How is LED life-span related to temperature?

The brightness at various temperatures is clearly outlines in the spec sheet, but not the life-span. Everyone knows that keeping the LED cooler is better, but how much does it really matter? Will doubling the operating temperature from 40C to 80C cut the life span in half, or is it just a certain small percentage? Does it rapidly decrease above a certain temperature? When referring to life span, I guess I mean a noticeable decrease in brightness. Such as 80% of the original rated brightness, kind of like how they rate batteries.

I'm specifically interested in behavior for an XM-L or XM-L2.

The XM-L's lifetime was specified at 25C, while the XM-L2's is spec'd at 85C. There isn't really a huge change in the spec. But it's well known that the lifetime plummets after 125C. From that I think it's safe to say there's a highly non-linear curve of life vs temp that has a very shallow slope below 85C, and a rapidly increasing (negative) slope starting somewhere after that. There's also a temperature at which the life will suddenly drop to zero. I also suspect that there's a substantial uncertainty in the temp at which the life will drop below a specific value, which I think is why all manufacturers specify minimum, not typical life.

Since they're spec'd at 85C, does that mean it's safe to run them continuously at this temperature? I don't understand why they started binning them at 85C rather than room temperature.

It seems really hot to me. I was a bit worried because my light was running at 120 degrees F (50C) and didn't want the light to dim over a short period of time. But that's quite a bit lower than the binning temp.

I just use a simple rule with all/any of my lights.Once the light gets warm(not to the point of too hot to hold) but you know that temp when it starts making you think"getting quite warm this" i either drop the level(if still need to use) or let it cool down by switching off.

Keeping hold of the light does help conduct some of the heat away from the body.Tail standing on high imo is not good practice if its a high output light(2-3000lm for example).Not only does it prolong led life but also protect the cells from being exposed to high temps........just an example

If out side and temps are cool then this obv aids longer run time without the risk of overheating as soon.But most high powered lights have timers or thermal temp regulation to step it down off "turbo".I dont depend on this as its just too much to rely on,nothing is fool proof imho.Common sense,judge off body temp and the chances are it will be shelved or sold and have been replaced with a newer version/model in the time,well before the leds go fubar.........

On a slight tangent,what difference does it have to the led switching it on low and ramping the lm up compared to straight into 3000lm for example....
This must have some effect down the line,like starting your truck with the gas pedal to the floor,red lining it cold..........just another thought on led life

Degradation of the LED output over time at temperature can be found by searching for the manufacturers LM80 data (or at least the published summary). Lifetime data is generated for three operating at elevated ambient temperatures and projected using the TM-21 standard. The complicated part for most users is getting an accurate operating temperature. The operating temperature is specified for the die not ambient. There is typically a measurement point on the MCPCBA that is directly connected to the mounting surface of the thermal interface. A thermal couple needs to be attached at that point for measurement. Thermal imaging or the gun type of surface measurements are not accurate for this. The best solution is a proper sized heat sink.

Yes, the XM-L2 is specified to run continuously at 85C, and last for years. But that's the temperature at the die, not at the heatsink. Understanding the difference, and more importantly, calculating the difference, takes some basic knowledge of thermodynamics. That's way beyond trying to post. Maybe we can come up with some rules of thumb for a couple of common configurations.

The reason they specify it at 85C now is that it's really a more useful and realistic number. No power LED is going to be run at 25C. They run hot. It's just a fact of life, and we need to know what happens when they run hot.

As far as turning them on gradually vs instantly, it makes no difference. In some applications, they are turned on and off very rapidly to modulate the apparent brightness. This is what PWM dimming is. They can be switched on and off, at full power, hundreds or even thousands of times per second, for long periods of time, without shortening their life.

DIWdiver said:

As far as turning them on gradually vs instantly, it makes no difference. In some applications, they are turned on and off very rapidly to modulate the apparent brightness. This is what PWM dimming is. They can be switched on and off, at full power, hundreds or even thousands of times per second, for long periods of time, without shortening their life.

Click to expand...

Thank you DIWdiver for that information:twothumbs it does make sense but was unsure:thinking:,you have now clarified that for me:thumbsup:

Here's a few links on the subject:

http://www.yegopto.co.uk/expertise/Drivers_and_Controllers

http://www.digikey.com/en-US/articl...g-the-cause-of-fading-in-high-brightness-leds

You can increase LED lifetime enormously by dropping junction temperatures even by 10 degrees. Many of the studies I've seen put the upper limit of LED L70 lifetime in the 200,000 to 400,000 hour range if you keep junction temperatures close to room temperature. This is obviously not practical for now but you can still obtain lifetimes of 50,000 to 100,000 hours at typical operating temperatures. In the long run, as LED efficiency increases further, junction temperatures will indeed be close to ambient temperatures, and we might expect useful lifetimes of several hundred thousand hours.

kuksul08 said:

Since they're spec'd at 85C, does that mean it's safe to run them continuously at this temperature? I don't understand why they started binning them at 85C rather than room temperature.

It seems really hot to me. I was a bit worried because my light was running at 120 degrees F (50C) and didn't want the light to dim over a short period of time. But that's quite a bit lower than the binning temp.

Click to expand...

Because that is DIE temperature. And its a reasonable value if you have the LED running in a room temperature environement with a moderate heatsink.