Lux meter selection in industrial photometric work is governed by the DIN 5032-7 accuracy classes (L, A, B, C), cosine-corrected f2 error, and a spectral match f1' to the CIE V(λ) photopic curve [S1][S3].
The instrument — also called a lux meter, lumeter, or luxometer — measures illuminance in lux (1 lx = 1 lm/m²) [S3]. Typical 2026 industrial and commercial units span 0.01 lx to 400 klx across at least four decades, with handheld handheld Testo 540/545-class devices addressing indoor light-level compliance and laboratory-grade sensors covering photometric laboratory calibration [S2].
Spectral match to the CIE V(λ) photopic curve
The single most cited spec on a lux meter datasheet is f1', the spectral mismatch index against the CIE photopic luminous-efficiency function V(λ). DIN 5032-7 requires f1' ≤ 1.5% for Class A, ≤ 3% for Class B, and ≤ 6% for Class C, with cosine response f2 ≤ 1.5% / 2% / 3% for the same bands [S1][S3].
Buyers specifying for white-LED or horticultural-lighting audits should also check the near-IR rejection: an unfiltered Si photodiode responds to 850–940 nm radiation that the eye ignores, so a single-pass IR-cut filter is mandatory for LED measurement [S2]. A second technical gate is the directional (cosine) response: at 30° incidence the f2 error should stay below 3% (Class C) or 1.5% (Class A); many cheap units jump to 6–8% error at 60° because the diffuser is too shallow.
Accuracy class, calibration, and traceability
A defensible audit chain requires the certificate to name the reference standard (typically a cryogenic radiometer or a secondary standard photometer calibrated against one), the traceability number, and a 12-month re-cal interval — anything older than that is non-compliant for ISO 9001 audits.
For occupational-safety work under EN 12464-1 (indoor workplaces) the minimum practical instrument is a Class B cosine- and V(λ)-corrected device, while lighting designers certifying IES LM-79 luminaire reports typically move up to Class A [S1][S2]. The cost step from a Class C consumer unit to a Class B industrial unit is roughly 3–5×, and the step from Class B to Class A is another 2–3×, anchored by the filter stack and the reference-grade photodiode.
Range, resolution, and detector geometry
Range governs whether a single instrument covers both emergency-egress (1–50 lx) and outdoor daylight (up to ~100 klx) work. Most industrial lux meters auto-range across 0.01 lx to 400 klx, and the data-logging variants on the Testo 540-class line store 30,000+ readings with a USB or Bluetooth dump for trending [S2]. Resolution matters at the dark end: a 0.01 lx lower limit is the practical floor for safety-lighting and museum-display verification, and 0.1 lx is adequate for general office surveys.
Detector geometry has a real effect on measurement uncertainty. A 10 mm × 10 mm active area with a properly machined PTFE or opal-glass cosine diffuser gives the cleanest f2 profile; a smaller 3 mm × 3 mm photodiode in a slim wand is convenient but tends to under-read by 4–6% at 60° incidence [S2]. For street-lighting and roadway work — where EN 13201 requires field verification at the carriageway plane — choose a sensor head with a 20 mm or larger diffuser and a remote probe on a 1.5 m cable, so the operator's body and shadow do not contaminate the reading.
Selection map: Class C vs B vs A vs laboratory reference
For a selection-guide buyer, the four realistic options line up against accuracy, cost, and intended duty. (1) Class C handheld — typical accuracy ±5–6%, f1' 6–8%, used for general indoor light-level spot checks; not acceptable for EN 12464-1 verification. (2) Class B industrial — typical accuracy ±3%, f1' ≤ 3%, f2 ≤ 2%, used for workplace-compliance and IES LM-79 field audits; the default procurement target. (3) Class A laboratory — typical accuracy ±2%, f1' ≤ 1.5%, used for photometric laboratory calibrations, luminaire manufacturer QC, and primary standard laboratories. (4) Cryogenic-radiometer-referenced reference — sub-0.5% accuracy, used only at national metrology institutes and reference calibration houses; overkill for almost every industrial buyer. [S4]
Buyers specifying for non-LED incandescent/fluorescent duty can stop at Class B; buyers specifying for white-LED, blue-light-hazard (IEC 62471), or horticultural-spectroradiometric work should not pick a Class C consumer unit — the V(λ) match is too loose and the LED spectra are too narrow to be sampled correctly by a broadband-detector with poor f1'. The same logic appears in the sound level meter selection guide, where the IEC 61672 Class 1/2 split mirrors this Class A/B/C logic: cheap units do not fail loudly, they fail silently with a V(λ)-shaped bias.
Standards, datasheet red flags, and what to demand
Compliance language to demand on the datasheet: "DIN 5032-7 Class B", "CIE V(λ) match f1' ≤ 3%", "cosine f2 ≤ 2%", "NIST-traceable calibration certificate", and "EN 12464-1 field-verification suitable" [S1][S2]. JIS C 1609-1:2006 and the older JIS C 1609-1993 are the Japanese equivalents of DIN 5032-7 and appear on Testo, Konica Minolta, and Topcon datasheets for the Japanese and Korean markets; ANSI C78.81a governs fluorescent-lamp output labelling in North America and is referenced when a lux meter is used to verify lamp-performance claims rather than to certify a workplace [S4].
Red flags that should fail the procurement gate: any datasheet that quotes accuracy but does not name the standard (DIN 5032-7, JIS C 1609-1, or CIE S 023); any spec list that does not separately state f1' and f2; any certificate older than 12 months; any "resolution" claim without a stated range; and any unit that does not specify a diffuser material (opal glass, PTFE, or corrected white acrylic). For buyers who also calibrate their own plant, a side comparison with the sound level meter vs anemometer selection map is useful because the same cosine-correction and traceability logic applies to both, and a plant that already has an IEC 61672 Class 1 sound calibrator usually has the reference standard needed to verify a Class B lux meter in-house.
Procurement checklist and field-procedure gates
Five concrete gates to apply before a PO is cut: (1) DIN 5032-7 or JIS C 1609-1 class stated on the datasheet; (2) f1' ≤ 3% and f2 ≤ 2% with both values printed, not summarised as "spectrally corrected"; (3) NIST- or PTB-traceable calibration certificate dated within 12 months; (4) auto-range covering at least 0.01 lx to 100 klx with a cosine-corrected diffuser ≥ 10 mm; (5) EN 12464-1 / EN 13201 field-verification suitability explicitly listed [S1][S2].
Field-procedure gate: hold the sensor in the plane of the task, not at eye level; allow the unit to stabilise for 5–10 seconds under LED (driver warm-up) before logging; record lux, the operator, the supply voltage, and the luminaire model. These four data points turn a single lux reading into a defensible audit line, and the same logic is used for plant electrical measurement — see the energy meter and electricity meter reference pages for the equivalent traceability chain on the power side.