Fallingwater
Flashlight Enthusiast
We know that Crees & co. all last much longer than 5mm LEDs due to more efficient heat transfer, but has anyone conducted long-term fade testing to determine exactly how longer they last?
Has anyone conducted fade tests on high-flux LEDs?
- Thread starter Fallingwater
- Start date
Help Support Candle Power Flashlight Forum
I might have someyhing...Ledbar with 60 led in it. Ledbar is waterpoof. The leds are Piranha-led with 4 legs. About 5500K and 1800 mcd each..
I did start to "burn" it 26.5.2008. So its has been on about more than 5500hrs. The led is from Epistar or Kingbright, so its not chinese crap.I also have those same ledbars, which i dont burn, so they are new.
I have a place where i measured them time to time. I use luxmeter for this.Here are the results:
The one that have been on about 5500hrs: 170lx
New one: 154lx
So the new one has lower result. That is because of little colourchanging in the one that is burning. But i might say that its working very well.
I also have another results from ledbars with chinese-led..200hrs and they fade to half of the original. Maybe poor MCD and have to use high current,hard to say..
I did start to "burn" it 26.5.2008. So its has been on about more than 5500hrs. The led is from Epistar or Kingbright, so its not chinese crap.I also have those same ledbars, which i dont burn, so they are new.
I have a place where i measured them time to time. I use luxmeter for this.Here are the results:
The one that have been on about 5500hrs: 170lx
New one: 154lx
So the new one has lower result. That is because of little colourchanging in the one that is burning. But i might say that its working very well.
I also have another results from ledbars with chinese-led..200hrs and they fade to half of the original. Maybe poor MCD and have to use high current,hard to say..
HarryN
Flashlight Enthusiast
I have seen this kind of information published at LED conferences, but have not searched for it lately. The statements by the major producers (Cree / Lumileds / Osram / Nichia, some others) about lumen maintenance in the 70-90% after 50-100 K hours is backed up by third party testing.
There have been University students do this for a thesis project. Of course, the testing time would be quite boring.
Like so many reliability tests on semiconductor devices, there is "normal" testing as well as acceleration method testing which includes actually heating the device into its near destructive operating region under higher than normal current.
Here is a link to Lumileds web site page with some info on reliability. I didn't look, but I think Cree has similar published info.
http://www.philipslumileds.com/technology/lumenmaintenance.cfm
In summary, they last longer than the driver electronics.
jtr1962
Flashaholic
I've been running a Q3 Luxeon at 350 mA since January 2004 ( ~44,500 hours). I didn't have a light meter when I started the test, but it's been holding steady at 10.3 lux @ 1 meter for a while now.
R33E8
Newly Enlightened
I think the LRC at RPI has a few paper's regarding lifetime of high flux (I assume you mean >/=1w) LEDs..
JWP_EE
Enlightened
Cree XR-E 700 mA lifetime curves.
jtr1962
Flashaholic
Interesting chart. With a decent heat sink in a fixed lighting application you can probably get junction temperatures under 40°C at 700 mA. Extrapolating a bit that should mean 200,000 hours to 70% brightness. Or put another way, if the light is on 6 hours per day then you'll get over 90 years life.
