Warping is not one setting error. It is a competition between shrinkage stress developing inside the part and the forces holding the system in place. Freshly deposited polymer cools and wants to contract while the layer below or build plate restrains it. When stress exceeds first-layer adhesion or interlayer cohesion, a corner lifts, a wall bends or layers split.

Identify the symptom first

A bottom corner that gradually lifts is classic warping. If the whole object suddenly shifts, investigate adhesion, collision or mechanics instead. If only an overhanging edge curls while the base remains flat, local cooling and geometry are more likely. A horizontal crack high on a wall may be layer separation: related to the same thermal gradient, but not a release of the first layer.

Mark which corner and height fail and at what print time. Photograph the gap from the side and measure it. A repeated direction may point to a fan, draught, bed-temperature variation or machine position; random failure suggests contamination or unstable surroundings.

Material, geometry and thermal gradient

Higher-shrinkage materials printed at elevated temperature—such as ABS, ASA and many PC blends—are more sensitive than PLA. Long straight walls, sharp corners, thick solid bases and abrupt section changes collect stress. Rounded corners, ribs, more uniform wall thickness and less solid mass can reduce it.

A datasheet shrinkage value is useful but cannot predict a printed part alone. Nozzle temperature, layer time, airflow, crystallisation, fibre reinforcement and orientation interact. The same filament may be flawless on a small model and warp on a large flat component.

Gate one: a clean, flat build surface

There is no value in tuning enclosure or profile on a greasy, dusty or unevenly coated plate. Clean it using the manufacturer-approved method, then verify nozzle distance and flatness across the usable area. First-layer lines should meet without being crushed until polymer ridges beside the nozzle.

Adhesive increases grip on some surfaces and acts as a release layer on others. Follow the printer and plate maker's material pairing. A hotter bed is not a universal cure: it can create a soft lower zone, elephant foot or greater heat accumulation over a long build.

Gate two: environment and enclosure

Remove open windows, air conditioning, door movement and direct fans. For ABS, ASA and PC, an enclosure slows cooling and evens the surroundings. The objective is not necessarily the highest chamber temperature but stability. Electronics, motors, power supplies and spools should remain in a heated volume only where the machine documentation permits it.

Precondition the enclosure by letting the heated bed settle, and do not open the door during comparative tests. Record temperature at more than one location if possible. One sensor at the roof may not represent the air around the part.

Gate three: cooling, speed and layer time

Excessive part cooling increases thermal differences and can reduce layer bonding; too little cooling can curl small overhangs until the nozzle strikes them. Reduced fan for early layers followed by a gradual increase is usually a better baseline. Treat minimum layer time and small-feature speed separately.

Slowing down does not always help. It may allow a layer to cool further while increasing total time in the chamber. Change one variable at a time: fan, bed, nozzle, speed or chamber. Rewriting the entire profile seems fast but destroys causal evidence.

Brims, tabs, draft shields and placement

A brim expands the bonded area and is useful at sharp corners. Corner tabs add resistance where stress concentrates. A draft shield can reduce local airflow on an open machine. These are treatments, not substitutes for a clean plate or stable environment.

Place the model in a stable, evenly heated region, usually near the bed centre. Rotate it so the longest edge does not track a colder perimeter. For a large part, segmentation, mechanical joining or another orientation may be more reliable than endlessly increasing adhesion.

A repeatable troubleshooting experiment

Create a smaller coupon that reproduces the critical corner, wall thickness and footprint. Record filament lot and drying, plate, cleaning, nozzle, temperatures, fan, chamber, brim width and file revision. Measure corner lift, base flatness and layer integrity on every run.

Restore the adhesion baseline first, stabilise the environment second, tune cooling and temperatures third, and only then change geometry or aids. Three repeated successes are stronger evidence than one lucky part. Measure functional dimensions after cooling and conditioning because a warm part can continue to move.

Fast decision order

1. Identify and clean the plate. 2. Verify first layer and flatness. 3. Remove draughts. 4. Use material-appropriate chamber and cooling. 5. Add a brim or corner tabs. 6. Reduce stress-concentrating geometry. 7. Change temperature only in a documented test. 8. Repeat the successful condition.

Source types and editorial conclusion

Prusa's knowledge base highlights temperature difference, stable surroundings, clean surface, enclosure, controlled cooling, brim and model placement. Exact limits for a filament must come from its supplier data and a local test. LAYERTIMES concludes that warping is not solved by a single 'more adhesion' control. A durable solution balances thermal gradient, geometry and restraint through measurable experiments.