On Mon, 31 Oct 2016 19:08:08 +0000 (UTC), Algeria Horan
wrote:

But what do you get for a 70%-illumination lifetime in the real world?

Dunno. Your real world LED application is not the same as my real
world LED application and the science fiction world of marketing tech
products.

Perhaps you might be interested in "lumen depreciation", L90, L80,
L70, TM-21, LM-80, LM-40, etc.
http://www.p-2.com/blog/lighting-basics-lifespan-lumen-depreciation-l70-tm-21-lm-80-lm-40-and-other-confusing-yet-useful-terms/
There are some interesting methods of calculating LED life such as:
"Reported TM-21 values have an upper limit of 6-times the number
of LM-80 test hours. So if an LED chip is tested for 6,000 hours,
its max reported TM-21 lifetime would be 36,000 hours. If the
chip was tested for 10,000 hours, its max reported TM-21 would
be 60,000 hours."
Magic is a reallly nice way to produce bigger numbers.

One can also be devious:
"It’s worth noting that there are two different types of
TM-21 ratings, "reported" and "calculated" ratings."
Sigh...

Few run 30,000 hr tests. At 8,760 hrs/year, it would take over three
years to run the test, by which time the product is obsolete and
replaced by something new and improved. Instead, they run a HALT
(Highly Accelerated Life Test), which is faster, and presumably
produces the necessary inflated figures:
https://www.google.com/search?q=led+accelerated+life+testing
The basic idea is to test the LED at various elevated temperatures,
connect the dots on a graph, and extrapolate to how long it might last
at some specified operating temperature. At 25C it will last nearly
forever, which assumes that the LED has some form of cooling system.
One can generate amazingly large and impressive numbers this way.
Despite my derisive comments, it does work quite well when performed
honestly and where the test parameters are sane.

For example, IEC 62717 and IEC 62722 LED life testing standards both
demand 6,000 hrs of test time, with total output in lumens recorded
every 1,000 hrs. Measuring lumens accurately requires an integrating
sphe
https://www.google.com/search?q=integrating+sphere&tbm=isch
which might explain why they don't just use a common light meter and
why there are so few measurement points. Some detail on how Luxo
specifies its lifetimes:
http://glamox.com/gsx/led-lifetime-and-the-factors-that-affect-it
I doubt if we will ever see detailed life test results from Costco
LED's as you might from higher priced LED luminaires.


--
Jeff Liebermann
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558