LED Lighting As Cheap As CFLs Invented 553
mcgrew writes "New Scientist reports that a British team has overcome the obstacles to cheap LED lighting, and that LED lamps as cheap as CFLs will be on the market in five years. Quoting: 'Gallium nitride cannot be grown on silicon like other solid-state electronic components because it shrinks at twice the rate of silicon as it cools. Crystals of GaN must be grown at 1000C, so by the time a new LED made on silicon has cooled, it has already cracked, rendering the devices unusable. One solution is to grow the LEDs on sapphire, which shrinks and cools at much the same rate as GaN. But the expense is too great to be commercially competitive. Now Colin Humphreys's team at the University of Cambridge has discovered a simple solution to the shrinkage problem. They included layers of aluminium gallium nitride in their LED design... These LEDs can be grown on silicon as so many other electronics components are. ... A 15-centimetre silicon wafer costs just $15 and can accommodate 150,000 LEDs making the cost per unit tiny.'"
Whoopie for cold light! (Score:4, Interesting)
Cheap by what measure? (Score:3, Interesting)
So what is the energy consumed during production for one of these LED lights?
If we're just using more energy per unit during manufacture, then what is the energy payoff balanced vs. the number of hours these will remain in service?
Gallium (Score:3, Interesting)
Aren't we already in short supply of gallium? Do we really want demand for this rare metal that already has so many uses? We have plenty of ways to generate light, let's use one that doesn't require one of our rarest and most useful materials.
Re:If they are still not dimmable they still suck (Score:3, Interesting)
As long as your dimmers are the electronic PWM based kind, LEDs will work with them. (CFL's have issues when they're only given part of the sinewave). If you use the crappy rheostat dimmers LEDs won't work with them.
The only thing I question about LEDs is... will they give us headaches from the flicker? It's my current annoyance with LED christmas lights - they can flicker quite horribly. CFLs not so much, and neither the standard flourescent lamps (flicker at 120Hz, plus smoothed out by phosphors).
I forsee cheap LED lamps not using a full bridge rectifier (= 120Hz flickering), and just hooking them straight up in series, so they're off half the time.
It's also one of my concerns with OLED displays - they have these really fast refresh times, but if you don't refresh them fast enough, flicker! (I've seen it on cheap MP3 players, annoying. I've also seen it on the Sony OLED TV... Sony sales guy blamed it on the 1080p24 source, but I'm not sure.)
What about the production? (Score:5, Interesting)
I'm all for getting away from CFLs, as their production alone is NOT environmentally friendly (most of the mercury in the world is mined in China, where HALF of it is "lost" to the environment during production, which means "polluted"), not to mention the ratio thrown out.
But what about the LEDs? How toxic (or not) are the materials they're talking about? And what about the production of such? And heck, back on the pollution thing, WHERE they are produced makes a big difference, since if it's in China, forget any environmental disposal of chemicals used, whereas if it's in a developed country, it'll probably be OK.
Not insurmountable problems, but I do want to know how those things will work out.
Big advancement (Score:5, Interesting)
I'm not sure everyone is completely aware of how big an advance this is. I'm going to buy Philips' stock as soon as I can. I'm sorry people have been screwed by some misleading marketing, but LEDs are the future of lighting... and the big green movement.
And yes, they're really easy to dim, either by converting to DC and modulating current, or by using a PWM - I'm not sure which is more efficient/cheaper.
I can't wait for CFLs to go away. Eventually you'll see commercially available, color-selectable LED bulbs.
Anyone know if the process was patented/sold to a specific company? Pretty obvious why...
Solar panels too? (Score:3, Interesting)
Aren't some solar panels made with GaN as well? Will this help them too?
Re:you sir are incorrect (Score:3, Interesting)
PWM.
Basic EE: http://letmegooglethatforyou.com/?q=PWM&l=1 [letmegoogl...foryou.com]
Good for displays too (Score:4, Interesting)
This is great news not just for lighting, but also potentially for ILED TVs (basically LED - the "I" stands for inorganic. It would be simpler than even OLED, and the lifetime would be amazing of course.
Re:What about the production? (Score:3, Interesting)
Re:What about the production? (Score:3, Interesting)
Wow, that looks awesome... but I'm just not prepared to spend $70 for a flashlight.
Don't need to. Go to target or walmart, they will have lights in the $15-$30 range that do 120 lumens too.
The bare emitter is only a buck or two, so even "cheaper" flashlights are starting to use them.
Unfortunately, even the brightest emitters can't do much more than 800 lumens and still cost in the $10-$15 range, so real incandescent replacements for the home are still either underpowered (a 90 watt incadescent can do roughly 1800 lumens) or prohibitively expensive.
Comment removed (Score:2, Interesting)
Re:What about the production? (Score:2, Interesting)
The mercury in CFLs happens to be in an organic form meaning its already bonded with Carbon. That is the best form to enter the waterways and the food chain. Its on the order of more than trillions of times more likely to enter the food chain than elemental mercury.
Lumens per watt is? (Score:4, Interesting)
Granted this is an as yet unrealized technology, but I really wonder what the lumens per watt will be.
For reference a standard 48" T8 fluorescent is about 80 lumens per watt,
Compact fluorescent is around 65 lumens per watt,
and a 60w A19 bulb (the "normal" light bulb) is about 15 lumens per watt.
Cree has some Gallium nitride LEDs that (they claim) produce a record breaking 130 lumens per watt,
but they also have some 100 lumens per watt LEDs and they're also gallium nitride.
Cost of production isn't totally meaningless, but over it's 20 year life, 5 watts means more than $100.
