12

I have a 336VDC source and have been connecting it to an AC extension lead (UK, 3 Pin Sockets) and then connecting various AC devices (which have no DC input specifications on them where the power requirements are listed) such as TV's, phone chargers, network switches, NVR's etc. All of the devices work fine as presumably they all use some form of full bridge rectifier followed by a step down converter to operate.

To install this in a home or otherwise what are the safety implications or regulations of using this as opposed to AC? Obviously certain devices that use AC may not function, but in theory most modern devices should work?


The 336VDC source is from a receiver unit from a company called VoltServer. It outputs 336VDC and takes input from one of their transmitter units.

19
  • 36
    Certain devices that use AC may in fact cause a fire. Arcing across switch contacts will be a significant problem and ditto fuses. Forget it.
    – Andy aka
    Feb 7, 2022 at 11:13
  • 19
    DON’T! Fuses and switches are not rated for that high DC voltage.
    – winny
    Feb 7, 2022 at 14:57
  • 22
    NO! Just NO. Experimenting like this is a good way to end up dead. DC is deadlier than AC and much nastier to fuses, breakers, switches. Answer your questions FIRST - THEN connect power. Doing it backwards forces you to learn all the lessons the hard way. Some of those hard lessons are survivable. Others are not. Don't roll the dice.
    – J...
    Feb 7, 2022 at 20:00
  • 16
    Can you tell us a little more about why you have a 336 V DC source? And why you're running AC devices off this supply? I feel this is an XY problem. Is the power source even legal?
    – Criggie
    Feb 8, 2022 at 6:03
  • 10
    @Andyaka Here is a YouTube video showing the arc when opening a 240V DC circuit. Quite impressive, didn't know it was so bad. (The 240V AC demonstration before that just showed a minimal, short-lived spark.) Feb 8, 2022 at 10:24

6 Answers 6

25

Most fuses in 230VAC appliances are not rated for 336VDC. Some 230VAC fuses are only rated for 28VDC. Arcs are not blown out on the zero crossings with DC because there are no zero crossings. This means that the internal fuses could just keep arcing getting very very hot when a fault occurs. You can get larger more expensive fuses that are rated for DC that can be relied to blow when a fault occurs.

1
  • 11
    You can get larger more expensive fuses - but just to emphasize this point even more, that doesn't fully help unless you can open up every device and find / change its fuses, which is pretty impractical. We're not just talking about the fuse-box / breaker panel for the house wiring. Feb 8, 2022 at 18:41
19

Looks like someone has a Tesla battery lying around.

A lot of hidden obstacles and dangers here.

  1. AC-rated devices may look like working, but fail after a while (hour, day) because of some subtle part of their circuits depending on the input being AC (as in thyristor-based phase regulation) or the input being 240V RMS (see p.2).

  2. 240V-rated devices that use resistive heating of some sort generating twice the intended heat. You may or may not notice at first, until something melts down.

  3. AC-rated fuses and mechanical switches are generally not good for DC. DC-rated ones are much bulkier. Especially bad in regard to fuses because it is the fuse that separates the mild failure mode from the major fire.

Of course, everything explicitly depending on AC like transformer-based or capacitive-divider-based power supply or an AC motor is not going to work. Transformers and motors will blow spectacularily.

An universal cure for all these difficulties: an inverter. It will convert the DC into 220-230-240V AC that will be good for everything.

6
  • 1
    A tesla makes a lot of sense. Does OP have the battery or the full vehicle? I'd presume that the "house battery" mode of a tesla would have to use some kind of smart+aware wall charger, or has OP simply wired onto the positive and negative bus leads in the battery pack?
    – Criggie
    Feb 8, 2022 at 9:16
  • 1
    I was thinking a solar panel string ... which might be better, as inherently limited to about 10A. The thought of an unkillable arc from a 90kWh battery is a bit more than tingly. Feb 8, 2022 at 18:18
  • Solar panels are almost pointless without a battery - unless you have very, very specific load profile.
    – fraxinus
    Feb 8, 2022 at 19:27
  • 1
    @fraxinus: One very reasonable such load profile is cooling, where demand may be directly proportional to available sunlight, if you have a compressor that can run at variable power. Feb 9, 2022 at 1:20
  • @R..GitHubSTOPHELPINGICE agree, but only in general. They need either a good spare power or battery buffering even for cooling.
    – fraxinus
    Feb 9, 2022 at 10:56
19

Accessories specified for use in the UK are rated for 250V AC. Plugs, switches, thermostats and relays may arc excessively on DC. Their life may be a lot shorter than normal.

