Designing print-in-place hinges: clearance, orientation, and test coupons

Three mechanisms that all get called a hinge

Before you open your CAD tool, decide which hinge you actually mean. The BOSL2 documentation separates several objects inside its hinges.scad file. The knuckle_hinge module builds a two-part pivot hinge with a hole for a separately supplied pin, by default a piece of 1.75 mm filament, or a screw if you pass a specification such as M3. That is the classic hinge with a pin inserted after printing.

The same module accepts an in_place option. The documentation says this mode creates interlocking cones instead of a pin hole. It remains a pivot hinge, but whether the segments separate after printing depends on settings and execution; BOSL2 says to adjust gap to obtain separation. By contrast, living_hinge_mask creates a thin web that bends. The two mechanisms do not share the same geometry.

This page covers the second case, the print-in-place pivot. The clearance and orientation rules below do not transfer directly to a living hinge, which follows a web-thickness logic. If your part needs a thin strip to fold many times, you are on a different topic. If two leaves must rotate around an axis with no assembly step, keep reading.

How interlocking cones replace the pin

The principle documented by BOSL2 is straightforward. Each knuckle segment ends in a cone. Cones on the inner segments nest into cones on the outer segments with a void between them. After printing, that nesting retains the hinge while allowing rotation. In this mode, pin_diam no longer describes a hole but the diameter of the cone bases. It defaults to one unit less than the knuckle diameter, which itself defaults to 4. The documentation notes this default suits larger hinges and that a small hinge may need a larger pin_diam.

Cone angle is the second lever. The default cones are 45 degrees, and you can pass an angle to in_place, measured from vertical. According to BOSL2, a larger angle makes a pointier cone that is harder to print and leaves a smaller clearance gap inside for the same gap value. In other words, the cone angle affects both printability, because it is the overhang angle, and how freely the hinge moves.

The third lever is gap, the space between segments, which defaults to 0.2. The documentation tells you to adjust it until you get something that separates after printing. These three values define an example hinge geometry. Nothing in the library documentation promises that a default value will print successfully on your machine, with your filament, at your layer height.

Clearance is a starting point, never a universal rule

The Prusa knowledge base is explicit: there is no single universal tolerance value. It lists the most common factors that drive fit: model size, horizontal or vertical orientation, the geometry of the interlocking parts, calibration, settings, and material properties. It adds that different materials can warp or shrink during printing, so the same geometry can fit differently in a different filament.

In the context of its own FFF printers, Prusa states that an Original Prusa is accurate to at least 0.2 mm and gives at least 0.3 mm as an initial good measurement for movable parts such as a hinge. That figure belongs to that context. It is not a rule for every filament, nozzle, or machine. On a printer you have not characterized yet, check the dimensional accuracy of your prints before you lock in a clearance.

Two more factors deserve attention. For the clearance parameter that shifts the rotation axis, BOSL2 says it should probably equal one or two layer heights. Layer height therefore changes that starting reference; see the layer height page. Prusa says elephant-foot compensation can improve tolerances in some areas. Inspect the coupon’s first layer before changing the model.

Print a clearance coupon before the final part

Prusa puts it plainly: you will have to make test prints before you can consider a design final, and you should expect to tweak tolerances and geometry. For a hinge, that coupon is a short version of the mechanism with the same knuckle diameter, the same cone angle, the same material, and the same orientation as the final part. Changing any of those between the coupon and the part weakens the test, since Prusa lists each one among the factors that determine fit.

A parametric approach makes this loop fast. Prusa recommends working parametrically and saving versions before merging objects or cutting cavities. With BOSL2, gap, pin_diam, and in_place are already exposed as parameters, and the documentation examples declare their values at the top of the file. You can print two or three coupons side by side with different gap values. In conventional CAD the idea is the same with named variables, for instance in FreeCAD.

Read the result as an observation limited to that coupon. If it separates by hand and turns, record the settings and geometry as a candidate to carry into a final part under the same conditions; BOSL2 does not guarantee that transfer. If it remains joined, the documentation says to adjust gap to obtain separation. Binding or excessive play is a reason to review one parameter at a time and reprint, not evidence of a service life or a single cause.

Orientation, overhangs, and wall thickness

Orientation determines the direction of cone overhangs relative to the bed. The BOSL2 examples place the rotation axis parallel to the bed with both leaves lying flat on either side, and rotate the design geometry, for example with xrot(90), to move from the modeling view into print orientation. One example states that the opening angle must be set to zero for printing.

The default 45-degree value describes BOSL2 cones; it does not alone determine overhangs on a model placed differently. BOSL2 warns that a larger cone angle produces a pointier shape that is harder to print. Prusa separately gives overhang references of 45 to 60 degrees depending on nozzle and settings: do not infer a hinge-printing guarantee without checking orientation and the angle convention. For a thin plate, BOSL2 says clear_top removes arm material above the mounting surface.

Also consider the walls around the knuckle. Prusa reminds you that a wall thinner than one perimeter is not printable, and that in its MK3S reference with a 0.4 mm nozzle, one perimeter in PrusaSlicer is about 0.45 mm and two are about 0.9 mm, as approximations that depend on slicer settings. A small knuckle with internal cones leaves little material, so preview the sliced segments to confirm they are solid. The page on 3D printer nozzles explains how nozzle size relates to fine detail.

Freeing, cleaning, and adjusting after the print

The BOSL2 function says to adjust gap to get a hinge that separates after printing; it does not provide a universal procedure for freeing a joint that remains fused. If your coupon does not pivot, avoid forcing it or levering it apart: inspect the segment clearance and first layer, then adjust the model or slicer before another trial. That observation alone does not identify a cause.

Check the first layer and elephant-foot compensation as two items to inspect: Prusa says this compensation can improve tolerances in some areas, and the elephant-foot page explains the defect. If a hinge separates but drags, that may indicate clearance or geometry to revisit. BOSL2 says a larger cone angle reduces internal clearance at the same gap; compare settings on a coupon rather than assume one correction will work.

Finally, do not generalize from one material to another. Prusa points out that filaments meant for outdoor or mechanically demanding use, such as PC, ASA, or PETG, do not reproduce small details as well as PLA. A hinge validated in PLA is therefore not automatically validated in PETG, so rerun the coupon. Keep the project with its parameters rather than only a mesh, as explained in the STL vs 3MF guide, so you can replay the adjustment loop later.

Sources and limits

• BOSL2 — hinges.scad, wiki edited October 2, 2026
• Prusa — Modeling with 3D printing in mind

We organize manufacturer and community documentation; we have not measured these products ourselves. The manual for your exact model takes precedence over general guidance.