Lux Meter

What a Lux Meter Measures

A lux meter measures illuminance, the photometric quantity that describes the density of luminous flux arriving on a surface. In the SI system illuminance has the unit lux (lx), defined as one lumen per square metre. Illuminance is the radiometric quantity irradiance, the power per unit area in watts per square metre, weighted across the spectrum by the CIE photopic luminous efficiency function V(lambda). That weighting is what turns a physical power measurement into a perceptual one: it models how the light-adapted human eye responds to each wavelength, peaking at 555 nm in the green and falling to near zero at the violet (about 380 nm) and deep-red (about 780 nm) ends of the visible band.

This distinction matters because illuminance is fundamentally different from the lumen output of a lamp. Lumens describe the total luminous flux a source radiates in every direction, a fixed property of the lamp. Lux describes only the fraction of that flux landing on a particular surface, which falls off with the square of the distance and with the cosine of the angle of incidence. The same 800-lumen bulb produces hundreds of lux directly beneath it and only tens of lux across the room. A lux meter answers the practical question that lighting codes actually ask: is there enough light on the work surface, here, now.

Three physical functions must be combined inside any honest lux meter. First, a photodetector converts incident photons into a proportional electrical current. Second, an optical correction filter reshapes the detector's raw spectral sensitivity so that it tracks V(lambda), because bare silicon is far too sensitive in the near-infrared and would count invisible heat as light. Third, a cosine corrector, typically a diffusing dome or disc, reshapes the angular response so that light arriving off-axis is weighted by the cosine of its incidence angle, satisfying Lambert's law as illuminance physically demands. A device missing any one of these three is a light-sensitive gadget, not a calibrated illuminance meter.

Historically, illuminance was first quantified with visual photometers in which an operator balanced a test field against a reference lamp by eye, a tedious and operator-dependent method standardised in the early twentieth century. Selenium photovoltaic cells in the 1930s gave the first self-powered electrical light meters, widely used in photography, though their spectral response and temperature stability were poor. The shift to silicon photodiodes with engineered V(lambda) filters from the 1970s onward, together with national metrology institutes realising the candela and lux on cryogenic-radiometer-based scales, produced the traceable, class-rated instruments used today. The modern reference framework is ISO/CIE 19476, which defines the performance indices, and CIE 69, which defines how those indices are calculated.

In scale terms, useful illuminance spans roughly eight orders of magnitude. A moonless overcast night sky delivers on the order of 0.0001 lx, full moonlight about 0.1 to 0.25 lx, a well-lit office 500 lx, an overcast day several thousand lux, and direct midday sun up to 100,000 lx or more. No single detector and range covers all of this with full resolution, which is why field meters use auto-ranging across several decades and laboratory photometers add neutral-density attenuators or specialised low-light detectors at the dark end.

Detector Types and Construction

The heart of a lux meter is the detector head. Several technologies have been used, but the silicon photodiode with a V(lambda) correction filter has displaced almost everything else for general illuminance work. Understanding the alternatives clarifies why. The table below compares the main detector approaches on the metrics that drive a buying decision.

Selenium cells generate a small current directly when struck by light, needing no battery, which made them popular in early exposure meters. Their inherent spectral response is a rough approximation of the eye, so some cheap meters skip a separate filter. The drawbacks are decisive for measurement work: selenium suffers fatigue (a slow droop under sustained bright light), ages noticeably over a few years, and has a strong temperature coefficient. They are effectively obsolete for traceable illuminance measurement.

Filtered silicon photodiodes are the workhorse. Silicon is linear over many decades of light level, fast, mechanically robust, and extremely stable over time. Its native responsivity, however, extends well into the near-infrared and is too high in the red, so a multilayer glass or gel correction filter is bonded over the active area to pull the combined response onto V(lambda). The quality of that filter, and how tightly it is matched across the spectrum, sets the f1' index discussed in Chapter 3. The photocurrent is converted to a voltage by a transimpedance amplifier, then digitised; auto-ranging switches feedback resistors to cover the full span from fractions of a lux to over 100,000 lx.

