What Voltage for the All-DC House?

The war of the currents was fairly decisively won by AC. After all, whether you’ve got 110 V or 230 V coming out of your wall sockets, 50 Hz or 60 Hz, the whole world agrees that the frequency of oscillation should be strictly greater than zero. Technically, AC won out because of three intertwined facts. It was more economical to have a few big power plants rather than hundreds of thousands of tiny ones. This meant that power had to be transmitted over relatively long distances, which calls for higher voltages. And at the time, the AC transformer was the only way viable to step up and down voltages.

No, not that AC/DC

But that was then. We’re right now on the cusp of a power-generation revolution, at least if you believe the solar energy aficionados. And this means two things: local power that’s originally generated as DC. And that completely undoes two of the three factors in AC’s favor. (And efficient DC-DC converters kill the transformer.) No, we don’t think that there’s going to be a switch overnight, but we wouldn’t be surprised if it became more and more common to have two home electrical systems — one remote high-voltage AC provided by the utilities, and one locally generated low-voltage DC.

Why? Because most devices these days use low-voltage DC, with the notable exception of some big appliances. Batteries store DC. If more and more homes have some local DC generation capability, it stops making sense to convert the local DC to AC just to plug in a wall wart and convert it back to DC again. Hackaday’s [Jenny List] sidestepped a lot of this setup and went straight for the punchline in her article “Where’s my low-voltage DC wall socket?” and proposed a few solutions for the physical interconnects. But we’d like to back it up for a minute. When the low-voltage DC revolution comes, what voltage is it going to be?

Resistive Heating

The problem with lower-voltage wiring is simple physics. For a given power demand, P=I*V, so a lower voltage means pushing more current. But substituting in Ohm’s Law, more current also means dramatically higher resistive losses P=I^2*R in the wires. Reducing the resistance of the wire by using more copper is one alternative, but you get more bang for the buck by focusing on the current-squared term.

electric-arcThat’s the reason that, for instance, power over Ethernet (PoE) schemes use around 48 V to transmit something like 30 W of power — those thin Ethernet cables can only carry so much current without wasting most of it away as heat. Even around 50 V, PoE schemes count on a loss of three to five watts in the cabling. So whatever cabling is used for the low-voltage DC segments of your home’s electrical system, it’s going to be thicker than Cat-5.

But copper costs money, so there’s always going to be some upward pressure on the voltage exerted by resistive heating effects.


00077Electricity starts getting dangerous to humans somewhere around 30-50 V. That’s where the current levels that push through the human body resistance start to become troublesome. But while everyone says “safety first!” it’s also worth noting that you’ve got 110 or 230 V AC in your walls right now. Clearly it’s “washing machine first” in the real world. Which is to say that although sub-30 V DC would be safer, we suspect that the safety will be designed into the connectors, or into circuit breakers.

Switches and Relays

Which brings us to the last concern. Have you ever arc welded? How much DC voltage does it take to strike up an arc? Something in the neighborhood of 24 V is a pretty common value for a professional unit, but people have been able to weld with 20 V tool battery packs or even 12 V car batteries. Some spot welder designs that we’ve seen only use two or three volts, but they develop the required current by pressing the workpieces together very hard to make the low-resistance path.

dscf9252Have you ever looked on a relay and noticed that it has ratings for DC and AC use? For example, these relays are rated for 10 A at 250 V AC, but only 10 A at 30 V DC. Where does this factor of ten come from? Relay contacts can spark as the two contacts get close together, and they are prone to weld themselves together at higher DC voltages in a way that’s just not the case for AC, because AC arcs self-extinguish 100 or 120 times per second.

Making mechanical switches that work for your DC home electric system is going to be a problem, then, and that’s going to put downward pressure on the voltage. The average car contains a lot of relays, and they seem to function most of the time, so maybe 12 V is a good pick here. Don’t take my word for it, though. Here is an automotive engineer’s take on DC in the home. It’s a little bit dated, but he complains about the extra design concerns in working on a 24 V diesel vehicle. We take that as a vote for lower voltages.

The X factor here is progress in MOSFET or IGBT manufacture. Solid state DC circuit breakers aren’t as cheap as mechanical (AC) breakers yet, but at voltages like we’re considering inside the home, they’re getting there. The higher price may also simply reflect present lower demand. Maybe the downward voltage pressure will evaporate in the near future?

W.A.G. Time

Now we’ve come to the end of the article, so let’s see if we can make any sense of all this. If solar is going to play a role in our future energy demands, it’s inefficient to round-trip from DC to AC and back again. It’ll be more efficient to stay DC from the panel to the battery to the end device, maybe only changing voltages once or twice with highly efficient DC-DC converters along the way.

sepeda_listrikIf there were to be a complementary DC standard, heating losses push the voltage level up, switching constraints push the voltage down, and safety is, we think, a wash. Solar panels are essentially costlessly configurable for either high voltage or high current, and we think it’s telling that newer installations tend to run in the 24-50 V range. That says a lot about the importance of heating losses. Batteries are similarly flexible, so there’s not much problem matching them to the source.

We’d be stoked to have DC sockets, and devices that plug into them, all powered by a medium-sized panel on our roof and stored in medium-sized batteries in our basement. Whether this goes from the panel to the battery to the plug at 48 V or 12 V is going to depend on the relative prices of copper and hefty FETs, but we’re betting that FETs get cheaper and copper more expensive. We’d personally like to see this relatively high voltage stepped down at the plug for safety, say to 12 V, but we won’t quibble. It would make the perfect complement to our existing AC infrastructure.

What do you say? What factors are we missing? Any of you have a DC side of your home already? What voltage(s) for DC?

212 thoughts on “What Voltage for the All-DC House?

    1. 42+-20% is about right. Keeps under the 50V “High Voltage” PPE limits for UL and other stuff, and is compatible with extant DC supply stuff (like PoE).

      Probably also have to mandate arc detection and interruption. (It’s relatively easy to detect arcing with a DSP and shut down from there)

      1. +1 to this

        If you can have tightly regulated, go for 48V. That’s already a standard.

        If you can have less tightly regulated, go for 42V +/- 20%. That’s also a standard.

        There are a plethora of products designed to work at those voltages already, so why reinvent the wheel?

        1. 48v is good for this, because it is ideally suited for battery configurations, there is a large platform already for stable and long MTBF inverters, converters, etc.
          I think that this is a very good candidate for a DC home, but the arcing is a problem for live circuits, so devices would need to be plugged in and be controlled to only engage the circuit after a good connection.

          Maybe a combination of something not unlike a standard AC socket keyed completely different, and a 1-wire network in place of the ground pin, as ground wouldn’t be needed in the same capacity as with an AC system.
          Then use a “smart” power bar to tell the panel what the max current could be from the outlet, to allow load-shedding and fault protection.

          Also, the current selection for 48v fuses and distribution sucks, and would need to be replaced with something a bit more durable.

          1. Ground is still needed for safety. If this becomes a standard there has to be something to measure against for reference(Think GFCI). I am an Electrician by trade of 16 years. The ground connection in your home is the most important connection you have because it saves your life every day. We will never see the end of the Ground pin on a safe electrical outlet of any voltage or oscillation.

          2. @Andrew Stoner

            AC will always need a ground. But DC is inherently safer, and at the voltages we want, we could replace all but a handful of ac appliances like the fridge, stove, laundry, microwave, etc.
            These would have hidden or obscured ac plugs, which would put them out of easy reach, and therefore have reduced risk.

            But on a socket or box dedicated to DC, there would be no benefit to a dedicated ground pin on the sockets. The low side return wire has the same function as a ground, and there are not the SAME risks present with a DC system as an AC system for a shock hazard. There are still other risks, but they would be mitigated in different ways.
            You would not need a GFCI type breaker on a DC system, because there are no risks of contact shock and propagation like in AC.

          3. 1-wire network in lieu of ground?

            1-wire has its uses but I find its use in wired systems that might experience many mating cycles questionable judgement at best. Much like the poorly designed USB connectors with its low mating cycles being installed alongside wall sockets with higher mating cycles. I’ll end up replacing the “hybrid” outlets long before the AC side wears out. Nevermind any electrical problems from the internal converter.

            I would think that if any new house tech wants to gain a foothold then it needs to reduce complexity and increase reliability in the household wiring even if it means increased complexity in the end device.

          4. I think for the sake of reliability, there should not be a data connection or microcontroller necessary. What I would suggest is to use an extra (low power) control contact, which closes late in the mating cycle of the connector and opens early. This way the power connection is already made and the power is switched on electronically afterwards and no arcing on the power pins can occur.
            Another possibility would be a manually operated mechanical DC switch which is locked in “off” position when no plug is in the socket and locks the plug in “on” position.

