Google Contest Builds More Efficient Inverters

A few summers ago, Google and IEEE announced a one million dollar prize to build the most efficient and compact DC to AC inverter. It was called the Little Box Challenge, with the goal of a 2kW inverter with a power density greater than 50 Watts per cubic inch.

To put this goal into perspective, the DC inverter that would plug into a cigarette lighter in your car has a power density of about 1 or 2 Watts per cubic inch. Very expensive inverters meant for solar installations have a power density of about 5 Watts per cubic inch. This competition aimed to build an inverter with ten times the power density of what is available today.

Now, the results are in, and the results are extremely surprising. The best entry didn’t just meet the goal of 50 W/in³, it blew the goal out of the water.

The winning entry (PDF) comes from CE+T Power, and comes in a package with a volume of 13.77 in³. That’s a power density of 143 W/in³ for a unit you can hold in the palm of your hand. The biggest innovations come from the use of GaN transistors and an incredible thermal management solution.

Other finalists for this competition include Schneider Electric Team from France that managed a 100 W/in³ and a Virginia Tech team that managed a power density of 61.2 W/in³.

Thanks [wvdv2002] for the tip.

109 thoughts on “Google Contest Builds More Efficient Inverters

    1. As I understood they choose the one with the highest power density which fulfils all the requirements. So I assume that all the other solutions, which were not mentioned in the final ranking, don’t fulfil one or more requirements. This is not very astonishing since those are prototypes with all kind of teething problems.

  1. I really do not understand the quest for power density in a solar farm where the panels are which delimit the size of the system for a much greater margin than current inverters.

    As far as I know, power density and energy efficiency are not related, more over, when power density raises, it is necessary to use active cooling, it will have some (marginal) cost in energy efficiency.

    1. “Active cooling” in this case is a very small DC fan; And also the proble they are trying to solve is not the one of Solar farms, it is the one of home (garage) Solar-to-grid converters.
      Rembember this is Google, and Google does it because Google gets it!.

      1. Close.. But no.
        Sure home installs are on the rise and will hopefully eventually saturate. If this product helps that effort, I am sure Google will market it that way. Not saying they are heartless corporate profit mongers.
        But they are. As is every company to some extent.
        Their real target always has been and always will be their own server farms. The current generation packs hundreds of server blades into shipping containers. Hook up backbone fiber. Hook up (solar) power. Turn on. Walk away.
        Their interest is to make not simply access to the Internet, but the internet ITSELF, truely mobile.

        1. But why do they want to use an inverter for powering computers rather than arranging solar battery systems at the right voltage to directly power the DC rails used by their comptuers internally?

    2. Not certain of this, but I think that one of the requirements to high power density is high efficiency. Lower efficiency means more volume consumed by heat sinks and cooling that contribute nothing to actual electrical output power. It is sort of an indirect way to drive that goal though.

          1. Can be?

            They have an estimated run time before failure. This means they will fail. Anytime you EVER see a project with a motor in it, be it a pump, fan, etc., the first thing that should cross your mind is, what happens when the rotor gets jammed. Not what happens IF, but what happens WHEN it gets jammed. Everything better shut down nicely and not burst into flames.

            Power density usually matters for servers where space is extremely constrained.

    1. Not this time, read the docs quality of materials and fabrication are critical to it functioning so well. Hmmm then again perhaps you are right, some of the knock-off makers only care if it looks about right and lasts 6 months, after that bad luck you got what you paid for.

      But could they source the GaN transistors?

      Anyway the real thing looks like it will pair very well with this, http://cleantechnica.com/2016/02/26/new-energy-storage-solution-could-hit-magic-54-mark/

        1. Do they? Do you really know what the supply chain from raw silicon to finished product looks like? Even if the chips were right next door why would counterfeiters be able to get hold of them? What are your assumptions?

    2. We will not see cheap copies of this for at least 10 years, and here is the reason: There is no need for it! You can get an off-the-shelf solar converter from China for 1/10 of the price and 10x the volume that works the same way.

      1. I disagree.
        Somewhere up there is my comment about Google’s server farm in a shipping container plans.
        Google will be the first customer. They have in the past been willing to invest millions into totally bespoke server hardware in order to fit their physical constraints.
        Most of their “server farm in a box” hardware runs DC for comparability to the most power input options. But they are always looking to shave a few degrees of temperature, a few amps of power, and cubic centimeters of space. On their scales, that all adds up very quickly. The investment cost in this research is pocket change compared to potential savings.

