I love that piece, they had it set up at burning man this last year too. It really was stunning. I also agree that it's more appropriately called "LED Art" than what we did... the artistic part of what we did wasn't building the light, it was the paintings we made to work with it. Most mixed-media artists don't have the EE expertise to do it on their own, and I sure don't have the painting expertise to do it on my own. The goal of this article is to eventually create a howto so that artists will know exactly what paints to buy to work well with particular LED colors, techniques they can use to make neat effects, with examples. I'd call what we did "LED+Art" instead of LED Art, but I guess that's what happens on translation.
Color Kinetics (now Phillips) is the standard for RGB lighting fixtures. They're just expensive. If you want to buy the lights I made in the article, I'm up for making some custom fixtures that are a bit nicer on the hardware end than the CK ones, but not quite as polished from a software standpoint. Certainly they'd work fine for just cycling through the rainbow in circles endlessly, and I'm setting it up to be designed for permanent installations (a hole in the back to just mount the light on a wall). Send me an email if you're interested - neltnerb@mit.edu, and we can talk about details.
Sorry, I don't know where it went and wasn't maintaining it. I assume it's still around somewhere, but I wasn't involved in setting it up and don't even know what computer it's saved on.
Sorry, this is simply not true. The eye can see any color inside the boundaries of the CIE chromaticity diagram (http://hyperphysics.phy-astr.gsu.edu/hbase/vision/cie.html). RGB colorspace is a subset of human vision. Colors which you patently cannot produce with RGB mixing, but can see in a rainbow, include: that awesome purple that you see at night on a really clear evening, true deep cyans, and true yellows. This can be seen very well on the illustration on the GSU website I linked to.
The eye is more awesome than you give it credit for.
We purchased a wide variety of high-end pigments -- http://www.goldenpaints.com/ was one we looked at a lot. The biggest challenge was finding paints that had only one pigment. I actually asked most of the big paint supply companies for absorption spectra for their paints, but none of them ever replied to my messages. I suspect that their salespeople don't even know what an absorption spectra is...
So the goal was to take the acrylic-style paints you'd find in an art store, and find the ones which were actually fairly close to being monochromates. I toyed briefly with the idea of synthesizing my own pigments (since I'm a material scientist and all), but I decided that it would be too expensive and difficult to reproduce.
The purpose of the first prototype "Color Flower" was actually to act as a poor-man's spectrophotometer. I'd turn on one color only at a time, and see what the intensity of the reflection was -- it wasn't great, but it gave me a decent picture of the absorption spectra at the points that really mattered (and cost $100 instead of however much a real spectrophotometer is).
I'm very confused as to why you got modded -1, but thanks for the explanation of the saturation. I'm still confused as to how a CCD seems to be able to take violet and magically interpret it as magenta -- is there special software that does this somehow? I know that when I use real violet LEDs, they frequently look blue in the CCD; perhaps most of the "purple" flowers we see are actually magenta?
I should type up the list we found of good pigments. Yellow and green were by far the hardest to find good monochromatic reflectors in. Almost all of the greens were dichromic, and the yellows just turned out dull and flat.
Sorry, I take it back. You're right as long as your primary absorbing colors are only Red, Cyan, and Yellow. It really boils down to the absorption spectra of the pigment materials. If I have a pigment that is a "true" yellow (reflects in the yellow, absorbs everywhere else -- including red, green, and blue), than it will absolutely appear to be black if you shine red+green LED light on it. You had me confused for a second, but I see where the misunderstanding is now. This is the entire reason we had to spend so much time finding pigments that were more than just mixtures of the primary colors.
I'll clarify the theory section to try to make this more clear.
Nah, djiwhatsit is right, I was wrong. Yellow pigment absorbs blue light strongly, so it should reflect both green and red fairly well and appear yellow. Thanks for pointing it out, I'll correct it and try to come up with a better way to explain it to a relatively non-technical audience. Examples with this sort of thing are fairly hard to come by, and will take some time to fine tune.
