# Precise double curvature from curves

**URL:** <https://discourse.mcneel.com/t/precise-double-curvature-from-curves/81849>\
**Category:** Rhino for Mac\
**Tags:** surface\
**Created:** [April 14, 2019, 12:00pm UTC](https://discourse.mcneel.com/t/precise-double-curvature-from-curves/81849 "2019-04-14T12:00:29Z")\
**Posts on this page:** 17\
**Page:** 1

<div class="post-metadata">

**Author:** ![Brunosouth](https://sea3.discourse-cdn.com/mcneel/user_avatar/discourse.mcneel.com/brunosouth/32/514705_2.png) [@Brunosouth](https://discourse.mcneel.com/u/Brunosouth)\
**Post date:** [April 14, 2019, 12:00pm UTC](https://discourse.mcneel.com/t/precise-double-curvature-from-curves/81849/1 "2019-04-14T12:00:29Z")

</div>

Hi , I am trying to build a surface from a closed curve (double curvature) and two support curves (simple arcs).

 ![Rhinofourm_screenshot](https://global.discourse-cdn.com/mcneel/uploads/default/original/3X/8/8/884c0880e40812e5fdc2c9a5f7a3e92d76e2f6e5.png)

I am using Patch command = good results, but when I use the intersect command between the curves used for the patch…they are not quite touching the patched outcome. I guess this has to do with the patch parameters I can set within the patch window dialog. I am trying to find out if there is any tutorial or vide or maybe you have insights to teach how to get these type of surfaces. I am a designer and I am constantly using these approach (specially in furniture seats). I am trying to finally achieve a precise surface that I can deliver to the producer. My goal is to have a surface that could be intersected by the generative curves and they will touch in each of the points.

Find attached the file, and thanks always for your invaluable insights.  
[RHINOFORUM.3dm](https://global.discourse-cdn.com/mcneel/uploads/default/original/3X/0/2/024981d1c29f7a7ad7b32cfb2edcdf5f26fd1b22.3dm) (3.0 MB)

---

<div class="post-metadata">

**Author:** ![markintheozarks](https://sea3.discourse-cdn.com/mcneel/user_avatar/discourse.mcneel.com/markintheozarks/32/678563_2.png) [@markintheozarks](https://discourse.mcneel.com/u/markintheozarks)\
**Post date:** [April 14, 2019, 12:09pm UTC](https://discourse.mcneel.com/t/precise-double-curvature-from-curves/81849/2 "2019-04-14T12:09:57Z")

</div>

hi Bruno, I tried a Network surface after splitting the blue curve at the long end. it worked okay . maybe this isn’t the outcome you are looking for.—Mark

---

<div class="post-metadata">

**Author:** ![Brunosouth](https://sea3.discourse-cdn.com/mcneel/user_avatar/discourse.mcneel.com/brunosouth/32/514705_2.png) [@Brunosouth](https://discourse.mcneel.com/u/Brunosouth)\
**Post date:** [April 14, 2019, 12:29pm UTC](https://discourse.mcneel.com/t/precise-double-curvature-from-curves/81849/3 "2019-04-14T12:29:03Z")

</div>

Like this?

 ![22](https://global.discourse-cdn.com/mcneel/uploads/default/original/3X/d/f/df1952c8908ad93d9221f6f5da16a6eb2092d77c.png) ![31](https://global.discourse-cdn.com/mcneel/uploads/default/original/3X/a/e/ae7876e4a95ed65fce5f7d83893d814fd0d89aca.jpeg)

It worked better than with patch, but the curves are not intersecting 100%

---

<div class="post-metadata">

**Author:** ![jeremy5](https://sea3.discourse-cdn.com/mcneel/user_avatar/discourse.mcneel.com/jeremy5/32/493741_2.png) [@jeremy5](https://discourse.mcneel.com/u/jeremy5)\
**Post date:** [April 14, 2019, 2:21pm UTC](https://discourse.mcneel.com/t/precise-double-curvature-from-curves/81849/4 "2019-04-14T14:21:31Z")

</div>

If you split the blue curve by the short axis instead of the long one you will get closer - intersection curves are:

 ![image](https://global.discourse-cdn.com/mcneel/original/3X/8/3/83818b0c72595c0591fdf074ff68bece1f88cffd.png)

But…  
Your patch surface deviates by 0.02mm or less. That’s surely more than good enough for a chair back. So going further is just a lost cost.