          5. @ChanuckFire

            ” But DC is inherently safer,”

            No it’s not. DC can do a lot of nasty stuff, AC can’t. Like keeping sparks much longer, or separating your blood through galvanic processes.

          6. @Canuckfire

            You said “But on a socket or box dedicated to DC, there would be no benefit to a dedicated ground pin on the sockets. The low side return wire has the same function as a ground”. But there is NO difference between AC and DC in this regard – in an AC system it’s entirely arbitrary which line line is connected to ground, and in a DC system it’s entirely arbitrary whether positive or negative is connected to ground. Earth grounding as a safety mechanism is necessary – otherwise unintended earth connections pose a shock hazard. This applies to ANY dangerous level of voltage, be it AC or DC. (If *unintentional* earth connections could be entirely eliminated, then electrical distribution systems would actually be safer without the *intentional* earth connections we now use).

            Also, for the sake of small children, and because DC can sustain an arc, we ARE going to need GFCI’s in household DC distribution systems in the 42-to-48 volt range. Note that a GFCI for DC is both more complicated and more expensive than one for AC, both in the detection circuitry involved, and because of the greater spacing required between the breaker contacts.

        1. Maybe. For a very long time consumers were not allowed to install their own telephones. I was too young then so I’m unsure of the logistics behind it. In any case, I think the rule change occured as early as 1969 or as late as 1974. As a kid, I remember seeing stacks of those old formica(?) telco connectors at my dads recycling business.

          I think the NEC was first published in 1987 though a Google shows some dates cited as early as 1983.

          1. NEC was first published in 1897. Telephone installation restrictions stemmed from the fact that Bell had a monopoly and literally owned the lines and equipment in your home.

      1. Oregon doesn’t care about fiddly details like that. If it’s wiring and you’re paying someone else to run it, that someone else has to be licensed for that type of wire. If you’re an electrician who wants to run network cable, then you have to get a license for that too. The labor unions run Oregon with a very tight iron fist.

        Thankfully, I don’t live in Oregon so a lot of that bolshoi doesn’t apply.

  1. “Because most devices these days use low-voltage DC, with the notable exception of some big appliances.” – there is your answer. Most POWER is taken by devices running happily on AC. Especially asynchronous motors which are perfect for AC.

    So it’s not going to happen, because the infrastructure cost would be, to put it mildly, HUGE.

    OTOH, companies pushing DC/DC switching will tell you that this is the future. Follow the $, as always.

        1. Most modern household machines that uses ac motors ( Refrigerators, air condition, washing machines) are now using some kind of inverters for better speed control and to bee “power greener”. There is already some sort of DC chopped somewhere inside, so jumping to DC powered household items isn’t that far.

          1. In terms of jumping to a largely DC powered household, from a pure physics perspective, I agree. From the perspective of most every house right now only has AC power already installed in their walls, I find it difficult for manufacturers to start to make DC only appliances or even bother to make either AC or DC appliances when the market size is, well, basically 0 right now.

            Chicken and the egg problem. Who wants to double their SKUs and deal with customer confusion?

            It’s the same reason why the US still has absolutely horribly designed electrical plugs. Exposed metal when partially plugged in? Check. No fuse? Check. Ground on the bottom (sometimes)? Check. Heck, even polarized plugs are not always a certainty. But that’s how things are done and that’s what is cheap and available and meets code so we have to keep doing it that way.

          2. Inverters for larger motors don’t operate at low voltages… they rectify high voltage AC and then chop it back up at the desired frequency….. it’s still high voltage AC in the end even if there is a short bit of rectified high voltage DC in the middle (note that it is a short bit minimizing losses).

        2. Air conditioning? – Not in half the world…
          Vacuum cleaner? – ok that runs on a battery
          Dishwasher? – if it’s just heating… why?
          Garbage Disposal? – we have a bin outside…

          So america might get left behind, but the rest of the world can praise our new DC overlords!

          1. DC might be a suitable voltage type, but the higher voltage lines will still be necessary.
            As there, at least in northern Europe, are many houses which only energy source are electricity, and there for uses many kW in heating, cooking etc
            I don’t see a 48V DC line spew out the same amount of power as a 400V AC line does any time soon :)

          2. A typical dishwasher needs in the range of 2 to 3 KW for heating. That would be in the range of 50 Amps on 48V. Then there are other high power appliances like hairdryer, washing maschine, toaster, coffee maker and so on. Have fun wiring that properly and plugs rated for 50 amps are on the hefty side. Just imagine the step down converter you’d need…

            Here in 230V-Land, all these appliances can be run on a normal outlet anywhere you have one. If I feel like it, I can use the hairdryer in the living room. There is something to be said to use one type of outlet for almost all household appliances. The only exceptions are usually the oven/range and, if present, a flash heater for hot water (usually rated for 3phase, 32A). I would really hate to give that up.

          3. @Gerrit: Actually most ovens/ranges nowadays run on single phase – standard models with oven and four hot plates(?) got the oven on a dedicated phase as well as two plates for the other phases. As long as you don’t turn on everything at once they’ll happlily run on a regular 230V/32A or even 16A outlet. Ok, wouldn’t be quite a thing you wnat to show your electrical security inspector ;)

          4. Dishwasher? Because it’s heating! 2kW @50V would be 40A! @24V more than 80A. If you have a car then look at the wires to the starter. Do think it’s a good idea to run such bus-bar like wires through your home? This would be really expensive on copper and in case of a short circuit you can have glowing connections very fast.

            Vacuum Cleaner: I have a battery powered Dyson. It’s good for it’s size, but you can not compare it with the AC powered and the battery does not last to clean the whole apartment.
            OK, garbage disposal also for me means to carry the bag downstairs and out and I do not have air conditioning at home. Luckily here in the office.

        3. Dunno about air con and dishwasher, but vacuum cleaner and garbage disposal would work just as well on a torquey DC motor. Especially if it’s fed with a 48V line, so that’s not much current at all…

          But yeah. I totally agree that DC-only houses will be a long time in coming, if ever. I’m more interested in the intermediate half-half stage.

          1. If your vacuum cleaner uses 1100W today on AC, it will use 1100W on DC too.

            At 48V, you will need to feed it about 23A.

            There’s no way a DC power supply for a home will work on anything less than about 300V. It needs to be higher because the current needs to be reduced. Wall sockets will be intelligent DC-DC converters, and devices will specify what voltage and power they need. Plug in a 6-port USB-C hub and it will ask for 600W of power, approx 120V@5A, which it then down converts to 20V@5A for each USB-C port. As your batteries charge up, the USB-C hub will renegotiate the power supply, 120V@4A, then 3A and so on.

            As for the vacuum cleaner, it is going to ask for 300V@4A. Less current means less heat due to resistive losses, so the heat exchanger only has to dissipate the heat left over from actually doing work.

          2. Running Alex’s scenario further:

            for your average vacuum cleaner cord length of 25 feet, at 48 volts, 23 amps, and assuming a 2 percent loss is acceptable over its length, its going to need an 8-9 AWG cable.

            your average vacuum cord is wired with 17 AWG wire.

            adding more copper onto absolutely everything to account for losses is silly and wasteful.

          3. @ Alex: You are right about the voltage. And therefore I see no use for a DC home wiring. 300VDC are not safer than AC and the switches get more expensive (have to extinguish the arc). And bridge rectifiers are quite cheap. OK you could avoid PFC and the big electrolytics. But what mostly fails are the low voltage electrolytics on the output side of a DC/DC converter.

      1. A typical fridge uses about 60 to 120W, some efficient ones even lower. That is quite easy done with DC. You get even now already DC fridge compressors, e.g. 12V/24V for camping/off grid use, although they are more expensive than conventional ones. They use BLDC motors with soft start.
        Using a conventional fridge on an inverter requires a heavily overspec’ed one, as the asynchronous motor takes about 10 times it’s nominal current at start up. My 220W /600W peak inverter is not able to start a 95W fridge, it just blows the fuse (30A@12V).
        But any kind of electrical heating (washing machine, dishwasher, coffee machine, hot water, microwave,…) basically anything with a power consumption greater a few 100W or, depending on the used DC voltage, 10 to 20A, I see not suitable for a (extra) low voltage DC system.
        LED lights, computers, LCD monitors, battery chargers would be a suitable use.

    1. Breaking this into 2 parts: getting the power from the grid (I think that will say the current AC we have) and self generated DC (12/24/48V DC). With the later I see a local battery technology coming into play. Thing is some government body needs to figure out the standard for that as Big Grid and Big Battery won’t play well together until someone decides for them. And I’m not sure that’s a good thing.

