How to read a flashlight spec sheet

Most specifications on a reputable flashlight follow ANSI/PLATO FL1, a standard that defines not just what to measure but exactly how. Output is measured 30 seconds after switch-on with a fully charged cell, which is why the number reflects a warm light rather than the instant-on peak. Runtime runs until output falls to 10 percent of that initial figure. Beam distance is calculated, not paced out: it is the distance at which the beam drops to 0.25 lux.

Two further tests appear on better lights. Impact resistance is a drop onto concrete from a stated height in metres, and water ingress is the IP rating. Both are pass/fail at the stated level rather than a measurement of how much better the light might be.

The important caveat: FL1 defines the method but nobody enforces it. Established brands measure honestly; some marketplace listings quote the emitter manufacturer’s datasheet maximum, which the assembled light never produces. Independent reviewers with integrating spheres are the only real audit.

Lumens measure total luminous flux: all the visible light leaving the torch, in every direction. It tells you how much light there is, and nothing at all about where it goes. This is why it is a poor single number to shop on.

The figure you should care about more is sustained output, which almost never appears on the box. Turning electricity into light is at best about 30 percent efficient and the rest becomes heat, so a small aluminium body can only run at high power for a minute or two before thermal regulation steps it down. A pocket light advertising 2,000 lumens may settle at 400; a larger 21700 light with the same claim may hold 900 because it has more metal to move heat into.

How many lumens you actually need covers realistic levels — the short version is that most tasks live between 50 and 300 lumens.

Candela is luminous intensity: how much light travels in one direction, which for a flashlight means the centre of the hotspot. This, not lumens, is what determines reach. Beam distance follows directly from it — the FL1 figure in metres is simply twice the square root of the candela.

Because it is a square root, the returns are brutal. Going from 10,000 to 40,000 candela — quadrupling intensity — only doubles rated distance from 200 to 400 metres. That is why serious throwers grow such large heads: reflector diameter is the main lever, and it has to keep growing.

Across this catalog rated distance runs from 35 m (Olight Obulb Pro) to 1,674 m (Noctigon K1), with a median of 219 m.

No manufacturer publishes a beam angle, but you can derive it, because lumens divided by candela gives the solid angle the light is spread over. Every light on this site has its beam drawn from exactly that calculation. It is a first-order estimate rather than a measurement — total lumens include spill, so a light with a tight hotspot and heavy spill reads a little wider than its hotspot really is — but it sorts the catalog correctly.

The catalog median is about 26 degrees, which is the balanced zone most general-purpose lights occupy: a defined hotspot with enough spill to see your feet. Thrower vs flooder covers how to choose.

Colour temperature in kelvin is the hue of the white: 2700K is warm orange, 4000K to 4500K neutral, 6500K a cold blue-white. Counterintuitively, higher kelvin looks colder. Cool white posts better lumen figures because less phosphor means less conversion loss, which is why budget lights default to it.

CRI asks whether the colours it shows you are honest. The headline Ra figure averages eight pastel test colours out of 100; 90 and above is high CRI. What Ra hides is R9, the deep-red sample, which is excluded from the average and is exactly where cheap emitters fail — the reason blood looks black under a bad torch. 47 of the 133 lights here with a published CRI figure are 90 or above.

Duv is the axis nobody warns you about: how far the emitter sits from the black-body curve. Positive duv looks green and unpleasant; slightly negative looks rosy and clean. Two lights both labelled 5000K can look completely different for this reason alone. CRI and tint explained goes further.

The emitter model is the single most informative field on a spec sheet once you know the names, because it constrains everything else. A Nichia 519A tells you high CRI, rosy tint and a wide beam. A Samsung LH351D tells you high CRI and strong flood output. Cree XHP50 and XHP70 mean high output and cool, ordinary colour. Luminus SFT-40 and the Osram W1 and W2 mean a tiny die built for throw, and low CRI as an unavoidable consequence.

That last point is the fundamental tradeoff: a physically small, intense emitting surface is what a reflector needs to focus tightly, and the phosphor layer that would fix the colour rendering also enlarges the source. You can have reach or you can have honest colour; multi-emitter lights sidestep it by fitting both.

The cell format sets the ceiling on both output and runtime. A 21700 lithium-ion cell holds around 5,000mAh and delivers high current without sagging; an 18650 is the same idea in a smaller body at around 3,500mAh. AA and AAA lights trade a great deal of performance for cells you can buy anywhere, and built-in cells trade repairability for convenience.

The full battery comparison covers protected versus unprotected cells, button versus flat tops, and the handful of safety rules that genuinely matter.

FL1 runtime is the time until output falls to 10 percent of the initial reading. That definition is why a quoted runtime can be technically true and practically useless: a light that starts at 1,500 lumens and spends most of the stated three hours limping along at 200 has met the standard.

What you want is a runtime graph, which independent reviewers publish and manufacturers rarely do. It shows the shape: how fast turbo steps down, what the light settles at, and whether it holds that level or sags steadily. A light with a flat regulated curve at 600 lumens for two hours is far more useful than one that starts higher and falls away.

The user interface is the specification you will interact with most and the one most likely to make you dislike an otherwise good light. The main division is between fixed modes, where each click steps through preset levels, and ramping, where holding the button smoothly raises and lowers output.

Things worth checking: whether there is a genuine moonlight mode below a few lumens, whether strobe is buried where you cannot hit it by accident, whether the light has a lockout so it cannot switch on in a bag, and whether mode memory returns to the level you were using.

A large slice of the enthusiast market runs Anduril, open-source firmware with ramping, deep configuration and a simplified mode for people who want none of that.

IP ratings have two digits: the first is solids, the second liquids. IPX8 means tested for continuous submersion with no dust rating claimed; IP68 covers both. For most people IPX8 is ample. An IPX4 light is splash-resistant only and should not be taken out in real rain with confidence.

Impact resistance is stated in metres of drop onto concrete, and one metre is typical. Body material is almost always 6061 or 7075 aluminium with type-III hard anodising; titanium and copper appear on enthusiast lights, look excellent, and are worse at shedding heat than aluminium in the case of titanium.

Weight ranges from 0.4 oz (Nitecore TIKI) to 14.0 oz (Sofirn Q8 Pro) across this catalog — a useful reminder that the specification sheet does not tell you whether you will actually carry the thing.