        1. Google spends $1M on research and then buys the cheapest knock-offs from China? If Google is the only demand, and they demand quality, it will be a while before cheap trash has a market.

  2. Hey i work for Schneider Electric. Great to see they made it as finalist. Now i wonder if i know anyone that was on the project.

    Cant wait to see china’s super compact 2kw solar inverters :) i know my company builds things in china.

  3. I’m missing something, maybe someone can fill in the aspect I’m not considering here:

    Let’s consider the Brusa DMC524.
    http://www.brusa.biz/_files/drive/05_Sales/Datasheets/BRUSA_DB_EN_DMC5.pdf

    It can put out a continuous 79 kW and has a volume of 7.6 liters not including connector bulk.
    In other words, 79000 watts at 464 cu.in. That’s a power density of 170 watts per in^3.

    Their DMC544 has a density of 213 W/in^3. So why is 50 W/in^3 such a big deal? Am I doing my maths wrong?

      1. Because an EV motor driver is not an inverter. It simply switches the battery voltage to the right poles of the motor in the right sequence. (of course there’s some more advanced motor control stuff in there but that’s the gist of it) There is no step up/down of voltage. So no voltage regulation and no frequency/sinusoidal waveform regulation.

        So its much easier to cram high power motor drive electronics into a box than the equivalent power inverter, because it has to do much less.

        1. EV motor drive like the Brusa one is indeed an inverter; it creates 3 phase sinusoidal AC (since it can run any type of 3-phase motor, it must be sinusoidal) from DC. Those should not be confused to BLDC motors like hobby RC motors which are completely different design from industrial or EV motors. However, as compared to solar inverters, or others doing AC for common use, the power quality (ripple current etc) requirements are much lower, and thus switching frequency etc can be made lower.

          I bet the other requirements than power/size ratio did rule out typical motor inverters from the little box challenge; there most likely is a need for output filters etc that add to size and losses. In that sence you are right that power inverters do more, but it’s not that much more: both create sinusoidal voltage and/or current in the end.

          1. The other trick with a motor drive is that you can use the (leakage) inductance of the motor windings themselves as the filter inductor, i.e. your output filter is free, consumes no volume, and does not generate any heat in the inverter box. Whereas for grid-tied inverters, you need to put in your own filters (usually LCL).

          2. Eh, I’ve played with a few industrial 3 phase motor drivers. They’re *sort of* sinusoidal and many of them still have plenty of square-ness on the output. It’s possible I’m looking at a different sub-market and we’re still talking about two different things. But with the drivers I’ve played with, you definitely can’t assume the output is a pure sine wave, and they often are not. In fact, you’re often supposed to couple them with “inverter duty” three phase motors because the HV spikes from the drive supposedly require thicker insulation to prevent arcing inside the motors (or so the motor manufacturers say)

            Anywho, you’ll notice that an enormous design and volumetric component of the finalist/winning entries is (active) filter design – tells you how much of the volume is required in a typical inverter to meet the EMC/harmonics requirements for this competition.

      1. Ah I missed the “up to” of the 2 kVA part. And the “no more than 40 cu. in.” part apparently.

        I guess I just don’t understand Google’s premise here: why is “cooler-size” too big for a home?

        And I don’t know much about nuclear reactors, but I don’t think they run solely on 450 VDC and stay under 60 C. You know, because of the whole “nuclear” bit.

        1. I think the idea is to couple the inverters with the panels and have the units output AC more directly, so the cooler-size boxes would end up on the roof of someone’s house, which would be unsightly… more unsightly than a bunch of solar panels, I guess, anyway.

  4. I never understood why the goal of this was “small” rather than highest efficiency and/or lowest cost – why do you need a solar inverter to be small?

    1. Yes, that’s exactly the same question I had. The whole concept is questionable. I would have expected a noble aim behind such a competition, like a “one dollar inverter” for third world countries. But I can’t find any other explanation than promotion for wide band gap MOSFETs.

    2. I don’t think this is for stationary application. The recent article about 100kW radio transmitter might be related. Perhaps it will be for airborne or even orbital solar power plant beaming the energy down. This is a speculation, but plausible in my opinion.

    3. Cheap, and/or efficiency are already driving the power supply market. The contest is to encourage new design ideas and not asking for more of the same stuff. Same as most of the contests by DARPA, NASA or even HaD.

      If google wanted a power supply for an application, they could have solicited the usual suppliers and invite them to bid.

