A friend of mine did something similar. He put in a small solar array with only 3 panels, an inverter/charger, and 400 Ah of 12-volt batteries. This DC then feeds into a breaker panel and goes to lights in all the rooms. He uses LED spotlights, the little bulbs that plug into wires stretched across the ceiling. He said he just ran 10-gauge Romex wires through the attic, down to DC-rated switches and then to the ends of the cables.

He still has all the original 110 wiring and fixtures in the house, but this battery-backed solar system runs all his lights. It can easily go 2 days without sun before he needs to use the regular lights. Another benefit is that DC LED bulbs are by definition flicker-free, and also low consumption (4-6 watts each) but bright. He calculated he can light each room better than before with 6-8 bulbs, so 40-50 watts, thus only 4-5 Amps (@ 12 VDC) to each room.

All the big appliances in the house still run off of of AC, but he does have a 110 inverter so that he could run a critical appliance such as a fridge during an extended outage.

So this little system replaces 1-2 kWh of incandescent lighting per day (CFC or LEDs on 110 would be less, but not flicker-free), while providing lights during outages. He says it's not cost-effective (the usage displaced by solar and battery usage doesn't cover the cost of the system), but the backup powers is worth it to him.

I am actually thinking of something similar. Instead of 12v DC I probably would go for 48v or maybe even higher. With 12v your cables soon start to become too thick to be practical if you want to run serious things like garage doors / window shades etc.

My friend said he picked 12V because that's what the easily-available LED bulbs are rated at, and he just wanted a simple lights-only system.

I guess you could pick whatever voltage you want for your loads and cable runs, but you need to have the appliances you want available (or easily DIY) in that voltage. It would probably be possible to convert down the voltage once, but any more than that and you're introducing loss again.

120VAC LED lamps will just rectify that 120V to DC and then use a power supply to drop it to around 30V.

In other words, a 120VAC LED should work just fine on 48VDC with the same output.

It’s also explains why they don’t dim.

(Obviously avoid the dimmable ones in this application!)

On the other hand, you could connect 12V lights in series to get them the desired voltage.

That’s not always a good idea.

It is putting more current through each light.

And particularly problematic if each bulb has parallel LEDs and they start burning out: then more and more current is put through the remaining LEDs.

> Another benefit is that DC LED bulbs are by definition flicker-free, and also low consumption (4-6 watts each) but bright.

You still need PWM to control brightness, no? I'd you're using resistors efficiency goes down a lot.

Sure, but if starting with DC you can pulse on the order of hundreds if not thousands of Hz instead of 50/60Hz, which shouldn’t be noticeable.

Our Philips brand 120v AC LED lighting runs about 10watts per, is flicker-free. It's not 6W but I think that is largely due to the higher light output at max brightness (~800 lumen). Also has the advantage of using existing wiring. Integration with the voice assistant software and color modes has been a nice added benefit (ignoring privacy stuff).

12 volt wiring in homes isn't terribly efficient as once you get past 6-10 feet you need really expensive (heavy, thick) wires to not have much voltage drop. Most new boats use 24 or 48v for long haul and then local step down to 12v where needed (kitchen, living, sleeping areas etc).

You can find 120V LED lighting that is flicker-free, eg. anything sold under the Philips brand (in my experience).

Power losses in wiring are an issue. Many of your DC devices also happen to be low-voltage. (For example, USB devices.)

If you transmit the same number of watts at a higher voltage (lower current) rather than at a lower voltage (higher current) through a wire, you will lose less power.

So for moving the power around the house at distances more than around 10 feet (~3 meters), it's more energy-efficient to convert it near the point of use. And if you're already going to convert it, it's probably not worth the expense and complication of running two separate wiring systems.

The USB power supply is probably a 5-volt switchmode power supply which doesn't care whether its input is 80 volts or 300 volts, or ac or dc. You could very reasonably run 100 volts dc or 200 volts dc to things that use switchmode power supplies like that.

A normal worldwide USB power supply accepts 86-240v AC, so you can usually supply a DC voltage on the same terminals of the same voltage with no issue (convert from AC peak to peak to DC)

It has less to do with ac vs dc and more to do with the voltage and distance from source to load. If you wanted a dc bus inside a house you would need to have voltages in the same range as ac power 120/240V in North America. At this point you have the same challenges of how do you reduce the voltage to something low and easy to work with for an end device like a light bulb or cell phone. Stepping down voltage is easy to do with AC power but difficult for DC which is why I don’t think it will ever be practical.