Regards  
Jeremy

---

<div class="post-metadata">

**Author:** ![John\_Brock](https://sea3.discourse-cdn.com/mcneel/user_avatar/discourse.mcneel.com/john_brock/32/476283_2.png) [@John\_Brock](https://discourse.mcneel.com/u/John_Brock)\
**Post date:** [April 14, 2019, 2:24pm UTC](https://discourse.mcneel.com/t/precise-double-curvature-from-curves/81849/5 "2019-04-14T14:24:28Z")

</div>

I’d be inclined to Sweep the center profile to both ends of the long profile, then Trim the surface with the outside curve.

---

<div class="post-metadata">

**Author:** ![pascal](https://sea3.discourse-cdn.com/mcneel/user_avatar/discourse.mcneel.com/pascal/32/479395_2.png) [@pascal](https://discourse.mcneel.com/u/pascal)\
**Post date:** [April 14, 2019, 3:40pm UTC](https://discourse.mcneel.com/t/precise-double-curvature-from-curves/81849/6 "2019-04-14T15:40:55Z")

</div>

Hello - What John said, only even simpler - `Revolve` the red curve on the center of the green arc, trim with the blue curve:

 ![image](https://global.discourse-cdn.com/mcneel/original/3X/4/f/4f922f0dd15afb42493f4c7a84294ab661d05f0f.png)

-Pascal

---

<div class="post-metadata">

**Author:** ![jeremy5](https://sea3.discourse-cdn.com/mcneel/user_avatar/discourse.mcneel.com/jeremy5/32/493741_2.png) [@jeremy5](https://discourse.mcneel.com/u/jeremy5)\
**Post date:** [April 14, 2019, 4:46pm UTC](https://discourse.mcneel.com/t/precise-double-curvature-from-curves/81849/7 "2019-04-14T16:46:51Z")

</div>

Hi @pascal,

But the OP wants a surface that intersects his original curves. The blue curve does not lie on your revolved surface…

---

<div class="post-metadata">

**Author:** ![jeremy5](https://sea3.discourse-cdn.com/mcneel/user_avatar/discourse.mcneel.com/jeremy5/32/493741_2.png) [@jeremy5](https://discourse.mcneel.com/u/jeremy5)\
**Post date:** [April 14, 2019, 4:51pm UTC](https://discourse.mcneel.com/t/precise-double-curvature-from-curves/81849/8 "2019-04-14T16:51:10Z")

</div>

Hi @John_Brock,

But the OP wants a surface that intersects his original curves. The blue curve does not lie on your swept surface…

---

<div class="post-metadata">

**Author:** ![pascal](https://sea3.discourse-cdn.com/mcneel/user_avatar/discourse.mcneel.com/pascal/32/479395_2.png) [@pascal](https://discourse.mcneel.com/u/pascal)\
**Post date:** [April 14, 2019, 4:53pm UTC](https://discourse.mcneel.com/t/precise-double-curvature-from-curves/81849/9 "2019-04-14T16:53:29Z")

</div>

Hm - yeah, if the blue curve counts then… it’s wrong =). I am being facetious of course, but only partly - It depends on the context - (like, what does that curve come from, why does it win over the exact surface implied by the red and green?) but any noodling of the surface to make sure it hits that blue curve will compromise the surface I would say.

-Pascal

---

<div class="post-metadata">

**Author:** ![jeremy5](https://sea3.discourse-cdn.com/mcneel/user_avatar/discourse.mcneel.com/jeremy5/32/493741_2.png) [@jeremy5](https://discourse.mcneel.com/u/jeremy5)\
**Post date:** [April 14, 2019, 5:03pm UTC](https://discourse.mcneel.com/t/precise-double-curvature-from-curves/81849/10 "2019-04-14T17:03:10Z")

</div>

I guess that’s an aesthetic and/or ergonomic argument. I leave those to people better qualified… 😀

---

<div class="post-metadata">

**Author:** ![John\_Brock](https://sea3.discourse-cdn.com/mcneel/user_avatar/discourse.mcneel.com/john_brock/32/476283_2.png) [@John\_Brock](https://discourse.mcneel.com/u/John_Brock)\
**Post date:** [April 14, 2019, 5:19pm UTC](https://discourse.mcneel.com/t/precise-double-curvature-from-curves/81849/11 "2019-04-14T17:19:08Z")

</div>

Because of the oval shape, I think a trimmed surface without singularities will be cleaner and more fair than one derived from it.  
Obviously your call.

