"Because many modern devices operate internally on direct current (DC)"
But notably the high power consumers like electrical stove, air conditioning, washing machines etc. do not. And it's not that with DC throughout there are no conversions. Solar cells' output varies greatly, hence it needs to be DC/DC converted to a fix voltage to be of any use. The same, albeit to a much lesser extend, applies to batteries. Further, there is no one single voltage, AC or DC, which is optimal for all uses. Even for devices which receive DC voltage, e.g. laptops, there are internal DC/DC converters.
Removing the AC/DC conversion seems hardly worth the trouble. High efficiency AC/DC converters can be constructed. It's rather a question on how much one is willing to spend on such.
Yes, you cannot reasonably run these high-power devices on low-voltage DC (12-48V), this is also said in the article. But the reason is not that these devices cannot run on DC in principle, the problem is cable loss. For running devices which need >1kW on low-voltage DC, we are talking at least 20-80A, depending on the voltage. So you need thick cables, which are expensive and hard to deal with (amount of space needed, securing connections, avoiding bending, everything becomes problematic with these cables). Any little mistake done with cabling is immediately a fire hazard.
So in the end, the article says: say good bye to your washing machine, dish washer, electric kettle, electric stove, etc. I mean, the site is called "low tech magazine" for a reason...
I don't see any future in 12V systems, even in cars (with rather short cable lengths) there are discussions to switch to 48V, like Tesla is doing it in the Cybertruck.
I was thinking more like 200-400V DC systems, that could power most devices, but comes with a lot of challenges.
Yes, but this is not what this article is about, which is to avoid the loss of efficiency by converting low-voltage DC solar power to AC through an inverter, only to convert it back to low-voltage DC again for most devices. If we are talking high-voltage DC, you'd just replace the old DC-AC conversion with low/high-voltage DC-DC conversion, which I don't think is much more efficient. Also, I wouldn't want to live in a house which runs on high-voltage DC, purely because I don't want to instantly die if I accidentally touch a life wire.
I think we need to talk about the definition of "low voltage". In my understanding it's everything below 1000V. Solar power systems often provide a voltage up to 1000V, by serially chaining modules. Running them in parallel with ~50V would require quite massive cables.
Edit: About the danger of DC wires: There are technical solutions to mitigate risks. They may be too expensive though. I think we can reuse some of the strategies electric cars use. It's the only end-user application I know that uses DC powered cables with up to 900V (Lucid Air).
I (as someone who designs electric and electronic equipment every day) agree. Low Voltage is a very specific term and typically means <1kV AC and <1.5kV DC in the industry:
https://en.wikipedia.org/wiki/Low_voltage
It’s not a very specific term, at all - and is dependent upon context - per the first sentence, of the wikipedia article, you linked, stating:
> In electrical engineering, low voltage is a relative term, the definition varying by context.
The various standards, mentioned in that article, are in relation to “installations”, and power distribution. The standard in the UK (BS 7671), defines “high voltage”, as >600VAC difference between conductor(s), and earth.
A voltage greater than mains voltage, in any item of domestic electrical equipment, is absolutely “high voltage”. Hence why valve amp power supply terminals, are considered high tension (HT, i.e., high voltage), despite not exceeding the threshold, claimed in wikipedia.
This is a great piece of data, thanks for posting it. Without this comment I would never have realized low voltage referring to ~50 volts or less was primarily a US thing.
Since we’re talking about power distribution in homes, the National Electric Code and local zoning apply. That means low voltage for the purpose of the article will be limited to 0-49v. All solar panels for homes that I’m aware of are under 49v per panel so fair enough for the article.
The panels on my house are 44V each but are connected with 11 in series. That means you can expect to see 484V in an open circuit situation with full sunlight.
Switching losses are less than ohmic losses. A few chips cost less than a lot of copper. The crossover point depends on the cable lengths, cable size, load, and voltage in question.
Let's do a quick example. 16 AWG carrying 12V. 16 AWG is 4 mOhm/ft. Let's say you have a 1200 W load (100 A, 0.12 ohm). Ohmic losses per foot of cable are 3%. So if you have a 15 foot cable it will use as much power as the load.
Absolutely. But with the voltage 10x as high, the current goes down 10x, too, and the power loss goes down, too, while the power stays the same. That's how things run on AC normally.
I don't see how that would be any different from unprotected "high" voltage lines, or preferably protected ones for that matter. In house installations around 500V count as Low Voltage in power engineering, by the way.