    4. What does google make, other than ads? Phones and (soon) cars, and god-knows-what in Google-X. They both have a need for very high power-density power conversion technology, and power-density is an excellent proxy for efficiency because much of the “wasted” space is usually thermal management.

      If someone can build an epic mains inverter, the same tech can be applied to other power-conversion modes, like powering a CPU or cranking a traction motor.

    5. I’ve scanned some of the PDFs that document “how it’s made”.

      One of the requirements was “close to zero input ripple”. Some very innovative solutions were invented. Those can be used separately.

      One group made a controllable voltage source/sink. By adjusting the requested voltage “just right” they get away with a smaller input capacitor while drawing close to constant input current.

      Another group managed to use an output filter inductor to continue to draw power from the input when the output power was near the zero-crossing.

        1. Yeah but… so can a diode!

          For data centres, everything runs at 12V, 5V, and 3.3V. If you’re gonna have solar panels, like in this idea, or batteries that are charged from anything, it makes sense to use a battery voltage close to the voltages you’re after using. DC/DC convertors, and inverters, are both pretty efficient, but when you’re using 1.21GW, even small inefficiency percentages add up to a fair few watts. That means heat, which is another gigantic pain in the arse, and if they use air cooling in an air-conned building, it means some ludicrous amount of watts for the A/C to remove the watts of unwanted heat.

          Actually it bothers me that companies have servers venting air into air-conned buildings. Surely some direct method, even if it’s just piping cool air in from outside, has to be better than cooling the whole room down. Better yet would be liquid cooling, deionised water should be reliable enough. I wonder why they don’t do it?

          Of course you could go to ludicrous extremes and dump them in a lake. That’s when your company has too much money per management brain cell. For either reason.

          1. Watch some of the videos about state of the art data centers such as google’s. They are located next tor rivers to use that water for cooling into the server racks. They also use high voltage distribution to the racks to cut down on cable losses (and thus extra heat).

        2. Actually… further to your point, the switchmode PSUs that modern computers run on aren’t all that different from switchmode inverters anyway. So they sort of DO run these backwards.

        1. Something like 310V to the racks, then the racked equipment steps that down to whatever it needs. I don’t think there’s a standard yet, IIRC from the article I skimmed long time ago.

  5. The use of “Spread spectrum” methods as a route to EMC compliance is misguided in the extreme.

    That technique does not reduce the level of EMI emissions, only moves them arround in the frequency domain to “Fool” the measuring instruments* into showing a reduced level.

    * When the QP (Quasi Peak) detector is used. That was only ever intended to be used for impulse type emissions, such as those from IC engines that use spark ignition systems.

      1. I would think the spirit of the law was to reduce power emitted and the regulatory tests designed for the equipment available. Spreading the power in frequency may not mitigate the likelihood of causing another device problems.

        1. In theory it does, since if you’re tuned into X frequency, a spread spectrum device is only interfering with that frequency for a fraction of a second, then off to interfere with some other one.

          In practice though, I wonder. I think this idea was invented back when people used one frequency at a time. Now, wifi and mobile phones use spread-spectrum communication. They could easily just ignore a channel with too much noise on it. But spread-spectrum noise can take out a whole set of channels, particularly if they’re low power, like Bluetooth, or Bluetooth Low Energy in particular.

          The result is either your comms don’t work, you lose data rate, or you have to transmit louder and your phone eats up it’s battery quicker.

          Is this actually a problem in practice?

    1. DaveB?…hmm, you’re not the legendary synthesizer guru David Bing, are you? if so, I though you were dead!

      Spread spectrum methods for staying under the EMC limits is sort of like spreading dog crap. If you have a kennel and 500 lbs of dog crap to get rid of, it isn’t going to get unnoticed if you simply dump it in your neighbor’s yard. However, if you put 1 lb on each of the nearest 500 lawns, it might not “raise a stink.”

      But…if ALL 500 neighbors have 500 lbs of crap to spread (i.e. there is a high density of identical EMC polluters) then each lawn still gets 500 lbs of dog stuff.

    2. You are arguing against a method that gets past a test.

      The test is misguided, not the method that gets past it.

      FCC begins testing at 9kHz and above, you sure about that ICE spark ?

  6. Excuse my ignorance, but how is power per volume a measure of efficiency? I thought efficiency was output power divided by input power, and is a unitless percentage.

  7. Where did the author get “efficiency” as a design goal.

    “”An open competition to build a (much) smaller power inverter, with a $1,000,000 prize””

  8. The entire premise of this contest has me scratching my head. Google probably has ulterior motives like for their space internet project, sure, but why is the IEEE on-board?