      While I agree cost will be huge and follow the money as that’s a clear indicator of whose interest this really is in. I can see where homes could be retrofitted with newer battery technology. DC-to-AC converters (add grid AC-to-DC also) can be used but I wouldn’t be surprised to see USB ports directly on the battery for charging things. Now how that DC could be sent around the house without the AC conversion is an interesting question. What is going to be considered the cheapest way to do it and what will be considered reasonable power losses.

      1. USB ports on the battery are not very useful. Normally you want to charge your phone in the living room and the battery sits in the basement or some storage room. Although I have already seen 100Ah lead acid batteries in the living room: In India where there are so many power outages, that you had inverters with backup batteries (for a light and the ceiling fan per room) in many places.

    2. The stupidity of the whole debate is that DC-DC converters are internally DC-AC-DC converters.

      If you’ve got a DC source like a solar panel and batteries, and you’re running DC appliances which will inevitably have their own internal DC-DC converters because you can’t charge a cellphone at 48 Volts, then the path of power from the source to the device gets (DC-AC-DC)-(AC-DC) converted. If you have an inverter that feeds your house wiring with AC, then the power conversion path goes (DC-AC)-(DC-AC-DC).

      As you may notice, the number of conversion steps in both cases are identical. You gain no advantage by having DC wiring in the house.

          1. I was just thinking this, and scanning the comments about it… my thinking was more along, if the wire resistance is significant, but inverters are cheap, then why not have in-wall wiring continuing to be AC… while your house source can vary (AC if on old-style grid, DC if generating via photovoltaic, AC if coming from wind/water turbine, etc). Now maybe switching losses can become a concern, but as Dax mentions, most of these inverters go through an AC-stage (even if it is only single-ended, it still looks the same for the skin-effect, right?)… if there was concern about vampire/leakage currents, a simple ‘smart grid’ for your house seems like it could be a solution.

        1. To elaborate on the point: The flyback, boost, buck-boost, sepic etc. do not utilize the magnetic core very efficiently or have other issues that limit the power density. The remaining options are push-pull, half-bridge and full bridge, which you need when you’re converting power for a house, at say 1 kW. When you have a full-bridge topology, what you essentially have is a DC-AC inverter – albeit a high-frequency one. Then you add a rectifier and you get DC.

          Hence: DC-AC-DC


      1. Your right but your wrong. Your not seeing the forest for the trees. The true stupidity of the whole debate, is that it’s a solution in search of a problem. When something is not broken there’s no need to fix it. The current system isn’t broken. It’s serving it’s purpose quite well and the whole argument that we are “right now on the cusp of a power-generation revolution” is pure fantasy. The ideas being talked about here would add significant cost to every home and building built and for 99.9% of consumers would add no benefit whatsoever. That’s not even considering that it would add cost to new household appliances and would mean that people would either be faced with the cost of replacing all their existing appliances, buying some kind of an adapter for them, or retaining at least one or several legacy circuits in their household wiring to maintain them. When you look at all of this you have to ask why? Why should the average consumer put up with all this extra cost and inconvenience for something that will provide them no benefit? That’s why this is a stupid idea.

        if someone wants a DC wall outlet, simply build an outlet with a DC power supply inside. I’ll bet solutions already exist. I’ve seen wall outlets with usb power ports. For most people, who would only want to use such an outlet to charge their phone or tablet, this is enough.

    3. Yeah this topic is constantly on HaD and is just a pipe dream how its presented.

      It doesn’t matter what you can do with something its the things you *cant* do that matter.

      How will my 2kW space heater work as well drawing 83amps at 24v ?

      I think we may move to a DC household where its rare to draw power from your AC on ccertaim parts of the highly developed world and the rest of us will stick to ac.

      Like in Bhutan who got TV 1999.

      1. Dollars will be the decider, and it will decide AC. How much do I pay for an AC outlet? Can be as low as $5, fully approved. How much do I pay for a wall USB outlet? Currently about $30, and at all of 2 amps (10 watts). Might be convenient for charging my phone, but certainly not cost effective. I can buy an outlet and a wallwart for less than a third of what a “USB oulet” costs.

        Even if people really get hung up on efficiency, as Dax points out, when “DC to DC” converters get sprinkled around a house, efficiency will not be better, and reliability will be far worse; and due to poor reliability, replacement and therefore environmental impact will increase.

        This subject has been presented (pushed?) so many times by HAD I think they are running out of properly cool hacks to talk about.

        1. Standard duty 15 amp outlets and light switches in the USA can be bought for as low as 50 cents in quantity one. A construction company will buy so many that for an entire house the cost for those will only be a few dollars.

          “Smart” DC outlets, able to adapt their output to a wide variety of devices, and making the devices “smart” to communicate their power needs to the outlets – while preventing things like deaths and fires from things like kids poking hairpins into the outlets or malfunctioning devices – not likely to ever happen.

          What we do have are things like tablets and smartphones and their AC adapters which are semi-smart in that the adapters have a fairly crude method of communicating how much power they can provide at a fixed 5 volts, and the device detects that and adjusts its power draw accordingly while using internal circuitry to alter the voltage as needed. That works, even though the OEM AC adapters tend to be grossly overpriced.

          For appliances like vacuum cleaners, leaf blowers and others to have such “intelligence” is not so smart. You flip a switch, they work. When you’re done, you turn them off. Many of those things already use a DC motor and a simple bridge rectifier, or four diodes soldered together as a bridge rectifier. They could run off straight DC by switching out the rectifier but no way is there going to be 110~120 or 220~240 volt DC house wiring. That would take welding cable in the walls.

          1. Those 50 cent sockets and switches have astronomically low mating cycles. Enough that I’ve already replaced the majority of sockets and switches in my 2000 home about five years in with more expensive “industrial” rated sockets. About $3 or $4 each IIRC.

            Hell, the sockets in my 1940’s house outlasted the aluminum wiring I ripped out. Yes… I switched out the sockets. But I kept the cool Art Deco wall plates.

            I can’t imagine $30 USB sockets lasting much longer than the smart phone I plug in.

      2. The current drawn by a space heater is largely irrelevant, as I^2R losses merely shift a little bit of the heating into the walls – the house is still heated. And 41 amps at 48v (the house voltage most supported by the international market), for 2 kw is quite modest, with 48v/5 kw inverters being more upper end in the marketplace, and now bog standard commercial products in widespread use. Adequately dimensioned house wiring is just a little bit heavier than for high voltage. Ostriches claiming a thing can’t be done may profit from glancing at the many who have been doing it for some years now.

        But electric heating is just polluting coal-heating here in Australia. I’ve just split and stacked the first 9 cubic metres of firewood for the coming winter. That heats without adding any CO2 to the biosphere, allowing tons of million-year-sequestered carbon to stay out of it. (OK, I have my own forest, so _know_ it is being replaced – in spades.)

        What will advance the practicability of a DC-only house is an improved range of DC appliances and workshop tools.
        E.g. there are now small DC refrigerators incorporating state-change thermal banks so that they require _no_ power overnight. (Thermal storage is much cheaper then electrical.)

        1. >”the house is still heated.”

          That’s a dangerous proposition. You’re wearing out the insulation in the wires and causing a fire hazard by sinking a significant amount of power in the wiring.

        2. In a non-airconditioned house on the East Coast of Australia, and with gas instant-demand hot water and stove (and battery vacuum cleaners) only the washing machine seems to need AC. 12V fridges are pretty common here (caravan and RV types though; I have not searched on 12V household models). And I’d be happy to run an inverter for fridge/freezer/washing machine use; the best ones are over 90% efficient these days.

          Using gas to heat water and to cook, and wood to heat house in winter is what we do here. Going off the grid and on to a 12V system backed by batteries is becoming increasingly attractive as utility companies are privatised, and prices for power go up.

  2. 24v. Because that’s what my battery bank is configured to run at, and because some cheap 12v regulators can’t take an input voltage above 50v, which rules out nominally-48v batteries if you want to drop down to 12v at the end.

  3. I for myself went with 12V – mostly because I didn’t want to deal with chaining batteries and could use ready-made 12V car power supplies. I thought about 24V, but some of my laptops car PSUs didnt support anything over 15V and it would also requite additional controls instead of just shoving cheap 12V LED-strips on the lines. In the last years I converted most of my stuff to DC, I guess ~50% of my power usage is now running over it. A part I had to struggle with was protection – most larger and cheap car fuses are for the audiophiles and wont stand a chance if you actually try to use the current rating printed on them. I ended up with oversized ANL-fuses and some off-the-shelve circuit breakers which are thankfully rated for DC-use from the manufacturer I buy from. I even started deploying RCDs, however since a lot of devices do not have proper isolation after the factory PSU it’s not quite as easy.
    For plugs i use the britisch plugs – cheap to get, fused, reasonable power, polarized and over here nobody knows what the hell the plug is for anyway ;)
    A problem is of course wiring, but for me its still cheaper to throw some money on heavier wires than to get PSUs rated for larger voltages. I guess I will deploy some “transfer wires” in the future using higher voltages to connect local distribution units and allow for longer cable runs without too much voltage drop.

