If it's not another Fleischman & Pons, that may be pretty damn important; ganking!
Of course, it's hard to know just what "<$1/W" means, because Watts don't really cost, but Joules do (for convenience, they're often expressed as kilowatt-hours (1 kWh = 3 600 000 J) or other such large units, of course). The total expected energy output of a solar cell is given by efficiency times area times average incident solar power density times average lifetime; the cost of the cell divided by this total energy is the actual energy cost, and I would like to know how it stacks up against what the power companies are charging right now, both at the supply end and at the load (i.e., customer) end.
Solar is very different from coal/hydro/nuclear in that it makes sense to produce it in a distributed, rather than a centralized, way. This means that energy production can be situated closer to loads, improving efficiency by eliminating distribution system losses, but some users (generally heavy industry) will still have to draw large amounts of energy from the grid. The need for distributed generation has been known for some time, and the cost of energy generation has been cited as one of the significant barriers to this (see page 16 of this paper by the Ontario Ministry of Energy from almost exactly 3 years ago), but it's not the only barrier. Hopefully, some of the other barriers will also be taken down; technology can't help us as much there, though.
Also it's natively low-voltage and Direct Current-- a house using this stuff would have a native DC power supply probably in the one to fifteen volt range, DC, as opposed to the current 120 volt AC we use today. DC can be readily converted to AC, but at great efficiency cost. A thrifty solar user would wire for DC and power laptops, printers, and anything else that has those "wall wart" converter boxes with the DC (note: without using the wall-warts!) before starting to meet the thermal guzzling AC needs of, for example, a vacuum cleaner.
The cost and efficiency of home-sized inverters has improved on both points dramatically in recent years. You can now get true sine wave inverters that max out at just under 90% (IIRC) efficiency. Modified sine wave models reach 95%.
That's good news! One would still save energy keeping native DC over converting to sinusoid then back, but despite how attractive the idea is from an aesthetic standpoint converting all those wall warts to a special "DC co-power" might actually prove expensive and confusing for most people.
Markets are unlikely to react until practical applications are shown. Other technologies have been revolutionary: LED produces almost pure light without heat, for instance, so saves on both lighting and cooling. Yet they are too ugly for everyday use. Video conferencing is highly perfected, but people with CEO genes *like* to travel. Until we find out whether the new technology has an achilles heel, I'm not rushing to change my savings portfolio.
Would the fact that they've already presold their first factory's print run for the next four years help at all? Nanosolar right now is turning down contracts until they can build additional factories. The demand is there.
People will pre-order things that look good, but the tide can turn quickly if the actual products don't work as advertised or have flaws that were not advertised. Not saying it is not as good as it looks, just that we don't know yet. "If something looks too good to be true, it usually is." If they really expected those customers to stick, wouldn't they have taken a higher price instead? If I made a machine that could extract gold from seawater at half the market price, would I sell gold at half the market price and take a backlog? No, I would sell it at 5 cent below market price until I had the necessary production capacity. Suspicious.
To get *any* business they have to prove they can stay in business selling below market prices for other production means. If they sold these at 5% below standard solar panel costs, they might not even generate enough business to stay running. But with a large, demonstratable undercut, it's reasonable to attack at a low cost point. Especially if they're still making 60% gross margin as PopSci seems to be indicating...
Basic principles don't look particularly cold-fusion-y-- it's just a PN junction photocell that has a cheaper production method. Old tech, new fab system. Doesn't hurt I saw this a while back when it was still in the "I'll believe that when they get it running" phase a few years ago. I don't have a cell myself yet, but I'm not throwing the BS flag yet either.
It certainly doesn't look like an all-out scam, I just doubt it will have an economic impact for some years yet, and then only gradually. Once the hype dies down, there are very few revolutionary products and concepts made in a human lifetime. Is this one of them? Will coal and oil power finally go the way of the horse-drawn buggy? Perhaps, but not overnight, or even over ten years. If energy prices fall even slightly (due to increased supply), demand will increase. There are just too many fun things to use energy for. Add the fact that the developing world is, for a change, actually developing. It seems rather too early to sell shares in fossil fuel prduction.
My only reservation is the size issue-- because these cells are no more efficient than conventional ones (they don't turn 100% of incident light into electricity by a LONG shot) they won't cut coal out of the electricity picture.
What's really exciting is actually the potential for wider distribution and use: with a technology that more people are using, there'll be greater economic pressure (with or without Google's energy initiatives) to improve their efficiency and further reduce cost. Also a number of spot applications really excite me, such as: how many square cm to power an XO laptop?
I think the cognition problem here lies in the fact that you seem to think Nanosolar is a new, experimental thing and it really isn't. They've been in the testing and production stages for years now. Popular Science doesn't say you're an innovative company just based on press releases - they actually went and looked at the tech.
