Direct answer: PETG that splits cleanly along layer lines usually has an interlayer-bond problem, not a bed-adhesion problem. Check part cooling and room drafts first, then compare the active print speed with the supported PETG profile, review nozzle heat, prove the spool is extruding consistently, and finally inspect the wall geometry and load direction. Change one variable at a time and break the same small test part after each change.
This page covers one exact failure: a PETG part separates between stacked layers, often leaving a horizontal split or a wall that peels apart along a layer line. It does not cover corners lifting from the build plate, a single vertical seam opening, or a part snapping across solid material without following the layer stack.
The checks below are diagnostic starting points, not universal temperature or strength guarantees. Begin with the current printer and filament makers' supported PETG profile and safety instructions.
Confirm that the crack actually follows the layers
Look at both fracture faces under good light. Interlayer failure usually follows a mostly horizontal path and exposes recognizable deposited lines. The layers may peel apart more easily than the strands themselves break. A crack that begins at a sharp corner, screw hole, thin notch, or vertical seam may still involve weak bonding, but geometry or seam placement can be the trigger.
If the bottom corners curled upward during the print, use the PETG corner-lift guide. If the wall shows missing material or sparse, inconsistent lines before it breaks, diagnose under-extrusion first. A useful fix starts with the right symptom.
Match the evidence to the first likely cause
| What you see | Best first lead | First controlled check |
|---|---|---|
| The split is worse on the fan-facing or draft-facing side | Uneven cooling | Restore the supported fan baseline and remove the room-air disturbance |
| A faster profile created the failure | Melt delivery cannot keep up | Repeat the same coupon at the known supported speed |
| A temperature reduction preceded weaker walls | Effectively too-cold extrusion | Return to the filament maker's profile before fine tuning |
| Surface texture, popping, or flow varies through the part | Spool-condition or feed inconsistency | Dry and handle the spool as its maker directs, then repeat unchanged |
| Only a hole, notch, thin wall, or sharp transition cracks | Stress concentration | Test a thicker or radiused version with the same profile |
| A strong-looking part fails when pulled across the layer stack | Orientation-loaded seam | Reorient the load path or redesign the joint |
Check 1: compare the fan side and the room side
Before changing a number, mark the side that faced the part-cooling duct, an open printer door, a vent, a fan, or normal foot traffic. If the crack is concentrated on that side, the asymmetry is evidence. PETG needs enough cooling to preserve shape, but aggressive or uneven airflow can reduce the time adjacent roads have to bond.
Return to the printer or filament maker's supported PETG cooling baseline. Remove obvious room drafts from the test without blocking required machine ventilation or defeating safety systems. Do not jump straight to zero fan: small features and bridges may need cooling, and eliminating it everywhere can trade a bonding defect for soft corners and poor overhangs.
Check 2: undo the speed change before adding more heat
A nozzle setting is not the whole thermal story. When flow demand rises, the hotend has less time to melt each unit of filament. A profile that bonded well at a moderate rate can behave effectively colder after a speed or volumetric-flow increase even though the displayed temperature did not change.
If the cracking began with a faster profile, repeat the same test at the last supported speed. This is safer evidence than immediately raising heat to compensate for an unproven flow limit. If the slower coupon bonds while the fast one splits, keep the lower demand or use a hotend and profile whose supported flow range fits the job.
Check 3: restore the supported PETG temperature baseline
Review the actual sliced profile, not what you remember selecting. Confirm the filament preset, nozzle size, layer height, and any first-layer-only or bridge overrides. If heat was reduced to fight stringing or improve surface appearance, return to the filament maker's PETG baseline and test again at the supported speed.
Use small, controlled steps only after the baseline test proves that bonding still needs improvement. More heat can improve layer fusion, but it can also increase stringing, soft detail, residue, or hotend limits. The manufacturer's range and the printer's material restrictions remain the boundary. Prusa's PETG material guide is one example of model-specific material guidance; use the equivalent current source for your printer and filament.
Check 4: prove whether the spool is making extrusion inconsistent
Moisture is not the automatic answer to every weak PETG part. Treat it as the lead when the crack arrives with changing surface texture, popping, fuzz, intermittent bubbles, or flow that varies through the same wall. A uniform clean print that always splits at one notch points more strongly to cooling, heat delivery, or geometry.
If the evidence supports a spool problem, follow the filament maker's drying limits and the appliance maker's operating instructions. Then repeat the identical coupon from a controlled spool path. The wet-filament diagnosis guide helps separate spool clues from random tuning, while the PETG drying and storage guide covers recovery versus prevention.
Check 5: inspect wall thickness and stress risers
A process problem usually affects more than one ordinary section. A crack that repeatedly starts beside a screw hole, insert, notch, sharp inside corner, sudden wall change, or thin shell has a geometry component. The feature concentrates load at a small section of the layer stack.
Print a comparison with a radius, smoother thickness transition, more appropriate wall structure, or a larger local section while keeping the material profile unchanged. Do not hide a badly loaded joint under a large global infill increase. Walls, local shape, and load transfer usually matter more to the crack than filling untouched interior space.
Check 6: follow the real load through the layer stack
FDM parts are direction-dependent. A bracket can survive a load carried along continuous roads yet split when the same force tries to peel stacked layers apart. If the fracture follows the build plane under the expected service load, orientation is part of the design decision.
