Too many Breps with Solid Difference

Replies to This Discussion

Permalink Reply by A:RK on February 12, 2012 at 3:05pm

Here is the simple sphere Rhino File and the Definition.

Thanks in advance.

Attachments:

• Male and Femal Ribbed Test Sphere Scheduled.3dm, 49 KB
• Twisting planes.ghx, 659 KB

Permalink Reply by Hannes Löschke on February 12, 2012 at 3:17pm

Path lengths need to match for the sDiff component. Either you flatten and graft or just use simplify to reduce to the minimal path complexity.

Attachments:

• Twistingplanes_hl.ghx, 660 KB

Permalink Reply by A:RK on February 12, 2012 at 3:37pm

Hannes,

Thanks for this. I downloaded your attached definition and for some reason it doesn't seem to quite work. I have to flatten the A of the solid difference in order to get the notches in the new set of breps. Once i do that the simplify of the extrusions has no effect on the outcome of the solid difference. I still get the same result i get all the notches i need but each notch is on an individual brep!

Any thoughts? Did it work on your computer? Maybe it's the Service release of Rhino or the Build of Grasshopper that i am using?

Thanks Kevin

Permalink Reply by Hannes Löschke on February 12, 2012 at 4:14pm

Yeah right. Didn't really see what you wanted to do. My guess was you wanted to match and intersect.

You will need to flatten the input that creates the notches. So one rib matches up with all your cutting geometry.

Problem is that your Geometry seems to be problematic to intersect. For 4 ribs it works ok then some of the intersections start to fail. No valid result at all fo 17 ribs.

Looks like you are creating some weird/unwanted geometry on the way. Here's what the cutting breps look like

Permalink Reply by A:RK on February 12, 2012 at 4:57pm

yeah i noticed that erronious rib myself. I'll jave a look tomorrow and see where that's coming from. Can't imagine that is causing the multiple brep problem tho.

So how do i flatten the cutting geometry and the geometry to be cut? I've tried using flatten but it doesn't make a difference!

Thanks K

Permalink Reply by A:RK on February 13, 2012 at 2:01am

Is there anyone else who might have a minute to look at this issue? or have any suggestions as to what might be causing it?

All Help appreciated!

X Kevin

Permalink Reply by Hannes Löschke on February 13, 2012 at 3:18am

Grasshopper evaluates the component for every branch and matches it to the other input.

If you have a rib and want match it up to a list ot cutting geometry, then you need to have the cutting geometry in a (flattened) list. To apply this cutting geometry to every rib seperately, each rib has to be in its seperate branch (not flattened).

Here's what it looks like for 5 ribs.

Permalink Reply by A:RK on February 13, 2012 at 5:52am

Thanks for your help,

I'm not at home so i will have to check it later. Does it work for more than 5? And if not why not?

Thanks K

Permalink Reply by Hannes Löschke on February 13, 2012 at 12:39pm

Just tried some samples above 5. As mentioned above some of the solidDiff start to fail. Like in Rhino this has soemthing to do with tolerance settings and very small intersection areas.

Right at the bottom of the arcs you can see an area wher all the cutting arcs try to cut away from the ribs. Also with higher rib counts the voids start to cut with a very shallow angle. My guess is that there is your problem.Do you need those intersections there? Is the shallow angle intended?

Permalink Reply by A:RK on February 13, 2012 at 2:15pm

Hannes,

Thanks. I've spent the evening going through the definition and i've managed to rebuild it better without any erroneous geometry. I think the above problem is largely to do with a misplaced centre point when scaling curves. I have built it another way have a  look at the attached.

I have encountered 'boolean' tolerance issues before so i think i might have a look int the Rhino settings.

Have a look at the new definition and let me know if it seems better/works better?

Thanks for your help mate.

X K

Attachments:

• Twistingplanes_02.ghx, 720 KB

Permalink Reply by Hannes Löschke on February 13, 2012 at 3:05pm

1) you are trying to union a set of closed breps with a surface. (extrude doesn't automatically cap the extrusion as in rhino). extrusion and breps weren't on the same branch so no union at all.

1a) Still the union fails with corrected input. Typically this is due to incident seams or vertices. You can test this with baking a few of the cutting ribs and try them to union in rhino. I got 3 of them to union with a tolerance of 0.00001. 4 still won't join.

2) When extruding the cylinder in both directions and cutting it away before cutting out the notches it works. Though I'm not sure this is what you want.

3) The way the ribs intersect at the bottom still seems odd to me from a manufacturing point of view. The way the ribs intesect, the notches basically cut away most of the rib.

Attachments:

• Twistingplanes_02.ghx, 740 KB

Permalink Reply by A:RK on February 13, 2012 at 3:37pm

Thanks Hannes.

You're right it did work your way. I did it a very similar way where i reduced the length of the arc so that they didn't interset at the bottom - effectively the same as solid differencing the cylinder.

I can only conclude that it was a tolerance issue and that because of the intersection at the bottom erasing almost all of the rib (as you stated) the issue arose.

With regards to a manufacturing POV. I am not going to be manufacturing this directly i am using it as a model to discuss inherent triangulation properties of different thicknesses of different pieces of timber. The structure itself at this point is not braced - it requires triangulation - especially if we are going to apply a live load onto the structure.

I am hopefully going to apply this to a different form that we have designed in the hope of fabricating the lattice.

Anyways, Thank you very much for your help with everything - you have been very patient and kind.

We are a new architecture firm in london called A:RK and our website is going to go live in 2 weeks (hopefully). I'll send you the link if you want to have a look at the larger project for which i am building this.

Thanks again Hannes!

XK