technical limitations to single emitter LEDs?

[CouchTater]

The maximum power output of single emitter LEDs seems to be growing at a steady rate. It occurred to me that there may be a diminishing returns effect at some point, either in terms of total watts thru a single chip or in terms of watts per unit chip area.

I kind of expect there's going to be a thermal limit using just conduction from the LED module into the flashlight body or reflector, but are there limitations coming from either manufacturing cost or the solid state physics?

 

[sysadmn]

Interesting question. Given that cpus with massive heat sinks are in the 100+ W range, I suspect the limits will be economic. LEDs are much simpler than cpus or memory in terms of processing steps, but no one pays $900 / LED either

 

[Gunner12]

At 100% efficiency and 100 Color Rendering Index, a white LED(or any light source) should create 242.5 lumen per watt and produce no heat. But it is almost impossible to reach 100% so I'd guess the upper limit to be about 200 lumen per watt. At that time we might be able to drive LEDs at 10+ watts with out too much worry about heat.

 

[RustyKnee]

not sure if its common practice at the moment...but looking at datasheets the Power Disipation figure is the same as power consumed by the LED. Power Dis is what is used for thermal calcs. When LEDs get more efficient they should quote what the power that goes into heat is as well as overall and as light.

Stu

 

[who]

Multi Core is the solution here. The current LuxV is a multi-core LED with minor heat and throw problems. It is still a great solution as proven by the L4/L2.

I do not believe multi threading would work in this case.

 

[CouchTater]

Forget threading, vector processing is the way to go.

Its obvious that there's some point where one could sandwich a peltier device to the back side of a LED chip, or even install a micro cooling fan, for folks who will accept a lower net lumens/watt in order to obtain more lumens period. But I don't know much about SS physics, maybe we'll never have LEDs with a high enough power/area ratio to require forced cooling.

As for $900 LEDs, if somebody comes up with a $900 LED that can bring down geese mid-flight, I've no doubt that many of the hard core folks here will be purchasing waterfowl stamps.