Home › Guides › How to find laser settings for a material…
How to find laser settings for a material nobody has charted
Power runs 10–100%. Speed runs 1–600 mm/s. Guess both at once on a €20 blank and you will spend more on ruined stock than the job is worth. This is the method that gets you to a working setting in two or three test pieces — and none of them has to be the expensive one.
Price a job in your browser →Why guessing is worse than it feels
The search space is the whole problem. A typical machine gives you roughly 90 usable power steps and several hundred usable speed steps. Landing on a good combination by intuition, on the first try, on a material you have never run, is not a skill issue — the odds are simply bad. Every experienced operator you have watched do it instantly is recalling a number they already tested once and wrote down.
So the goal is never “guess better”. It is to spend your uncertainty on cheap material and arrive at the expensive blank already knowing the answer.
Step 1 — find a proxy you can afford to ruin
Most materials have a cheap stand-in that behaves closely enough to get you into the right range. You are not looking for an identical result, just the right neighbourhood.
| What you actually need to run | Test on this first | How close it gets you |
|---|---|---|
| Powder-coated tumbler / flask | Powder-coated steel offcut, or a cheap coated tumbler from the same seller | Very close — coating chemistry matters more than the object shape |
| Finished hardwood board | Same species scrap, sanded the same way | Close; finish and grain shift the result |
| Anodised aluminium tag | Any anodised offcut in the same colour | Close — colour changes the contrast far more than the alloy does |
| Cast acrylic sheet | Offcut of the same sheet, or same brand and thickness | Near identical if it is the same casting |
| Leather goods | Scrap from the same hide | Close; veg-tan and chrome-tan behave very differently, so match the tannage |
| Slate coaster | Any slate offcut | Close; slate varies a lot batch to batch, so still corner-test |
When you buy blanks, buy one spare per new product and mark it as the test piece. It is the cheapest insurance in the shop — one ruined €4 blank instead of a ruined €25 one.
Step 2 — bracket, don't crawl
Start from the closest published number you can find rather than from zero. Our power and speed chart and the per-machine pages give you a starting point for your machine class; take that number and test half to double the speed across a power band around it. A material test grid does this in one burn.
If every square on your grid looks the same, your range was too narrow. Widen it and re-run — a grid that tells you nothing cost you ten minutes; a wrong setting on a real job costs you the blank.
Step 3 — the corner test
Once the proxy gives you a candidate setting, you still do not run the full job. Run a small mark — a 5 mm square, a couple of characters — on a part of the real blank that will not be seen: the base of a tumbler, the back of a tag, inside the waste area of a cut, under where the label goes.
This catches the thing a proxy cannot: this specific blank's coating thickness, curvature, or finish. It takes fifteen seconds and it is the last gate before you commit.
Step 4 — change one variable at a time
When the corner test is close but not right, resist changing three things at once. The usual order:
- Focus first. Nearly every “wrong settings” result on a curved or unusually thick item is actually a focus problem. On a tumbler, focus on the surface at the point the beam strikes, not the tangent.
- Then speed. Speed changes the result more predictably than power and does not stress the tube or diode. Too dark? Speed up. Too faint? Slow down.
- Then power, in 5–10% steps.
- Then passes or line interval, last. More passes on an engrave usually means you should have slowed down instead.
Step 5 — write it down the moment it works
The setting is worth more than the part. Record machine, material and supplier, thickness, power, speed, passes, line interval, air assist, focus offset and the date. A note on your phone is fine; a shared sheet is better if anyone else touches the machine.
Two things make a recorded setting stop working later: a new batch of material, and a dirty or aged lens. Both are worth checking before you assume you wrote the number down wrong.
Materials where the answer is “don't”
No setting is the right setting for some materials. PVC and vinyl release chlorine gas that will corrode your machine and harm you. Polycarbonate, ABS, fibreglass, carbon fibre and any unidentified plastic belong in the same category. If you cannot positively identify a plastic, do not put it in the machine — see what a laser can and cannot cut.
Once you have the setting, price the job
A known speed is also what makes a job quotable. Drop the file into the calculator, pick your machine, and it converts cut length and engrave area into machine time and a price. If you are working out what your hour is worth in the first place, see how to price laser engraving.
FAQ
How do I find laser settings for a material with no chart?
Do not test on the expensive item. Find a cheap proxy that behaves similarly - a powder-coated offcut for a tumbler, same-species scrap for a finished board - and run a material test grid on that, bracketing half to double the speed of the closest published setting for your machine class. Take the winning setting to a hidden corner of the real blank for a 5 mm confirmation mark, then run the job.
Why can't I just guess the power and speed?
Because the search space is large. A typical machine offers around 90 usable power steps and several hundred speed steps, so hitting a good combination first try on an unfamiliar material is mostly luck. Operators who appear to do it instantly are recalling a number they tested once and recorded.
What should I change first when a test is close but not right?
Focus, then speed, then power, then passes. Focus explains most bad results on curved or unusually thick items. Speed changes the outcome more predictably than power and does not stress the tube or diode: too dark means speed up, too faint means slow down. Adjust power in 5-10% steps, and treat extra passes as a last resort.
My settings worked last month and don't now. What changed?
Usually the material batch or the lens. Plywood glue lines, moisture, veneer colour, acrylic cast vs extruded and leather tannage all shift how much energy reaches the material. A lens that needs cleaning, or a diode a few hundred hours older, does the same. Re-run a small grid on the new batch before assuming the recorded number was wrong.
How much material should testing cost me?
One spare blank per new product. Buy an extra when you order, mark it as the test piece, and keep it with your notes. That converts an unpredictable risk - ruining a 20-25 EUR customer blank - into a fixed few-euro cost you only pay once per material.
Are there materials I should never test settings on?
Yes. PVC and vinyl release chlorine gas that corrodes the machine and is harmful to breathe. Polycarbonate, ABS, fibreglass, carbon fibre and any plastic you cannot positively identify are also off limits. There is no safe setting for these - the correct answer is not to run them at all.