Can you Fasten a vertex to a surface?

jeff_mcaffer

I have a relatively simple surface (floor) and would like to "triangulate" it with a set of triangular parts (tiles). See porch+walk | walk last mile question

I'd like to model this to see how big I can make the triangular tiles and still get a reasonable walking surface. I've been manually fastening these together and tweaking the mate connector angles etc but it would be way easier if I could mate the three tile vertices to the surface like you would a vertex to a plane and then move them around/tweak sizes. Using a Tangent mate for each vertex and the surface seemed promising. Unfortunately, it appears you can't spec a surface in the tangent mate.

Is there another way to approach this problem?

eric_pesty

Hmm…
I'm not sure why the tangent mates aren't working for this surface…

Actually, I asked in the new "advisor" and got this, which basically says it wont' work for such an "arbitrary" surface unfortunately…

There are several reasons why your tangent mate might not be working:

1. Tangent mates have specific limitations on what surfaces can be used. They only support:

- Swept faces like torus and cones [3]

- Faces, edges, or vertices from two different parts [3]

- They do not support offset surfaces [3]

2. The tangent mate requires careful selection of surfaces. Some key tips include:

- Select continuous sets of edges or faces [4]

- Only swept faces are supported, not generic faces like splines [3]

- The mate does not work with any Relations [3]

3. When selecting surfaces, you need to ensure:

- You're selecting faces, edges, or vertices from two different parts [6]

- The surfaces are truly tangent to each other [7]

- You can optionally use the "Flip primary axis" option to adjust alignment [5]

4. If you're experiencing issues, try:

- Checking that you've selected appropriate surfaces

- Verifying the surfaces are continuous

- Using single path or adding fillets to create smoother connections [4]

If you continue to have problems, it may be best to consult Onshape support directly for specific troubleshooting.

You might be able to get something workable if you split your face with a grid, then you can set tangent relations between the face of your triangle and the "nearest" "point" or "edge" of the face.

jeff_mcaffer

Thanks @eric_pesty. Interesting suggestion. Turns out I don't need a grid, just some "stripies" and that helps some. I'm sure I don't fully understand how the Tangent mate works as there are still a bunch of cases where parts of the split surface "should" work but do not. I seems that if the tile is too big for the set of tangents, it won't work but if you change pick a point/edge/face outside the tile, it works. For now I'm just clicking around until something works. Further study is required. I'll report back if I figure something out.

eric_pesty

"Stripes" as you call them can help but a "grid" gives you more options as it 's easier to satisfy the requirements of a grid point being tangent to the bottom of your triangle than a line.

Basically it should take less random clicking if you split it into a grid.

jeff_mcaffer

Hmmm, I did a grid initially as you suggested and that seemed to not work so well so I remove the vertical edges to make strips. Perhaps I did the grid wrong? The surface is 3D but the grid was just a sketch that I used in the Slice feature — so 2D. That seemed to "project" the grid straight down onto the surface. The net result is that the "squares" of the sliced surface, are not square. Looking straight down sure, but the edges are not all the same length in 3D space.

As such, and naively since I don't really understand how the tangent mate works, it would seem impossible for an equilateral right triangle (i.e., diagonally cut square) to have all three vertices tangent to both horizontal and vertical edges of a non-square. I suppose this might be my problem even with stripes. Unless the stipes are exactly the right distance apart (the length of the equal legs in this case) then all three vertices cannot be tangent to the edges. While I can draw the slicing tool that way, the surface is essentially a twisted plane, so the resultant sliced edges are almost definitely NOT the desired distance apart in 3D space. They's most likely not even equal.

Is there a way to slice a surface with lines that are a uniform 3D distance apart?

In any event, I really appreciate your help. I'll try the grid again and research some more on the Tangent mate to see can understand what it's expecting in this context.

Stepping back a bit, do you have any insight as to why the surface cannot be used in this case? It's really a pretty simple surface. Is there something I can do to it to make is usable? Like create it differently? I used some 3D Points matching topological measurements, made planes for the measurement cross-section grid, drew lines between them on sketches and then lofted them together.

vondes_vodes

You’re definitely not alone in running into that issue — when projecting a 2D grid onto a twisted or non-planar surface, the resulting “squares” will inevitably distort in 3D space. The key lies in how the slicing tool interprets projection vectors and surface curvature. Using true 3D spacing requires defining your slicing planes based on the surface normals rather than a flat projection, which can be tricky depending on the software’s geometry engine.