Printable threads: buttress and trapezoidal profiles

You take a lid that threads beautifully onto a shop-bought jar, measure it, model it faithfully to the V-shaped metric profile, and print it. It won't thread. Either it goes in one turn and jams, or the crests come out furry and fragile, and strip on the third tightening. The fault isn't in your dimensions or your clearance: it's in the profile. The metric V is designed to be cut in metal with a tap, and the very features that make it excellent in steel — the sharp crest, the pointed root — are exactly what a 0.4 mm nozzle can't lay down. For printed threads there are far better profiles, and they were invented a century ago: the buttress thread and the trapezoidal thread. This article explains why they win and when to use each.

Why the sharp V fails in FDM

The ISO metric profile is a symmetric V with a 60° included angle, 30° per flank. On metal it's ideal: the tip cuts, the straight flanks share the load in both directions, and the tap leaves a clean root. Shrunk to the size an FDM machine prints, that same profile fails on three fronts at once.

The first is the crest. The nozzle doesn't draw an edge: it lays down beads of bounded width, with some die swell at the exit, and as layers stack the pointed tip rounds off. The slicer can thin a line below the nozzle diameter — down to half of it — but even then it won't reproduce a sharp vertex; the sum of the minimum bead, the layer stepping and the extrudate swell leaves the crest blunt and starved of material.

The second is the root. With the axis vertical — the only sane orientation, as Modelling threads explains — each layer is a nearly circular ring, and the pointed bottom depends on how the ring radius changes from one layer to the next. The layer stepping and the rounded bottom of the bead never close that vertex: you get a smeared root instead of a defined one. It isn't that the head steers badly within a layer; it's that a vertex between layers never forms.

The third is the overhang. With the axis vertical, the lower flank of each turn hangs over the root of the previous turn. On a 60° V that overhang is still tolerable — the descending flank sits at about 60° from vertical — but the cross-section of each turn is so small that any imperfection wipes out half a thread.

The result is a thread that's almost all void: little material per turn, a fragile crest and a dirty root. At a coarse pitch — 1.5 or 2 mm — the V survives because each turn is large and forgiving. At a fine pitch it smears into a rippled cylinder that threads into nothing. So, unless you're working at a clearly coarse pitch, the sharp V is the profile to avoid, not the one to start from.

Buttress: sawtooth for load in one direction

The buttress thread — the sawtooth — trades symmetry for asymmetry. It has two distinct flanks: a load flank, almost perpendicular to the axis, and a trailing flank that slopes away in a gentle ramp. The load flank takes all the thrust when you tighten the lid; the trailing flank bears hardly any force — it just closes the profile.

That asymmetry is a gift for FDM, but it's worth seeing exactly why. With the axis vertical, "perpendicular to the axis" means almost horizontal, not vertical. You orient the part so that the ramped trailing flank faces down: it becomes a gentle, self-supporting overhang that comes out without drooping. The load flank, almost horizontal, faces up, where it prints well — not because it stacks like a wall, but because it is effectively a flat top surface, a roof the nozzle covers without trouble. Unlike the V, which hangs on both sides, no face is left as a severe overhang.

On top of that, because the buttress packs more material per turn than the V, a badly formed crest here or there doesn't ruin the thread: the neighbouring turns keep gripping.

The price of asymmetry is that the buttress only holds well in one direction: the load flank carries, the trailing flank doesn't. A lid doesn't care about that — it always tightens the same way — so the buttress is the ideal profile for lids, caps and collars you turn by hand in a single direction.

Trapezoidal and ACME: symmetric flanks that share the load

When the thread has to work in both directions — a lead screw that goes up and down, a large lid you might force either way — the buttress's asymmetry is no longer any use. That's where the trapezoidal thread comes in: symmetric ramped flanks, flat crest and root. It's the trapezoid shape that gives it its name, and it's the robust general-purpose profile.

The metric version (Tr threads, ISO 2904) uses a 30° included angle, that is, 15° per flank. The American ACME uses 29°, 14.5° per flank: practically the same.

