Reusable jars and lids: designing the thread-and-cap pair

When you print a lid for a jam jar that already exists, you have nothing to decide: you measure the jar's thread and copy it with just enough slack. But when you design the jar and the lid from scratch — a kitchen canister, a hardware box, a bespoke container — you are the one inventing the thread. And that is where people make the same mistake every time: they copy a fine metric pitch, draw both parts to the same nominal size, and wonder why the lid won't go on, or goes on skewed and wrecks the first turn. The problem isn't printing; it's design. A thread you invent has four degrees of freedom — pitch, starts, flank clearance, and lead-in — and all four are settled before you slice. That's what this article is about.

The two parts are the same thread

The first idea to nail down is that here you aren't copying a standard: you're defining one. The jar carries the male thread (or the female, depending on the design) and the lid the opposite one, and the two are the same thread seen from two sides. They share the nominal diameter, the pitch, the number of starts, and the profile. If the jar climbs 3 mm per turn with two starts and trapezoidal flanks, so does the lid, no exceptions. The moment one dimension drifts between the two parts, they won't engage: no print tolerance will rescue a 2.5 mm pitch mated to a 3 mm one.

The only thing male and female differ in is the clearance. That's the point worth getting straight from the start, because it changes how you model. You don't draw two independent threads and pray they mate; you draw one thread geometry, use it as the male on one part and as the female on the other, and give the female the gap. This leaves you a single dimension to touch when the first print comes out tight, and it guarantees that profile, pitch, and starts are identical by construction. Modelling each part on its own is the fastest way to end up with two threads that won't mesh. How the helix geometry and the profile are built is covered in Modelling threads.

Coarse pitch and several starts

A common instinct is to copy the profile of a commercial glass jar: a perfectly decent pitch, around 3 or 4 mm per turn, but spread over a low, shallow thread with very little crest-to-root height. In FDM that shallow profile comes out blurred — not because of the pitch, which is more than coarse enough, but because there's almost no thread height for the nozzle to draw. As Modelling threads explains, with the axis vertical the helix climbs layer by layer, and each turn needs to stack enough layers to draw crest, flank, and root. And what sets that resolution is the pitch — the axial distance between two threads — not the lead per turn: at a 0.2 mm layer, a 2 mm pitch gives you ten layers per thread, plenty; a 0.8 mm pitch leaves you four, right at the limit. Since here you get to decide, go generous: a pitch of 2 to 4 mm, with a tall thread, prints clean, grips well, and forgives layer defects.

The tall profile has one drawback: with a coarse pitch, the lid takes many turns to travel down. You fix that with several starts. A two- or three-start thread wraps that many parallel helices around the same cylinder, so the lid advances two or three pitches per turn and closes in under a turn — sometimes half. Watch out for a common misunderstanding: even though with two starts the lid climbs two pitches per turn, the profile resolution is still set by the pitch of each individual thread, not by how far the lid travels; that's why coarse pitch matters even when you use several starts. And, above all, several starts centre sooner: with several lead-ins spread around the rim, the lid finds a thread at almost any angle you set it down at, instead of having to hunt for the single point of a one-start lead-in.

Flank clearance: neither seizing nor slop

The flank clearance is the gap you leave between the male flank and the female flank, measured perpendicular to the contact surface. It exists because the printer biases the two parts in opposite directions. The male comes out fattened, its bead bulging the contour outward; the female comes out pinched, its bead biting the gap inward. If you draw both to the same dimension, the flanks interfere before they can seat and the lid seizes on the first turn. It's the same bias that Real printed clearances explains, where you'll also see why it always runs in the tightening direction.

On fine printed hardware threads, 0.1–0.2 mm is enough — and 0.1 mm only if your machine is well calibrated. On a jar lid designed from scratch you can be a bit more generous — of the order of 0.3 mm, between 0.2 and 0.4 mm depending on your machine — because the diameter is large and a lid you close by hand doesn't need the fine fit of an M8. The honest way to split that gap is to leave the jar at its nominal size and open up the lid: that way you have a single part to tweak if the first print grabs, and the jar stays as the reference.

Careful not to overshoot the other way. Too much clearance and the lid wobbles: it loses flank contact, threads with play, and strips with barely any torque, because the plastic crests hardly mesh. Too little and it grabs, jams half a turn in, or has to be forced. The goal is that it threads with your fingers, no hard spot but no loose-nut slop either. A numerical detail that's often overlooked: work the clearance on the pitch diameter (the mean diameter where male and female touch), not by subtracting outer diameters. On an inclined profile, the perpendicular gap you're after and the diameter shift you apply are not the same figure.

The lead-in: chamfer it or it cross-threads

A thread that starts on a knife edge — the first thread beginning abruptly, at full height, with a live edge — is the most common cause of an otherwise correct lid cross-threading. When you set the lid down and turn, that live edge on the male catches on the one on the female, rides over it instead of dropping into the root, and you mesh a thread out of phase: the crossed thread. On the second or third go, the crossed crests strip and the lid stops gripping.

The fix is the lead-in chamfer: taper the start of both threads. On the male, thin the first quarter or half turn of the thread with a conical chamfer, so the thread starts from zero height and grows to the full profile. On the female, flare the entry with a small cone. With both parts chamfered, the lid self-centres as you set it down and the first flank slides until it finds its root before anything meshes. Combined with the multiple starts, this is what makes the lid catch the thread first time without you having to feel for the spot.

Orientation, profile, and seal

The two parts are printed with the thread axis vertical, perpendicular to the bed. It's the rule from Modelling threads and it doesn't change here: vertical, the helix climbs layer by layer, each turn rests on the one below and comes out rounder with better load-sharing between layers. The jar prints upright — mouth up or mouth down depending on where you want the best finish — and the lid mouth-down, with its female thread pointing upward during printing. Printed on its side, the part loses half the thread to overhang, goes out of round, and the fit you calculated suffers.

A caveat about the female: even vertical, the downward-facing flank of each internal thread sits cantilevered over the bore, with nothing beneath it. It doesn't come out clean simply because it's printed vertical. It comes out acceptable if the flank runs at a gentle ramp and the pitch is coarse — which is exactly why the trapezoidal profile and the generous pitch work so well here; with a near-horizontal flank or a fine pitch, that underside droops.

For the profile, steer clear of the sharp metric V. Model trapezoidal or buttress flanks: they spread more material per turn, the flat crests and roots come out of well-defined layers without the point the nozzle rounds off, and a stripped crest here and there doesn't kill the thread because the neighbouring turns still grip. The buttress profile — one near-straight flank that carries the load and one on a ramp — works especially well if the lid always tightens in the same direction; just orient the thread so its load flank, the near-straight one, faces down, resting like a floor on the previous turn. If you leave it facing up, that near-horizontal flank sits like a cantilevered ceiling and prints badly.

With these four decisions — coarse pitch, several starts, flank clearance put on the lid, and a chamfered lead-in — you have a jar-and-lid pair that threads first time and stands up to the rough handling it's going to get. Before you print the full pair, run off a test turn of each thread and tune them to the real figures of your machine following Real printed clearances; a test lid costs ten minutes and saves you reprinting the whole jar.