hckrnws

So we're saving on the wire only. I would probably pay for extra wire just to make the system simpler, unless I was a builder aiming to save on costs.

What the world needs is a 150 watt low voltage power over ethernet standard for lighting an low power appliances.

There is, shockingly enough, 100W PoE lighting...

https://www.cisco.com/c/en/us/solutions/enterprise-networks/...

Imho a totally crazy waste of expensive gear, but whatever floats your boat...

Isn't it cheaper and more reliable than a dozen ZigBee bulbs?

Could be, I don't have any price info on all of the PoE lighting stuff.

But let's say i've got an office area, where I would need 2kW lights in total (which isn't much, I think 200W per desk is considered bright enough). This would need 20 PoE ports, 20 cables, 20 PoE-capable light fixtures with the right transformers/bulbs. Just the 2kW/>20port PoE-Switch will set you back 2k€ or more, no idea what the light fixtures will cost. Maybe a 50W ZigBee bulb will cost 50€, you need 20, then you are at the price of just the switch. And you can just use plain electrical wiring, no need to do the whole more expensive Cat6+Patchpanel thing.

And my personal favourite, even better than overly expensive ZigBee bulbs: https://www.shelly.com/de/products/shelly-dimmer2 Wifi-controllable dimmer, speaks MQTT or HTTP, for a load of 200W at 35€, works with any dimmable light fixture. Also available as a relay if you aren't interested in dimming, then you can do 2kW in 2 channels at a lower price. And if you are worried about WiFi, there is a cabled version as well.

200W is a huge number for modern lights.

You don't care about the power consumption, You care about the light level produced (Watts != Lumen).

Most advice says 500 to 1000 lux is fine for an office desk (en 12464).

For a 1 sq meter desk, that's about 800 lumens worth of light, or maybe 50 watts max in modern LED lighting (two 800 lumen bulbs, 14W each).

LEDs are just way better than older lights.

I could hope for a future PoE-over-Data Line (PoDL) version using an 18 AWG twisted pair that would work for this, providing something like 48V @ 3 or 4 A, and 10BASE-T1L for data. Hopefully a standard way to do this gains traction.

The existing limitation is with the 2x23 AWG (or 4x23 AWG) of Cat5e/6/6a not providing a lot of current-carrying capability at a safe voltage.

Big concern IMO with 10BASE-T1 and PoDL for long runs / in wall installations etc is the lack of galvanic isolation. Maybe I'm just paranoid but there's a ton of advantages to not having to worry about things like ground loops and common mode voltages etc.

What I wouldn't mind would be a beefier 100BASE-TX 2-pair cable spec, and some extended PoE profiles. You'd need high current magnetics or something but surely that's possible.

Was just thinking about it for my home. Googling showed that it's only for commercial applications now (like "request a quote" for a lightbulb).

There is a good opportunity for a startup IMHO.

With 100W PoE, one can daisy-chain 10 light bulbs on one cable (or put them on different pairs of wires within one cat6 cable).

But what I really wanted is controllable devices, so that I could re-assign lights<->switches in my home hub.

UPDATE: and you don't need an electrician license to work with low voltage.

A managed switch will allow you to power on/off specific ports.

When I was looking at home automation systems and all the insanity found there. at one point I said "if I were to design a home automation system from scratch what would it look like?" my answer for the physical layer was poe and managed switches.

> A managed switch will allow you to power on/off specific ports.

As somebody that's looked into this... yes, but. The contemporary light switch is instant. Within a few nanoseconds of your brain feeling the "click", you see the light. There's a noticeable latency from an API request to the switch's control plane which then talks to the power control which then turns power to the port off and then the light goes out. You have the additional delay of poe negotiation for the "turn on" operation, too.

That is how I managed to talk myself out of investing in a home automation setup, none of the incompatible systems felt good and more importantly, they were all worse than having a physical switch I can toggle.

