Flip variable fillet

[summerson1990]

Hi! Each time I use variable fillet command I get no idea how does it detects the start and end points. I mean when I start selecting edges one by one I put in the values for the the start and the end and I see that the start point goes from the top and the end point lies in the very bottom of the edge and that’s ok. But when I’m continuing selecting another edges I see that the values I put in to the start now refers vice versa, that means that if for the first edge I had to put for example 1mm fillet as the start value and for the bottom 0,2mm now I have to do it vice versa because of F360 detects the start from the bottom of the def and the end from the top. This is not a big problem of course as long as I can visually determine that I need to flip start and end values but maybe there is some workaround that’ll help with that. Thanks! P.S. please, delete same my question from the support thread where I’ve posted it right now, didn’t find an option to remove it by myself.

[jeff_strater]

can you post a screencast of what you are seeing?  It's not completely obvious from the description, and a video would certainly help clarify it.  Thanks.

[summerson1990]

Here is an f3d file attached. Please look at the last fillet command in the timeline bar. As you can see here is variable fillets applied to 4 edges. But note that for the first 2 edges (green frame in the screenshot) and for the second 2 edges (red frame) start & endpoints are going vice versa. That means that I had to use user parameters vice versa for example for the first pair of edges I had to use "sa_vertical_bottom_fillet" parameter as a start point of the variable fillet but for the second pair of edges now I had to use "sa_vertical_top_fillet" as a start point in order 4 edges will look the same.  

I just tried to see if this also happens with simple shapes and made a cube to test it but it works as expected so start & end points match for all 4 vertical edges of te cube. 

[jeff_strater · Accepted solution]

the short answer is:  each edge has an inherent direction, and VR fillet uses that direction to determine start/end.  Here is a view in our internal kernel application.  The marks are hard to see, so I added arrows

the strange part is, VR fillet follows them backward, but at least it is consistent.  Regarding a simple cube, it depends how you look at it.  If you look from the top, you have the same situation:

Basically, in order to have a consistent set of edge directions, some faces will seem "wrong".  That's just the way it is.