41

Manufacturers routinely claim that their LED light bulb can last from 20,000 hours to 50,000 hours:

Philips LED OSRAM LED Feit electric

Anecdotes, such as this one suggest otherwise. Daily mail wrote:

  • More than a quarter did not meet claims of a 15,000-hour life
  • Disappointing result comes despite claims of them lasting 25,000 hours
  • Some even fell below the legal minimum of 6,000 hours

Are LED manufacturers exaggerating the life span of these bulbs?

Sklivvz
  • 78,578
  • 29
  • 321
  • 428
Question Overflow
  • 1,003
  • 9
  • 14
  • Of the article referenced in the first link one should only accept the observed statistics of failed LEDs. The "theory" part that relies on the claim that every single electronic part has the same failure rate as a light bulb does not hold - an average CPU has millions of transistors and according to that theory your brand-new computer should be dead before you have finished pushing the power button ... – Hagen von Eitzen Jan 17 '16 at 14:41

1 Answers1

39

In an October 2015 test report by the Dutch 'ConsumentenBond' (consumer organization/watchdog) they report back on an earlier test started in 2013 (no link available) where 5 lights each of 24 types from 13 brands were tested.
At that moment they had been running 20000 hours. They were switched 8 times per day: 2:45 hours on, 15 minutes off.
75% of those lights were still running after 20000 hours. They mention the results for specific brands. Although not available everywhere, I'll list here how many died per 5 per brand (fractions can occur when several types per brand were tested):

  • Albert Heijn, Kruidvat, Hema, Calex, Panasonic: 0
  • GoGreen: 1
  • Verbatim (one type): 3
  • Philips, Osram, GP: 1.5
  • Megaman : 3
  • Ikea, Pharox, Verbatim (other type): 5

I was unable to find the life expectancy claims of the 2013 batch (only the phrase with claims up to 25000 hours). Assuming that the current test period does not substantially exceed the claimed periods, my tentative conclusion would be '75% meet the claims.'

The 2015 test results are also available in that report (in Dutch of course), but those lights had only burned 2000 hours by then (1 had failed by then).

  • Could you provide a translation for the headers of the table on the final page of the report? I tried running it through Google translate, but its choking on the diagonal text and showing mangled gibberish instead of something minimally understandable. http://imgur.com/SZLX0lP – Dan Is Fiddling By Firelight Jan 15 '16 at 22:34
  • @Dan That imgur link did not work. Here are the diagonal texts: Price, Test result, Durability and reliability to switching (counts for 30% in test result), Efficiency (light yield) (counts 30%), Light quality (20%), Luminosity measured divided by specified number (15%), Luminosity in lumen on packaging, Measured luminosity, Efficiency (yield) in lumen/watt, dimmable. Note: these are the results for the 2015 batch. –  Jan 15 '16 at 22:54
  • The missing 5% is the result for the 'blindstroomfactor', Google translates that as 'reactive power', and I have no idea what that means. Something like the non-light producing current drawn by the light. –  Jan 15 '16 at 22:57
  • Odd, the imgur is loading for me at home. Reactive power is probably [power factor](https://en.wikipedia.org/wiki/Power_factor). It's a measure of how the power drawn by something deviates from a pure resistive load (in the simplest case, due to capacitors dis/charging and inductors fighting changes in current flow). Bad power factor makes things harder for the power company, and can in some circumstances result in a higher power bill. In the US it's generally only an issue for commercial/industrial customers; but IIRC that residential customers some parts of Europe are also charged for PW. – Dan Is Fiddling By Firelight Jan 15 '16 at 23:31
  • 3
    @DanNeely: Reactive power and power factor are related -- the first is an absolute quantity, the second is relative to the real (dissipated) power drawn. The actual equation is `PowerFactor = RealPower / sqrt(RealPower^2 + ReactivePower^2)` It's also equal to the cosine of the phase angle between voltage and current. – Ben Voigt Jan 16 '16 at 00:08
  • 1
    "Still running" is not the metric they use to say that a bulb is still within it's life span. It is based on lumen depreciation. Usually that threshold is 70% of initial output. They usually say something like "lifespan estimates determined once 50% of a sample of bulbs drops below 70% original lumen output". –  Jan 16 '16 at 00:31
  • "For certain lamp types other criteria than failure to light can be used; for example, the life can be based on the average time until the lamp type produces a given fraction of initial luminous flux." [source](http://apps1.eere.energy.gov/buildings/publications/pdfs/ssl/lifetime_white_leds.pdf) –  Jan 16 '16 at 00:33
  • @fredsbend: Many LED lamp units have a number of series strings wired in parallel; a problem with such designs is that while they work well when all LEDs have matching characteristics, an small imbalance in characteristics can cause some LEDs to be subjected to much greater stress than others, which can in lead to increased imbalances. – supercat Jan 16 '16 at 01:57
  • @fredsbend LED lamps don't behave like CFLs. Their light output does not deteriorate over time as all phosphor-based lights do. So this criterion is irrelevant. Mostly with LEDs it is the electronics that fails and then the light will not light at all. – matt_black Jan 16 '16 at 13:26
  • @Matt I don't know if your right or wrong, but that is what they do. Read the link I provided. It's near the end. –  Jan 16 '16 at 15:36
  • 2
    @matt_black, Re, "Their light output does not deteriorate over time as all phosphor-based lights do." Most "white" LEDs _are_ phosphor based. In virtually all household LED lamps (not counting the expensive, variable color kind), the light source is a broad-band yellow phosphor mix excited by a deep blue LED chip. https://en.wikipedia.org/wiki/Light-emitting_diode#Phosphor-based_LEDs – Solomon Slow Jan 16 '16 at 19:59
  • @jameslarge Fair point for some LED bulbs, but not all. And the effect seems to be much slower than for fluorescent tubes, probably because the mechanism of deterioration of the phosphors is different and less severe than in mercury-driven discharge lamps. – matt_black Jan 16 '16 at 20:46
  • i wager that LED bulbs last longer in homes with Whole House Surge Protection. – rjt Jan 16 '16 at 22:09
  • @rjt, prob'ly right. A cheap-o LED lightbulb replacement probably will fail either because the LED module was not adequately heat sinked, or because its cheap-o built-in power supply either was not adequately heat sinked or, unable to withstand a transient. – Solomon Slow Jan 16 '16 at 22:15