If I made a machine that could extract gold from seawater at half the market price, would I sell gold at half the market price and take a backlog? No, I would sell it at 5 cent below market price until I had the necessary production capacity.
Except that the entire point of the product is its lower manufacturing cost and market price, which makes that kind of strategy bad.
LEDs are also expensive (at least compared to tungsten filaments), and up until very recently only came in red or (a bit later) green. Advances in semiconductor photonics recently made possible the creation of blue LEDs, which thanks to phosphorous, can be used to make white lights. Of course they're still no good for growing plants as narrow-band as they are. You will however notice we've started seeing LOTS of white-LED lighting solutions popping up since then.
As to this technology I see two limits: first, area (since the panels aren't any more efficient than the old ones, they can't get anything like that mythic "power America with a field in Arizona" energy flux) and second, materials. However I'm already pretty hopeful since the materials don't need to be ion-implanted into a crystaline matrix the way silicon dopants need to be.
Mmm, "recently" is a subjective word of course, but when I started looking at LED lighting about six years ago the hue of the lighting could be described as yellow-orange. Not great for home lighting in general. Then about three years ago I started seeing ads for LED lighting with a "moonlight" color - white with a bluish cast to it. It still wasn't great for home lighting, but it was a remarkable improvement, and I could definitely see using it for accent lights.
If the claims are to be believed (I've only been checking about every six months to see the current state of the art, so I can't verify the current crop personally) we have now achieved real white light LEDs. The problems that remain are butt-ugly designs (really guys, if you're going to compete with the light bulb don't make it uglier than a light bulb), training in LED light use (electricians tend to get shy when it comes to doing things that might get them sued, and contrary to what has been said LEDs for home lighting can get quite hot), and of course general resistance to change.
You'll notice they're replacing flashlights at a good clip these days, but I'm pretty sure that to get a 100W-incandescent-equivalent LED array you'd need, ah, a LOT of them, which would be expensive.
LEDs also suffer from the fact that they're natively DC and low voltage (as diodes, they tend to "clamp" at 0.7v per diode junction) but hey, guess what solar panels are?
Indeed, these are two technologies that could match up very well, although infrastructure issues (e.g. parallel AC and DC wiring) are non-trivial.
Just a note: the voltage can be modified by running them in series. A typical red LED runs at a voltage of 1.4V DC, but if you have a 24V DC supply, you can run 16 of them in series with a relatively small current-control resistor. Similarly, solar panels can be linked in series to get more voltage, which is good for distribution systems. (Resistive power transmission loss drops as voltage increases.)
Cool -- I hope that this heads off the construction of new coal plants.
I still think that to transition into a post-oil energy economy, we're going to have to base the power grid off nuclear reactors. The reasons why are round-the-clock energy availability and energy production density; a solar-based energy system needs to have some massive energy-storage system behind it to deliver power at night or on cloudy days, and you still have to cover over a lot of landscape to provide power at the densities required by factories and conurbations. The likeliest energy system by mid-21st century would feature:
Main power source: nuclear fission (probably breeder), with some nuclear fusion plants coming online (probably deuterium-tritium).
Secondary power source: ground-based solar, with some solar power satellites coming online.
Tertiary power sources: various, ranging from tidal to wind to geothermal; maybe some old oil and coal plants still running.
Mobile power sources: Large vehicles (ships, possibly some trains) use nuclear fission (in the case of trains, minipiles like the Sony). Small vehicles (cars, some airplanes) use zero-emissions fuel cells, rechargable from the power grid. Some high performance vehicles (mostly airplanes, some cars and watercraft) use exhaust-based chemical reaction systems of various kinds.
![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
(no subject)
Date: 2007-12-22 03:36 pm (UTC)Of course, it's hard to know just what "<$1/W" means, because Watts don't really cost, but Joules do (for convenience, they're often expressed as kilowatt-hours (1 kWh = 3 600 000 J) or other such large units, of course). The total expected energy output of a solar cell is given by efficiency times area times average incident solar power density times average lifetime; the cost of the cell divided by this total energy is the actual energy cost, and I would like to know how it stacks up against what the power companies are charging right now, both at the supply end and at the load (i.e., customer) end.
Solar is very different from coal/hydro/nuclear in that it makes sense to produce it in a distributed, rather than a centralized, way. This means that energy production can be situated closer to loads, improving efficiency by eliminating distribution system losses, but some users (generally heavy industry) will still have to draw large amounts of energy from the grid. The need for distributed generation has been known for some time, and the cost of energy generation has been cited as one of the significant barriers to this (see page 16 of this paper by the Ontario Ministry of Energy from almost exactly 3 years ago), but it's not the only barrier. Hopefully, some of the other barriers will also be taken down; technology can't help us as much there, though.
