An injection-moulded part is one continuous piece of polymer. A printed part is a stack of welds, and the strength of each weld was decided in the fraction of a second while the new bead was still hot enough to fuse with the one under it.
That is why a print can look flawless and still snap in the hand along a line you can see. Nothing went wrong visually; the joint was simply never made properly.
Start with when it broke
- It split during the print, usually partway up a tall part, with the crack running along a layer line. The part was pulling itself apart faster than the welds could hold.
- It survived printing and snapped later under load, cleanly, along one layer. The welds held their own weight and nothing more.
- It came off the plate fine and cracked overnight with no load at all. That is stored stress relaxing, and it is a different mechanism from the two above.
- The filament broke before it ever printed, between the spool and the extruder. That is not a printing fault at all — the polymer degraded on the shelf.
Those four chronologies map onto the four pages in this category, and telling them apart takes seconds once you know to ask.
What actually makes the weld
Two thermoplastic surfaces fuse by diffusion: chains from each side wander over the boundary and tangle up with their neighbours. Diffusion is fast when the polymer is hot and effectively stops once it is not, so the strength of any joint is decided by how long the interface stayed warm. Nothing reheats it afterwards. That single fact is why a cold room, a fan aimed at a small part, and a slender cross-section all show up as weakness rather than as a visible defect.
This is also why tall, thin parts are so much weaker than their volume suggests. A layer that takes four seconds to draw gives the previous one four seconds to cool.
Orientation beats every setting
The strongest lever here is not on the slicer's temperature page. A printed part is markedly stronger along the beads than across the welds between them, so a bracket rotated so the load runs within layers rather than across them can survive many times the force with no change to any profile value. When a functional part keeps failing, look at how it is standing on the plate before looking at anything else.
Degradation is a separate problem with the same appearance
Some materials lose strength before they are printed at all. Absorbed moisture attacks the backbone of polyamides and polyesters while they are molten; daylight and slow oxidation do comparable damage over months on an open shelf. The outcome either way is a shorter average chain length, and short chains make poor welds however carefully the print is run.
The practical difference is recoverability. Moisture is reversible — dry the spool and the strength comes back. Chain scission from ageing is not, and a spool that stays brittle after a full drying cycle is finished. That test is worth running before spending money replacing a printer that is working correctly.
What raises strength, in the order worth doing it
- Reorient the part so the load does not cross layer boundaries.
- Check the spool's moisture history first if the material is hygroscopic — every polyamide, copolyester and polycarbonate on this site is.
- Work upward through the material's temperature window, breaking a test piece at each step.
- Cut part cooling on the materials that do not want it, and keep the chamber warm.
- Add walls rather than infill. Perimeters carry load; interior lattice mostly does not.
Infill density is where most people start and it is the weakest of these levers. Past roughly a third, adding more contributes far less strength than one extra perimeter, while adding print time and material cost to every copy you make.