Round coil won't over lap but others will?

[t_lamourine]

so playing with the coil feature, I would like to simply make a coil using the round shape. I also want to "compact" it a tad so basically the coil overlaps a bit each revolution. Seems if I use a square or either triangle it has no issues overlapping but the round/circle profile if it even touches it fails to compute.  why? I'd really like to be able to do a round coil body that basically overlaps a tad. I know I can do it with a like 20 step method with splines and all, but why can the other profiles overlap with no issue and the round can't? 

[etfrench]

Instead of modeling the positive, try sweeping the negative.  In other words, start with a cylinder and cut the space between the (coil)circles.  You may still get an intersection error, but you can leave a tiny gap and use a patch and boundary fill to fill the gap.

[TrippyLighting]

I don't believe the coil feature allows self intersecting geometry.

While the process shown in the image below contains more than one feature, it isn't 20 and it is straight forward and reliable.

[t_lamourine]

that's just it, it DOES allow self intersecting with anything but the round profile! 

[t_lamourine]

sure, again, a hack/work around.. I know a few different ones, but WHY?  both square and triangle profiles allow self intersecting so why doesn't the circle? 

[t_lamourine]

sorry, you're right, not 20 steps, just 10, to do what should be 1. 🙄. and you are goinginto a whole different modeling "system" which is in reality another 2 steps in and out.. plus it's a rather complicated method to be honest.  not picking on you guys, I get there are work arounds... but WHY!! 

[TrippyLighting]

I have no insights into the underpinnings of ASM (Autodesk Shape Manager) , the geometric modeling kernel that is used by Fusion 360 and Inventor. I believe that is where we’d need to look for a conclusive answer. 

[MichaelT_123]

Hi Mr. T.Lamourine 

You seem to be a curious man ... opposes... person.

WHY ... triangles and squares but not circles?

I will try to explain ... telepathically ... going back in time and traveling tens of thousands of km. 

The underlying algorithm calculates the respective shape in UV 2D space and maps the result into the 3D we enjoy ourselves in. 

Such calculations/mapping is pretty straightforward when the 3D shape is of developable topology (can be rolled out to the flat form). It will be the case for coil surfaces created by sides of triangles, squares, polygons, ... line segments. Manifolds created by 'moving' in space, not straight curves, are generally not developable. The respective construction of such a topology is much more expensive (regarding computations and individual data sizes). Things get even more difficult if non-developable surfaces interact/intersect, so different, even more sophisticated, computationally, and data-intensive algorithms must be implemented.

That WHY ... curvy objects (in general) are more expensive than squarish simpletones, aren't they?

 ... so they must be implemented/dealt with with additional caution.

Regards 

MichaelT

[t_lamourine]

I get it's more computer intensive, but fusion computes things on a regular basis that are WAY beyond this. If it can generate the shape without even hesitation with a copy paste feature, I find it hard to believe that fusion can't calculate it, as it does calculate it.  Now the "script" or what ever they call it for the coil function may not be able to, because it has limits I can understand, but then that function should easily be able to be corrected. (OK, NOT easily, but it should be doable) as it can calculate it, without even a spinning beach ball moment. so the geometry isn't too complicated for fusion. And calling people names because they are asking a simple question others can't answer is rather despicable. 

[jeff_strater · Accepted solution]

Thanks for reporting this, @t_lamourine.  This is a valid question.  I did some investigation, and the short answer is:  The geometry of a circular coil is just more complex than the other types.  Lines swept along a helix produce a simpler surface than a circle.  The modeling core can more easily compute the intersections between the swept surfaces from a line than from a circle.  That's all there is to it.  There is a bug for this:  ASM-11551, but in all honesty, I would not expect it to get addressed any time soon.  As far as I know (and I've been working with Fusion for over 10 years), you are the first person to report this issue, so it is not a high priority issue for the modeling team.

[t_lamourine]

Thanks for you response, kinda what I thought but was wondering if there was some other issue I wasn't seeing!