Its advantage in FDM isn't the overhang, and that's worth saying plainly. The flank angle is measured from the plane perpendicular to the axis, so 15° per flank means the descending flank sits only 15° above horizontal: it's a flatter overhang than the 60° V's, not a gentler one. For every 0.2 mm layer, that flank advances close to 0.75 mm in radius, against the V's 0.35 mm. What saves the trapezoidal is something else. The flat crest and root come from well-defined stacked rings, without the sharp tip that the nozzle can't resolve. And the generous cross-section per turn makes it robust and fault-tolerant, just like the buttress.

So don't read the trapezoidal as "the profile that escapes the overhang". Its descending flank, at 15° from horizontal, hangs over the previous root almost as much as a square profile, whose perpendicular flank leaves the underside fully horizontal: there's only 15° between the two. The real difference is made by the flat crest and root and the amount of material, not by a much kinder overhang. With either of these profiles, vertical orientation and flank clearance are not optional.

Profile, clearance and axis orientation

The profile alone isn't enough: a male thread and a female thread modelled to the same nominal profile won't thread in FDM, however buttress or trapezoidal they are. The male comes out fattened because the outer bead pushes the contour outward. The female comes out narrow for several reasons: the inner perimeter sits slightly inward on concave curves and the contour is approximated by a polygon, and, secondarily, the part shrinks as it cools. Whatever the cause, the flanks bind before they can seat.

That's why, whichever profile you choose, you have to give flank clearance. Start with 0.3–0.5 mm of total gap measured on the pitch diameter, perpendicular to the contact face, and come down from there if the thread wobbles. That gap makes room for the fattened male and the narrowed female, and leaves margin for the roughness of the lower flank. A start of 0.1 mm — 0.05 mm per flank — falls below the normal position-and-flow error of an FDM machine: with a fattened male and a narrow female, it will almost certainly seize.

The sensible split is the usual one: leave the male at its nominal dimension and open up the female. That way you have a single dimension to adjust if it threads too tight, and one intact reference part. The full physical explanation of that clearance — why the female comes out narrow and the male fattened — is in Real printed clearances.

Once you've chosen profile and clearance, don't throw it away at the orientation step. Always print the cylinder with the thread axis vertical, perpendicular to the bed. That way the helix climbs layer by layer and each turn rests on the one below, with the crest and root coming from well-defined stacked rings. Lay the cylinder down and the helix becomes a succession of overhangs and bridges: half the thread hangs in the air, droops, and the fit you measured is gone. No profile saves a thread printed on its side.

Mind, too, the two places where printing ruins the thread even when the profile is right. The first is the bed: if the thread starts flush against it, the first layers are squashed by elephant's foot and deform that opening turn, right where the other part goes in. Chamfer the thread entry or lift it a few millimetres off the plate on a smooth skirt, and the first good turn ends up clear of the squashed layers. The second is the Z seam: vertical printing leaves a small layer-start blob running up a flank for the whole height, and one extra bead there can stop it threading. Put the seam on the outer diameter of the male or the inner diameter of the female — away from the contact flank — or scatter the seam so it doesn't form a continuous ridge.

Choosing the profile by load direction

The rule boils down to the direction of the load. If the thread always tightens the same way — a lid, a cap, a collar you turn by hand — use buttress: its load flank holds well and its trailing flank prints with no severe overhang. If the thread works in both directions or moves something — a lead screw, a large lid you might force either way, a ring that goes up and down — use trapezoidal or ACME, which share the load symmetrically and come out just as defined. Save the sharp V for when the pitch is clearly coarse and you want to copy an existing metric thread; at a fine pitch, the bed won't forgive it.

All three are gentle-load threads for hand use: for what a plastic thread really withstands, and when to put metal in its place, go back to Modelling threads. Once you've settled on a profile and want to take it to a lid that seals well against its jar, carry on with Reusable jars and lids: designing the thread-and-cap pair.