led life

My friend and I are curious about the lifespan of LED's. I have read from several sources that led's get dimmer overtime, whereas my friend thinks led's remain a consistant level of brightness pretty much right through their life until nearly the end.
If they do in fact get dimmer.. at what rate? will they remain 100 % bright until the last few hours of their life at which point they will rapidly get dimmer and cease functioning.. Or do they gradually get dimmer over time, so for example an led light with a 10 year life, after 5 years will be 50% as bright.

I would appreciate any answers (along with proof) that you can find.
Thanks
Anthony.

Instead of just blowing out like an incan, LEDs will gradually lose brightness over their lifetime.

Of course, their lifetime is so freakishly long that most people simply won't manage to use an LED enough to notice anything!

EDIT: The Lumileds website promises 70% lumen maintenance over 50,000 hours for their Luxeon I Star. 50,000 hours is equal to ten hours a day, every single day, for over thirteen years, or 24 hours a day for over five years.

Hello pontypool and welcome at CPF

The answer is more complex. There are different leds, so the lifespan depends on many factores.

The manufacturer claimed the lifespan of a Lux I,III at 100000hours, driven at specs.

The new K2, driven at 1000mA with 50000hours.

A not optimazed heatsink or to overdrive the led will reduce the lifespan.
Over the lifespan the output will get dimmer. To underdrive the led will slow down this process.

I hope this helps.

Best regards

____
Tom

so what does the level of dimming amount to per solid 24 hours of use? like 0.5 %
I suppose another factor to consider is how often you turn the device on/off Since turning the power on uses an extra amount of power. but maybe thats getting far too technical. However if one was to consider which would be cheaper multiple flourescent lamps (which are very cheap) or a single led unit of the same luminosity over the same period, you would need such data to decide which is truly more economically sound.

I really don't think this is an issue you're going to have to worry about. The diode life is looooooooong.

Law

pontypool said:

so what does the level of dimming amount to per solid 24 hours of use? like 0.5 %
I suppose another factor to consider is how often you turn the device on/off Since turning the power on uses an extra amount of power. but maybe thats getting far too technical. However if one was to consider which would be cheaper multiple flourescent lamps (which are very cheap) or a single led unit of the same luminosity over the same period, you would need such data to decide which is truly more economically sound.

Click to expand...

In 24hours permanently running you will measure a reduced output due to heat and lost batterie capacity but not the reduced lifespan (like I said, if all factors are right).
You can overdrive a led and the lifespan is only 1 second:laughing:.
To switch the led on and off does not reduce the lifespan, so far I know.

In this times, the price for the leds drops down, so I'm sure that the leds are more economically.

Best regards

_____
Tom

Most Luxeon LEDs will actually get somewhat brighter during the first few hundred hours driven at spec (though for LuxV this may be different for various reasons).

After that initial increase, they will very slowly dim, but at a rate much slower than 0.5% per 24 hours.

LawLight said:

I really don't think this is an issue you're going to have to worry about. The diode life is looooooooong.

Law

Click to expand...

I didn't say this was a concern for me, i,m motivated entirely out of curiousity.

BTW longjohn . Switching the unit on/off repeatidly will reduce the lifespan quicker. According to reports I have read, the exact amount of power required to run a single diode in the split second that it is being powered ON equates to about 3-4 seconds of continous use (don't quite me on exact specs) this is the same as a lightbulb and for this reason lightbulbs that are used to flash repeatidly will run out far quicker than the same bulb being used as a constant source of illumination.
Think about it.. if it requires for example 50% more power to run a diode in that split second when you turn it on, then repeatidly turning on/off will result in a significant increase in power usage and decrease in lifespan for the diode. Not that anyone would want to sit there repeatidly turning it on and off, but you must get the general idea of where I am coming from ; )

With 70% at 50,000 hours for the Lux I star, the figure is about 0.05% every 24 hours.

Some lights are dimmed by something called PWM, or Pulse Width Modulation. This flashes the LED extremely quickly (faster than human eyes can usually detect) to appear dimmer. If on/off cycles decreased LED life, PWM would hardly ever be used - I hope.

But no, you probably won't notice it. Anyway, after several years, wouldn't you want a new, cooler light anyway?

pontypool said:

I didn't say this was a concern for me, i,m motivated entirely out of curiousity.

BTW longjohn . Switching the unit on/off repeatidly will reduce the lifespan quicker. According to reports I have read, the exact amount of power required to run a single diode in the split second that it is being powered ON equates to about 3-4 seconds of continous use (don't quite me on exact specs) this is the same as a lightbulb and for this reason lightbulbs that are used to flash repeatidly will run out far quicker than the same bulb being used as a constant source of illumination.
Think about it.. if it requires for example 50% more power to run a diode in that split second when you turn it on, then repeatidly turning on/off will result in a significant increase in power usage and decrease in lifespan for the diode. Not that anyone would want to sit there repeatidly turning it on and off, but you must get the general idea of where I am coming from ; )

Click to expand...

