BRIO-compatible wooden train track

Your child loses a curved piece, or you want a crossing no brand ever made: the obvious fix is to print one. The hard part isn't the shape of the track — you can draw that in an afternoon — it's the two centimetres at each end where one piece grips the next. That's where a part printed to the dimensions you read off the callipers goes in only with a mallet, or won't go in at all, or goes in and falls straight back out. Wooden track looks like toy geometry, but the connector is a genuine friction fit, and FDM skews it in exactly the wrong direction.

What you're really copying

What we call "wooden train track" is a de facto standard: nobody published a master drawing, but BRIO popularised it enough that dozens of brands — IKEA, Hape, Lidl, the no-name pieces from the bargain bin — make parts that fit together. That compatibility doesn't come from a written standard; it comes from everyone copying the same physical part until the tolerances converged.

The track is a beech board of constant cross-section with two lengthwise grooves in its top face: the channels the wheel flanges run in. The grooves sit a fixed distance apart, and that spacing is what keeps the train on the track through the curves. The underside is flat, so it sits level on the floor or on the piers of a bridge.

What makes the track a system rather than a loose board is the ends. Each piece finishes in a male connector at one end and a female socket at the other. The male is a round head joined to the body of the track by a narrower neck; the female is the matching recess, with a wide mouth and a narrow throat. You push the head of one piece into the mouth of the next and the two lock together. The whole standard is decided at that joint.

The dimensions you can measure, not the ones you invent

Be honest with yourself here: there is no published standard with official dimensions. BRIO doesn't hand out a dimensioned drawing, and the compatible brands work by reverse engineering, exactly as you're about to. Any table you see with exact tenths of a millimetre is someone who measured their own set, not a standard. So what follows are values measured on real parts, with all the spread that implies: wood is made to generous tolerances, and two pieces from different brands can differ by several tenths and still fit, because the joint is designed to tolerate slack.

Treat that table as a starting point, not a drawing. Measure your set before you print a batch, because the piece your part has to mate with is the one you have at home, not the industry average.

Why the nominal connector won't go in

This is where the joint collides with everything you already know about printed clearances. The wooden connector is a friction fit, but it pays to separate two measurements that don't do the same job. One is the pocket the head slides into and seats against — the head against the back of the mouth; the other is the throat, the narrowing the head passes through and is then held behind. The only thing keeping two pieces joined is that the head is wider than that throat.

FDM's bias always runs the same way: holes come out small and bosses come out large. Your printed female mouth ends up narrower than you drew it, and the male head comes out thicker. If you copy the nominal calliper dimensions to both sides, you stack the two errors in the same joint and the result is a male that won't enter the female, or that is forced in and splits the mouth open. It's the same mechanism Real printed clearances explains: zero on screen is already interference in the part.

The fix is the usual one: reason per side and open a gap on purpose, but not the same gap in both places. In the sliding pocket — the head inside the mouth — you want the play Choosing the fit: clearance, transition, interference calls "locating, hand-separable": in PLA, that's on the order of 0.15–0.25 mm per side between the head and the wall of the mouth. Subtract it from the head radius, or add it to the mouth radius, but not both at once, or you'll end up with twice the play and a piece that rattles.

The throat is another matter entirely. There you don't want clearance: you want retention. If you open the throat until the head passes loosely, you kill the one thing holding the pieces together and the joint comes apart the moment you push the train. The throat has to stay narrower than the head. In wood, the fibre flexes to let it through and then hugs it; in rigid PLA, which barely flexes, that margin is more delicate, and one tenth too much in the throat turns "holds when you pull" into "lets go at the first shove". It isn't an elastic snap fit: it's retention by geometry, and you decide between the two with a single tenth of a millimetre.

Print the track flat

Orientation on the bed decides two things at once: the quality of the grooves and the behaviour of the connector. Print the track lying flat, groove face up. That way the wheel channels form as horizontal pockets, with clean vertical walls and a flat floor coming straight off the layers — no overhangs to collapse and no support to scrape out of the inside of the groove.

That same orientation puts the connector with its axis in the XY plane, which is what you want for the head and the throat: they print as a plan-view contour, where the slicer controls wall width well and you can predict the bias. Print the track on edge instead and the head comes out as a hole and a boss lying horizontal, ovalised, sagging on top, with an effective clearance that has nothing to do with the one you measured.

But lying the track flat has a cost you have to face head-on. The connector lives at half the 12 mm height, because that's where the mouth of the wooden piece it has to mate with sits. Printed flat, that leaves two critical surfaces in mid-air: the underside of the male head, which juts out from the body with nothing but air beneath it, prints as a horizontal overhang; and the roof of the female socket, those three millimetres of material over the cavity, prints as a bridge. Both land exactly where the fit is defined, and a delamination in the bridged roof is the most likely crack if you force the joint — far more so than the side wall.

Elephant's foot comes into play here too, and on both sides of the joint. The first layer's spread thickens everything that touches the bed at its base: the underside of the male head ends up a hair wider than its top, and the floor of the female socket narrows at the bottom and squeezes its mouth. If the connector won't go in, start by shaving there or turning on the slicer's elephant's-foot compensation, and correct it in one place only so you don't subtract it twice.

Two more contour details. Round the connector head in plan: the wooden head is a nearly complete circle precisely so it goes in at whatever angle a child brings the two pieces together. Draw it with sharp corners "to grip more" and it will only enter in one orientation and catch in the rest. The curve isn't an aesthetic whim: it's what makes the joint tolerant of clumsy assembly. And place the seam (z-seam) outside the contact zone, on the back of the head that never touches the mouth; the little bump where the perimeter starts, if it lands on the sliding contour, catches and leaves you with a connector that "won't go in at one angle" for no reason you can see.

With a flat profile, the sliding pocket opened a tenth or two per side, the throat narrower than the head, and everything calibrated against your own wood, your printed pieces will fit the ones in the set with no difference you can feel. If this is the first time you're turning a nominal dimension into a real gap, read Real printed clearances first to fix your per-side number in PLA, then come back here with the connector head already sized.