One interesting application for super-bright LED's is home theatre projectors. Until very recently such projectors used lamp modules (usually metal-halide) that typically have an operating lifespan of a few thousand hours at most. That's a very rough estimate. Failures can happen sooner and they occasionally implode. Also, brightness usually drops off gradually as lamps age. New lamp modules usually cost a few hundred dollars. These lamps also produce a large ammount of heat and require active cooling, which makes projectors noisy unless carefully designed for quiet operation. For these reasons, home theatre projectors probably remain more of a niche product than they might otherwise be.
LED based lamps are starting to show up in this market sector. Current LED-based projectors are mostly portable projectors that offer low brightness and poor image quality, but some home theatre models of decent quality are starting to appear. At present, they're expensive, less bright than most projectors based on traditional lamps, and still require fans for active cooling. However, as LED's become more power efficient and economical, these projectors will hopefully become brighter, passively cooled, and significantly cheaper.
Projectors are not appropriate in many environments, especially those with high ambient light levels, but LED's may help them make major inroads into the big-screen market.
Another interesting application: growing vegetables very efficiently (potentially in urban areas).
"Plants mainly need blue and red light for photosynthesis and far-red, a colour not even visible to the human eye but visible to the plant….."
This makes LEDs very interesting for this purpose - the ideal light spectrum can be achieved with LEDs. This will make more and more sense over time as the world urbanises further and LEDs get ever cheaper.
For more stationary projectors, couldn't you just make a larger LED array and focus the light through a lens, rather than try to do cram super bright, and presumably expensive LEDs into a regular bulb sized package?
It makes the compelling argument that people are addicted to light. "Tsao calculates that, as a result, light represents a constant fraction of per capita gross domestic product (GDP) over time; the world has been spending 0.72 percent of its GDP for light for 300 years now. If there are other energy markets that show a constant percentage of GDP expenditure over time, Tsao doesn’t know of them."
Certain things are cheap in the rich world that are expensive in the 3rd world. Sewage, electricity, and water are so cheap in the 1st world, that we don't think about the cost.
The first time I saw 100% adoption of compact fluorescents was in Cambodia. They pay $.40/KWhr. That is insanely expensive (unless you live in Germany).
Lighting technology disproportionally benefits the poor rather than than the rich. Anyone that works on it is my hero. My family bought power for $.04 KW/hr, so I could study at night.
I'm genuinely curious whether new homes will start wiring lighting for DC to accommodate all these new LED bulbs? I just don't have a good sense whether its more efficient to transform AC to DC at a central hub in the home to distribute to all rooms, or more efficient to do it at the bulb itself.
The Emerge Alliance[1] is an industry group that has been working on this issue, though they've started with a spec for datacenters (as stephen_g alludes to). I think the consensus is that it is definitely more efficient to convert to DC at a single point, but one located as physically close to the loads as is practical, because of the line losses mentioned. Definitely something that will make sense for commercial buildings/campuses and multi-unit residential, and single family homes are not too far behind, I think. I can picture homes sometime soon just being wired for AC in the kitchen/garage/utility rooms with a LV DC grid elsewhere. The lower safety requirements for low-voltage really open up possibilities for system design and integration in the built environment.
Line losses through the cables are usually a killer with really low voltage DC (like 5V or 12V), so I think that AC would be more efficient there.
But some datacentres are using 48V DC now and then just piping that through a switching regulators in the servers instead of going from AC to DC. That could be workable.
Interestingly, I'm looking to do exactly this in a new home (new to me, not newly constructed).
I'll soon be moving to a park home (a relatively-permanent mobile home) that is in need of a fair amount of renovation.
I'll be running all lighting, computers and the TV off 12v DV circuits powered by bank of vehicle batteries and where the batteries are charged from 12v solar panels.
They emit more blue light than incandescent bulbs [1]. If people light their households with LED lights at night, it might shift their circadian rhythms [2]. Screwed up circadian rhythms can have all sorts of negative health/productivity effects.
I'll be trying my hand at some LED strip / arduino programming to build a dawn alarm that goes from dim red to bright white progressively. I have a Philips halogen dawn alarm which naturally goes from orange -> red as it ramps up, but unfortunately doesn't get quite as bright as I would like for that "wake up at the cabin" experience.
I'd love to see more applications of LEDs providing the right color temperature and intensity for the time of day, as well as more applications that avoid the bulb form factor. I'd love for my ceiling to emit light like the sky...
That thought occurred to me as well. Circadian phase delay already seems to be a problem re: the increasing brightness of our computing devices, almost all using LEDs.
New phones, tablets, TV's, etc. should come with a warning to keep them "dimmed down" in the evening, or better, keep use to a minimum late in the day. Sure, I know, the odds are about zilch people would actually heed such advice, but it ought to be out there anyway.
Another thing is the the relatively low CRI. For many purposes (industrial, medical, artistic) the spectral output of bright LEDs is far less than ideal. LEDs will probably get closer to "full-spectrum" over time, and no doubt will easily beat the discontinuous spectrum of fluorescent lamps.
LEDs now just need to get cheap enough, and really need to work with existing dimmer controllers.
Interesting link. The contrarian in me says it only looks at a couple bulbs, and in my (limited) experience with a variety of new low-end LED bulbs the light quality varies greatly and has also been improving, so maybe this won't end up being much of an issue on circadian rhythms at all. But it's an interesting factor I hadn't been considering (how do LED's compare in this to the CFL's that have already replaced incandescents for many, I wonder?)
