GoPro mount: ears, fingers, and screw

You clip the camera into a printed GoPro mount, tilt it to frame the shot, and hear a sharp crack: one of the ears comes away in your hand, split cleanly along a horizontal line. You didn't break it by being heavy-handed. You broke it because you printed the ears standing up, and the weight of the camera levering off the end of the arm pulls in exactly the direction where FDM is weakest. The interlocking-ear mount is one of the most-printed parts there is — adapters, arms, helmet brackets — and, for exactly this reason, one of the parts people most often buy ready-made rather than risk printing. With the right orientation and a couple of tenths of clearance in the right places, a printed ear is more than strong enough to hold a light camera. Get either wrong and it won't survive the first fitting.

That's what this article is about: what the interface actually measures, why the ear splits along a layer line and nowhere else, and how to give the fingers and the screw the clearance the process is going to eat.

The interface and its dimensions

The GoPro mount is not a published standard. GoPro has never released a dimensioned drawing: it's a de facto standard, measured by the community with callipers, which is why every figure here is a measured, rounded value rather than an official spec. If you need your part to mate with a specific accessory, measure that accessory; tolerances between clone manufacturers vary a fair bit.

The geometry is simple, and that's the beauty of it. The mount is two parts with fingers (the "ears") that interlock like the teeth of two combs, run through by a through-screw that clamps them together. One part has two fingers; the other, three. When they interlock, the two fingers of one part drop into the two gaps of the other, and the assembly acts as one the moment you tighten the thumb screw. The joint pivots around the axis of the screw until friction between the faces locks it at whatever angle you want.

The axle hole is around 5 mm and it is a clearance hole: it doesn't thread the screw; it just lets it pass through. The thread lives at the other end, in a nut or a blind thumb screw. The exact thread is the least documented part of the whole standard; replacement M5×0.8 thumb screws fit and work, but don't assume the original is exactly M5 without measuring it. If you're going to print the thumb screw yourself instead of buying hardware, Modelling threads explains how to generate the profile so it grips.

Why an ear splits along the layer

Here is the failure that ruins most printed ears, and it's purely a matter of orientation. FDM is strong within the layer — the beads pull well in tension along their length — and weak between layers: the bond between one layer and the one above is a diffusion weld that never reaches 100% of the material's strength, and it opens under the slightest perpendicular tension. An ear almost always breaks along that plane and nowhere else.

Think about the two loads it takes. The first is the screw clamp, which pushes the fingers against each other along the axis: that's pure compression, and compression between layers is precisely the strong direction of FDM, so however hard you tighten the thumb screw, that isn't what delaminates it. The second load is the one that actually breaks it: bending at the root of the arm. The camera and its lever hang off the end, and all of that moment concentrates where the ear grows out of the body. If that bending pulls perpendicular to the layers, it opens the part along a layer join like unpicking a seam.

That's why the rule is to print the ear lying down, with the arm and the plane of the fingers laid along the bed, so the layers run in the direction of the arm and the root bending falls with the layers, not across them. Printed that way, the tension from the bending is carried by the beads along their length, which is where the material is solid. Printed standing up — the arm pointing skyward — the layers sit crosswise at the root and the camera's first load peels them apart.

Two more things help the root. Fillet the base of each finger with a generous radius instead of leaving a sharp corner: an internal corner concentrates stress right at the most-loaded join — exactly where the crack starts. And push the perimeters to 4 or 5 with high or solid infill around the joint; on a small ear the infill barely weighs anything, and it's the walls that do the work. More wall in the right direction beats any infill trick.

Clearances for the fingers and the screw

With orientation sorted, two clearances are left to decide, and both go the same way: FDM tightens the fit, so you have to open it up on purpose. The hole comes out small and the fingers come out oversized, exactly the bias Real printed clearances sets out: holes shrink and features grow, always in the same direction.

The interlocking fingers need room per face to enter and, above all, to pivot once assembled. On paper, a 3 mm finger drops into a 3 mm gap, but printed, the finger grows half a bead per face and the gap narrows another half. What was a light rub on paper becomes a part that won't go in. Leave on the order of 0.2–0.3 mm per face between the finger face and the gap face. Reason it per face, not per diameter: 0.2 mm per face is 0.4 mm of total clearance across the interlocked assembly, and that's the difference between a joint that pivots smoothly and one you have to hammer in and then can't reorient.

The axle hole is the other place the bias bites. A nominal 5 mm clearance hole comes out under 5 mm printed, and then the screw either won't pass or only goes in by force, cracking the finger as it's driven through. Model it with generous clearance — 5.3–5.5 mm for a 5 mm screw — because it's a clearance hole and gains nothing from a tight fit: all it has to do is let the shank pass. And since you print the part lying down, that hole comes out horizontal and a little oval; at 5 mm diameter the oval is tolerable, but it's one more reason to err on the loose side rather than come up short.

Quarter-inch adapters

The other half of the craft with these mounts is crossing between ecosystems: going from GoPro fingers to the tripod thread everything else uses — ball heads, poles, articulated arms, photo mounts. You make that jump with an adapter that has GoPro fingers on one face and a male or female 1/4-inch thread — the universal camera-shoe thread — on the other.

Here it pays not to improvise the thread. The tripod thread isn't metric: it's 1/4"-20 UNC, a quarter-inch nominal diameter with twenty threads per inch, and its bigger sibling, the 3/8"-16, turns up on larger mounts. The exact dimensions, the thread-profile angle, and how to bring them to a printed part — printing the thread directly, leaving room for a captive nut, or fitting an insert — are covered in Tripod thread: 1/4"-20 and 3/8"-16. The underlying recommendation is the same as for any thread that holds something heavy: a directly printed 1/4" thread withstands little torque and wears out, so for an adapter that's going to carry a camera, capture a 1/4" nut in the model or fit a metal insert — Designing for heat-set inserts covers it — rather than trusting the plastic with everything.

Before you hit print

To recap what decides whether the part survives: lying down so the bending runs with the layers, filleted root so you don't concentrate stress at the weak join, plenty of wall around the joint, 0.2–0.3 mm per face on the fingers, and a generous axle hole so the screw passes without forcing. And above all, the right material: no PLA in a part that's going to spend hours in direct sun, clamped on top of your head or on a drone. None of these is optional in a mount that holds a camera.

The exact clearance number comes from your machine, not this page: measure it once with a tolerance tower in the material you'll use, as Real printed clearances explains, and reuse it across all your mounts until you change filament or nozzle. From there, you print GoPro ears that interlock first time and take the pull. And if you want to understand fully why layer direction matters so much in a part like this, Layer adhesion and anisotropy goes into it.