If you use standard plugs and sockets, nothing will stop people plugging in devices that are not capable of running on DC. Anything with a linear power supply, a synchronous motor, or resistive components such as incandescent lamps may fail very rapidly on DC. Possibly with smoke and flames.

5
  • 14
    Also anything with a big passive PFC choke on the input will seem to work just fine and then produce the mother of all sparks when suddenly unplugged.
    – TooTea
    Feb 8, 2022 at 7:49
  • 4
    Also brushed AC motors might start to "run away" when run with DC. They rely on the inductance of the coil to limit the current. If you have an angle grinder you can try connecting it to a simple 9V battery: many will start to rotate! Also vacuum cleaner motors are often brushed AC (because they can have higher rpm than the mains frequency)... SO: HVDC is a great Idea but not for todays households!
    – kruemi
    Feb 8, 2022 at 15:16
  • 1
    Nothing will stop you from plugging things in, but they will display a suprising amount of attitude when UNplugging them.... Feb 9, 2022 at 17:41
  • 2
    @kruemi just tested this with a non-speed-controlled electric drill and a bench power supply ...9V will not do the trick, but at 24V you actually get slow rotation with a surprising amount of torque... Feb 9, 2022 at 17:48
  • May arc excessively is an understatement. Any circuit breaker not rated for DC may literally catch on fire after tripping just a few times. People have tested them on youtube. For example: youtube.com/watch?v=PRT6lv8RKvA
    – mkeith
    Feb 10, 2022 at 6:19
13

All of the devices work fine as presumably they all use some form of full bridge rectifier followed by a step down converter to operate.

Well, it's not that simple because not all the home appliances or the devices-supposed-to-run-from-mains are designed to operate in DC even if they have a bridge rectifier at their inputs to convert the incoming AC to DC. The bridge allows you to apply a DC voltage to the input regardless of its polarity but there's an important thing to consider.

schematic

simulate this circuit – Schematic created using CircuitLab

When a DC is applied to the input, the whole current will flow through only two of the diodes (Either D1-D4 pair or D2-D3 pair). Depending on the design of the device, the input current may exceed the average current rating of the rectifier diodes. So they may burn eventually or instantly. This is not a problem when operating from AC because the polarity changes 50 or 60 times a second.

ADD: Luckily, supply from DC will not be a problem if there's a PFC pre-regulator in the device. Most of these devices' PFC pre-regulators generate an ENABLE signal for the post regulator (or in other words, for the rest of the circuit) when it sees the 90% of PFC boost voltage (typically 385-400VDC). So when they are supplied from 390VDC directly, this will not be a problem. But still, the main issue that I tried to express above still persists.

2
  • One rarely sees rectifiers working at the edge of their specs. A generous 3x-5x is more common. p.s. 1N4001 is a 50V diode.
    – fraxinus
    Feb 8, 2022 at 19:32
  • 2
    One rarely sees rectifiers working at the edge of their specs. A generous 3x-5x is more common. Who knows? The design might be size- or cost-limited, the designer engineer might be a dumb, or the entire company might not be that generous :) 1N4001 is a 50V diode. Yeah you're right. I just didn't bother change the sch editor's default. The intention was to visualize the possible issue.
    – Rohat Kılıç
    Feb 8, 2022 at 19:55
9

Electric railway mechanic here. This is a bad, bad idea.

Higher voltage DC is a mean drunk.

Here's a nice arc display from an inherently current-limited source - series-connected solar panels.

Here's about twice your voltage. Note the car is energized the whole time, and the arcing restarts - twice!

That could be your house.

AC crosses zero volts and zero amps 100-120 times a second. That is extremely useful for snuffing arcs. Inductive loads resist changes in current and will spike voltage to do so, and if they're doing that across a switch, in a few milliseconds it will be up against an opposing current.