    Another sidenote: Most AC-equipment with universal switchmode PSUs will happily run on DC if the voltage is large enough – 40 to 75V sould be the lower limit for those. This would open up another way of doing this kind of stuff and allowing for a smooth transition.

    1. For plugs i use the britisch plugs – cheap to get, fused, reasonable power, polarized and over here nobody knows what the hell the plug is for anyway ;)

      What’s a britisch plug? If I search for “britisch” I get pictures of cats… if I add the word “plug”, I get the BS 1363.

    2. I hope your using wire nuts to secure the connections, not sure if people outside the UK can handle the safety of our glorious plugs, need to offset it with something dangerous :)

      1. One design flaw is when you stand on the pins when unplugged bare footed, there is no pain like it, Apart from that when I see other plugs from other countries and think they look dangerous compared to ours.

        1. Funny, I look at the UK plugs and wonder if they were serious when designing them for the currents used around a home. I reckon I could push 100 amps through one of those pins! At 220v, that would be over 20kW. Good for heating, which I believe is something well needed in the UK though…

          1. I think the pins could take that no problem would need some thick wiring though. UK homes *normally* have between 80-100amp mains supply fuse @ 240v so I don’t think I’ll try it but the pins are really beefy.

          2. I’m sure normal people in the UK can’t afford 20kW electrical heating.
            Power prices in Europe are not the lowest for normal consumers. In many countries 5 times higher than in the US I understand.

      2. Those UK Super Safety plugs should make excellent melee weaponry. Cut the cord off a hair dryer, tie a loop in the end then give it a whirl so it has plenty of momentum for a good THWACK.

  4. I’ll stick with AC for my house wiring, thanks. I’d rather get a brief jolt of AC that throws my digits off of the live wires, than a brief jolt of live-fed DC which could lock my digits onto the live wires. My heart has strong preferences in this regard. :)

      1. Throw / trigger your nerves to spasm and release (rather than DC which can trigger a spasm and hold)… it’s a muscle reaction, but somewhat involuntary, depending on how you look at it.


        1. It’s still the wrong way around. AC causes a muscle to lock up because it’s repeatedly triggering the nerves. DC just polarizes the nerves one way and they stop working, which makes the muscle pull once and then go limp.

          It takes four times more current to get the grabbing effect out of DC than AC, and with DC the muscle will relax after the spasm.

        2. http://emedicine.medscape.com/article/770179-overview#a5

          “High-voltage DC often causes a large single muscle contraction that throws the victim away from the source, resulting in a brief duration of contact with the source flow. In contrast, AC of the same voltage is considered to be approximately 3 times more dangerous than DC, because the cyclic flow of electrons causes muscle tetany that prolongs victims’ exposure to the source. Muscle tetany occurs when fibers are stimulated at 40-110 Hz”

          1. So we have just the “right” mains frequencies for this to happen :-( But 400Hz whine would not be nice and the losses would be higher and the really big transformers they need for railway power (16,7Hz) would also be expensive.

          2. The dangerous bit with DC is the presence of commutator noise from motors and other switching ripple on top of the DC. That has the same effect as AC in that it can make you grab the wire and not let it go.

  5. There has to be a tipping point where the inefficiencies of using an AC inverter are overcome by the efficiencies of using AC appliances and wiring. As Miroslav pointed out, there will be lots of things requiring DC-to-DC conversion if you go strictly DC which means you could have just efficiently used AC from an inverter.

    There are two things to be defined as “efficiency” — the cost efficiency of investing in the necessary upgrades, and the purely physics-related efficiencies of the energy transfers of the two systems. I can already hear the purists lauding the efficiencies of going straight from a battery to appliance, but if a DC appliance costs five times as much as its AC counterpart, there is still an efficiency loss there. It’s just making the human work harder to earn the money to pay for it.

      1. Crack a book Dax… That is absolutely not true. Buck converters are essentially a PWM circuit that feeds into a back-end LC filter. There is a feedback path through an error amp relative to a fixed voltage reference (eg. band-gap) that controls the on/off frequency of the PWM load switch in-order to maintain a target output level. Boost converters switch an inductor to ground to charge it up then back into a series path with a tank cap to dump that energy and build up voltage. There is never any reversal of polarity with respect to ground reference in either design. Nothing even close to AC.

        1. Sorry, I’m with Dax here. Switched DC is AC, there is absolutely no argument.
          Even if PWM goes from 0 volts to 48 volts it is 48v p-p AC with a 24v DC bias. You should check out some Fourier.

          1. And that might well be the answer we’re looking for. 170V or 340V DC pulsed at 100Hz or so with controlled rise/fall rates and an 80% or so duty cycle. The pulsing interrupts arcs like AC does, but it is much simpler to go from straight DC to pulsed DC than all the way to AC.

          2. “but it is much simpler to go from straight DC to pulsed DC than all the way to AC.”


            The big DC-DC converter that feeds the house wiring with several kW capacity will be a full-bridge switch that feeds a transformer with AC, which is then rectified back to DC. That being the case, it would actually be simpler just to have the AC.

            Of course the circuit would need to be re-designed for 60 Hz operation, but that’s besides the point.

        2. You can describe it like that. But of course the voltage on the coils change polarity between the two parts of the PWM cycle. In the buck converter the input end of the coil is switched between battery voltage and ground (the switching to ground can be done by a diode or a MOSFET (synchronous rectifier). The output end is connected to the output cap, which has an intermediate voltage, so the polarity switches. In the boost converter the coil voltage also changes polarity.

    1. Problem with grid tied inverters is that they must – at least in my region – shut off if the grid goes down. And since I saw how much money was cut from infrastructure maintenance I kinda prefer having a backup…

      1. SunnyBoy has a new feature where they have a socket on the inverter itself that remains up during a grid outage. I feel like that’s a big sell for those places that can have multi-day outages caused by weather or some such. Where I live, AC power by my reckoning is around four nines – more than acceptable.

      2. I thought newer inverters were designed to cut power the grid in that scenario but keep piping power to the house (provided the solar array is generating enough power to provide power for the house.) Then when the grid comes back… they adjust their sin wave to match the grid and start pumping power back out.

        Thought I had read that was how they were being built now… I might be mistaken.

        1. Depends on the method of connection to your panel. Different countries/municipalities handle these things differently.

          My system has a “smart” meter which senses when grid power is lost and cuts off outbound flow. So, my panels continue to insert power into the panel leg to which they are attached.

          Others require fully manual cut-offs.


      3. There are ways to isolate your home from the grid in these situations, and it is the only way I am planning my home.
        It is a massive waste to invest in a solar system that sits idle if the grid goes down, because my house still needs power and I dont want to rely on the power companies incompetence.

        Most of the solutions consist of a grid-aware inverter or transfer switch before the inverter.
        This solution even works with the microinverter solutions if the control panel sits on the house-side of the inverter, and will continue to function if the grid goes down.

    2. “Grid connected solar arrays, problem solved.”

      That’s another approach. The grid helps out as much as a) centralized storage is more efficient than distributed storage or b) the problem is balancing out loads over distance rather than over time. Local is also more capital-intensive, and may require more maintenance by the end-user.

      But, particularly for solar, the power that’s going to be used overnight has to be stored somewhere, so the question is just centralized (plus distribution) versus local. If local DC storage were as efficient as centralized storage (not yet), the equation would go the other way, and we may see that happen in our lives.

      1. What makes you think that local storage is going to be more efficient than centralised storage (the “not yet” comment)? Local storage is limited to pretty much batteries – I really can’t think of anything else.
        Centralised storage can use anything from heat (molten salt/metal/silicon), to hydro systems that pump up and generate down, or esoteric battery systems (eg: redox flow batteries). All these become economically viable when talking MW or even GW. Traditional batteries can’t compete at these scales.

        In Australia it is cheaper to burn more coal than it is to store energy, and that is predominantly because of government subsidies to the coal industry. But some people are still pushing against the trend, thank god.

        1. @ Fred: I don’t think that local storage will ever be more efficient than centralized storage when compared straight up. The wedge between them is the round-trip for solar to LV DC, though, which costs something like 7-15%: invert, transmit, store central, transmit, transform, rectify.

          Home storage (the battery part) only has to get within this 7-15% of remote storage (the “battery” part) to make sense.