Lutron offers a variety of styles that work near instantly. I’ve been using them for years.

It really isn’t- 20-30ms minimum for most switches/lights, on average. Incandescent likely a bit more.

Which is still ‘instant’, but with some non-ridiculous implementation, also easy to do with a ‘smart home’.

Of course most implementations are pretty ridiculous, so there does that idea.

You would likely need a low voltage license to permanently install something like this in most states. There are standardized limits on how many of these can be run in parallel due to heat constraints which is one of the thresholds that necessitates a license

I keep telling myself one day I’ll design a PoE DC to 120VAC adapter board so I can run PoE to a standard outlet box or light switch.

Yeah you get it. What triggered it for me is there is a PoE standard for automotive that can deliver 50W over a single twisted pair.

If you bumped that to 125 to 150W then it would yes be great for residential for all the reasons you can think of.

Cheap cable, crimp connectors. No license required to install.

Low voltage and generally much safer than 120V and especially 240V. Consider with power negotiation a default max power of a few watts which isn't going to start fires.

Wanna replace a light fixture? Unscrew the old one, unplug it. Plug the new one in and screw it back on.

Also imagine kid safe outlets that are standard world wide.

USB-C fulfils that already.

It doesn't work out economically because big networks of usb devices (ie. Hubs) all need the power components if one is to maintain the generalisability of being able to plug anything into the grid anywhere.

USB has significant length limitations and the cabling is generally not plenum rated. It's also silly to have a cable with wires that will never be used beyond negotiating for power.

USB-C is for short distances (<4m) and is more expensive than CAT6 cable.

https://www.cablematters.com/Blog/USB-C/how-long-can-a-usb-c...

As an American, I'd just be happy if we switched to 240v rather than 120.

This is a common misconception. You do have 240V(or 208V in some apartments if we're being pedantic). You primarily use it in the kitchen(range receptacle), laundry, and for your HVAC equipment or an EV charger if you have one.

Someone who knows what they're doing can also add a 240V outlet to any room.

We don't serve 120V to homes, we take 240V and split it. In fact, that's not the only time you use a mix of different voltages. You take that 120V, and then further transform it to 18-24V for things like doorbells and thermostats, or you use a power-brick at the wall to convert it to direct current to charge your laptop, phone, and other items. In none of those use cases would 240V really change your life. It's really only better at boiling water in counter-top kettles. I'm guessing we don't use more 240V kettles because we're not a tea drinking country and culturally people tend to use coffee machines that are "fast enough", but not slow enough to be worth switching over.

> We don't serve 120V to homes, we take 240V and split it.

Ackchyually, we deliver 120V to the poles, not 240V. Then transformers at the poles near the homes split it to two 120V phases, 180° apart. So the difference between +120V and -120V becomes 240V. But the pole wires are still 120V.

three phase, If we are going with unrealistic dreams, I want three phase to every house.

"avoids eye contact as you patently explain that the only real benefit to three phase is for motors"

But really the US is a 240V system. we just split the phase and wire the house in two halfs. you should not even need to rewire your house. Just hook the neutral to the opposite phase in the panel, like magic you now have 240V in every outlet.

Don't actually do this it will end up killing someone.

As the sibling comment notes: The US/Canada are already split-phase 240v/120v systems.

As near to 100% as makes no difference every single house is already wired for 240v. In fact wired for it and using it: electric ranges, stoves, dryers, etc are all 240v.

NEMA even defines 240v receptacles/plugs for normal amperage: NEMA 6-15 (two horizontal blades: meh face) and 6-20 (one vertical, one horizontal: wink with other eye closed). Unlike the common 240v dryer/large appliance cords that are huge and bulky 6-15 and 6-20 are about the same size as our current 120v plugs. They have the same compatibility as 5-15/5-20: The 15 amp version fits in the 20 amp receptacle but the 20 amp only fits 20 amp. And it is impossible to plug 120 into 240 or vice-versa. Everything made for 120v (to a rough approximation) is also rated for 240v.