    -From their website, in regards to why solar adoption isn’t greater:
    “The problem is household inverters are too big…”

    I thought ‘the problem’ was financial. When they actually have a ROI, it’s frequently measured in decades. It takes a large up-front investment if you want to see a return.

    I understand and would love to see a sustainable power grid, but this is NOT the way and we shouldn’t be putting resources toward it.

    1. “I thought ‘the problem’ was financial. When they actually have a ROI, it’s frequently measured in decades.”

      My solar panels on my roof + the inverter combined pay for themselves in 7-8 years at current electrical prices, assuming the power company buys from me at the same rate as selling. (which they do). The invertor is warrentied for 10 years, the panels 25 years (at within 80% power generation of first year).

      Cant say the physical size of the inverter was ever a priority though.

      1. It sounds like you’re happy so please don’t read the rest.

        Best case, financially you’re barely breaking even and your contributing to higher energy prices for everyone through the additional burden you’re putting on the grid. A better choice would have been a low risk stock market investment with some of the proceeds being donated to your favorite ‘green’ charity.

        After 26 years you’ll have likely replaced your inverter twice and be ready to replace your solar panels. After you compensate for inflation, every dollar invested has AT MOST made a few dimes. Compare that to an ultra LOW risk stock investment which would have returned at least three dollars per dollar invested over those same ~30 years.

        Few consider their lost opportunity cost, lack of compound growth or inflation. They’ll ignore time lost, labor costs, replacement costs, make apologetic comments and skew numbers just to get a hypothetical return.

        I’m unaware of any circumstance where a small home solar array mounted to the roof had a real ROI outside of unsustainable factors like crazy high costs of electricity and government assistance. Dave Jones’s setup is a perfect example of this. Even these border cases still have zero compound growth …which translates to a massive lost opportunity cost. i.e. He can’t make more each year unless he adds more solar panels. Whereas he can reinvest stock market earnings to make more each year.

        1. Unfortunately you are probably correct, given the real-world mean time to failure of currently purchasable technology, but if we examine the cost performance improvement curves that will not be the case within a decade, which is why I have a lot of cash in the bank but have not committed to installing anything yet. At the moment shares are a bad idea, wait 3 to 4 years.

        2. Depends if your goal is to just make money…

          Solar won’t make you cash like the stock market could. After inflation, as you point out, maybe it won’t make you any. But will putting your cash into the stock market reduce consumption of non-renewable energy? Exceedingly unlikely unless you are *exceedingly* picky.

          Even though Gordon Gekko might not be in the deck chair when the Titanic goes down, if he or his relatives have kids, they will be.

        3. Well, over here in Europe electricity is about 30ct per kWh so that makes for a payback period of about seven years without other subsidies.
          Rough math puts that at a ten percent roi which is very hard to find in low risk stocks.

          1. Google “compound growth”.
            Highly unlikely optimistic solar scenario: Your $1k solar system yields $143 profit (14.3%) each year and requires zero maintenance cost. ROI year seven. From year eight onward a fixed $143 annual profit eternally.

            Highly unlikely pessimistic stock scenario: $1k investment, 4% return and all profits reinvested during those seven years. End of year seven = $1316 in stocks and >$50 annual profit.

            WITHOUT reinvesting further stock profits, the solar system would catch up around year 22 and pull ahead.
            If stock profits were reinvested, the solar system would take ~30 years to catch up and even then only for a brief moment until the stock’s compound growth inevitably comes out ahead.

            Keep in mind:
            – No solar system sees 14% return. A ROI less than ten years involves fudging numbers and ignoring inflation.
            – No solar system has zero maintenance cost.
            – Your time is worth something.
            – The stock market has never seen an index return 4% over any ten year span, typically 5% ‘real return’ (INCLUDES inflation) is a worst case scenario. Even 6% is typical through the recent recession.
            – Compound growth will always win.

            A more realistic, still optimistic, example is
            – 5% solar return after accounting for maintenance costs
            – 6% stock return with compound growth.

          2. “– No solar system has zero maintenance cost.”

            When devices are warrantied *that means you don’t pay for maintenance*
            That’s factored into the cost/payback times already.

            I would also point out in all your “maths” you seem to be assuming that the cost of non-solar energy will stay the same. 30%~ of my energy costs are now paid for at least the next 25 years. Upfront, Ive paid for it.

            Meanwhile on your stock market theory, your still paying, and at the wimms of energy supply and demand, investments in nuclear and the supply of fossil fuels etc.
            Your millage may vary – but to assume that energy costs will stay the same seems a huge leap to me. Especially as many cases are currently projected for shortfalls.