The idea I was trying to get across is that dichromates do not appear as you would expect when they are shone on absorbing paints, and that if you mix paints together, you can very frequently just get flat out ugly results. I didn't put up any movies of the paint splotch examples, but in the case of red and yellow paint mixed together to make orange, it *definitely* only appears red when both red and green lights are on =) Experimental evidence! But yes, the theoretical foundation should be clarified some more.
I think a much better example would be green paint. Green paint should absorb strongly both blue and red light. So, if you shine dichromatic red+green=yellow light on that green paint, it will appear to be green only. The effect that many non-technical people would expect is that if "yellow" light is shining on the green paint, it will appear dark -- this is the misconception that I'm trying to correct.
The energy density is probably too low, but there are a variety of devices which use UV LEDs to sterilize bacteria and parasites in water. These UV LEDs were designed for use in curing polymers, but the total light output of two LEDs is still only about the same as a smallish blacklight fluorescent bulb. The neat part is that you can get fading UV light (so your fluorescence fades in and out), but the total power output still isn't even close to that of a fluorescent bulb.
Frostbyte worked at Color Kinetics, by the way... and no, everyone agrees that if I ever wanted to sell the lights, I'd need to pay 5% of the proceeds to Phillips. Unfortunately, no one out there seems to have come up with a silver bullet to kill the PWM patents yet.
Hiya,
I like the way you explain the behaviour of the light, and if you give me permission, would like to adapt parts of your explanation to improve the theory section of the page.
I'm trying my best to explain this so that someone with about the level of understanding of "light has a frequency" will at least sort of grasp the idea. Someone else mentioned that my understanding of RGB additive synthesis versus subtractive synthesis of colors is wrong, so I'll be rereading the chapter from Feynman's book on color vision again anyway to try to make it clearer. I've thought about it a lot over the last few years, but I fully acknowledge that there are some holes in my model of how the whole thing comes together.
In particular, this one bugs me. Why does red + blue look similar to purple? After a lot of training, I can distinguish between the two (now, magenta really looks like red+blue to me). However, it seems bizarre that a CCD can capture the color purple at all. It's certainly not actually emitting purple light, so it must somehow be converting the purple frequencies of light coming from, say, a flower, into magenta. How on earth would this happen? I can grant that a purple light would excite the blue pixels in a CCD, but it's hard to swallow that they'd magically also excite the red pixels in the appropriate proportion.
I feel like if I got an explanation of this, everything else would fall into place.
-Brian
I've never had the slightest problem before with assembling my boards. I'm usually too busy to spend lots of time on routing traces by hand, so I just carefully place components, let Eagle handle the autorouting, and clean up the worst of its mistakes. When someone starts paying me to design electronics full time, I'll worry more about aesthetic details that don't seriously change the functionality of the devices.
The paintings have fluorescent paint in them. It's pretty cool looking to have bright red light at the same time that your fluorescent paints are glowing -- very wild.
I know that at my university, and it seemed many similar universities, a huge draw for getting the best and brightest individuals around into the doctoral program is that if you invent something, you can actually make money on it. For instance, at MIT, one third of the proceeds of any patents go directly to the inventors, one third goes to fund the lab that they work in, and one third goes to MIT. This makes it massively more attractive for inventors such as myself to come here. It is true that my work is funded in part by the government, but the amount I am being paid is a pittance compared to the value I am creating here. The only reason it makes sense for me to do my research in an open, academic setting is because when I'm done, I can go and commercialize what I invent and make a living out of it.
I know that the same is true for many professors. How could a university reasonably convince an amazing professor and technologist to work there if they force the professor to give up all rights to anything they create to the public domain? No one in academia makes nearly enough money to justify removing their ability to capitalize on inventions they develop.
Beyond the need to attract top-tier students and faculty, universities being able to make money from inventions means that less public money goes to fund the research that goes on there. Instead, companies who benefit from that research are forced, through patent licensing, to fund the universities. It makes a lot of sense -- if a university does amazing mechanical engineering work, it makes sense that companies that want to capitalize on that work subsidize it instead of being able to use it and put the burden of funding that research more heavily on the public as a whole.
