A 50,000 hr flashight LED duty cycle specification is useless...

[wweiss]

What if, instead of 50,000 hours of life for a flashlight LED - (which is actually silly because it is way longer than most of the rest of our lives) - we had an option for a 5-10K hour rated LED that would have a more efficient design and higher output. Perhaps drivers could be made that pushed the limits on such a lower duty cycle design so we got X+% more output and less useless lifespan. A 50,000 hour LED (or even 25K) for a flashlight that is used perhaps 5 hours/week, is like buying a car that can go 5 million miles before its functional life is over - there isn't enough lifetime or generational years to drive it. The weakest part of the light is usually the driver and not necessarily the LED - so reduce the duty cycle for the LED and harden the driver to crank the output?

LED / Driver technology is due for an upgrade soon I'm thinking. Perhaps LEP will be the next bump.

 

[desert.snake]

Large companies actually don't give a damn about hand lanterns. The main task is industrial lighting. The street LED lamp shines for about 9 hours a day constantly, it's only 15 years, too little for me. These are ideal conditions. In reality, the lights have to be replaced every 4-5 years, because diodes burn out. Perhaps manufacturers will come to a shorter service life, as you want. This is a planned aging, the more often they break, the more often have to pay for new devices. So it was with incandescent lamps - https://spectrum.ieee.org/tech-history/dawn-of-electronics/the-great-lightbulb-conspiracy

 

[bykfixer]

Thank you for sharing that classic story DS. Little by little LED makers are whittling away at projected life spans too.

My first LED light said "100,000 hours", then a few years later "50,000 hours". Now to see 30,000 hours is not unusual.

 

[wweiss]

Very good link DS. I like the story it has the trappings of the car industry as well.

What I meant to say in my OP statement was that it was like buying a car that had headlights that would last 5,000,000 miles but the car itself would only last 10 years or 300,000 miles.

 

[asdalton]

I wouldn't take any of these nominal lifespans, or differences between them, too literally. Especially not if they are part of advertisements.

Back in the Luxeon days (pre 2007), there was a widespread problem of manufacturers advertising the performance of their flashlights based on the spec-sheet numbers for the emitter. The result was often greatly exaggerated lumen output (due to under-driving, optics losses, or poor heatsinking) or shorter lifetimes (due to over-driving or poor heatsinking). Also, I suspect that there was industry folklore in which people claimed "100,000 hours" because that's what everyone else claimed.

The exaggerations of lumen output lessened when the Cree LEDs came out; the boost in real-world performance was enough that flashlight manufacturers started providing more realistic numbers.

I still wouldn't trust the spec-sheet lifetimes, because the real emitter lifetime will depend on things like the driving current and the effectiveness of the heatsinking. Flashlights sold today might very well have true emitter lifetimes that are matched to the useful life of the whole flashlight, but it's hard to know.

 

[Hugh Johnson]

LED life expectancy is a not a failure rating. Rather, it is a degradation of output to a percentage of original. This means that the led in your example could last much longer than 50k hours and will have reduced output much sooner. I don't see how designing an led with a shorter lifespan can improve efficiency.


I'm also guessing that lifespan is calculated as a best case scenario in lab conditions. I see data sheets with ratings at specific temperatures. I'm not sure if lifespan is factored based on temp or a specific amperage. I suspect that even our 1,000 lumen single 18650 lights are operating beyond lifespan testing. We further have an abundance of lights driving led's much, much harder. It seems we already have what you're asking for.


I'm using a flashlight as a nightlight. It will soon reach 10,000 hours in under 3 years. I'm pretty happy it's not now end of life. I also just purchased a duplicate light and notice the new one is quite a bit brighter. How much of that is led lottery vs. wear I cannot say. It also provides hundreds of hours of light on a single AA. I'm thrilled to have this level of efficiency and durability.

 

[wweiss]

Hugh: Do you have a photo light meter to measure the new v old output? Any consistent measurement would do to get a % change value. Your points are well made - I was looking for the possibility of "over locking" as an option instead of longevity.

Has any study been done about flashlight LED degradation on these forums? Your 3 year usage decline would be very interesting. Unless there is also the function of driver degradation present as well.

 

[archimedes]

The real specs are actually more complicated than "50,000 hours" or whatever.

I can't remember all of the exact details but there are L, B, C, and F ratings, I think.

L70 is something like average burn time down to 70% initial output (at some specified temperature)

B10 means something like only 10% (or less) will fail to meet the L rating

C5 would be 5% (or less) with sudden / complete failure prior to reaching the L value

And I think F values are some type of summary of various failure modes

 

[wweiss]

Thanks for the specs, Archi...

 

[archimedes]

Thanks for the specs, Archi...

Here's an interesting, brief, but moderately technical discussion (commercial site, but linked for reference material) ....

https://www.digikey.com/en/articles...g-the-cause-of-fading-in-high-brightness-leds

 

[wweiss]

Ohhhhh, Yesssss! Love the science. Much thanks for this.

 

[archimedes]

It's interesting for sure

 

[BattleBrat]

I remember back in the day you would have to change the lightbulb once every couple battery changes, it was nuts! If they want to create LED assemblies that will last 10,000 hours and then say you know for peak performance you should replace every 10,000 hours I'm totally down for that. As long as they can increase reliability and efficiency I believe it's worth it .

 

[wweiss]

In my Boy Scout days (which were the best times of my young life) the bent standard issue BS light was heavy, clunky and dim. I did love that thing...

 

[bykfixer]

Changing the LED bulb every 375,245 charges? That's ridiculous. Somebody should do better.
- our great grand children.

By then the term palm pilot will likely be literal in that your autonomous cars pilot system will be a hologram on the palm of your hand. "Siri, take me to Neptune"……

I imagine the folks who inovated the LED when announcing the invention went something like this:

Scientists; "Today, we come here to show you the next great advance in light bulbs".
(shows a really dim puke green light source.) "Someday this will be brighter than a light bulb and last 50,000 hours".

Audience; "BWAHAHAHAHA" "Boooo" "hucksters, get them outta here"
Crowd starts throwing stuff at the scientists.

 

[wweiss]

Can I see this video on YouTube?