        4. “barely breaking even”

          = 7-8 years to pay for itself followed by 18-19 years of 30% lower energy bills.

          Could I pick some good stocks and get better returns?
          Maybe.
          But its still utter bollocks the idea that solar “barely pays for itself”

    1. To have a really high power density you got to take care of getting rid of the heat and/or not having too much heat in the first place. Efficiency comes for free as part of solving that problem as there is only so much surface area or cool air that the box can draw in.

  9. Hah, and then in the comments Slobodan Ćuk, who the Ćuk converter is named after, is taking them to task for not publishing efficiency numbers for the converters…

  10. While all this is interesting, we would (as a species) be better spending our time getting away from AC and converting our homes to DC with a focus on home storage of energy. Almost(almost) everything in your house just converts the AC back to DC for use anyway. Why not focus on eliminating the middle man? Leave the electric companies to generate AC for use by big industry, where it is actually needed.

    Just looking around me right now, I’m typing this on a laptop(DC), My TV is a newer flat screen (internally converts to DC). I’ve converted to all LED light bulbs (internally converting to DC). My fridge and major appliances are all that are running on AC, and they COULD be converted with a bit of effort.

    1. The appliances are the last problem. Most motors are really inherently AC rather than DC devices (and then 3 phase rather than single phase devices). Remember, all DC motors basically AC motors with an mechanical switching converter (commutator), which usually throws EMI left and right and has all sorts of reliability issues. “One-phase” motors are (in most designs) just three phase motors with a capacitor to start them (and of course that capacitor and switch mechanism itself have all sorts of reliability / EMI issues as well). In fact, in many homes you’ll find that the peak total power consumption is going to be from high power appliances, most of which are motor driven, rather than DC appliances.

      Also puts some perpsective into another reason why AC won over DC distribution, especially a century ago.

        1. As I said, we need to get away from the communal idea and move towards localized generation. As stated, the bigger AC driven appliances could be converted with some effort. Most homes aren’t running in the MW-GW scale.

          I’m not a Tesla vs. Edison guy. AC has its uses (big industry). DC has its own (electronic based items and small power requirements)
          I just see a simpler way to do things. Of course simpler is always a relative term.

      1. You could make multistranded cables that include a DC voltage(s), and you just wire each receptacle for what you want. Then you just need a DC standard. 5V USB receptacles might be neat, though you’d need some big gauge wire to handle all the current draw if you use it for a lot (i.e. if you wire your PC straight to the DC voltage(s) with just a little regulator / filter).

        The advantage would be you could upgrade a single AC/DC (or DC/DC in the case of solar) converter in your garage and optimize your efficiency rather than having 100 crappy AC/DC converters all around the house (some of which drain power even when the device isn’t in use, as far as I understand). More devices would need a DC input standard, though.

        1. I might have misread what you wrote…but…

          No, you can’t have a 5V bus mixed in with 120VAC or high voltage DC wiring, the wiring cannot even be co-located in the same housing: The wiring must be physically separated, otherwise you violate the NEC:

          http://www.lutron.com/TechnicalDocumentLibrary/213%20Combining%20low%20and%20high%20voltages%20in%20box.pdf

          A 5V USB line can’t tolerate much voltage drop. It would require a pretty significant gauge wire to be able to support a few amps, and be up to 75ft long and stay within the voltage drop spec of USB. I think the change to USB 3.1 will make this more realistic, where the device tells the USB Hub that it can handle a higher voltage so the hub raises the voltage on that port to accommodate it.

      2. Going from 120V AC to 170V DC requires a bridge rectifier and a cap. You can easily get 99% efficiency. Going the other direction from DC to AC is not as simple. With DC, now you have to worry about the polarity too. By the time you put in diodes for reverse protections, might as well stick with AC. Basically it is change for the sake of change without any real benefits.

        You problem is not AC vs DC but rather 120V/220V vs much lower voltages.

          1. Mike Lu: 98% is only accounting for the diode voltage drop it does not account for any of the other effects of the current and voltage curves not lining up, while the apparent efficiency may be equal to (Line Voltage-Diode Drop)/Line Voltage the standard bridge rectifier causes Power Factor issues. That cause the real power to be much larger and while residential customers are charged only for apparent power commercial customers pay for real power or are changed extra if their PF is off by more then 0.05. the most common way to passively fix power factor is put a capacitor across the line, this cleans it up but ends up causing any all the Reactive power to be converted to heat lowering the efficiency.