48v dc bus and buck converters work just fine. generally power supplies that work off AC rectify and go through the same business - except they need larger inductors to sort out the zero crossing.

i think the difference you're looking for is in volume of copper.

Maybe we should instead invent a domestic 3 phase distribution system. No more zero crossings requiring energy storage, just flat constant power. It would require a new, small four conductor plug to be designed. Total volume of copper would be a less for the same available power, since the three wires deliver twice the power as two single phase wires for the same conductor size. There would be more insulation involved so the wiring might grow or might shrink if outlets were held to 1500watts.

The big savings would be in longer lived electronic devices. A sizable portion of digital electronics ultimately fail because the capacitors in their power supplies exceed their designed lifetime. Electrolytic capacitors (ticking time bombs) are cheap and compact for the capacity so they are used for the energy storage required in single phase systems, but if designing a three phase power supply less storage is required and maybe manufacturers wouldn't have to optimize the smaller, capacitors with electrolytics. I've had a 4k TV and a rather expensive musical instrument amp fail for capacitors this year, about $3000 of gear there. Looking back further there are a couple of professional powered sound reinforcement speakers (twice), an iMac G5 (three times). It's probably safe to say over the decades I've had $10k of electronics fail prematurely from electrolytic capacitor failure. These are all items which were not obsolete or worn out. Scale that up from one person to worldwide and it's easily in the hundreds of billions (10⁹) of dollars which could be saved.

The problem is you can't get there from where we are. No one wants to drag more wires to residential houses, no one wants to shell out >$1k for a household phase converter and then rewire their house. You won't convince people to add three phase to a new house because it might be the future. And you won't convince the semiconductor houses to make a line of three phase power supply controller chips just because, well, maybe someone might use it.

We're probably trapped at a suboptimal local maximum.

Edit: In response to atoav's comment I found IEC 60309 and its extra low voltage variant which is a 42mm (1 2.3") 3 phase plug and connector in two compatible sizes good for 1.3kW and 2.6kW. Also, that Germans and Swedes did the engineering first and then built an electrical system and they get 480V 3 phase at 16A to their cooktops and stoves through a plug for a whopping 13kW!

I am not exactly sure what you imagine, but I (in Germany) have two 3 phase outlets (3x 16A @230V) in my flat and this is very common here. We use it for cook tops and ovens in the kitchen typically, and it makes the difference between 15 mins for boiling water (old single phase cook top) and (on my induction cook top) 1 min.

Edit: and as an aside, why we probably don't do that everywhere: it makes for bigger plugs, and keeping safety standards can be a bit more expensive and space intensive with the potential 400V between phases.

I think Germany is one of the few countries which bothers to wire up three phase power to ordinary houses, not sure why.

It's quite common in the Nordic countries as well.

Well that's news to me! I'm jealous. I'll bet you can preheat an electric oven in less than 20 minutes.

I see the successor to the 16A Perilex connectors you are referencing is IEC 60309 which has an "extra low voltage" version with a smaller, 42mm diameter (1 2/3") plug good for 16A or 32A. 16A devices will fit in a 32A socket, but not the other way around. So that seems to be a 1.3kW and 2.6kW solution. Sign me up!

"So that seems to be a 1.3kW and 2.6kW solution"

Germany is 230v so that would be more like 3kw and 6kw.

I'm in the UK (also 230v) and a standard outlet can do 3kw, I think fuseboard fuses are 32amp so I don't know if 6kw is a practical upper bound though (edit: On 2nd thoughts I'm not even sure this is relevant to 3 phase).

Pedantic: Standard outlets are 16A in mainland Europe, so 3.7kW.

Modern single ovens can also be connected to a standard single phase outlet rather than a dedicated circuit. In my kitchen the induction cooktop is on a 16A three phase circuit, and all the other outlets (fridge/freezer, oven, dishwasher and appliances) are on a 16A single phase circuit.

5:11 from 20°C to 220°C

Wouldn't pedantic be 3.68kw?

I've never actually come across anything higher than 3kw, also a 16amp fuse isnt that precise, you could probably get up to 20 amps without it failing. The hotter you run it the quicker it fails though, so if you're regularly running 3.7kw through a 16 amp fuse, you may well find it failing, not through an underlying fault of anything plugged in to it.