Unfortunately no one bothers to install the 6-15/6-20 plugs. There is nothing stopping any builder from doing it standard, especially in the kitchen. There is nothing stopping a homeowner/buyer from asking for it either. But no one does. Therefore there is no market for appliances that use these plugs. And thus no demand to wire for them.

It would likely take a push from government, manufacturers, the NEC, etc to push for supporting 240v for common appliances. Start installing them in new homes. Offering the 240v version of electric kettles and such.

Note: some European appliances can be wired up this way and will run fine because they tolerate 60Hz but not all of them.

> Unfortunately no one bothers to install the 6-15/6-20 plugs

Almost no one, but I had a 6-15R put in my kitchen. Then I imported a 3kW tea kettle from the UK, lopped off the plug, and put on a 6-15P. And now my wife doesn’t have to wait very long for her hot water.

NEMA 6-15 and 6-20 seem to be pretty uncommon. This is no doubt in part because most stuff we run on 240 in the US is higher amperage anyway, and probably in part because some stuff also wants a neutral?

The only NEMA 6 I have is on my table saw, and maybe the planer? Everything else is NEMA 14 (two hots, neutral, and ground). The upside of that is the bulb in your dryer, for example, can be a standard 120 volt bulb.

I'd also like to note that the locking connectors plug and unplug a hell of a lot nicer than the non-locking ones, at least at the 30 amp size (which is common for a dryer). They just operate much more smoothly, presumably because they aren't relying on friction to keep them plugged in.

Eh, if we're going to be making big changes, might as well switch to the type C and type E plugs / sockets as well. They're safer, and the distinguishable nature from the type A / 120v devices currently in use (esp the ones that really can't handle 240).

POE+++ is 90W so maybe POE++++?

Hard to go up in voltage as its now no longer considered "low" voltage wiring and hard to go up in current without increasing conductor size of ethernet cable.

POE+=4 is a clear win at this level. ;-)

Wouldn't that be POE+=2?

Also, if this is anything like USB it will be PoE 3.14 Gen 2 type E.

PoE seems excessively wasteful for dumb loads. 8 conductors for lighting and a dozen bytes of "turn off bulb 1234" is crazy. Not to mention that you need almost as much electronics on either end of the wire as you do with any 120VAC to DC adapter.

We definitely need domestic DC systems, but not loaded down with the hardware and software complexity of Ethernet. You could run three wires for DC plus half-duplex RS232 or similar, or you could do two wires and impose a high frequency communication signal on top of the DC. There are many options that don't require your lights to have IP addresses.

Even though it's 8 conductors, you still put only one cable, that's much easier than even two wires.

Low voltage means you don't need an electrician license to DIY, because it's not going to burn the house.

Mass producing cables and electronics becomes cheap. The limiting factor is then materials, not complexity. Look at ESP32 controllers for $7.

Also, I d like my cable to serve 10 light bulbs, each of which may not only turn on/off, but change brightness and color temperature.

Not sure sharing the neutral buys much as conductor size needs to go up vs just running separate wire.

Most hybrid inverters now days are high frequency and have a HVDC bus sitting between battery DC/DC, solar MPPT and AC inverter/converter. Since solar comes down from roof with HVDC it can be inverted without a transformer to AC then the LV battery uses a physically smaller HF transformer on the buck/boost to that bus.

Seems to be lots of talk to tap that HVDC bus for say EV charging as its around 400v, also some inverters moving to HVDC batteries as well simplifying the DC/DC even more so.

Could see a future where houses have a HVDC panel with battery, solar, inverter and EV charger hanging off it, perhaps even HVDC from pole with solid state transformers. If/when EV become popular then you will have much more readily available/cheap HVDC switching gear at 400/800v. Hard to beat the simplicity and reliability of AC using transformers though for distribution.

> Not sure sharing the neutral buys much as conductor size needs to go up vs just running separate wire.