            Once again he max efficiency of any Full Wave Rectifier with 100% ideal diodes with 0 voltage drop is 8/(pi^2). When you add standard rectifier diodes it becomes (Line Voltage-Diode Drop)/Line Voltage * 8/(pi^2). That extra 18.8% is lost in a capacitor, motor windings or parasitic capacitance in any long wires.

      3. I disagree,

        Interrupting a high current DC fault is probably the last serious problem. What do we use? Recloser switch gear in a home? Contactors aren’t rated for this action.

        AC motors in appliances are getting replaced by BLDC motors because the BLDC motors are more efficient, therefore smaller, therefore cost less.

        A microwave is another story…

        Dishwashers and washing machines are already switching over, or have switched over to 3 phase BLDC motors either with or without hall effect sensors. ST makes pretty much owns this market. If the input was DC, you would just hook it up.

        Energy storage is indeed a difficult problem to solve, but the first step is energy efficiency upgrades — they help immediately, then develop a power budget to size the energy generation and storage system.

        Maybe someday we’ll all have laminar flow batteries in our basements, or molten salt batteries, or extremely insulated superheated water tanks that collect solar thermal energy for heating purposes. There are certainly a lot of interesting areas to refine, but none of them can be usable if they aren’t made safe.

  11. From my understanding AC was the winner in the AC/DC war because it was exceptionally easy to convert voltages. A few minor other advantages (3 phase motors), but conversion was the primary reason.

    It’s much easier to convert DC to DC these days than it once was, but *all* conversions require AC in the middle… unless you just want to go resistor… but then how do you convert up without AC? How would I get my 240vDC (at 20A) oven to operate from my 24vDC battery pack? I could run 24v at 200 amps, but that is exceptionally lossy over more than a couple of meters, besides the wire being damn thick and expensive to boot!

    AC is still the *best* distribution method for anything more than a few tens of meters. Conversion is cheap, easy and above all, reliable. No DC-DC converter can compete with the reliability of a transformer (most DC-DC include transformers/inductors in them anyway).

    And yes I know that there are stupidly high voltage DC distribution systems around (eg: New Zealand), but they are used in very specific cases with very specific requirements.

    1. Conversion of voltages is easy in AC, true. But the main reason for AC use was transmission losses. With DC system you can generate at 120 V for example, but to supply 120 kW of power to a small neighboorhood you need wired capable of carrying 1000 A (huge $$$$). With AC, you just step up the voltage to 120000 V with transformer, and your wire carrries only 1 A (cheap wire, small $). At the point of use, another transfomer drops it down to 120 V. So your costs are basically only transformers. It is quite obvious that for any longer distance, AC wins financially. If your generating plant is tens or hundreds of kms away, AC system is probably 10 or 100 times cheaper than the old style DC installation.

      1. Exactly. The driving factor behind *nearly* everything that humans do is cost, and that is the reason AC was adopted over DC – it is easy and *cheap* to transform AC to different voltages. The losses are identical… I2R… regardless of AC or DC, it simply becomes a matter of reducing the I.

        Back when electricity was a new thing most loads were either motors or resistors (incandescent light bulbs, stoves, heaters), and an AC motor is easier to make (that cost thing again). A resistor don’t care. Add in ability to transform voltage and our equation becomes : +3 for AC, +1 for DC. Looks like AC wins.

        @evad : yes, I could put 240 volts worth of batteries in series, but then how could I pull 12v (or 5v or whatever) out to power my laptop or USB phone without either un-balancing the cells or without a DC-DC converter? And if I need a DC-DC, why wouldn’t I go AC from the outset?

  12. Anybody know what the licence the projects are required to release under? It would be FANTASTIC if everything was open source or MIT licensed so that all this research would now be freely available to use by everyone. If google owns the rights then they may choose to licence it but then we will pay a premium for “new” technology for a decade.

  13. I feel that Google is leading us into a path of throw away insanity with this contest. Many inverters built today can be more easily repaired, sometimes by just replacing the power transistors and blown fuses. Future inverters will be as cheap and disposable as modern cell phones or laptop computers. How useful will these be in a zombie apocalypse? (just kidding about that) No prepper in their right mind will like this. I work in 2-way radio repair and these radios have become almost impossible to repair any more. Most need to be repaired at the factory, if the manufacturer will support them and the customer will pay the huge flat-rate fees. I am glad I am retiring in a decade since I probably won’t be able to repair any of these products much longer

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.