All answers, links & opinions are welcome.
I think this is led model specific, and theres no magic number but an average, and it would follow a normal distribution.
I’d love see the ramifications for emitter life to 70% output as a function of drive current.
It will have far less to do with drive current than it will with junction temperature.
PPtk
That’s what I figured. It’s more precise for me to say that due to match’s latest emitter testing roundup I want to analyze the extent to which high (read: not so extreme as to cause immediate, catastrophic failure) drive currents contribute to metal migration, and their impact on bond wire integrity and output over time.
i look forward to seeing all this data with great anticipation!
Alas, I’m just a student today :nerd_face:
But one day, I tell you! One day! H)
Don’t need a degree to follow the scientific method for testing theories. Get some samples, a start them cooking…
PPtk
but the lifetime of LEDs is so high it would take years, as i understand it Cree releases their LEDs with about 3000 hours of testing even though they are rated for 25-100 thousand hours.
Mind you, if i had the money and equipment, i would like to test some CFLs to see how many on/off cycles a particular model can take so i can determine how many hours of life (equivalent) i lose every time i switch it on and off.
Typically the failure rate of electronics doubles for every 10 degree C increase in temperature…
[/quote]
interesting results, some of the numbers don’t seem to make a lot of sense though, and these are extrapolations as well, but lets hope conservative ones.
Some info on life and accelerated life testing
http://reliawiki.com/index.php/Arrhenius_Relationship
The 8C to 10C applies to many chemical reactions but I’ve seen it mostly applied to motor insulation life from testing done a century ago, probably at GE.
Since case and ambient temps are the same, the table data used an infinite heatsink, but theta J-C is necessary to figure junction temp.
A silicon diode junction Vf drops 2.2 mV per degree C with good linearity but LEDs might have a different coefficient.
Thanks for all your inputs 
As far as I can tell, by looking at long-dead posts on this forum and others; and with a little evidentiary support from my own junk box, I would like to offer this as a general theory:
The LED in question will last about 3 times longer than it will remain useful in real life.
A corollary assumption would be that, therefore, it will NEVER die unless you or I take action to kill it.
But I’m just
Dim
up to me to scrounge up some data sheets and do some homework.