>I think fuseboard fuses are 32amp

Most cables are 16A, so the outlets are 16A as well. Also UK uses 240V, not 230V. 32A circuit breakers would make the cables burn when overloaded.

Flex cables might be 16amp, individual sockets are 16 amp though, so if you have more than one of those on a ring main...

"Use a 30amp fuse or 32amp MCB for larger radial circuits, ring circuits up to 100m, cookers and electric showers up to 7200W"

http://simplifydiy.com/electrical/basic-electrics/fuses

normal fuses (in the fusebox) still have to be 16A (B or C class), the main fuse can be anything 32-50A or even tri-phase 50A.

>ring circuits up to 100m, cookers and electric showers up to 7200W

Normally those tend to be 3phases.

Of course it's possible to get over 16A on a single phase and a circuit breaker but that's quite a non-standard option. The issue is that if the fuse/circuit breaker is rated for 32A in a standard application (i.e. an outlet) the cables are likely to burn 1st.

I think we might be getting our wires crossed :)

I'm talking about the UK here, which I understood to be similar to the rest of Europe(?).

That's a UK website and 3 phase isn't common here. I've just double checked my fuse box, and I have a 32 amp fuse for my socket ring main.

Don't forget for a ring main the current is spread over 2 wires, so if theres 32amps power draw, theres only 16amp on each side of the ring.

UK here:

We don't have three phase (well its not common) but we can heat an oven up to 220c(420f) in 7 minutes. I think its a 3kw oven. its on a 32 amp (isolated) breaker.

I've never ever seen it in the UK, but I did live in a house with a flat on a separate meter, that as it turned out was on a separate phase. I only discovered this because there was a fault on the line for only one phase, and only after checking and changing every fuse (wire), and unplugging every electrical item. Grr.

Had something to toast so I tested mine too (3x 16A @230V). Took 4:52 to heat up to 220°C

I'm in Romania and it heats to 180C in about 3 and a half minutes.

Do we need to invent anything? My understanding is that most of the grid in the US operates on 3 phase, and some buildings (including residential) already have a 3 phase supply to operate AC and industrial applications.

All we are missing is for it to be standard to wire the 3 phase distribution throughout buildings to every outlet (preferably in a backwards compatible manner; in theory converting 3-phase to single phase is just a matter of ignoring a wire. I would assume that a) the grid can handle people doing this and b) doing this with out current 3-phase standard would result in our current single phase standard,

I hate those 3 phase residential buildings, because they usually deliver 208V to ovens and dryers instead of 240V.

While the appliances are probably approved fo 208V, I doubt they’re drawing more peak current to make up for it.

Wouldn't it be easier to mandate separated replaceable power supplies vs changing the electricity standard in buildings world wide?

Or make power supplies some sort of standard, so you can use your own expensive capacitor power supply instead? Or just shop for things that have separated power supplies?

Dc to dc voltage voltage conversion is easy kind of, you simply need to convert it to AC and back.

So as its going to become AC at some point anyway, why not use the existing wiring.

For the devices that run off of low voltage DC today, lots of other commenters make good points about needing to regulate the voltage anyway, and I'm inclined to agree. Those devices generally use very little power anyway, so it probably isn't worth running a second set of wires because copper is expensive.

On the other hand, the biggest power users in a house (heater, air conditioner, dryer, stove, refrigerator) seem like they are a different story. In a future home with solar power and a high voltage DC battery pack, I think there is opportunity to make appliances that operate from the raw battery voltage, which is in the 300-500v DC range. That way, there isn't any conversion required at all, and those applications should be able to handle slight variations in voltage as the battery discharges.

Any high voltage DC distribution system would require every load to have a single run, with intelligent connectors paired with remote switches to shut off the circuit at detection of any detected alteration of the circuit (eg. Ground fault, capacitance change, etc.)

High voltage DC is spectacularly more dangerous than high voltage AC.

Could you give more details of this danger?

An arc on AC is more likely to be extinguished since the current drops to zero many times every second (so eg, on 50 Hz AC, an arc is not likely to last more than 1/100 of a second). DC switches and circuit breakers are designed with greater concern for arcs.