Instantly consume the solar electricity instead of increasing the amount of oil you need to instantly consume to then convert back into oil with solar. The result is the same, except that you have more fuel afterwards.
(Yes, I know I'm talking about an oil plant. Electricity doesn't care what the source is, and for every grid powered by coal, there's probably an oil plant somewhere contributing that could be removed or have its load decreased)
I feel like I should point out that this does not remove CO2 from the atmosphere.
Fuel made by this, when burned, releases the CO2 back into the atmosphere. I suppose the only reason they talk about doing it on coal plants is because the concentration of CO2 is higher where it's being used. However, as usual there's no way that the energy input is equal to the energy in the fuel, so what is the point? If you put that solar electricity on the grid, you need to burn that much less *actual* oil, so you have no energy loss at all...
I'm fairly experienced with these things, so I figure I'll offer up what it looks like to me in terms of efficiency.
The 80% number is the ratio of the energy contained in the hydrogen gas to the energy contained in the acetic acid plus the energy used in the form of electricity. However, the stated claim that this is more efficient than ethanol is not really justified based on the actual paper (which I read).
While it may well be better than ethanol (most things are), if we calculate out the actual efficiency of conversion between ethanol and hydrogen typically reported in the literature (Kugai and Deluga are the best papers IMO) we get around -165kJ/mol theoretical energy loss per mol of ethanol converted to hydrogen and carbon dioxide. For comparison, ethanol contains about 1145kJ/mol of energy to start with. That gives an efficiency of around 86%.
Of course, the "hidden" (not very well) cost of ethanol is that it takes massive amounts of energy to produce before you get to the stage of conversion into hydrogen. In the end, you get something between -30% and 30% efficiency converting seeds into ethanol (one well-known paper reported an actual energy loss, ignoring the energy from the sun of course, which will always mean an energy loss in conversion). This is not very good. However, this sort of analysis was *not* done in this paper, and wasn't claimed to have been done in the paper. It is very likely that this process is better than ethanol (most things seem to be), but the summary (and press release) are overstating the case, unless there is information they have that wasn't included in the actual paper.
And to clarify, carbon-neutral does not mean it produces zero CO2. It means that all the CO2 produced at one point came from the atmosphere instead of from fossil fuels. As long as the process is net energy positive, you can use the extra energy to fulfil the energy needs of the process, and remove the need for fossil fuels.
You should email this story to as many people as possible. I think I've heard of some places you can go that'll distribute your message to millions of people for pennies.
That's funny, I didn't really expect anyone else to know about that. Well played sir.
It's actually pretty awesome how they did this, although I'm more familiar with the use in stained glass windows. The red tint in stained glass windows from that era was from the surface plasmon of gold nanoparticles. They made it by adding a gold salt to the molten glass solution, and as it cooled the glass became viscous fast enough that diffusion was too slow to form bulk gold. Instead, gold nanoparticles nucleated and sucked in the gold ions in the very local area, but couldn't conglomerate and drop out of solution because the glass was too thick. The history of this stuff is very cool, and I thank you for mentioning it!
I do think it's fair to say that hard drives are the first nanomaterial which we purposefully and knowingly created understanding that it was a nanomaterial.
Hard disks are absolutely, with no qualification, nanotechnology. In fact, hard disks were the *first* nanotechnology we ever used, anywhere. Each bit on a modern hard disk is literally nanometers on a side, the read head is a thin film nanometers in thickness, flying above the disk less than a micron above the surface! Saying that nanotechnology will replace that is like saying that wheat will replace rye as the best sandwich containing substance. Moronic.
When I was helping with a proposal to the EPA for regulating the environmental effects of nanotech, I needed to come up with a definition for nanotechnology. The *only* definition that exists for nanotechnology is a system where the relevant controlled length-scale is less than 100nm. Hard drives are the most advanced nanotechnology on earth!
Slashdot Mirror
I love that piece, they had it set up at burning man this last year too. It really was stunning. I also agree that it's more appropriately called "LED Art" than what we did... the artistic part of what we did wasn't building the light, it was the paintings we made to work with it. Most mixed-media artists don't have the EE expertise to do it on their own, and I sure don't have the painting expertise to do it on my own. The goal of this article is to eventually create a howto so that artists will know exactly what paints to buy to work well with particular LED colors, techniques they can use to make neat effects, with examples. I'd call what we did "LED+Art" instead of LED Art, but I guess that's what happens on translation.
