Hi everyone!
I’m working on a small university project where I used the Sweep2 command to create the angled surface. The problem is, as shown in the screenshot, that the surface has a lot of curves really close to each other at some point. This results in a strange fillet behaviour later on when I want to fillet the circle.
Am I doing something fundamentally wrong here and/or has anyone some tips how to make this surface cleaner?
The file is attached.
Thanks
David
kerzenhalter.3dm (3.3 MB)
give a try with _loft
do a _loft for each segment.
arc of different angle might not behave as expected because the are rational curves (cvs have weight)
the inner circe here is messy (_pointsOn)
kerzenhalter_02_tp.3dm (4.2 MB)
my guess these are the surfaces you re after.
all black surfaces are loft.
the pink one also is a loft, but the arcs have different angle, so the loft / surface has unequal parametrisation in u-direction… this will fool the game.
therefor the blue one is tricky.
ok another approach is to
_rebuild degree 5 with 6 cvs
deviation is less then 0.001 and therefore within document tolerance.
with the correct split positoin a standard loft will do its job:
Thanks for the detailed analysis Tom! The circle was messy because it was created with _DubEdge. Tried _Rebuild to simplify it but that wasn’t actually working accurately.
Loft does the job for this model – but the transition isn’t perfect. Just out of curiosity, would there be a way to make this more accurate or is this not possible on the given shape?
One problem I’m still faceing is the filleting. Is there a way to flip the surface edge arrows? I think thats whats causing the problem. The other half of the circle faces in the other direction, towards the surface.
kerzenhalter-V2.3dm (972.4 KB)
and the surfaces are not tangent - the issue i was talking about in above post.
at which university are you studying ?
kind regards -tom
Now I see, thanks Tom!
Industrial Design at FH Joanneum in Graz, Austria.
Here’s another approach you may want to consider. Use degree 5 circles. Then build simple single span surfaces from them, and rebuild multi-span surfaces where possible, because Rhino likes to build overly complex surfaces. You also need to only build the primary domain and then mirror. Then you have a surface model with the least amount of spans, thus CPs, and subsequent tweaks become more straightforward; see a basic beginner guide here.
That looks promising Lagom!
Never heard of single span curves and surfaces before – just had a quick look at the guide and definitely have to dig deeper into that.
Maybe it’s an obsolet question because I’m not familiar with concept but after everything is in degree 5, split up etc, did you then built the surface with _Loft? Because when I try that on your edited model, I still get isocurves in the surface.
Thanks
David
I used Loft and EdgeSrf, then made sure the surfaces are single span, see “rebuild multi-span surfaces where possible” in my reply above. For industrial design surface modelling, it’s a very good habit to get into single-span curves/surfaces thinking as soon as possible, as outlined in the guide, to minimise or even avoid downstream problems later.
Unfortunately, Rhino doesn’t name surfaces with their creation type automatically, so other users can’t know by just selecting them to see what surface building tool was used.
A degree 5 circle has 12 spans. A degree 5 quarter circle has 3. See below that the Loft should have 3 spans also, but Rhino refuses to build it like so, thus Rebuild. However, with 1 span, for this type of object, it may suffice, hence I rebuilt it with 1 span only.