---

<div class="post-metadata">

**Author:** ![Brunosouth](https://sea3.discourse-cdn.com/mcneel/user_avatar/discourse.mcneel.com/brunosouth/32/514705_2.png) [@Brunosouth](https://discourse.mcneel.com/u/Brunosouth)\
**Post date:** [April 15, 2019, 5:24am UTC](https://discourse.mcneel.com/t/precise-double-curvature-from-curves/81849/12 "2019-04-15T05:24:38Z")

</div>

Thanks a lot, I think you are right. I can live with 0.02 mm margins. Just that I did not understand why the surface can not be 100% touching the surfaces. I started asking about the patch command because I encounter a problem: The closed curve its actually en extraction of a surface (the side of the seat). After the patch surface outcome I could not join it with the surface from the sides of the seat. So I was wondering why the patch will not be fixed to the closed curve at least.

Thanks so much

---

<div class="post-metadata">

**Author:** ![Brunosouth](https://sea3.discourse-cdn.com/mcneel/user_avatar/discourse.mcneel.com/brunosouth/32/514705_2.png) [@Brunosouth](https://discourse.mcneel.com/u/Brunosouth)\
**Post date:** [April 15, 2019, 5:25am UTC](https://discourse.mcneel.com/t/precise-double-curvature-from-curves/81849/13 "2019-04-15T05:25:50Z")

</div>

I will give a tray to this, sounds promising. Thanks !

---

<div class="post-metadata">

**Author:** ![Brunosouth](https://sea3.discourse-cdn.com/mcneel/user_avatar/discourse.mcneel.com/brunosouth/32/514705_2.png) [@Brunosouth](https://discourse.mcneel.com/u/Brunosouth)\
**Post date:** [April 15, 2019, 5:29am UTC](https://discourse.mcneel.com/t/precise-double-curvature-from-curves/81849/14 "2019-04-15T05:29:20Z")

</div>

Thanks !! I am not native English speaker. Do you mean that its cleaner to sweep 1 rial and then trim with the curve instead of recreating the surface with the closed and the two open arcs?

I am interested in your opinion. Thanks so much for helping me to all/

---

<div class="post-metadata">

**Author:** ![Brunosouth](https://sea3.discourse-cdn.com/mcneel/user_avatar/discourse.mcneel.com/brunosouth/32/514705_2.png) [@Brunosouth](https://discourse.mcneel.com/u/Brunosouth)\
**Post date:** [April 15, 2019, 2:22pm UTC](https://discourse.mcneel.com/t/precise-double-curvature-from-curves/81849/15 "2019-04-15T14:22:45Z")

</div>

I tried today cutting the closed curve by the two arcs. Then I grabbed all the curves and make a NetworkSrf and worked pritty well. The boundry curve was still touching the surface where it came from.  
So no gaps between the seat and the sides 🙂 Thanks so much for all your insights.

---

<div class="post-metadata">

**Author:** ![Tone](https://sea3.discourse-cdn.com/mcneel/user_avatar/discourse.mcneel.com/tone/32/718939_2.png) [@Tone](https://discourse.mcneel.com/u/Tone)\
**Post date:** [April 16, 2019, 10:33am UTC](https://discourse.mcneel.com/t/precise-double-curvature-from-curves/81849/16 "2019-04-16T10:33:06Z")

</div>

![saddle](https://global.discourse-cdn.com/mcneel/uploads/default/original/3X/8/6/860b79d513fdfa2ed11a81af10ff8e021cc5e09c.jpeg)  
[saddle.3dm](https://global.discourse-cdn.com/mcneel/uploads/default/original/3X/e/b/eb1f7c80550fd8eb2551ca85239e68c5ce70b77f.3dm) (96.7 KB)  
another way is to use the arcs to create a torus then trim the torus with the outer blue ellipse (you might want to move the seams out of the way before you trim).

---

<div class="post-metadata">

**Author:** ![Brunosouth](https://sea3.discourse-cdn.com/mcneel/user_avatar/discourse.mcneel.com/brunosouth/32/514705_2.png) [@Brunosouth](https://discourse.mcneel.com/u/Brunosouth)\
**Post date:** [April 17, 2019, 7:46am UTC](https://discourse.mcneel.com/t/precise-double-curvature-from-curves/81849/17 "2019-04-17T07:46:04Z")

</div>

This is a great tip. Thanks so much .