That might very well be, but I have touched 240V AC accidentally, and I'm fine. Don't think that would still be the case for 400V DC, whether it counts as "low-voltage" or not.
EDIT: After reading a bit, it seems that the "what is more deadly" discussion regarding AC/DC is much more complicated than I thought, so the above might very well be false...
Any arc source is vastly more destructive with DC. That makes switching and circuit protection more difficult/expensive.
AC helps you because the voltage crosses zero twice every cycle so normal voltages don't tend to be sufficient to sustain much of an arc in air. Every time you flip a switch you cause an arc, but the same DC voltage eats up the contacts more quickly. Even silly things like unplugging a running appliance does way more damage to the receptacle and the plug under a DC load.
200-400V DC distribution systems are dangerous, both for electrocution and fire. Though maybe modern GFCI and arc detection breakers could make them adequately safe. Could someone who knows the state of the art comment?
>I don't see any future in 12V systems, even in cars
why not? it works, it has industry wide support and a huge range of products, and it's safe without excess need for shielding and isolation. it's closer to the actual operating range of most equipment, so it produces , in most cases, less intrinsic need for shifting voltages around, and the wiring that would get substantially smaller is already near size limits for dealing with vibration and harshness.
aside from trying to get closer to the operating voltages of whatever arbitrary energy pack we're specifically talking about , what's the point? if anything it would just wreck any hopes of cross-industry compatibility for a long time for very little manufacturing efficiency gains or energy savings.
tl;dr: the vast majority of car 12v is lights and logic, and they're closer to 12v than 200v in the vast majority of cases.
You're massively underestimating just how much wiring is in a modern car. The car industry has been trying to move to 48V for decades to save money on wiring and Tesla just succeeded with Cybertruck. High consumption devices like electric power steering, electric A/C, and active anti-roll bars all require huge amperage to run at 12V and are only becoming more common on cars.
Car makers project big savings in moving to 48V and it's going to become a lot more common.
I guess that really depends on how you define “consume”. For example do you do your laundry by hand or do you use a machine that lives outside your van?
In-sync AC grids are an engineering marvel; high-voltage AC lines are comparatively lossy due to the skin effect among other things; a frequency of 50 or 60Hz yields what are, by modern standards, comically huge step-down transformers; etc. It's worth considering whether the relative ease of running AC motors makes all the difficulties of AC supply worth it.
(It's worth remembering that the "modern", i.e. 1930s, three-phase power grid is built primarily to power motors on factory floors. I don't know if these are the highest load these days, but wouldn't be surprised if not.)
> It's worth remembering that the "modern", i.e. 1930s, three-phase power grid is built primarily to power motors on factory floors.
Small nitpick, I don't believe 3 phase power was picked primarily because it was convenient for powering 3 phase motors - it was picked because it's convenient for generating and transmission. There is no need for a neutral conductor in a balanced AC transmission system for example and it can very efficiently be converted to high voltage and stepped down again. It also results in modest power pulses though the generator compared to single phase etc.
Feels like it's both to me. 3 phase, 3 wires is optimal for the amount of conductor material needed to distribute it. But it's also optimal for motors for a few reasons. Easy to reverse direction of the motor, self starting without a capacitor, and higher overall power and efficiency.
Motors consume around half of the world’s electricity, and nearly all of the power consumed by motors connected to the grid is consumed by three-phase A/C induction motors. Motors aren’t just in factory equipment, HVAC systems have all sorts of motor driven equipment.
VFDs have made A/C motors far more efficient than one driven by an across-the-line starter.
> a frequency of 50 or 60Hz yields what are, by modern standards, comically huge step-down transformers
Should switch to 400Hz like aircraft do. Not just comically huge transformers, but also motor/generator magnets.
> A special generator was designed to create an output of 400 Hz. This allowed a motor which was the size of a watermelon to be replaced by one the size of a one-pound coffee can which could do the same work.
> The saving of weight allowed increased cargo capacity and decreased fuel consumption. Power at 400 Hz for aviation was a success and became the standard of modern AC-powered aircraft.
> Airports all around the world standardized on the same power system. This included the physical plug and cable as well as the 400 Hz power so that aircraft from anywhere in the world could land and be serviced wherever they landed.
The aviation power system of 400 Hz became one of the first worldwide-adopted standards.
All of the DC/DC boost converters that I am aware of use inductors to transform low voltage to high voltage. It's technically pulse-width-modulated DC instead of AC but there isn't a lot of difference between the operation of a boost converter and the operation of a high frequency inverter.
Semiconductor factories are not far down the list either. Texas will probably see quite a few come online within the next decade.
