9
\$\begingroup\$

I've been thinking about this for a while. Now that we have means to generating electricity in DC, what would it take to do a DC wiring for an house to work along the traditional AC wiring.

Specifically, I would be very interested in understanding the formulas to compute dissipation, and get advise on how high should be the voltage for distribution to avoid wasting too much power due to omh's law. It is fair to assume that every outlet will be no further away than 30 meters (90 feet) from the source. Also, worth assuming that the outlets would be anything from a 5v usb, to a USB-c that can negotiate up to 20v.

Power wise, each outlet will need to support from 10w to up to 100w.
What would be the formula for computing resource waste of the buck conversion? DC Solar (13v) -> DC (distribution) -> DC Outlet.

I think the final question boils down to: can this be more efficient than a system that has to do DC -> AC -> DC conversion (which I believe has a 89% energy efficiency).

\$\endgroup\$
  • \$\begingroup\$ Comments are not for extended discussion; this conversation has been moved to chat. Any conclusions reached should be edited back into the question and/or any answer(s). \$\endgroup\$ – Dave Tweed Feb 20 at 19:54
22
\$\begingroup\$

Loss at voltage conversion is just a function of how good your converter is, 5% to 20% losses could be expected at each conversion.

Losses in distribution are the real killer. With 120/240 volt distribution, we don't really think about distribution losses until we are running many 10s of metres of cable down to an outbuilding. If the cable is thick enough for the current, the volt drop is negligible, or at least a negligible fraction of the total supply voltage.

However, with low voltage distribution, you get hit by a double whammy. The currents are higher, to transmit the same power. Any voltage drop is a larger fraction of the voltage you start with, so your power losses are proportionately larger.

There are three practical regimes to consider for DC distribution, 12v, 48v, and 'high' voltage.

12v is worth considering as there is so much stuff around that uses that voltage. However, you need stupidly thick cables to supply an area much more than a single room.

48v is around the highest voltage that's deemed to be 'touch safe'. As it's four times higher voltage than 12v, you only need 1/16th of the cable cross section to supply the same power to the same area with the same losses. It's used extensively in the comms industry, and so there is a significant amount of equipment available for dealing with it, inverters, converters etc.

'High' voltage, so 250 and up, is often used to connect solar panels, and in electric cars. Compared to AC, it's more difficult to handle, arcs don't go out as easily, so you need special fuses and switches. I wouldn't really recommend it from a safety and ease of use point of view for distribution in a home.

| improve this answer | |
\$\endgroup\$
  • 1
    \$\begingroup\$ And having 48VDC wiring around your house wouldn't really be useful would it? (Unless you're using comms equipment that can accept it directly). You'd still need DC->DC converters to power real stuff (like phone chargers, laptops, etc), so you're just going to get more ohmic power loss compared to mains voltage on top of still getting DC->DC conversion losses. \$\endgroup\$ – Peter Cordes Feb 19 at 19:57
  • 2
    \$\begingroup\$ I'd go with 48V without even thinking more about it. Low enough to be safe, common enough to get off-the-shelf equipment, high enough to power most of what you want. \$\endgroup\$ – pipe Feb 19 at 20:19
  • 2
    \$\begingroup\$ I'd love if I could get the formulas to compute the DC->DC conversion loss vs the DC->AC->DC loss \$\endgroup\$ – Snick Feb 19 at 21:19
  • \$\begingroup\$ @Snick DC->DC conversion is DC->AC->DC. It's just that the intermediate is generally at much higher than mains frequency, and probably (but not necessarily) lower than mains voltage. \$\endgroup\$ – hobbs Feb 20 at 1:36
  • 1
    \$\begingroup\$ There are two things you ned to know about the efficincy of DC-DC converters. A) general and B) specific. A) In general, the efficiency of most commercial units is in the 80% to 95% range. B) Once you've established the two voltages you want to use and a power level, trawl suppliers for units of that size, and data sheets for them. Read the efficiency for that particular model from the data sheet at the power level you want to operate at. If there's no data sheet, don't buy it, or if you do, you get to test it. Pick the one with the highest efficincy that you want to afford. \$\endgroup\$ – Neil_UK Feb 20 at 12:27
5
\$\begingroup\$

FWIW, I played with a 12V distribution around a small flat (apartment). I had a big ATX power supply serving up 12V and setup a sort of ring of thick low voltage cable. I then "tapped" off that ring to provide drops to the places that needed it.

Pros:

  • You can do what you like - it's entirely unregulated so no "code" to follow, and it's touch-safe so it's pretty easy to work on (even when running)
  • Most stuff is actually pretty low current, so volt drops along cables wasn't much of a concern (and a lot of stuff has a power supply inside it to drop from 12V to even lower, so even if you had quite a bit of volt drop, it all still works)
  • You can avoid having any sort of power supply nearby to whatever you're powering

Cons:

  • It's a lot of work to put in, and quite expensive (although I did it with a lot of recycled cable)
  • It's non-standard, so if you move out, you may as well pull it all out because it's unlikely anyone else would take it on
  • Actually, power supplies for small devices are cheap and plentiful and are mostly pretty small, so not hard to put somewhere nearby to whatever you're trying to power

I've concluded it's really not worth it - at least not on a building scale (even a small one). Where it does makes sense is in (say) a media cabinet (or perhaps behind your TV). There, having a single power supply can supply a load of USB devices, stuff that uses a 12V "wall wart" or brick power supply, and if you have the outputs, 18V stuff too. Even there though, you're not doing much that you couldn't do relatively easily by more traditional means, but it's quite fun.