ETA Fleischman & Pons link
(no subject)
Date: 2007-12-22 05:07 pm (UTC)(no subject)
Date: 2007-12-22 07:08 pm (UTC)(no subject)
Date: 2007-12-22 08:01 pm (UTC)(no subject)
Date: 2007-12-22 04:15 pm (UTC)(no subject)
Date: 2007-12-22 04:31 pm (UTC)(no subject)
Date: 2007-12-22 06:01 pm (UTC)(no subject)
Date: 2007-12-22 06:16 pm (UTC)To get *any* business they have to prove they can stay in business selling below market prices for other production means. If they sold these at 5% below standard solar panel costs, they might not even generate enough business to stay running. But with a large, demonstratable undercut, it's reasonable to attack at a low cost point. Especially if they're still making 60% gross margin as PopSci seems to be indicating...
Basic principles don't look particularly cold-fusion-y-- it's just a PN junction photocell that has a cheaper production method. Old tech, new fab system. Doesn't hurt I saw this a while back when it was still in the "I'll believe that when they get it running" phase a few years ago. I don't have a cell myself yet, but I'm not throwing the BS flag yet either.
(no subject)
Date: 2007-12-22 08:34 pm (UTC)(no subject)
Date: 2007-12-22 08:59 pm (UTC)What's really exciting is actually the potential for wider distribution and use: with a technology that more people are using, there'll be greater economic pressure (with or without Google's energy initiatives) to improve their efficiency and further reduce cost. Also a number of spot applications really excite me, such as: how many square cm to power an XO laptop?
Not a Singularity KoolAid drinker, are we? X3
(no subject)
Date: 2007-12-22 07:48 pm (UTC)If I made a machine that could extract gold from seawater at half the market price, would I sell gold at half the market price and take a backlog? No, I would sell it at 5 cent below market price until I had the necessary production capacity.
Except that the entire point of the product is its lower manufacturing cost and market price, which makes that kind of strategy bad.
(no subject)
Date: 2007-12-22 05:02 pm (UTC)As to this technology I see two limits: first, area (since the panels aren't any more efficient than the old ones, they can't get anything like that mythic "power America with a field in Arizona" energy flux) and second, materials. However I'm already pretty hopeful since the materials don't need to be ion-implanted into a crystaline matrix the way silicon dopants need to be.
(no subject)
Date: 2007-12-22 09:34 pm (UTC)If the claims are to be believed (I've only been checking about every six months to see the current state of the art, so I can't verify the current crop personally) we have now achieved real white light LEDs. The problems that remain are butt-ugly designs (really guys, if you're going to compete with the light bulb don't make it uglier than a light bulb), training in LED light use (electricians tend to get shy when it comes to doing things that might get them sued, and contrary to what has been said LEDs for home lighting can get quite hot), and of course general resistance to change.
Lesse, links: There's a PDF here from an on-line magazine [LEDs] ... are important enough to have their own category within the Lighting For Tomorrow (LFT) contest.
Huh, it appears that white LEDs are not necessarily made by mixing red, green and blue LEDs anymore, if Wikipedia is to be believed.
(no subject)
Date: 2007-12-22 10:26 pm (UTC)LEDs also suffer from the fact that they're natively DC and low voltage (as diodes, they tend to "clamp" at 0.7v per diode junction) but hey, guess what solar panels are?
Natively DC and low voltage.
Because they're diodes TOO.
(no subject)
Date: 2007-12-24 06:56 pm (UTC)Just a note: the voltage can be modified by running them in series. A typical red LED runs at a voltage of 1.4V DC, but if you have a 24V DC supply, you can run 16 of them in series with a relatively small current-control resistor. Similarly, solar panels can be linked in series to get more voltage, which is good for distribution systems. (Resistive power transmission loss drops as voltage increases.)
(no subject)
Date: 2007-12-31 02:00 pm (UTC)I still think that to transition into a post-oil energy economy, we're going to have to base the power grid off nuclear reactors. The reasons why are round-the-clock energy availability and energy production density; a solar-based energy system needs to have some massive energy-storage system behind it to deliver power at night or on cloudy days, and you still have to cover over a lot of landscape to provide power at the densities required by factories and conurbations. The likeliest energy system by mid-21st century would feature:
Main power source: nuclear fission (probably breeder), with some nuclear fusion plants coming online (probably deuterium-tritium).
Secondary power source: ground-based solar, with some solar power satellites coming online.
Tertiary power sources: various, ranging from tidal to wind to geothermal; maybe some old oil and coal plants still running.
Mobile power sources: Large vehicles (ships, possibly some trains) use nuclear fission (in the case of trains, minipiles like the Sony). Small vehicles (cars, some airplanes) use zero-emissions fuel cells, rechargable from the power grid. Some high performance vehicles (mostly airplanes, some cars and watercraft) use exhaust-based chemical reaction systems of various kinds.