How to Fix Stringing on 3D Prints: Retraction, Temperature, and Travel Settings

Stringing — those fine plastic hairs bridging the gap between towers or spanning open air — is one of the clearest diagnostic signals in FDM printing. Learning how to fix stringing on 3d prints comes down to three settings working together: retraction, nozzle temperature, and travel speed. Get any one wrong and the other two won’t compensate fully.

This guide walks through each lever systematically, with exact numbers for direct-drive and Bowden setups, and ends with a calibration sequence that locks in a stable result rather than one that drifts with the next filament swap.

What Is Actually Happening

The nozzle is a pressurized orifice. When a travel move begins, molten filament is still under pressure inside the melt zone. Without intervention, that pressure pushes plastic out through the nozzle as it moves across open air. It cools mid-air into a hair.

Retraction addresses this by pulling filament backward before the travel move, reducing pressure in the melt zone. Temperature addresses it by making the plastic less runny — higher-viscosity melt doesn’t drool as easily. Travel speed addresses it by minimizing the time the nozzle spends crossing open air.

Three conditions consistently worsen stringing:

• Nozzle temperature above what the filament actually needs
• Retraction length or speed dialed too low
• Travel speed slower than the printer can reliably handle

Retraction: Length and Speed

Retraction length is the most commonly tuned setting, and the number that works depends entirely on whether you’re running direct drive or Bowden.

Direct drive (Bambu X1C, Prusa MK4, Bambu A1 Mini, most modern CoreXY machines): start at 0.4–0.8 mm for PLA. The Bambu Lab stringing guide ↗ recommends 0.2–0.4 mm as a baseline, which is conservative but safe. Going past 2 mm on direct drive risks pulling a molten plug into the cold zone and causing a clog. More retraction is not always better.

Bowden extruder (Creality Ender 3, Ender 5, older cartesian machines): the filament path from drive gear to nozzle is 300–700 mm of PTFE tubing. Compressibility and delay mean you need 4–7 mm retraction to achieve the same pressure drop. The Prusa Knowledge Base ↗ notes the MINI/MINI+ uses up to 3.2 mm on its semi-Bowden layout.

Retraction speed matters just as much as length. Pulling filament back slowly still drops pressure slowly — the ooze window stays open. 35–45 mm/s is a good starting point for PLA on direct drive. Increase in 5 mm/s increments until you hear extruder skipping, then back off 5 mm/s. On Bowden printers, 40–60 mm/s is typical, but the PTFE tubing sets an upper limit; too fast and the tube flexes, reducing actual retraction effectiveness.

Temperature: The Often-Ignored Variable

Nozzle temperature controls melt viscosity. PLA at 215°C runs noticeably thinner than at 200°C, and thinner plastic oozes through any gap more readily.

Step your temperature down 5°C at a time and reprint a stringing test after each change. Generic PLA typically prints clean at 195–210°C. Branded engineering-grade PLA blends may need 210–215°C for layer adhesion — dropping below their floor causes under-extrusion, which is worse than stringing.

PETG behaves differently. Its printing window is 230–245°C and it strings even at minimum retraction if temperatures are too high. For PETG, prioritize temperature reduction over retraction increases; aggressive retraction on PETG causes the filament to stretch and snap rather than retract cleanly. Keep retraction at 0.5–1.0 mm (direct drive) and bring temperature down before touching retraction.

TPU is the edge case: use near-zero retraction (0–0.5 mm) regardless of stringing, because flexible filament compresses in the extruder instead of retracting, which causes tangles and jams. Accept minor stringing on TPU and remove it post-print.

Travel Speed and Combing

Travel speed directly controls how long the nozzle spends over open air. Most modern slicers default to 150–200 mm/s. If your profile is at 80 mm/s, that’s leaving significant ooze time on the table. CoreXY machines (Bambu, Voron) handle 250–300 mm/s travel cleanly. Bed-slingers with heavy moving beds should stay at 150–180 mm/s to avoid resonance during travel.

Combing (labeled “Avoid Crossing Walls” in OrcaSlicer and BambuStudio, “Combing Mode” in Cura) routes travel moves inside the perimeter of the model wherever geometrically possible. When the nozzle never crosses open air between outer walls, stringing doesn’t appear on the exterior even if there’s slight ooze. Enable combing for every PLA and PETG print. The cost is marginally longer travel paths — a few seconds per layer on typical parts, which is negligible.

Z-hop lifts the nozzle vertically before travel and lowers it after. It does not reduce ooze; it reduces the chance that a hanging drip drags across the surface during travel. Useful for tall, detailed models with many separate towers. Skip it for blocky parts; it adds measurable time per layer with little payoff.

Calibration Tower Method

Rather than guessing at retraction values, print a retraction calibration tower. OrcaSlicer has one built in under Calibration > Retraction Test ↗: it sweeps retraction length from a start to end value, stepping up each column. The shortest column with no visible hairs is your setting.

Recommended sequence:

1. Set temperature to your typical print temp.
2. Print a retraction tower (start 0.2 mm, end 2.0 mm, step 0.2 mm for direct drive).
3. Identify the cleanest column — note that retraction length.
4. Print a temperature tower (225°C to 190°C in 5°C steps for PLA).
5. Find the coolest layer without layer separation or surface roughness.
6. Lock both settings.
7. Reprint the retraction tower at the new temperature.

Steps 6 and 7 matter: temperature changes melt viscosity, which changes how much retraction pressure drop is needed. A 10°C temperature reduction sometimes lets you cut retraction length by 0.5 mm, which improves print speed and reduces retraction-related clogging risk.

Material Reference

Filament	Nozzle Temp (°C)	Direct Drive Retract (mm)	Bowden Retract (mm)
PLA	195–210	0.4–0.8	4–6
PETG	230–245	0.5–1.0	4–6
ABS	235–250	0.5–1.5	5–7
TPU	220–235	0–0.5	N/A (avoid Bowden)

Verifying the Fix

Print an octopus or spider-leg test model — any geometry with multiple narrow towers separated by open air. Pass criteria: no visible hairs between towers, or hairs thin enough to snap off cleanly with a fingernail rather than requiring a heat gun or knife. Surface fuzz on PETG is normal and polishes off. Persistent hairs on PLA after the above steps usually mean temperature is still above the optimal range.

If stringing persists after correct retraction and temperature, check for partial Bowden tube degradation (PTFE that has compressed or deformed at the coupler end reduces effective retraction) and verify that your minimum travel distance for retraction is at least 1 mm — short travel moves shouldn’t trigger retraction at all.

Sources

• Prusa Knowledge Base — Stringing and Oozing (link ↗): Official Prusa documentation covering retraction length limits by printer type, lift Z, and minimum travel settings. Authoritative for MK4, MINI, and XL users.
• Bambu Lab Wiki — Stringing and Oozing (link ↗): Bambu’s official guidance on retraction and temperature tuning, covering AMS multi-material considerations alongside standard single-filament setups.
• OrcaSlicer Wiki — Retraction Calibration (link ↗): Documents the built-in retraction test tower and explains the wipe sequencing options (retract before vs. after wipe and their trade-offs).
• All3DP — 3D Print Stringing: Easy Ways to Prevent It (link ↗): Broad overview covering coasting, wipe, and advanced slicer options not covered in vendor docs, with comparisons across multiple slicers.