That 190 Lumen 3W part....

Comment from Don Klipstein, http://members.misty.com/don/, on sci.engr.lighting:

>>>Take a look at their latest version of DS46, the datasheet for "3 watt"
Luxeon Star:

Characterized light output is achieved with 1400 mA of current, and the
typical voltage drop at that current is 2.95 volts. This indicates a
typical power consumption of 4.13 watts to achieve this. 190/4.13 works
out to about 46 lumens/watt.
The "1 watt" "lambertian radiation pattern" red-orange part achieved 55
lumens with a typical voltage drop of 2.95 volts at its characterization
current of 350 mA (1.03 watts), or typically 53 lumens/watt, according to
a version of the DS23 datasheet that I got last year.

Furthermore: According to these datasheets, this is with the junction
temperature being at 25 C. Check out the thermal resistance figures in
these datasheets, and you will find that achieving that requires a
heatsink temperature of about 4 degrees C for the 1-watt part and about
-16 C for the "3-watt" part.
Light output for the red-orange "3-watt" part with 1400 mA and heatsink
temp. 25 C has typical voltage drop of about 2.87 C, power input 4 watts,
junction temp. 65 C give or take less than a degree for "center value of
typical" (my words), and light output about 64% of that achieved at same
current and 25 C junction temp. (Look for light output as a function
of junction temp. on page 9 of the version of DS46 mentioning red,
red-orange and amber parts.) This works out to about 30.5 lumens/watt.

The "1-watt" version fares better: With heatsink temp. 25 C I figure
junction temp. close enough to 45 C and typical voltage drop of 2.91 volts
and this at 350 mA is 1.02 watts. Typical light output in this case is
about 79% of that with junction temp. of 25 C, for about 43.5 lumens or
about 43 lumens/watt. This is according to a version of DS23 that I
downloaded last year.

Also please consider that a heatsink temperature of 25 C is awfully
optimistic. It surely appears to me that a heatsink has to be of close to
"overkill" size to achieve 35 C temperature of the heatsinking surface of
the part in question, and for a Lumileds red-orange LED that means about
8-10% less light than is achieved with 25 C temperature of the heatsinking
surface of the LED.

Red LEDs are worse than reddish-orange ones and amber ones are not as
bad as reddish-orange ones in terms of output and efficiency/efficacy
varying inversely with temperature. Better still are green, bluish green
and blue ones and white ones, although temperature sensitivity of white
and green ones is significant. Blue ones have performance varying least
with temperature. <<<

Pass me the LN2 tank just like Lumileds et al, use in the lab ;-)

Adam

Heheh, made much the same point to others.

Still, the 1 Watt is pumping out 43 lumens, and the "3 Watt" at the four watts is pumping out over a hundred.

I don't like the Tj way LumiLEDs uses to rate parts. IMHO, the slug temp of 25C would be much closer to reality for most folks.

The Luxeon IV or whatever they will call it, will be nice with it's 6 C/W thermal resistance. This should go a long ways to maintaining the output.

I do recall a number of recent "high output" Luxeon III LED lights with the body of the lights reaching temperatures of 160-165 F (or 71C to 74C). The LED slug is going to be even warmer than that.

Toss those numbers into your calculations...it bumps the die temperature on up by another 50C, putting the die into the 110C range, which causes a drop in output of 60%, or down to 40% of the original. 140 lumens becomes 56 lumens of actual output.

Yeah, so heatsinking these puppies starts getting really important, things like hotlips in a D mag and brute force heatsinking really pays off.

Fans. They're not just for Megaray anymore. /ubbthreads/images/graemlins/grin.gif

TECs could come in handy here, too.

How would you go about mounting a fan in a mag head with a chunka heatsink in there already?

Also, it's all good to move air around but does it help all that much? Would it be feasible to have some kind of air-vent or one-way valve system to maintain relative water resistance on the whole while incorporating a vent for the heated air, thus (theoretically speaking) allowing greater heat reduction?

Am I onto something here or just on something? /ubbthreads/images/graemlins/wink.gif

You're on something. /ubbthreads/images/graemlins/smile.gif

You can't have an air vent without an air intake. An intake that will let cooling air in will also let in water; an intake that won't let water in won't let air in either.

The way to ensure water resistance is to keep the cooling airflow isolated from the LED and electronics. In short, the LED rejects heat into a heat exchanger which in turn rejects heat into the atmosphere. The atmosphere side of the heatsink will be sealed with respect to the LED side of the heatsink.

What we need is liquid cooling.

Ya a little tube of liquid nitrogen with a small bleeder- make it refillable like a butane lighter. Install a thermal switch so light quits when out of coolant.

"what we need is liquid cooling"

-> or stuff like this

or even better effiencywise, something like this !

Olivier

Heatpipes could offer some increase in the efficiency of a heatsink, but only by so much.

Start adding active cooling elements, and you'll probably end up expending more energy on cooling than lighting b ythe time you're removing enough heat to justify active elements...

I am most intrigued by the transformation of heat to electricity (succinctly!) described in the second link above; it almost sounds too good to be true from a thermodynamic point of of view!

Olivier

Yes, the second link mentioned 50% Carnot efficiency. My only questions are when can I buy one and how much? /ubbthreads/images/graemlins/wink.gif Although they mention power generation, a 50% efficient thermoelectric device would be able to reach very low temperatures. Today's devices can get down into maybe the -40°C range with a light load and liquid cooling. I would love to see a single stage device which can reach liquid nitrogen temperatures at 50% Carnot efficiency.

Getting back to my vent idea... /ubbthreads/images/graemlins/wink.gif

What if you incorporated one of those little pumps (or the pump idea) into a flashlight salvaged from the blood-pressure test the nurses etc use to pump up the pressure bandage thing that goes around your arm and inflates.

Doesn't that have a 2-stage valve so if you pumped it it'd... hmm maybe I need to think this thue some more.

You know where I was going generally with this tho? Hot air out, cool air in. (Use an umbrella or something to keep the water out)!