My Laser-Cut Pieces Never Fit Together — Then I Learned the Word 'Kerf'

My Laser-Cut Pieces Never Fit Together — Then I Learned the Word 'Kerf'


Par Miles Harper
4 min de lecture


The beam was eating my tolerance

For six months I blamed my measuring. I'd design a press-fit box in software, send the job to my desktop laser engraver, and every time the joints came back just a hair too loose. Tabs wobbled in their slots. Corners gaped. Small pieces fell straight through the holes they were supposed to lock into. I re-squared my stock, re-checked my dimensions, re-cut the same design three times — and the fit still wasn't right. Then a friend at a maker meetup said a single word that rewired how I think about cutting: kerf.

Kerf is the width of material the laser beam vaporizes as it moves through the stock. It sounds insignificant — a fraction of a millimeter — but on a laser engraver for wood, that sliver is a real physical gap you introduce on every single cut line. When two parts are meant to interlock, the beam removes material from both sides, and your supposedly perfect fit is suddenly loose by twice the kerf width. You didn't measure wrong. The machine is quietly eating your tolerance.

Measure it once, stop guessing

The first fix was to stop guessing and measure my actual kerf. I cut a ten-millimeter square, then measured both the hole it left behind and the slug it dropped out. The difference between the two, split in half, is your kerf. On my machine with three-millimeter basswood plywood, it came out to about 0.2 millimeters. That's roughly the thickness of two sheets of printer paper — and it's more than enough to make a press-fit joint fall apart.

If you're just starting out, none of this needs a fancy setup. A pair of cheap calipers, a scrap of your favorite plywood, and five minutes at the bench will give you a kerf number you can trust. The mistake most beginners make isn't bad equipment — it's assuming the beam cuts exactly where the line is drawn.

Build the offset in

Once I had the number, the solution was simple but slightly counterintuitive: I stopped drawing parts at their final size. Instead I built the allowance into the design — tabs a touch bigger, slots a touch narrower — so that when the beam removed its 0.2 millimeters, everything landed exactly where I wanted it. Most laser software calls this kerf offset or kerf compensation, and LightBurn has it built straight into the cut settings, so the toolpath adjusts itself instead of you hand-editing every dimension.

Then there's the subtle gotcha I missed for weeks: the offset points different directions depending on what you're cutting. On an outside cut, the beam eats into the material outside your part, so you compensate outward. On an inside cut — like the slot a tab slides into — the beam eats inward, so you compensate the other way. Mix those up and your tabs get skinnier while your slots get wider, which is exactly backwards. Getting the direction right matters as much as getting the number right.

Here's the part nobody warned me about: kerf isn't a fixed number. Denser plywood cuts wider than soft basswood. A slow pass or multiple passes removes more than a single fast one. Even a slightly out-of-focus beam spreads and takes a wider bite. The value I measured on a Monday morning wasn't quite right for the oak I ran that Friday. So I keep a sticky note of kerf numbers for every material I cut regularly, and I re-measure whenever I switch stock.

I saved my kerf offset as a named preset per material — “3mm basswood,” “3mm birch,” “5mm oak” — so I'm not re-typing numbers from memory every time. It feels like overkill until you realize a typo in a single offset box is the difference between a drawer that slides and one that jams.

The snap test

I also stopped trying to nail it with one expensive test. Now I cut a tiny calibration comb — a scrap strip with notches at 0.18, 0.2, and 0.22 millimeters — and press two offcuts into each until one snaps together cleanly. That one-minute check has saved me more boards than a week of re-measuring ever did. The goal isn't a number on a screen; it's that satisfying click when two pieces lock together without glue.

Kerf wasn't the only thing sabotaging my cuts, by the way. For a long time I also blamed charred, ugly edges on the wrong settings, when the real culprit was something entirely different — and entirely fixable. Figuring that out is what finally made my cut pieces look as clean as they fit.

Learning kerf changed what I'm willing to design. I used to avoid interlocking projects — puzzle boxes, press-fit lamps, flat-pack organizers — because the joints never held. Now they're my favorite things to cut. The difference between a pile of parts that rattle apart and a kit that clicks together like a well-made toy is, most of the time, a single number sitting quietly in the software settings.

If you're chasing that same satisfying click, the lesson is short: measure your kerf, build the offset into your files, and re-check it every time you change material. I do my interlocking work on a Laservii L1 Pro, and once I dialed in the offset, the loose-fit mystery disappeared for good.