How much more room for improvement is there with LEDs? Will they become much better?

Vox Clamatis in Deserto

Flashlight Enthusiast
Joined
Jul 7, 2006
Messages
1,126
You are wise, but for superconductors to work, they have to be at ridiculously cold temperatures anyway - correct?

Yes they would, at least as far as we know right now. And discussion of quantum effects here on a hobbyist flashlight forum is perhaps usually not a good idea. But, we were talking about theoretical limits to LED efficiency so I was questioning the blanket premise that current flow always implies ohmic heating.

Maybe those electrons travel on Cooper tires now that I think about it...
 

wrf

Newly Enlightened
Joined
Jan 7, 2009
Messages
124
This thread needs a skeptic. I'll volunteer.

I think the advent of the blue LED gave a boost to LED lighting similar to how the advent of jet engines gave a boost to air travel.

I think most of what we are seeing is the maturation of that technology and also the related phosphor methodologies. Given that, I expect advancements to level out. Sure, there will be lab examples that do wonderous things. But when I got off the plane last week, I was getting off a 737, not some supersonic wonder.

I grew up in the dark ages. The 5mm Panasonic red/yellow/green LEDs I mail ordered from the Digi-Key flyer (yes flyer, not thousand page catalog like today) would not overdrive at all. My first LED light lasted all of seven seconds.

It has been a very long wait for any meaningful advancement, which in this case was the blue LED. You just can't take these advancements for granted.

[...]The electrons flow in Cooper pairs, right?:)

Dang, now I have to Google "Cooper pairs"!!!!
 
Last edited:

jtr1962

Flashaholic
Joined
Nov 22, 2003
Messages
7,505
Location
Flushing, NY
This thread needs a skeptic. I'll volunteer.

I think the advent of the blue LED gave a boost to LED lighting similar to how the advent of jet engines gave a boost to air travel.

I think most of what we are seeing is the maturation of that technology and also the related phosphor methodologies. Given that, I expect advancements to level out. Sure, there will be lab examples that do wonderous things. But when I got off the plane last week, I was getting off a 737, not some supersonic wonder.
Yes, we're already seeing some leveling out in advancements. For example, we're never again going to see efficiency double practically overnight. I think we'll be on a long, fairly slow evolutionary path from this point onwards, where we might see gains of 10% or 15% a year, until we finally plateau for good near theoretical limits.

I'm not sure if the analogy with air travel is apt here. If one takes the speed of sound as a sort of limit then maybe. Jet engines did allow us to fly fairly close to the speed of sound and then progress, at least in regards to increasing speeds, stopped there. However, the speed of sound is more an economic limit than a theoretical one. If we had some inexpensive energy source, I've little doubt we would be flying in hypersonic transports now on suborbital trajectories. LEDs on the other hand have no such barrier to improving all the way to their theoretical maximums. The only reason they won't is if we can't figure out how, or if the only available means to gain a few extra percent results in much higher costs.

I grew up in the dark ages. The 5mm Panasonic red/yellow/green LEDs I mail ordered from the Digi-Key flyer (yes flyer, not thousand page catalog like today) would not overdrive at all. My first LED light lasted all of seven seconds.
I remember LEDs back in the "dark" ages. So-called "sunlight viewable" LEDs were actually a big deal when they first started appearing in the 1980s. I remember how I actually considered those 50 or 100 mcd LEDs really bright (they were compared to everything else available at the time).

It has been a very long wait for any meaningful advancement, which in this case was the blue LED. You just can't take these advancements for granted.
I played with my first blue LED in the mid 1990s, and my first whites around 2000. My thoughts on the latter were one day these will take over the world. I knew that with anything electronic it's only a matter of time before we figure out how to tweak it to obtain maximum possible performance. We went from efficiencies under 1% to 60% with blue LEDs in about 15 years. I think we'll hit 90-95%, but it'll probably take another 15 years.
 

blasterman

Flashlight Enthusiast
Joined
Jul 17, 2008
Messages
1,802
Using the jetliner analogy; while the typical cruising speed of airliners at altitude has stayed about the same for several decades, what has continued to improve at a regular pace is fuel efficiency and engine reliability.

