VSS X-Vector Rules?

[jeff4136]

Except for trajectories, the two VSS's are identically defined.
Does anyone know how to predict when the section will be rolled over?
The behavior can be demonstrated by changing the angle value in datum
sketch POINT_DEF a degree or so one way or the other.


2007-01-05_001912_vss-roll_wf2_.zip


View attachment 2981

 

[vlad1979]

That is strange, but I think it's abouthow the trajectory is made (the curve). I tried to make an intersection from two curves and then make the VSS using this trajectory. Like this the section will not be rolled over for any angle.


View attachment 2982

 

[jeff4136]

Thanks, Vlad. I think for that particular curve it won't ever roll
over because the results of the intersection will be planar, hence no
twisting of curve normal vectors. I feel certain the behavior is
associated with normal vector twisting (Help says it'll try to
minimize the twisting) but I can't fathom the rhyme or reason to the
threshold. I also think it's interesting that it seemingly blends to
another quadrant rather than simply tracking normal vector orientation.
I guess that's done to make the results more predictable?


(I learned a little something interesting playing with the idea. If
you mess around in the VSS Tangency options fly out tab (pass the
cursor over Side 1, Side 2) you can actually highlight the,
transparent to user, surfaces used to create the intersection curve.)

 

[mgnt8]

hmm... pretty interesting. There's no z axis control for normal to trajectory plane control so it looks like where it falls depends on the location of the end of your trajectory in relation to the x & y of the origin, which could be all over the map. Your results look like a bug. Did you submit it to PTC tech support? This is one more reason why VSS & swept blend need to be combined.