Color Kinetics (now Phillips) is the standard for RGB lighting fixtures. They're just expensive. If you want to buy the lights I made in the article, I'm up for making some custom fixtures that are a bit nicer on the hardware end than the CK ones, but not quite as polished from a software standpoint. Certainly they'd work fine for just cycling through the rainbow in circles endlessly, and I'm setting it up to be designed for permanent installations (a hole in the back to just mount the light on a wall). Send me an email if you're interested - neltnerb@mit.edu, and we can talk about details.
Sorry, I don't know where it went and wasn't maintaining it. I assume it's still around somewhere, but I wasn't involved in setting it up and don't even know what computer it's saved on.
Sorry, this is simply not true. The eye can see any color inside the boundaries of the CIE chromaticity diagram (http://hyperphysics.phy-astr.gsu.edu/hbase/vision/cie.html). RGB colorspace is a subset of human vision. Colors which you patently cannot produce with RGB mixing, but can see in a rainbow, include: that awesome purple that you see at night on a really clear evening, true deep cyans, and true yellows. This can be seen very well on the illustration on the GSU website I linked to.
The eye is more awesome than you give it credit for.
We purchased a wide variety of high-end pigments -- http://www.goldenpaints.com/ was one we looked at a lot. The biggest challenge was finding paints that had only one pigment. I actually asked most of the big paint supply companies for absorption spectra for their paints, but none of them ever replied to my messages. I suspect that their salespeople don't even know what an absorption spectra is...
So the goal was to take the acrylic-style paints you'd find in an art store, and find the ones which were actually fairly close to being monochromates. I toyed briefly with the idea of synthesizing my own pigments (since I'm a material scientist and all), but I decided that it would be too expensive and difficult to reproduce.
The purpose of the first prototype "Color Flower" was actually to act as a poor-man's spectrophotometer. I'd turn on one color only at a time, and see what the intensity of the reflection was -- it wasn't great, but it gave me a decent picture of the absorption spectra at the points that really mattered (and cost $100 instead of however much a real spectrophotometer is).
I'm very confused as to why you got modded -1, but thanks for the explanation of the saturation. I'm still confused as to how a CCD seems to be able to take violet and magically interpret it as magenta -- is there special software that does this somehow? I know that when I use real violet LEDs, they frequently look blue in the CCD; perhaps most of the "purple" flowers we see are actually magenta?
Neat work by Bruce, my personal inspiration was a good friend of mine: http://sub-zero.mit.edu/fbyte/ledart/
I should type up the list we found of good pigments. Yellow and green were by far the hardest to find good monochromatic reflectors in. Almost all of the greens were dichromic, and the yellows just turned out dull and flat.
I put up what I think is a clearer explanation of what is going on. Please let me know if it addresses your confusions.
Sorry, I take it back. You're right as long as your primary absorbing colors are only Red, Cyan, and Yellow. It really boils down to the absorption spectra of the pigment materials. If I have a pigment that is a "true" yellow (reflects in the yellow, absorbs everywhere else -- including red, green, and blue), than it will absolutely appear to be black if you shine red+green LED light on it. You had me confused for a second, but I see where the misunderstanding is now. This is the entire reason we had to spend so much time finding pigments that were more than just mixtures of the primary colors.
I'll clarify the theory section to try to make this more clear.
Nah, djiwhatsit is right, I was wrong. Yellow pigment absorbs blue light strongly, so it should reflect both green and red fairly well and appear yellow. Thanks for pointing it out, I'll correct it and try to come up with a better way to explain it to a relatively non-technical audience. Examples with this sort of thing are fairly hard to come by, and will take some time to fine tune.
