Finally 300 Lumen Per Watt !!

[jugg2]

Great for all those Gandalf moments.

"Let me risk a little more light..."
:huh:

LMAO! Then people would be convinced I have lost my mind... Until I bumped it to high lol.

 

[Jarl]

Then we wouldn't see any improvement by adding the carbon nanotubes.

The idea is that you start with your emitter, which gives you x lumens/watt, then replace the phosphor coating with the nanotube coating, which will give you (say) 2x lumens/watt.

 

[joema]

...Give it 5 years...and we should be seeing 400+.

We'll never see much over 400 lumens/watt using our current physics -- it's an absolute limit -- like the speed of light. Even 400 is based on limited research; the normally accepted limit is somewhat lower, around 300.

100% efficient conversion at 555nm green monochrome light produces 683 lm/W. Of course no conversion is 100% efficient, and white light emitters are less efficient than monochrome.

These limits are bounded by basic physics. To exceed them would require a revolution on the scale of zero point energy or warp drive.

http://ledsmagazine.com/features/2/5/4/1
"The theoretical maximum for any light spectra is 683 lm/W (for 555 nm monochromatic radiation). For white light, the maximum is typically 300 to 350 lm/W....Yoshi Ohno's work at NIST has shown that it is theoretically possible for a RGB white LED to have a maximum of over 400 lm/W"

 

[RustyKnee]

We'll never see much over 400 lumens/watt using our current physics -- it's an absolute limit -- like the speed of light. Even 400 is based on limited research; the normally accepted limit is somewhat lower, around 300.

100% efficient conversion at 555nm green monochrome light produces 683 lm/W. Of course no conversion is 100% efficient, and white light emitters are less efficient than monochrome.

These limits are bounded by basic physics. To exceed them would require a revolution on the scale of zero point energy or warp drive.

http://ledsmagazine.com/features/2/5/4/1
"The theoretical maximum for any light spectra is 683 lm/W (for 555 nm monochromatic radiation). For white light, the maximum is typically 300 to 350 lm/W....Yoshi Ohno's work at NIST has shown that it is theoretically possible for a RGB white LED to have a maximum of over 400 lm/W"

going on that...Once/if those levels are reached I won't be fussed about impossible lumen/watt gains....i will be after better batteries that have higher capacity (if they haven't happened at the same time). They will be so efficient that they don't get hot even when driven hard (assuming the process is suitable for high lumen leds).

the more efficient the led the less heat it produces. more battery watts will be converted to light watts instead of heat watts.

Stu

 

[DM51]

the more efficient the led the less heat it produces. more battery watts will be converted to light watts instead of heat watts.

This is the most important aspect of greater efficiency. It enables higher light output without increased heat management problems.

 

[Mr. Bruin]

While my knowledge of flashlights is definitely lacking, I do understand nanotech and basic physics.

There are a few issues here with people's understanding of what's going on.

The paper uses nanocrystals, not nanotubes. Nanocrystals are inorganic and easily made by the gallon, as opposed to carbon nanotubes which suffer from the issues of sorting different types (semiconductor and metallic).

The theoretical limit for power to light conversion is simply a conversion of what the eye's response is to a perfect energy conversion of electrical power to white light. The easy part of this calculation is the energy conversion, but it's much harder to address the eye response. That's where all those different numbers come from.

We continue to close the gap on power to light conversion with LED improvements. However, this is only part of the equation. In a simple view, LEDs only emit a single color. That single color has to be made into the 3 visible colors, traditionally done using phosphors. This paper uses phosphor changes (switching to nanocrystals) to improve the spectral tuning of the LEDs after emitting.

I'm guessing one of the big changes is that phosphors are slow at doing the conversions. These nanocrystals are quicker. That's my guess. I can read the paper and find out -- though they might not know. It's a pretty obscure journal.

It's odd to me this is presented as "news." I met with Nichia back in the early days of blue LEDs about using nanocrystals as phosphors...