          1. But there you’ll have great difficulty. The manufacture of e.g. lithium batteries alone will equal in energy about 10% their lifetime storage capacity, which puts the idea in jeopardy right from the beginning. Check the concept of “ESOEI” for different battery types.

            The whole storage business is a great big “IF” because there just is no good storage solution that wouldn’t double or triple the base cost of electricity – whether it’s remote or local.

  6. Touring caravans already work in this way; 230V AC is used for heating and higher power appliances. With 12V DC used for lighting and low power applications, this is often supplied via PV solar.

  7. I would go with one terminal being +42VDC and the other being -42VDC. This is similar to a centre tapped supply so that if the voltage you are exposed to by either terminal is below 50V but the combined voltage is 84V.ji

  8. “So whatever cabling is used for the low-voltage DC segments of your home’s electrical system, it’s going to be thicker than Cat-5.”

    This is a little misleading, though, since CAT-5 has 8 conductors in it, each of which is at most 22 gauge. A single solid 14 gauge wire is way thinner than the whole set of UTP wrapped in CAT-5.

      1. In this “gauge” system lower number is bigger wire, so 8 parallel #22 must be something like 22/8 or~#2 or #3 :-) But it seems to be more similar to a logarithmic scale like dB.

        1. The “units” utility sorts that out effortlessly:

          $ units -1
          You have: awg(22)
          You want: mm
          * 0.6438033

          The diameter increase for 8 times the area will be sqrt(8) ~ 2.83, and

          You have: awg(22) * 2.83
          You want: awg

          I.e. around half the gauge for nearly 3 times the diameter.
          To read up on the formula, type “help awg” at the prompt:

          You have: help awg

          then hit ^B (control-B) to go back a page in the help. There’s history there, and brwiregauge if you were thinking of BSW, not AWG.

          (If there is no linux command-line within easy reach – we hope the power comes back on soon. ;-)
          ($ man units # is worth a read.)

  9. The standards are multiples of 12v. I have some reconditioned batteries (we have some heavy equipment that if someone lets the batteries discharge – 1700Ah – it takes too long, so a local place does them as a refurb for half price).

    I have an Outback 80A charge controller. Note that Amps matter for the reasons stated above, but even worse, P=I**2*R, so cutting in the current in half cuts the losses by a factor of 4. If you have only a few solar panels or a windmill and a short distance, you can feed it in at low volt, high amp. The 80A is at any voltage (there is an upper limit on the wattage though), which means 24V 100A needs to be reconfigured as 36v, or 48V 50A to use the 80A charge controller.

    FETs are limited (at the top) to about 150V, so that is the upper limit for solar panels in series.

    12V – all your car accessories work including larger inverters, but you see the same amperage problem. Cars and trucks tend to be of limited size so the wiring can be short at 12V. Check out winches that can do over 10k pounds for what is required to wire them. Similar for high capacity inverters.

    The 100W ham radios often work on 12v, but you aren’t going to get your 1500W linear going on that. (Think Field Day).

    24V also seems common, I got a 24V pure sine inverter and a larger stepped one and I have my battery bank at 24v because of the lengths of wire and the currents. There are fairly inexpensive 24V switches and efficient DC-DC converters taking 24V down to 12V.

    Relays are mentioned, but you are looking at (automotive starter) relays to switch things. Also, if you have tiered pricing, you might want AC to charge DC during the cheap hours.

    Another consideration is when your batteries are fully charged, you might want a grid-tie “push” inverter to move your electric meter backward. I’m not in a place where they will pay me for electricity, but I can pay far less by pushing power back onto the grid. These are also in 12V/24V variations.

    1. FETs work up to medium voltage without issue (IXTL2N450 is a part I’ve used recently. 4.5kV, 2A). Most the PFC stuff is using 650V super-junction FETs at low power (<300W or so). As you get above about 200V and a few amps though IGBTs often become lower loss at least in silicon. If you go to GaN or SiC it gets even better.

    1. Exactly.. Up to 100W, and usb-c sockets in the wall. All tooth brushes, shaving machines etc just start using usb-c plugs instead.. (which would make sense anyway). For households not having native DC voltage wiring, usc-c wall warts or sockets converting AC to DC make sure products are not limited to DC households.

      It would be interesting if even the DC wiring in the walls was smart like USB PD, requesting different voltages depending on needs directly from the batteries.. maybe even AC. Would decrease the number of DC-DC conversions needed, and keep AC out of the sockets unless necessary.

  10. I would go for a dual voltage system, maybe +48V and a +12V tap on the system.

    Most wallwart driven stuff needs less then 12V, and heavier loads could be connected to the 48V system.

    Also there is a lot of 12V stuff for cars and RV’s already, like waterboilers, microwawe owens, coffe makers, fridges, soundsystems, tv’s and so on .
    (sure there is also 24V stuff for for trucks)

    Then you could have something like 4 different taps on the battery bank, with a smart balancing system, and wire up the solar panels for a 48V system

  11. Computers and a lot of other domestic electronics run generally at 12v & 5v (with some 3.3v), yet if you use 12VDC the resistive losses around the house will be huge. Even at 50VDC suggested above the losses will be greater than your 230VAC. I think something that the article doesn’t quite address is that to drop the higher DC voltage to something more usable there will be a DC to DC stage which is no more efficient than the AC to DC stage currently used – so really I’m not seeing the advantage.

    There is the trend that old electric car batteries will make their way into houses as ‘behind the meter storage’ to smooth utility demand, and create a dynamicly priced smart grid. Perhaps there is a way to utilise the PV panels to charge the battery strings within the ‘behind the meter’ storage directly (with no voltage conversion) rather than changing a house full of legacy wiring?

  12. 48v is a good compromise between safety and efficiency, and there are plenty of off-the-shelf converters. Circuit protection is still nontrivial – a big 48V PSU isn’t very different from an arc welder.
    Another issue is maintaining efficiency of the incoming mains to LV conversion over a wide range of loads.

  13. Use automotive standards, be it 12v or the newer 42v (or the .mil 24v). There are already plenty of components.

    Some say DC has too much loss but thats just in lines to/from your house, the runs inside homes are close to that of a big RV. We need to get a standard plug developed soon. HAMs informally standardized on PowerPoles so maybe thats an option. But we need to get a standard so it can be incorporated into newer houses alongside the 120v before 120v can be phased out.

    1. What about high power appliances like a 2400W air conditioner. The wiring would need to be thicker than car jumper leads and copper isn’t cheap.

      The reason we have 120V / 240V systems is simply economics.

      1. And industrial use generally uses three-phase at 415 volts (Australia) and 380 volts (EU & others). Don’t know about the US. Very simply put, less power loss over thinner wires which are cheaper and quicker/easier to install, and therefore cost savings on all sides.
        Dollars will be the decider, not idealists.

        1. EU standardized for 230V/400V many years ago. Basically as a mean value between continental European 220V and British 240V. Perhaps the 415V is the delta-voltage of the 240V system?

  14. Well if it weren’t for the High Voltage I think you could switch to from AC to DC with just the Flick of the Switch, much like switching from dynamite to TNT, they’ll both “Blow Up Your Video”. But it would take some convincing of the guys in the neck ties to work on the Contracts to get Back In the Black (but I’d love to be a Fly on the Wall during those meetings). It’s not like Jail Breaking a phone here. This is not a dirty deed, done dirt cheap. Money Talks. BS walks. Just thinking about it all leaves me Thunder Struck. I’ve been on this Highway to Hell before. Not to mention getting across some Live Wires Shook Me All Night Long one time. You’ve got to have Big Balls for messing with this sort of thing. Really skating on the Razors Edge or Black Ice. You’ve got to keep a Stiff Upper Lip in field of energy. I think you all need to calm down and have a drink on me. I know some of you are saying “Let There Be Rock”, so I guess For those About To Rock, I Salute you?

  15. Thinking about yet another factor at play – how about AC for ‘centrally coordinated’ consumption, and DC for on-demand consumption?

    Dishwashers, A/C and the like – they can pretty much run whenever, and that includes when the grid is providing baseline load. I think it might become attractive to have meters/contracts that would disconnect if the spot price is above a certain threshold. And for the things we need now – we’ve got DC and batteries!

    And my answer is 12-48V. Appliances should accept as low 12V (most of them should just do step-down DC/DC anyways), the building can choose to produce up to 48V – the extra cost on the appliance side is quite minimal, and it would allow people to choose their own local voltage as they see fit.

    The only problem are dumb (resistive) appliances – LED strips etc. They could just be HF-switched for thermal reasons.

