cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
Showing results for 
Search instead for 
Did you mean: 

Sketcher constrain for arc length

Sketcher constrain for arc length

At sketcher, arcs which are (tangentially) connected to surrounding lines and arcs there is two (mathematical) solutions for resulting geometry. Currently users have to push in some effort to guarantee the desired result for highly prarametrized models.

 

Product Idea

Provide additional sketcher constrain, which will guarantee the (shorter/longer) arc during any dimensional change.

 

Flipping ArcFlipping Arc

5 Comments
mneumueller
17-Peridot
Status changed to: Clarification Needed

thanks for your input. Interesting idea. Do you have some sketch examples that would exhibit this behavior? If so, please send to mneumueller@ptc.com

Thank you very much

S_Edgenear
14-Alexandrite

@mneumueller In the screenshot one could for instance dimension both lines with a 90º angle and specify the radius value of the arc. If we needed to parametrically change the angle between both lines, to an angle greater than 180º, for instance, 270º, the sketch regeneration would fail, because the arc would not change the tangent direction. In this case, we ould have to delete the arc, drag the lines for the other side, and then re-apply another arc. We could not even replace it to maintain the same entity ID. So, there is room indeed for improvement to deal with this situations. Here, I think the problem is that the fillets (or chamfers) in sketch are not parametric operations, meaning, the arc is applied, specified which side of the lines its origin belongs to (quadrant), and the lines are trimmed, but then, the arc and lines stay as regular sketch entities, and not entities to which an operation (fillet) was applied. It's like a "dumb" solid in 3D operations. I think to guarantee the desired behaviour, the sketcher would have to add fillets and chamfers as internal feature operations and kept in an internal model tree (like the offsets operations). If this operations would be kept as an internal 2D "feature tree", operations like fillets, chamfers, offsets, mirrors would be parametric, and would allow us (finally) to also have parametric 2D patterns (which would also allow us to have parametric polygons instead of having to copy pre-defined polygons form the pallette, which we cannot parametrically change the number of sides as a parameter selected by the user, which as workaround forces us to do certain operations as patterns in 3D that could more easily be solved in 2D, with faster regeneration times).

 

Even if there could be added "intelligence" to the sketcher without adding an internal 2D operations feature tree to solve this particular problem as shown in the original poster screenshot, one thing that could be added is the possibility to simply internally hide the arc if the angle between both lines becomes too small (close to 0º or 180º), meaning, the radius would tend to infinity. In this case, both lines should be trimmed to each other at a vertex, and the arc internally supressed, to be restored if the angle between both lines is changed to allow the arc to be re inserted with reasonable radius value. One could even go further and try to allow both lines to have a 0º angle if needed in certain regeneration "configurations", but that would be more difficult to solve, even taking into account for instance a previous vertex location for the intersection of both lines, when they ar at an angle to each other.

S_Edgenear
14-Alexandrite

@MikkoHinkkanen 

 

In this case it's easy to add that new constraint, smaller arc length. It's easy to calculate it's value, and it would guaranteee that we would not get "sharp" vertex at the tangency points, and only "smooth" vertex.

MikkoHinkkanen
13-Aquamarine

Sample model shipped to mneumueller@ptc.com via email.

mneumueller
17-Peridot

Got it - thanks!