Laser doe ablation
Skill victorzhu-eng/industrial-laser-principles/skills/laser-doe-ablation
Step-gated, reference-grounded Claude skills for laser process engineering — DOE for ablation/dicing, diagnostics, and operational SOPs.
npx -y skills add victorzhu-eng/industrial-laser-principles --skill laser-doe-ablationAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Build a 3-step ablation-recipe DOE (surface threshold → depth line matrix → volume hatched / multi-pass) with the operator, gated at each step. Use when the user says "DOE for ablation", "ablation recipe", "I need to mark/scribe/drill X material", "set up a marking parameter sweep", or otherwise wants to develop an ablation recipe.
SKILL.md
7.8 KB, as published. Nobody here has run it
Laser Ablation DOE
Operates under the rules in
../../INSTRUCTIONS.md(reference-first, reference-only, interactive mode).
When to use this skill
When the operator needs to develop a new ablation recipe — for a new material, new wavelength, or new geometric target. This skill enforces the surface → depth → volume progression from laser_ablation.md §12. Each stage has a defined experiment, and each experiment pauses for the operator to run it and report results before the next stage begins.
What this skill is NOT
- Not for cutting all the way through (use
laser-doe-dicing). - Not for joining (welding skill not built — yet).
- Not for diagnosing an existing recipe gone wrong (use
laser-defect-diagnose).
References to load at start
references/laser_ablation.md§4 (threshold), §9 (multi-pulse), §12 (workflow) — primaryreferences/laser_process_calculations.md— fluence, peak intensity, focal spotreferences/gaussian_beam_theory.md— M², spot size derivationreferences/laser_optics_selection.md§5 — damage-threshold sanity check on chosen fluence
Inputs to gather first
Ask the operator:
- Material — composition, surface state.
- Goal — surface mark / shallow scribe / depth groove / drilled hole / area cleaning?
- Tolerance constraints — kerf width, depth target, HAZ tolerance, throughput target.
- Available laser — wavelength, max power, PRF range, pulse-width range, M².
- Optical chain — focus-lens focal length, beam expander mag, focal-spot diameter (or input beam diameter so we can compute it via
laser_process_calculations.md). - Throughput requirement — parts per hour or scan-speed target.
Compute and show derived values to the operator:
- Focal-spot diameter (from
gaussian_beam_theory.mdformula). - Maximum single-pulse fluence at full power.
- Average fluence at mid-PRF.
- Whether any of these exceeds optic damage thresholds (per
laser_optics_selection.md§5).
→ Pause: confirm all inputs and derived values. Address any damage-threshold flags before proceeding.
Workflow — 3 gated stages
Stage 1 — SURFACE: single-shot threshold
Per laser_ablation.md §12 Step 1.
What to do
- Set parameters for non-overlapping single shots:
- Low PRF (e.g., 1 kHz).
- High scan speed so pulse-to-pulse spacing >> focal-spot diameter (e.g., 100 mm/s).
- This keeps rise-time tail effects (
scanner_optimization.md§6) out of the threshold reading.
- Run a fluence sweep — typically 5–7 shots at logarithmically-spaced energies, bracketing the literature/estimated threshold.
- Inspect the resulting craters under microscopy.
What the skill computes
- Required pulse-spacing for true non-overlap.
- Suggested fluence sweep range:
- From any literature value the operator has (cite source explicitly — note that literature values are outside the reference docs and must be flagged as such).
- From general material-class anchors when literature isn't available.
- Test-pattern parameter list (PRF, speed, pulse energies for each shot).
After running — Liu plot threshold extraction
Per laser_ablation.md §12:
- Operator measures crater diameter at each above-threshold fluence.
- Compute (crater diameter)² vs. log(fluence).
- Linear extrapolation crosses the x-axis at the ablation threshold fluence.
- Cross-check against any literature value.
The Liu plot can be built in the user's existing Process Parameters Calculator Liu Plot sheet, or computed by hand here.
