Twist in curve / parabola

Hello world!
I tried using spiral but couldnt flip two circle as in image.
how to control the twist with points?
Advice me world.


twist_2023Aug19a.gh (11.3 KB)

P.S. Oops, that was four curves - this is two:



twist_2023Aug19aa.gh (18.3 KB)

1 Like

Playing with code like this can be instructive. The BiArc curves I used have an ā€˜Rā€™ (Ratio) input so I added a slider (0 to 1) for that. When set to zero, the ā€˜Angleā€™ slider becomes useless.

I changed the range of the ā€˜Angleā€™ slider from ā€˜0 To Piā€™ to ā€˜-Pi To +Piā€™, which means that to work correctly, the end tangent vectors (ā€˜Teā€™) from the white circle must be reversed. So I added an Eval component to do that. Then I started playing with MultiPipe (MS), which I donā€™t know much about.

I wanted to set ā€˜NodeSizeā€™ radii for both ends (the same) and the the midpoint at the top (larger), so added two sliders for those values. Playing around, I saw some weird flaws so added a ā€˜Countā€™ slider and Graph Mapper to see if more than three ā€˜NodeSizeā€™ values made any difference.


twist_2023Aug19b.gh (29.4 KB)

It looked like some of the issues were with BiArc so I replaced it with Arc SED, which has a start tangent vector (ā€˜Dā€™ input) but no end tangent vector, only an endpoint. It behaves differently, so doesnā€™t match the goal image you posted. But itā€™s nice in a different way.


twist_2023Aug19c.gh (28.4 KB)

Very interesting. :thinking:
Oh, in both models, flattening the ā€˜Curvesā€™ input to MultiPipe has a very peculiar effect :question:

3 Likes

Thanks Joseph, your last GH file above really got my attention and I knew I had to turn it into something unique that I could print. It didnā€™t take very long before I came up with this:

This is my slicerā€™s sliced view - the GH produces a Closed Brep so there are no errors in the exported STL file. Even though it slices OK itā€™s got some areas that are too thin, so Iā€™ve got a bit more work to do on it. Consequently I canā€™t predict yet what the end result will look like.

Five ribs instead of four? Without your GH code, I canā€™t tell which version you used (ā€˜bā€™ or ā€˜cā€™?) or if you flattened the ā€˜Curvesā€™ input to MultiPipe? If you did flatten, be aware that the result is two different breps (SubDs) that intersect each other. Iā€™d like to understand MultiPipe better :question:

Iā€™ll post the final result - hopefully later today. I used only one curve from c, Lofted it, scaled that, used edges for Ruled surfaces, joined all together, then made 5 item Polar array. I like using prime numbers for geometric arrangements.

So you didnā€™t use MultiPipe at all? Weird looking thingā€¦ With all the parameters I provided, why would you scale it?

The Loft surface looked nice - fairly unusual for sure. But it needed thickness. Offset didnā€™t work because of all the curvature; scaling worked better. The whole thing is quite small so I doubled itā€™s size in the slicer. Iā€™ll put that back into the GH file along with some other stuff I think will produce something useful.

Fun isnā€™t it? What happened to @super_mario ?


twist_2023Aug20a.gh (29.5 KB)

SUnion is quite slow and fails sometimes so it is disabled.
Wish I understood MultiPipe better.

P.S. This is why SUnion is slow - that busy center is a confluence of round ā€˜Capsā€™ from MP.

Hi, Joseph
Iā€™m incorporating youā€™re script with an rectangular base rather than original circle, so that extended curve can rest on rectangular base and can act as tensegrity stc.
Iā€™m new here but yes I love it here.
Thank you Joseph and all.
And wondering to use kangaroo to check it as tensile stc

Until you post your code, I wonā€™t know what you did but did it occur to you that you can have multiple inputs, like this?


twist_2023Aug20b.gh (13.9 KB)

Uh-huhā€¦ Until you post some code, you are likely to be ignored. After you get the support hoops the way you want, you might need a separate thread for Kangaroo and the mesh fabric surface.

Hereā€™s my version 0.8:

I donā€™t like this enough to print it, but it is Closed Breps and will print OK. The GH file is ā€œunnamedā€ because it doesnā€™t seem like itā€™s quite finished yet. When it is Iā€™ll put an LED puck light inside and it can be a table lamp or something like that.

unnamed.gh (24.2 KB)

TWIST.gh (15.3 KB)
image

I tried to collect point as in the image. Plan is to rest the curve end point at the desired points along the rectangle boundary.

missing

:-1:

1 Like

Sorry about that - Iā€™ve been using the trial Rhino 8. You should be able to replace it with the Rhino7 Midpoint component with no problem.

Yeah, I already did that. Rhino 8 isnā€™t the problem, itā€™s Wombat.

I see no resemblance at all to my code. :yawning_face:

The slider connected to ā€˜Ratioā€™ should be 0 to 1 with default of 0.5

Yeah - I donā€™t know how Wombat got installed. Iā€™ve also got Heteroptera - whatever that is.

After numerous iterations the only parts of your code left are the initial ones that establish the planes for the Loft control curves. I still canā€™t quite visualize how that works.

Your change from a base circle to a rectangle has several problems. I fixed some of them.

TWIST_SM_2023Aug20a.gh (13.4 KB)

You could use the base circle and make a square from the points, like this:


twist_2023Aug20c.gh (13.2 KB)

If you insist on a rectangle, youā€™ll have to figure it all out. Starting position is different from a circle, you probably want the Discontinuity points (corners) instead of Divide Curve.

Please keep this discussion here instead of using PMs.