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.
In 2013 Philips showed a 200 lm/W LED tube prototype. What it would take to replace fluorescent tubes and CFLs: efficacy, ballast compatibility, retrofit safety and mercury.
In April 2013, Philips announced a prototype LED replacement for the fluorescent tube that produced 200 lumens of white light per watt, roughly double the efficacy of the best fluorescent lighting then available. The company said the technology would reach the market for offices and industry in 2015. By June 2013 the obvious question was being asked in offices and plant rooms everywhere: could LEDs replace every fluorescent tube and compact fluorescent lamp, and how soon?
Unlike the incandescent bulb, the fluorescent tube was not an easy target. The US Department of Energy’s 2012 figures put a modern linear fluorescent lamp at up to 118 lm/W, or about 108 lm/W including ballast losses, with a colour rendering index around 85 and a rated life of 25,000 hours. The best LED troffers of the time reached about 110 lm/W. Compact fluorescent lamps (CFLs) were weaker, around 63 lm/W, and were already losing ground to LED lamps of 70–100 lm/W.
So in mid-2013 the efficiency case was clear for CFLs but marginal for good T5 and T8 tube systems. A 200 lm/W tube would change that.
Worked example. Delivering 3,000 lumens at one tube position:
Fluorescent system at 108 lm/W: 3,000 / 108 ≈ 28 W
LED tube at 200 lm/W: 3,000 / 200 = 15 W
Saving ≈ 13 W per position, about 46%
For an office floor with 400 tube positions lit 3,000 hours a year, that is 400 × 13 × 3,000 / 1,000 = 15,600 kWh a year. An LED tube at the 70–100 lm/W the DOE reported for typical LED replacement lamps in 2012 would have saved little or nothing, and at the low end would have used more power than the fluorescent system it replaced.
Philips described the lamp as combining blue LEDs made by its Lumileds subsidiary with red LEDs and a phosphor, achieving 200 lm/W with a colour rendering index above 80 and colour temperatures in the warm-to-neutral range. It was a laboratory demonstrator; the efficacy was achieved under controlled conditions, and commercial products typically trail lab results by several years.
A fluorescent tube cannot run directly from the mains; it needs a ballast to strike the arc and limit current. That control gear stays in the fitting, and how an LED tube deals with it is the central retrofit question. In 2013 three approaches were in use:
| Approach | How it works | Advantages | Drawbacks |
|---|---|---|---|
| Ballast-compatible tube | LED tube runs from the existing ballast | No rewiring, quick | Must match the specific ballast; ballast losses and failures remain |
| Ballast bypass (direct mains) | Ballast removed, mains wired to the lampholders | No ballast losses or maintenance | Requires an electrician; fitting is modified; mains voltage at lampholders |
| Remote driver | Ballast replaced with an external LED driver | Driver can be efficient and dimmable | Rewiring needed; driver and tubes must be matched |
Other considerations apply whichever method is chosen:
In many cases, replacing the whole fitting with an LED luminaire designed as a system was the better engineering answer, and often not much more expensive than buying good retrofit tubes.
Every fluorescent lamp contains mercury. Under the EU’s RoHS rules in force in 2013, the limits per lamp were 2.5 mg for a CFL below 30 W, 3 mg for a standard T5 tube, 3.5 mg for a standard T8 and 5 mg for a long-life tube. The amounts are small, but they are released when lamps break or are landfilled, and fluorescent lighting was installed in huge numbers. LEDs contain no mercury.
As of mid-2013 the answer was: yes for CFLs, soon; for tubes, once efficacy clearly exceeded 110–120 lm/W at a competitive price and retrofit products matured. The barriers were cost, compatibility with installed control gear and confidence in lifetime claims, not physics. The related story of the filament lamp is covered in Malaysia’s incandescent ban, and the basics of efficacy and colour quality in LED lighting in 2012. For why wattage no longer indicates brightness, see what determines the brightness of an energy-saving lamp.
Commercial LED tubes and luminaires steadily overtook fluorescent systems on efficacy and price, and the retrofit approaches described above were formalised in safety certification, where the industry came to refer to ballast-compatible, ballast-bypass and remote-driver tubes as Type A, Type B and Type C. Regulation then closed the door on fluorescent lighting: parties to the Minamata Convention on Mercury agreed in 2022 to phase out CFLs for general lighting by 2025, and in 2023 to phase out all fluorescent lamps for general lighting by the end of 2027. The question the 2013 post asked has, in effect, been answered by law as well as by technology.