How to Create Laser Cut Files: A 7-Step Checklist from a Quality Manager Who's Reviewed 200+ of Them
- What this checklist is for
- Step 1: Lock down the material before you open your design software
- Step 2: Check what your machine can actually do—not what the listing says
- Step 3: Set file format and units before anything else
- Step 4: Design around the kerf, not on top of it
- Step 5: Test cut on scrap—and look at the edges
- Step 6: Nail the nesting and fixturing
- Step 7: Build your deadline buffer into the file review
- Common mistakes to watch for
What this checklist is for
I'm the quality and brand compliance manager at a laser equipment company. Every file that goes to production or customer samples crosses my desk—roughly 200 a year. Maybe 180, I'd have to check the system.
In Q1 2024, we rejected about 15% of first submissions. The surprising part? Most of those rejections had nothing to do with the machines. They were file problems. Wrong units. Missing kerf compensation. Materials that didn't match the settings.
This checklist is for anyone sending a design to a laser cutter—whether you're cutting wood signs on a CO2 machine, engraving metal plates, working through xtool f1 ultra gun engraving samples, or running a small job shop. Seven steps. Follow them in order. Skip any of them, and you'll probably be redoing the job.
Step 1: Lock down the material before you open your design software
This sounds obvious. It isn't. I've seen designers build an entire layout for "3mm acrylic" and then learn we only stocked 2.8mm. That 0.2mm difference changed the fit of every interlocking joint.
Write the material down with exact spec: species, grade, thickness, moisture content. "Wood" is not a material. "4mm Baltic birch plywood, B/BB grade, 8% moisture" is a material. For laser engraving materials, the same rule applies—leather weight, acrylic cast vs. extruded, anodized vs. bare aluminum. Each one behaves differently.
Same for metals. When you're running xtool f1 ultra metal cutting on the fiber source, 304 stainless and 316 behave differently at the same power setting. The spec sheet matters. So does the surface finish. Brushed steel marks differently than polished.
One more thing: if you're sourcing material yourself, order one extra sheet. Not for production. For testing. You'll thank yourself later.
Step 2: Check what your machine can actually do—not what the listing says
I reviewed a batch of marketing claims in early 2024 for our own product line. The lesson stuck: rated capability and repeatable production capability are not the same number. A diode laser that etches aluminum oxide coating is not the same as a fiber laser that marks bare steel. Same word—"engraves metal"—completely different process.
If you're using a dual-laser system like the xtool f1 ultra, know which head does what. The fiber source handles bare metals, deep engraving, and gun engraving on steel or aluminum. The diode handles wood, leather, acrylic, and coated metals. Mixing them up in your file plan costs time and material.
Before you commit a production run, run a 20-minute capability test: three materials, three power settings each, inspect the results under magnification. The scratches and burn zones will tell you more than any spec sheet. That step is boring. It also saved us a 300-piece acrylic order in 2023.
Step 3: Set file format and units before anything else
This is the step most people skip. In my first year reviewing files, I made the classic specification error: assumed "standard" meant the same thing to every vendor. It doesn't. Now I check units first, every time.
Here's the short version:
- Vector files for cutting: SVG, DXF, or AI. Raster files (JPG, PNG) are for engraving only, and their DPI determines quality—use 300+ DPI, no smaller.
- Set units to millimeters or inches—but declare which one, and make sure the file carries that unit natively. A DXF exported in inches and opened in millimeter space will be off by a factor of 25.4.
- Convert all text to outlines. A missing font turns a clean logo into a default serif mess.
- Close all cut paths. An open path becomes a score line, not a cut.
That step takes 90 seconds. Skipping it has cost me, personally, a $600 redo on a batch of acrylic display pieces. Not ideal.
Step 4: Design around the kerf, not on top of it
The laser removes material. That gap is called kerf—usually 0.1 to 0.3mm depending on the machine, lens, and material. If your file draws two pieces to butt perfectly edge-to-edge, they won't fit. They'll rattle or bind.
