underdriven leds

i heard that underdriven leds were more efficient
if i run a 3W luxeon on 1W, do i get more light than with a 1W luxeon?

would running a 3W luxeon on 3.4V instead of 3,7V make much difference?

would it be safe to just put a 3W luxeon on a terralux ministar2 (it's all i have to play around with at the moment), and what would probably happen? (could it overdrive the circuit?)

Interesting - I was about 30 seconds from posting a similar question about running a 3W at 350mA to see if it would put out more light than a 1W at 350mA... Maybe someone will know.

The way I understand it, the Lux III is essentially the same as the high-dome Luxeon, except that it's designed for better thermal tolerances and heat dissipation. I wouldn't expect an underdriven Lux III to be more efficient than a Luxeon.

Differences in output from luxeon to luxeon aside, a Q bin Lux I will have about the same output as an S bin Lux III and a R bin Lux I will have about the output of a T bin Lux III. About the only advantage that you'll get with a Lux III is that the tint binning with the Lux IIIs has been pretty good. I haven't had enough experience with Lux Is to tell.

Just be careful when driving a Lux III to 350mA because the tint binning will really show. X1s will look yellow, XOs and YOs will look blue, etc. I've been EDCing a TWOJ running at 350mA and it's pretty nice.

LEDs are current devices. They regulate their own voltage fairly consistently. So long as the voltage is adequate (and they don't release the magic smoke!), you need only watch their current consumption.

That being said, LEDs are almost always more efficient at lower power levels. Many CPFers have reported reaching the point with LEDs where more current just leads to more heat (and rapid failures). Lumensower is not a linear function in LEDs - likely because of less-than-perfect heatsinking.

A LuxIII might put out more light than a LuxI at the same level due to better heatsinking, but the difference would probably only be observable with a light meter.

I'll let more experienced folk correct me if necessary.

another bonus is that a luxIII at 1 w levels will be garunteed to be equal to a high flux 1W. So if you get a 1W you might not get a Q bin or better, but 3W you'll be fine.

Lux III running at a lever where it's dissipating 1W, will give a slightly higher output than a Lux I running at the same level. PROVIDED that in both cases, the package is mounted in such a manner that external thermal management isn't affecting things.

This is due to the fact that the LuxIII has better internal design, meaning that the thermal resistance between the die and the package is lower.

Therefore: If 1W of power is being dissipated in both dice, the package itself is maintained at the same temperature in both cases, the LuxIII die will be running cooler than the LuxI. This will mean that the die will be producing light more efficiently (refer to the lumen/temp charts), hence, the light output at that power level will be grater.

I doubt at 350mA the difference will be huge, but it'll be there.

Lifetime of the LuxIII at that power level will probably be greater than the LuxI too. Again, charts on the datasheets of the devices should confirm that.