Don - can you please post the manufacturer and part number of such an
LED?
Nichia NCCW022, of brightness rank Jc - the rank that includes the 42
lumen output mentioned in:
http://www.nichia.co.jp/product/led-smd-powerled.html
That web page mentions 42 lumens as some sort of typical, and the
minimum for brightness rank Jc is 36.
This is at case temperature of 25 C, as opposed to Lumileds specifying a
junction temperature of 25 C, according to Nichia's datasheet.
Voltage drop is typically 3.6 volts, max. for the upper-middle voltage
rank drop is 4.0 volts, and max for the highest rank (worst) voltage drop
is 4.3 volts.
36 lumens at 4.3 volts at .35 amp works out to... 23.9 lumens/watt
This is a real worst case assuming you get the brightness rank that
includes the lumen figure mentioned in the above-mentioned web page.
"Typical" (maybe a bit optimistic for now) is 42 lumens at 3.6 volts at
..35 amp, which works out to 33 lumens/watt.
I would be a bit concerned about these figures being a bit optimistic
because most of the major manufacturers appear to have been one-upping
each other in their claims recently now that Cree is in the 350 mA white
LED game.
However, Cree has been advancing somewhat steadily at least their
claimed blue LED chip performance over the years, so I expect white
LED output to advance enough for the current claims to soon enough be
conservative rather than optimistic. Optimistic would then be the next
round of even higher claims for output and overall luminous efficacy.
(Cree has a 350 mA white LED with output minimum 40 typical 45 lumens
supposedly hitting production next month. I believe, based on the
datasheet for the "dice" (chips) that I believe they use, that the
voltage drop is typically 3.4, maximum 3.9 volts. This works out
to luminous efficacy of minimum 29.3 "typically" 37.5 lumens/watt. This
is one of their "XL7090" LED "lamps".
Cree specifies ambient temperature of 25 degrees C for their current
XL7090 products in
http://www.cree.com/Products/lmp_7090.asp
-?? specify ambient temperature rather than case temperature for a product
that requires a heatsink??)
When I look at the data for the most efficient white LED that can be
purchased, which are the Lumileds Luxeon III LEDs operating at 700 ma,
and then adjust for minimum guaranteed output instead of "typical"
output and adjust for maximum guaranteed operating voltage instead of
"typical" operating voltage and finally calculate the junction
temperature for a heat sink temperature of 25C and adjust for the
decrease in output as a function of junction temperature I get an
efficacy of slightly less than 16 lm/W, which is about the same
efficacy as a normal 100-watt, 750-hour incandescent lamp.
Close to what I got for Luxeon III "star"... I figure about 16.1
lumens/watt if the worst case voltage drop decreases with increasing
temperature the same as typical voltage drop does. With a thermal
resistance of 17 degrees C per watt and 3.06 watts, junction temp. is 77
degrees C. Looks like you get 82% of the light output that you get with
the junction temperature at 25 degrees C.
The Luxeon III "emitter" has a thermal resistance of 13 degrees C per
watt as opposed to 17 for the "star", making the worst case junction
temperature slightly over 12 degrees C cooler, increasing light output
slightly over 4% (and voltage drop .5-.6%) higher - let's say worst case
overall luminous efficacy is 3.5% better than that of the "star", or 16.6
lumens/watt. Of course this assumes cooling the smaller heatsink surface
of the "emitter" to 25 degrees C.
I do not like the way Lumileds has their nominal performance specified
at a junction temperature that requires a heatsink temperature of
typically -19 degrees C (-2 degrees F) (for Luxeon III "star") if operated
continuously.
- Don Klipstein (
[email protected])