Why do 3D prints always break along the layer lines?

Because a layer line is a weld, not solid plastic. Each new line of filament has to melt the surface of the layer beneath it enough for the polymer chains to tangle across the boundary, and it only has a second or two at temperature to do that. Everywhere else in the part the plastic is continuous. The result is that a printed part is meaningfully weaker in the Z direction than across the layers — often somewhere between half and three-quarters of the in-plane strength — so any crack starts at the weakest available interface and runs along it.

Does more infill make a print stronger?

Much less than people expect, and it is usually the wrong lever. Above roughly 40 percent infill the returns flatten out sharply, while perimeters keep paying. For a part loaded in bending, the outer walls carry almost all of the stress, so going from two perimeters to four does more than doubling infill and costs less time and filament. Spend the material on wall count, a wider nozzle and print orientation first. Raise infill afterwards, and only if the part is loaded in compression.

Can drying filament really stop parts snapping?

Yes, and it is the cause people rule out last. Absorbed water flashes to steam in the melt zone and leaves microscopic voids through the extruded line, so each layer has less solid material bonding to the one below and the part fails at a fraction of its rated strength. Nylon, PETG and polycarbonate reach that state in days in an ordinary room. Dry the spool at the manufacturer's stated temperature for four to six hours and reprint the identical file before changing any setting — if the part survives, storage was the whole problem.

Should I turn the part cooling fan off to get stronger layers?

For ABS, ASA and polycarbonate, largely yes — cooling is what stops those materials bonding, and they should print with the fan off or very low in a warm enclosure. For PLA, no: it needs full cooling to hold overhangs and bridges, and switching the fan off trades a strength gain for a dimensionally poor part. PETG sits in between and usually wants 30 to 50 percent. The general rule is that cooling improves geometry and harms layer bonds, so set it by material rather than by habit.

Why did my print survive the bench and snap in a hot car?

Because it was almost certainly PLA. PLA begins to soften around 55 to 60°C, and a car interior in sun passes that easily, so the part does not break so much as creep and then fail under a load it held all winter. This is a material choice, not a printing fault, and no setting fixes it. Move the part to PETG for a straightforward upgrade, ASA if it also sees sunlight, or polycarbonate where it must stay rigid at high temperature — and expect ASA and PC to need an enclosure.

When is it cheaper to have the part printed by a service?

When the part has to be trusted rather than merely finished. Chasing layer adhesion costs several test prints, often a dried spool and a new material, and a desktop machine in a cool room has a real ceiling in ABS, PC and nylon that no setting clears. If the part is structural, carries a person or a load, or is due to someone, a shop with a heated-chamber machine and a calibrated process is the shorter path. The directory at /providers lists providers by location and process, and one quote often costs less than the spools already broken.

Why 3D Prints Snap: Layer Adhesion, Heat & Filament

The bracket held for a month and then came apart in your hand, and the break is suspiciously tidy — a flat face across the part, exactly on one horizontal line, with no stretch or whitening around it. That is not a weak plastic. That is a failure of layer adhesion, and it means the two layers either side of that line never properly welded to each other. FDM parts have a grain, the same way timber does, and almost everything that snaps snaps along it. This guide reads the break first, works the fixes from free to expensive, and ends at the point where the honest answer is having the part made by a service.

This post contains affiliate links. If you purchase through these links, 3D Prototyping Hub may earn a small commission at no extra cost to you.

This post also contains Amazon affiliate links. As an Amazon Associate we earn from qualifying purchases.

A Layer Line Is a Weld, Not Plastic

The slicer draws a part as a stack of two-dimensional slices, and the printer builds it by laying a hot bead on top of a cooler one. For those two beads to become one solid, the new bead has to re-melt the surface of the old one deeply enough that polymer chains wander across the boundary and tangle. That process needs heat, and it needs time at heat, and it gets about a second of both.

Everywhere else in the part the plastic is genuinely continuous. Along the Z axis it is a stack of welds. That is why a printed part is reliably weaker in the build direction — commonly somewhere between half and three-quarters of its in-plane strength, depending on material and how hot the print ran. Every crack starts at the weakest interface it can find, and in an FDM part that is a layer boundary, every time.

So the useful question is never "why is this plastic so weak". It is "why did that particular weld not happen".

Read the Break First

Pick the broken part up and look at the fracture surface in good light. It narrows the cause before you change a single setting:

What the break looks like	What it usually means
Flat, clean, exactly along one layer line	Layer bonding — too cold, too much cooling, or too fast
Rough, running across layers, with white stressed plastic around it	The material itself is at its limit — the bond was fine
Crumbly, chalky, snaps with almost no force	Wet or degraded filament
Starts at a hole, corner or screw boss	A stress riser in the design, not a printing fault
Part sagged or distorted before it broke	Above the material's softening point — a material choice, not a defect
Splits between the perimeters and the infill	Under-extrusion or too little overlap, not adhesion

Two of those rows are not adhesion problems at all, which is why looking first matters. A break that runs across layers means the welds outperformed the plastic — the part needs a tougher material or more material, not a hotter nozzle. A break that starts at a hole is a geometry problem no printer setting will fix. And a print that skipped or gapped its way up the wall is under-extruding, a separate fault with its own fixes.

Fix It in This Order

Cheapest and most likely first. Most parts stop breaking before step five.

