complex sweep

Replies to This Discussion

Permalink Reply by Dave Anderson on October 4, 2012 at 6:28am

Without a 3DM or at least screens I'm afraid I don't understand the scenario well enough to offer my help. Care to upload?

• ▶ Reply

Permalink Reply by Charles Fee on October 5, 2012 at 8:05am

Yes. Here's how.

Draw the centerline of your handrail. Make it a continuous straight segment polyline. Turn it at will, take it in any direction - even 90d turns, not only descending (ascending) but wherever you need that handral segment to continue to. (bad english, but so what ?). Use PIPE. Set the diameter to EITHER a proper handrail diameter or an exagerated diameter to make it barely visible from where you are viewing it from (again, but so what?) . Fillet radiused turns BEFORE using PIPE. Experiment with diameters if you get fails at turns. Ok, this gives you round handrails. What if you need a complex cross section instead ?

Interesting note - you might not be able to see the complex cross section at the zoom distance you're using for imaging, so a round pipe may meet the need.anyway. If not -

2 common scenarios - angle and sculpted cross sections. Create polyline as above. For non 90d turns, radius ALL turns that require it with fillets of radius LARGE ENOUGH to allow 1-Rail sweep to succeed. BE SURE your single cross section is NORMAL to the START of the centerline you're sweeping with. Keep ALL 90d bends SIMPLE 90d turns of the polyline centerline (don't go around the corner with lots of points close together - a common mistake). You may still have sweep fails. Wherever you have a sweep fail (RH still needs work in this area) break the polyline, and start a new cross section from there or , you may need to use brute force to fix those transitions "manually". You may discover that at the zoom distance of your imaging, the break is too small to see. If that's the case,

do a simple cleanup, and DON'T get too caught up in detail you can't see. You have far more important work to do instead. RHINO is a POWERFUL partner in your modeling efforts, but its just that - a partner, or better put - a "tool".

This will require some experimenting, but the basic principles outlined here will get you going in the right direction.

 

Good luck and post back here with how it worked out ?

 

Charley.

• ▶ Reply

Permalink Reply by Pascal Golay on October 5, 2012 at 8:58am

Hi Bill- generally, try to break up the shape into simple, if possible, sections. Make the straight parts as straight extrusions, the curves as sweeps or railrevolves on just the bendy parts, etc..  Simple example attached. Deforming arcs in the vertical direction only often requires a ChangeDegree to 3 in order to maintain tangency at both ends.

-Pascal

pascal-at-mcneel-dot-com

Attachments:

• SimpleRailing.3dm, 532 KB

• ▶ Reply

Permalink Reply by Charles Fee on October 5, 2012 at 3:12pm

Pascal - are you referring to the handrail centerline ? Interesting. Sounds like Bill was hoping to get his sweep to happen in one smooth motion, start to end, tho. Interesting approach. I'm going to give it a try, especially setting the degree to 3. What's the default degree ? This suggestion might be a great way to handle ends that need to end truly normal to the sweep curve.

Bill - let us know how it goes. This is one I wrestled with myself awhile backso I'm genuinely interested !

Thanks -

Charley.

• ▶ Reply

Permalink Reply by Bill Amaya on October 10, 2012 at 8:13am

Ok here are some pictures of the HR turn.  I am trying to draw these turns exactly as they will be when built, partly because we are using the drawings to cut parts on our CNC.  Any errors or ambiguity will result in inaccurate or un-usable parts.  I have drawn the centerline of the rail, and the cross sections.  To get the sweep correct I have to place multiple cross section at the correct attitude.  That is the part that takes time.  Here check it out. 

Thanks Bill

Attachments:

• plan view sections.jpg, 744 KB
• perspective sections.jpg, 549 KB
• view 1.jpg, 525 KB

• ▶ Reply

Permalink Reply by Bill Amaya on October 10, 2012 at 8:15am

Here are a couple more pictures of one of these turns.  You can see that a simple one rail sweep does not work.

The last picture is the best

Best regards,

Bill

Attachments:

• view 1.jpg, 525 KB
• view 2 1rail.jpg, 803 KB
• perspective HR.jpg, 717 KB

• ▶ Reply

Permalink Reply by Bill Amaya on October 10, 2012 at 8:19am

Ok here is a picture of another turn on this project.  You can see all of the cross sections and also see some of the steps needed to get here.  I am looking for a quicker and easier way to do this.

Best regards,

Bill

Attachments:

• sweep 2 perspective.jpg, 869 KB

• ▶ Reply

Permalink Reply by Dave Anderson on October 10, 2012 at 9:19am

So the problem is lack of rail/cross-section perpendicularity? I struggle with that a lot, too. Even with two rails, this is usually not enough information for Rhino to know how to orient cross sections. If you're okay with a high point count surface you can add rails and cross sections to use for NetworkSrf.

What I'd do: it looks as though the shape of the handrail remains the same throughout, so what you can do is "ExtrudeCrvAlongCrv" instead of sweep. This will give you minimum and never changing cross sections. As they will be fewer (depending on your input curves) the orienting won't take that long anymore. Judging from "sweep2perspective.jpg" though you have too many points.

• ▶ Reply

Permalink Reply by Bill Amaya on October 10, 2012 at 9:27am

Dave,

Thanks I will give that a try.  The other challenge is keeping the cross section "level".  For handrails, and especially helical handrails,  cross sections need to be oriented to point to the center of the circle.  And  each one of those cross sections need to be level.  So unlike a road, handrails do not "bank".  Never the less let me try your suggestion and thanks again.

Bill

• ▶ Reply

Permalink Reply by Bill Amaya on October 10, 2012 at 8:27am

Pascal,

That is exactly what I am up to. And that is exactly the approach I was using.   Seems like Rhino can do this within certain parameters.  If the curve is to tight or the change in height is radical, rhino gets unhappy.  Unhappy being non-fair surfaces, at times with gaps in the surface.  Again thanks for the response. 

All the best,

Bill

• ▶ Reply

Permalink Reply by Pascal Golay on October 10, 2012 at 9:48am

Hi Bill- please send me an example, with all input curves and the ugly result...

thanks, 

-Pascal

pascal@mcneel.com

• ▶ Reply

Permalink Reply by Chris Tietjen on October 14, 2012 at 7:43pm

Upload your rail and section curve if you need more help.  It's hard to get the nuances from a jpeg.

Chris

• ▶ Reply

• ‹ Previous
• 1
• 2
• Next ›
• Page