When I was working at SAS in ATX circa 2013, the 2 manufacturing lines were responsible for something like 10% of the city's average power consumption. I can't imagine the bleeding edge will have improved much - EUV light sources aren't exactly Energy Star compliant.
So all efficient (switching) voltage conversion involves AC. Your stock plug-in DC power supply is actually a AC->DC->AC->DC device yet it's trivial to get above 97% and miniscule quiescent (minimum) load. High-frequency AC isn't a problem for efficiency.
What's NOT easy is efficiently making 60 Hz AC at low quiescent power. Most inverters always burn 2-5% of maximum power with no load, while 1% would be considered far excessive if the output was DC (maybe 0.1% would be generally okay?). Part of this is lack of regulation; consumer devices have low-quiescent power supplies originally because of government mandates and, now, economies of scale. The government mandate didn't apply to inverters so there isn't an initial push to shift to low-quiescent topologies and prevent the first movers from being undercut.
So just for that reason, lower-voltage DC distribution can overall be more efficient if you are off-grid, even if some other portions of the system have increased losses. 120Vdc distribution would be superior from an efficiency standpoint in every way, but you're going to spend a lot on switches and protection equipment.
I'm not even sure how to reply to how wrong you are. You could only argue that with flyback topology (I'd still say you are wrong) but try a standard full-bridge and tell me the current isn't alternating. Then check out some ZCS variants and tell me which of those current waveforms don't look like Alternating Current.
Since all such power supplies prominently feature transformers, they can't operate on DC alone. But the frequency they use is high (a few kHz), the transformers are small, unlike the huge beasts in traditional 50/60 Hz power supplies.
AC can mean "sinusoidal" or it can more generally mean "any non-constant voltage [or current]." It was clear from the context that the parent meant the latter.
Most of the high power consumers could run on DC power with only slight modifications (in design, not aftermarket).
Heating elements can run on DC more or less unmodified. Also electric motors can run on DC efficiently (see electric cars). Back in the days 3-phase power was needed to run strong electric motors (very common in Europe, but i think it exists also in the US for factories and businesses). But there are many easy solutions nowadays to just run them on DC.
> Also electric motors can run on DC efficiently (see electric cars).
It's quite likely that the motors on electric cars (and other efficient electric motors) are actually using an inverter to turn the DC into multi-phase AC and power an AC motor; this allows varying both the voltage and frequency of the AC fed into the motor.
Bingo. The frequency is proportional to motor/vehicle speed. The inverters are at least 98 percent efficient at high power, and a good EV motor is in the 90s at high power.
High voltage DC is hard to switch with mechanical switches, and you have to convert it back to AC to run motors(Unless you're willing to accept brushed motors), so it wouldn't be trivial.
A lotta motors are moving to variable drive to have a continuous duty cycle and do that by internally converting whatever is coming in into AC. They should happily accept DC since they're just rectifying the AC anyway.
My fridge is "inverter drive" and just runs continuously at a low duty cycle all day. Most heat pumps are like this too to dial up and down instead of strictly being on or off.
(Unsure what voltage the motors are running at though, so maybe AC-in is preferred to step it down before rectifying and chopping)
That's pretty neat! Nobody I know has many appliances from the last 20yrs so I'm a bit out of the loop on those...
I guess all we'd really need is reliable switches and breakers that could be made affordably, and connectors that didn't create a huge arc every time someone unplugged a running space heater?
Still a pretty big challenge, the lack of arcing seems like a really big advantage.
> Heating elements can run on DC more or less unmodified.
..but to get the same power (heat) output you need more current in a low voltage DC system. And power loss in the supply cables scales with the square of current, so you'll need thick (expensive) wiring in the wall or you'll risk the walls becoming an equally effective heating element.
Electric motors that run on DC are either brushed or have extra driver/conversion circuitry to pulse their coils in the right order. Both of these are not perfectly efficient.
Increasingly, heat pump devices, like fridges, air conditioners, heat pumps, water heaters, are inverter drive, so they just rectify all the AC into DC and start chopping it to get to their desired frequency.
(Also dumb that electric codes require dedicated circuits for many of these things, despite them not having the surge like they used to)
> But notably the high power consumers like electrical stove, air conditioning, washing machines etc. do not.
An electric resistance stove (radiant or exposed coil) would work just fine on DC, although, if it’s controlled using a TRIAC, that would need to change.
An induction stove uses rather high frequency AC, and that could be generated just fine from a DC supply.
