LED life, not hours...

[scottaw]

This may seem like a silly question, but i haven't found a straight answer yet, I know that led's have ridiculously long lifespans, are imprevious to shock, etc, but will quick flashing burn them out? Is this just an incan problem or has anyone thought about this for led's? All the new fenix lights have a strobe mode, which seems worthless anyway, but will using them shorten the led lifespan?

 

[mdocod]

I have yet to see much evidence that LEDs actually do have a long life. I'd seen a lot of them burn out in normal use, lol...

but to answer your question....

flashing an LED is fine, it's a diode, being turned on and off rapidly is what solid-state electronics are best at

 

[Atomic_Chicken]

Greetings!

What mdocod said... LEDs handle On/Off cycles quite well. A few years ago, I did a project for a client that measured lengths of wood for sorting, using a series of sequentially strobed LEDs and a single photo detector. The LEDs were turned on/off at over 100HZ (100 on/off cycles per second!). As far as I know, it's still operating to this day...

Best wishes,
Bawko

 

[IMSabbel]

Switching doesnt matter at all.
In fact, due to the fact that its off a lot of the time, it will live longer.

Think about it: LEDs and laser diodes( _very_ close cousins, technically) are used in telcom applications where they are swiched 100s of millions of time per second, for years.

 

[TorchBoy]

LEDs don't have ridiculously long lifespans if they're run with too much current. But we sooo want to do that because we get more light out of them that way. :rock:

 

[2xTrinity]

I have yet to see much evidence that LEDs actually do have a long life. I'd seen a lot of them burn out in normal use, lol...

I've never seen an LED burn out in normal use. I've seen them dim rapidly or burn out by say accidentally feeing them too much voltage, or not mounting them to a heatsink properly. I've also seen crappy driver circuits for LEDs break in normal use, but I've never seen an actual LED burn out in normal use. In genreally LEDs will gradually lose ouptut, but not dramatically burn out like an incandescent.

While direct color emitters have extremely long life (there have been indicator LEDs on continually for decades with minimal degredation, also, there have been LED street lights on at about 50% duty cycle for many years with minimal degradation) I don't believe the phosphors used in white LEDs have the same kind of reliability. If we start making light using RGB instead of blue plus phosphor, I expect lifespan to improve dramatically.

Also, the hotter you go, the faster the LEDs will degrade. Regularly running an LED well over specified temperature could cause it to fail (read: dim to a fraction of its original output, not suddently burn out) in a matter of a few hours, as opposed to tens of thousands of hours for an LED that is underdriven and kept cool.

 

[asdalton]

The only LEDs that I've ever seen die are overdriven 5mm LEDs. Heat is the usual culprit for either sudden death or slow degradation of output.

Charts of LED output (5 mm) over time have been posted several times on CPF, and they show that 5 mm white LEDs degrade rapidly (hundreds or even tens of hours) when run on much more than 20 mA. Even when not overdriven, their lifetime is on the order of 5000 hours, not the "100,000 hours" often advertised for the early LED flashlights.

The good news is that the popular high-power LEDs such as the Luxeons and Crees are much more durable if properly heatsinked.

 

[mdocod]

Charts of LED output (5 mm) over time have been posted several times on CPF, and they show that 5 mm white LEDs degrade rapidly (hundreds or even tens of hours) when run on much more than 20 mA. Even when not overdriven, their lifetime is on the order of 5000 hours, not the "100,000 hours" often advertised for the early LED flashlights.

this is the effect I have observed... for example: I have a friend who bought a LED lighting kit for the backyard. They're little clusters in nice little metal housings. We tested current and figured the LEDs are seeing less than 20ma each, they don't even get warm to the touch, but after a few months, leds in the clusters started failing. I've also seen many lower power LEDs in devises like phones and various other household devises "burn-out" for no good reason.

 

[asdalton]

Some commercial LED devices contain questionable circuits that contribute to early failure. For example, they might deliver an average current of 20 mA per LED, but the actual waveform could be spiky with 100+ mA peaks at a high frequency. An incandescent filament won't notice this kind of imperfection, because of its thermal inertia. But for an LED, these transients are deadly. I've heard that some city governments got burned when they purchased cheaply designed LED traffic lights, because they didn't know to check the quality of the current source.

Another bad design is multi-LED clusters in parallel, unless each LED has its own separate resistor. The parallel configuration drives all of the LEDs from the same voltage. LEDs are very sensitive to small changes in voltage, and no two LEDs are exactly alike. The result is that one or more of the LEDs will see a lot more current than the others, and these will fade or burn out early.

 

[asdalton]

By the way, "a few months" is not too surprising a time frame for failure, especially if there are deficiencies in the circuit:

6 months = 4320 hours.

Even if it's not on 24 hours per day, that's far more use than most flashlights will see.

 

[2xTrinity]

this is the effect I have observed... for example: I have a friend who bought a LED lighting kit for the backyard. They're little clusters in nice little metal housings. We tested current and figured the LEDs are seeing less than 20ma each, they don't even get warm to the touch, but after a few months, leds in the clusters started failing. I've also seen many lower power LEDs in devises like phones and various other household devises "burn-out" for no good reason.

A lot of LED lights powered from the line essentially have no regulation -- they just throw a capacitor in series (limits AC current). However, minor fluctutations in your houses power supply, such as an air-conditioning motor switching off and on, could send a voltage spike through that can blow an LED without good regulation. IMO good switching power supplies are a must. Also, 5mm-type LEDs are the absolute worst for fixed lighting applications --that is because there is no thermal path at all -- the plastic domes act as insulators, thus any overdrive at all, especailly for extended periods, leads to huge amount of excess heat at the junction, whereas heatsinked LEDs are a lot more forgiving in that respect. If manufactuers sold products using say underdriven Luxeons or Crees (mounted to a heatsink) longevity would be much, much longer. OF course, those don't cost <$0.02 each like the 5mm LEDs...

 

[frank777]

I have experienced many 5 mm LED failures, but never due to strobing. The problem is mostly cheap Chinese manufacturing. I have never had a 5 mm Nichia burn out. The power LEDs seem even more durable. Strobing will not affect them as long you stay below the manufacturerers maximum forward voltage specification.