Laser Level Accuracy: mm Per 10m by Class, and What That Costs You Across a Room
Accuracy is quoted per 10 m, but almost nobody works at 10 m. A ±3 mm/10 m level drifts 0.9 mm across a 3 m kitchen run and 3.6 mm across a 12 m ceiling. Divide the quoted figure by 10 to get millimetres per metre, then multiply by your actual span — and add the beam's own line width, which the spec never counts.
Divide the quoted figure by ten and you have millimetres per metre. Multiply that by the span you're actually working across and you have the number that matters. A ±3 mm at 10 m level — which is where most of the DIY shelf sits — gives you 0.3 mm per metre, so 0.9 mm across a 3 m cabinet run and 3.6 mm across a 12 m ceiling grid.
That's the arithmetic in full. What follows is the chart with the work done, plus the three errors the spec sheet quietly leaves out.
The chart, with the conversions done
Tier Typical quoted accuracy mm per metre Drift at 3 m at 6 m at 12 m at 20 m Self-levelling range Manual bubble-and-beam, no compensator ±10 mm at 10 m 1.00 3.0 mm 6.0 mm 12.0 mm 20.0 mm None — you level it Entry self-levelling cross-line, red ±4 mm at 10 m 0.40 1.2 mm 2.4 mm 4.8 mm 8.0 mm ±4° Mid DIY self-levelling, red or green ±3 mm at 10 m 0.30 0.9 mm 1.8 mm 3.6 mm 6.0 mm ±4° Trade 3 x 360 green ±2 mm at 10 m 0.20 0.6 mm 1.2 mm 2.4 mm 4.0 mm ±3-4° Premium 3 x 360, pro brands ±1 mm at 10 m 0.10 0.3 mm 0.6 mm 1.2 mm 2.0 mm ±3° Interior rotary ±1.5 mm at 30 m 0.05 0.15 mm 0.3 mm 0.6 mm 1.0 mm ±5°, motorised Exterior grade rotary with detector ±1 mm at 30 m 0.033 0.10 mm 0.20 mm 0.40 mm 0.67 mm ±5°, motorised Find your working span in the columns, then read up to the cheapest tier whose drift is smaller than the tolerance your job needs. Do this before you shop, not after. Note how flat the top of the table is in practical terms. Across a 3 m span, an entry-level unit is out by 1.2 mm and a premium unit by 0.3 mm — a difference of under a millimetre, on a job where you're marking to a pencil line. The gap only opens up as the span grows, which is exactly why the accuracy figure is quoted at a distance most people never use.
Three errors the spec never counts
Beam width comes first. A red line projected at 5 m is typically 2-3 mm wide and a green one 1.5-2.5 mm, and if you mark to the top edge at one end of the room and the bottom edge at the other you've just introduced an error larger than the instrument's entire tolerance. Pick one convention — centre of the line, always — and stick to it.
Mounting is second. A tripod that settles into a soft floor, a magnetic bracket on a flexing stud, or a pole clamp that slips a quarter of a turn will all beat the instrument's spec comfortably. On a long ceiling job, re-check your reference mark before the last run.
Temperature is third and least obvious. A pendulum compensator takes time to settle after you bring the unit in from a cold van, and readings taken in the first few minutes can drift. Give it five minutes in the room before you mark anything you can't undo.
Add these up and a ±2 mm/10 m level in real site conditions behaves like a ±3 mm one. That's not a defect. It's the difference between a lab figure and a delivered figure.
Green isn't more accurate, it's more visible
Nothing about beam colour changes the compensator, the optics or the tolerance. What changes is how bright the line looks to your eye, and that comes from the eye's own photopic sensitivity curve, which peaks around 555 nm and falls away steeply into the red.
The multipliers below are the ratio of those sensitivity values at equal optical power. They're the reason a 1 mW green line looks dramatically brighter than a 1 mW red one across a lit room.
Beam Wavelength Eye sensitivity, CIE photopic Apparent brightness at equal power Typical indoor working range Outdoors, no detector Red diode, budget units 650 nm 0.107 1.0x (baseline) 8-12 m 1-3 m Red diode, better units 635 nm 0.220 2.1x 10-15 m 2-4 m Green direct diode 515 nm 0.607 5.7x 15-25 m 3-8 m Green DPSS 532 nm 0.880 8.2x 20-30 m 5-10 m Any beam with a line detector n/a n/a Irrelevant — the detector sees what you can't n/a 30-60 m line, 100 m+ rotary Use the multiplier to decide whether green is worth the premium for your work, not to judge accuracy. If you work outdoors at any distance, the last row beats every row above it. Two caveats worth having. Direct green diodes at 515-520 nm run cooler and start instantly, while DPSS green at 532 nm is brighter but historically fussier in the cold — check the operating temperature range if you work in an unheated space over winter. And green units draw more current for the same run time, so budget for the battery.
