Change One Thing at a Time

Reliable print troubleshooting is an experiment: keep everything the same, change exactly one variable, and reprint a small, repeatable test so you can trust the comparison. This prevents “false fixes” where a second change hides the real cause or where random variation looks like improvement.

TL;DR

Pick one variable (for example nozzle temperature or retraction distance), change only that, then reprint the same small test model and compare to the previous sample side-by-side. If you change multiple settings at once, you won’t know what actually fixed (or caused) the defect.

Single-Variable Troubleshooting LoopTopic-specific diagram for the concept, checks, and tradeoffs in this lesson.SymptomDescribe, don’t diagnoseLast changeWhat changed since good?Pick 1 variableTemp or speed or filamentSmall testSame model, same setupCompareBetter, worse, sameKeep or revertLock in, undo, repeat
A simple loop: hold everything constant, change one variable, re-test, then keep or revert based on evidence.

Why “one change” works (and multi-change doesn’t)

Most print defects have overlapping symptoms. If you increase temperature and slow down at the same time, an improvement could be from either change, their interaction, or just normal print-to-print variation. A single-variable test turns guessing into a clear cause/effect check: you can say “this specific change improved this specific symptom on this specific test.”

What counts as “one thing”

One thing is one variable in one category. Examples: Material: switch to a different spool, dry the same spool, or raise/lower filament diameter override (one of those, not all). Machine: tighten one belt, clean the nozzle, change the nozzle, re-level, or adjust extruder tension (pick one). Environment: close the enclosure, move away from a draft, or raise room temperature (one change). Model: rotate orientation or move the part on the bed (one). Slicer: temperature, speed, cooling, retraction, flow, layer height (one variable at a time). Temperature plus speed is two things.

A simple troubleshooting loop you can repeat

  1. Name the symptom in observable terms: where it appears, when it starts (layer/time), and what it looks like.
  2. Return to (or note) the last known-good setup and write down what changed since then: filament, nozzle, bed surface, profile, room draft/temperature.
  3. Choose the most likely single variable to test next (the one with the highest chance to affect that symptom).
  4. Run a short, consistent test print: same model, same filament, same bed area, same profile except for the one change.
  5. Compare the new print to the previous one. Keep the change only if the symptom clearly improves; otherwise revert and pick the next variable.

Symptom clues to pick your first single change

A defect appears suddenly after many good prints

Likely cause: Something external changed: dirty bed, partial nozzle clog, loose hardware, wet filament, new draft, or a profile change

Fix: Go back to the last known-good slicer profile and material. Clean the bed, inspect the nozzle/extruder path, then run a small repeatable test before changing settings.

The defect repeats in the same location or along one direction

Likely cause: Geometry/orientation/cooling/motion issue rather than random settings noise

Fix: Check slicer preview at the problem layers. Change only one factor (for example rotate the model 90° or adjust cooling) and reprint the same test.

Several fixes seem plausible and you feel tempted to “tune everything”

Likely cause: Stacked changes masking the true cause (common with under-extrusion vs cooling vs speed)

Fix: Pick the highest-probability variable and test only that. Keep both best and worst samples so you don’t convince yourself a tiny change was a big win.