Design for Additive Manufacturing — A Complete Guide

A model that is good for milling is usually a bad model for a printer. A practical guide to designing parts for the layer process — wall thickness, orientation, supports, clearances and material choice.

Most parts sent for printing do not fail because the printer is bad. They fail because the part was designed for a different process and then exported to STL.

Design for Additive Manufacturing means shaping the geometry from the start around the fact that it will be built one layer at a time.

Why a milling model prints badly

A part designed for a mill has rounded internal corners (because the cutter is round), thick uniform walls, accurate round holes, and no concern about orientation.

The same part on a printer:

  • Rounded internal corners are pointless and only add print time
  • Thick walls mean hours of extra printing without proportional strength
  • A horizontal round hole needs support and comes out oval
  • How it sits on the bed changes strength by up to 50 percent

1. Wall thickness comes from the nozzle

This is the most common mistake. Wall thickness should be a multiple of nozzle diameter, not a round number you liked.

With a standard 0.4 mm nozzle:

Wall thickness Result
0.8 mm Two full perimeters — solid and clean
1.0 mm Two perimeters + 0.2 gap — hollow centre
1.2 mm Three full perimeters — solid
1.6 mm Four full perimeters — solid

A 1 mm wall comes out weaker than a 0.8 mm wall, because its middle stays empty. The round number is your enemy here.

2. Orientation is the biggest decision

Layers bond thermally, not with a full molecular bond. The part is measurably weaker across the layers — typically 40 to 60 percent of its in-plane strength.

The rule: orient the part so load runs along the layers, not across them.

A hook pulling a vertical load carries roughly three times more printed lying down than printed standing up.

3. Design supports away instead of tuning them away

Supports are not the problem; removing them is. A supported surface stays rough, and if it is a mating face it has to be machined afterwards.

Three simple changes that remove the need:

  • Make horizontal holes teardrop-shaped. Replace the top of the circle with a peak. It becomes self-supporting.
  • Put a 45° chamfer under overhangs. Anything steeper than 45° prints unsupported.
  • Split long overhangs or carry them on a thin breakaway rib.

4. Clearance for moving parts

If two parts must slide or rotate against each other, you have to leave a gap. How much depends on the process:

  • FDM: 0.3–0.5 mm for free movement, 0.15–0.2 mm for a hand press fit
  • Resin: 0.1–0.2 mm — the process is more accurate

Leave no clearance and the two parts fuse into one.

5. Fasteners and threads

Directly printed threads are weak and strip after a few cycles. Three better options:

  1. Nut trap — a hexagonal pocket holding a standard nut. Cheapest and strongest.
  2. Heat-set insert — a brass bushing melted in with a soldering iron. Best where the joint is opened repeatedly.
  3. Tapping — print a plain hole and cut the thread. Acceptable for coarse threads, M5 and up.

6. Controlling warp

Warp comes from thermal contraction and is severe in ABS, ASA and nylon. Design can reduce it:

  • Round the sharp corners at the base (2–3 mm radius)
  • Reduce bed contact area, or split it into separate islands
  • Avoid abrupt changes in cross-section

7. FDM or resin

FDM Resin
Large structural parts ✓ ✗
Fine detail below 1 mm ✗ ✓
Impact resistance ✓ ✗ (brittle)
Dimensional accuracy ±0.2 mm ±0.05 mm
Cost Lower Higher

8. Choosing a material

  • PLA — prototypes and fit checks. Softens below 60 °C; not for functional parts.
  • PETG — a good default for functional parts. Good to about 75 °C, tough, easy to print.
  • ABS / ASA — up to 95 °C, UV stable (ASA). Warps; needs an enclosure.
  • Nylon (PA) — tough and wear resistant. Absorbs moisture and must be kept dry.
  • PC and PEEK — high temperature and heavy load. Require an industrial machine.

And most importantly: when not to print

If the part carries heavy structural load, if its service temperature is above 100 °C, or if thousands will be made — 3D printing is the wrong route. Choosing the manufacturing method is the first decision, not the last.