LED lights get dimmer over time?

I have two different sets of white LED night lights, both bought from costco. All of them worked quite well when new, but become noticeably dimmer over the course of one to two years. Some are so dim they are useless as a night light.

LEDs are supposed to last a long time. Could this be caused by overdriving the LEDs? It's very disappointing.

Now I'm thinking of buying LED flood lights from hardware store eqivalent to

60W incandescent blub. Would they also become noticeably dimmer after a year or two? If so, I'm not buying.
Reply to
bob
Loading thread data ...

LEDs operated at ratings last little longer than electrolytics -- a few months. Extreme life is only obtained near room temperature, and, you know, heatsinks are expensive, after all...

Tim

-- Deep Friar: a very philosophical monk. Website:

formatting link

Reply to
Tim Williams

Did the lights come with a guarantee of light output over time? If not, that was a clue to lack of quality. If so, get them replaced.

There are a lot of crappy things that are built with LEDs, poorly, and at very poor lumens per watt, for sale to the uncritical masses that see LED and look no further. If you avoid those, there's also some good stuff coming on the market, and some stuff that might at least be worth giving a sample item a try to see if it's any good yet. But you have to be able to smell the pure crap and avoid it.

Aging is normal. How bad (or how much) is somewhat influenced by what you buy. Inadequate cooling and/or overdriving the current will make it worse, yes.

Two major-brand name not terribly cheap (well, one is actually somewhat expensive) LEDs I have are guaranteed to produce at least 70% of rated output at 5 years of age (or 50,000 hours). That would be noticeable, but not (as I use them anyway) useless. The companies in question will probably actually be around in 5 years, if the lights fail. Failure is more likely to be a defective individual unit than a design flaw affecting all units, if it does occur.

Chinese "night lights" from "company changes name every two weeks to avoid liability for the junk they make" are a different story. Most are crap from day one, and they certainly don't improve with age.

Likely the same with "flood lights" if they are house-brand or a name you've never heard of, and don't have a 5 year or 50,000 hour guarantee. House brand is better than unheard of if there is a guarantee, since the hardware chain is likely to be around longer than a chimeric low-quality company.

IMHO, if you are looking for a "flood light" (particularly an outdoor one) and want LED, it's better to look at a fixture than a bulb. Both the cooling and the throw of light are done better with a fixture that's designed from the ground up as an LED fixture. But if you want an outdoor fixture for sub-$100, I think all you're going to find at this time and that price point is junk. Plenty of it, but it's junk.

One simple criteria (you may need to have a calculator handy, as it's often not presented) is lumens per watt. If you stick to things that are

50 lumens per watt or better, you avoid a lot of the junk without having to look too closely at it, or know much beyond "how many lumens" divided by "how many watts"...
--
Cats, coffee, chocolate...vices to live by
Please don't feed the trolls. Killfile and ignore them so they will go away.
Reply to
Ecnerwal

I think this is a significant point -----^^^^^^^^^^^^

Most (higher power) incandescent replacement LED lights have a "depreciation" rating -- how much light output falls over time. Usually, something like 30% of 50,000 PoHrs. If you are only running the light "at night" (i.e., to appease young children sleeping in an otherwise dark house), that's >10 years!

Depends on where you are located and the sorts of ambient temperatures in which they will be operated. Heat is the killer, as always. I have some 7W LED lamps that I've NOT used simply because of the color temperature. The "lamps" are roughly the size of BR30's but, instead of "glass", it's just a huge chunk of aluminum (heat sink).

Reply to
Don Y

If you have an IR temp measurement device, check the temp of each unit.

LEDs degrade over time at higher heat, not light output.

But, light output is based on the amount of current thru the LED.

Temp can tell you indirectly the amount of heat being generated.

Let us know what you find.

Reply to
hamilton

Usually caused by poor thermal management. Particularly bad with lamps that are "designed" (ie seriously compromised) to fit into existing lamp fittings. What waste heat there is from an LED device ends up mostly as heat. Filament bulbs can at least radiate a fair proportion of their waste heat away in the near infrared and are much more tolerant of being very hot than semiconductors.

Quite likely unless the heat sinking is done properly.

--
Regards,
Martin Brown
Reply to
Martin Brown
[...]

Can't argue with that! :-)

--
~ Adrian Tuddenham ~
(Remove the ".invalid"s and add ".co.uk" to reply)
www.poppyrecords.co.uk
Reply to
Adrian Tuddenham

The cheap ass ones I got from Fry's (two for a buck) have lasted for years.

However, I bought some "Lights of America" at Costco. They didn't make it more than 6 months. Costco refunded my money.

For floods, the person who suggested getting a fixture with the LED installed is probably on the right track. I think putting all the step down and regulation into something the size of an Edison socket is asking too much.

Reply to
miso

I wonder. Do these use batteries?

Just because they plug in to the mains doesn't mean they actually use mains power.