The idea I was trying to get across is that dichromates do not appear as you would expect when they are shone on absorbing paints, and that if you mix paints together, you can very frequently just get flat out ugly results. I didn't put up any movies of the paint splotch examples, but in the case of red and yellow paint mixed together to make orange, it *definitely* only appears red when both red and green lights are on =) Experimental evidence! But yes, the theoretical foundation should be clarified some more.
I think a much better example would be green paint. Green paint should absorb strongly both blue and red light. So, if you shine dichromatic red+green=yellow light on that green paint, it will appear to be green only. The effect that many non-technical people would expect is that if "yellow" light is shining on the green paint, it will appear dark -- this is the misconception that I'm trying to correct.
The energy density is probably too low, but there are a variety of devices which use UV LEDs to sterilize bacteria and parasites in water. These UV LEDs were designed for use in curing polymers, but the total light output of two LEDs is still only about the same as a smallish blacklight fluorescent bulb. The neat part is that you can get fading UV light (so your fluorescence fades in and out), but the total power output still isn't even close to that of a fluorescent bulb.
Frostbyte worked at Color Kinetics, by the way... and no, everyone agrees that if I ever wanted to sell the lights, I'd need to pay 5% of the proceeds to Phillips. Unfortunately, no one out there seems to have come up with a silver bullet to kill the PWM patents yet.
Hiya, I like the way you explain the behaviour of the light, and if you give me permission, would like to adapt parts of your explanation to improve the theory section of the page. I'm trying my best to explain this so that someone with about the level of understanding of "light has a frequency" will at least sort of grasp the idea. Someone else mentioned that my understanding of RGB additive synthesis versus subtractive synthesis of colors is wrong, so I'll be rereading the chapter from Feynman's book on color vision again anyway to try to make it clearer. I've thought about it a lot over the last few years, but I fully acknowledge that there are some holes in my model of how the whole thing comes together. In particular, this one bugs me. Why does red + blue look similar to purple? After a lot of training, I can distinguish between the two (now, magenta really looks like red+blue to me). However, it seems bizarre that a CCD can capture the color purple at all. It's certainly not actually emitting purple light, so it must somehow be converting the purple frequencies of light coming from, say, a flower, into magenta. How on earth would this happen? I can grant that a purple light would excite the blue pixels in a CCD, but it's hard to swallow that they'd magically also excite the red pixels in the appropriate proportion. I feel like if I got an explanation of this, everything else would fall into place. -Brian
I've never had the slightest problem before with assembling my boards. I'm usually too busy to spend lots of time on routing traces by hand, so I just carefully place components, let Eagle handle the autorouting, and clean up the worst of its mistakes. When someone starts paying me to design electronics full time, I'll worry more about aesthetic details that don't seriously change the functionality of the devices.
The paintings have fluorescent paint in them. It's pretty cool looking to have bright red light at the same time that your fluorescent paints are glowing -- very wild.
I know that at my university, and it seemed many similar universities, a huge draw for getting the best and brightest individuals around into the doctoral program is that if you invent something, you can actually make money on it. For instance, at MIT, one third of the proceeds of any patents go directly to the inventors, one third goes to fund the lab that they work in, and one third goes to MIT. This makes it massively more attractive for inventors such as myself to come here. It is true that my work is funded in part by the government, but the amount I am being paid is a pittance compared to the value I am creating here. The only reason it makes sense for me to do my research in an open, academic setting is because when I'm done, I can go and commercialize what I invent and make a living out of it.
I know that the same is true for many professors. How could a university reasonably convince an amazing professor and technologist to work there if they force the professor to give up all rights to anything they create to the public domain? No one in academia makes nearly enough money to justify removing their ability to capitalize on inventions they develop.
Beyond the need to attract top-tier students and faculty, universities being able to make money from inventions means that less public money goes to fund the research that goes on there. Instead, companies who benefit from that research are forced, through patent licensing, to fund the universities. It makes a lot of sense -- if a university does amazing mechanical engineering work, it makes sense that companies that want to capitalize on that work subsidize it instead of being able to use it and put the burden of funding that research more heavily on the public as a whole.
