Confirm which material is leaving the nozzle

Fine threads between separate features suggest material escaping during travel. A rough surface with popping or inconsistent flow deserves a material-condition check first. Retraction pulls filament back temporarily; it does not remove absorbed moisture or repair a damaged nozzle. Establish reliable extrusion with the normal material profile before turning a stringing test into a long series of unrelated changes.

Use a test with repeatable travel

Choose two or more narrow features with repeated moves between them. Keep spacing, orientation, layer height, temperature, and cooling constant. Save a baseline print and note the current retraction length and speed, including filament overrides. Check the toolpath to confirm that the travel moves actually trigger retraction. A minimum travel threshold or perimeter-only condition can prevent the setting from affecting the move you are evaluating.

Test length and speed separately

Start from a validated profile for your extruder type. Adjust length within the documented range while holding speed fixed. Compare string count and restart quality. Restore the best length before trying a speed change. Direct-drive and Bowden paths differ; transferring a large Bowden retraction into a short direct-drive path may cause trouble. Stop if clicking, grinding, or interrupted extrusion appears.

Judge the beginning of each new line

A cleaner gap between towers is only half of the result. Look at the first segment after travel for thin lines or missing material. Excessive retraction can trade strings for restart defects. Compare the same layer and viewing angle on each sample. If temperature tuning is needed, return to the baseline and run a separate comparison within the filament and hotend limits rather than changing it halfway through the retraction series.

Transfer the result to a real part

Run a small section of the intended object with its actual travel distances and density of features. Avoid assuming that an extreme test model represents every print. Travel planning, wipe behavior, and cooling can change the outcome. Keep the smallest reliable adjustment that improves strings without gaps, then save it as a named material profile. If no supported setting helps, investigate material condition and flow rather than continually increasing retraction.

Score the comparison before choosing a winner

Use the same viewing angle and light for each sample. Record threads in the open gap separately from defects in the new extrusion path. A simple descriptive score is enough; there is no need to invent a precision measurement from a photograph.

Sample observation Interpretation to test next
Fewer strands, continuous restarts Repeat the candidate setting before adopting it.
Fewer strands, new gaps after travel Reject the tradeoff and return to the previous value.
Similar strands across supported adjustments Confirm the tested moves trigger retraction; revisit material and temperature.
Clicking or ground filament Stop the series and investigate feed resistance or an unsupported setting.

Do not judge the very bottom of a tower against the top of another. Cooling time, travel, and feature size can change with height. If a model varies settings automatically, verify which value applies to each section using your slicer’s supported method. A number embossed on the mesh does not prove that the printer used that number.

Example: a clean gap with a damaged restart

Imagine two samples. The baseline has fine hairs but complete tower walls. A trial leaves fewer hairs but has short missing segments after travel. The trial has not passed: it has moved material loss from the gap into the object. Restore the baseline and choose a smaller supported adjustment, or investigate a different cause. These are hypothetical observations, not measured PrintFix3D results.

Keep Bambu calibration terms separate

Bed leveling, flow-ratio calibration, flow-dynamics calibration, and retraction answer different questions. Do not assume that changing a retraction field replaces a printer’s documented calibration. Start with the correct Bambu printer, nozzle, and material profile, and check the instructions for your software version before using a third-party test or custom commands. This guide supplies no universal Bambu retraction distance.

If the defect consists of marks at perimeter starts rather than threads between separate features, use the seam guide. If the material pops or becomes rough, consult filament storage and drying before extending the retraction series.

Comparison record

Check Evidence to keep
Travel event Confirm retraction actually occurs on the inspected move.
One parameter Record length or speed, preserving the other value.
Restart Score missing lines after travel as well as remaining threads.

Reference and limits

The linked Prusa stringing guidance explains retraction and the distinction between extruder types. Its numerical settings belong to its listed hardware. Our scoring table and hypothetical samples are comparison methods, not measured retraction limits for Bambu or another manufacturer.