One thing to remember or think about when talking about the lack of large-scale production of nanotubes is the fact that you aren't trying to make something very large (say a bicycle frame) entirely out of a nanotube composite. The relative amount of nanotubes needed to 'coat' a tiny LED phosphor would be miniscule and a small lab quality set-up for nanotube production could probably realistically produce the amount needed for substantially large runs of the new LEDs to be produced. Probably .

 

[jrminky]

I wonder if this nanocrystal coating is suitable for use by fluorescent lights, too, and if so if it yields a similar increase in luminous efficacy?

A 28W T5 tube that outputs 8k+ lumens for $4 would do wonders for garages around the world.

 

[jtr1962]

Is it just me? I think LEDs and there "lunar white" are the purest whitest light we can produce currently. People just think incandescents are white because, it's what's they are used to, and that's what the package says. They are really more of a warm white or ivory.

It's not you at all. I've often questioned the logic of producing off-white LEDs myself when sunlight is really what we should be trying to emulate.

 

[Crenshaw]

I think there was a thread about this in the LED forum....

I cant wait for the future, imagine,

"Man arrested for lighting up large portions of the sky, police still searching for the impratically big spotlight they suspect he used, suspect insists they will never find it."

Crenshaw

 

[TorchBoy]

I've often questioned the logic of producing off-white LEDs myself when sunlight is really what we should be trying to emulate.

From an article in a recent New Scientist, it seems that blue light triggers our diurnal rhythms (or something). It seems that too much blue light late at night might keep us awake, so LED lighting after dark would be better if it's not high in blue.

The article also mentioned that even blind people can be affected by blue light in their eyes, and keep in step with the day because of it. Obviously, blind people without eyes are not affected in that way, and (it was claimed) have trouble sleeping because of it.

 

[Canuke]

I'm surprised no one has brought this up yet. Isn't the theoretical limit for white light(full spectrum) 242.5 lumen per watt? The theoretical limit for visible light is 683 lumen per watt at 555 nm(green).

Of course, it would be different depending on how much of the spectrum included.

Yup. that 242.5 lm/w limit is for full spectrum. To get past that point, you would have to sacrifice color rendition -- try to emit more photons closer to 555nm at the expense of those at the ends of the spectrum, in the deepest reds and violet/blues. It will be interesting to see what color of lights we end up seeing used for "high-efficiency" lighting, relatively speaking, in the days of near 100% conversion efficiency.

 

[mdocod]

DING DING DING DING DING

Post #37 and #51 By Gunner12 and Canuke hit the points that need to be made here.

Everyone is jumping at "300lm/w" with all this great excitement, and it took that many responses for someone to point out that an ideal white light source can never exceed 242lm/w???.... I'm sure some sacrifices can be adopted in light quality to achieve higher figures, but whether it's worth it or not is up for substantial debate... My gut says we'll see some major barriers around 175-200lm/w for ideal white light sources that will bring the advancements to a screeching halt.

 

[TorchBoy]

Yup. that 242.5 lm/w limit is for full spectrum. To get past that point, you would have to sacrifice color rendition -- try to emit more photons closer to 555nm at the expense of those at the ends of the spectrum, in the deepest reds and violet/blues.

How many lumens/watt if the light is all emitted at just the three frequencies we see at? And wouldn't that make colour perception of many objects really awful? If they are only coloured an in-between colour will they look black, or at least darker than they really are?

 

[jtr1962]

Everyone is jumping at "300lm/w" with all this great excitement, and it took that many responses for someone to point out that an ideal white light source can never exceed 242lm/w???.... I'm sure some sacrifices can be adopted in light quality to achieve higher figures, but whether it's worth it or not is up for substantial debate... My gut says we'll see some major barriers around 175-200lm/w for ideal white light sources that will bring the advancements to a screeching halt.

