Aligning joints on geodesic sphere

[cyangaze]

Been having issues aligning these joints and dowels for a geodesic sphere structure. I'm guessing it's an issue of circular patterns and axis points...been trying for days and it just seems to create a mess even more.

I know the length of the dowels are off in some parts but even where the length is proper, the dowel doesn't insert into the adjacent joints properly:

I've attached the file.

[jhackney1972]

If you are trying to use Joints, you are going to have to convert your Bodies to Components.  Joints will not work very well with bodies.

[cyangaze]

Is that why the circular patterns made the joints not aligned?

[jhackney1972]

Not sure what you are talking about.  You must have Components to create Joint between them.  If you do create a group of joined rods and a socket, a pattern will not carry the joints along with the pattern.  You may be able to use a Rigid Group after a pattern but I do not know about the accuracy of your geometry. 

I also noticed that your geometric dome does not have matching seam lengths, this can be a real problem.

[cyangaze]

What I meant was I was thinking maybe my sockets aren't aligning properly because I didn't make them components like you mentioned. The dome is 2v geodesic, so I believe that's why the sides do not have matching seam lengths? I do see how that is more of a challenge regarding aligning sockets on each vertex though; I was hoping to use less sockets, which is why I chose the 2v dome.

[davebYYPCU]

My fault, I (we?) presumed the 6 way joiner was symmetric - or at least I didn't check that it was.

So the first yellow joiner is the culprit, 5 legs are symmetric on it, (coloured yellow), but the black leg (@ 9 o'clock left view) is not 

End result is that you have a handed 6 way joiner, but only on the hemispherical ring, 10 of these will assemble with 5 inverted.

The other blue 6 way joiners are symmetric but have differing angles to the yellow.

Sorry about that.  Fixed now, but no time to follow through with placements of the subsequent ones.  May have messed up that yellow sketch, and certainly the rest of the timeline will not behave without updating manually.

Might help...

[Drewpan]

Hi,

How accurate are your angles on the connectors? There are two types of Geodesic

Dome/Sphere methods. Those that are calculated and actually touch the inside of

a sphere. Great but you end up with MANY different lengths of connector rods. Or

there is the type that Buckminster Fuller designed that uses only three different

rod lengths that is extremely close to  a sphere but some touch, some are a little

inside and some are a little outside the sphere. Method 1 is basically a custom

design that is hard to scale. Method two is absolutely scaleable as long as the

angles are correct and the ratio of the rods are exact. You should be able to scale

a 2V at an Order of Magnitude for each Decimal Place. I.E. 1M sphere - 1 decimal;

10m sphere - 2 decimals; 100m sphere - 3 decimals; the size of the earth - apparently

only 6 decimals but I haven't checked that. It wouldn't surprise me as you only need

six decimals to work out the radius of the Universe to within 18 inches. (Trivia I have

picked up somewhere).

When I first started on this Forum I had a long discussion about this. Do a search and

there is some very good resources linked to it. I was VERY new, it may have been my

first post.

When I attempted this project manually on AutoCAD I made Pentagonal assemblies

and then joined them together one by one and then filled in the Hexagonal assemblies

which is what you should get with a 2V.

A few comments.

"Red and Yellow Icons in the Timeline should keep you awake at night."

You have too many Move commands. Do NOT use Move unless you THOUROUGHLY

understand what it does and when and why to use it. Move Bites the unwary in

Fusion. It is NOT a simple tool for moving things around.

While it looks very pretty, unless you are going to do a Combine Command at the

end of the exercise all of those bodies are simply floating in space and are NOT

joined at all. You can click and drag every one unless you have Rigid Groups. This

design in its present form would be useless if you actually wanted to fabricate it.

You need to have Components to make Joints that will join everything together.

To do this properly you will need two types of connector Components and three

types of rod Components as your basics. You then create an Assembly and copy

each of the Components you need into the Assembly and Join everything up into

a Hexagonal and a Pentagonal basic Assembly. You then copy these assemblies

and use Joints to connect them together and build your Sphere. You "could" use

Circular Pattern if you wanted but you only need a rod and two Joints to connect

the Hex and Pent Assemblies and it is straight forward.