Spectroradiometers and illuminance spectrophotometers take a fundamentally different route. Rather than approximating V(lambda) with a filter, they disperse the incoming light with a grating onto a detector array, measure the full spectral power distribution, and compute illuminance by numerically integrating that spectrum against the exact V(lambda) table. Because the weighting is done in software against the true curve, spectral mismatch error is eliminated, which is why instruments such as the Konica Minolta CL-500A are preferred for LED and display measurement where filtered meters struggle. The trade-off is higher cost and the need for periodic wavelength and stray-light verification.

Construction of the head also governs accuracy. The cosine corrector, a white opal diffuser dome or flashed-opal disc, sits on top of the filtered detector and is shaped so that the assembly's angular response follows cos(theta). A detached head on a cable is preferred for serious work because it lets the operator place the sensor flat on the workplane without their own body or the display casting shadow or bouncing light into the dome. Integrated meters where the detector is in the same body as the readout are more convenient but more prone to operator shadowing error. Connector quality, temperature compensation of the amplifier, and the linearity of the analogue-to-digital chain round out the parts that separate a class B field meter from a class L laboratory photometer.

Accuracy Classes: DIN and JIS

Unlike many instruments, a lux meter is not adequately described by a single accuracy percentage, because its dominant errors are spectral and angular, not just gain. Two classification systems formalise this. DIN 5032-7 defines four photometer classes, L, A, B, and C, evaluated against roughly fifteen quality indices. JIS C 1609-1 defines a top precision class plus general classes AA and A, where AA outranks A. Both systems are reconciled by the international standard ISO/CIE 19476, which defines the indices and the standard calibration conditions, and by CIE 69, which prescribes how each index is computed. The table below summarises the practical hierarchy and the headline limits.

The single most important index is f1', the spectral mismatch index, defined in CIE 69 and DIN 5032-7. It integrates the absolute deviation between the meter's actual relative spectral responsivity and the ideal V(lambda) curve, normalised and weighted to CIE Standard Illuminant A. A smaller f1' means the detector behaves more like the standard eye. Crucially, f1' is the error that explodes under narrowband sources: a meter with a high f1' may read perfectly under the incandescent light it was calibrated against, then read more than ten percent off under a blue-pumped LED or a sodium street lamp, because the spectral gaps line up with the source's peaks.

The second index is f2, the directional or cosine response error. It measures how far the angular response of the corrected head departs from the ideal cosine. A high f2 causes a meter to over-read or under-read depending on whether light reaches the workplane from steep or shallow angles, which in a real room with ceiling-spread luminaires is always a mixture. DIN class B and JIS class AA both hold f2 to roughly 3 percent. Additional indices in DIN 5032-7 and ISO/CIE 19476 cover linearity (f3), fatigue, temperature dependence (f6), modulated-light response, polarisation, range change, display rounding, and others; the overall class is set by the worst index, so a meter cannot earn class A by excelling on spectral match while failing on linearity.

In practical procurement, the mapping is roughly this. Class L and the JIS precision class are laboratory grades for calibration houses and lighting research, rarely needed and priced accordingly. Class A is the right target for commissioning measurements that feed a defensible EN 12464-1 or ISO 8995 report. Class B / JIS general class AA, which covers the widely used Konica Minolta T-10A and similar professional handhelds, is the workhorse for routine workplace and facility surveys. Class C is acceptable for quick relative checks and rough indication but should not anchor a compliance document. Note that a published class is only meaningful with a current, traceable calibration certificate behind it.

Quantities, Units, and Lighting Standards

A lux meter reports illuminance, but lighting practice involves several related quantities that buyers routinely confuse. Knowing which quantity a code asks for, and which unit a meter reports, prevents reading errors of a whole order of magnitude. The four core photometric quantities are luminous flux (lumen), luminous intensity (candela), illuminance (lux), and luminance (candela per square metre), and only the third is what a lux meter measures.