I've seen (owned a house having) dual-fuel natural gas + electric light fixtures, mounted to walls, gas pipes to every room right alongside the (knob & tube) electrical wires. So we've done weirder shit in the past to deliver two types of power throughout houses.

And no, I never tried lighting one. Hopefully none of them had had an active gas supply in decades.

I wish more devices would run on PoE. It's DC, it's easy to install, it carries power and data. There are some (but few) PoE light fixtures, but not as many as I'd like.

That does make sense, especially for IoT devices (literally anything low power). However, I haven't seen PoE used outside of business use. The problem is that Wifi is becoming ever more prevalent (on my network I have 12 devices using wifi, and only my desktop using an ethernet cable - ironically the only device that isn't suitable for PoE).

I use PoE UniFi gear, because running a single Ethernet cable to power e.g. a wifi AP gives me a lot more flexibility in placing it cleanly than having to worry about power; that's for residential use.

Given the degree of fire risk in houses that's associated with 230V/10A wiring and devices, I do wonder how much safer it would be to have fewer AC/DC transformers scattered around the house with DC wiring.

If you're using DC to transmit the same amount of power, you'll probably be using similar voltages (otherwise you're going to need thicker cables). I suspect DC is probably a greater fire risk in this case due to arcing. An arc is less likely to persist on AC since the potential keeps passing through 0 V (hence DC switches are designed to have a bigger air gap).

If it's just a matter of providing alternative circuitry for low-power devices, presumably you can already improve safety by just using lower-current circuit breakers.

> If you're using DC to transmit the same amount of power, you'll probably be using similar voltages (otherwise you're going to need thicker cables).

A good point - I guess I'm mostly thinking about the volume of things in the house that don't really need access to even a fraction of the 2.4kW available; having moved from incandescent bulbs to LEDs, I've got a bunch of unnecessary power supply and per-bulb transformers, for example.

This is a thing!

The ISS runs 120v DC for this reason...higher efficiency and much less complex since you don’t need inverters and and power supplies.

My burning man camp of 65 people also only has a DC grid that runs on 24v

30kwh daily solar harvest

It’s a thing of beauty :)

Converting DC voltage is unlike AC with transformers not easy or practical.

Very few AC-powered devices even use traditional transformers anymore. Mostly they have some kind of switch-mode power supply which immediately rectifies the AC to DC. (Better and higher-power supplies have complicated active power factor correction circuits to do this.) Then they convert that high-voltage DC to low-voltage DC.

> I wonder if there's any future to running both an AC and a DC bus through the house

I used to have a house built around 1950 which had combination 117/220AC and 25VDC wired throughout the house. The electrical box was sectioned and the 25V had a large transformer. The house had low voltage DC ceiling fixtures that were tricky to source, and everything in the house that was switch controlled was on relays that allowed multi-point control, smart timers, and various other functions. It was a chore to get replacement parts but fortunately the attic had gangways and such so replacing blown relays was not horrific once you had the replacement parts. It was perfect for upgrading to smart control which I did for about half the house before selling it. Here's the thing. Tons of houses back then were wired this way. It was the wiring of the future and they explicitly realized in the 50s that someday they would be able to have computers in the home, perhaps as small as a single room, control everything. So instead of an electrical closet there was an electrical room with enough space for a refrigerator sized computer should such a thing ever come to market. 1950s.

Conk has a good response but you also have to realize that it's still too expensive to setup a battery as well. So, since the grid will be needed, you'll need conversion to be able to handle the fluctuating demand your home requires. And don't install new devices, as that can change your equation but connecting back to the grid allows much needed simplicity.

Other companies are working on this (energy sovereignty) issue, my guess it (stand alone affordable systems) will be solved within the next twenty years and be affordable. DC use might be apart of it but I'm doubtful since there is a level of complexity that each device is using different amps/volts/etc that will take new products to solve and the independent grid system still takes a lot of capital.

New (>2005) AC/DC converters (switching PSU), have such a high efficiency than I would think it's not worth the hassle

Pretty much all switching power supplies would run off DC as well - from 110V - 350V DC.

Is there any safe way to test this for a server PSU? As in "unlikely to blow the PSU", potentially requiring slight pre-checking by looking at the circuit?

Yes, a noop and declaring the "test" successful.

I would be very surprised if you could find any non-ancient (and I really mean ancient, like, pre-PC, probably even pre-home computers) computer PSU that you couldn't feed with DC. What voltages will work will vary, but as long as you stay below the peak voltage corresponding to the effective voltage printed on the label, you probably won't damage anything.

