Sweep2 - a final(?) redo

OK friends,

I'm trying to help a good friend of mine and GH novice with this "jewelry" definition (Eric does such things). But (as usual) the issue here is that @#$@#^%@ sweep2 thing (wonder when this could be finally fixed, he he).

focus to the saved view: 5 "ways" to skin a cat (all failed).

case1: the Python does nothing.

case2: Python turns red : what's the 6th element?

case3: the standard GH sweep2 component (does an open poly - no close sweep option available). In fact and since the sweep2 is the most critical component of them all (at least for AEC purposes) I wonder ... er....the obvious (i.e. "some" critical options are MIA, that is).

case4: forget this

case5: does nothing

BTW: yellow is made by baking the data and using the Rhino sweep2 (close = true). Turquoise is baked via the case3 GH component no close option available anyway (spot the "gap").

any ideas welcome

best, Peter

Attachments:

object1_for_Eric_by_Peter.gh, 110 KB
object1_for_Eric_by_Peter.3dm, 498 KB

Replies to This Discussion

Permalink Reply by peter fotiadis on June 9, 2014 at 2:41am

BTW: just discovered that case1 and case5 they work (this means that a polysurf is made but not closed) for certain (higher) numbers of segments. A bit odd that one.

But making the surface periodic (and closing the "gap") has the usual "sharp" edges removal result (not wanted that one)

Permalink Reply by Pieter Segeren on June 9, 2014 at 5:58am

Hi Peter, I had a quick (diagonal) look, I would sweep the 'segments' individually, then join these. For now I used only GH components in my approach btw. Hope I understood your intension correctly~

Attachments:

object1_for_Eric_by_Peter_PSG.gh, 63 KB

Permalink Reply by peter fotiadis on June 9, 2014 at 6:43am

Hi Pieter

But of course this is Plan B (in this simple case due to the symmetry is very easily doable ... not so in numerous other cases, small membrane modules without tension worth the name etc etc).

My intension is to rise some interest for that poor sweep2 thing (as a whole) and ... well there's always the weird fact that scripts work OK for a number of segments and up (but the "close" sweep in all cases still leaves a gap).

Made some corrections (see case2 - as regards the "s" type, identical to case1 anyway).

Attachments:

object1a_for_Eric_by_Peter.gh, 114 KB

Permalink Reply by Marios Tsiliakos on June 12, 2014 at 11:18am

Peter,

The gap is just visual. try to bake the geometry.

As for the Rhinocommon syntax in any scripting environment I am using the following.

Will try to take alook at the sweep2 method. but a rough guess is that needs to be cated to a brep.

Best,

Permalink Reply by peter fotiadis on June 12, 2014 at 2:28pm

Marios

1. The gap is real (slightly bigger than the Greek national debt, he he) - see Rhino test with 3/4/5 segments. Rhino sweep2 (close = ON) fails with the very same way, thus obviously it's a Rhino issue.

2. All scripts fail to output something when some "small" number of segments is used (3/4/5/6/7...) - use Stored View "segments". I guess that's due to some tolerance issue.

best, Peter

Attachments:

object1b_for_Eric_by_Peter.gh, 111 KB
object1b_BAKED_for_Eric_by_Peter.3dm, 438 KB

Permalink Reply by Marios Tsiliakos on June 13, 2014 at 3:07am

The RHcommon command gives the same result. So apparently that's the one used in the gh component.

It can be a tolerance issue as you mention. However I would never try to approach this geometry as a single sweep, rather than individual sweeps. It is problematic geomtery to begin with. And geometricaly you can understand the reason why rhino gets congfused with small segments.

I can't seem to be able to find the state..:(

RSS