I made some experiments with catia V5 non-positionned (a.k.a simple implicit) sketches The supporting planes of the sketches were the 3 planes of reference of a freshly created part named xy, yz and zx. As a result the local unit vectors of the sketch denoted H and V were implicitely defined as
All the basis are right-handed ... but I would have naturally expected,
(H,V) = (+z,+x) : ((H,V,y) right-handed
as the third solution by considering a circular permutation of the x,y and z axes in the two previous ones. Is it a CATIA bug ? Does anyone know about the way local sketch axes are computed in implicit mode ?
Thanks
I too have done a lot of experimenting with positioned and non-positioned sketches(implicit). Long story short, if you want something stable and reliable, always use positioned sketches that are based on good references that you understand and can control.
It seemed to me that the local sketch axes were somehow based on the global coordinate system closest to X+, Y+, and/or Z+. In my experience, I had several issues with document templates and when instantiating one, the sketch axes would flip on their own even if the target geometry was the same as the template. So, to get something stable, always use positioned sketches and make sure the directions of the references of your positioned sketch are stable and consistent as well.
In the beginning, it seemed like a bug to me too, but it behaves as it should when you use a fully positioned sketch.