Knuckle hinge (piano and barrel): interleaved knuckles on a shared pin

A knuckle hinge doesn't share the load out of elegance — it does it because it has to. Take the moment that tries to tear a door off its frame: instead of letting it land on a single central joint that would concentrate everything at one point, you chop it up along a shared axis and spread it across a row of short cylinders that interlock like the fingers of two hands. That's the piano hinge taken to the extreme; the barrel hinge is its compact cousin. And as with almost everything in FDM, what decides whether it turns for years or seizes on the very first lid comes down to a few tenths of axial and radial clearance — and the orientation you printed it in.

Two combs, one axis

The kinematics are those of any revolute joint: a single degree of freedom, pure rotation about an axis. What's distinctive is how that axis is built. Each leaf contributes a comb of cylindrical knuckles, each a short hollow tube, and the two combs interleave: the gap left by a knuckle on leaf A is filled by one from leaf B, and vice versa. When you line them all up, their holes form a continuous, straight channel; you run a pin down that channel and now you have the physical axis of rotation. The two leaves share that one pin, so they can only rotate relative to each other about it.

Spreading the load is what makes this worth the trouble. A load applied far from the axis generates a moment that translates into a force on the knuckle carrying it. If you have a single long knuckle holding everything, that force concentrates and, worse, any misalignment pries against a narrow base. Chop that long knuckle into six short ones spread along the edge and each takes a fraction of the force, each with its own anchoring base to the leaf. The total length of contact between plastic and pin grows, the pressure per unit area drops, and the assembly resists bending all along the axis instead of at a single point. For a given hinge length, many short knuckles generally beat a few long ones.

There is a floor, though. A knuckle that's too short — below roughly 2–3 mm — has so few layers and perimeters that its base comes out fragile, and the axial clearance it needs from its neighbors eats up much of its usable width. It isn't an open-ended "more is better": there's an optimum between having many short knuckles and giving each one enough material to anchor well to the leaf.

Piano versus barrel: where the knuckles go

The two families are the same mechanism with knuckle density pushed to each extreme, and each solves a different problem.

The piano hinge runs knuckles along the entire edge, end to end. It's made for lids and long, flat doors — a chest lid, a fold-down panel, the cover of an elongated box — where the problem isn't only the weight but the buckling of the edge itself. A long lid supported only at its ends sags in the middle; the edge bows and the joint stops closing. The continuous row of knuckles stiffens that edge and keeps it straight, on top of spreading the load. It's the hinge on the lid of a piano; hence the name.

The barrel hinge does the opposite: it concentrates a few thick knuckles, often just two or three, each one solid and generous in diameter. You choose it when the load is local and concentrated, when you want a robust, discreet joint at a single point, or when the edge doesn't need stiffening along its whole length. Each knuckle is thicker, its wall around the pin is more solid, and it stands up better to a point load, at the cost of not covering the whole edge.

Print-in-place or assembled

Here FDM forces a basic choice, and both options are legitimate depending on the case.

The print-in-place option prints both combs already interleaved, as a single piece, with a designed-in clearance between the knuckles and the pin. That pin you either print in place too, or leave as a hollow channel, so that nothing prints fused and everything turns the moment you pull it off the bed. It's elegant: zero assembly, zero hardware. The price is that the rotation is plastic against plastic and the tolerance depends entirely on your calibration; too tight and it comes out stuck and won't turn, too loose and it ends up sloppy.

Within print-in-place, the trickiest case is printing the pin itself inside the channel. In that case the clearance is doubled — the entire surface of the pin against the entire hole — and it's the geometry most prone to fusing on the bed: if a single interface welds, the pin becomes one with the knuckle and the hinge won't turn. Leaving the channel hollow and inserting a separate shaft is considerably more reliable.

