Nominal Values when picking from model not correct???

[cappleby1]

Ok I am setting up a program to measure a fixture plate with dowel pins in it.  I have imported the Model and instead of using the world coordinate system I have made my own (to match the print's X & Y Zero is the lower left corner).  I used the perpendicular planes method to create my new coordinate system.  The system looks good and X points to the right and Y is pointing up to the top (just like the print).  When I start to pick the circles on the pins, I notice that some of the nominal values are off by a few .001.  The error looks like it only affects the Y?  This is a fairly simple inspection however I am getting funky numbers.  I really dont see what is casuing this error in my cordinate system, the only work around is to go back and look at every one and manually define the nominal using the number method.

[cappleby1]

Update,  The same problem exists when using the world coordinate system and picking it as a cylinder or a circle.  I have to guess there is something wrong with the data from the model somehow.  If anyone else out there has run into this please post a reply.

[jvenlet]

Did you verify that the nominals of the top, and sides of your base plate are 0 along the appropriate axis?  Where are you getting numbers from for the base plate?  Basic dimensions?  I see quite often that basic dimensions will be rounded to 2 or 3 decimal places.  Which can account for differences in nominals, in which case, CAD would be master.

[jrayself]

I run into that quite often. I've always chalked it up to sloppy designers. Odds are that the Y planeisn't actually zero, but 0.00052589612511", or something like that. IMINspect is rounding up to display 0.001"

I usually just turn on more decimal places in the annotations, and manually align the cad to the coordinate grid.

[cappleby1]

It would be nice to be able to control each dimension displayed precision separately.  Sometime on a print you get two place dimensions or three place dimensions.  I have only found setting in the software that control the precision of everything all at once.

[jvenlet]

It is more necessary on a print, because the number of decimal places is useful in conveying tolerancing information with title block tolerances.  A lot of times, if you go out 4 or 5 decimal places, you'll start to see rounding, all over the place.  Typically though, the rounding is orders of magnitude smaller than the tolerance and present very little in the way of conflict.   Then, occasionally you will find places where the print numbers were overridden manually, and then you start seeing issues measuring from CAD.  I usually just report everything to whatever my largest number of decimal places happens to be, but I also am on PC-DMIS 99% of the time, and I can change decimal places back and forth throughout the program, anytime I want.