  16. I vote for a wiring system where the wall plug is usb type-c with QC3.0. This allows the receiving device to decide wall voltage. On the back end from wall plug to breaker box a similar method can be used such that whole circuits collaborate to adjust line voltage. At the breaker box we would take into consideration generation voltage, conversion cost, time of year (is ambient heat bad in winter?) , and the overall wattage requirements of the circuits. Maybe come up with a timing system with stepped voltage where each plug picks what voltage their turn will be.

  17. am i the only one that thinks the basic premise here is flawed?

    with the amount of appliances and possible load an average family home can be under i think there would have to be far too many circuits or the wires are going to have to be far too thick to be economical.

    even if you take away the large built in appliances like washer, dryer, dishwasher, oven, microwave etc.
    there is still a lot of possible use at any one point in a house, a couple of tv’s and computers, the infrastructure to run them plus the available capacity to allow for household tools like a hand drill still leaves you with a dc system that need to handle at least a kilowatt, more probably 3, to be useful, even at 50 volts one would be looking at 20+ amps on the conservative side, you would need conductors 3mm squared to safely transport 1kw at 50v.

      1. But 3 mm (1/8″ in the old money) conductors are skinny. All the solar panel cabling starts at 4 mm diameter, and for just a few tens of metres around a house, it’s cheap and easy to run a few in parallel for higher currents. At half the current per wire, it’s 1/4 of the losses, three give 1/9. How far is it from the batteries to the aircon, anyway?

    1. Well, I guess that totally depends on your lifestyle. Dryers – well, it’s warn enough to dry by itself. Dishwasher? Takes more time to load than to just rinse something off. Ovens mostly run on gas over here. Computers nowadays are usually tablets or for powerusers laptops – not quite power hungry and TVs? Is anybody still using these? You see: Power consumption vastly varies depending on where and who you are…

        1. However AC not the standard for those who don’t have economical access to reasonably price AC service. You want AC far from the grid you pay for the installation and maintenance from the grid to your chosen location. A person living in a rural location may find it less expensive using the DC power they provide for themselves and still live comfortably.

      1. that was sort of my point, one system fits all brings the economics of scale to bear in a different way.
        it also makes it easy for people on the other end to design for their customers, in a lot of devices the voltage requirements are fairly strict, sure 10-20% wont make much of a difference in most cases, but going from a 12v rated device to a 42v? how about conductor sizing, should they be overkill so you know they wont burn out or should we have several different sockets for different wattages and voltages, how many sockets do we really want?

        sure we could make them smart, but then every single product needs to be compatible with that specific protocol and any legacy protocols that accrue over time.

        a simple kettle can be a few kw in itself, everything adds up, the real issue is that for this concept to work it has to be more than just a bit more effective or economical, it has to justify the cost of changing, plus the cost of standardization and compliant design, as it is today if i made a small low voltage gadget that i wanted to sell i could buy power bricks that were already certified and just deal with the much less stringent low voltage certifications, plus i wouldn’t have to include a wide range voltage converter in my device, for something portable every mm2 matters.

        in short there is a ton of reasons why i don’t think dc will ever be the de facto standard, sure there are niche use cases where it makes perfect sense but there are a lot of issues to deal with before dc power distribution is practical around the house.

  18. As others have pointed out, split the problem and go DC for the lighting and other low current devices that are wired in, the rest may have to wait. One thing with motors is the starting vs running power, perhaps putting super capacitors into appliances could help? Also people tend to design houses, then work out how to run the power, perhaps we need to let power efficiency considerations influence floor plan layout more? Get it right and you may find larger low resistance DC bus-bars between areas then become economically viable?

  19. “it stops making sense to convert the local DC to AC just to plug in a wall wart and convert it back to DC again.”

    But that’s exactly how a DC/DC converter works! It converts DC to AC at some frequency predetermined to create the most possible RFI on all bands, runs it through a transformer, and then rectifies it back to DC, often leaving all of the harmonics intact for the rest of the world to enjoy as reduced radio, TV, and phone reception. Brilliantly stupid.

    1. As a practicle matter currently 12 volts would be the choice for anyone deciding to live for any reason off the grid and is going to have use DC. So many devices that many want to use use 12 or 5 VDC. Their current demand is low enough that 14 or 12 gauge wire should suffice, even if more circuits may have to be installed. Of course things item as central HVAC, electric hot water, kitchen ranges and cloth dryers are out of the question. Bottled gas can be transported in to most area easily enough to provide for a kitchen range hot water and even operate an absorption cycle refer, and supplemental heat if needed. An inverter may have to be included if a water well pump used. I belie 12 VC of those are available, but very expensive. While they are something stores will keep in stock one would have to purchase their own back up unit. That’ thing about being your own utility company, it’s up to you to keep backup components in stock, so you can do your own repairs, or hire another to do so if you are unable. Only in the event you unwilling to change your easy and convenient lifestyle would you have to use more that 12 VDC provide enough electrical power to live better than most on the planet.

      1. No I am not willing to change my lifestyle, I want it similar (or a little better) as the other people in my country (in central Europe). “live better than most on the planet.” is not enough. I do such things like you describe sometimes on weekends. For camping, festivals or in the garden. There I only have a small 12V solar system for some LED lights or music because I did not want to invest into the grid connection as I don’t have a house there.
        So I know the limitations of a 12V system. Even the 12V compressor fridge drains the batteries in some hours.
        Heating and hot water can be done with gas. If you want, you can do without a clothes drier, but to have laundry to dry you have to wash it first. And a washing machine on 12V DC will not work.

    2. Humorous or satirical intent not indicated. Respectfully, at best it has to be brilliantly ignorant, to suggest that buck boost converters are purposely engineered to create as much RFI as possible.

      1. Have you ever taken a look at some of them on a spectrum analyzer? I have, and I haven’t found one single switch-mode device with a clean output, and most of the cheap ones which have flooded the market have little to no filtering at all.

  20. Has anybody ever been in a motorhome?

    We used to have a 1978 Winnebago.

    Everything worked when unit plugged into AC. At 5 MPG, Dad parked out in the country by a pond we fished in. Had a utility pole installed.

    I believe all lights ran off inverter, probably 220VAC to about 12.5VDC.

    The refrigerator was happy with 110 VAC, 12 VDC, or propane (motor home must be level, not hard, wish I knew how many BTUs

    My ~1,900 sq. ft. residence was recently flooded with 6′ 4″ of nasty unsanitary water in August 2016. I have been thinking a lot about lighting, doing away with wall switches & remove Cu wiring; replace with a momentary switch connected to a Raspberry Pi, Adriuno, etc. Not sure as I am not familiar enough with the different manufacturers. Perhaps no visible switch as a sensor behind the STD light switch location could work. With LED lighting, relay 4 power, or possibly a transistor.

    All I know is that I prefer the entire system to be based on OPEN SOURCE HARDWARE & SOFTWARE.

    Does anyone know of a way to find the right LED(s) that do not require:
    A fan / heat sink.
    5,000 Candle Power (one or combined)
    180° 3-D light dispersion
    Doesn’t cost an arm and leg.
    The industry changes pace so fast, more light, less heat, less power, lasts,.

    About a month ago saw a 5k candle power LED w/fan & heat sink for $20.

    Don’t know how to find it’s older cousin, say 1K candle power each with a good temp (near full color spectrum, decent light dispersion (ceiling mounted). Priced around $1.00 EA.?

    Or suggestions on any known Open Source Home Control Hardware / Software.

    Need to be able to have a few electrically-controlled water valves.

    THE #1 residential energy hog is the electric hot water heater that we insist on supplying with COLD ground water. I’m going to split the system before it enters the home, think of every way I can to preheat 40 gallons like a roof-mounted tank/radiator & try to put inside a glass or clear plastic containers, painted black, perhaps (as I need a new roof), but if there was a way for the water to remove the heat from the +100 metal roof, every BTU counts.

    Perhaps based on some of the cooling tech for computers.
    My dream of repeating a cooling system for a generator with no water pump, a living off the grid article used water or coolant whose flow of hot water away from generator (closed loop) & had a tank with enough water to ensure the Genny never overheated.

    Now we have self-contained chains of plastic squares that use 0 lade, float in ocean, bacteria introduced inside & produces more power per unit than solar. I may have a few words wrong, check out Nasa Tech Briefs.

    They also have a closed cooling system with refrigerant that requires no compressor. I was gonna patent that idea, been thinking about it since 10th grade before I won 1st place at the annual Mississippi Private School Education’s Association’s annual state science fair competition. Without my knowledge, but on my behalf, they contacted the equivalent state of MS entity to request I be allowed to participate with the public school students who won 1st place in their respective states so that I may compete on a national level. Basically, “Yeah, right… when hell freezes over ya “boy” might have the same chance as a snowball before hell frroze over.”. That was over 30 years ago. I see my project, based on aiming the solar panels towards the sun as it passes overhead & upon dropping below preset threshold, return to initial AM recorded position to make adjustments for next AM.