Modern air conditioners use variable frequency drives, and those generally work by first converting the AC supply to DC.
I imagine that many modern washing machines also use some sort of variable frequency drive or DC-powered motor.
> Solar cells' output varies greatly, hence it needs to be DC/DC converted to a fix voltage to be of any use.
This is simply wrong. It is true that modern MPPT solar regulators use PWM to extract the maximum power, but for years, a simple on/off regulator was used to prevent overcharging.
The reality is that a solar panel is essentially a "Constant Current" supply, so the panel puts out the same current over a wide range of battery voltages. It is the battery which puts out a Constant Voltage at whatever current the load requires.
With more and more DC generation and consumption on the network, I think the AC mains is becoming very noisy, i.e. there are spikes, which we are not so used to handling.
That is true. There is a whole field of power factor correction and filtering that tries to reduce the noise injected into the mains (hello, EMC regulations) but some devices do that better than others.
It is easier to turn AC at the wrong voltage into DC than it is to turn DC at the wrong voltage into the right voltage. Particularly if you care about efficiency of the device.
> Dutch researchers managed to reduce total cable length in a house down from 40 metres to 12 metres. They did this by moving the kitchen and the living room (where most electricity is used) to the first floor, just below the roof (where the solar panels are), while moving the bedrooms to the ground floor. They also clustered most appliances in the central part of the building, right below the solar panels
That's not great, since entertaining guests it is preferable to have these rooms on the ground floor, and the bedrooms which can be private on the upstairs floor.
While upstairs has a better view, more light coming in through the windows, and heat rises up (so the bedrooms can be kept comfortable cool). I guess it depends on the house and it's surroundings:-)
What's happening in their living room to use so much electricity? I figured TVs, computers and lighting use a fraction of what they used to use.
My mom is in a typical suburb tract housing development, and the kitchen is right above the service entrance and the HVAC/mechanical room is right at it.
Great they're applying the same logic to solar-first systems, but it's not a new concept at all.
a combination of a 60" TV and an xbox series X or a PS5 is probably the most power consuming thing in a living room, an xbox of any type when actively rendering a game puts out a decent amount of waste heat. Nowhere near as much as a more costly desktop gaming PC, but still a fair amount.
It’s actually not that crazy. Many multi story houses have their plumbing and hvac built this way to save on copper pipe and get you hot water faster. All the bathrooms etc backing a central plumbing chase.
Something I didn’t find in the article is that in AC systems sparks are self-extinguishing because of the zero-crossing.
There are a few domestic circuit breaker teardowns on youtube. I suggest watching one and then asking the question: how can we break a high current _DC_ surge?
Not mentioned so far: AC->DC adapters also provide galvanic isolation. In many power supplies, that electric isolation comes 'for free', as you don't want device(s) connected to high voltage wiring anyway. AC vs. DC as input isn't that big a deal really.
In situations where power source (solar!), storage and most consuming devices are low-voltage DC, just use a DC based setup if more practical.
I'm on a boat & most everything electric here is low voltage DC. But I do have a 12V->230V AC converter when needed.
In short: AC & low voltage DC can live happily side-by-side. The DC vs. AC debate is kind of a moot point these days.
Few devices other than switching power supplies are designed to work over a wide range of DC input voltages. There's an electric pump which will run slowly on an low input voltage and faster at a higher voltage.[1] Those are driven from windmills and solar. It's not cheap. It has to be self-protecting against too high and too low voltages.
This is part of the reason why a fully compliant USB-C cable is so expensive and stiff. It should carry thick power lines, and also a number of high-speed serial lines, and some electronics to configure and drive all that on either end.
If DC in houses became popular, it seems like they'd follow the same model as RV's -- low power devices like lighting, water pump, maybe refrigerator all run on 12VDC, high power devices like air conditioning, microwave, etc run on 120VAC (sometimes inverted from battery).
That limits the high gauge wire to the short run from battery to inverter, otherwise you'd need very expensive and heavy large gauge cable running to those high power devices.
I think that, if DC appliances as used in RVs became genuinely popular, they would move to reasonable voltages. 12VDC is absurd. 48VDC would be much, much better.
12V is reasonable for lighting on a small scale. It’s not really fit for purpose for lighting on a house or garden scale. (It’s not even really awesome on a small scale - nice LEDs chips and board vary between 4 and 50V or so and should be driven at constant current. With a 12V supply, this requires a buck-boost converter. With a 48V supply, most bases are covered with a simpler buck converter.)