It's not that big of a sacrifice for much higher efficiencies. First off, the limit would be closer to 200 lm/W if we're talking about a light source identical in spectrum to the visible portion of sunlight. This would have a CRI of 100. However, by going to a CRI of 98 using 4 colored emitters we can get as high 360 lm/W while sacrificing very little color rendering. A CRI 98 lamp is plenty good enough for even critical color matching applications. By going with 3 colored emitters and sacrificing still more CRI, we can get as high as 400 lm/W with a CRI of 85. This is perfectly acceptable for all residential and commercial interior lighting. There is really no reason at all that we need to try to emulate a so-called ideal white light source when much higher efficiencies are possible with not much sacrifice in color rendering. Basically, what is lost is mostly very deep reds or purples, both colors for which hardly anything has pigments for.

So in conclusion, yes, if we can get to 75% efficiency then 300 lm/W is easily possible by properly chosing the center frequencies and bandwidths of the three colored emitters. Even using higher quality CRI 98 light we can get to 270 lm/W with a 75% emitter efficiency. A good idea would be to use the CRI 98 lamp for things like desk lamps or reading lamps where color perception might be more important. Outside of that, no good reason we need to use CRI 98 general lighting for our homes or offices, much less highly less efficient CRI 100. The difference between CRI 98 and CRI 85 is imperceptible to most people. And only sensitive spectrometers can tell the difference between CRI 98 and CRI 100, not our eyes, so why take a 45% efficiency hit for an improvement nobody can see?

 

[jtr1962]

From an article in a recent New Scientist, it seems that blue light triggers our diurnal rhythms (or something). It seems that too much blue light late at night might keep us awake, so LED lighting after dark would be better if it's not high in blue.

The article also mentioned that even blind people can be affected by blue light in their eyes, and keep in step with the day because of it. Obviously, blind people without eyes are not affected in that way, and (it was claimed) have trouble sleeping because of it.

You can get rid of the blue spike in LEDs without going to the extreme of having LEDs which look like incandescents. In fact, that's exactly what 3 color emitters will do for us. I think the rationale here has more to do with initial market acceptance of LED lighting. There is still a negative association of higher color temperatures with poor color rendering due people over about 25 remembering the old halophosphor cool whites. This is in spite of better fluorescents having been around for over 25 years. This seems to be gradually being overcome with the increasing use of neutral white CFLs (3500K). I'm also seeing more use of 5000K when I go for walks. Since the perception that whiter light = poor color rendering seems to be gradually going away, and shouldn't exist at all for the under 25 crowd, why take a step backwards by making LEDs imitate incandescent lamps? Granted, almost nobody wants the ~8000K to 10000K of earlier white LEDs, but we're now making nice 4000K to 5000K whites perfectly acceptable for interior lighting.

One thing I don't get is people will go through great pains to choose the colors in their homes, yet by using incandescent or similar lighting they must by necessity compromise either the way it looks during the day or at night. Sunlight type LEDs would make this problem go away.

 

[TorchBoy]

I don't think that really addresses the possible problem of too much blue light when we're trying to go to sleep, but quite apart from that, we simply may not want white light in the evenings. If someone came up with a new kind of candle that burned with a pure white light how romantic would they be? (The candles, that is.) People at the south end of this country in winter really do like warm coloured light - it helps them feel warmer - and I'm told that people in the tropics (not in this country) have a much higher preference for cool white CFLs than those southerners do.

 

[EntropyQ3]

I don't think that really addresses the possible problem of too much blue light when we're trying to go to sleep, but quite apart from that, we simply may not want white light in the evenings. If someone came up with a new kind of candle that burned with a pure white light how romantic would they be? (The candles, that is.) People at the south end of this country in winter really do like warm coloured light - it helps them feel warmer - and I'm told that people in the tropics (not in this country) have a much higher preference for cool white CFLs than those southerners do.

While there may well be a preference for warmer light through the association chain of warm light - evening - relaxation, that doesn't mean we need new candles.
We already have candles. Turn the electric light off and light the oldfashioned firesticks, and they'll add live flame, flicker and smell to the atmosphere that no electronic light will ever do.

