1. Print orientation — free, and worth more than everything else

COST: $0 · EFFORT: ONE ROTATION IN THE SLICER

Rotate the part so the expected load runs along the layers, not across them. A hook printed flat on its side (layers running along the hook's curve) can hold dramatically more than the same hook printed standing up, where the load tries to peel the layers apart — this is the most consistently replicated result in community strength testing. Ask one question before every functional print: "Which way does the force go, and does it cross my layer lines?" If the answer is yes, rotate — or split the part and print it in two pieces oriented correctly. Orientation also changes where supports land; if reorienting creates overhangs on critical surfaces, modeling your own supports keeps those faces clean.

2. More walls, not more infill

COST: A LITTLE FILAMENT · EFFORT: ONE SETTING

The counterintuitive finding that community testing keeps confirming: perimeters add far more strength per gram than infill. Like an I-beam, most bending and torsion stress is carried at the outside of the part — exactly where walls live. Going from 2 walls to 4 typically buys more real-world strength than doubling infill percentage, often for less material. Our shop defaults for functional parts: 4 walls, 4–5 top/bottom layers, then infill as the supporting cast. 100% infill is almost never the answer — past roughly 40–60%, testers consistently report steep diminishing returns while print time and weight balloon.

3. Better layer welds: heat, speed, and dry filament

COST: $0–$80 · EFFORT: CALIBRATION + A DRYER

Layer adhesion is a welding problem, and welds like heat. Printing at the upper end of your filament's temperature range generally improves layer bonding measurably — community testing repeatedly shows hotter nozzles producing stronger Z-direction results, at the cost of slightly more stringing. Slow down for small parts so each layer stays warm enough to bond, keep part-cooling fans moderate on functional prints (max cooling is for looks, not strength), and — the San Diego special — dry your filament. Moisture boils in the hotend and leaves micro-voids in every weld; wet spools are one of the most common causes of parts that snap "for no reason." Coastal humidity makes this a when-not-if problem here — our filament dryer guide covers the fix.

4. Wider lines, bigger nozzle

COST: $0 (LINE WIDTH) OR ~$15 (0.6MM NOZZLE) · EFFORT: LOW

Fatter extrusions mean fewer seams and bigger weld area between lines and layers. Bumping line width to 110–120% of nozzle diameter is a free strength gain most slicers allow with zero hardware changes. The bigger move: a 0.6 mm nozzle, which lays down beefier perimeters, prints faster, and — as a bonus for anyone eyeing carbon-fiber filaments — resists the clogging that fiber-filled materials cause in 0.4 mm nozzles (though CF has its own handling considerations, covered in our carbon fiber safety piece).

5. Infill — pattern and percentage, in that order

COST: FILAMENT + TIME · EFFORT: ONE SETTING

Infill still matters — it supports the walls, resists crushing, and gives screws something to bite. Pattern guidance the community broadly agrees on: gyroid or cubic for loads from multiple directions (and gyroid prints fast with no crossing lines), grid/rectilinear when the load direction is known and in-plane, and honeycomb variants for stiffness at the cost of print time. Percentages that cover most real parts: 15–25% for general use, 30–50% for hard-working brackets and fixtures. We used exactly this logic speccing the frame parts in our 3D printed drone build — 40–60% gyroid where motors bolt on, lighter elsewhere.

6. Brick layers — the new trick that removes the weak plane

COST: $0 · EFFORT: MODERATE (SCRIPT/SLICER SUPPORT VARIES)

Normally every wall extrusion sits directly on top of the one below it, so the seams between layers align into one continuous weak plane. Brick layering offsets alternating perimeters by half a layer height — like staggering bricks in a wall — so no single plane runs through the part. In CNC Kitchen's initial testing the technique improved tensile strength across layers by roughly 10% (PETG) to 14% (PLA), and proponents argue tuned implementations can do better. Open-source implementations (notably Roman Tenger's "Bricklayers" post-processing script for PrusaSlicer and OrcaSlicer) made it accessible to everyone.

