Print Settings That Matter to Design

A slicer has hundreds of settings, and as a designer you can ignore most of them. But a handful of them aren't print preferences at all: they're the continuation of decisions you already made in the model. They change how your geometry behaves once it's turned into beads: its real strength, the finish on a curve, whether an overhang sags or holds. Ignoring them doesn't make them go away; it makes you blame the model for a problem that was really a misplaced control. These are the ones that actually move the part coming off the bed, and I'll walk through the physics behind each.

Layer height: finish and fit, not strength

Layer height is the thickness of each slice the machine stacks: 0.2 mm as a working value, 0.1 to 0.3 mm as the useful range with a 0.4 mm nozzle. It's the most visible trade-off in the process. Thin layers take longer: going from 0.2 to 0.1 mm doubles the number of passes and nearly doubles the time (acceleration and travel moves don't scale the same way). But they buy two concrete things that do matter to design.

The first is vertical resolution. Every sloped or curved surface is built by stacking the edges of the layers, so on a gentle slope thick layers leave visible stepping — the stair-stepping. If your part has a dome you want clean or a fine engraving on a non-horizontal face, a 0.1–0.15 mm layer earns its keep; if it's a solid block, 0.2 or 0.3 mm is plenty and far faster. The rule isn't habit: match the layer height to the smallest vertical detail you care about.

The second is less obvious: the mouth of vertical holes improves with thin layers, though not for the reason you'd expect. A straight, vertical hole has no stepping to refine — its wall is perpendicular to the bed, not a slope — and its roundness is set by the number of segments in the perimeter in the XY plane, not by layer height. Where a thin layer does help is at the start: the first layer, printed squashed to stick to the bed, leaves an elephant's foot that narrows the hole's mouth by a few tenths. With thinner layers that elephant's foot takes up a smaller fraction of the hole's useful height, so the pinched neck is shorter and the effective fit lands closer to nominal. It's an effect of the first layer, not the whole wall.

What layer height does not fix is anisotropy. Thinner layers mean more layer welds, not better welds: the weak plane perpendicular to the Z axis is still there, just as weak per unit area. Refining the layer smooths the skin; it doesn't reinforce the Z axis. That weakness is attacked with orientation and with temperature, not with resolution. See Layer adhesion and anisotropy.

Perimeters: the real strength lever

If you can change just one setting to make a part stronger, change the number of perimeters. It's the strength lever that actually responds, and it does so ahead of infill for a physical reason: a perimeter is continuous plastic that wraps the entire contour, a solid wall that works in tension and in bending along the bead, where the material works at nearly full strength. Infill is a mesh of crossing threads with gaps; it contributes little in tension, and its real job is to support the upper layers so they don't sag, not to resist load.

That's why the trade-off is lopsided. Going from two to four perimeters adds a couple of solid wall beads right where the stress is born, for little plastic and little time. Raising infill from 20 to 40% spends nearly twice the material and time in exchange for much less than twice the strength. Start here: 15–20% infill for most parts, and when it gives or cracks, raise perimeters before you touch infill. The full argument, and sizing walls as whole multiples of the bead width, are in Walls, perimeters and infill.

The connection to your design is dimensional, and it starts with a decision that lives in the slicer, not the model: the bead width. With a 0.4 mm nozzle you can ask for beads anywhere from 0.4 to 0.6 mm; it's not a constant, it's a setting you have to marry to your walls. Set it, say, to 0.45 mm, and then the math works out: a 1.35 mm wall is exactly three perimeters. Draw it at 1.2 mm with that same width and the slicer lays down two perimeters and leaves a thin gap in the middle that it either fills badly or skips; your solid wall stops being solid. Size the wall as a whole multiple of the bead width you're going to use, not the other way around.

Temperature and cooling: an explicit trade-off

This is the setting that confuses people most, because it pushes in two directions at once and there's no value that wins at everything. Nozzle temperature and the part-cooling fan together govern how long each bead stays hot enough to fuse with its neighbor. More heat at the nozzle and less fan keep the interface hot longer, so the new bead melts the one below better and the layer weld climbs toward the high end of its range. That figure is a rough guide and depends on the material and the profile — on the order of 40–80% of the solid bead's strength, higher in a well-printed PLA, lower in a PETG or an ABS that's poorly cooled — and it collapses when cooling is aggressive or the plastic comes out cold. It's the direct lever on anisotropy: if a part cracks along a layer line under load, a few more degrees and less fan usually do more than any change to the model.

But the same heat that bonds also makes overhangs worse. A bead that hangs partly in the air needs to cool and solidify fast so it doesn't sag out of place; that's exactly what the fan does and what heat delays. The degradation is gradual: on demanding geometry it shows up well before the geometric limit of 45° without support. Raise the temperature and drop the fan to weld better in Z, and overhangs and fine detail come out coarser and saggier. Lower them so overhangs come out clean, and you'll bond worse between layers.

There's no free lunch: it's a trade-off you decide based on what the part is for. A structural lever that's going to take load wants the hot side of the range and little fan. A housing with overhangs and fine raised text wants the cold side and full fan. PLA tolerates the air better than PETG or ABS, which is why it prints clean overhangs; even so, on a structural PLA part, dropping the fan measurably improves the Z bond. PETG and ABS weld worse with a lot of air and prefer less fan, which costs them overhangs. Temperature isn't a number the preset gets right for you: it's a design decision dressed up as a setting.

Speed and flow: the scale the rest works at

Speed and flow (the extrusion multiplier) aren't independent settings: they shift the real dimension at which all the others work. Printing faster gives each bead less hot-contact time with the one below, so the same temperature that welded well at normal speed bonds worse at high speed; as a rule of thumb, raising the speed is like lowering the temperature, though the equivalence isn't exact — at higher speed you usually have to add a bit of heat to keep the melt flow up. That's why two places call for going slow on purpose: the first layer, where bed adhesion is decided, and tall, thin geometry, which wobbles when the head hits it on every pass.

Flow moves the other axis: the dimensional one. Bead width is flow made geometry: extrude too much and the beads fatten, the walls come out thick and the holes tighter; extrude too little and gaps appear between beads that weaken the part. Since all your tolerances rest on the real bead width, a miscalibrated flow shifts every clearance and every press fit you sized, all at once.

The practical upshot is simple: when you change material, recalibrate. Every filament has its own temperature window, its own shrinkage on cooling, and its own flow behavior. A PETG at your PLA's speed and temperature bonds differently, strings more (stringing), and spreads wider. It can shift a clearance by 0.05–0.15 mm — exactly the margin between "slides" and "seizes"; badly miscalibrated, the shift reaches 0.2–0.3 mm. The numbers you measured and saved hold as long as you don't change material or nozzle; the moment you change one, they're a guess again until you reprint the calibration.

The first layer deserves its own chapter, because almost every slicer gives the first layer its own height and temperature, and both bed adhesion and the elephant's foot that pinches your hole mouths depend on it; The first layer and bed adhesion covers it.

Five controls, each tied to a decision you already locked into the geometry: layer height to your vertical detail, perimeters to strength, temperature and fan to the trade-off between bond and overhang, speed and flow to the effective dimension of everything else. Learn how they respond to you and they'll stop being a black box — a set of deliberate levers, each answering to a decision you already made in the model.