Laser Cut Settings for Plywood: Start Here, Then Test

Everyone starts the same way: you find a number somebody posted, type it in, press go, and the cut comes out half through with a black smear around it. Then you find another number, and it burns. The frustrating part is that both numbers were probably correct — on somebody else’s machine, with somebody else’s sheet.

This page gives you starting points that are worth typing in, and then the part that actually matters: what to change, in what order, when the first cut is wrong.

The one rule that comes before every number

Every setting below is a starting point, not an answer. The same 3 mm plywood cuts differently on two machines in the same room: different lens, different tube age, different air pressure, different humidity in the sheet. Every source that publishes settings — including the machine manufacturers — says the same thing, so it is not a disclaimer, it is the method.

Run a test cut on an offcut of the exact sheet you are about to use. Not the same brand. The same sheet.

What is actually happening in the cut

A laser does not slice wood, it burns and vaporises a narrow line of it. Everything else follows from that.

The cut has width. The burnt-away strip is called the kerf, and on plywood it runs about 0.2–0.3 mm. So a cut part comes out slightly smaller than drawn, and a hole slightly larger. On a box with finger joints that difference decides whether the joint is tight or loose.

The cut is not a perfect rectangle. The beam converges to a point and diverges again, so on thick material the top of the cut is wider than the bottom. That is taper, and on thick sheet it can be enough to make a part jam in its own hole.

Smoke gets in the beam’s way. Combustion products absorb energy and stain the edge, which is why air assist stops being optional.

A dark edge is normal. Some char always happens. The question is whether it is a thin line or a wide dirty band.

Focus beats power. For 3 mm plywood the beam should be focused on the surface of the material. A lost focus is one of the most common reasons a machine “stopped cutting through”, so check it before you blame a worn tube.

The sheet matters more than the settings

This is the part beginners skip, and it costs them a weekend.

Cheap construction plywood breaks the process in three ways. Voids — air pockets between the veneers; the beam crosses the gap with nothing to burn and then hits the next layer with less energy, so the cut stops halfway along an otherwise perfect line. Uneven glue — the glue line is harder than the wood, and in construction sheets it is laid unevenly, so settings that cut one sheet cleanly leave an uncut corner on the next sheet from the same pack. Knots and density changes — a hard knot slows the beam exactly like a glue line.

So when a cut wanders, suspect the sheet before the machine.

What people with clean results use is Baltic birch: even veneer thickness, even glue, predictable burn. Basswood is softer and cuts on weaker machines. Bamboo is very consistent, though it usually costs more. Poplar and okoume are cheap and soft with more variation. Laser-grade plywood is sold specifically for this — fewer voids, glue chosen to burn predictably, often with an E0/E1 emission class, which means lower formaldehyde emission from the glue.

Avoid anything varnished, filmed or pre-finished: it burns badly and releases things you do not want to breathe.

Grades and thickness

Grades, if you are buying Baltic birch: the grade is written as two letters, face over back. B/BB is the usual “for laser” grade — clean face, a few oval patches on the back. BB/BB has patches on both sides. C allows open knots, veneer cracks and voids in the core — this is construction grade, and it is the sheet that produces the mystery uncut sections.

Measure the thickness. Plywood thickness is nominal. It varies between sheets and along one sheet, and slot-fit guides report differences of half a millimetre and more. If your design has slots, measure the actual sheet with callipers before you cut, not after.

MDF is a different animal: no layers, no voids, so it cuts more predictably than plywood — but it releases noticeably more formaldehyde and wants a carbon filter, and being dense it burns readily if the beam ever stops moving.

What each machine can realistically cut

| Machine | Practical ceiling in plywood | |—|—| | 5 W diode | 2–3 mm, with difficulty and many passes | | 10 W diode | 3 mm reliably, good sheet only | | 20–40 W diode | 3–6 mm | | 40 W CO2 | 6 mm in one pass | | 50–60 W CO2 | up to 10 mm | | 80–150 W CO2 | 10–18 mm, slowly and with taper |

Past 18–20 mm a laser stops being the sensible tool for plywood: time, soot and taper eat the advantage.

Fibre lasers belong to metal. On wood they are a poor choice and carry a higher fire risk.

