Fix PETG Layer Separation: Why the Filament That Resists Warping Still Cracks

PETG is supposed to be the material that does not do this

Prusa’s own knowledge base is blunt about it: PETG “has similar advantages” to ABS, but “does not have the same tendency to crack/de-laminate.” That reputation is earned. PETG does not shrink much as it cools, so it is not fighting the same internal stress that splits ABS parts open. Most of the time, a PETG print that comes off the bed intact stays intact.

So when a PETG part does crack along a layer line, the usual advice does not help much, because most delamination guides are written with ABS and PLA in mind. PETG’s resistance is a default, not a guarantee, and a handful of PETG-specific choices are enough to erode it: print speed left ahead of nozzle temperature, a flow ratio nudged down to fix an unrelated problem, or a cooling fan set the way you would set it for PLA. Each one is explained below with the fix, in the order worth checking first.

One more thing worth knowing before troubleshooting settings: the crack does not always show up while the printer is still running. Prusa’s own definition of layer separation describes it as forces building up while layers cool at different rates, until those forces finally exceed how well the layers are bonded. On a part with borderline adhesion, that threshold can be crossed minutes or hours after the print finishes, once the whole object has cooled and settled and internal stress has had time to relax through the weakest layer. A print that looked fine coming off the bed and only cracked later is not a mystery, it is the same underlying weak bond, just discovered on a delay.

1. Print speed and nozzle temperature are moving targets, not fixed numbers

Bambu Lab’s own PETG troubleshooting guide ties layer separation directly to this: when print speed goes up, the nozzle has less time to melt each millimeter of filament, so the temperature needs to go up with it, or the plastic reaches the layer below only partially melted and never really welds to it. Run the reverse and you get the opposite problem: slow the print down without lowering temperature, and PETG has more time than it needs at the nozzle, which shows up as stringing and oozing instead.

Bambu’s own numbers make the pairing concrete. Their suggested fallback for PETG HF that has not been dried, Generic PETG HF, runs a lower volumetric flow rate of 16 mm³/s at 220°C, specifically because it is more forgiving than their faster default Bambu PETG HF profile at the same temperature. That is the same coupling in miniature: drop the speed (and the volumetric flow demand that comes with it) without touching temperature, and the print gets more stable, not less, because the nozzle finally has enough time to fully melt what it is pushing out.

In practice this means a slicer profile is not a fixed pairing. If you have bumped up print speed for a faster job, or you are running a fast third-party profile at PETG’s default temperature, check whether temperature moved with it. Test in 5°C steps, the same increment Prusa recommends for PLA cracking, and watch for stringing as the ceiling. That ceiling is exactly what the Fix PETG Stringing guide covers in more depth if that is the side you hit first.

2. A flow ratio pushed too low starves the bond

Bambu Lab publishes a specific working range for PETG-family filaments: a flow ratio of 0.93 to 0.96. Drop below 0.93 and you get under-extrusion, which weakens layer adhesion along with everything else. Push it above 0.96 and the risk flips toward oozing and clogging instead. Their own default sits at 0.95 for a reason, it leaves margin on both sides.

This matters because flow ratio is a common target when someone is chasing a clogging complaint or a slightly over-extruded outer wall, and dialing it down is a reasonable first instinct. If that adjustment happened recently and delamination showed up afterward, that is not a coincidence to rule out. Nudge flow back up in small steps, 0.01 at a time, rather than resetting it blind, and note that this exact band comes from Bambu’s own AMS-fed profiles. The underlying idea, that under-extrusion starves layer bonding, is the same principle Prusa documents for its extrusion multiplier on any printer, but the precise edges of the range may sit slightly differently on a Bowden setup from another brand.

3. The cooling habit that carries over badly from PLA

PETG bonds better between layers when it stays warm a little longer, which is the opposite of how most people are trained to think about cooling after printing PLA. FLG’s own PETG settings guide already puts this in one sentence: check for cooling that is too aggressive before anything else if a PETG print delaminates under light stress. Prusa’s guidance backs it up with numbers, fan off for the first few layers, then roughly half power afterward, not the near-full cooling PLA wants almost immediately.

The most common way people end up here by accident is fixing a different problem with the fan. Bumping cooling to clean up a sagging bridge or a rough overhang is a reasonable instinct, but applying it globally pulls heat out of every top surface before the next layer has a chance to fuse to it, not just the overhang that needed it. Limit stronger cooling to the specific features that need it in the slicer’s per-feature settings, rather than raising the whole print’s fan curve.

4. Wet filament turns the melt porous instead of solid

PETG absorbs moisture readily, and Bambu’s documentation describes exactly what that does at the nozzle: trapped water flashes to steam at print temperature, and the expanding melt fills with pores instead of flowing solid. The result is stringing and surface holes, but also a genuinely weaker bond between layers, since a porous melt has less material actually welding to the layer beneath it. This is the same mechanism the site’s generic layer separation guide flags as showing up most with PETG and nylon.

The tell is usually audible before it is visible: a faint crackling or popping at the nozzle as the print runs, on top of the stringing and rough surface you would expect from wet filament anyway. If a spool has been open for a while or stored somewhere humid, dry it before changing any other setting, since chasing temperature or flow adjustments on wet filament just moves the problem around instead of removing it. Bambu’s published parameters are 60 to 65°C for 8 hours in a convection oven, or roughly 80°C for 12 hours on a heated bed, flipping the spool every 6 hours so both sides dry evenly. The site’s full filament drying guide covers oven and dedicated-dryer methods in more detail than fits here.

5. Drafts do quietly what an aggressive fan does directly

PETG does not need the enclosure that ABS effectively requires, but it is not completely indifferent to its environment either. A cold draft from an open window or an air conditioning vent pulls heat away unevenly, and the side of a tall print facing that draft can bond noticeably worse than the side that does not, even with temperature, flow and cooling all set correctly. This matters most on tall, thin prints, where the top layers being extruded are furthest from the heated bed and most exposed to moving air, and least on short, wide parts that finish before ambient conditions have much chance to shift. Closing doors and windows during a print, moving the printer away from an AC vent, or adding even a loose enclosure for tall PETG parts removes a variable that is easy to miss because nothing about the slicer profile looks wrong.

6. Rule out warping before you chase a bonding fix

Layer separation and warping share a root cause on tall prints, layers cooling unevenly enough to build internal stress, so it is worth confirming which one is actually happening before spending time on settings that will not move the needle. Corners lifting off the bed while the print is still running is warping, not delamination, and the fix lives in a different guide. PETG does warp, just less dramatically than ABS, and Fix PETG Warping covers bed adhesion and first-layer settings specific to that failure mode. Fixing one problem does not automatically fix the other, since a print can warp without ever cracking along a layer, or crack along a layer without any corner lifting at all.

PETG layer separation checklist

• Nozzle temperature matches current print speed, tested in 5°C steps against stringing as the ceiling.
• Flow ratio sits within roughly 0.93 to 0.96, nudged up in 0.01 steps if it was recently lowered.
• Cooling fan follows PETG’s own curve (off for the first layers, ~50% after), not a PLA-style aggressive setting, with stronger cooling limited to overhangs and bridges only.
• Filament has been dried if it has been open or stored somewhere humid.
• Drafts are ruled out, or a loose enclosure is in use for tall parts.
• Confirmed as delamination and not warping before changing settings further.

If the crack shows up on ABS or ASA instead, the same checklist does not transfer directly, since those materials are fighting shrinkage-driven warping stress that PETG mostly avoids. The site’s generic layer separation guide covers those materials and the drafts/speed/temperature basics that apply across the board.