Which print settings actually make a part stronger?

An FDM part is not a solid. It is a stack of welded beads, and almost everything about its strength comes down to how well those beads fused and which direction the load runs relative to them. That reframes the settings question: you are not tuning a material, you are tuning a weld.

1. Orientation — the setting that is not in the slicer

Between layers, a printed part typically reaches 30–50 % of its in-plane strength. That is a bigger spread than any two filaments in our catalogue. A PLA part loaded along the layers will out-perform a PETG part loaded across them.

• —Find the load path first, then rotate the part so it runs along layers, not across them.
• —A bracket that carries a bending load should be printed standing up, so the tension face is a continuous wall, not a stack of layer boundaries.
• —Where a load path is unavoidably across layers, add material: a fillet, a rib, or a gusset costs grams and buys back most of the loss.

2. Wall count, then wall count again

Walls are continuous extrusions that follow the perimeter, so they carry tension and bending far better than infill can. In our bend tests, going from two walls to four does more for stiffness and failure load than raising infill from 20 % to 60 %, and it prints faster.

The practical rule: set walls so the loaded section is solid, and use infill to stop those walls buckling. For a 3 mm rib with a 0.4 mm nozzle, four walls is already a solid rib and infill is irrelevant.

3. Nozzle temperature

Interlayer strength is a diffusion problem: polymer chains have to cross the boundary and tangle while the material is above its glass transition. Hotter melt stays mobile longer, so more of that happens. Ten degrees is usually worth 10–20 % on interlayer strength, which is a large return for a free change.

The limit is surface quality and stringing, not the polymer. Raise temperature until stringing appears, then come back 5 °C. Cooling works against you here: cutting part cooling to the minimum the geometry allows keeps the interface hot enough to fuse.

4. Layer height (and why thinner is not stronger)

Thin layers give better surface detail and more accurate curves. They do not reliably give stronger parts. Thicker layers put more material in each weld, cool more slowly and often bond better — the trade is resolution and overhang quality.

The setting that matters more is extrusion width. Widening it to about 120 % of nozzle diameter increases bead overlap and squashes each bead harder into the one below. That is a direct improvement to the weld.

What barely moves the needle

• —Infill above roughly 40 %. Diminishing returns, then weight and time only. Solid parts are also more likely to warp.
• —Infill pattern, for most load cases. Gyroid, cubic and grid are within a few percent of each other once wall count is right.
• —Print speed, within reason. It changes surface finish and dimensional accuracy long before it changes strength — provided the hot end can keep the melt at temperature.
• —Annealing PLA, in most cases. It raises heat resistance and stiffness but shrinks the part 1–2 % unpredictably, which usually costs you the fit.

A working recipe

1. 01Orient the part so the load runs along the layers.
2. 02Set walls so the loaded cross-section is fully solid — typically 4 for a 0.4 mm nozzle.
3. 03Infill 25–35 %, gyroid or cubic. Do not go higher without a reason.
4. 04Nozzle at the top of the published range; cooling at the minimum the overhangs tolerate.
5. 05Extrusion width 0.45–0.48 mm on a 0.4 mm nozzle.
6. 06Dry filament. A wet spool undoes every one of the changes above.

If the part still fails, the material is the next lever — that is where nylon and composites and polycarbonate earn their price. Settings first, though: they are free.