Sweep Shapes Along Compound Curve


Hello. I am looking for some guidance as to how to sweep the two shapes shown in the image along the compound curve shown. These shapes form an arch wrapped on a cylinder. I am very new to Rhino and I cannot get it to sweep correctly. Any help would be most appreciated!

Thanks!

Hi Tom - please post a file with those curves.

-Pascal

Compound Curve.3dm (1.1 MB)
Here is the drawing. Thanks!

Tom

Hi Tom - is this it:

Compound Curve_Maybe.3dm (1.3 MB)

Sweep1, from relocated profiles (rotated in Front) with 'Roadlike on the axis shown.
Actually, @Tom_Haws - I take that back - Freeform Sweep1 may be better here, not Roadlike.
File is updated

-Pascal

Yeah, that is what I got too. From the top it needs to look like the attached image. The back face of the profiles need to stay vertical - no rotating back and the back face needs to cut back perpendicular to the front face at the ends.

Top View:

Front View:
Capture2

Yeah, you may end up needing to build this in parts if all the cylindrical faces need to stay parallel to the cylinder.
Hm - is this any better:
Compound Curve_Maybe.3dm (1.6 MB)

?

-Pascal

Very Close! The only thing is the ends need to cut back perpendicular. Your model looks very clean. Did you just sweep it in one plane and then ā€œbendā€ the whole model? Iā€™m so new, Iā€™m not sure if thatā€™s possible :slight_smile:

Thanks,
Tom

Yep, there is a bend and it looks great. Iā€™m going to head in that direction now. Thanks, Pascal for your help!

Tom

Hi Tom - Bend may work as well but I used trickery -

ProjectObjects.rhp (20.5 KB)

See the attached plug-in. Download and unblock the rhp file in Windows Explorer, then drag and drop it onto Rhino to load it and add the command ProjectObjects
The base objects are Revolved from the profiles; these are then projected onto that section of a cylinder from the Front view with density set to 16 both in x and y, though y could probaly be less there is no harm done.

-Pascal

1 Like

Wow, that plug-in is great! Is it possible to add a Preview option so that the settings could be changed if necessary before building the end result? Also, history enabled maybe? :slight_smile:
I noticed that the plug-in produces some weird looking results when the user-set direction is at angled other than parallel to the Worls axes.

Hi Bobi - a preview should be possible - initially it would simply be the result added to the document and deleted if you donā€™t like it and change the settingsā€¦ that is the dumb but quick way, Iā€™ll have to see if I can make a proper preview.
History is not possible Iā€™m afraid, outside of an actual command, for which this does not really count.
Iā€™ll look at the skewed axis thing - probably doing something dumb.

Note the major fly in the ointment here is that the ā€˜footprintā€™ of bounding box of the projected objects must fit on the target in the projection direction - that is a major weakness that I may or may not be able to fix.

-Pascal

1 Like

Compound Curve.3dm (1.4 MB)
Hi Tom , I just wanted to see if I did what your intent was. thanks,Mark

Both your top and lower profiles are tapering out away from the cylinder. If you look at the bottom view, you can see that clearly. The intent is for them to ā€œhugā€ the cylinder. The back side of each profile should be flush.

Thanks,
Tom

I keep getting this error:

Yep - the target cylinder surface needs to be big enough that the biounding box of the objects hits it completely - in the example I posted the cylinder is big enough.

Does that make sense?

-Pascal

I can easily see many applications of that tool in the architecture and product design, and jewelry design as well. I hope to see it as a native tool of Rhino some day. :slight_smile:

Itā€™s hard, I guess - this script/plug-in has been around for a good few years and is useful, but it is nonetheless a hack that is hard to do in a better way - the bounding box footprint restrction is pretty severe for a real command. I have some thoughts about how to overcome it but have not typed it yet.

-Pascal

1 Like

Hello Pascal,

Well whenever you get a chance to type these out I (and I suspect others) would be quite interested to learn of them.

Yes what you are stating does make sense and the obvious workaround that I believe follows from this and might work (and I have ā€œdipped my toeā€ into) is to scale the object (s) so as not to get the error message. And then in theory scale the projected object result back to the original dimensions.

This is a really good plug-in irrespective of itā€™s current limitations.

Thank you,

Andy

1 Like

Hi Andy - a better way(s) may be to ExtendSrf or UntrimBorder if it is a single surface, or use the Fin command set to make a tangent surface and join the results to the target, and so on, - the idea being that the geometry of the actual target not be altered. All are regrettable workarounds, there is no getting away from that.

-Pascal

Hello Pascal,

Got it.

Thank you,

Andy