Most efficient LED?

[1996alnl]

Just curious as to what's the latest most efficient LED is on the market now.
My guess is the Cree R2, or maybe the Cree Q5?

Take care

 

[Sgt. LED]

The one that will be released tomorrow.

You will not tell the difference in 1 single bin upgrade. You could tell between Q2 and R2 but not between Q5 and R2. The tech moves so fast it is hard to get the latest and greatest, you finally snag an R2 and the week after that the R2's can be had in a much better and way more preferred bin!

It is endless I tell you, and it can drive you crazy & broke.

You ask for what is the most efficient? R2 in single emitter and I have no idea in multi emitters, perhaps the MC-E?

 

[Marduke]

Cree XR-E-Q5 is the highest efficiency, production level LED.

R2 bin is higher, but not available in production level quantities.

 

[OrchidLight]

Cree XR-E-Q5 is the highest efficiency, production level LED.

R2 bin is higher, but not available in production level quantities.

So the Q5 is more efficient than the OSRAM GDP?

 

[Sgt. LED]

Right

 

[Woods Walker]

I wonder if at lower outputs say 5-10 lumens if the Q5 is more efficient than the 5MM GS leds that are in the E01 etc?

 

[Marduke]

I wonder if at lower outputs say 5-10 lumens if the Q5 is more efficient than the 5MM GS leds that are in the E01 etc?

Yes, by a large margin.

At 20mA, the XR-E-R2 for example manages 145 lm/W where the GS-K1 manages about 97 lm/W

The Q5 would be just slightly lower, at least 140 still.

 

[vincebdx]

Cree MC-E / seoul P7 (low cri) with low current :shrug:

 

[1996alnl]

Yes, by a large margin.

At 20mA, the XR-E-R2 for example manages 145 lm/W where the GS-K1 manages about 97 lm/W

The Q5 would be just slightly lower, at least 140 still.

Incredible! If someone told me these figures 10 years ago i would of never believed it.
Makes one wonder what we'll have in 10 years from now.

Take care

 

[Marduke]

Cree MC-E / seoul P7 (low cri) with low current :shrug:

As you can see from your graph, the Cree XR-E R2 is still higher at those lose currents.

 

[dbvanhorn]

Interesting the way the curve falls with current.
Is this a thermal effect? IOW, if the LED were sufficiently cooled, would it remain at or near the highest level?

 

[2xTrinity]

Interesting the way the curve falls with current.
Is this a thermal effect? IOW, if the LED were sufficiently cooled, would it remain at or near the highest level?

I believe part of the effect may be thermal, but while the diminishing efficiency effect may be reduced with good thermal management, it won't be eliminated completely.

As semicnoductor devices, there are only a limited number of many generation and recombination sites to go around. Beyond a certain current, the LED will begin to saturate, and the probability of an electron "finding" an efficient pathway to get through and generate a photon will get lower and lower, where the probability of an electron losing energy due to resistance (from impurities, or the semiconductor crystal structure) will go up as the current goes up.

Heat will have the effect of increasing the lattice vibration in the crystal, which will make the effective resistance higher. However, heat or no heat the LED will still saturate above a certain current.


On the low end, somewhat more interestingly, there is an effect where LED efficiency goes down for extremely LOW currents as well. What is happening here is that it takes at least 2.8V to generate a blue 450nm photon (assuming a perfect LED, in a real device the necessary voltage may be higher) If you drive a blue (or white) LED at less than that voltage however, there still will be some measurable current, and there still will be some photons made -- but at a longer-than-ideal wavelength. Thus, both the emitter and the phosphor will operate less efficiently, and the device will actually have lower efficiency. This is also the threshold at whcih noticeable color shifting of the LED will start to take place.

A way around this if you want extremely low output without efficiency and color shift, is to use PWM.

 

[JohnR66]

I recall seeing graphs of LED output vs. temp. The power LED, Luxeon I think, showed huge increase in efficiency at very low temps.

While chilling a LED is not practical in normal use, It would be interesting to see how efficient A chilled LED can be across the current range.

LED headlights would be very bright in the winter unless the controller takes temperature into consideration.

 

[jirik_cz]

With 2,5A current and very low temperature you can get 500 lumens from single Cree XR-E R2. https://www.candlepowerforums.com/posts/2546760&postcount=200

jtr1962 makes really great job in that thread. I'm surprised that it didn't get more attention.