So my construction from your "control point measure" is false :)
It was the hasard :)
But my measure by numeric calculate is of course good :)
And I can build my curve anyway
Each line is perpendicular from the curve to the circle
Do you think now that your system of control point "past" to the real curve?
I have taken one more so I estimate that the 8 first are good, the 9 maybe not yet...