Illuminance versus luminance. Illuminance (lx) is light arriving on a surface and is what a lux meter with its diffusing cosine head reads. Luminance (cd/m2) is light leaving a surface toward the eye, measured by a luminance meter with imaging optics and a narrow acceptance angle. Glare, screen brightness, and road-surface brightness are luminance quantities; task lighting levels are illuminance quantities. Using a lux meter where a luminance meter is required, or the reverse, is a category error that no amount of accuracy can fix.

Lux versus foot-candle. The imperial unit of illuminance is the foot-candle (fc), one lumen per square foot. The conversion is fixed: 1 fc equals approximately 10.76 lx, so a 500 lx office target is about 46 fc, and a US IES recommendation of 50 fc is about 538 lx. Many field meters, such as the Extech LT40 and LT45, switch between lux and foot-candle on the same detector. Always confirm which unit a report uses before comparing numbers, because a tenfold-looking discrepancy is usually just lux read as foot-candles.

Lighting codes are where lux meters earn their keep. The European standard EN 12464-1 ('Light and lighting, Lighting of work places, Indoor work places') sets maintained illuminance Em on the task area, together with uniformity, glare (UGR), and colour rendering (Ra). The ISO equivalent is ISO 8995-1 / CIE S 008. In the United States the IES Lighting Handbook expresses targets in foot-candles. The table below lists representative EN 12464-1 maintained-illuminance targets used to plan and verify installations.

Two procedural points follow from these codes. First, Em is a maintained value, defined at the end of the maintenance interval when lamps have aged and surfaces dimmed, so a freshly commissioned installation is normally designed to read above Em. Second, the values apply to the task area, which means measurements should be taken on a defined grid across the relevant plane, with the cosine-corrected head flat on that plane, not held at a convenient height or angle. The choice of calibration illuminant also matters here: meters are calibrated to CIE Illuminant A at 2856 K, so under LED lighting a colour-correction step or a low f1' meter is needed before a number is recorded against a code.

Key Specification Parameters

Reading a lux meter datasheet means looking past the headline range to the parameters that actually bound a measurement's trustworthiness. The same instrument may list a dozen lines; the eight below drive selection. Worked reference points come from the Konica Minolta T-10A, a class B / JIS class AA meter whose published specification is representative of the professional field tier.

Measurement range. Field meters span from roughly 0.01 lx to 200,000 to 300,000 lx across several auto-ranged decades. The T-10A is specified 0.01 to 299,900 lx (0.001 to 29,990 fcd); the Extech LT40 and LT45 cover four lux ranges to 400,000 lx and four foot-candle ranges to 40,000 fc. Check both the floor (for dim or emergency-lighting work) and the ceiling (for daylight and high-bay work), because resolution and accuracy can degrade at the extremes of the span.

Accuracy. Field-meter accuracy is quoted as a percentage of reading plus a digit count, separate from the class indices. The T-10A is specified at plus or minus 2 percent plus or minus 1 digit of the displayed value; the Extech LT40 and LT45 quote a basic accuracy of plus or minus 3 percent. Treat this gain accuracy as additive to, not a replacement for, the f1' spectral and f2 cosine errors, which dominate under real, non-incandescent sources.

Spectral mismatch f1' and cosine response f2. These class indices, covered in Chapter 3, belong on the spec line and not just in marketing copy. The T-10A publishes visible-range relative spectral response within 6 percent (f1') of V(lambda) and cosine response within 3 percent (f2). A datasheet that omits f1' is hiding the parameter that matters most under LEDs and discharge lamps.

Linearity and repeatability. Linearity (f3) is how faithfully output tracks input across the decades; a good meter holds a fraction of a percent. Repeatability is the scatter of repeated readings of the same steady source. For relative work, such as comparing two points or tracking a dimming curve, repeatability can matter more than absolute accuracy.