Cynically yours,

--
Syd
Reply to
Syd Rumpo

mer

a
g

to

ear

Yes, probably, but this information should be available from the manufacturer. From this rather wretched PhD thesis* on white LED reliability:

formatting link

The most significant causes of failure in white LEDs (actually blue LEDs whose light emission stimulates various phosphors to produce wide spectrum light) are encapsulant darkening, current and temperature acceleration of defect propagation (increasing the fraction of current that does not cause the junction to emit photons) and open circuits caused by such things as die attach delamination or wire bond lifts. Thermal stress caused by the differential between ambient and die temperature caused these mechanical failures.

*More decent journal articles require a subscription to say, the IEEE database. This was openly available.
Reply to
spamtrap1888

They are not overdriven, they just operate at the absolute maximum constant ratings :-).

Thus, these LEDs run very hot, close to the maximum junction temperatures, which for any electronics in general drops the lifetime to one half for every 10°C of increased temperature.

"White" LEDs are actually blue LEDs with a fluorescent material that converts some blue light to red and some yellow. Unfortunately due to the hot temperatures and the intense blue radiation, the performance of the material is deteriorated, causing a fast drop especially on red/yellow radiation and hence total radiation. The blue might even be more dominant, thus increasing the colour temperature as the device gets older, perhaps after a few thousand hours.

In general, the light output from a LED drops at high junction temperatures. In addition, the efficiency [lm/W] at high currents (even momentarily at +25°C junction temperatures) also drops significantly at high currents. Check for any reputable manufacturer, such as Cree or Philips that specify the operational characteristics at 1/2 or 1/3 of the maximum current and publish derating tables for other temperatures and currents.

To build a highly efficient (up to 100 lm/W) light source that will produce nearly the same amount of light (lumens) for years, one should take "3 W" white LEDs (rated for 1000 mA max) and run them at 350 mA (1 W). This significantly reduces aging, while simultaneously helps heat management (smaller Rth(j-c)) Since the LED is now operating at a higher efficiency, for the same amount of light output, less electric input is required, generating less heat.

Thus:

  • cheap 1 W lamps are made of "1 W" LEDs operating at 1 W
  • expensive 1 W lamps might be made of "3 W" LEDs operated at 1 W
Reply to
upsidedown

An ordinary 60 W bulb will produce 700-750 lm and depending of CRI and colour temperature needed, assuming 80-110 lm/W efficiency (possible only well below maximum current), this would require 6-9 W input into the LED array.

In order to produce the same light output for years, 6-9 LEDs each claiming to be "3 W" but actually operated at 1 W would be required.

Reply to
upsidedown

mmer

a

ng

For LED degradation, use of Arrhenius is often erroneous.

Further, these devices have failure mechanisms that have nothing to do with defect propagation in the die.

LED degradation depends far more on current density than on junction temperature.

Less current density to propagate defects, which increase the number of non-radiative recombinations. (So-called "dark current")

This is likely true.

Reply to
spamtrap1888

I agree with you that these phenomenons are going to be an issue after

100 000 hours at reasonable junction temperatures, but the drop in blue/red conversion efficiency at 10 000 h might make the LED useless in a few months.
Reply to
upsidedown

Sounds like Costco is selling junk, an initial wow factor but ultimately a non-performer. You can't go wrong with name brands like Philips, GE, Sylvania who actually engineer their products.

formatting link

Reply to
bloggs.fredbloggs.fred

snipped-for-privacy@downunder.com wrote in news: snipped-for-privacy@4ax.com:

white LEDs depend on PHOSPHORS to create the "white" light from blue/UV emissions of the semiconductor. that is why white LEDs appear yellow when off,you're seeing the phosphors.

Phosphors eventually wear out. the harder you drive them,the sooner they wear out.

--
Jim Yanik
jyanik
at
localnet
dot com
Reply to
Jim Yanik

The inorganic phosphors are generally pretty robust. It is cooking the chip and the epoxy nearest the chip going brown that wrecks output. eg.

formatting link

Reports on accelerated ageing tests.

--
Regards,
Martin Brown
Reply to
Martin Brown

o

en

A die dissipating one watt will definitely cook the epoxy that encapsulates it. Coating the die with a liquid silicone (as is used in optoisolators) before molding in epoxy might decrease the amount of browning, as silicone is a good thermal insulator. But the back of the die should be attached to a good heat sink -- maybe a heat pipe.

Reply to
spamtrap1888

Cheap LED's often use inferior epoxy for their packaging. This discolors over time and goes dark - increasing the normal aging effects.

--
We have failed to address the fundamental truth that endless growth is  
impossible in a finite world.
Reply to
David Eather

mer

a
g

Some time ago, testing, I had cheap Chinese white LEDs decay from full brightness to useless in two weeks' continuous operation at ordinary currents, and room temperature.

Nichia (NSPW500) LEDs were superb.

-- Cheers, James Arthur

Reply to
dagmargoodboat

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.