Reorient the part so critical tension and bending are carried through continuous paths where practical. If orientation is constrained by surface finish, support access, or dimensional accuracy, redesign the joint, enlarge the section, add mechanical fastening, or choose another manufacturing approach. A hotter profile is not certification for a safety-critical load.
Apply only the fix the check proved
- Draft or fan asymmetry: restore the supported PETG fan baseline and remove uncontrolled room airflow from the test.
- Fast profile fails, baseline passes: reduce flow demand or use a supported higher-flow setup rather than stacking random heat and fan changes.
- Temperature reduction caused the split: return to the maker's profile, then tune inside its supported range.
- Spool evidence is present: dry and handle the PETG as specified, then retest before touching the slicer.
- One feature always cracks: soften the stress transition, increase the local section, or change the joint.
- The load peels layers apart: reorient or redesign the load path; do not assume profile tuning removes anisotropy.
Run one controlled layer-bond test
- Use the same PETG spool, nozzle, layer height, and printer as the failed part.
- Slice a small wall or bracket coupon that reproduces the crack direction without wasting the full job.
- Record the filament profile, fan behavior, nozzle setting, speed or flow limit, room condition, and orientation.
- Print the baseline coupon and mark which side faced the fan or draft.
- Change only the highest-confidence lead from the ordered checks.
- Let both coupons reach the same condition, then load them in the same direction and compare where the fracture travels.
- Confirm the winning change on a second coupon before returning to the full-size part.
This test does not create a certified strength value. It tells you whether one controlled change improves the exact failure you are trying to stop.
Use limits: what a passing layer-bond coupon actually proves
A coupon that stops splitting proves only that the tested spool, profile, orientation, geometry, and room conditions produced a better comparison. It does not assign a strength rating to PETG or qualify the finished part for a different load or environment.
- Cosmetic cover or lightly handled indoor part: accept it only after the cooled part has no visible opening, peeling layer, loose wall, or fit change at the real mate.
- Fastened, loaded, impact, or repeated-flex part: a hand bend is not enough. Print the real orientation and critical feature, then run a representative load and cycle test. Reject a part that opens at a layer, loses fastener preload, changes alignment, or grows a crack.
- Warm, outdoor, wet, or chemical-exposed service: the room- temperature coupon does not prove retained dimensions, creep resistance, weather life, or chemical compatibility. Qualify the exact PETG grade under the job's real temperature, exposure, duration, and load.
- Sealing, pressure, lifting, food-contact, medical, electrical- safety, or other consequence-heavy use: visible layer bonding is not proof of pressure integrity, sanitation, load capacity, dielectric safety, or compliance. Use an appropriate validated process, material, inspection, and safety factor.
- The crack returns with clicking, missing lines, nozzle contact, a growing hotend blob, material leaking above the heater block, smoke, or an electrical smell: stop the print and follow the printer or hotend maker's safe inspection procedure. More temperature or more walls is not a substitute for correcting a feed, motion, or hotend fault.
Release sequence: first pass the controlled coupon, then reprint the actual critical feature in its intended orientation, let it cool, inspect both sides and the layer stack, verify the real mate, and run the representative load or exposure test. Repeat the job at least once before treating the fix as stable; one passing coupon can hide a draft, spool, or flow condition that later returns.
Common fixes that waste time
- Changing fan, heat, speed, flow, and walls together. A passing reprint will not reveal which change mattered.
- Drying every spool before inspecting the fracture. Drying is appropriate when spool evidence exists, not a substitute for checking one-sided cooling or a sharp stress riser.
- Turning the fan off for the whole print. This may help bonding but damage bridges, small details, and shape control.
- Raising infill globally. It adds time and material while leaving a thin wall, notch, or peeling load path unchanged.
- Testing a different model after every change. Different geometry destroys the comparison.
What to do next
If the baseline-speed coupon passes, keep the lower flow demand and review the printer's supported hotend limits before trying to recover speed. If the draft-side coupon alone fails, stabilize airflow and repeat the same orientation. If a maker-compliant drying cycle restores consistent walls, improve between-print storage instead of retuning around a drifting spool.
If the same notch or load direction still creates the split after the process baseline is stable, stop treating it as a slicer-only problem. Redesign the feature or orientation and test the actual service direction. For broader diagnosis, use the weak-layer adhesion guide. If the real issue is wispy travel strings rather than a fracture, move to the PETG stringing guide.
Common questions
Should I raise nozzle temperature when PETG layers split?
First restore the filament maker's supported profile and known-good speed. If the baseline still splits, a controlled adjustment within the supported range may help. Do not use extra heat to disguise excessive flow demand.
Should PETG use no cooling fan?
Not as a universal rule. PETG often needs a balance between bonding and shape control. Start with the supported profile and use the fracture pattern to decide whether cooling is excessive or uneven.
Does wet PETG cause weak layers?
Spool condition can make extrusion inconsistent and muddy layer bonding, but it is not the only cause. Look for popping, texture change, bubbles, or variable flow before making moisture the first diagnosis.
Will more walls stop PETG from cracking?
More appropriate wall structure can help a thin section, but it will not fix a room draft, unsupported flow demand, or a joint loaded directly across the layer stack. Match the change to the evidence.
When should I redesign the part?
Redesign when the same sharp feature or peeling load path fails after the material process is stable. Add a radius or local section, change the joint, reorient the part, or use a more suitable process for critical service.