Temperature characteristics. The detector filter and amplifier drift with temperature, so meters publish an operating range and a temperature coefficient. The T-10A is rated for operation from -10 to +40 degrees C at 85 percent relative humidity or less without condensation. Outdoor and cold-store surveys can leave the rated window, so confirm the temperature spec for the actual environment.

Response and sampling. Steady-state surveys are forgiving, but pulse-width-modulated (PWM) LED dimming and flickering sources require a meter whose integration time and sampling are designed for modulated light, otherwise readings beat against the PWM frequency and jump around. Several modern meters explicitly state PWM compatibility. Functional extras such as data hold, min/max/average, analogue or USB output for logging, and a detached head on a cable round out the practical line items.

• Range: floor and ceiling in lux and foot-candle, number of auto-ranged decades.
• Accuracy: percent of reading plus digits, stated against the reference illuminant.
• f1' spectral mismatch: the index that governs error under LED and discharge lamps.
• f2 cosine response: the index that governs error from off-axis light in real rooms.
• Temperature range and coefficient: must cover the real survey environment.
• PWM / modulated-light handling: essential for modern LED and display work.

Selection Decision Factors

To turn the preceding chapters into a specific purchase, work through the sequence below. Most selection mistakes come not from one wrong line but from skipping straight to range and price before fixing the class the measurement actually requires. These steps double as an RFQ template.

1. Define the measurement's purpose first. A defensible compliance measurement against EN 12464-1, ISO 8995, or a workplace-safety code needs at least DIN class A or the JIS precision class. Routine facility and survey work is well served by class B / JIS general class AA. Rough relative checks tolerate class C. Buying class before range avoids the common error of a wide-range but poorly matched meter that cannot support a report.
2. Match the detector to the light source. For incandescent, halogen, and broad-spectrum sources, a low-f1' filtered photodiode is fine. For monochromatic LEDs, sodium lamps, signage, or displays, either choose a meter with a very low f1' (class A) or move to an illuminance spectrophotometer such as the Konica Minolta CL-500A that measures the full spectrum and removes spectral mismatch.
3. Confirm range covers both extremes. Verify the floor for dim or emergency-lighting work and the ceiling for daylight or high-bay measurement, in both lux and foot-candle if the report uses both. Check that accuracy holds, not just that the range exists.
4. Require f1' and f2 on the datasheet. Treat any meter that does not publish its spectral mismatch and cosine error as unrated for serious work, regardless of its headline percent accuracy.
5. Choose head geometry. A detached cosine head on a cable lets the operator lay the sensor flat on the workplane without self-shadowing, the right choice for grid surveys. An all-in-one body is acceptable for spot checks where convenience outweighs the shadowing risk.
6. Check the environment and modulation. Confirm the operating temperature and humidity cover the actual site, and that the meter handles PWM-dimmed LED if that is what will be measured. Cold stores, outdoor work, and flickering sources each rule out otherwise-adequate meters.
7. Verify calibration and traceability. Insist on a calibration certificate traceable to NIST, PTB, NMIJ, or an equivalent national metrology institute, stating the reference illuminant (2856 K Illuminant A), the uncertainty, and the recommended interval, typically 12 months. A class rating without a live certificate is not auditable.
8. Total cost of ownership. Include the annual recalibration fee, the cost of any colour-correction factors for specific LED sources, and spare detector heads, against the cost of an inadequate measurement that has to be repeated or that fails an audit. A meter that is cheaper to buy but cannot be calibrated locally often costs more over a five-year life.

One last dimension is serviceability: whether the manufacturer or a local laboratory can recalibrate the meter to a traceable standard, whether replacement cosine heads and filters are available, and whether logging software and firmware are supported. Established names such as Gossen, Konica Minolta, LMT Lichtmesstechnik, Gigahertz-Optik, HIOKI, Testo, Yokogawa, and Extech / FLIR maintain calibration and service paths, which over a multi-year service life matters as much as the original purchase price.