[Robotbeat]

So, I used to be "team DC" as well, but after learning up a bit, I realized there are still a lot of advantages to AC in the 21st Century.

One reason DC power is more viable nowadays is (as others here have mentioned) the use of DC-DC converters. They internally actually use AC to do the conversion (usually at much higher frequency to save weight and cost on inductive elements). But they're not super cheap (not cheaper than transformers) and they can also be inefficient.

But part of my problem with this article is it kind of sandbags the efficiency. We can get pretty high efficiency inverters and rectifiers. >90% is common and 95-98% is feasible (and not uncommon). Low voltage DC (like 12V) requires MUCH thicker cables for the same power, which means a lot more copper (and copper mining). Typical line voltage can be very lightweight.

With DC, nothing is really at the same voltage, so you need DC-DC converters all over anyway, so you're not saving anything (although not losing much, either!).

Another important thing that really drives some of the advantages is: breakers, relays, and (much less important) current measurements of existing cables. All these are feasible for both DC and AC, but cheaper and easier with AC. AC is self-extinguishing as it crosses zero 120 times a second. That means you can switch a circuit on or off when the applied voltage is very low, meaning you can use cheaper and lighter power electronics. Plus, at a given voltage, there's a safety advantage as arcs will more easily stop with AC.

I actually think it's funny to see this in a "low tech" blog. DC ubiquity is something really only practical for significant use (i.e. beyond automotive or RV) with modern 21st century power electronics.

[throw0101a]

> Low voltage DC (like 12V) requires MUCH thicker cables for the same power, which means a lot more copper (and copper mining). Typical line voltage can be very lightweight.

-48V DC has been a thing with telco equipment for decades: it's what a landline telephone uses for signalling.

* https://www.servertech.com/blog/48vdc-power-and-the-backbone...

If we'd use anything DC, it would probably be that.

[MisterTea]

DC is used where you want direct battery backup, meaning the battery bank voltage is directly supplied to the equipment. This makes the system more reliable as there are no power supplies or inverters to fail. The system can then be composed of simple fuses, breakers, switches and relays to control power flow.

You can even design the system so that the batteries are in parallel with the power supply and load making the system uninterruptible by default.

A lesser known standard is 110V DC used in electrical substations and switchgear. They cover a lot of ground and cable runs can be hundreds of meters so the higher voltage allows longer distances with minimal line loss.

[JKCalhoun]

12V has been standard in auto/marine for quite some time. There are plenty of consumer items that already run on 12V.

Power hogs like refrigerators, window-AC units would no doubt still run on 110/220.

But I already have mixed voltage AC in my home: otherwise 110V but purpose-placed 220V for electric stove/oven (in kitchen) and electric clothes dryer (in garage).

In my perfect future only those high-current outlets would remain — the rest of the house would have 12V plugs and lighting.

[throw0101a]

There is more and more 24V in marine systems to deal with higher loads in recent years. Your 28' (8m) day sailor may not need much more than a starter battery, but once you're at 40' (12m) and above things start changing. E-motors in the marine world generally use 48V (AFAICT).

48V is also being looked at in the auto world as well, at least for "low voltage" stuff:

* https://www.automotiveworld.com/articles/48-volt-likely-to-b...

The IEEE's 802.3bu PoDL can handle sending 12, 24, and 48V, and automotive is one of its deployment spaces:†

* https://blog.siemon.com/standards/ieee-std-802-3bu-2016-powe...

* https://en.wikipedia.org/wiki/Power_over_Ethernet#Standards_...

† 10BASE-T1, 100BASE-T1, 2.5GBASE-T1, 5GBASE-T1, and 10GBASE-T1; .3cz task force is working on 25/50/100 Gb/s.

[mindslight]

DC-DC converters can be much simpler than PFC [0] AC-DC converters.

There are actually scant few things in modern houses that inherently want to run on 60Hz single phase AC - basically just fixed-speed induction motors [1]. And the amount of those are shrinking as more and more appliances get "inverter" technology for the energy savings.

