How the clip latch stays shut
26 August 2026
The rugged box on this site closes with a separate printed clip. No screws, no metal pins, no film hinges that crack after twenty cycles. The clip is a flat plate, 4.5 mm thick, that snaps over the seam between the lid and the base. Four small bumps hold it there. That sentence took me about five printed versions to earn, so here is what each part of it does.
Bumps on the clip, holes in the box
The first decision sounds backwards. The nubs sit on the clip, and the walls of the box get shallow dimples for them. Doing it the other way around, bumps on the box, would mean every box wall has little studs sticking out, catching on everything in a drawer. It would also ask the box wall to flex, and a box wall is the one part built to be stiff.
This way the flexing part is the clip. It is a flat plate with nothing else to do. When you press it on, the plate bows outward by about a millimeter, rides over the wall, and the nubs drop into their dimples with a click you can hear.
Why four bumps and not two
Version one had a single nub per row, centered. It clicked on fine. Then I picked the box up by the lid, the clip rotated around that one nub like a compass needle, and by the time I looked down it had walked halfway off the seam. A single contact point is a pivot.
So each row has two nubs, 5 mm wide, set out at 28 percent of the clip width from its center. On the default 18 mm clip that puts them about 5 mm to each side. Two contact points block the rotation, and the clip stays square.
The bottom pair is smaller on purpose
Both rows could have identical nubs. They do not. The top pair reaches 1.0 mm into the lid, the bottom pair only 0.6 mm into the base. The bottom is meant to lose.
When you flip the clip open, some pair has to let go first. If both pairs hold equally, which one gives up is decided by layer adhesion and luck, and the whole clip pops off in a random direction. On a workbench that is fine. Outdoors it means crawling in the grass looking for a yellow plastic rectangle. With the weaker pair at the bottom, the bottom always releases first and the clip swings up and hangs on the lid, still attached, until you press it back down.
Slopes that do not quite match
One detail I only understood after a print jammed: the nub climbs at a 45 degree slope, but the dimple it lands in is cut at a gentler 0.8 ramp. With both at 45 the two surfaces would mate along their entire length with zero margin, and a single layer of over-extrusion is enough to turn a snap fit into a press fit. The mismatched slopes touch along a line instead of a plane, and the 0.2 mm of clearance per side stays real.
The clip also never touches the wall face itself. It floats 0.3 mm off it, so only the nubs make contact, and the dimples are cut for what actually arrives rather than for the nominal nub. A 1.0 mm nub crossing a 0.3 mm gap presents 0.7 mm to the wall. That, plus clearance, is the depth the dimple gets.
The fingernail flare
The last 3 mm of the clip flare outward. A ledge for a fingernail. That is the entire opening mechanism.
Before the flare existed I opened test boxes with a flat screwdriver, which works right up until it slips and gouges a groove across the print you spent nine hours on. A shape that needs a tool to open is a shape that gets pried open badly.
There is a second latch style in the generator, SnapLock, that uses spherical studs instead of long nubs. There the rule is different again: a stud must be a slice of a sphere, because a ridge extruded through the clip offers no resistance along its own length and slides right off. But that one is a story for another post, together with the hinge.