Instantly consume the solar electricity instead of increasing the amount of oil you need to instantly consume to then convert back into oil with solar. The result is the same, except that you have more fuel afterwards. (Yes, I know I'm talking about an oil plant. Electricity doesn't care what the source is, and for every grid powered by coal, there's probably an oil plant somewhere contributing that could be removed or have its load decreased)
I feel like I should point out that this does not remove CO2 from the atmosphere.
Fuel made by this, when burned, releases the CO2 back into the atmosphere. I suppose the only reason they talk about doing it on coal plants is because the concentration of CO2 is higher where it's being used. However, as usual there's no way that the energy input is equal to the energy in the fuel, so what is the point? If you put that solar electricity on the grid, you need to burn that much less *actual* oil, so you have no energy loss at all...
Well son, when a mommy law of nature and a daddy law of nature like each other very, very much...
Hiya,
I'm fairly experienced with these things, so I figure I'll offer up what it looks like to me in terms of efficiency.
The 80% number is the ratio of the energy contained in the hydrogen gas to the energy contained in the acetic acid plus the energy used in the form of electricity. However, the stated claim that this is more efficient than ethanol is not really justified based on the actual paper (which I read).
While it may well be better than ethanol (most things are), if we calculate out the actual efficiency of conversion between ethanol and hydrogen typically reported in the literature (Kugai and Deluga are the best papers IMO) we get around -165kJ/mol theoretical energy loss per mol of ethanol converted to hydrogen and carbon dioxide. For comparison, ethanol contains about 1145kJ/mol of energy to start with. That gives an efficiency of around 86%.
Of course, the "hidden" (not very well) cost of ethanol is that it takes massive amounts of energy to produce before you get to the stage of conversion into hydrogen. In the end, you get something between -30% and 30% efficiency converting seeds into ethanol (one well-known paper reported an actual energy loss, ignoring the energy from the sun of course, which will always mean an energy loss in conversion). This is not very good. However, this sort of analysis was *not* done in this paper, and wasn't claimed to have been done in the paper. It is very likely that this process is better than ethanol (most things seem to be), but the summary (and press release) are overstating the case, unless there is information they have that wasn't included in the actual paper.
And to clarify, carbon-neutral does not mean it produces zero CO2. It means that all the CO2 produced at one point came from the atmosphere instead of from fossil fuels. As long as the process is net energy positive, you can use the extra energy to fulfil the energy needs of the process, and remove the need for fossil fuels.
You should email this story to as many people as possible. I think I've heard of some places you can go that'll distribute your message to millions of people for pennies.
That's funny, I didn't really expect anyone else to know about that. Well played sir.
It's actually pretty awesome how they did this, although I'm more familiar with the use in stained glass windows. The red tint in stained glass windows from that era was from the surface plasmon of gold nanoparticles. They made it by adding a gold salt to the molten glass solution, and as it cooled the glass became viscous fast enough that diffusion was too slow to form bulk gold. Instead, gold nanoparticles nucleated and sucked in the gold ions in the very local area, but couldn't conglomerate and drop out of solution because the glass was too thick. The history of this stuff is very cool, and I thank you for mentioning it!
I do think it's fair to say that hard drives are the first nanomaterial which we purposefully and knowingly created understanding that it was a nanomaterial.
Mod parent up...
Hard disks are absolutely, with no qualification, nanotechnology. In fact, hard disks were the *first* nanotechnology we ever used, anywhere. Each bit on a modern hard disk is literally nanometers on a side, the read head is a thin film nanometers in thickness, flying above the disk less than a micron above the surface! Saying that nanotechnology will replace that is like saying that wheat will replace rye as the best sandwich containing substance. Moronic.
When I was helping with a proposal to the EPA for regulating the environmental effects of nanotech, I needed to come up with a definition for nanotechnology. The *only* definition that exists for nanotechnology is a system where the relevant controlled length-scale is less than 100nm. Hard drives are the most advanced nanotechnology on earth!
Some MIT hackers did just that. It's beeping instead of transmitting, but ya know =)
http://hacks.mit.edu/Hacks/by_year/2007/sputnik/
No, that's not distracting. Biting off his toenail while one of the other speakers was talking... that was distracting.