT/CONTACT/VECTOR POINT/DEFAULT,CARTESIAN
            THEO/<5,10,1.08>,<0,0,1>
            ACTL/<5,10,1.0805>,<0,0,1>
            TARG/<5,10,1.0805>,<0,0,1>
            SHOW FEATURE PARAMETERS=NO
            SHOW CONTACT PARAMETERS=YES
              AVOIDANCE MOVE=NO,DISTANCE=0
            SHOW HITS=NO
GEN_A1     =GENERIC/POINT,DEPENDENT,CARTESIAN,$
            NOM/XYZ,<A1.TX,A1.TY,0+0>,$
            MEAS/XYZ,<A1.X,A1.Y,A1.Z-1.08>,$
            NOM/IJK,<A1.TI,A1.TJ,A1.TK>,$
            MEAS/IJK,<A1.I,A1.J,A1.K>
$$ NO,

A2         =FEAT/CONTACT/VECTOR POINT/DEFAULT,CARTESIAN
            THEO/<-5,10,1.32>,<0,0,1>
            ACTL/<-5,10,1.3185>,<0,0,1>
            TARG/<-5,10,1.3185>,<0,0,1>
            SHOW FEATURE PARAMETERS=NO
            SHOW CONTACT PARAMETERS=YES
              AVOIDANCE MOVE=NO,DISTANCE=0
            SHOW HITS=NO
GEN_A2     =GENERIC/POINT,DEPENDENT,CARTESIAN,$
            NOM/XYZ,<A2.TX,A2.TY,0+0>,$
            MEAS/XYZ,<A2.X,A2.Y,A2.Z-1.32>,$
            NOM/IJK,<A2.TI,A2.TJ,A2.TK>,$
            MEAS/IJK,<A2.I,A2.J,A2.K>
$$ NO,

A3         =FEAT/CONTACT/VECTOR POINT/DEFAULT,CARTESIAN
            THEO/<5,-10,1.22>,<0,0,1>
            ACTL/<5,-10,1.22>,<0,0,1>
            TARG/<5,-10,1.22>,<0,0,1>
            SHOW FEATURE PARAMETERS=NO
            SHOW CONTACT PARAMETERS=YES
              AVOIDANCE MOVE=NO,DISTANCE=0
            SHOW HITS=NO
GEN_A3     =GENERIC/POINT,DEPENDENT,CARTESIAN,$
            NOM/XYZ,<A3.TX,A3.TY,0+0>,$
            MEAS/XYZ,<A3.X,A3.Y,A3.Z-1.22>,$
            NOM/IJK,<A3.TI,A3.TJ,A3.TK>,$
            MEAS/IJK,<A3.I,A3.J,A3.K>
$$ NO,

A4         =FEAT/CONTACT/VECTOR POINT/DEFAULT,CARTESIAN
            THEO/<-5,-10,0.5>,<0,0,1>
            ACTL/<-5,-10,0.5>,<0,0,1>
            TARG/<-5,-10,0.5>,<0,0,1>
            SHOW FEATURE PARAMETERS=NO
            SHOW CONTACT PARAMETERS=YES
              AVOIDANCE MOVE=NO,DISTANCE=0
            SHOW HITS=NO
GEN_A4     =GENERIC/POINT,DEPENDENT,CARTESIAN,$
            NOM/XYZ,<A4.TX,A4.TY,0+0>,$
            MEAS/XYZ,<A4.X,A4.Y,A4.Z-0.5>,$
            NOM/IJK,<A4.TI,A4.TJ,A4.TK>,$
            MEAS/IJK,<A4.I,A4.J,A4.K>
$$ NO,

PLN_A      =FEAT/PLANE,CARTESIAN,OUTLINE,NO
            THEO/<0,0,0>,<0,0,1>
            ACTL/<0,0,-0.0002>,<-0.0001,0.000025,1>
            CONSTR/PLANE,BF,GEN_A1,GEN_A2,GEN_A3,GEN_A4,,
            OUTLIER_REMOVAL/OFF,3
            FILTER/OFF,WAVELENGTH=0
$$ NO,

I forced some deviations in -Z- measured axis. I recreate the points and offset the [BSC] depths of my target locations to make them 1 equal location. Now when the plane is created, it is reacting to the deviations.

This can also be completed using an iterative alignment. I prefer to SEE the information rather than rely on software interpreting information I can't re-create later.
HTH GL