Match the tier to the span, not to the price
Job Span you work across Error you can live with Minimum tier that gets there What goes wrong if you undershoot Kitchen wall unit run 3 m ±1 mm Trade 3 x 360 Doors line up at one end and show a wedge gap at the other Floor tile layout, large format 4-5 m ±2 mm Mid DIY Cumulative lippage; the last row won't fit the wall Suspended ceiling grid 12 m ±3 mm Mid DIY, or interior rotary for speed Grid visibly dips mid-room under lighting Stud partition plumb 2.4 m ±2 mm Mid DIY Plasterboard sheets don't meet flush at the joint Shower tray and wet room fall 1.5 m ±1 mm Trade 3 x 360, checked against a spirit level Water pools instead of draining Decking or joist levels 6 m ±3 mm Mid DIY Noticeable bounce and standing water Patio drainage fall at 1:80 8 m ±3 mm Interior rotary Fall reverses somewhere in the middle Shopfitting shelf run 5 m ±1 mm Premium 3 x 360 Shelf line reads as sloped against the ceiling Fence or wall line 20 m+ ±10 mm Exterior rotary with detector Nothing else is visible outdoors at that range Work out your longest regular span before you look at a single product. Everything above the tier this table names is money spent on precision the job cannot use. Check your own level in ten minutes
Every self-levelling unit drifts eventually, usually after a drop nobody admits to. The reversal test catches it without any equipment beyond a pencil.
- Find two parallel walls about 5 m apart. Set the level on a tripod roughly 0.5 m from wall A, facing along the span.
- Mark where the horizontal beam lands on wall A. Call it a1. Mark where it lands on wall B. Call it b1.
- Move the tripod to roughly 0.5 m from wall B, facing back the other way. Adjust the tripod height until the beam lands exactly on b1.
- Mark where the beam now falls on wall A. Call it a2.
- The gap between a1 and a2 is twice the error over that 5 m span. Halve it, divide by 5, multiply by 10, and you have your unit's true millimetres per 10 m.
- For the vertical line, do the same trick against a doorway: mark the beam top and bottom, rotate the unit 180 degrees in place, and see whether the line still covers both marks.
If the measured figure is worse than the quoted spec by more than about 50%, it's a calibration job. Most manufacturers will recalibrate a pendulum unit for a modest fee, and it's usually cheaper than a replacement.
Self-levelling range is a site spec, not a lab spec
The ±3° or ±4° figure describes how far out of level the base can sit before the pendulum hits its stops. Inside that range the unit corrects itself and stays accurate; outside it, most units flash the beam or beep, and a few older ones simply project a confidently wrong line.
Four degrees sounds generous until you set a tripod on a screeded floor with a fall in it, or on a scaffold board. Wider range costs nothing extra on the spec sheet and saves genuine time on uneven ground.
One feature worth checking separately: a pendulum lock for manual mode. Without it you cannot project a deliberate slope, which rules the unit out for drainage falls and ramp work.
Beam class, briefly
Almost every cross-line and multi-line level sold for interior work is Class 2, meaning under 1 mW, where the blink reflex handles accidental exposure. Don't stare into it, don't point it at anyone's face, and that's the whole precaution.
Higher-output rotary heads and some outdoor line lasers run Class 3R, up to 5 mW. Those have a genuine nominal ocular hazard distance and the figure belongs on the label. If a unit is sold as Class 3R and no hazard distance appears anywhere in the documentation, that absence tells you something about the manufacturer.
What to ignore on the box
- "Working range up to 50 m" with no mention of a detector. That figure is almost always detector-assisted; the visible range is a fraction of it.
- Accuracy quoted without a distance. ±3 mm means nothing on its own, and the omission is rarely accidental.
- IP ratings quoted as a single number. IP54 tells you dust and splash; the second digit is the one that matters if the unit lives in a wet van.
- Line count as a proxy for quality. Three 360-degree planes are genuinely useful for first-fix; twelve lines on a budget chassis are twelve chances for one to be out of square.
- Battery capacity in mAh with no run time. Green units eat current, and 4000 mAh in a green 3 x 360 is not the same day's work as 4000 mAh in a red cross-line.