    Anyways, suggestions welcome.

  21. I’d like to see a dc wall socket carrying multiple voltages, like 5, 12 and 19V or a socket that allows the device to select the voltage it needs, just like some of those psu bricks with resistor selectors.
    The thing is, almost all dc devices use an internal regulator or dc/dc convertor to change the voltage once again, even multiple times in cascade. By eliminating one or more conversions may save a lot of power, it adds up quickly. A centralized converter may be easier to make efficient, and easier to replace when defective or when more efficient designs become available.

    As for dc only for solar powered circuits, it may be that there’s much more power available than usually small currents of dc devices uses so what to do with the surplus if not pumping it back into the grid? It may very well be that your neighbor will use it so the distance your supplied power covers is much lower than that of a powerstation.

  22. Hi Guys, I guess you all know about the following:

    * Series arc fault and arc fault risk in DC systems
    * Thermionic emission and field emission of electron and ion current
    * How to solve nonlinear equations for finding the DC arc resistance

    For those who don’t: This is relevant for all DC voltages above 30 volts, give or take. Google “series arc fault in DC systems” will give you an idea of the problems involved.

    Even though a high voltage is required for ignition of an electric arc over a certain distance, this is not the case for sustaining an arc if you seperate two contacts. E.g. when opening a switch or in case of a loose contact.

    Do not use more than 24V unless the system is specially designed for handling series arc-faults.
    If it was so easy, the automotive industry would have used more than 24 volts long ago.

    Except you want to build yourself an electric arc welding system instead of an electric power system.

    ==> Use 24 Volts and fuse the circuits with 15 Amps.
    ==> Or use 12 Volts and fuse the circuits with 30 Amps.

    Unless you have proper DC arc fault protection in place.

    1. Of course one shouldn’t arbitrarily fuse circuits based on the supply voltage. The current carrying capacity of the conductor being protected determines the fuse used. There is reason the fuse panel in a motor vehicle has mix of fuse values The current requirement of a device determiners conductor size used and the value of the fuse used.

  23. I would opt for a dual output standard. 500W 60V for laptops, garden lights, small pc’s, phone chargers, flatscreen tv’s and stereo’s. 4kW 400VDC for gaming pc’s, large flatscreens, aircon, vacuums, dish washers, welding equipment,driers, EV charging.

    60V means easy integration in about all low power system. Non Selv voltage, so all equipment must still pass regular electrical safety / regulatory standards. (should keep dangerous equipment flooding our homes)

    400V means you can rip out the entire PFC portion of about all power supplies. All sockets will need a relay or mosfet to disable the output prongs for save making/breaking the connection. some sort of power request communication may be required to prevent serous overloading of the x-kW PFC.
    Also smart-grid / solar panel balancing should be easy!

    1. This just-above SELV voltage is no good idea. Or do you work for some kind of regulatory body and have the wish to generate more business (or have the fear of losing business)? So if it is a low voltage make it SELV.
      Further I don’t think that one “Mega PFC” at the entry point of the house which is most of the time in light load operation, can be very efficient.

      1. If you check, I think you’ll find that 60v _is_ SELV in Australia. Please remember that the 4x12v lead-acid batteries, used for a century in telephone exchanges (Am. central offices), will be above 50v much of the time, and approaching 60v when charging.

        If your neck of the woods only allow 50Vdc for SELV, then you can only have 3x12v (nominal) batteries. That would be singularly daft.

        Yes, connection to the battery bank must be via HRC fusing and a rated isolator. (A HRC fuse is typically ceramic bodied and sand filled. It is rather good at quenching the rupture arc.)

  24. 1. RFI is a very real problem.
    2. Component costs will plummet with usage. i.e. LED bulbs.
    3. Microwave ovens are power pigs.
    4. My yurt does well on 12V (but it’s only 16 ft in diameter.)
    5. Alternate sources are important in the north, mountains etc. Most generators don’t have high DC output.
    6. DC vacuum sweepers are wimps.

  25. Btw, another severe cost issue with DC systems above a few tens of volts is that you need newly developed plugs with an electrical/mechanical interlock and/or dual pilot contact in addition to the individual MOSFETs / DC contactors for each point of load. You really cannot plug or unplug a DC load without equalizing the voltages and interrupting the current first.

    1. Not sure what you mean by”few tens of volts”, but most of this article was concerning >50 VDC, and relatively low currents. On par with current levels being switched in motor vehicles every.

  26. IEEE Spectrum has done several very interesting articles on real-world DC microgrid deployment in very remote locations. By microgrid, they mean about 1-4 houses. The power source is solar (with batteries for nighttime), and the area is sufficiently remote to make grid connection difficult, expensive and unreliable when it does exist. These are situations where the ‘last mile’ (or 10) is literally by pack animal and foot.

    This system uses a 48V DC system in the house: http://spectrum.ieee.org/energy/renewables/innovative-direct-current-microgrids-to-solve-indias-power-woes

    And this seems to be either 12 or 24 volts: http://spectrum.ieee.org/green-tech/solar/lights-for-the-enlightened-an-engineering-trek-in-the-himalayas

    Obviously, solar generation in the range of 125-250W is tiny by developed world standards, but this size of system can power a TV, computer, phone charger, ceiling light, task light, and a fan for a more comfortable sleeping environment.

  27. I too, have been giving this topic a lot of thought lately. I bought a bunch of surplus solar panels about a year ago. It seemed silly to me to convert DC to AC since most things that I wanted to run would run on either AC or DC. A light bulb doesn’t care. Even most LED light bulbs rectify the AC before applying to the LEDs. Even most switching power supplies we all use to charge cell phones, and run most appliances will run on 100-240 AC or DC. I am going to use 120VDC for my application because, with the right screening, of course, most things I want to run (lighting, phones and even my LED TV) will run just fine on DC so it makes sense to use a voltage that is acceptable to those things.

  28. I vote for SMART VOLTAGE SYSTEM that can change its voltage by negotiating connected device.
    Hey IT IS 2017!
    Why do we have to use only ONE? Think like Quick Charge 3.0 or USB-IF…

    For old fashioned people, I advice DC V400, because of compatibility.
    Indeed almost every device we use right now (of course that doesn’t include an alternative motor in it) is compatible with it. Your TV, Laptop, PC, Printer, Scanner, Powersave CFL or Led bulbs, Boilers, Also Modern {ACs, Refregirators, dish Washers} that has DC motors or drivers in it can usable with DC400V. Because this machines already convert AC220 to DC 400 before use. They have DC 400V rated caps and bridge diodes.

    1. No, I prefer a reliable power system, with no SMARTs and as little attack area for malicious hackers and failure modes as possible. If this “smart” crap fails it can easily destroy the connected appliance with overvoltage. So keep the system simple, have fixed voltage at the wall socket and a (wide input range) PSU in the device.
      I also have many wall sockets in the house or appartment which are not all used at the same time. So there should not be electronics in the socket which increases cost and standby losses

  29. Dual 12V (+12V GND -12V). When I was doing reno on my house I ran 12V for the lights, regret now not running the dual setup. Gives twice as much power with only one extra wire plus the option of running 24V loads. 14-3 wire is cheap, giving you 360W of power. More then enough for most common loads. AirCon, fridge and stove run on 120V/230V.

  30. 48V is already a widely used standard, in photovoltaics, in PoE, and in enterprise data centers.
    It remains under typical regulatory limits for ELV/SELV at 50VDC, and yet is about as high as you can get away with, therefore minimizing ohmic losses.

    1. +0.9

      Here in Australia, IIRC, SELV is up to 60 Vdc. That’s still only 4 x 12v lead-acid batteries, as they typically need 14.2v for charging, and heading toward 15v each for equalisation. I.e. 48v (nominal) battery is up to 60v exposure.
      Importantly,I can legally wire that without an electrician’s licence, given appropriate experience.

      As for the other consideration – arc extinguishing (and the concomitant fire danger), the battery bank must be equipped with a HRC fuse, and an isolator capable of arc extinguishing. (I have an isolator with the HRC fuse built in.) I’m not much inclined to switch high current DC loads with relays, as MOSFETs do it without arcing, and if 5 milliohms RDSon isn’t good enough, just parallel two. (With individual series gate resistors, I suggest.)

      A small 24v system is good for LED lighting, running a computer, modem, and soldering iron, and a 24 vdc pressure pump for an off-grid holiday home, but not much more.

      The decider for my pending off-grid tree-change is that 48v battery-inverters are common and inexpensive. Converting battery energy to 240 Vac is still needed for most appliances, even though 48 Vdc aircon is available now.

