How FDM Shapes Your Design

An FDM printer is, jargon aside, a hot glue gun mounted on a robot: it melts a strand of plastic and draws your part out of it, one thin bead at a time, each layer stacked on the one before. That one image is the most useful thing you can know about 3D printing, because almost everything else—the strength, the precision, the overhangs, the clearances you have to open up—none of it is an arbitrary quirk of the process. It all falls straight out of how that plastic gets laid down. Understand the how and the rules stop being a list to memorize and become consequences you can work out on your own.

This article builds that mental model. The rest of the path—Orientation and Overhangs, Supports and Bridging, Walls, perimeters and infill, Real Printed Clearances—is just this model applied to concrete problems.

It's Built Bead by Bead, Not Carved

If you're coming from a milling machine, or picturing a sculptor with a block of marble, drop that image. What you're picturing is subtractive manufacturing: you start from a solid and take away what you don't want. FDM does the opposite: it's additive and, more importantly, sequential. It starts from nothing and adds material exactly where the part should be, tracing its shape with a moving nozzle. The nozzle never cuts a detail: it follows a path and squeezes out molten plastic as it goes. That plastic solidifies as it cools, and the solidified bead is the part.

Everything else follows from that. Every surface you see on a finished print is the edge of a path the nozzle traveled. Every mechanical property is a consequence of how those beads welded to each other. That's why the design question isn't "what shape do I want?" but "what beads does the nozzle have to lay down to make this, and in what order?" Ask yourself that systematically and you solve most printing problems before they happen.

Layer by Layer, Bottom Up: Nothing Hangs in Midair

The nozzle moves freely in X and Y—side to side and front to back. To gain height, the part (or the nozzle) rises by one layer height and the next layer is drawn on top of the one just finished. That step is typically between 0.1 and 0.3 mm, with 0.2 mm as the usual working value. The maximum isn't fixed: the physical rule is to keep layer height under ~75% of the nozzle diameter, so with the standard 0.4 mm nozzle you top out around 0.3 mm.

The consequence you need to internalize is physical, not aesthetic: each new bead can only rest on something that's already there. The plastic comes out hot and soft; if there's nothing beneath to catch it, it droops. The bed is the foundation for layer one, layer one is the foundation for layer two, and so on all the way up. When a print fails, you can almost always see it by picturing the part drawing itself slab by slab up from the bed: the first layer with no material under it is the one that's going to droop.

The Bead Sets the Resolution: Nothing Is Finer Than a Line

The bead has a width, and that width is your minimum unit in the plane. The 0.4 mm nozzle—the standard one—lays a line of configurable width, typically between 0.4 and 0.6 mm, because the plastic squashes out a little as it exits and the slicer (the software that cuts the part into layers) usually aims for 100–120% of the diameter. In X-Y nothing can be narrower than one bead, and in Z nothing thinner than one layer height. That minimum grid decides which details survive and which don't.

Anything smaller than one bead won't resolve reliably. A sharp edge doesn't come out sharp: the nozzle can't trace a zero-radius corner, so it rounds it off to the radius of its own tip. Engraved text 0.2 mm wide doesn't resolve—there's no way to fit even half a bead. Walls come out better if you size them as whole multiples of the bead width you've chosen, because then the nozzle fills them with complete perimeters and doesn't leave a thin gap that the slicer fills poorly or ignores. Vertically, layer height rules: sloped and curved surfaces turn out stepped, and dropping the layer from 0.2 to 0.1 mm smooths them at the cost of doubling print time.

Anisotropy: Strong Within a Layer, Weak Between Layers

Here's the fact that governs more design decisions than any other. Within a single layer, the bead is continuous plastic extruded hot: along the bead, the part has essentially the material's full strength. Between layers there's no continuous material, only a weld: the new bead is laid down hot onto the one below, which is already cooling, and the two fuse only across their contact zone. That bond runs about 40–80% of the strength the material has along the bead—PLA usually lands at 50–70%—and it drifts toward the low end when cooling is aggressive or the extrusion temperature is low.

That's why an FDM part isn't an isotropic solid but something closer to a stack of paper: hard to tear in the plane of the sheets, but easy to pull apart sheet by sheet. The between-layers plane is the weak plane, and it's always perpendicular to the Z axis. That makes orientation the single most important structural decision you make: it doesn't change the part's shape, but it decides which way the weak plane faces. A lever laid flat so the layers run along its length carries the load; the same lever printed upright splits along a layer line under the first load. Two articles go deeper here: Layer Adhesion and Anisotropy and Orientation and Overhangs.

Curves, Overhangs, and the First Layer: The Bead Eats Into Your Dimensions

On curved geometry, bead width translates directly into dimensional error, and always in the same direction: holes come out small and shafts come out big. The nozzle centers the bead on its path, so to make the wall of a hole it traces a circle smaller than the finished hole, and the inner half of the bead eats into the opening and narrows it. On a shaft it's the reverse: the path runs outside the finished outline and half a bead bulges outward. Add the polygonal approximation of the outline on small circles and the shrinkage on cooling, and the effect gets worse the smaller the hole. The net result closes the hole and fattens the shaft, which is why a fit drawn at zero clearance comes out tight: the parts end up interfering and jam together. Real Printed Clearances quantifies it.

Overhangs are the flip side of "nothing hangs in midair." As long as the wall rises near vertical, each bead rests almost entirely on the one below and you're fine. Past about 45° from vertical, each bead overhangs the previous one so far that it has nothing under it: it cools crooked, the underside comes out rough, or it fails outright. That threshold isn't arbitrary: it's geometric—the bead-to-bead overlap stops being enough—and it depends on the ratio of layer height to bead width, so with finer layers you can exceed 45° comfortably. For typical settings, 45° is the boundary, and it's the reason supports exist (Supports and Bridging). And right at the bottom, the first layer is printed deliberately squashed so it sticks to the bed; that squash, combined with excess bed heat and a nozzle sitting too close, leaves that telltale rim, the elephant's foot: the base widens by a few tenths of a millimeter and vertical holes pinch in right at their base. You fix it by lowering the bed temperature for the first few layers, adjusting the Z-offset, or turning on the slicer's elephant's-foot compensation.

You Design the Sequence, Not the Shape

Go back to the sentence at the top: a nozzle squeezing out beads, layer by layer, bottom up. Every property you care about starts there. Resolution is the size of the bead. Strength is the quality of the weld between beads, and orientation decides which way the weak plane faces. Precision is the bead width biting into your curves. Overhangs are beads reaching for support that isn't there.

So you're not drawing an ideal shape for a machine to reproduce: you're defining the sequence of beads and layers that will build it, with all their physical limits. Get that model right and the rest of the path is just detail. The first detail worth tackling is the one that ruins the most parts: which way to stack the layers. Continue with Orientation and Overhangs.