Re:What about the production? (Score:3, Interesting)
It's all about the tight beam. LED lighting is much dimmer overall than CFL, but if you only shine light in a 12 degree arc it will be 100 times as bright as the same light in a 120 degree arc. If you *want* a narrow beam, LEDs are a clear winner. If you want to light a room, not so much.
In any case, gallium is scarce and there's not an easy way to increase production (there's no such thing as a gallium mine, we get gallium as a byproduct of aluminum production). If there suddenly a spike in gallium demand there will be a huge spike in gallium price. Any LED or solar technology that needs gallium or indium will be quite expensive if it becomes popular.
not everybody's power comes from coal (Score:3, Interesting)
There are a lot of places where the majority of power comes from nuclear and hydro.
Yes, and both of those power sources use a lot of energy to build. Then for nuclear there's pollution from mining as well as the waste.
Also, the real risk of CFLs is caused by the fact that any pollution from it is local and concentrated as a point source
Over a period of years I replaced almost all of my incandescent light bulbs with CFLs. As one burnt out I got a CFL to replace it. A point source of mercury is easier to handle than air born mercury. As for possibly breaking one, I try to use measures to reduce any possibility of breakage.
Also, consider that the plumes from garbage dumps invading your water supply
CFLs are supposed to be recycled and not thrown in the trash. Of course some people do throw them away, either because they don't care or because they don't know better. Even so, that's still less mercury in the environment than the amount of mercury that would be emitted to produce the power to light today's incandescent lights.
Which is why I hope these LEDs, or others, that are good for area lighting come onto the market within a couple of years.
trace mercury emitted into the atmosphere a hundred miles away from the city....
Not all power plants are 100 miles away from the city. There's more than one power plant in South Bronx [cnet.com]. NYC has 25 plants serving it [worldchanging.com]. The first ones built by Con Edison, used the used steam to heat neighborhood buildings.
Falcon
Re:We need lumens ratings (Score:3, Interesting)
I will verify this experimentally when I get home; I have diffraction gratings, I have "pure" (amber and orange-red) power LEDs available. If you're lucky, I'll get a picture.
Re:Ok, let's get this thread straightened out. (Score:4, Interesting)
Problem is: any recommendation I make is going to be based on the LED's we have, and most of those are either not for sale yet, or are stupidly expensive. I honestly don't know what commonly available lightbulbs have which LED's, either, because only a few LED's intended for general illumination are easy to recognize. I can tell you that I have great respect for Philips LumiLEDs, Osram Dragons, and most of the stuff Cree makes. I've worked with all those extensively and they're beautiful.
A white/red combo might look okay. It's good for general illumination but still not particularly great for really good color rendition. The problem is that color rendering is dependent on a *lot* of weird stuff. (For instance, part of the reason a ruby is so red is because it's absorbing shorter-frequency light, and re-emitting it as red light, so even if you have a good red source that's lacking in the short-wavelength area, the ruby will look dead. Fish have lots of fluorescent/iridescent coloration, so a standard color rendering index (CRI) test might not tell you what you want to know: will this light look good with this situation?)
With the li-ion, they'll burst into flame, so the manufacturers *have* to put in good charge-controller circuitry. LED's are currently like cars in 1910: you can make anything that sort of vaguely works, and someone will buy it. It's no secret that the LED crowd are hoping so many people will buy them just because they're cool, that they'll jump the early adopter chasm before they've realized that they've mostly bought turkeys.
All the stuff we work with does, essentially, some sort of rectification, then dc-dc conversion, either buck, boost, or SEPIC, and ends up as a constant-current supply for the LED. We (and I'm sure a bunch of other people) have made drivers for really good dimming using standard triacs (ours is just about to hit the market) and they work *amazingly* well. You're limited in part by how well the triac works: a lot of the ones I've worked with have such crappy triggering circuits that the best you could ever get is roughly 60% dimming -- as in, you turn it on, keep coming up, and suddenly the light comes on at 40% brightness. You can then finesse it down to maybe 20%, but it's a pain. Thing is: we can digitize that and do clever stuff in software and come up with something that can essentially accommodate for the crappy triac, learn what it's doing (by sensing how it's chopping the AC) and produce a fabulous dimmer. We have one setup that can dim 10,000:1. If you get 5:1 out of a so-called dimmable compact fluorescent you're lucky. A good dimmer with an incandescent can do something similar: it can start producing heat before there's visible light coming off the filament. But I think only our (and similar) driver can get good performance out of cheap triacs.
With all that said, a cheap triac dimmer driving a cheap unregulated LED will work, but they're (in my opinion) objectionably flickery. A triac dimmer driving a good switching converter (without any detection stuff to try and decode what the dimmer's trying to send) will still dim because you end up starving the switcher when the power's off, but it's icky-looking. In my experience it's even more non-linear than just the cheap triac: when there's enough power it drives well (but as you say at full current/full brightness), and when there's no power it turns off, and in the between situation it thrashes. Most LED driver chips have dim pins so you can apply an external dim signal derived somehow from the signal if you want to add a bunch of external circuitry. (We're building stuff that integrates all that junk into the chip.)
>the idea that you're trying to keep the LEDs at the peak of their efficiency, it seems like a no-brainer.
It is. But very few people will actually *pay* for better circuitry, apparently. So instead we have to add bullet-point functionality to justify more research, and then add in good current regulation and the like as a side-effect of (say) cold-we
Re:Wrong bulbs (Score:3, Interesting)
(*tinfoil* as a crude reflector? Gimme an effing break! Have you ever actually installed any outdoor lighting??)
Re:Wrong bulbs (Score:3, Interesting)