HELP! How to create a script to keep the tube length the same all over the design.

Replies to This Discussion

Permalink Reply by peter fotiadis on February 8, 2017 at 4:47am

Well ...

... that's not that difficult (with Kangaroo1 or 2 (better)) BUT in order to "equalize" something (a collection of lines in our case) you need something else to act as the pulling "constrain" (i.e. a surface/mesh made out of the distorted points).

This outlines this approach PRIOR Kangaroo.

But ... BEFORE I'll provide a similar Kangaroo solution ... and speaking as Engineer I would strongly suggest to forget equal stuff/puzzles (your concern numero uno is NOT equality but planarity) in similar cases and go for a classic truss (with W depth) that can handle big overhangs with easy (say a MERO KK ball system). The nodes and the struts for similar stuff come at no extra charge since Germans (MERO) do these things via CNC and thus it makes absolutely no difference.

Here's one out of a million solutions using similar systems (W is due to a variable option: less at perimeter, more at "center" (a Math.Sin distortion is applied) > that yields a very airy appearance). Other freaky options also available (random this, random that) resulting ... er ... freaky stuff.

Here's a semi freaky option (random "bottom" W distances):

And here's a full freaky thing (both "layers" random):

Shown an abstract visual solution (no balls and no composite struts [i.e. sleeve + cone + rod]).

So ... what you want to do? Decide and a solution could arrive soon.

Permalink Reply by peter fotiadis on February 9, 2017 at 1:25am

BTW: Departing from "trivial"/classic MERO solutions here's the very same "distorted" (fields) surface as above yielding 3 different tensegrity (i.e. push via tubes pull via cables)  types of trusses (a VERY expensive solution, mind):

That said most people feel a bit nervous with tensegrity trusses probably because they don't understand how they work (but lot's of people refused to cross Golden Gate in the old days).

If the rigid parts are made from Carbon (carbon fiber is the concrete/metal of the future) then trusses of that type are actually "barely visible".

Permalink Reply by Gjergji Prendi on February 9, 2017 at 10:33pm

Thanks a lot Peter! This is very helpful! Thank you for your time and knowledge! 

Permalink Reply by peter fotiadis on February 10, 2017 at 12:16am

Well ... what are (finally) your thoughts with regard the desired orientation?

Single layered truss:

Pros: thinner (in theory), suitable for vaults/domes.

Cons:heavy members (in reality), tricky nodes, questionable overhangs, questionable seismic behavior.

Double layered truss:

Pros: thinner (in reality), low cost, hybrids possible (VERY costly), WOW topology, all issues resolved these days, dependable systems available in the market, structurally stiff, seismic behavior, BIG overhangs (up to 20+ meters).

Cons: clash issues (C# required to address that one)

Permalink Reply by Gjergji Prendi on February 23, 2017 at 12:05am

Sorry for the late reply as I was out the country. 

For my design I used the double layered truss as I had to have a flat(horizontal) top layer. 

Thank you!

Permalink Reply by peter fotiadis on February 23, 2017 at 12:54am

Wise decision. Post some pics of the truss (you did it with components?). And given the opportunity what truss type?

So the "deformation" (bottom layer) is used in order to beef-up the "mounting regions"?

And what are your plans with regard the envelope/roof ?

Permalink Reply by Chapulin Colorado on March 2, 2017 at 1:00pm

HI Peter,

really nice work, like always!

I was just wondering about is there anyway to get a structure like this printed/watertight in rhino?

regards!

Permalink Reply by peter fotiadis on March 2, 2017 at 1:19pm

By "structure" you mean the truss? (i.e. struts + balls [+ sleeves if we go by the book] as ONE "decorative"/demo/toy object?)

Permalink Reply by Chapulin Colorado on March 2, 2017 at 2:14pm

exactly i mean the truss, saw the models coming out of exoskelton,-wireframe are often damaged meshes at the nodes

Permalink Reply by peter fotiadis on March 2, 2017 at 11:41pm

Spend some time by studying a real system (say a MERO KK) and spot how "tiny" the sleeves are in comparison to tube/ball radii (since we deal with axial forces and no torque etc etc etc).

This means that your 3d printed model it could be extremely fragile (unless you want to 3d print an indicative portion of some node). Of course it would be impossible to make a demo toy from a tensegrity truss.

On the other hand the scope of a truss is to support an envelope of some sort. That has usually components FAR tinier than the truss members. For instance: imagine planar glazing brackets due to one node (Google planar glazing). 

But if you feel lucky and you want to spend your dollars for no reason > just boolean union all things (Rhino would require days/weeks/months/years to do that).

Suggestion: spend your cash elsewhere.

Permalink Reply by Chapulin Colorado on March 3, 2017 at 7:12am

I already tried some balls/tubes intersections without the tiny sleeves it works more or less well. Other trusses like welded ones were causing enormous pain. Get a solid intersection out of 4-6 tubes is nearly impossible i think or would last days/weeks/months/years which is time i`d like to spend elsewhere.I was searching for another program/way cuase sometimes i need things to be more than just rhino-solid.

Thinking of some autocad software or maybe dp? but to be honest still not read enough.

Hoped you could give me some advice.

Permalink Reply by peter fotiadis on March 3, 2017 at 8:02am

Well

1. Attempting to do a toy/demo model of a real-life truss (even sleeve-less) AS one object ... is a bit kinda tuning a Harley Davidson (I can't think any reason).

2. Other than that in order to unite a zillion of "solids" (closed polysurfaces in Rhino speech) you need a decent solid CAD app. Rhino is a surface modeller ... meaning that you should narrow your search towards the right girl.

3. Personally I work with Microstation (same 3d core engine as Siemens/NX [ParaSolids]) and CATIA/NX. The difference in speed for doing things like these ... well ... find a friend who works with any of these and experience it first hand.