And 12V in a vehicle makes little sense if the vehicle’s battery operates at 24V or 48V — it requires DC/DC conversion for no real benefit.
Less than what? If you're converting DC->AC->DC to step it up, why not just convert DC->AC next to the battery, then AC->DC or just use the AC directly like you would in a motor at the point of use?
Are you talking about skin effect losses? This isn't a long distance high current grid transmission system, for the lengths and sizes of conductors in a home (or RV), I think it's safe to disregard skin effect.
Just sitting here speculating, not even a web search yet...
In the USA we have both ground and neutral wires at every outlet along with the hot (120VAC nominal) wire.
Why couldn't we raise the ground wire to +12VDC (or whatever) with regard to the neutral? The ground would have a very sensitive detector so that if electrical conditions exceeded some preset threshold it would collapse to zero nearly instantly, allowing the circuit protection to function as expected.
You could make a plug that only has a neutral and ground pin to tap into the low voltage side, or use all three pins to provide low power standby mode.
Now I'll do a search to find out that it is impossible because of a, b or c reasons or that it has already been patented somewhere and abandoned. Yay Internet...
The GND (or Earth/GND) is there for safety and equipotential bonding. It’s not meant to carry current and running current through it will trigger your RCD, because that’s how a ground fault is detected. And detecting that is important so when you trip with your hairdryer into the bathtub you won’t kill yourself.
Even if plug is polarized (plug goes in only in 1 orientation), you can't assume any of the high voltage wires to be touch-safe. Wiring mistakes (swaps) can & do happen.
In-house AC systems are designed to remain safe when that occurs. But it prevents using any of those wires in ways like what you suggested.
One problem is that many older houses don't have grounding. Half my outlets are ungrounded and I use GFCI to get three-prong plugs. Or the grounding is done through pipes and conduit which aren't suitable for running current.
The standards for grounding make them unsuitable for running current. The ground wire in box are frequently bare. Metal boxes are grounded. The center screw is grounded. If you run current over the ground, those are live.
You have invented a way to electrocute lots of people.
That's why this could work - the panel is the logical place to insert the DC supply. If ground and neutral were bonded in multiple places then it would just be a DC short.
They are bonded there for safety reasons and that bond needs to be low impedance.
The reason is that you don’t want to allow a fault which puts a line voltage (120VAC) onto the chassis of a piece of equipment waiting for a human to come along and complete the circuit to ground. So, you use the protective earth to connect to the chassis, such that if a power line ever did contact it, it would be a dead short to ground, which would trip the OCPD (over-current protection device, typically a circuit breaker), clearing the fault.
If you replace this safety mechanism with something that’s high-enough impedance to maintain a 12VDC differential, you’ve eliminated this important safety mechanism in your AC distribution system.
I think this would also make another possible fault more dangerous: a break in neutral somewhere between an AC device and the panel.
If there was also a DC device on the same side of the break as the AC device, then you'd have a possible return path for the AC current that goes through the DC device, the ground wire, and the DC power supply.
In addition to that, would it also cause problems with GFCI?
The 12 VDC return current would be on the neutral. I'd guess that a steady DC current on the neutral wouldn't cause problems, but whenever a DC device is switched on or off, or you use a DC device that has a fluctuating instantaneous power need, that should induce current in the GFCIs sensing coil that won't be canceled out by an opposite current induced from the load wire.
I don't know if that mismatch would be long enough or large enough to trip the GFCI.
Currently the ideal balance of price, performance and availability of off the shelf components is 48v dc power.
There is a telecom industry standard (perhaps only loosely standardized) for DC converters which are called "isolated DC-DC converters bricks."[2]. These are reasonably priced and reasonably high efficiency DC converters that fit into a set of standard form factors, e.g. 1/4 brick, 1/2 brick, etc. They are small sealed modules and a variety of outputs ratings are available for < $100 with efficiency > 85% and many newer examples are > 90% efficient.
The setup I've settled on is the following, based on 48v nominal[2] system.
Solar Array
80-100v
|
Solar Charger
|
LiFePo4
Battery
[40-53v]
|
Circuit
Protection
| <- Standard house wiring to point of use
|
| --- [12v DC Brick] -- short wire --> 12v devices
|
| --- [5v DC Brick] -- short wire --> 5v devices
Note that multiple devices could run on a single DC brick. It can, for example, provide an array of USB ports for devices to source power. The 5v or 12v power wiring is short so there isn't a need for oversized wire to combat voltage drop.
