Can sweep-as-surface, but not as-surface?

matthew_stacy

I am struggling to sweep a circular profile along a helical path with tangential entry and exit tails.  As a swept-surface this is successful, and can be subsequently enclosed to make a solid. 

However the sweep fail when swept as a solid.  Any thoughts or ideas as to what I might be doing wrong?  Both sweep operations utilize the same sketch profile ("wire") and the same path (composite-curve named "path").

I suspect that the problem lies at the intersection of tail and helix, but I am troubled by the fact that the intended geometry can be created by enclosing a swept surface.

S1mon

Make sure there is some amount of clearance between the coils. If they are tangent, Parasolid won't like that for a solid.

matthew_stacy

@S1mon, that's a good point and I have already taken that into consideration.  Thanks. The helix pitch is already defined at 110% of wire diameter, so I should be alright on that front.

I am starting to suspect that the problem lies in establishing tangency between tail and helix.  When I apply the SELECT > EDGES > TANGENT CONNECTED tool it does not recognize these entities as being continous.  Still working on that.

S1mon

I've had weird failures in the past with sweeps which had "too many" turns in Solidworks (also based on the Parasolid kernel). There may be certain upper internal limits to the number of isoparameters (i.e. the degree of the surface) which could cause it to fail. It may be that if you break it up into more than one piece, and either add each one or boolean them together at the end, that it will behave. 

bryan_lagrange

@matthew_stacy we will be covering surfacing at the User Group meeting on 9/15. You could present your question there.

S1mon

I just made it work by doing 3 sweeps (along tail 1, Helix 1 and then tail 2). The first one uses the sketch as the face to sweep (along tail 1), and after that I just picked the previous end of the solid.



I've definitely seen this issue before in Solidworks, so I assume it's a Parasolid issue, but it does seem like a basic enough thing that should just work. I would report this to Onshape as a bug.

matthew_stacy

Mental error on my part.  I sketched the tail parallel to the x-axis ... totally neglecting the the helix angle.  The tail cannot be constrained "horizontal" (parallel to x-axis), but instead must be constrained TANGENT to the helix.

Problem solved.  But I still don't understand why a swept surface is more forgiving than a swept solid with regard to discontinuity of tangency.

Thanks for the help everyone.

S1mon

@matthew_stacy

Oops. I didn't look that carefully at the tails and assumed you had made them tangent. That is an issue in terms of modeling a real torsion spring, but I agree, if it can sweep the surface, why does it barf on the solid?

Evan_Reese

This doesn't answer your main question, which I'm also curious about, but you could make the tails with a single thicken feature after sweeping and get the tangency for free like this.

matthew_stacy

Thanks @Evan_Reese, that's a useful modeling approach.  I appreciate your suggestion.  But, since figuring out the tangency continuity error in my helix-with-tails path I still favor the continuous sweep.