Skip to content

Designing for FDM

A model can be geometrically perfect and still print terribly. FDM builds objects one horizontal layer at a time, out of squished plastic strings — and that process has opinions. Design with them, not against them.

Walls: multiples of the line width

A printed wall is made of parallel perimeters (also called walls/loops), each one line width wide — usually 0.4–0.45 mm from a 0.4 mm nozzle. Design wall thicknesses as whole multiples:

Perimeters Wall thickness (0.4 mm nozzle) Use for
2 0.8 mm Light-duty, prototypes
3 1.2 mm General purpose — the default
4–5 1.6–2.0 mm Load-bearing, screw bosses

A 1.0 mm wall on a 0.4 mm nozzle forces the slicer into awkward fractions (two perimeters + a gap, or a wobbly single fat line). A 1.2 mm wall is three clean perimeters. Model to the process: pick 0.8 / 1.2 / 1.6 / 2.0 and move on.

(If you run a 0.6 mm nozzle, the multiples shift to 1.2 / 1.8 / 2.4. Same idea.)

Overhangs: the 45° rule

Each layer needs something underneath it. A wall leaning out at 45° or steeper from vertical prints fine — each layer is supported by most of the one below. Shallower than 45°, and you're printing in mid-air.

Consequences for design:

  • Replace 90° ledges with 45° chamfers where you can. A 45° chamfer prints cleanly; a flat overhang needs supports.
  • Arches beat flat ceilings: a 20 mm wide flat ceiling sags; an arched one (even a shallow arch) bridges much better.
  • Bridges — horizontal spans between two supports — work up to roughly 10–25 mm depending on filament and cooling. Design short bridges; the slicer stretches straight strands across them.

Orientation: choose what the layers do

Parts are strongest along the layers (X/Y) and weakest between layers (Z) — layer adhesion is the weak axis, often half the strength. So:

  • Put mechanical loads across layers, not along Z. A hook that holds weight should print lying down, not standing up.
  • Put the prettiest face up or vertical — the bottom gets the bed texture, overhangs get rougher.
  • Put holes with their axis vertical when roundness matters — a vertical hole is a clean circle; a horizontal hole prints as a slightly saggy oval (and needs supports above ~10 mm diameter).

The L-bracket in the project chapter prints flat on its base for exactly these reasons: no supports, strong layer orientation for the upright's loads, clean vertical holes.

Supports: a cost, not a crime

Sometimes you need them. But supports cost time, filament, and surface quality, so treat them as a design smell: can I add a 45° chamfer instead? Can I split this into two parts that bolt together? Splitting a complex part into two simple printed parts plus screws is often faster and stronger than one heroic print.

Threads: don't model what you can buy

  • Printed threads work for coarse, low-load jobs (a jar lid, a knob) — model them big (M12+) and with generous clearance. Small printed threads (M3–M6) are fiddly and weak.
  • Heat-set inserts are the pro move for real fasteners: model a plain hole (sized per the insert's datasheet, typically screw Ø + ~1.5 mm), press the brass insert in with a soldering iron. Strong, repeatable, removable screws.
  • Captive nuts are the free alternative: model a hexagonal pocket, drop a nut in mid-print (pause) or slide it in after. Size the pocket nut-flat-to-nut-flat + 0.2–0.3 mm.

Text and details

Embossed (raised) or debossed (recessed) text needs to be at least 0.6–0.8 mm tall/deep and chunky — fine serif fonts won't survive a 0.4 mm nozzle. Sans-serif, bold, ≥ 4 mm letter height is the safe zone. Details smaller than your line width simply don't exist to the printer.

The pre-print checklist

Before you export (see Exporting and Sharing), run through this:

  1. Walls are multiples of line width (0.8 / 1.2 / 1.6 / 2.0)?
  2. Overhangs ≥ 45° from vertical, or supported, or bridged short?
  3. Weak axis (Z) isn't carrying the main load?
  4. Holes oversized for shrinkage (tolerances)?
  5. Bottom face flat on Z = 0, no floating bits?

Five yeses and the slicer is on your side.

Next: Project: The L-Bracket