Transformers on DC are already a solved problem with various switched-mode converters that can boost or decrease voltages at will. In fact, many AC-DC supplies (phone charger, etc.) are already using such converters, first rectifying the incoming AC. Not needing a bulky transformer makes them smaller and lighter.
Most stuff is switched mode nowadays anyway, so this becoming less and less relevant. Not sure if there really are any transformer based appliances in my house anymore.
DC power never went away in serious ISP/telecom applications. It's more popular than ever. And things like the DC power distribution within an open compute platform standard rack, for large-scale numbers of individual small servers.
Same here. The problem is that there's no consensus on what voltages. I might prefer 48v for lower losses but you might want 24v since it's "safer" and my neighbor might want 192v since he's already an electrician.
There are multiple things that we really need to make it viable. A standard for the plug that separates it from any AC standard, a standard voltage and current and then standard DC to DC convertors for adjusting the voltage. I guess also we need a completely different light bulb fitting socket too that only fits DC light bulbs. Lots of different commercial DC convertors of varying sizes to replace the AC convertors in all sorts of devices.
I do wonder how much in practice would be gained given the DC to DC conversions are guaranteed for every device.
When I built out a van into an RV, I got to run a new electrical system for it — one based around the 12V automotive battery.
It turns out that there is a small industry out there that serves both RVs (and boats as well) that is 12V/DC based. Lights, pumps, refrigerators, fans, all running straight off a 12V power bank. "Cigarette lighter" outlets stand in for wall outlets for USB-style chargers, etc.
The 12V/DC battery system in the van is of course charged from both solar on the roof of the van and from the van's alternator (when the van battery is topped off and the van is under power of course).
I did add an inverter to supply 110V/AC for a pair of traditional electrical outlets I installed in the kitchen area of the van/RV. These are primarily used for plugging in wall-wart style chargers for the laptops.
(My Kill-A-Watt suggests that my rice cooker and even the electrical kettle would, just barely, function on the current provided by the inverter but RV-life tends toward minimalism so those extra appliances I've left behind.)
In any event, the whole experience did have me wondering if I could run a parallel 12V/DC electrical system in a new home and do away with a lot of the step-up/down of AC.
Definitely could do with some kind of modern outlet (USBC?) rather than the cigarette lighter outlets, ha ha.
What would be nice is a residential power over Ethernet standard. With standard crimp connections.
Some advantages power limited and low voltage means it's safer[1]. Ethernet means you can control lights and other devices. Smaller diameter wiring means it's cheaper. Not needing a licensed electrician to install it saves $$$.
[1] Can imagine for a non North American not having to deal with 230VAC would be a big bonus.
What would make the most sense to repurpose AC wiring for DC. The big problem is that only circuit goes to each spot and putting in duplicate circuits would be expensive.
The standard AC wiring is probably enough for 720W with 48V. The problem with choosing 48V is lock to medium power. Appliances would require 240V or 480VDC to get enough power.
The big problem is that there is no standard for DC power outlets or plugs. Would need to be different from AC and would also need to be different for different voltages.
There's already a somewhat common standard for DC wiring in homes and offices, though it's probably not exactly what you had in mind: it's Power over Ethernet (PoE), which provides around 48V over standard twisted pair Ethernet wiring, while still allowing it to be used for data.
PoE is too low power to use for many things that would want DC power. 802.3bt tops out at 71W. You couldn't run a 100W USB-C adapter on it.
Also, PoE is pretty high losses running power over Ethernet cable. The max power at source is 100W. PoE puts in higher voltage for account for the losses which means it isn't a straight 48V input to 48V output.
I think it could make a difference. A lot of devices have power factor correction (pfc) now. To draw a current without harmonics from the grid. You wouldn’t need a pfc stage anymore.
Some kind of voltage conversion would be needed anyway. Most electronic devices run at 5V, 3.3V or 1.8V. And that's limiting cable lengths to a few meters (probably less than 5m).
Yeah of course, but the pfc stage boosts the voltage first up above the ac peak voltage and then a second stage steps it down to whatever voltage is needed. If one has a dc grid one doesn’t need the pfc stage (but of course still the second stage).
DC currents emit a static magnetic field with a frequency of 0Hz. This is not my concern since the magnetic field is so weak from a 12v system. Frequency is the issue with AC currents. because the EMF from AC lines are time varying, they can induce currents and voltages in nearby conductive objects, like people.
The WHO seems to think ICNIRPs concerns are valid:
Removing the AC/DC conversion seems hardly worth the trouble. High efficiency AC/DC converters can be constructed. It's rather a question on how much one is willing to spend on such.