Exponential growth is awesome! However, I was curious about the maximum theoretical lumens per watt to determine when the growth will fall flat. A current LED bulb gives about 60 lumens per watt. While an ideal monochromatic light source could give 683.[1] So it looks like there is _only_ one order of magnitude left in efficiency gains.
That 683 lumens per watt number is, as you say, for a monochromatic light. This is not useful for humans to see with[1]; the actual limit for human vision is about 300 lumens per watt[2]. That's an upper bound even ignoring losses, and some LEDs are around 70 now, so the remaining efficiency gains we can make seem limited to about 3x.
He makes some dubious efficiency claims. It's true the colored LEDs have always been pretty efficient, but that's no use lighting your home. White LEDs have surpassed incandescent bulbs but still have a way to go to reach flourescent tubes and the grand-daddy of efficiency - gas discharge lamps.
One (the?) reason for low power levels is they generate so much heat and are difficult to cool. LEDs are still primarily heaters - useful in the winter.
I think you need to update your figures. High-end consumer luminaires have been coming in above 100 lm/watt total system efficiency for at least a year or so now, and with CRI superior to fluorescents.
Several years ago I designed 1,500 W LED-based light source using extremely tightly packed LED's. Thermal management was a huge challenge. It took over three months of constantly running FEA thermal tests as well as physical tests to zero-in on an innovative approach to cooling the array. Crazy project. The surface of the emitter was measured at over 60,000 candelas per square meter. You simply could not look at the thing directly, it was really dangerous, almost like working with lasers.
beloch|11 years ago
LED based lamps are starting to show up in this market sector. Current LED-based projectors are mostly portable projectors that offer low brightness and poor image quality, but some home theatre models of decent quality are starting to appear. At present, they're expensive, less bright than most projectors based on traditional lamps, and still require fans for active cooling. However, as LED's become more power efficient and economical, these projectors will hopefully become brighter, passively cooled, and significantly cheaper.
Projectors are not appropriate in many environments, especially those with high ambient light levels, but LED's may help them make major inroads into the big-screen market.
patrickk|11 years ago
"Plants mainly need blue and red light for photosynthesis and far-red, a colour not even visible to the human eye but visible to the plant….."
This makes LEDs very interesting for this purpose - the ideal light spectrum can be achieved with LEDs. This will make more and more sense over time as the world urbanises further and LEDs get ever cheaper.
http://www.bigpictureagriculture.com/2011/02/plantlab-nether...
grimman|11 years ago
Scaevolus|11 years ago
It makes the compelling argument that people are addicted to light. "Tsao calculates that, as a result, light represents a constant fraction of per capita gross domestic product (GDP) over time; the world has been spending 0.72 percent of its GDP for light for 300 years now. If there are other energy markets that show a constant percentage of GDP expenditure over time, Tsao doesn’t know of them."
jcampbell1|11 years ago
The first time I saw 100% adoption of compact fluorescents was in Cambodia. They pay $.40/KWhr. That is insanely expensive (unless you live in Germany).
Lighting technology disproportionally benefits the poor rather than than the rich. Anyone that works on it is my hero. My family bought power for $.04 KW/hr, so I could study at night.
witty_username|11 years ago
seanalltogether|11 years ago
ghostly_s|11 years ago
http://www.emergealliance.org/About/Vision.aspx
stephen_g|11 years ago
But some datacentres are using 48V DC now and then just piping that through a switching regulators in the servers instead of going from AC to DC. That could be workable.
webignition|11 years ago
I'll soon be moving to a park home (a relatively-permanent mobile home) that is in need of a fair amount of renovation.
I'll be running all lighting, computers and the TV off 12v DV circuits powered by bank of vehicle batteries and where the batteries are charged from 12v solar panels.
RoboTeddy|11 years ago
They emit more blue light than incandescent bulbs [1]. If people light their households with LED lights at night, it might shift their circadian rhythms [2]. Screwed up circadian rhythms can have all sorts of negative health/productivity effects.
[1] http://www.designingwithleds.com/measuring-light-quality-phi... [2] https://justgetflux.com/research.html
peatmoss|11 years ago
I'd love to see more applications of LEDs providing the right color temperature and intensity for the time of day, as well as more applications that avoid the bulb form factor. I'd love for my ceiling to emit light like the sky...
jrapdx3|11 years ago
New phones, tablets, TV's, etc. should come with a warning to keep them "dimmed down" in the evening, or better, keep use to a minimum late in the day. Sure, I know, the odds are about zilch people would actually heed such advice, but it ought to be out there anyway.
Another thing is the the relatively low CRI. For many purposes (industrial, medical, artistic) the spectral output of bright LEDs is far less than ideal. LEDs will probably get closer to "full-spectrum" over time, and no doubt will easily beat the discontinuous spectrum of fluorescent lamps.
LEDs now just need to get cheap enough, and really need to work with existing dimmer controllers.
jp555|11 years ago
Regarding circadian rhythm, light from near-sleep-time television watching is much more of a concern, but that's been a problem for at least 30 years.
joshuahedlund|11 years ago
montecarl|11 years ago
1: http://en.wikipedia.org/wiki/Luminous_efficacy
jzwinck|11 years ago
[1] http://www.cool.conservation-us.org/byorg/us-doe/color_quali...
[2] http://phys.org/news202453100.html
Yardlink|11 years ago
One (the?) reason for low power levels is they generate so much heat and are difficult to cool. LEDs are still primarily heaters - useful in the winter.
ghostly_s|11 years ago
robomartin|11 years ago
johnm1019|11 years ago
jsilence|11 years ago
jonathansizz|11 years ago