Alignment for curve

[[To...]]

I should know this but my brain is a bit foggy.  I have a face groove detail on a cylindrical part.  View shown below is from the -Y direction.  The Z depth of this groove has been problematic for reasons I don't know.  The drawing is calling for profile of 2 radii within the curve.  If I take a point in the -Z direction at the bottom, where the 2 radii are tangent, I believe that will give the curve enough information to locate it for measuring. Will a geometry best fit alignment from the point do the job? 

 

Edited April 17

[[Ri...]]

Sure? You could also just create a new Alignment and fill in that point as the Z Origin - it should net the same exact result, but with less steps. 

[[To...]]

Please sign in to view this username.

 But doesn't setting that point as Z origin establish a new z origin, which would offset the curve the difference between the BA and the 2nd alignment?  Where the geometry best fit should report the difference between where the point's z nominal and actual.

[[Ri...]]

Please sign in to view this username.

Yes, you are correct. The GBF does not move the origin where as a new alignment would change the nominal origin. 

[[Ma...]]

Well nothing is changing - only nominal representation. So you only link deviation in Z against nominal point and measured point.
Unless you want to keep some numbers, then point in new alignment is the easiest and the most predictable.

Of course you can move origin in new alignment to keep your old nominal value and still keep that correction in needed axis.