5. 3D Printing & Scanning

Learnings

Our lab has two main types of 3D printer and it’s worth understanding the difference between them.

• FDM (Fused Deposition Modeling) — the standard plastic extrusion printer. It works by melting a plastic filament and depositing it layer by layer to build up a form. It’s accessible, relatively fast, and good for structural parts. The material is typically PLA or ABS — petroleum-based plastics, which is something I’m not particularly comfortable with from a sustainability standpoint.

• Resin (SLA/MSLA) — instead of melting plastic, these printers use a UV light source to cure a liquid resin layer by layer. The resolution is significantly higher, making it better for fine detail. The tradeoff is that resin is more toxic to handle and the post-processing is messier.

Both are additive processes — building up material from nothing — which is fundamentally different from subtractive methods like milling, where you start with a block and remove material.

My reservations about petroleum-based materials aside, what 3D printing offers that almost nothing else can match is the ability to iterate quickly on complex geometry. A perfect example from my own work: I used the printer to test gears for the kinetic mechanism I’m developing for my final project. Printing a gear — with its precise tooth profile, specific pitch, and tight tolerances — would be essentially impossible to produce by hand or through basic subtractive methods at this scale. The printer just does it.

One important lesson was that tolerances matter enormously. When designing parts that need to connect, fit together, or move against each other, you can’t just model the theoretical dimensions — you have to account for how the material actually behaves when printed. Plastic has a slight flexibility to it, and layers bond in ways that affect the final dimensions. I tested this directly when designing grooves to snap-fit rods into place in preparation to make a Jig that can hold rods at a specific angle. The first prints were either too tight or too loose. Testing the tolerance before committing to the final design saved a lot of time and material — and is now something I factor in from the start.

[Photo of snap-fit groove test here]

Anglular Jig Photo [Photo of snap-fit groove test here]

Designing the Lattice and Why It Could Not Be Made Subtractively

The object I chose to print was a lattice structure — a repeating geometric grid made up of nodes and connecting struts. These structures are found throughout nature (bone tissue is a good example) because they are lightweight yet structurally efficient. I wanted to take the dodecahedron I had already modeled in Rhino and use it as the repeating cell of a lattice.

To do this I used the Intralattice plugin for Grasshopper. The key steps were:

• Setting the custom cell — using the Cell/Custom Cell component, I assigned my dodecahedron geometry as the base unit of the lattice.

• Choosing the grid — I connected the cell to a Basic Box component, which tiles the cell across a cartesian grid. Two number sliders control the size of each cell and the number of repetitions along X, Y, and Z.

• Giving it volume — at this stage the lattice is just curves — lines in space with no thickness. The Multipipe component wraps those curves in geometry, turning each strut into a pipe. Two more sliders control the radius and thickness of the pipes.

• Baking — once the proportions felt right, I baked the geometry out of Grasshopper into Rhino as a solid .3dm file.

• Converting to mesh — the baked geometry needs to be converted into a mesh before it can be printed. In Rhino I used the Mesh command to convert the polysurface into a mesh, then exported it as an .STL file — the standard format for 3D printing.

Why this could not be made subtractively: A lattice structure like this is essentially impossible to produce through CNC milling or any other subtractive process. The internal geometry is entirely enclosed — a milling bit simply cannot reach inside the structure to carve out the nodes and connecting struts. The undercuts, the interlocking geometry, the hollow internal volumes — none of it is accessible from the outside. Additive manufacturing is the only practical way to produce this kind of form.

Preparing the File in Chitubox

Once I had the .STL file, I opened it in Chitubox — the slicing software for the resin printer. Slicing software is what sits between your 3D model and the printer: it takes the mesh and translates it into the layer-by-layer instructions the machine actually reads. The key steps in Chitubox before slicing:

• Import the .STL — open the file and position the model on the build plate. Orientation matters — you want to minimize the surface area of each layer to reduce peel forces during printing.Tilt up slightly.

• Add supports — resin printing is bottom-up, meaning each layer is pulled away from the FEP film at the base of the vat. Any overhanging geometry needs supports to stop it from detaching mid-print. Chitubox can generate these automatically, but it’s worth reviewing and adding manual supports to critical areas.

• Set infill — for lattice structures the geometry itself provides the structural logic, so infill is less of a concern. But for solid parts, infill percentage controls how dense the interior is, balancing strength against material use and print time.

• Add drainage holes — this is specific to resin printing. Hollow parts trap uncured liquid resin inside, which adds weight and can cause parts to crack over time as the resin tries to cure. Small holes (typically 2–3mm) are added to the model to let the resin drain out after printing.

• Slice — once all settings are confirmed, Chitubox slices the model into layers and exports it as a .CTB file (the resin printer’s native format). This file goes directly to the printer via USB or network.

[Screenshots of Chitubox settings and sliced model here] [Hero shot of final printed lattice here]

Printing the lattice on the Resine Printer

Due to the structure of the lattice it was best to print it on the resine printer. Resine printing is favorable due to the strucutre of the lattice, the need for supports that and percision that a 3D printer might struggle with. It also produces a seemless print, there each layer is no visible like a regular print. That was the look and asthetic I was aiming for so I went with the resine this time.

Chitubox export

After the designs are ready I used the ‘mesh’ command in Rhino to make them into meshes. Then I ‘exported selected’ to export into STL.

From there the process is straigt forword in Chitubox. After openning the file and placing on the bed you have to go through these major steps before slicing the file and saving it:

• Set the supports: Use Autosupports. There are other ways that use less material for the supports but that required a special subscription.

• After this under the prepare tab I set no infill just due to the size and design of the lattice. I wanted it to be really strong.

• In this case since there is no infill there was no need to dig holes.

Other than this I just went with the stranadrd settings and sliced the print. The results looked really good.

Fresh off the print

The supports looked great

It was satisfying to take them off

Final dodecahedron Lattice (bone lattice)

Reflection

It would have been imposiible to make this form subtractively due to the complexity of its geometry. 3D printing and in this case resine are prefect for such difficult cases. It still amazes me that we can do this using relatively cheap printers!

3D SCANNING & FIXING MESH ON BLENDER

I used scaniverse to scan an object. First I scanned a coffee mug but due to the reflection and poor scanning technique I had a lot of holes in the cup. I did not bother to fix it.

1

Second, I scanned a small 4 faced sculpture that sat on my desk. Although the object is metallic and has some reflection, the scan turned out relatively well on the app. When I moved it into the blender, my definition was significantly worse.

I looked at some tutorials on how to refine the form using the sculpting tools. I would need much more time on this to get good results so I just played around.

Group Assignnment for this week is here

This week the group tested the design rules and capabilities of the Bambu Lab X1-Carbon printer using a standard capability test print — evaluating infill patterns, surface textures, bridging, and overhangs. The results showed clean bridges up to 40mm and reliable overhangs up to 45°, giving a useful baseline for what the machine can handle. The most practical lesson was around tolerances and material behaviour

Design Files

Download Rhino file

Download Grasshopper file