Starting points: diode lasers, 3 mm plywood

Speeds here are in mm per minute, with air assist on.

| Power | Power setting | Speed | Passes | Notes | |—|—|—|—|—| | 5 W | 90% | 150–200 | 10–15 | barely worth attempting, do not leave it alone | | 10 W | 90–100% | 250–300 | 4–6 | the realistic minimum, and only on a good sheet | | 20 W | 85% | 350 | 2–3 | the reference point most people land on | | 33 W | 75–80% | 500 | 2 | power comes down to keep the edge clean | | 40 W | 75% | 600 | 1–2 | hold it at 70–80% |

Without air assist the same machines have to go slower and produce far more soot — a 10 W module drops to around 200 mm/min and 6–10 passes.

If you like a single number to sanity-check against: multiply the watts by the power setting and by 60, then divide by the speed in mm/min. That gives joules per millimetre for one pass. The 10–40 W rows above work out to roughly 2–3 J/mm per pass, and the total for a cut is that figure times the number of passes.

Starting points: CO2, plywood

Speeds in mm per second, power around 90%.

| Thickness | 38 W RF | 55 W | 80 W | 100 W | 130 W | 150 W | |—|—|—|—|—|—|—| | 3 mm | 10 | 40 | 50 | 52 | 55 | 60 | | 6 mm | 5 | 8 | 15 | 20 | 25 | 30 | | 10 mm | 5 | 5 | 10 | 13 | 18 | — | | 15 mm | — | — | 1 | 2 | 3 | 3 | | 18 mm | — | — | — | 1 | 2 | — |

Two independent points to cross-check against: 3 mm Baltic birch on a 40 W CO2 usually lands at 60–100% power and 20–30 mm/s, and the manufacturer tables for 3 mm plywood give 30 speed / 60 power at 1000 Hz, with 10 / 100 / 1000 for 6 mm.

The spread between sources is wide. That spread is exactly why the test cut exists.

Three things that matter more than the numbers

Air assist is not an accessory. It blows smoke and sparks out of the kerf, puts out flare-ups, keeps the edge clean and lets you cut thicker at the same power. Typical pressure is 10–20 psi on a diode, 30–50 psi on a CO2 when cutting.

Faster with more passes beats slow in one pass. This is the single best trick against a charred edge. When a cut does not go through, add a pass rather than lowering the speed — slowing down pours more heat into the same line and widens the burn.

Get the sheet off solid metal. Heat pools where the material touches the bed and scorches the back. A honeycomb bed helps, but it also reflects some energy back up and prints a honeycomb pattern on the underside, so the real goal is less contact area: a knife bed, pin supports, magnets, or a strip of scrap under the part.

Reading a bad cut

| What you see | Usual cause | What to do | |—|—|—| | Cuts through in some places, not others | voids, glue line, a knot | change the sheet, add a pass, check focus and a dirty lens | | Does not cut through anywhere | focus is off, dirty optics, or you are at the machine’s ceiling | clean the optics and refocus before touching power | | Wide black band along the edge | too slow at high power, no air assist | go faster with more passes, turn on air assist | | Soot film on the face | smoke settling around the kerf | mask the face — and the back — with painter’s tape before cutting |

Notice the order in that second row. Power is the last thing to change, not the first.

How to run the test properly

Draw a row of small squares — 20 mm is plenty — and cut each one with a different pair of settings, labelled. Cut them on the actual sheet, in the actual place on the bed you will be using, with the same air assist and the same masking you plan to use for the real job. Then push each square out with a finger.

The one that falls out with no force, with the thinnest dark line and no flare marks on the face, is your setting for that sheet. Write it on the offcut with a marker and keep the offcut. Next time you buy the same plywood from the same supplier, you start from that number instead of from a forum post.

Something to cut while you test

Test squares are a waste of good plywood if that is all you cut. Files linked on this site are free to download with the Creative Fabrica free trial (see the note at the bottom), so use a small one as your test piece — a keychain, a single ornament, an earring pair. Small parts show up bad settings faster than big ones anyway: the fine details fill with soot first, and the tiny islands are what fall out when the edge is over-burnt.

Good candidates: wooden earrings — small, thin, and unforgiving of a wide kerf; Christmas ornaments — flat, layered, and forgiving if the first attempt goes dark; night lights and lanterns, where the light passing through the cut-outs shows you exactly how clean the edges really are.

And when the settings are dialled in, the whole catalogue is here: free laser cut SVG files for wood.

Cutting words and names from plywood? The lettering matters as much as the settings: see how to choose fonts for laser cutting and engraving.

Before you load a sheet, the laser cut sheet planner shows how many parts fit on it and what happens to a slot when you resize a design.

Good to know: the files in this guide are hosted on Creative Fabrica. Downloading is free with their free trial: a payment card is needed to start it, and you can cancel the subscription at any time.