Diagnosing Banding and Ghosting Artifacts in Fiber Laser Galvo Engraving

A fiber laser galvo marking machine (EZCAD-driven, mirror-steered rather than gantry-driven) has a completely different set of failure modes than a belt-and-gantry diode or CO2 laser, because the beam is steered by two fast-moving mirrors instead of a physical XY carriage. When a galvo-marked part comes out with regular banding across a fill, a faint ghost outline offset from the real mark, or inconsistent fill density edge-to-edge, the cause is almost never a mechanical belt or rail problem the way it would be on a gantry machine — it's galvo tuning, timing, or lens geometry. This is a different diagnostic process than the setup and lens-selection content elsewhere on this site covers, and it's worth having as a dedicated troubleshooting reference because galvo defects genuinely don't map onto the gantry-laser mental model most people bring to them.

Banding: Regular Stripes Across a Filled Area

Visible banding in a filled area, where density alternates in a regular pattern, almost always traces back to one of a small number of causes:

• Hatch/fill line spacing too coarse for your spot size. If the fill line pitch in your software is close to or larger than the actual focused spot diameter, you get visible gaps between passes. Tighten the fill spacing, or re-check spot size if you've changed the lens.
• Mismatched scan speed and laser response (jump delay / mark delay settings). EZCAD's delay parameters (laser-on delay, laser-off delay, polygon delay) exist specifically to compensate for the time it takes the galvo mirrors to actually reach commanded position and the laser to actually reach commanded power before marking starts. Wrong delay values produce exactly this kind of periodic density variation, especially at direction changes in a fill pattern.
• Power supply or Q-switch instability on the laser source itself, producing periodic pulse energy variation independent of the galvo at all — less common, but worth ruling out by checking whether the banding direction/pattern follows the fill pattern (points to galvo timing) or appears regardless of fill orientation (points to the source itself).

Ghosting: A Faint Offset Duplicate of the Mark

A faint secondary image offset from the real mark is a classic galvo mechanical response symptom:

• Galvo mirror overshoot and ringing. The scan mirrors are driven by a servo loop, and if that loop is under-damped for your current scan speed, the mirror overshoots its target position and rings briefly before settling — marking happens slightly before the ring fully damps out, leaving a faint duplicate. Reducing scan speed, or tuning the galvo controller's damping if your system exposes that setting, addresses this directly.
• Insufficient settle time after a jump. Similar root cause to banding's delay issue — if the software starts marking before the galvo has physically arrived and stabilized at the new position after a jump move, you get a ghost at the jump's intermediate position. Increase jump delay incrementally and retest.
• F-theta lens field distortion at the extreme edges of the work field can produce a secondary distortion that reads as a faint double-image specifically near the corners of large marks, while the center of the field stays clean — if ghosting is field-position-dependent rather than uniform across the whole mark, suspect the lens and field calibration rather than the galvo timing.

A Practical Isolation Procedure

1. Run a test grid, not a real job. Mark a simple grid of filled squares at varying speed and power across the field, the same basic diagnostic approach used for dialing in diode/CO2 laser settings, adapted for galvo-specific parameters (speed, delays) instead of speed/power alone.
2. Check if the defect is field-position-dependent. Mark the same test pattern at the center and at the far corners of your work area. If the defect only appears away from center, it's a lens/field calibration issue, not a timing issue.
3. Vary scan speed alone, holding everything else constant. If banding or ghosting gets worse at higher speed and better at lower speed, that's a strong signal toward galvo response/damping and delay settings rather than a fixed hardware fault.
4. Vary hatch spacing alone. If banding specifically changes with fill line pitch while everything else is held constant, that isolates it to fill overlap rather than timing.
5. Check your current EZCAD delay values against the manufacturer's recommended baseline for your specific laser source and lens combination before hand-tuning from scratch — most sources ship a sane starting point, and defects often trace back to those having been changed (sometimes accidentally, through a corrupted or copied parameter file) rather than needing tuning from zero.

When It's Actually a Lens or Mirror Contamination Issue

If banding or ghosting is accompanied by an overall drop in mark contrast or an uneven haze across the field, check the protective window and galvo mirrors for contamination before chasing software settings further — a dirty optical path produces its own density inconsistency that can look superficially similar to a timing problem but won't respond to any amount of delay tuning. Clean per the laser source manufacturer's specific procedure; galvo mirrors are front-surface and far more delicate to clean incorrectly than a standard CO2 laser cutter's rear-surface mirrors.

SymptomMost likely causeFirst thing to try Regular stripes in fills, follows fill directionHatch spacing or mark/jump delayTighten hatch spacing; adjust laser-on/off delay Faint offset duplicate, worse at high speedGalvo overshoot/ringingReduce scan speed; increase settle/jump delay Defect only near field edges/cornersF-theta lens field distortionCheck field calibration; verify correct lens for work area size General contrast loss plus bandingOptical path contaminationInspect and clean protective window and galvo mirrors

Galvo systems reward a methodical, one-variable-at-a-time diagnostic approach precisely because so many of their parameters interact — speed, delay, and hatch spacing all affect the same visible fill quality, and changing two at once makes it nearly impossible to tell which one actually fixed (or caused) a defect. Isolate before you tune, and keep a record of your working baseline settings per material so a future defect has a known-good reference to diff against.