Need help to constrain panel

[mjh310]

I posted a similar question before, but still having problems so will try to ask a different way.  I am trying to create a panel that behaves as pictured.  The corner at point 2 will always be perpendicular, but the other corners could be any angle.  I am having trouble figuring out how to constrain the reference lines to be perpendicular to the reference lines extending from reference point 2, but still connecting to the line opposite. Any tips for me? 

[ToanDN]

It certainly can be done as regular family or as a 2-point adaptive family.  What is the application of this panel?

[mjh310]

This is for a panel for a curtain wall.  This panel will only be used along the ends and top of the curtain wall.  The wall will always be planar, though will be sloped.

[ToanDN]

---

@mjh310 wrote:

This is for a panel for a curtain wall.  This panel will only be used along the ends and top of the curtain wall.  The wall will always be planar, though will be sloped.

---

Custom shaped curtain panels won't work unless you place them independently.  Why not use system panel types or basic wall types as panels so that they can adapt to any shape and size?

[mjh310]

I will be placing these independently (with some help from Dynamo).  I am using regular system panels for the majority of the curtain wall. The system panels are all symmetrical, but the panels at the edges and top will be asymmetrical.

[ToanDN]

---

@mjh310 wrote:

I will be placing these independently (with some help from Dynamo).  I am using regular system panels for the majority of the curtain wall. The system panels are all symmetrical, but the panels at the edges and top will be asymmetrical.

---

If you use Dynamo to place then just use a 4-point adaptive family.  The shape should defined by the curtain wall  or curtain system and the adaptive panel just need to adapt to it.

Again, sans knowing in and out of you model, my concession is that it is overly complicated approach.  System panels or basic wall panels work just fine and are much more flexible.

[mjh310]

Where I am stuck right now is constraining the reference lines to be 90 degrees.  I can't change the angular dimension  and can't figure out why.

[barthbradley]

Create cross-through Ref. Lines via Spine Through Points that are hosted to the Perimeter Reference Lines. 

 

[mjh310]

Okay, I gave it a go.  I create a point on the bottom perimeter reference line with Measurement Type "Segment Length" so I can control the distance from Point 2.  I then created a point on the perimeter reference line opposite, and set that point to be by "Driving Curves" figuring I would let the cross through reference line define exactly where it hits that line.  Use Spline through points to create the curve, but still can't change the angle of the cross through reference line. It allows me to select it and type in a value, but the angular dimensions doesn't change.

[mjh310]

From your screenshot, I can see the difference.  I do not want the cross-through lines to go through the midpoint of the bottom line, I want it to be a specific distance from Point 2, so I specified the Segment length.  That doesn't work on the top reference line though because it might be angled and not sure what the segment length would be.

[mjh310]

Thought it might help if I showed what I am trying to achieve.  Here is a picture of the wall I am trying to construct.  This panel is the one on the end.  I will also want to use the same or somewhat similar along the top and opposite side. Ideally I would like the crossbar to extend from the left of the circle too.

[barthbradley]

Points are measured from Beginning or End. See the "Measure From" Parameter under the Point's Properties. 

[ToanDN]

Try this skeleton.  It should work unless your shapes are extremely stretched.

[mjh310]

Understood.  I am able to get the point on the bottom line just where I want it, but given the various potential shapes and sizes of the panels, how do I define the top point? Depending on the angle of the top line, the distance from the end (or beginning) will vary.

[mjh310]

Thanks! Looks promising. I will give it a go.

[mjh310]

I exercised this skeleton by moving point 3 to the left and point 4 upwards and it didn't behave like it needs to.  The center of the circle needs to remain in the same position relative to point 2 regardless and the vertical and horizontal crossbars need to stay connected to the left and top of the outer frame. I tried create my own panel a bit ago and this is just the type of problems I had.

[ToanDN]

---

@mjh310 wrote:

I exercised this skeleton by moving point 3 to the left and point 4 upwards and it didn't behave like it needs to.  The center of the circle needs to remain in the same position relative to point 2 regardless and the vertical and horizontal crossbars need to stay connected to the left and top of the outer frame. I tried create my own panel a bit ago and this is just the type of problems I had.

---

Yeah I said if not stretching it too much.  Do you even have such condition or it is a hypothetical one?  Based on the image you post, you don't have any of such extreme condition.

[mjh310]

Maybe it will be okay.  I will add some solid geometry and try it out.  Thank you!

[mjh310]

Unfortunately, the skeleton fails in some of the panels needed.  I'll keep exploring options.  Thank you so much for your time helping me through this.

[ToanDN]

You can create the divided surface by intersecting ref planes instead of UV grid so that you can fixate the 2 ref planes at the circle to point 2 with parameters.