Mesh Basics
Before a printer can build your model, it has to understand it. And what it understands is surprisingly simple: points, lines, and triangles.
What is a 3D model made of, from the printer's point of view?
Every 3D model is built from three things: vertices (points in 3D space), edges (straight lines connecting two vertices), and faces (flat surfaces enclosed by edges). Together they form a polygon mesh — the skin of your object.
A vertex is just a coordinate: an (x, y, z) position. An edge joins two vertices. A face is the flat patch of surface bounded by edges. Stack enough faces together and they describe any shape — a cube, a gear, a vase.
This matters because the printer never sees your design the way you do. It sees only this mesh of points and flat faces. Everything else — how it prints, whether it prints at all — follows from the quality of that mesh.
Why are 3D models made of triangles?
When you model in a tool like Blender you often work with quads (four-sided faces), because they are easier to edit. But the moment you export for printing, every face is broken down into triangles. This is called triangulation (or tessellation).
Triangles win for one reason: three points always lie on a single flat plane. Four or more points might not — they can twist in space, leaving the surface ambiguous. A triangle is never ambiguous. That reliability is why the standard print format, STL, stores nothing but triangles.
Curved surfaces are an illusion. A cylinder or sphere is approximated by many small flat triangles. The more triangles, the smoother the curve looks — but the curve is never truly round. You will see this directly in code-based modeling: in OpenSCAD the `$fn` value controls how many facets a circle is built from.
Why does the printer see a surface, not a solid?
This is the idea that surprises most beginners: a mesh describes only the surface of an object — its outer shell — not the solid material inside. The mesh is a boundary, like the skin of a balloon. There is nothing 'in' the model; there is only the surface that separates inside from outside.
The printer (really, the slicer software) infers the solid from that surface. If the surface fully encloses a volume — no gaps, no holes — the slicer can confidently say 'this region is inside, fill it; this region is outside, leave it empty.'
This is exactly why the next concept, being watertight or 'manifold', is non-negotiable. If the surface has a hole, the slicer can no longer tell inside from outside — and the print fails or behaves unpredictably.
What are normals, and why do they matter?
Every face has a normal: a direction, perpendicular to the face, that says 'this side points outward.' Normals are how the software knows which side of the surface faces the outside world and which faces the material inside.
When normals are consistent — all pointing outward — the slicer reads the model correctly. When a normal is flipped (pointing inward), that patch of surface looks 'inside out,' and the slicer can misinterpret what is solid and what is hollow.
Most slicers try to auto-repair flipped normals, but relying on automatic repair is risky. Clean, consistent normals are part of a healthy mesh.
How does a printer turn a mesh into layers?
A 3D printer cannot print a whole object at once — it builds it one thin layer at a time. The software that prepares this is the slicer, and the process is slicing.
The slicer takes your mesh and intersects it with a stack of horizontal planes, one per layer height. Each intersection produces a set of 2D outlines — the cross-section of your object at that height. The slicer then plans the printer's movements to trace those outlines (the walls) and fill the interior (the infill).
So the chain is: mesh → horizontal slices → 2D contours → toolpaths. Your beautiful 3D surface becomes hundreds or thousands of flat layers stacked on top of each other.
How does mesh resolution affect quality and file size?
Resolution is the number of triangles used to describe your shape. More triangles mean smoother curves — but also larger files and slower processing. Fewer triangles mean faceted, angular surfaces but lightweight files.
The trick is matching resolution to need. A flat box needs almost no triangles. A smooth curved vase needs many. Adding huge triangle counts to a small or flat part just wastes file size without improving the print, because the printer's own precision has limits too.
In OpenSCAD this is the `$fn` parameter you have already met: `$fn = 8` gives a chunky octagon-like cylinder, while `$fn = 128` gives a visually smooth one. Raising it makes curves smoother and the exported file larger.
STL or 3MF — which file format should you use?
STL is the classic, universal format. It stores only triangles and their normals — nothing else. No units, no color, no metadata. Because an STL has no built-in unit, software has to assume one (commonly millimetres), which is occasionally a source of scaling surprises.
3MF is the modern format. It is richer and more robust: it can carry real-world units, color, multiple objects, and metadata, all in a single well-structured file. It avoids many of STL's ambiguities.
As a rule of thumb: 3MF when your toolchain supports it (it usually does today), STL when you need maximum compatibility or a service bureau asks for it. Either way, the geometry inside is still the same mesh of triangles.