In terms of LED technology we've continued to see an improvement in lowered forward voltages and improved thermal management even though absolute efficiency improvements are obviously dealing with diminishing returns. Also realize that there has to be economic incentive for companies like Cree, Phillips, etc. to keep dumping $$$ into R&D rather than the same investment being used to improve existing production efficiency. Righ now the gulf between hyper efficient LEDs and mass produced fixtures that can utilize them is as big as it's ever been. A 3-5% efficiency improvement won't sell any more reels, so it's often not worth pursuing.
 

SemiMan

Banned
Joined
Jan 13, 2005
Messages
3,899
:welcome:

Cree makes the LED die (the part that actually makes light) in the US and assembles the whole LED (die + case + optics + wiring) in both the US and China. Because the most important part is not made in China the Chinese are unable to make illegal bootleg copies. (There are identical looking products but they do not perform as well.)

-

EDIT
The old Luxeon is about 30 lumens per watt.
The Cree XRE is about 80 lumens per watt (P4 bin) to 114 lumens per watt (R2 bin)
The Luxeon Rebel is 60-90 lumens per watt.
The Cree XPG is about 139 lumens per watt.
The Cree XML is about 160 lumens per watt.
As mentioned earlier:
Cree experimental 200 lumens per watt.
Theoretical limit for blue/phosphor LEDs 300 lumens per watt.
-
While RGB LEDs can go higher they only have red, green, blue. No orange, yellow, cyan, or violet. Not a problem in LED TVs and computer monitors but may be a problem elsewhere.


Rebel, is 110-120 lumens/watt, Rebel-ES, 120 -130+ lumens/watt.
 

cdrake261

Enlightened
Joined
Jun 26, 2011
Messages
484
Location
Columbus, IN
Using the jetliner analogy; while the typical cruising speed of airliners at altitude has stayed about the same for several decades, what has continued to improve at a regular pace is fuel efficiency and engine reliability.

In terms of LED technology we've continued to see an improvement in lowered forward voltages and improved thermal management even though absolute efficiency improvements are obviously dealing with diminishing returns. Also realize that there has to be economic incentive for companies like Cree, Phillips, etc. to keep dumping $$$ into R&D rather than the same investment being used to improve existing production efficiency. Righ now the gulf between hyper efficient LEDs and mass produced fixtures that can utilize them is as big as it's ever been. A 3-5% efficiency improvement won't sell any more reels, so it's often not worth pursuing.

That's providing all the led manufactures do not complete for having the most efficient led... The "my products is better than your's" comes to mind here.
 

slebans

Enlightened
Joined
Mar 1, 2010
Messages
457
Location
Moncton, NB Canada
Yes, we're already seeing some leveling out in advancements. For example, we're never again going to see efficiency double practically overnight. I think we'll be on a long, fairly slow evolutionary path from this point onwards, where we might see gains of 10% or 15% a year, until we finally plateau for good near theoretical limits.

From a luminaire/bulb standpoint, I believe we will see annual efficacy increases exceeding 20%. The 25-50% annual cost reductions of the LEDs themselves will allow the manufacturers to use more LEDs at lower power/higher efficacy levels. Integration of the various luminaire/bulb components and increasing power supply efficiencies will also aid in reducing costs.

Stephen Lebans
 

blasterman

Flashlight Enthusiast
Joined
Jul 17, 2008
Messages
1,802
That's providing all the led manufactures do not complete for having the most efficient led

They aren't competiting as much as people think.

LED makers make money selling large batches of LED's to fixture/bulb makers, with Cree being a rare exception in that they make some of their own fixtures.