Reprint it rotated. If the load runs along the Z axis, turn the part so the layers run across the load instead. This is free and frequently doubles the strength that matters.
Raise the nozzle temperature by 10°C and reprint. Hotter plastic wets the layer below more deeply. Print a temperature tower and take the highest temperature that does not string or sag.
Set part cooling by material, not by habit. Full fan for PLA, 30–50 percent for PETG, off or near-off for ABS, ASA and PC.
Dry the filament if the spool has been open more than a couple of weeks.
Add perimeters — two to four — before you touch infill.
Slow the outer walls to around 30–40mm/s so each bead spends longer hot.
Enclose the machine, if you print engineering materials.
Change the material, once the machine is not the limit.

Heat Is the Bond

Almost every genuine adhesion failure is a heat failure somewhere. The three places to look are the nozzle, the fan and the room.

The nozzle. Manufacturers quote a range because the range is real: printing PETG at 230°C rather than 245°C is a visible strength difference in a bend test. Work up in 5–10°C steps until you see stringing or drooping overhangs, then step back once.

The fan. Cooling exists to freeze plastic before it sags, and freezing plastic is exactly what prevents a weld. For PLA that trade is worth it. For ABS, ASA and polycarbonate it is not — those materials want the fan off and the heat kept in.

The room. This is the one people miss. A layer that has cooled to 25°C before the next one lands has to be re-melted from cold; a layer sitting at 50°C in a warm chamber is already halfway there. That is the entire reason ABS parts that split up the side on an open printer come out solid in an enclosure. A fabric enclosure tent over an open-frame machine costs a fraction of a new printer and changes the outcome, and our round-up of printer enclosures covers what fits which frame. The same warm, still air is what stops the corners lifting, so if your parts are warping as well as breaking, one change fixes both.

If the hotend simply cannot reach the temperature the material needs — many stock hotends stop at 250 or 260°C, which rules out polycarbonate — then no setting will do it. Both the ELEGOO store's FDM machines and hotend spares and Anycubic's FDM range and enclosed models list maximum hotend temperature and chamber type by model, and Flashforge's enclosed printers ship the enclosure as part of the machine rather than as an accessory.

Wet Filament Breaks Parts

A spool that has sat open in a normal room absorbs enough water to matter, and nylon, polycarbonate and PETG get there in days rather than months. In the melt zone that water flashes to steam, foaming the extruded bead. A foamed bead has voids through it, so there is less solid plastic in contact with the layer below and less of it to weld — the part comes out looking almost normal and breaks at a fraction of its expected load.

The tell is audible: faint crackling or popping at the nozzle, sometimes with a rough, pitted surface. The test is cheap. Dry the spool at the manufacturer's temperature for four to six hours and reprint the same file before changing anything else. A heated dryer that feeds the printer while it runs is what keeps that fixed on a long print, and our comparison of filament dry boxes covers the practical options. Between prints, vacuum bags with fresh desiccant stop you drying the same spool every month.

Walls Beat Infill, and a Wider Nozzle Beats Both

For a part loaded in bending — which is most brackets, clips, mounts and handles — the outer perimeters carry nearly all of the stress and the infill contributes little. Going from two walls to four is a larger real gain than going from 20 percent infill to 60, and it costs less time and material.

The overlooked lever is nozzle diameter. A 0.6mm nozzle lays a wider, hotter bead, which carries more heat into the layer beneath and presents more welded area across every seam. Thicker layers help for the same reason: fewer interfaces, each one better fused. For functional parts, a 0.6mm or 0.8mm nozzle is one of the cheapest genuine improvements in Z strength available, and it prints faster. Keep the 0.4mm for detail work.

When the Material Is the Problem

If the break runs across the layers rather than along them, the welds were fine and the plastic was outmatched. That is a material decision:

Material	Fails by	Move to
PLA	Snapping without bending; softening above ~55°C	PETG for load and warmth, ASA outdoors
PETG	Creeping under sustained load; softening around 80°C	ASA or PC for heat, nylon for impact
ABS	Splitting between layers on an open printer	Same material in an enclosure, or ASA
Nylon	Absorbing water and going soft and dimensionally vague	The same nylon, dried and kept dry
Carbon-fibre blends	Snapping suddenly — stiffer, but less forgiving	Unfilled nylon or PETG where impact matters

That last row is worth stating plainly, because it is a common and expensive mistake: carbon-fibre-filled nylon is stiffer and holds its shape better, but filled materials generally trade elongation for stiffness. If your part is snapping under impact, adding carbon fibre can make it snap sooner — our guide to carbon-fibre filament sets out where the filler genuinely helps. For most people the real upgrade path is simpler: PETG for anything that has to take a load, ASA for sunlight and weather, and polycarbonate only when you have the hotend temperature, the enclosure and a dry spool to support it.

Design the Break Out

Some parts break in the same place every time because that place was designed to break.

Sharp internal corners concentrate stress — a fillet of even 1–2mm moves the failure elsewhere. Screws driven straight into printed plastic act as wedges and split the layers apart, which is why so many parts fail at a boss; heat-set brass inserts spread that load into the surrounding material and survive repeated assembly. And a thin section printed flat is strong, while the same section printed upright is a stack of small welds waiting to peel. Our design guidelines for 3D printing cover wall thickness, fillets and orientation in more detail.

When To Stop Tuning and Send the File Out

Layer adhesion is worth solving on a machine you will keep using, and most of the fixes above are free. It is not worth solving on a deadline, and it is not always solvable at all: a desktop printer in a cool room has a genuine ceiling in ABS, polycarbonate and nylon, and that ceiling is a property of the setup rather than a fault in it.

If the part is structural, carries load or a person, or is due to somebody on a date, send the file out. A shop running heated-chamber machines gets layer bonding as a matter of process rather than as a tuning project. Browse providers by location and process, send the same STL to two or three, and compare the quotes against another weekend and another broken bracket.

Hero photograph by Tom Claes on Unsplash.