Why would a 3D print crack when nothing is touching it?

Because the load is already inside the part. Plastic shrinks as it cools, and in a printed part the bottom layers cool and set while the upper ones are still hot and contracting. That pulls the finished part into tension against itself, and the stress stays locked in after it comes off the plate. Plastics also creep — they keep deforming slowly under a constant stress — so a part can sit at a stress level it survived on day one and fail at the same stress on day four. Nothing has to touch it. The crack almost always starts at a stress concentration: a sharp internal corner, a hole, or the boundary between a thick section and a thin one.

Which filaments crack on their own most often?

The high-shrinkage engineering materials, in roughly this order: ABS, ASA, polycarbonate and nylon, with PETG and PLA far behind. ABS shrinks on the order of 0.5-0.8 percent as it cools against PLA's 0.2-0.3, and all of that difference becomes internal stress when the layers cool at different rates on an open machine. This is why the same model prints fine in PLA and splits in ABS on the same printer, and why an enclosure is treated as a requirement for ABS rather than an accessory. If a part must be made on an open-frame machine and must not crack, PETG is the lowest-stress material that still has useful heat and chemical resistance.

Does an enclosure actually stop delayed cracking?

It addresses the cause rather than the symptom, which no setting change does. The mechanism is differential cooling, so anything that slows cooling and evens out the temperature across the part reduces the stress stored in it. A passive tent helps substantially on ABS and ASA by trapping bed heat and stopping draughts. An actively heated chamber goes further and is the specification professional machines sell for high-shrinkage materials. What an enclosure does not fix is a part whose geometry concentrates whatever stress remains: sharp internal corners, abrupt thickness changes and unfilleted bosses still crack in a chamber, just later.

Why do my parts crack around the screws a few days after assembly?

Because a screw driven straight into plastic puts a permanent hoop stress on the boss around it, and plastic relaxes under that stress rather than holding it. As the material creeps, the crack grows from the root of the thread outward until the boss splits, typically along a layer line. It feels like the part got weaker, but the stress was applied at assembly and the clock started then. The fixes are mechanical: a heat-set brass insert gives a metal thread and spreads the clamp load, a thread-forming screw designed for plastics displaces material instead of cutting it, and a washer under the head stops a conical screw head acting as a wedge.

Can superglue make a printed part crack later?

Yes, in two distinct ways, and both are delayed. Cyanoacrylate cures to a hard, brittle glass line, so a joint that stays under load has no give in it and lets go at the first thermal cycle or knock. Separately, CA vapour and some solvent cleaners craze stressed thermoplastic surfaces, leaving fine cracks that grow into a real fracture under a stress the part was previously carrying happily. If the joint is structural or stays loaded, a two-part epoxy is the better chemistry, and the part should be stress-relieved or fitted without preload before anything is bonded.

When is it worth having the part made by a service instead?

When the part has to hold a load for months, live in an environment that cycles, or be made as a batch of identical units that all have to behave the same way. Desktop FDM builds internal stress by the nature of the process, and the fixes available to you — enclosure, geometry, material — reduce it rather than remove it. Processes like SLS nylon, MJF and industrial FDM in a heated chamber produce parts with far lower residual stress and much better repeatability across a batch, and injection moulding removes the question entirely. Providers are listed by location and process in the directory at /providers.

All articles

Buyer GuideTroubleshootingFDMMaterialsEquipment

Why 3D Prints Crack Days After Printing

3D Prototyping Hub·November 8, 2026

The bracket came off the plate clean. You checked it, liked it, put it in a drawer. On Thursday you picked it up and there was a split running from one of the holes, straight along a layer line, with nothing to explain it. Nobody dropped it, nothing was bolted to it, and the printer was behaving. When a 3D print cracks days after printing, the useful conclusion is immediate: the load was already inside the part, and it has been working on it since the moment the print finished.

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 I earn from qualifying purchases.

Read the Crack First

Delayed cracks are not random, and where the part split narrows the cause down faster than any test print will.

Where and how it cracked	Most likely cause
From a corner or hole, across a large flat part, in ABS/ASA/PC	Residual stress from uneven cooling
Around a screw boss, days after assembly	Preload plus creep
At a bonded or press-fitted joint	Brittle adhesive, or crazing from solvent and glue fumes
Cleanly along one layer line, part unloaded	A weld that was never fully made — usually wet filament
All over, in a part that lives in a garage or a car	Thermal cycling against a stress that was already there
A resin part that was fine, then went glassy and split	Continued cure and UV embrittlement

Each of those needs a different fix, and only the first two are about the printer.