The transformers on the poles are bona fide 60 Hz devices (and I'll take the electricity I can get). Industrial customers with large 3 phase induction motors want 60Hz AC. But for household use, DC would be straightforward.

I agree about the self-extinguishing arcs, although it seems like moving to solid state switches would solve much of that.

[0] power factor correction

[1] And actually, "single phase" induction motors don't even want to run off of a single phase. They all contain capacitors to create a second phase so the motor rotates.

[mastax]

> One reason DC power is more viable nowadays is (as others here have mentioned) the use of DC-DC converters. They internally actually use AC to do the conversion (usually at much higher frequency to save weight and cost on inductive elements). But they're not super cheap (not cheaper than transformers) and they can also be inefficient.

DC-DC converters are cheaper than the classic transformers. There's a reason that you almost never see the classic "big chunk of iron" wall-warts anymore: not only are they inefficient, they're more expensive.

> Low voltage DC (like 12V) requires MUCH thicker cables for the same power, which means a lot more copper (and copper mining). Typical line voltage can be very lightweight.

Yes, a low voltage DC bus doesn't make much sense except for specific applications. Even in small spaces like recreational boats it's more common to use 24V or 48V these days. I think around 50V makes sense, as it's not very dangerous so there are fewer safety regulations. Look at PoE, for example. Unprotected runs of fairly thin cable transmitting up to 52V.

> With DC, nothing is really at the same voltage, so you need DC-DC converters all over anyway, so you're not saving anything (although not losing much, either!).

DC-DC buck converters are the simplest, cheapest, and most efficient category of power converters at the moment. (Of course a specific converter may be worse than a specific converter of a different category, but this is broadly true) Buck converters that support up to 50V are only marginally more expensive than lower-voltage ones. As an electronics engineer, I would be very happy if I could get away with only needing buck converters from now on; they're very easy.

If the DC source is required to be galvanically isolated and to have it's own transient suppression than the DC devices can get away with much less input protection than AC devices require. Could be a significant cost savings.

> Another important thing that really drives some of the advantages is: breakers, relays, and (much less important) current measurements of existing cables. [...]

Absolutely. I don't see high voltage/high power DC becoming popular in homes any time soon for that reason. There shouldn't be any difficulty with arcing at 50V though. (Yes, switch contacts still arc but it's not any more difficult than 120VAC) Also high power applications in houses tend to work just fine on AC (motors and restive heaters), though that may be a chicken-and-egg thing.

[lovemenot]

>> With DC, nothing is really at the same voltage, so you need DC-DC converters all over anyway,

This may be the case now, but it's just through historical lack of standardisation, resulting from AC having been the one true standard. It's not inherent to DC.

USB PD is attempting to address this issue. It is standardising voltages for DC-powered devices.

[SilverRed]

Even internally there is no standard. You may put 12V in your user facing input and the device then splits that in to 12v, 5v, 3.5v for all the different components which run at different voltages.

[Robotbeat]

But internally, devices operate at different voltages due to fundamental differences in required voltages for their transistors, LEDs, battery cells of different chemistry and charge state, etc. There’s not much getting around it in a strict sense, although standardization can help.

[bluGill]

USB PD works because they can assume very short distances. It wouldn't work for anything longer than a few meters, and it isn't clear you can safely scale it up to longer distances or higher voltages.

[rini17]

There are inherent factors, though.

1. Any time you mix motors and electronics, motors produce brownouts and spikes which can cause directly connected electronics to malfunction. DC-DC converter mitigates that. LEDs are particularly prone to overload with just minute increase in voltage and dim/flicker with decrease.

Not inherent to DC but same thing happens with poorly designed appliances that emit RF which then interferes with everything connected to the circuit.

2. Voltages. You can't standardize electronics power supply to 12V because semiconductors are most efficient at lower voltage, which depends on particular technology.