The article mentions a system which will balance the shared current. But that sounds like needless complexity that was introduced to entice power companies by saving a few bucks not needing two insulators and conductors installed on each pole.

> The article mentions a system which will balance the shared current.

Not quite.

It mentions the asymmetric Neutral current, and this is because in a balanced multiphase system, there is (almost) no current flowing on the neutral conductor, i.e. if you are consuming 10 amps from L1, 10 amps from L2, and 10 amps from L3, then there are zero amps flowing on Neutral. This is also why inherently-perfectly-balanced multi-phase loads (e.g. 3-phase motors with identical phase windings) can be constructed such that they do not require a Neutral connection at all.

In the UK for example it's quite common that houses only get connected up to one of the phases, so if you have 60 houses on a street, 20 each would be connected to only L1, only L2, and only L3. This achieves a reasonable balance in practice; the current flowing on the street's Neutral conductor(s) is/are very small in comparison to the currents on the Line conductors.

As in, it is already fairly well balanced, and there is nothing to be done in this regard.

Within the single-phase installation the currents on Line and Neutral are of course equal under fault-free conditions (typically up to 100 amps for a residential installation). There's nothing you can do to reduce that.

A lot of countries in mainland Europe wire up residential properties to all 3 phases (with much smaller permitted supply currents; I've seen as low as 20-amp main supply fuses), but again they go to some degree of effort to balance their phases with respect to single-phase loads. This is for grid stability reasons.

For the street example again, they could have wired up every house identically, but connected the phases in a round-robin fashion, such that the socket outlet circuit(s) are on a different phase in each house, and so on. If there were ever a severe imbalance (which would be detected automatically, or by routine maintenance and inspection of distribution transformers, depending upon how bad it is), typically it would be arranged for some rebalancing work to be performed, for example by disconnecting an outgoing feeder line serving some dozen properties or more and rotating their phases (connecting the old outgoing L1 to the transformer's L2 output, L2 to L3, L3 to L1). This still maintains the phase rotation order (very important for 3-phase loads, like some EVSEs and motors) but allows a more equal distribution of loads.

See https://docs.openenergymonitor.org/electricity-monitoring/ac... for why this is important.

This is somewhat simplified though. Real-world loads vary their consumption dramatically, and also feature reactive components that shift the current forward or backward with respect to the voltage. A 10-amp capacitive load on L1 and L2 will not cancel out a 10-amp resistive load on L3. Harmonics are also a problem, as the article mentions.

Well aware of how three phase systems work. You're not guaranteed perfect balance which the article addresses by stating that DC current injection is monitored and controlled which is what I was referring to.

Really not sure what/who this is for. There aren't many 3-phase LV AC + DC use cases apart from industrial load + control. I can't imaging the DC part can be transmitted at a voltage that will make the system efficient while avoiding an expensive step-down conveter.

Amusingly, the first thing I thought of was "phantom power" for microphones, where a power supply for the microphone is carried along the same pair that returns the audio signal.

I don’t see a world where this is possibly cheaper than just putting a AC-DC supply at the point of load. Looks like a big safety hazard too.

IDK, but I wouldn’t even want to debug circuits for this thing.

Cute. This is one of those websites that disables selecting text.

They explain that the DC injection must be scaled down to keep current through the neutral conductor at safe levels when the AC is under load.

That's worse than useless. Nobody wants their lights to dim when they plug in a toaster. Most people would rightly assume that's a dangerous overload condition and their wiring is faulty.

All DC loads must be capable of actively responding to DC brownouts without spiraling the system to collapse. That means every single item on your DC bus needs complex electronics and software. Every lightbulb every phone charger, every widget must be able to safely handle unpredictably variable levels od DC power and all loads need to agree on what they're doing so that no one device sees the voltage creep up and suddenly draw more power than is available.

I don't think this concept is at all feasible.

This is nuts. Why make such a complicated system with another fault point?