Three practical rules from our production line:
- For slots and tabs, offset by half the kerf on each mating side. A 3mm slot in 3mm material becomes a finger joint that actually holds.
- For holes, compensate inward. A 5mm pin needs a 5.2mm hole if you want it to move; 4.9mm if it needs to press-fit.
- Test with a scrap cut before the real run.
One detail that surprised me when I started: material moisture changes kerf. We ran a batch of 12mm plywood in August 2022—summer humidity—and the same file that worked in January was cutting 0.15mm wide. We had to re-cut 40 parts. Now we keep the material in a humidity-controlled area for 48 hours before production runs. Simple. Effective.
Step 5: Test cut on scrap—and look at the edges
I know. You've heard "always test." Here's what testing actually means at our shop: cut a 100mm straight line, a 10mm circle, and a 3mm slot. Then inspect three things:
- Edge appearance: charred or yellowed edges mean too much power or too slow a speed, especially on wood.
- Dimensional accuracy: measure with calipers. If your 10mm circle comes out 9.8mm, your kerf offset needs adjusting.
- Cut-through: lift the piece. No fibers holding on, no scoring on the backing surface.
For metal cutting on the xtool f1 ultra, our test adds a fourth check: heat discoloration on the back side. If it's visible, the parameters are running too hot and you'll warp thin stock.
Never expected the cheap material to outperform the expensive one on our first comparison. Turns out moisture content mattered more than price. We kept the cheaper supplier and added a drying step. That was 2022. Still using them.
Step 6: Nail the nesting and fixturing
Two files look identical on the canvas and perform completely differently in the machine. The difference is nesting and fixturing.
Keep parts at least 3-5mm apart—tighter than that and thermal stress can crack thin sections. Nest close to the origin (top-left) so if a cut fails halfway, you're not remaking 20 parts when 4 would do.
For small parts, use tabs or double-sided tape. We had a run of 2mm acrylic charms in 2023 where loose parts migrated under the laser head and got re-engraved with a stray line. Cost us the whole batch—maybe 400 pieces, give or take. Now every small-part file gets a tab check before we send it.
Step 7: Build your deadline buffer into the file review
Here's the part nobody puts in a design tutorial: when the delivery date is firm, the file review schedule becomes part of the job. In March 2024, we paid $400 extra for a rush shipment of metal plaques. The alternative was missing a $15,000 event. That wasn't a speed premium. It was a certainty premium.
What I do now: add the inspection and re-cut window to the schedule before confirming a date. If a customer says "we need it Friday," and my file hasn't passed Steps 1-6, we're already at risk. I'd rather tell them "next Wednesday, guaranteed" than "probably Friday" and hope.
A cheap, uncertain delivery costs more than a firm, slightly pricier one. Every time. Even after choosing the rush option in March, I still second-guessed it for two days. Didn't relax until the shipment landed on time with clean edges. Worth it.
Common mistakes to watch for
Even a clean file can fail on these. Check them before you ship the job:
- Over-claiming material compatibility. FTC guidelines (ftc.gov) require that product claims be truthful and substantiated. Our marketing team learned to write "tested on X material at Y settings" instead of "works on everything." So should your quoting.
- Assuming the vendor has your material. Verify stock and thickness before you send the file. A file built for 6mm stock cannot be rescued if the shop only has 5.5mm.
- Not inspecting the first article. Compare the first offcut against your drawing with calipers. If it's out of tolerance, stop production. Don't "run a few more to see."
- Forgetting the shipping buffer. USPS First-Class Mail letters are $0.73 as of January 2025, and even a 100-piece small package takes 2-5 business days. If your deadline assumes same-week delivery, check the actual transit window before you promise anything (usps.com).
The checklist isn't glamorous. Follow it, and you'll reject fewer of your own jobs—which, honestly, is the only quality metric I care about anymore. Done.