If you really want to test it, you could measure the resistance at the mains input, if that gives you a non-infinite value that stays at that value (rather than increasing towards infinity), you are dealing with a mains transformer. But you really won't find that in a computer PSU.

~240V fully rectified to DC is 340V. Since AC varies it's safe to say virtually any PSU that's not exclusively Japan or North America will take 350V DC.

Thanks!

It's a hot-swappable 300-1897 (aka X6328A aka Type A217) 1050W Sun PSU. I'm primarily worried about the active PFC I expect in it.

Is there a way to test such a PSU without risking the entire server? (save for taking everything out that isn't needed to post into BIOS)

Well, you probably can somehow power it up without a server connected to it?

Active PFC tends to be simply a boost converter at the input stage that boosts the (rectified but still unfiltered) input voltage to (DC) 350 V or more, modulating the current to follow the mains voltage and should generally have no problem with DC input, especially so with wide-range inputs that have to be able to deal with different voltage levels anyway.

>modulating the current to follow the mains voltage and should generally have no problem with DC input

I'm particularily worried about confusing that following/regulating circuity. If the 50/60Hz it senses are caused by screen refresh feeding back somehow, it might end up behaving "weird".

But I guess I'll just try, and hope it won't break anything.

If it weights 5KG+ (12pounds), likely it is not a switching power supply as it has a bulky transformer (I can't imagine a server PSU not being a switching one). Other than that laptops, phone chargers, etc. would work off DC. Most of them use a small transformer and high frequency (above hearing range)to drop the voltage.

The transition is something like: 90-240V AC (50-60Hz) -> DC (1.414*AC) -> AC high frequency ->(transformer) AC low voltage -> DC 12V

Very cheap electronics that use capacitor+resistor voltage drop need AC.

---

To answer the question directly, use a small fuse 1-3A and run the PSU idle. Like advised by sibling, measuring the resistance would be sufficient as well.

Looked around a bit, and the link below says DC has a lot of problems. One is that you's still have to convert your home DC voltage to whatever voltage each DC device uses.

https://www.quora.com/Why-do-we-use-AC-but-not-DC-in-househo...

Realistically 25-30A with thick copper wire are doable. 20V is a good ground but even then it could power a TV + couple laptops, perhaps lights too. That would be a solution to have dedicate DC lane in-house but the conversion to low voltage DC must have in-house (office buildings would require tons of copper). Galvanic corrosion might be an issue to a certain degree.

Carrying high voltage DC, comes with its own issues - likely doable though. Many household devices would run perfectly fine on high voltage DC (instead of AC).

Honestly, just use aluminium cables. Those only need proper crimping, which isn't too expensive.

The cost for the cables themselves would be lower than copper AC, even.

A number of off grid cabin setups use DC distribution within the house and purely DC appliances (fridge, LED lighting, etc). Usually in the context of places that are very far from a grid, with wood heat, or tiny home size off grid setups.

Think of it something like planning the DC loads on a 55 ft sailboat.

People who do this will usually have one small 300 to 500W sized DC to ac pure sinewave inverter for laptops and chargers of small consumer electronics.

It’s happening even more than you think.

Your fridge is probably running “variable frequency drive”, so instead of being off and on at full blast, it just runs the compressor at 20hz or 30hz or whatever it needs continuously, and dialing that up or down depending on demand.

Your air conditioner too.

For bigger customers, this saves them a fortune in demand charges.

For others, it prolongs life by vastly reducing stop/starts.

Most things that use DC use pretty low voltage, like 5 or 12 volts or so. High voltage allows you to transfer energy much more efficiently. So, for low power devices it might make sense to run a low voltage line but you wouldn't be able to run high-power devices out of it unless you wire your house with really thick wiring.

Also a layperson. How many different specs are running around the house, and what sorts of electrical variations can they withstand? I think you'd need to have a house-wide AC-DC converter to feed in energy when your solar supply drops down. Or a battery storage system probably makes more sense.

It's actually a good intuition. See: https://solar.lowtechmagazine.com/2016/04/slow-electricity-t...

tldr: it would make sense for an office building but actually not that much in a residential one.

Many of your AC devices are actually DC after transformer.

You could do power over ethernet to get connectivity too.