The honest asterisk: the technique has a messy patent history — an original Stratasys patent from the 1990s expired years ago, but a 2020 patent filing by another company created enough uncertainty that major slicers were slow to ship it natively. Industry coverage generally concludes hobbyist and non-commercial use is not the concern, and no equivalent block exists outside the US — but if you're printing commercial products, that's a question for an IP attorney, not a blog. We're a print shop, not lawyers.

7. Annealing — bake your PLA (carefully)

COST: $0 · EFFORT: MODERATE, WITH REAL TRADE-OFFS

Annealing means heating a finished print below its melting point so the polymer crystallizes, which can meaningfully raise heat resistance and, in many tests, stiffness and strength — it's the standard answer to "my PLA part sagged in a hot car." The catch is dimensional: standard PLA typically shrinks a few percent in X–Y and can grow in Z during annealing, so precision parts come out the wrong size unless you scale the model to compensate or use a filament designed for annealing (several "HT-PLA" formulations target low-warp annealing). It's a great trick for chunky brackets, a poor one for snap-fits. The basic process:

1. Check your filament's guidance first. Annealing temperatures and shrinkage vary by brand and material; the manufacturer's numbers beat any generic recipe.
2. Preheat an oven to roughly 80–100°C for standard PLA — and verify with an oven thermometer, because home ovens swing wildly and overshooting turns your part into modern art.
3. Support the part on a flat tray (a bed of sand or salt helps prevent sagging on complex shapes).
4. Bake 30–60 minutes depending on part thickness, then turn the oven off.
5. Let it cool slowly inside the oven. Rapid cooling reintroduces internal stress and can crack the part.

Common-sense safety note: it's your kitchen oven and your part — stay nearby, don't anneal at temperatures that scorch plastic, and expect some plastic smell.

8. Brass heat-set inserts — stop threading into plastic

COST: ~$10–20 PER ASSORTMENT + SOLDERING IRON · EFFORT: LOW

Printed threads and self-tapping screws are the first thing to strip on any part that gets assembled more than once. Brass heat-set inserts fix that permanently: design a slightly undersized hole, press the knurled insert in with a soldering iron, and the plastic flows into the knurls as it cools — leaving real metal threads that survive repeated assembly and distribute load into the part instead of into three fragile plastic ridges. This is standard practice on every screwed part that leaves our shop, and it's the trick that makes rebuilt parts stronger than their injection-molded originals in our scan-to-print reverse engineering work. Two design rules: leave at least 2–3 mm of wall around the insert, and put inserts in holes printed vertically where possible so the insert's load doesn't pry layers apart.

9. Design for strength — fillets, ribs, and mass where it matters

COST: $0 · EFFORT: CAD SKILLS

If you control the model, you control the strength. Sharp inside corners are stress concentrators — a simple fillet at the base of any tab, hook, or wall can multiply its fatigue life. Ribs and gussets stiffen thin walls for grams of material. Thicken the part exactly where it broke last time (that's the whole philosophy behind our rebuild-don't-photocopy repair work). And design around the layers: a snap-fit clip should flex along the layer direction, never across it. These changes routinely outperform any slicer setting because they fix the geometry the settings are trying to compensate for.

10. When settings run out: change the material

COST: $20–70/SPOOL · EFFORT: NEW PROFILES, MAYBE AN ENCLOSURE

Sometimes the honest answer is that no PLA setting makes PLA into the right material. The practical ladder: PLA (stiff, strong in a straight pull, brittle under impact, softens around 55°C) → PETG (tougher, heat-tolerant to ~70–80°C, and with layer adhesion so good its Z-strength approaches its X–Y strength — the single easiest material upgrade for functional parts) → ABS/ASA (impact resistance and real heat tolerance, enclosure required) → nylon, polycarbonate, and fiber-filled blends (genuine engineering territory with genuine printing demands). The full property-by-property breakdown, with specific product picks, lives in our definitive 2026 filament guide.