Nov 18,2026
A 1.8 m aluminium panel, 0.6 mm thick, with 0.3 mm of allowed flatness deviation. The question is not whether the laser can clean it — it is whether the heat budget allows it.
Two ways to lower E: lower power or raise speed. Raising speed is free; lowering power costs throughput. Every aerospace parameter set is a negotiation between those two, and the answer is usually "both, a little".
There is a third lever most buyers miss: beam shape. A line beam 20 mm long spreads the same energy over a larger area per unit time, which lowers peak intensity at any one point while keeping the area rate respectable. On a 0.6 mm skin, a 20 mm line at 200 W outperforms a 0.06 mm spot at 150 W — same energy density, four times the coverage per pass. This is why aerospace heads look different from the handheld units sold to job shops.
| Parameter | Range that works | Why |
|---|---|---|
| Average power | 100–200 W | Above 200 W the required scan speed exceeds galvo stability |
| Pulse duration | 80–150 ns | Short enough to ablate, long enough to avoid plasma shielding |
| Repetition rate | 150–300 kHz | High PRF = small energy per pulse = less thermal load per spot |
| Scan speed | 2,000–4,000 mm/s | The main lever for staying under 0.8 J/mm² |
| Overlap | 20–35% | Above 40% the second pass re-heats the first |
| Standoff | ± 1.5 mm from focus | Defocus spreads energy but widens the heat-affected band |
Table 1 — Parameter window for 0.6 mm Al 2024-T3 with an anodised or primer layer. Every number is a ceiling, not a target.
Three energy densities on identical 300 × 300 mm coupons, 0.6 mm, clamped on a granite surface plate, measured with a dial indicator on a 100 mm span before and after.
| Energy density | Flatness change | Surface result | Verdict |
|---|---|---|---|
| 0.4 J/mm² | 0.08 mm | Primer removed, 2 passes needed | Safe, slower |
| 0.8 J/mm² | 0.26 mm | Clean in 1 pass | At the limit |
| 1.6 J/mm² | 1.10 mm | Clean, visible oil-canning | Scrap |
Table 2 — Measured flatness change against energy density. The curve is not linear: doubling from 0.8 to 1.6 quadruples the distortion.
| Method | Time per m² | Material removed | Waste stream | Distortion risk |
|---|---|---|---|---|
| Chemical stripping | 45 min + rinse | 0 | Solvent, chromate, rinse water | None |
| Manual sanding | 90 min | 0.05 mm of skin | Dust, discs | Local thinning |
| Laser, 0.4 J/mm² | 18 min | 0 | Filter dust only | 0.08 mm |
Table 3 — Three methods on the same coupon. Laser is 2.5× faster than chemistry and removes no parent material, which is the number that matters for a repairable skin.
Three limits, stated plainly, because aerospace buyers test them first:
| Case | Why it fails | Alternative |
|---|---|---|
| Skin under 0.4 mm | Even 0.4 J/mm² moves it; the tolerance is smaller than the measurement noise | Chemical strip only |
| Carbon-fibre composite | Resin ablates before the contaminant; fibre damage is invisible | Peel ply, solvent wipe |
| Deep scribe or corrosion pit | Laser cleans the surface, cannot restore section thickness | Blend out, then laser clean |
Table 4 — The three cases we decline. A supplier who quotes all three has not run a coupon.
You do not need our lab to find your number. You need a dial indicator, a granite plate and three coupons. It takes an afternoon and it settles the purchase before it is made.
Step one. Clamp a 300 × 300 mm coupon exactly the way the real skin is clamped — same fasteners, same torque, same support spacing. Distortion is a function of restraint, so an unclamped coupon will lie to you in the optimistic direction.
Step two. Measure flatness on a 100 mm span at nine points, write the numbers down, then clean at the setting you intend to use. Re-measure at the same nine points. The delta is your distortion, not the absolute reading.
Step three. If the delta exceeds a third of the drawing tolerance, drop the energy density by 30% — raise scan speed first, because it costs nothing but time — and repeat. Two iterations usually find the window; three means the panel is too thin for the process.
One more measurement worth taking: the reverse side. Aluminium conducts heat well enough that a 0.6 mm skin can show a faint temper colour on the back face even when the front measures flat. If the back is visible in the assembly, that colour is a rejection criterion in its own right, and it appears at roughly 60% of the energy density that causes measurable distortion.
A 0.6 mm skin has a 0.3 mm distortion allowance and a 0.8 J/mm² heat budget. Stay under both by running 100–200 W at 3,000 mm/s with 25% overlap, and expect 0.08 mm of movement at the safe setting.
Above 1.6 J/mm² the panel is scrap. Below 0.4 mm thickness there is no safe setting at all. Those two numbers decide whether laser is the right tool before any brochure does.