Cause 1: The Part Was Printed Into Tension

This is the big one, and it explains most self-cracking in ABS, ASA, polycarbonate and nylon.

Plastic shrinks as it cools. In a printed part the bottom layers reach room temperature and stop moving while the layers above them are still hot and still contracting, so the upper material pulls against a base that will not follow. The part ends up in equilibrium with itself — one region in tension, another in compression — and that balance is locked in when it comes off the plate. It looks finished because nothing has moved.

The reason the failure is delayed rather than immediate is creep. A thermoplastic under constant stress keeps deforming slowly, and the stress redistributes toward whatever feature concentrates it most: a sharp internal corner, the edge of a hole, the step where a thick boss meets a thin wall. Give it three days of that and a microscopic flaw at one of those points becomes a visible crack. The same mechanism is why a loaded shelf bracket slowly bows instead of snapping, described in why 3D printed brackets sag.

Material decides how much stress gets built in. ABS shrinks roughly 0.5-0.8 percent on cooling against PLA's 0.2-0.3, which is why the identical model prints happily in PLA and splits in ABS on the same open-frame machine. What follows from that:

Slow the cooling. A printer enclosure traps bed heat and stops draughts, which is the single most effective change available for ABS and ASA. Turn part cooling down or off for those materials while you are at it.
Design the risers out. Fillet internal corners at 2-3mm, avoid abrupt thickness changes, and put a radius or a chamfer at the mouth of every hole that sits in a loaded area. Geometry is free and it does not expire.
Reduce the shrinkage instead. If the part does not need ABS specifically, PETG shrinks much less and tolerates an open machine, at the cost of a lower service temperature. The trade-offs are in why PETG prints worse than PLA.
Anneal only with your eyes open. Heat treatment does relieve locked-in stress, and it also moves the part while doing it — see why annealing warps 3D printed parts before trying it on anything that has to fit something else.

Where the material is not negotiable, the chamber is the hardware answer. A passive tent is a real improvement; an actively heated chamber is the specification that professional machines sell for exactly this problem, and QIDI's enclosed, heated-chamber range is where to read stated chamber temperature as a number rather than inferring it from a panel. If the honest conclusion is that your machine cannot run the material at all, the ceiling specifications across ELEGOO's FDM range are a fairer comparison than another retrofit — the full version of that argument is why your printer can't print ABS or nylon.

Does It Actually Need Drying?

Tell moisture from the faults that imitate it. Then dry it without wrecking it.

Cause 2: You Preloaded It at Assembly

If the crack is at a screw boss, a press fit or a bearing seat, the clock started when you assembled it rather than when you printed it.

A screw driven straight into plastic wedges the boss apart and leaves a permanent hoop stress around the thread. A pin pressed into a hole does the same thing from the inside. The part carries it easily on the day — and then creeps, because every thermoplastic does under constant stress. The material relaxes away from the load, the stress concentrates at the root of the thread, and a crack walks outward from there over hours or days until the boss splits, usually along a layer line.

The fixes are mechanical rather than material:

Heat-set brass inserts replace the plastic thread with a metal one and spread the clamp load over a much larger area. This is the default answer for anything that will be opened and closed more than once, and the rest of the thread question is in why 3D printed threads strip.
Thread-forming screws for plastics displace material rather than cutting it, which leaves far less stress behind than a wood screw or a machine screw forced into an undersized hole.
A washer under the head, so a conical head cannot act as a wedge on a countersink that was printed slightly tight.
Clearance, not interference. Design press fits with a real interference figure you have tested rather than a nominal dimension, and remember that printed holes come out undersized by default — the arithmetic is in why 3D printed parts don't fit together.

Cause 3: Something Chemical Got to It

A part that cracks a few days after being glued, cleaned or sprayed has usually met a chemistry it cannot carry stress in.

Two separate things happen. Cyanoacrylate cures hard and brittle, so a bonded joint that stays under load has no compliance in it and releases at the first thermal cycle — the failure modes are catalogued in why glued 3D prints come apart. And CA vapour, isopropyl alcohol and many household cleaners craze a stressed thermoplastic surface, opening fine cracks that grow into a fracture under a stress the part had been carrying without complaint. That is environmental stress cracking, and it is covered in depth in why 3D printed parts fail around oil, fuel and solvents.

For a joint that stays loaded, two-part epoxy is the better chemistry — it has some give, it does not outgas onto the surrounding surfaces as it cures, and it fills the gap that a printed joint always has. Bond parts that are relaxed, not parts that are being clamped into position by the glue.