Whereas in DC, absolutely nothing stops an arc. It can’t be bargained with. It can’t be reasoned with. It doesn’t feel pity, or remorse, or fear. And it absolutely will not stop, ever - unless you kill it or it destroys its own current pathway.

Like that poor tram.

Arc detection and suppression

If you ever have unhooked speakers "live", you're heard the crinkle-crunch sound of arcing (the waveform is the same on the wire). Better trolley substations have a "Rate of Rise" detector (AFCI basically) to use digital signal analysis to "listen" for that "sound", and trip power. You need something like that. The Odessa tram system didn't have it, as you can "hear" from the audio.

Now, on trolley cars, DC switches (even for small loads like a vestibule light) have two essential parts. First, the switch has a "snap" mechanism that throws the contacts far apart when they open. Second, there is a "blowout" mechanism that re-directs the arc into an "arc chute", made of ceramic or hard-board asbestos. The blowout is typically 2-3 turns of magnetic coil; the collapsing magnetic field pulls the arc into the chute.

So you need your switches and circuit breakers to be rated for interrupting DC. Your breakers need to be rated to interrupt a bolted dead short - AC power breakers are typically rated 10kA or 22kA. Further, you need a disconnect switch! You can't rely on pulling the plug, because if the fuse in a British plug blows, it will just arc across the opened fuse, and much like the poor tram, burn up the plug entirely, and even spread the arc to the melting socket. At which point you'll never get it out.

You see in Peter's video here (from comments).

DC ratings are much lower on average equipment

And if you read your equipment's spec sheets, you'll see that their DC ratings are lower than AC by as much as 90%. This is because of the sheer difficulty of snuffing a DC arc. For instance a switch rated 250V might be rated only 28 volts DC.

I remember seeing a light rail vehicle where they just could not obtain a 700VDC rated contactor. They used a 3-phase contactor rated for 2000VAC and wired all three phases in series.

Rearrange the battery pack for a lower voltage.

Split the pack and re-stack it in series-parallel for a lower voltage. Aim for voltage around 36 volts, although you'll be limited by the sub-pack sizes of course. Don't go much over 48 volts, even that is starting to head into DC's dark side.

Anything you do with DC of even 36 volts needs to be properly BSI/TUV/UL listed equipment. None of that "cheap Chinese" stuff, get gear from quality electronics supply, or use AC mains gear that is cross-rated for DC. For instance Schneider rates some of its consumer-tier lines for DC up to 48V (120V available on specialorder).

-1

None of these answers seems to have checked up on VoltServer, and have gone off the assumption that this is an uncontrolled DC rectifier or a battery. I think we're all in agreement - a custom power system for 336 V DC has to be designed carefully by someone who knows what they're doing, AND checked by someone else who does as well.

Now VoltServer claims to be someone who knows what they're doing, and claims to create a DC circuit of a type called Class 4 Power-limited FMPS circuit, which seems to be digitally monitored power and which can stop power supply the moment any arcing or other anomalous condition is detected.

So overall, it seems to be fine and reasonably safe to use this 336V system because of the seemingly certified fault management inbuilt into the VoltServer power supply. Many Active PFC devices officially support DC input (95-400 V typically) in their spec sheets. The ones that don't, might have issues relating to overheating of 2 of the 4 diodes though.


Now I'm not from the US, and I am not familiar with NEC - so I myself have several questions on this new power class - such as why it reads like company marketing material, or why it needs to be digital (a DC RCCB should do nicely as a quick analog device).

Here's one article on what the NEC regulation on Fault Managed Power System (FMPS)s apparently says.

4
  • Hi, I dont see how this answers the question. Aug 5 at 19:48
  • @RohitGupta Thanks for notifying me, I forgot to put the punchline.
    – Milind R
    Aug 6 at 10:08
  • This does not provide an answer to the question. To critique or request clarification from an author, leave a comment below their post. - From Review
    – FreeMan
    Aug 6 at 18:23
  • @FreeMan I don't understand - the question is, should this work? And what are the safety implications? I answered, yes it would work. And that it should be safe and legitimate in the eyes of the regulator. Almost all the other answers don't actually address this, and talk only about the danger involved. What am I missing here?
    – Milind R
    Aug 6 at 19:31

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