Retraction length, retraction speed and travel path describe different things. Increasing a value at random does not identify why strings appear between regions. Start with the profile intended for the extrusion system and material.
Read each setting with its unit
Prusa documentation distinguishes the amount withdrawn, motor speed during withdrawal and travel conditions triggering retraction. Its suggested values concern specific machines and vary by extrusion system. They are not a universal recipe.
| Setting | Question | Confusion to avoid |
|---|---|---|
| Length, mm | How much filament is withdrawn? | Print-head travel distance |
| Speed, mm/s | How quickly is it withdrawn? | Wall printing speed |
| Travel condition | When is retraction triggered? | Assuming it occurs on every move |
| Z lift | How does the head avoid the part? | Treating it as first-layer adjustment |
Describe the defect before comparing
Locate strings: between objects, inside a cavity or after a specific region. Keep a screenshot of corresponding slicer travel moves. Record filament identity and the initial profile. These observations support comparison, not a unique diagnosis.
Prepare a small representative project and retain other conditions. For each variant, record the changed parameter and expected result. Improvement in one area with a new defect elsewhere is not automatically an overall improvement.
Keep a decision record
A test table can contain initial value, selected value, unchanged conditions, observations and whether to return to the initial profile. Do not copy a length from another extrusion system without justification.
Prusa warns that excessive retraction speed can cause skipped motor steps and explains Z-lift/collision trade-offs. Do not disable travel protection merely to remove a string; safe movement still needs checking.
Broaden the investigation only when needed
If behaviour does not fit the hypothesis, return to general diagnosis. The storage/drying guide organizes another set of checks without attributing all stringing to moisture. The PETG profile explains preparation and use limitations.
LayerBench3D’s G-code analyzer does not reconstruct retractions and cannot validate these settings. This article provides a reading and comparison method without optimal values or invented physical tests.
Sources consulted
- Stringing and oozing — Prusa Research
Accessed