A problem measuring distances

[bherrin]

I had a part come in to me with a number of slots that they needed the width of. So i take a point on one side, a point on the other, no problem. Except the results aren't right. An example:

 ASSIGN/Z=-3.99
PNT6       =FEAT/POINT,CARTESIAN
            THEO/<8.0851,5.8909,-3.99>,<0,-1,0>
            ACTL/<8.0861,5.8918,-3.9898>,<0,-1,0>
            MEAS/POINT,1,WORKPLANE
              HIT/BASIC,NORMAL,<8.0851,5.8909,Z>,<0,-1,0>,<8.0861,5.8918,-3.9898>,USE THEO=YES
            ENDMEAS/
PNT7       =FEAT/POINT,CARTESIAN
            THEO/<8.3187,5.4223,-3.99>,<0,1,0>
            ACTL/<8.3176,5.4215,-3.9902>,<0,1,0>
            MEAS/POINT,1,WORKPLANE
              HIT/BASIC,NORMAL,<8.3187,5.4223,Z>,<0,1,0>,<8.3176,5.4215,-3.9902>,USE THEO=YES
            ENDMEAS/
DIM DIST3= 2D DISTANCE FROM POINT PNT6 TO POINT PNT7     (CENTER TO CENTER),NO_RADIUS  UNITS=IN,$
GRAPH=OFF  TEXT=OFF  MULT=10.00  OUTPUT=BOTH
AX       MEAS    NOMINAL       +TOL       -TOL        DEV     OUTTOL
M       0.5242     0.5200     0.0110     0.0100     0.0042     0.0000 ---#-

The problem being that i can take a .533 gauge block and slide it in this slot with a little wiggle room to spare.

Can someone please help me to see what i'm doing wrong here?

[anthony.alfaro]

check your workplane.

2d distances are measured relative to a workplane.

[bherrin]

I just re-ran it under every workplane, and the results were even further off the mark

[Darrollh]

Measure the slots as either a round or square slot.

[Matthew D. Hoedeman]

Measure the slots as either a round or square slot.

with 3 sample hits

[kingsld1]

Instead of "Center to Center" on your distance dimension, try using the appropriate axis, i.e. X, Y or Z. You will have to be in the right work plane for this. If you're looking for X or Y then you'll need to be in a Z work plane.

[michaelwildschutz]

it is because of your dimension..
re do it and select the axis that it is in Y or X

[UKCMM]

Your slot appears to be at 26.5 degree but your hit vectors do not reflect this as they are 0,1,0 this will cause probing error.

[bherrin]

If you don't mind the noob question, is there a way to figure out the appropriate angle and change the vectors accordingly. I have 12 of these notches spaced evenly around a circle.

[UKCMM]

As you have 12 slots I would think the centre line of 4 off of the slots are in line with the X and Y datum alignment if this is the case then you can use the paste with pattern function.

To do this pick a slot say the one at the 12 o'clock position and make a hit on the left face setting the X,Y to the correct distance and the vector at 1,0,0. Now do the same on the opposite face with the correct X,Y and vector at -1,0,0.

It is critical that the X and Y values are identical and the vectors correct.

Now go to the pattern function and were it has Angle enter 30 degree. Cannot remember the exact wording but there is a box for the number of repeats at 30 degree for 12 slots that will be 11 as you have already one done.

Highlight the lines of code you want to copy , right click mouse and select copy. now un-highlight the code and make sure your curser is at the bottom of the programme.

Go back to the pattern function and select paste with pattern. this should now equally space the hits with the correct vectors.

You will have to add the navigation between the groups of hits using clearplane or move points.

I would add that I would check slots as Schlag suggested as is more robust method the pattern function would work with that as well.

[Schlag]

How big is the slot ? I always tried to do a line and then measure 1 line perp to the other. his will give you better reading if your slot isnt perfect in orientation to your alignment ?

[Quality ish]

Just touch off the walls and radii in multiple places on each aspect of the slot, and it'll tell you what the length and width are, as long as you touch off enough points for it to register as a slot...

Also, if the orientation of the slot is anything other than perfectly along the X or Y axis, you're going to have dimensional issues. Orient off the slot itself and check the X or Y dim. further, if the cut through the material is other than perpendicular to the plane it's going through, you may have issues. If the material is thick enough, maybe measure each as a plane and check the distance between them that way?

I'm just spitballing here. Hope my thoughts are helpful.

If that doesn't give you better results, check and make sure you've got the right probe on the unit.

[AndersI]

If your PC-DMIS is new enough, you can construct a max inscribed 2D WIDTH (BFRE) feature from two lines in the slot (one on each side).

[JacobCheverie]

bherrin You should know the angle to the feature from your print or from your CAD. I have typed up a small paper that shows how you can get the basic equations necessary to calculate the normal vectors of your points, planes, etc... Hope this helps.

Normal Vectors.pdf

[bherrin]

That's one of my bigger issues, I have no print or CAD, just a small note in an overhaul manual. I do look forward to that paper though when i get a minute.

[JacobCheverie]

bherrin You can orient your part so you have four slots along your X/Y directions. Create an alignment by level on face, rotate to line (X) formed by four points (one on each slot face from slots along X axis, just don't change Y value), origin Z on face and XY on center of part. Now you can choose any slot, probe a point on each face, construct midpoint, and find the angle to the midpoint.

As has been mentioned, if you have twelve slots, assuming that they are equally spaced will imply 360/12 = 30 deg apart.

Once angle is determined you can calculate your approach vectors and you should have a little more luck.

[louisd]

seems like you need to level and rotate to something normal to the plane and square to the slot in which you need to take the 2d distance along.
think of it as setting the part on a surface plate, then using the calipers, laying on the surface plate, to measure the width.

--Right now you are just shoving the calipers in and might have them crooked.

Measure a plane on the top of your part. (3 pts min)
measure something to rotate to on your part (2 pts min), but it can even be a couple holes, then construct a line between the holes
then measure a point on the edge or center of your part to origin.
press CTL+ALT+A to open the alignment tool.
level to your plane (should be Z+ is my guess)
rotate to your line
origin x y and z respectively (z would normally be the plane, X your line or point, and Y being whatever you didn't select for X). if your part is round, origin x/y to the most important circle that makes the center of your part.
click OK.
now measure your distance output relative to x or y axis, not point-to-point.