Cause 4: The Weld Was Never Really There

A crack that runs cleanly along a single layer line in an unloaded part usually means that layer never fully bonded, and it took a few days of handling and thermal movement to finish the job.

The common cause is moisture. Filament that has absorbed water flashes it to steam at the nozzle, leaving voids and a rough bead that cannot key properly into the layer below. The part survives the build and the first handling because the stress is low, then separates at the weakest plane later. A heated dryer box fixes the stock going in — though drying is not a cure-all once the bead geometry is wrong, which is the point of why drying filament doesn't fix your prints.

The other contributors are a nozzle temperature at the bottom of the material's range and part cooling running hard on a material that does not want it. Both reduce how much the new bead re-melts the one beneath it. The full diagnosis for layer-line failures under load is why 3D prints snap along layer lines.

Cause 5: The Shelf Is Not a Neutral Environment

A garage, a loft, a car and an unheated workshop all cycle through a wide temperature range every day. The part expands and contracts, any metal fitted into it expands at a different rate, and each cycle works the existing residual stress a little further. Below about 10°C most printed thermoplastics also lose toughness and become easier to crack from a stress they shrugged off in a warm room — the detail is in why 3D printed parts break in the cold.

Resin parts have their own version. An SLA or MSLA part keeps curing slowly after it leaves the wash-and-cure station, and sunlight accelerates it, so a part that was tough on day one can be glassy and crack-prone a fortnight later — see why resin prints go brittle and yellow.

A Ten-Minute Triage

Work it in this order before changing any setting:

Find the origin. Follow the crack back to its narrowest end. That point is the stress concentration, and it names the cause.
Ask what was applied, and when. A screw, a press fit, a glue, a wipe with isopropyl, a move into a cold garage — a delayed crack almost always has an event three to seven days before it.
Check the material against the machine. High-shrinkage material on an open frame is cause 1 until proven otherwise.
Fillet the geometry. 2-3mm radii on internal corners and relief at hole edges, in CAD, before any reprint.
Reprint one coupon hot and slow with cooling off, then leave it a week. If it survives where the original failed, the cause was thermal rather than chemical.
Only then change filament brand. Material brand is the last variable worth touching, not the first.

When the Honest Answer Is a Different Process

Some parts should not be made this way. If the thing has to hold a static load for months, live outdoors through a winter, seal against a bearing, or arrive as fifty identical units that all behave the same, then the residual stress in a desktop FDM part is a property of the process rather than a setting you have not found yet.

That is the point to price a quote against another month of reprints. SLS and MJF nylon parts come out of a powder bed with far lower locked-in stress than an extruded part, industrial FDM runs a genuinely heated chamber, and machining or moulding removes the question. The shops that run those processes are listed by location and capability in the provider directory, and how to prepare files for a 3D printing quote covers what to send so the first quote is the real one.

Everything else is the same short list: slow the cooling, round the corners, stop wedging plastic apart with metal, and keep the filament dry. A part that is not fighting itself does not crack on a Thursday.

Hero photograph by Tom Claes on Unsplash.

Does It Actually Need Drying?

Tell moisture from the faults that imitate it. Then dry it without wrecking it.

Recommended Resources

Disclosure: Some links below may be affiliate links. We only recommend services we have personally evaluated or that are used by providers in our directory. Clicking earns us a small commission at no cost to you.

3D printer enclosure tent

The direct fix for the most common cause on this list. A passive tent holds enough chamber heat to slow cooling, which is what stops ABS, ASA and PC from locking stress into the part as they print.

QIDI's enclosed, heated-chamber printers

Where a tent is not enough. An actively heated chamber is the specification that matters for high-shrinkage materials — judge the models on stated chamber temperature, not on the enclosure panel alone.

ELEGOO's FDM range

The mainstream shelf to compare if the honest conclusion is that your open-frame machine cannot print the material this part needs. Read hotend temperature ceiling and enclosure as published specifications.

Heat-set threaded inserts, M3-M5

What stops a screw boss splitting three days after assembly. A brass insert gives a metal thread and spreads the clamp load into a larger area of plastic instead of wedging it apart.

Two-part epoxy for plastics

The adhesive to use on a joint that stays under load. Cyanoacrylate cures hard and brittle and its fumes craze some printed surfaces, which is a delayed crack waiting for a reason.

Heated filament dryer box

For the parts that crack along a layer line with nothing on them. Wet filament prints welds full of steam voids, and a weld that is half there survives the build and fails later.