Sweet Spot

Last night, I had the pleasure of seeing Penn State’s solar decathalon house. These guys are brilliant, not insofar as the solutions they propose are all that novel or even so well executed (I’m hoping their design is fully realized, i.e., done, a few short weeks away when the house gets carted off to the National Mall in Washington DC.) The thing that is great is that the design is so multifaceted, and yet relatively coherent. It’s a web 2.0 type of achievement: smartmob design. Take a bunch of eager students and a few fac/staff from a bunch of disciplines and do a mashup. People clearly worked hard to take their expertise and integrate it through the solution. And it takes advantage of a peculiarity of Penn State—we have departments for everything from turf grass management to engineering mechanics, in the true spirit of land grant universities, covering ‘the practical and liberal arts.’ So all of a sudden you can have nutritionists and restaurant management students designing low carbon footprint meals to be cooked in a kitchen designed by architects, engineers, and even communications and marketing majors trying to figure out how to “sell” the meal to people. Very cool on a number of layers. And perhaps most tellingly, the students take it as a given that the house has to change over time. Systems will be swapped out, reconnected, interconnected, evolved. The mission of the house is to be both demonstration and platform for experimentation, in true mashup style. Amazing.

And again it got me thinking about the wall warts (a.k.a., power bricks, a.k.a. AC to DC transformers). See, this house had two distinct solar PV systems. One was for generating AC for the household appliances. One was for generating DC for a low power LED lighting system. As I understood it, these systems weren’t necessarily interconnected. In the old twentieth century days, most devices were big, people wanted reliable high voltage to power them, and the best way to get high voltage places without too much loss was AC. Washing machines. Stereos. Incandescent lights. Cathode ray tubes. Desktop computers. Laser printers. You couldn’t do much with low voltage except feed it into a transformer and up the voltage (fluorescent tubes, for instance, had to take wall voltage and make it even higher). Low voltage was for malibu lights, and that’s about it.

Nowadays two things are different. First, we have a lot of devices which are less about moving motors and gears, or charging up glass tubes, and more about computing (which is a lower energy activity)–you just don’t need a lot of volts to run a router, or the bulk of a stereo system, or a well designed computer. And secondly, we have a lot of small devices. Devices that are often portable, and therefore need batteries. Cell phones. Ipods. DVD players, and so on. Batteries are DC. The devices want a DC power system, and they have some flexibility in when they get ‘fed’. Even my vacuum cleaner went DC when I bought a roomba. So why is my house designed to provide only AC? And why do my electronics all consume power with transformers that use energy in the AC to DC conversion even when the darn things are off, or nothing is plugged into them?!!

My sister bought me a solio for Christmas. It’s a sort of universal solar charger, and it’s great device for two reasons. First, it can charge a wide variety of devices’ batteries: cell phones, ipods, etc. The thing has a brutally large assortment of plugs, but at least I have only one brick. Second, it uses not-too-expensive solar to charge itself up. It is a power reservoir/battery. It ‘feeds’ on sun when sun is handy. Then it shares the wealth with whatever needs it, either now while it’s in the sun or later. They’ve even thoughtfully provided a wall wart in case the sun is insufficient.

There’s a very important lesson hiding here. The only reason for AC electricity is to transmit it, over long distances. Which you need to do if you need lots of power, coming from a centralized source. If we want to think about decentralized power (and so many things these days really benefit from decentralization), then we benefit from going back to DC. And there’s no reason you can’t have two systems for the two sets of devices.

I have a hunch that not only could direct-to-dc solar be a lot cheaper than solar that goes through an inverter to provide AC, but that many more homeowners would be willing to install a solar PV system if it was designed not to provide all energy needs for all the spiky-demand high energy devices (old school AC devices like the electric baseboard heater or the blender), but instead to provide only for things that are ‘natively’ DC. Imagine, perhaps, twelve volt automobile-style cigarette lighter sockets or USB plugs as wall sockets in the place where you charge your twentyfirst century devices. No need for AC and the inherent losses in the wall warts. Rather you have some kind of smart charger that treats everything plugged into the wall as a battery. True, it’s a battery connected to a device, but the battery buffers whatever the special needs of the device itself are. Mostly we just care about the battery. We make the batteries and the chargers aware of each other through some protocol–maybe a modified USB, maybe something else. “Hi, I’m a Li-ion cell with 47.5 watt-hour capacity at 53% and a penchant for trickle charging after 90%, who are you?” “I’m a nominal twelve volt socket coming from solar, but it’s 7:30pm, and I’m not gonna have much power for you in about 45 minutes, so get what you can now. We can get the rest out of my reserves after the sun goes down and trickle charge you all night.” “Well, my owner is in a rush, can we use reserve plus incoming now? I can take up to 24V.” “I’ll see what I can do.”

This way of managing power allows a lot of things to happen. First, and perhaps most importantly, it allows people to start taking advantage of solar without needing a way to ensure a solid 120V AC at all times (or buy expensive DC-friendly washing machines, refrigerators, etc.) Let the grid handle what the grid is good at, and let solar + batteries do what batteries are good at. Second, it circumvents losses in the wall warts. And third, if we put the smarts in both the batteries (devices) and the wall, we can imagine upgrades as power management techniques get more sophisticated. My outlets could get smarter over time with software patches. I would be like the solar-d house, a mashup of devices that can evolve with the times to fit better and better with needs and constraints. And ultimately the power system would be tuned to the needs of charging and discharging, not providing some fixed, dumb voltage.

Right now my laptop battery does something somewhat like this. But the smarts are all in the battery and powerbrick, on the load end. My wall socket is stupid (120V, all the time!), and the grid has to be very smart to allow that wall socket to be so dumb (regulating the voltage in the face of fluctuating supply and demand). That regulation is expensive and needless—my roomba can charge during daylight or nighttime equally well, saving some juice for an air conditioner somewhere if drawing from the grid, or taking advantage of a sunny day if on solar. I wonder whether a more standardized way of managing this would allow the batteries + powercord to be a little stupider and the socket a little more sophisticated.

The US Department of Energy says “In the average home, 75% of the electricity used to power home electronics is consumed while the products are turned off.” How much of this phantom load goes away if we use a smart DC-to-DC charger system? The substantial savings, combined with the different profile of these 21st century devices, might make all the difference for solar, even if you live in a place where feeding excess decentralized power back into the grid isn’t allowed.