The assembled option prints the combs separately (or together but without a pin) and then runs a metal pin — a rod, a nail, a steel wire — down the channel. It almost always turns better and lasts longer: metal is smooth, hard, and doesn't wear against itself the way plastic would, and a printed hole rubbing against a steel shaft ages far better than two plastic surfaces chafing against each other. If the hinge is going to cycle a lot or carry a real load, this is the serious option.

The two clearances that decide the rotation

A knuckle hinge has two independent clearances, and it pays not to confuse them: they do different things.

The axial clearance is the gap along the axis between a knuckle and its neighbor: the space you leave between the faces of the interleaving cylinders. Its job is to keep the knuckles from rubbing sideways as they turn; without it, the faces chafe and the hinge seizes or squeaks. But overdoing it has its penalty: too much axial play leaves the hinge loose along the axis, the combs slide over one another, the leaves wander sideways, and the joint never closes properly. You want the minimum gap that guarantees they don't rub, no more. And be careful in print-in-place: separating two facing knuckle faces by less than about three or four tenths usually welds them or leaves them stuck after the first movement, so here the floor isn't yours to choose; your nozzle sets it.

The radial clearance is the gap between the knuckle hole and the pin, measured on the radius, and it's the one that defines the rotation itself. It's exactly the problem of any pivot: too tight and it won't turn, too open and it wobbles. Compute it per side, not per diameter, and size it using your measured sliding clearance — the one you get off a calibration coupon — and not from a table, because your material and your nozzle shift that figure. Tolerances for moving parts explains why the printed hole comes out narrow and the shaft comes out fat, which is exactly what you have to budget for here.

When it's the right hinge

The knuckle hinge is the answer when a single central joint isn't enough. If you have a lid or a door that has to spread the load along its entire edge — because it's long, because it's heavy, because it would buckle held only at the center, or because the joint has to close straight along its whole length — this is what you want. It's also the choice when you need continuous, wide rotation with no stops or travel limits. The angular range, though, isn't free: with ordinary geometry, the thickness of the leaves and the knuckle overlap limit travel well before 180 degrees. Reaching 180 degrees or more is possible, but it takes a geometry designed for it — thin leaves, recessed knuckles — and deliberately designing where the leaves bottom out as a stop.

Don't use it where a single pivot would do: the knuckle hinge adds knuckles, clearances to calibrate, and a pin to fit. Its complexity only pays off when load spreading or edge stiffness is genuinely the problem.

How a knuckle hinge fails

Three things break a knuckle hinge, and it's worth keeping them in mind before you print.

The first is a single knuckle failing from load concentration. If the load isn't spread evenly — because one knuckle is poorly anchored, because the load arrives off-center, or because you made a few long ones instead of many short ones — that knuckle takes a disproportionate fraction of the moment and gives way. It usually fails at its base, where it joins the leaf. That point concentrates the maximum bending stress and, in FDM, also tends to be the weakest because of layer orientation. The defense is to add more knuckles and spread them better.

The second is seizing from axial rubbing: if the knuckles touch sideways, from insufficient axial clearance, the hinge stops turning smoothly and starts forcing. It's the failure most often misread as "it needs more radial clearance" when in reality there's too much axial contact. Identify which of the two clearances is the culprit before you touch the wrong one. Thermal expansion can also close that gap and seize the hinge, but it only matters in hot environments — direct sun, near a motor; at normal use temperature, the expansion of PLA or PETG is negligible against clearances measured in tenths.

The third is slow, relentless wear: the plastic hole widens with use. Plastic turning against plastic, or plastic turning against a metal pin cycle after cycle, eventually opens up the hole, and the radial clearance you calibrated so carefully grows on its own until the hinge wobbles. A smooth metal pin already slows this wear dramatically compared with plastic against plastic. For it to genuinely stop aging, what you need is a bushing of a different material — brass, PTFE — or a bearing embedded in the hole: a bushing that's also printed in plastic solves nothing — it's still the same material pair wearing out. When a metal pin is enough, when a bushing is worth it, and when a bearing is the answer is the subject of Embedded hardware: magnets, bearings, and inserts.