→ Pause: confirm threshold value. Sanity-check against literature if available. Confirm to proceed to Stage 2.
Stage 2 — DEPTH: single line, log-spaced matrix
Per laser_ablation.md §12 Step 2.
What to do
- Use the Stage 1 threshold as the lower-bound fluence anchor.
- Build a Power × Scan Speed log-spaced matrix.
- Mark the matrix as a grid of single lines on the same coupon.
Log-spaced series
Per
laser_ablation.md§12 — use logarithmic spacing because the eye easily distinguishes 0.5 m/s from 1 m/s, but cannot reliably distinguish 4.5 m/s from 5 m/s.
The 1.6× series:
0.1, 0.16, 0.26, 0.41, 0.65, 1, 1.6, 2.6, 4.1, 6.5, 10, ...
- Power axis: span ~ 1× to ~ 5× threshold fluence (typically 4–6 levels).
- Scan-speed axis: span useful range (typically 4–6 levels).
What the skill computes per cell
- Fluence (J/cm²).
- Pulse-to-pulse pitch in µm.
- Multiple of ablation threshold.
- Optic damage check — flag any cell that exceeds the lens damage-threshold limit (with
laser_optics_selection.md§5 + 2–10× safety factor).
Output
Generate the matrix as:
- CSV for now (simple, clean).
- Or markdown table for quick review.
- Excel output via shared library when
lib/ppc_writer.pyis built — eventually matching theProcess Parameters Calculatorcolumn layout.
→ Pause: review the matrix before committing. Operator can edit cells, drop ones flagged as out-of-envelope, or expand the range.
After running
Operator inspects each line for cut quality, kerf, HAZ.
→ Pause: ask the operator to identify the best-quality cell at the highest scan speed that meets the requirements. That cell becomes the unit recipe for Stage 3.
Stage 3 — VOLUME: hatched area / multi-pass
Per laser_ablation.md §12 Step 3.
What to do
- Take the unit recipe from Stage 2.
- Apply across an area by stacking parallel lines at hatch distance (typically 0.5× to 1× spot diameter — same overlap math as
laser_ablation.md§9). - For depth targets exceeding what one sweep clears, stack passes: 2× / 4× / 8× repeats.
What the skill computes
- Hatch overlap factor at the chosen hatch distance.
- Estimated depth per pass (calibrated from Stage 2 line cross-section).
- Estimated cycle time per area for each repeat count.
After running — the heat-accumulation diagnostic
Per laser_ablation.md §12 — Diagnostic.
Plot achieved depth vs. (number of repeats × inverse scan speed):
| Relationship | Meaning | Action |
|---|---|---|
| Linear | Heat-tolerant process — doubling repeats halves speed-to-target | Use the highest-throughput cell. Done. |
| Non-linear (diminishing returns) | Strong thermal accumulation | One of: add Pause between passes (test 0.5 s, 1 s, 2 s); or lower per-pass power (drop one cell on the matrix). |
→ Pause: operator reports which case it is. Loop with adjustment, or finalize.
Final output
The skill produces:
- DOE matrix file — Stages 1, 2, 3 each as a sheet (CSV with stage labels for now; xlsx when ppc_writer is available).
- Recipe summary (markdown) — final selected cell with all derived parameters, threshold value, validation samples, reference citations.
- Reference index — which reference section drove each decision.
Save to outputs/laser-doe-ablation/<YYYY-MM-DD>-<material>/.
Cross-skill links
- If a defect appears in Stage 2 or 3 that the matrix can't explain →
laser-defect-diagnose. - If power readings drift during the experiment →
laser-power-measurement(sanity check). - If an optic damage check fires in Stage 2 → stop and warn (a
laser-damage-sizerskill is planned but not built; for now, hand back to the operator with the warning). - If position-specific defects appear (corner burn-in, line-end short) →
laser-delay-tuner.