Constraining a spline to a point fails

[Anonymous]

Hello,

Many times when I try to "coincident" constrain a spline to a point it fails. It's physically possible, because when I play with the handles the spline goes through the point. Is this a software limitation? Any workarounds, tips to get it constrained? See attached image. I want the spline to touch the point (circled in red). 

I tried adding a "Fit point" but then the new handle creates an ugly "Curvature comb". Or I could leave it just looking like it's constrained to the point, but for parametric design purposes I want to fully constrain it because I plan to derive it later.

Thanks,

— Yevi

[g-andresen]

Hi,

Can you upload the file?
On the screenshot I can't see which elements are already defined.
I also can't see how the elements are assigned to the sketches.

regrda

günther

[Anonymous]

I exported the sketch only as I cannot provide the whole work file. I hope it's enough. The angles of handles and the position of fit points are constrained. The only thing that is adjustable are the lengths of two handles. If you play with those you can see that a coincident constrain should be possible with the point in question.

[TrippyLighting]

I cannot get this to work either. Your best bet is to use an additional spline fit point with the result of having to fine fiddle the curvature.

[cekuhnen]

@jeff_strater Jeff Look at the file

if you remove the dimension of the middle line

coincident it to the spline

then add the dimension back to the line and enter the previous value

it moves the line point away from the spline while still being coincident.

[jeff_strater · Accepted solution]

created bug FUS-47180 for this.  The problem is that this is a 2-point spline, so the only way for the system to create this constraint would be to adjust the tangent handles at the spline ends, and I don't think the solver will do that.  I would not expect a quick fix for this, to be honest, so I'd recommend looking into the workarounds.  Adding another fit point, as @TrippyLighting suggests is probably the best approach, IMO.

[Anonymous]

 @jeff.strater thanks for looking into it. I moved to using the Control Point Spline. Works better constrainwise, but has some problems with jumping directions at ends when changing some dimensions.