Best Fit using the new GD&T engine

[[Ja...]]

I’m not understanding how the new GD&T engine functions when doing a best fit. It’s putting the center of the bolt circle out in left field somewhere so when you look at the nominals they’re not even in the ballpark. I can go into the special functions and offset each feature back to where it’s supposed to be according to the print, but then the actuals really aren’t going to be respective of the best fit. What am I doing wrong?

[[Ha...]]

Are the actual results showing large positional deviations, or is it just the nominals not matching the print specifications?

Edited Monday at 08:25 PM

[[Ja...]]

In the picture below, you'll see Calypso thinks the center of the bolt pattern is located a couple inches above and slightly behind the highest bolt hole in the pattern. Consequently on the report (bottom picture), the nominals are way out in left field. The amount of deviation looks to be correct but when the program actually runs, we won't be able to tell which hole is off, what direction, and how much. The machinist won't have the required information to make the adjustments.

[[Ri...]]

It goes to the center of the DRF. If you only have a Plane as the Primary, then it will go to the center of the plane. Very annoying indeed. 

[[Ja...]]

Any way to fix it?

[[Ri...]]

None that I'm aware of, short of putting a bogus secondary datum in there. 

[[Ja...]]

Ya, that's what I thought. I'm going to use the constructed bolt circle as the center. Although it won't be a true best fit, it should get me reasonably close.

[[Ri...]]

Please sign in to view this quote.

If you switch over to Loose, you can use Alignments and then apply Best-Fit constraints. 

[[Iv...]]

Could you not add a special alignment offset/rotation? See my screenshot below. This offset allowed me to reposition the nominals to be 0,0. 

[[Ja...]]

Richard - I'll give that a try. Thanks.

Ivan - That would still be a pre-defined starting point. It would be the same as the original scenario except now it would be in a different spot. It still wouldn't be a true best fit. A best fit allows the bolt circle centerpoint to move around so that all of the error from all of the individual holes is minimized. That will give a better picture than just an average, or a location to a pre-defined point. Make sense? Did I explain that sufficiently? Hope that helped.