Elsewhere, I'm told the datacentre industry is looking at low-voltage cabinets (along the same lines). The idea being that pretty much all computer and network gear uses 5 and 12V, so having a couple of big power supplies in the cabinet, you can actually supply the whole lot - thus saving dozens of mains power supplies in the cabinet. In a datacentre that helps because that probably less cost and heat, at the possible expense of ease of deployment and choice of vendor.

Some Maths

Back to your actual question... You need to decide what voltage you're going to use and what you'll accept at the socket. Let's say you pick on 12V. You also need to decide what voltage range you'll accept at the socket.

To throw some maths onto this... this 42A cable (the biggest they sell) has a conductor cross section of 4.5 mm2. This calculator suggests that it will have a resistance of 3.733 milli-ohms per metre, so over 30 m, that's 112 milli-ohms (0.112 ohms). I=P/V, so 100W at 12V is 8.33A. V=IR, so the volts drop at that current over 30m of that cable will be 8.33 x 0.112 = 0.932V.

That means if you decide that the open-circuit socket voltage has to be 12V and no more, then at 100W, you're only actually going to be getting about 11V (even less if you use cheaper, thinner cable). You may then want to compensate, perhaps by making the socket voltage 12.5V, and thus only drop to 11.5V at full load. That also means you're going to have to make sure every device you might ever plug into any socket can handle 12.5V rather than 12V. You should also consider the heat generated by the cable - don't stuff it into tight, insulated spaces or else the voltage drop will be even higher.

Oh, and by the way, that one run of 30m to the socket will cost you £85 in cable. For that money, you could (almost) have an electrician come to your house and add in a new mains socket in that location and plug in a 12V wall wart for you ;-)

| improve this answer | |
\$\endgroup\$
  • \$\begingroup\$ It's not entirely unregulated, but it's pretty lightly regulated. Now, that sourcing problem you're having is because you're unaware of the mains wiring parts bin. They can get you 4-4 Al URD for $2/metre and it has only 1.36 milliohms per metre. \$\endgroup\$ – Harper - Reinstate Monica Feb 20 at 20:30
4
\$\begingroup\$

I've thought about this for years. I ended up at why not just use Power over Ethernet? Whether or not you use the network. Cat 5e cable is pretty cheap.

| improve this answer | |
\$\endgroup\$
  • \$\begingroup\$ I wish we saw a lot more POE devices on the market. \$\endgroup\$ – chicks Feb 20 at 18:49
  • \$\begingroup\$ This is an excellent idea - I run POE out to an outbuilding where I run a POE-enabled raspberry pi for backups (not exactly off-site but at least they're out of the home building) There are products like the Adafruit splitters that produce 5 or 12 volts from a POE source. Downside, POE switches aren't that common and do cost a bit more. But when you have one, everything can be POE-powered for small extra costs. \$\endgroup\$ – Criggie Mar 30 at 22:16
1
\$\begingroup\$

what would it take to do a DC wiring for an house to work along the traditional AC wiring?

One way to learn about some of the practical implications of this would be to do some research on areas where this is already commonly done. The power systems that you describe are commonly used in boats and RVs. There is a fair amount of information available for people looking to build these systems themselves. More importantly, you can find discussions about the drawbacks and pitfalls to watch out for.

The obvious difference to me is in the amount of space that you are attempting to cover. You comment that every outlet would be within 30m of the source. In my RV, the longest run I have is about 8m from the solar panels to the charge controller, and I had to run 8AWG wiring for that. It seems to me that scaling that up to larger distances would quickly become prohibitive, but I could imagine a design involving multiple batteries scattered around the house, and a higher voltage from power source to batteries could be devised. Basically a scaled down version of how the AC power grid uses high voltage power transfers for long runs with transformers at the destinations. But then you have to wonder whether all that complexity actually gains anything.

| improve this answer | |
\$\endgroup\$
1
\$\begingroup\$

Even if you buy enough copper to move the power you are trying to move, you have a bigger headache - safety. DC arcs are i credibly hard to extinguish compared to AC. A switch rated for 15A AC may well be rated for only 2A DC. Look into the physics and safety aspects before you burn your house down. ...and I doubt your insurance company will back you if you aren't using fully rated UL components.

| improve this answer | |
\$\endgroup\$
  • \$\begingroup\$ That's why you need to use rated equipment (UL not RU) and overcurrent protection. Square D "QO" panels are perfectly common (albeit upscale) service panels sold in Home Depot etc. but they are cross-listed for DC power. \$\endgroup\$ – Harper - Reinstate Monica Feb 20 at 20:36

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.