What determines the brightness of an energy-saving lamp? Lumens, watts and efficacy
Why lumens, not watts, measure lamp brightness; luminous efficacy of incandescent, CFL and LED lamps as of 2012; colour temperature; and how to choose a replacement.
Efficacy, lifetime, price and light quality of LED lighting in 2012, using US Department of Energy data, with the colour, heat and cost caveats buyers faced at the time.
By August 2012, light-emitting diodes had moved from indicator lamps and torches into street lights, downlights and retail displays, and governments, including Malaysia’s, were planning to move away from incandescent bulbs. The case for LEDs rested on three numbers: how much light they produce per watt, how long they last, and how fast their price was falling. In 2012 the first two were already impressive; the third was the main obstacle.
An LED is a semiconductor junction that emits light when current flows through it. A single junction emits a narrow band of colour, so almost all white LEDs in 2012 used a blue gallium-nitride chip coated with a yellow phosphor; the mix of transmitted blue and re-emitted yellow is seen as white. Adding red-emitting phosphor gives a warmer colour and better colour rendering, at some cost in efficiency.
One popular description of LEDs as “cold light” needs correcting. LEDs emit almost no infrared in the beam, so the illuminated object is not heated, but a large share of the electrical input still becomes heat at the chip. That heat has to be conducted away through the lamp body and heat sink, which is why thermal design dominates LED engineering.
Efficacy is the light output divided by the electrical input:
Efficacy (lm/W) = luminous flux (lm) / input power (W)
The US Department of Energy’s April 2012 solid-state lighting plan tabulated typical products of the time:
| Light source (2012) | Efficacy | Output | Power | CRI | Rated life |
|---|---|---|---|---|---|
| Incandescent | 15 lm/W | 890 lm | 60 W | 100 | 1,000 h |
| Halogen | 22 lm/W | 1,100 lm | 50 W | 100 | 5,000 h |
| Compact fluorescent | 63 lm/W | 950 lm | 15 W | 82 | 12,000 h |
| LED A19 lamp (L Prize winner) | 93 lm/W | 910 lm | 9.3 W | 93 | 25,000 h |
| Linear fluorescent, lamp and ballast | 108 lm/W | 6,100 lm | 56 W | 85 | 25,000 h |
| LED 2 ft × 4 ft troffer | 110 lm/W | 4,000 lm | 36 W | 90 | 75,000 h |
The DOE noted that typical LED replacement lamps achieved around 70 lm/W, with the best, such as the L Prize lamp, reaching 90 to 100 lm/W. Commercial LED packages, the bare light-emitting components, had reached 144 lm/W for cool white and 111 lm/W for warm white. In other words, LEDs had decisively beaten incandescent and compact fluorescent lamps but were only level with a good linear fluorescent system.
Worked example. A lamp providing about 900 lm, used four hours a day:
Replacing the incandescent lamp saves about 74 kWh a year; replacing the CFL saves about 8 kWh. That second figure explains why, in 2012, LEDs were an obvious upgrade from filament lamps but a harder sell against CFLs. For a household view of these numbers, see what is your household electricity use?
LED lifetime is not the point at which the lamp goes dark. It is normally quoted as L70: the operating hours until light output falls to 70% of its initial value. Packages tested at controlled temperature were rated around 50,000 hours, but a complete lamp is limited by its hottest and weakest parts, usually the driver electronics and the phosphor at high temperature, which is why the DOE listed 25,000 hours for the L Prize lamp. The L Prize test required lumen maintenance above 70% at 25,000 hours; the winning Philips lamp was projected to exceed 99%.
The DOE reported that the typical US retail price of a 60 W-equivalent LED lamp fell from about US$40 to US$25 during 2011, and that warm-white lamps were selling at around US$16 for 400 lm and US$24 for 800 lm, roughly US$30 per thousand lumens. A 13 W CFL cost about US$2 per thousand lumens. At component level, cool-white LED packages had fallen to about US$8 per thousand lumens and warm-white to about US$12. The DOE expected further cuts in 2012, and its long-term target for packages was US$0.50 per thousand lumens at 266 lm/W.
Fluorescent lamps do contain mercury, but in milligram quantities: under EU rules in force from 2013, a compact fluorescent lamp below 30 W was limited to 2.5 mg and a standard T8 tube to 3.5 mg. That remains a disposal problem, but not “huge amounts”. And LEDs do waste energy as heat; they simply waste much less than a filament.
The DOE’s expectations proved conservative. LED lamp prices fell far below the 2012 levels, efficacy of mainstream products passed that of fluorescent systems, and LEDs became the default for new lighting in homes, offices and streets, including Malaysia’s large-scale street-lighting conversions later in the decade. Policy followed: Malaysia’s incandescent phase-out and, globally, the Minamata Convention’s decisions to phase out compact fluorescent lamps by 2025 and general-purpose fluorescent lamps by the end of 2027. The question of whether LEDs could replace fluorescent tubes is taken up in can we replace all fluorescent bulbs with LEDs? The quality issues described above, colour rendering, heat and driver reliability, are still the points that separate good products from poor ones.