Print-in-place captive hinge: born articulated off the printer

You design a hinge so it comes off the bed already turning, with no pin to insert and no assembly step: you print the knuckles and the pin interleaved, separated by a clearance the printer can't close; peel the part off the bed and the joint already works. It's the same idea as a print-in-place pivot, but spread across several knuckles along one axis, which multiplies the number of interfaces that can fuse. The whole hinge rides on a narrow band of clearance: too little and the knuckles weld to the pin somewhere and lock the rotation; too much and the joint goes slack and the lid sags. This article is about where that band lies and how to orient the part to hold it all the way around.

The principle: a captive pin inside its knuckles

A classic knuckle hinge is a straight pin onto which the knuckles of the two leaves are threaded alternately: some belong to the left leaf, some to the right, and all share the same axis of rotation. The kinematics are those of a pure revolute pair — a single degree of freedom, rotation about that line. Printing it in place means there's nothing to thread: you model the pin and the knuckles already interleaved, in their final position, separated only by a gap the nozzle never fills.

That gap is what makes the pin captive. The printer lays down the knuckle material and the pin material in separate passes, and between adjacent beads that don't touch, there is no weld: you get two independent walls with air between them. When you peel the part off, the pin isn't bonded to anything, only geometrically trapped by the knuckles around it. It turns because it was never fused, not because you freed it from something. The design consequence is immediate: the clearance isn't a finish that improves the feel — it's the condition that decides whether you have a hinge or a solid block shaped like one. Tolerances for moving parts walks through the logic that governs any printed fit: the hole comes out narrow, the pin comes out fat, and the nominal gap on screen is not the real gap in the part. Here it applies several times in a row, once for each knuckle-pin interface.

Every interface can weld

The risk of a print-in-place part isn't spread evenly along the hinge: it concentrates at the worst point. Each interface between a knuckle and the pin is a potential fusion surface, and it only takes one spot on the contour too tight for the knuckle bead and the pin bead to touch, weld, and lock the joint. The whole hinge doesn't lock all at once; it locks at a bridge the size of one bead, and the result is the same: it won't turn.

That's why the clearance has to hold all the way around every knuckle, not on average. And this is where the way material is deposited comes in. On a knuckle lying down, the roof of the annular gap — the top face of the gap sitting just under the captive pin — bridges unsupported: the printer spans that stretch with no support, and a bead that hangs a few hundredths, or a layer that droops a little, closes the gap right there. The positions at three and nine o'clock, where the wall is nearly vertical and the bead rests on itself, are more forgiving. Layer height governs that roof more than any other setting: the thicker the layer, the more material hangs on each unsupported pass, and the more gap you need so the roof doesn't touch the pin. Over-extrusion is the common enemy: a flow calibrated on the high side fattens every bead, and that thickening is absorbed precisely by the gap. What stays invisible in a solid wall is, in a print-in-place clearance, the difference between turning and welding.

Pin horizontal, knuckles lying down

Print orientation matters more than any other decision. You have to print the axis of rotation horizontal, parallel to the bed, for two reasons that reinforce each other.

The first is to avoid the overhang at the top closure of the gap. If you stood the pin vertical, the knuckles would be stacked rings and the annular gap would have to close over the top with nothing under it: the printer would have to bridge a curved roof with no support along the full height, and that closure prints drooping, poorly welded, and fragile. With the pin horizontal, by contrast, each knuckle is a cylinder lying down: the unsupported overhang shrinks to the roof of the gap under the pin, a short stretch the layers resolve by bridging, and the rest of the contour rests layer on layer. The second reason is strength: a knuckle lying down has its layer lines crossing the hinge axis, not stacked along it, so the joint load doesn't pull layers apart. It's the same decision that governs any moving part — which way the interlayer weakness faces relative to the load — and it's reasoned out in full in Layer orientation for motion.

In exchange, the rotation comes out a little faceted. A cylinder printed lying down isn't a smooth cylinder: it's a stack of layers, and its surface is a polygon of fine steps. The pin turns inside the knuckles rubbing facet against facet, and at first you'll feel micro-adhesions, small spots where the first layer stuck slightly. It's not a defect to correct in the model. The micro-adhesions break with the first movement: what tears away are tiny spot welds and the burrs left from peeling, not material lost to wear. The faceting itself stays, but stops being noticeable once the two surfaces have worn against each other over the first few cycles.

The clearance: between welding and rattling

For most printers with a 0.4 mm nozzle, the gap between pin and knuckle is around 0.2–0.4 mm radially, and it's not a value you can inherit from a table: it depends on your machine, your material, and above all your layer height. The useful rule is that the gap shouldn't drop below the thickness of a couple of layers, because below that the roof bridged over open air droops and closes the gap before the hinge ever exists. At 0.12 mm layers you can push toward 0.2 mm; at 0.28–0.3 mm layers you need the high end of the range, or the roof will close the gap. You want to err toward loose rather than tight, because the two failures aren't symmetric.

If you come up short, you weld: one fusion point and the hinge is born locked, unrecoverable short of brute force, which almost always ends up breaking the part. If you go too far, you don't weld but radial play appears: the pin rattles inside the knuckles and the joint loses firmness. Don't confuse this with axial play — the pin shifting along its own line — which isn't caused by the radial clearance but by the side gap between the faces of adjacent knuckles. They're two distinct clearances and you calibrate them separately: the radial one decides whether it turns or welds; the lateral one, how much the lid wobbles side to side. To stop axial shift without tightening the rotation, the safest option is a pair of stop knuckles at the ends. A shoulder on the pin itself also stops it, but note: it creates a new knuckle-pin face in the axial plane that needs its own print-in-place clearance; skimp on it and it welds and locks just like any other interface.

When to use it and how it fails

The print-in-place captive hinge is the choice when you want the part to come off the bed already articulated and working instantly, with no separate pin to insert and no assembly step: lids, boxes with their own closure, housings that open, single-body parts that have to arrive assembled. Against a metal-pin hinge it's less precise and less durable, but in return there's nothing to lose, nothing to buy, and nothing to assemble. If your part is already plastic and the load is moderate, it's usually the cleanest solution.

It has three failure modes, all of them foreseeable from the design. The first is knuckle fusion, from short clearance, thick layers, or over-extrusion: the hinge is born welded and won't turn. If the first movement meets firm resistance, don't force it — that's the weld, and forcing it breaks the part rather than freeing it; only soft resistance that gives way is legitimate micro-adhesion. The second is excessive play, from large clearance: the pin rattles radially, the side gap lets the lid wobble, and on long hinges with a thin pin, it flexes between knuckles and worsens the droop. The third is roof failure of the gap: if that overhanging stretch printed badly — from standing the pin vertical, from excess speed, or from insufficient cooling — it becomes the fragile point, opening at the first load and releasing the pin. All three are avoided with the same decisions: clearance measured on your machine at your layer height, the pin horizontal, a pin of adequate diameter, and a flow calibrated without over-extrusion.

If what you need is exactly the opposite — for the pin not to turn, but to press in and stay fixed — the reasoning flips sign and takes you back to Tolerances for moving parts: the line between what slides and what grips is always the same handful of tenths — you just read it carefully, and in the right direction.