Even as someone who cares a lot about colour fidelity (photography is a big interest), I'm at least half-way with jtr1962 on this. While I feel colour fidelity is more important than wringing the last percentages of efficiency out of LEDs, small compromises in exchange for large gains in efficiency makes a lot of sense in many situations.

But the core issue is that daylight should be the reference, not incandescent light. People holding onto the notion that the the ancient light bulb is "right" is a problem that continues to hamper adoption of fluorescent lightning, and which will carry over into the age of LEDs.

The biggest problem from an energy conservation standpoint is still that most light is produced unnecessarily. I don't see this changing, unfortunately. Indeed, higher efficiency lighting is likely to make this part of the problem even worse.

 

[e2x2e]

I have a severe case of spoiled kid syndrome, but I'm patiently waiting for LED technological advances such as this.

 

[SpencerF]

I hate to point out a little reading comprehension test from the original press release:

"The final LEDs were also better than commercially available LEDs at creating visible light, giving off more than 300 lumens of visible light for every watt of all light emitted."

The 300 lumens/watt claim is NOT electical power->white light conversion effeciency. It is the conversion of the blue light output, which is by convention measured in watts (not lumens), to white light.

A top bin Cree royal blue LED produces 400 mw of light output for just over 1W of electical input. So 300 lumens/watt*.4w -> 120 lumens output for ~1W of electical input. Not bad, but certainly not up to the title line.

Have fun,
Spencer F

 

[2xTrinity]

A 28W T5 tube that outputs 8k+ lumens for $4 would do wonders for garages around the world.

Forget garages, I actaully have a couple T5 strips in my bedroom that could benefit from those... Anyway, I think it's unfortunate that so many people find linear tubes only suitable for the garage, and think it's "weird" to have 28W linear fixtures indoors that are tastefully installed (ie, no direct line-of-sight to the bulb, either positioned undercabined, or indirectly in a "wall sconce"), yet have no problem installing 100W-incan equivalent CFLs in the same room that are much less efficient overall (net fixture efficiency).

IMO promoting lamps, wall sconces, and "residential looking" fixtures that used T5 lamps would ultimately produce a lot better results than screw-based CFLs, which are almost always slightly "off" compared to the incan bulbs they replace in the way they distribute light. Designing such fixtrues with true dimming ballasts would be even better. Finally, separate lamps/ballasts have MUCH better reliability.

Yup. that 242.5 lm/w limit is for full spectrum

At what color temperature? 242.5 is unusually precise, so there should be some details to go along with that. I've yet to hear a clear definition on this. There's a big difference between candlelight and direct sunlight at the equator, even though both are "full spectrum". For continuous-spectrum sources, more lm/w can be achieved in a lower color temp (green and even red light are better detected by our eyes than blue).

I don't think that really addresses the possible problem of too much blue light when we're trying to go to sleep, but quite apart from that, we simply may not want white light in the evenings. If someone came up with a new kind of candle that burned with a pure white light how romantic would they be? (The candles, that is.)

Generally, there is a trend where the lower the illuminance is, the lower the ideal color temperature is. So comparing a candlelight to the typical illuminance of even a modern incan-lit house, let alone a store or office, isn't a fair comparison. I don't believe higher-color temperature is necessarily better for all, but IMO a great deal many houses can, and do, look better with 3500k-5000k lighting.

I prefer ~5000k for garage/workshop light, 3500k for general purpose indoor, 3000k for low-level ambient lighting.

A top bin Cree royal blue LED produces 400 mw of light output for just over 1W of electical input. So 300 lumens/watt*.4w -> 120 lumens output for ~1W of electical input. Not bad, but certainly not up to the title line.

Have fun,
Spencer F

In that case, Cree's ordinary phosphors in warm white, which are already at Q2 efficiency (almost 100 lm/W of input power), will probably catch up to this LED in less than a year. Naturally neutral and cool will be slightly higher.