Round Transitions

[jeff4136]

This has been bugging me for some time and
after trying to make heads or tails of it myself
have given up.


There are three valid transitions for the
intersection (round only 1 & 2, patch). I assume
I should(?) be able to specify and create any of
them, at will, at that particular intersection
but can't. By way of an example; I want to
recreate the yellow round only transition on the
remaining visible shape.


Is there a solution; either technique or version /
build? Or is manually building the transition as
good as it gets?


-TIA


2007-03-20_012431_rnd_transitions-wf2.prt.zip


View attachment 3285

 

[molsson]

I tried this, and Pro/E will automatically create the 2 round transitions you are after. However, since the radii are the same there will not be a patch available. To create the round below, create the round in two sets. Remember to shift-select to get the whole chain for each side. You should end up with an automatic transition like this.

View attachment 3286


Next we hit right mouse button, and select 'show transitions'


View attachment 3287


Now we will see the available transitions to choose from.


View attachment 3288

And now we have the other round case

View attachment 3289



You can also delete the transition, and control select the edges of the remaining rounds, then right mouse button and 'make transition' to create something else.

Have fun.

 

[jeff4136]

Molsson, thank you for taking a look.
If I understand you correctly, you could solve
either Round Only 1 or Round Only 2, at will,
using the same Sets?


What version and build are you using?


- - - - - - -


FWIW, there is no selection method; 1 set, 2 sets,
tangent chains or concave chain and convex chain
that will work for me (I ~think~ I've tried them
all, early build WF2). As you can see I get Round
Only 1 & 2, but the highlighted Round or Patch
are the only two transitions that will solve.


The only solution for me with that surface set
was to change from .0012 rel acc to .001 abs acc
AND Flip Normals (suppressed at the bottom of the
model tree in pic). It solves the same for 1 set,
2 sets, tangent chains or concave chain and
convex chain.


Still not sure if there's a dark cloud hovering
over me or a bug in this build but am leaning
toward the latter.


Thanks, again, for your input.


- - - - - - - -


View attachment 3290



- - - - - - - -


As an afterthought; I'm curous why you aren't
showing a patch transition in your pics. What
accuracy and model size are you dealing with?



- - - - - - - -
Edited by: jeff4136

 

[vlad1979]

I think it's a build thing I have WF2.0 m240 and it looks just like yours. If I select round 1 or round 2 I get the same round transition!?!


View attachment 3292


But if you don't want to have the same radius you could make individual radius for the four edges, and the transition will change direction. Just make it with 4 sets and select all 4 edges and the one that is indicated by the arrow give it a smaller value.


View attachment 3293

 

[jeff4136]

Thanks, Vlad.


> ... give it a smaller value


That's interesting, too. This isn't for a specific
application. Just trying to make sense of it all
(maybe rack up a few points toward an upgrade in the
bargain). I went for the longest time thinking there
was something about the process I was missing ...
Still would like Molsson, or someone, to verify my
understanding of Molsson's observations and WF3(?).


It (what I see) is truly squirrely. If I start with
two full cylinders and split down the middle each side
behaves differently. Same if I try to round the full
cylinders. Strange stuff ...