Sorry, but I don't understand at the moment what you mean (english is not my motherlanguage).
Do you mean to switch the light on and of during 3-4 seconds will reduce the lifespan. Or the strobe effects like at the Gladius?

Or do you mean, that the led is overdriven (for a short time) when the light turns on? This should not happen at full regulated leds, driven at specs.

The "old" unregulated incans (except the SFA2) could have this problem (with fresh cells), but IMO not a regulated one.

Best regards

_____
Tom

According to reports I have read, the exact amount of power required to run a single diode in the split second that it is being powered ON equates to about 3-4 seconds of continous use (don't quite me on exact specs) this is the same as a lightbulb and for this reason lightbulbs that are used to flash repeatidly will run out far quicker than the same bulb being used as a constant source of illumination.

Click to expand...

Please quote this source then. This information flys in the face of every accepted fact about LEDs and their characteristics. LEDs don't use any more power in the first few microseconds of operation than they do in steady state operation. They are not at all like incandescent bulbs which experience an inrush current spike while the filament warms up.

Many PWM dimming systems switch an LED on and off thousands of times per second. If what you say were true, then PWM dimmed LEDs would only last for a few hours. This is obviously not the case, so what you're saying can't possibly be true.

I think he's talking about "inrush current", a high current transient affecting both incandescent and LED emitters. Obviously PWM dimming would be more affected, but this is well understood. It's not like the engineers who devised PWM dimming forgot about that. I think PWM dimming circuits may include inrush current limiting. Maybe some of our tech gurus could comment further.

Regarding LED emitter life and lumen maintenance, this varies widely based on emitter type and drive level.

In general lower-powered emitters last longer and degrade slower. I think non-white emitters (red, orange, etc) degrade slower than white.

The Lumileds Luxeon V portable emitter has a spec'd life of only 500 hours: www.luxeon.com/pdfs/DS40.PDF. While that's much longer than typical incandescents, you'd hardly call it "infinite" operational life.

Do you have any documentation that describes the "inrush current" in LEDs?

AFAIK, the Vf of an LED is at steady state nearly instintaneously. Even if there is some kind of transient during the LED on-time where the Vf is lower than expected, that is only for about 100ns. Unless you have extremely low ESR bypass capacitors very close to your LED, and no parasitic inductance from things like wires, you're not going to be able to ramp up current fast enough to do something bad in that 100ns window.

Any circuit that uses a transistor to switch will have some slight form of inrush limiting due to the switch time of the transistor. An LED gets to steady state in 100ns - it's unlikely that the transistor switches fast enough to allow high current to flow in that timespan (assuming there is such a thing). Likewise, wires, switches, and springs in a flashlight introduce enough inductance into the circuit that pretty much eliminates the possibility of any inrush current in that first 100ns.

Anyone scoped the first 100ns of an LED powering up? Anyone have a 1GHz digital scope and a power supply / PWM switching system capable of ramping up in a few ns?

evan9162 said:

Do you have any documentation that describes the "inrush current" in LEDs?...

Click to expand...

There's a little info here (PDF): focus.ti.com/lit/an/slyt084/slyt084.pdf

However my point was inrush current either (a) doesn't materially affect LEDs, (b) doesn't happen due to other damping, or (c) is handled by the driver circuitry.

It doesn't sense that designers would use PWM dimming and stupidly forget it shortened emitter life by several orders of magnitude.

Where exactly does that document talk about inrush current in LEDs?

Means "inrush current" the same like "high voltage peak"?

Thanks and best regards

_____
Tom

evan9162 said:

Where exactly does that document talk about inrush current in LEDs?

Click to expand...

The subheading titled "Load Disconnect".

Uh, no.

That section is talking about the inrush current that the driver must supply to an output capacitor that is fully discharged. It has absolutely nothing to do with the transient response of an LED while it is powering up. If you removed the LEDs completely, you would still have to deal with the current inrush problem.

However my point was inrush current either (a) doesn't materially affect LEDs, (b) doesn't happen due to other damping, or (c) is handled by the driver circuitry.

Click to expand...

My point has always been (a), which is evident by the fact that there are PWM designs that allow switching speeds in the tens to hundreds of kHz, and that there are no explicit design elements in PWM circuits that are there to combat any kind of "inrush current" inherent to LEDs while they're switching on.

evan9162 said:

...there are no explicit design elements in PWM circuits that are there to combat any kind of "inrush current" inherent to LEDs while they're switching on.

Click to expand...

I was confused by the many references to "soft start to reduce inrush current" in various LED driver chips. I guess you're right, maybe that refers to only non-LED circuit elements such as capacitors in the same circuit.