The machines' words — one line, one source each. 34 terms, from the app's tested tables.
The width of material the beam burns away. Measure it with ten squares: K = S − M/N. Every edge of a cut loses K/2.
LightBurn's compensation: K/2, paths move outward on outer shapes and inward on holes — exactly what the cut removes.
The space between scan lines. DPI = 25.4 ÷ interval. The app's default 0.1 mm = 254 DPI, inside LightBurn's 120–300 window.
How a colour photo becomes one grey channel: 0.299 R + 0.587 G + 0.114 B (ITU-R BT.601).
Dithering that spreads each pixel's rounding error to its neighbours: Floyd–Steinberg ÷16, Jarvis–Judice–Ninke ÷48, Stucki ÷42, Atkinson 6×⅛.
A fixed pattern of thresholds at (index + 0.5)/64 of full scale — the screen-door look.
A grid of sized dots, screen angle 45° (the print standard); its density in lines per inch (lpi).
Alternate lines run opposite directions, killing the one-direction diagonal drift.
E = W · (p/100) · N / v in joules per millimetre. The app's fire caution lives at 1 J/mm.
Modules are sold by electrical input; the light out is several times less. Enter optical in the energy math.
The beam's width in the focus: 2 · (2·M²·λ·f / (π·D)). A 2" CO₂ lens example gives 0.118 mm.
The range the taper stays acceptable: 2·π·w0² / (M²·λ) — 1.71 mm for that same lens.
How far a real beam sits from perfect Gaussian; from the tube's data sheet, with the beam diameter.
CO₂ at 10.6 µm cuts clear acrylic; fiber at 1.064 µm marks metal; diodes (blue, ~450 nm) cannot do either task fully. The app refuses cross-type conversion.
PVC/vinyl/artificial leather (chlorine), polycarbonate, ABS, HDPE, the foams, fiberglass/FR4/G10. Not listed ≠ safe — read the safety data sheet.
The material's own chemical truth. The app's standing line: "a material not on this list isn't automatically safe; check its safety data sheet."
Ten slots from T−0.25 to T+0.20 mm in 0.05 steps, cut with the offset OFF; the slot that fits hands you the clearance.
Designed slack for a fitting part. No defaults exist: zero is a press fit, the gauge measures the rest.
Fingers per edge = the odd n ≥ 3 nearest your target; clearance applied as c/4 per boundary; odd counts keep both edge-ends on one panel.
Grow every outline by K/2 and shrink holes by K/2, so drawn files cut to true size.
A solid w×T strip twisted θ spans w·cos θ + T·sin θ; links needed = bend ÷ θ. Fenner's example: 3 mm sheet, 90°, 23 links.
A bank card: exactly 85.60 × 53.98 mm everywhere on Earth — the ruler in the bed photo.
Four marked bed corners give the photo-to-bed transform the nesting needs; tested to 0.2 mm on synthetic tilts.
The run-out beyond each raster line's edge; the raster = lines · (width + 2·overscan) / speed, doubled on one-way scanning.
The head returns without burning — at that return's cost; the app doubles raster time as the estimate.
Estimated ÷ real job time, per machine (525 s estimated → 600 s real = ×1.143).
A spool's empty weight; weigh total minus tare = grams of plastic left.
Mass per metre: the 1.75 mm cross-section × density = 2.98 g/m PLA → a kilo ≈ 335 m.
The material's drying temp and hours (PLA 50 °C/8h … nylon's honest estimate). PCL pen filament is never oven-dried.
The dryer's partner: bag a warm finished spool with it, or print straight away.
The line of plastic a 3D pen lays — about 0.7 mm for a hot pen; it sets every stencil's lanes.
The one-bead band inside a colour's edge; the triangle/honeycomb/crosshatch support inside; the dots pinning the piece flat as it cools.
A print that leaves the bed already assembled — hinged, jointed, moving.
One layer's engrave seconds, the estimator's building block (each layer's problems surface beside it).