All of those Axes look pretty and would have taken quite some time to create and

set up. You don't actually need any of them. If your connectors are accurate and

you use Joints then the sphere will build itself.

If you know some of the rods are off then some of your problem is here. Done correctly

a Geodesic object is a precision structure. It is very strong and very forgiving from a

structural sense but it needs to be accurate to work properly.

Cheers

Andrew

[cyangaze]

Hi Drew, I was following Kristian Laholm's method for a 2V Geodome https://www.youtube.com/watch?v=O0rqha3XeWY&t=256s. Do you know of a youtube tutorial that demonstrates the other method (i.e. 3 different rod lengths)? Yes, I am aware there were way too many yellow and red icons in the timeline and move commands, which is why I posted here because I ended up making a mess and likely this was only going to continue to get more inaccurate the way I was going about it, so you're absolutely right about that.

Agreed, I am aiming to get this to be components rather than bodies going forward.

The axes were set up in an attempt to figure out how to get the sockets to align with each vertex using a circular pattern, although I am not sure the way I was doing that was necessary. 

[cyangaze]

Thanks Dave! No problem, appreciate the follow up. Would you be able to attach the file you used? Or some more screenshots? I'm just wondering how you made the yellow joiners vs the blue ones and how you were able to align them to each vertex? If I understand, the yellow connectors have the joint at 9 'o clock but how did you know which vertices that one belonged vs the symmetrical blue ones? I could be wrong but is there also another (beige coloured) joiner?

[davebYYPCU]

I thought I sent the file, but seems the selection was a wrong attachment.  My edits to your file, all edits using the dome before the Shell.

As I understand it, the red and blue joiners in the pic are symmetric.

I used plane on Path or origin planes to cut the joiner mitre cuts.  I used sketch radial lines, similar to your axis’ that gives me a view normal to the vertex.  So the yellow / black joiner mitre construction on the left view - projected the panel edges, and the angles were under / over 60 degrees.  Made for some choice words, 4 individual mitre cuts.

Will send file later today.

Might help…

[cyangaze]

Thanks Dave, I'll have to see visually what you meant by that second paragraph when I see the file in the timeline, but I think I understand. Much appreciated

[Drewpan]

Hi,

I don't know of any YouTube videos but I did a Research Report as part of my

Diploma of Engineering on Geodesic Spheres. A couple of resources I found

very useful were "Dome Book One" and "Dome Book Two" that are Public Domain

now and explain the techniques required to design Domes and other Geodesic

like structures - there is a kool one for an Egg shaped object. There is also

software that is available and https://www.desertdomes.com has some good

resources.

In terms of getting everything to work, if you simply get the connectors with

the angles right, all you need to do is use a simple Rigid Joint and then rotate

the struts until they have the right orientation. As I suggested, if you build the

sphere in Assemblies of Hexagon and Pentagon structures, all you need to do

is join them and orient them and it all comes together. I managed to do this all

manually in AutoCAD.

It may be worthwhile to start again from scratch and set up your Components,

Assemblies and Sub Assemblies correctly. It is a very interesting topic to learn

about and a fun way to apply engineering principles.

Cheers

Andrew

[davebYYPCU · Accepted solution]

Attached here, but since this post I fixed another error with the lower blue 6 way connector, that involves the first triangle patch location.  two pairs at 144 degrees is correct, but the 3rd pair should be 148 ish degrees. 

Might help...

[TheCADWhisperer]

@cyangaze 

If this were my design, I would model single instances of the components in position and Pattern essentially the same as you did the original panels.

No Moves needed.

No Splits needed.

No Remove (bodies) needed.

[cyangaze]

Thanks Dave, sorry for the delay, still studying the timeline but makes sense! Thanks for the solution; I see where you added mirroring and circular patterns that work. What made you realize which parts had to be 148 degrees, was it after creating the red, blue and yellow and determining which ones didn't align?

[cyangaze]

Thanks! I'm still reading the dome books...really cool from a practical and historical perspective. You're absolutely right, it's all about the angles and the rigid joints.

[davebYYPCU]

Generally just Measure Tool, with crease lines from the surface half dome.  Where I was finding non symmetric angles, all originate from the triangle patches.  The 6 way joiners have differing mitre cuts - when done still have some symmetry.

Harder to describe it, than I thought.

Might help….