      The decades-old voltage standard for telephone exchanges (Am. “central office”?) seems to have some life in it yet.

  31. oh, come on. Those lousy humans can’t even agree on having the same type of hole in the wall that gives you electricity and you want to replace the whole infrastructure with something else now?
    (actually this might be an opportunity for a universal plug type)

    1. For lower current I like the DIN ISO 4165 https://en.wikipedia.org/wiki/ISO_4165
      power socket. It is a little smaller (12mm) than the cigarette lighter socket, but makes much better contact, as the center pin does not try to push the plug out of the socket but is connected and hold in by spring contacts at the side. Unfortunately it is not so widely used and I did not yet find it at cheap chines suppliers like on AliExpress.

      1. If you search for “hella plug socket” on ebay, you’ll find e.g. two of each for A$25. That’s probably as good as it gets for something which is used in boats and caravans. The quantities are perhaps too small for the Chinese suppliers.

  32. I lived for years in the 1970’s and 80’s in a home on 114V DC with a Dunlite wind generator and 19- 6 volt golf cart style batteries. We ran everything, lights, vacuum, tools, flat iron, etc. “F” rated normal switches work, normal wiring works, universal motors run, switches in some appliances need a capacitor to stop that spark when the contacts open like our Kitchen aid mixer and the vacuum cleaner. It was easy. In my RV I went to 12 volt as so much is available, like TV’s, chargers for the computer etc. Now we have lots of DC to DC voltage converters and electronic inverters that are efficient which didn’t really exist then or were too expensive for common use.

  33. No, no, no. This will never happen. In the future it will be the way of Nicolas Tesla, and it is already starting. Your wireless phone charger produces a magnetic field that the phone translates into DC with a simple coil. All you have to do is make the magnetic field the size of your house. Everything will run and nothing will need cords.

    1. Magnetic field strong enough for wireless powering of the whole house will be so strong that it’ll be dangeroous. Especially for people with metal implants, braces, earrings etc. Also, the RFI will be massive.

  34. I am going to go with the “48V is already a standard” crowd here. Tales of world ending DC arcs are nonsense if you are talking about modern electronics. Should you hookup a 48V battery directly to some wire and run it all through the walls of your house? No, that’s not the best idea. Every office building, and many data centers, are wired inside out with 48V PoE. It uses dark mysterious magical voodoo to determine when a fault occurs and it stops supplying power instantly. It’s called…resistance. The voltage is known, the max power capacity of the device is negotiated, and if the current exceeds the negotiated limit you disconnect the load. If you just PoE chipsets even without the networking you gain the power management benefits. Plus you still have Ethernet sitting there if you want to use it. One downside to this is that every outlet needs its own run back to the electrical panel. Though each room could have its own panel in this instance.

    These folks stopped by our makerspace last year and I helped them 3D print some parts for a prototype. They have a lot of interesting ideas on this topic and basically have done what I describe. They have a nice DC panel design that has various kinds PoE-like modules that plugin much in like breakers in an AC panel. It can supply Ethernet and other “features” to the outlet. (Not particularly a fanboy but I have some experience with 48V systems and PoE and generally like their ideas.)

  35. “Because most devices these days use low-voltage DC, with the notable exception of some big appliances. Batteries store DC. If more and more homes have some local DC generation capability, it stops making sense to convert the local DC to AC just to plug in a wall wart and convert it back to DC again.”

    Faulty premise.

    “Batteries store DC”
    To provide useful power without excessive I^2R losses, you need a battery pack of at least 24 volts, likely 36V or 48V.

    “most devices these days use low-voltage DC”
    Low-voltage DC at various voltages, 3.3V, 3.6V, 5V, 12V, 19V [and quite a few of the 5V & higher devices will be internally dropping to lower voltage(s)]. What you won’t find are a lot of devices that run off 24V DC or higher, so “Batteries store DC”, “devices use DC” … Profit! is bogus.

    The reality is that household power needs to be able to supply high consumption [kW range] devices, as well as a plethora of low-power devices that require a wide range of input voltages. AC does that job well. Switching to DC will just mean additional DC/DC conversion is needed for devices that don’t match you reticulated DC voltage.

    One place you *could* gain efficiency is switching to 400Hz AC, if only you could get appliances that could handle it.

    1. I am not sure if this is a good idea in the household. 400Hz is much more audible than 50Hz (or 60). So if devices are not built very well, the noise can be annoying. 400Hz AC would also produce more capacitve losses in the power transmission to the houses, I am not sure about transformers – they have more hysteresis losses per kg of iron but need less iron.

  36. In Argentina many houses on the countryside doesn’t have electric services, so the run generators or solar panels at 36V nominal.
    it’s been like that since the 1950’s or so.
    i went to a house with a really old diesel generator that still works now a days,
    Also the lightbulbs and fridges are meant for 36v (and also you find 12v ones)

  37. My home uses a time-sliced DC system.

    The DC voltage in the wires starts at zero volts (useful for devices that are turned off) then spends a short period of time at every voltage up to about 160 V, then comes back down to zero. It does the same thing for negative voltages, which is useful if you plugged in your device upside down (USB should learn from that). The whole process takes about 16.7 ms so you never have to wait very long for the right voltage to come along.

  38. Just what DC voltage to choose seems to have generated a split consensus.

    High wattage devices will have to remain on the AC grid. We don’t have enough copper for low voltage DC at high wattage so I won’t be going there.

    “42” has more agreement than any other response. Historically humorous (thank you!) and interesting, as well as appealing to the lower end of the range for copper costs.

    Multiples of 12v were suggested and seem a convenient use of off the shelf stock to have the commute convert more gasoline into energy, at the expense of your alternator life, to charge batteries to run the home… effectively about the same as mounting a windmill on the car to charge a battery. Believe this has already been tried… Red Green Show had an episode with it. In any event it’s moot because soon we’ll be charging the car from the house.

    Advise that I’ve already tried the solar cell option. Code here won’t allow such batteries in a living space due to outgassing so there is a shed or an outdoor box. You water the solar panels to clean them and the dead lawn underneath results in mud you track into the house. Fallen leaves must be swept off daily, then raked off the dead lawn. This was solved by putting the panels on the roof but now can’t get at them quickly to flip/cover when the hail storm 2x a year comes along.

    I want the DC for lighting… perhaps a small fan or two… a laptop… the router and modem… an alarm clock… weather radio… and two cell chargers.

    All that to get rid of wall warts?

  39. My 0.02 €
    DC in the home is a dumb idea. Maybe it could have made sense a few decades ago, when low-power power supplies were shit and inefficient. These days, modern switching power supplies almost don’t care if the source if you feed them DC or AC. And they are about as efficient with AC as with DC as a source.

    Switching to DC will just make us lose all the benefits from AC. Transformers are still a thing! It’s naive to say otherwise.
    Even if they are only used for galvanic isolation.

    Dumb idea, NEXT!

  40. Interesting mental exercise… but pretty far from being a Thing for most of us.

    I can see some sort of standard evolving around residential ‘secondary’ DC sub-circuits, probably for LED lighting, like we’d currently do when say using 12 v lights as under-cabinet lighting, but I don’t see much practicality in a house-wide DC standard with outlets, that would replace the AC wiring.

    If anything, the next step could be common PSUs for systems – like having ONE DC power supply that powers everything in your computer system – the PC, monitors, printers,, speakers etc. Or one DC power supply for your home theatre, for the tuner, power amps, big monitor etc.

    If you’re off-grid or have flaky power, sure you could design for DC distribution. In RVs and boats, 12v is the standard, and there’s lots of hardware and components for that use. though the “cigarette lighter” is not a great standard as an outlet.

    Our 19′ boat has two 12v sockets, some built-in 12v LED lighting, and six USB sockets driven by one cheap buck converter, for IKEA gooseneck lights and device charging.

  41. I know I’m not the first to mention it… but for my money, 12V DC at the outlet is the way to go. The recreational market already has tons of devices manufactured for that voltage, its home hobbyist friendly, and doesn’t carry a great danger.

    Depending on how much bank I had when I built a house, and how much things like DC-DC converters cost, I’d probably go with either a straight 12V system (probably rectified to 13.4V at the mains box with inline regulators in the jacks), or use 48V internal and step down at the jack. As the article mentioned, if I were doing solar power, I’d still have that choice to make…since with DC sources, final voltage is dependent on the wiring more than the source (in that solar cells are usually ~1.2V individually anyways, so unless that’s what you’re running, you’re chaining them together in some way).

  42. If it’s a small system I’d go with 24V as there’s plenty of low cost hardware such as DC-DC converters for use in some heavy trucks and aircraft.
    A large system 48 volts but expect to pay more for inverters though you might make up the cost in being able to use thinner wire.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s