An LED that's 10% brighter than a competitor doesn't mean much if the competitor costs 40% less. Or 80% less in the case of Chinese knock-offs {sigh}
 

Kinnza

Enlightened
Joined
Nov 15, 2005
Messages
223
Location
Spain
LED's manufacturers huge effort to increase efficiency past of that alternative technologies now is ended. Once all main LED manufacturers have high power white LEDs achieving 150lm/W on operating conditions, we still continue seeing efficiency improvements, but gradual and progressive, until, luckily, reaching 250lm/W on 10 years. As higher the efficiency level, the smaller in percentage the new improvements.

Now the focus is on lowering manufacturing costs. Many are working on the migration to silicon substrate, which technology is very well developed for mass production, and allowing the use of larger wafers, a key to lowering costs. But LEDs grown on silicon have lower efficiency, at least for the moment. But if they cost 1/3 for 1000 lm, they finally will displace other more expensive technologies, despite an small lower efficiency.

For sure than a flashaholic prefer to pay 3x for best LED and get the best color quality, the longer runtime and easier thermal management. But LED manufacturers thinks on the big market, not on us.

So it is to expect a slower rate on efficiency improvement from now on, but prices dropping continuously for a given amount of lm. Once LEDs have the performance to compete effectively with other lighting alternatives, new manufacturers, especially asians, currently producing silicon semiconductors will join, production capacity will grow very fast and later prices with go down as consequence of effective competence, that until now has been relatively limited.
 

Vox Clamatis in Deserto

Flashlight Enthusiast
Joined
Jul 7, 2006
Messages
1,126
Using the jetliner analogy; while the typical cruising speed of airliners at altitude has stayed about the same for several decades, what has continued to improve at a regular pace is fuel efficiency and engine reliability.

Cruise speed for airliners has actually dropped in the past three decades from what I can see. Legacy long haul planes like the 707 and 747 typically cruised at .86 mach or faster. 'Newer' designs like the 767 dropped to .82 or slower. Airbuses are usually slower than comparable Boeings, you gotta watch out for those trailing edge bird strikes.:grin2:

LED's are a game changing developing technology, like jet engines were when they were introduced to commercial aviation, eventually far more reliable and for some power ranges much more efficient than the earlier technology.

I would suggest an even closer analogy might be the transition from vacuum tubes (valves) to transistors in electronics. The heated filament active component is replaced by a more reliable and longer lasting solid state device.
 

Kinnza

Enlightened
Joined
Nov 15, 2005
Messages
223
Location
Spain
For those interested on projections of performance/price of LEDs, take a look on DOE page: http://www1.eere.energy.gov/buildings/ssl/information_resources.html, and specifically, at Technology Roadmaps.

Last projection I am aware of states

Efficacy (lm/W): Price ($/Klm) of coolwhite (4750-7050K, CRI 70-80) and warmwhite (2580-3710K, CRI80-90) (Klm=1000lm), OEM prices.

2010: 134lm/W:13 $/Klm (CW) 96lm/W:18 $/Klm (WW)

2012: 176:6 141:7.5

2015: 224:2 202:2.2

2020: 258:1 253:1


On the detailed roadmap, each topic able to be improved is analyzed in deep.
 

jd_oc

Newly Enlightened
Joined
Oct 15, 2009
Messages
114
For those interested on projections of performance/price of LEDs, take a look on DOE page: http://www1.eere.energy.gov/buildings/ssl/information_resources.html, and specifically, at Technology Roadmaps.

Last projection I am aware of states

Efficacy (lm/W): Price ($/Klm) of coolwhite (4750-7050K, CRI 70-80) and warmwhite (2580-3710K, CRI80-90) (Klm=1000lm), OEM prices.

2010: 134lm/W:13 $/Klm (CW) 96lm/W:18 $/Klm (WW)

2012: 176:6 141:7.5

2015: 224:2 202:2.2

2020: 258:1 253:1


On the detailed roadmap, each topic able to be improved is analyzed in deep.

If this is true, in 4 years we'll be able to have quad xpg-type led P60 drop in for $8 + electronics and reflector that produces 3000+ lumens. Count me in!
 
Top