Hacker News Reader

Comment by philipkglass | Hacker News Reader

The ratio of DC panel wattage to AC inverter wattage is called the inverter loading ratio, and it has been creeping up over time in utility scale solar farms. The reasoning is indeed to give more of a plateau shape to AC output (which is limited by transmission line capacity) and to perform more steadily under hazy or partly cloudy conditions. In theory, this ratio could go up a lot if panel prices continue to decline faster than inverter prices.

In practice, inverters probably also need partial redesign for a high loading ratio. Higher ILRs in large solar farms has appeared to contribute to faster inverter failures in recent years.

View on HN

> In practice, inverters probably also need partial redesign

I'd also love for appliances to entirely adopt DC.

- Homes should run on batteries.

- Batteries should be trickle charged from the grid with a cheap rectifier.

I agree on more DC appliances. Can't really escape having a failure-prone component in the mix though. For utility-scale you still need AC for transmission. For small-scale you need a switching regulator to even out the output voltage.

Been tossing around the idea of Poe Eth to 100w usb-c wall ports. I'd love to see more Poe DC gear. In aus we have laws where if it's <100v and x amperage you don't need a license to work on it. Would make rolling out a house or shop install/fitout able to be done using regular labourers not expensive sparkies.

Hell sneak some data in and you have smart control of all these connected usb-c cabled devices too. Make for game changing smart home rollout.

Gimme lamps, lights, speakers, fans, window motors, TVs all on usb-c power plus data pls.

Power over Ethernet can't do 100W. The 802.3bt supports 51W (Type 3) or 71W (Type 2). PoE uses pretty small currents, Type 4 is 960mA.

Also, PoE does have significant loss over longer distances. It isn't a problem with low power devices but would be a huge issue for higher power. It is almost certainly more efficient to send power as 120V AC and convert to USB-C.

The alternative is outlets with USB-C ports, or power strips with USB-C.

Places that use 12V/48V DC have much thicker cables. They also have shorter lengths on RV, boat, or cabin.

I think if there was large scale adoption of PoE to power stuff outside of network equipment the regs would eventually catch up and it would have similar requirements line voltage wiring. 100W is more than enough to start a fire. 100W at 100V is enough to kill someone. There just aren’t as many regs because it’s new and not widely used. Building codes and standards are written in blood. One an apartment burns down because a bad RJ-45 connector started a fire the codes will catch up.

> In aus we have laws where if it's <100v and x amperage you don't need a license to work on it

You're the first other person I've seen to point this out. Better even is negotiated power means normally the wire would only deliver 15W which isn't going to burn down anything.

I'm also seeing more and more cordless appliances. A cordless vacuum doesn't need much power when charging.

Extremely amusing to see that the war of the currents is back on, between cheap solar panels and the advent of high-voltage DC power transmission. Edison might still get the last laugh.

Today's DC electric components use such advanced electronics (like IGBT) that Edison would find very alien.

What’s the advantage of DC? You presumably still need to step down from high voltage or you end up with expensive thick wiring to avoid resistive losses.

You seem to imply that you can’t have high voltage DC. Why wouldn’t you be able to have 200v dc at home?

Because its nasty stuff! switching off a low Z 200V DC carrying 20 A in an emergency is not trivial.

AC arcs self extinguish every 100 or 120 Hz.

I'm just talking about DC inside the home. To the meter it should still be AC, then pushed into a battery bank. The house should only use the battery bank (i.e. different battery packs that are dedicated for different parts of the house - like the breaker box - the battery packs exchange power between each other and the grid as needed.

This way DC -> DC fast charging could be a thing, also my DC -> DC oven could pull a lot more Amps and pre-heat very quickly.

Inputs:

- Solar Panels -> Battery Bank

- Grid <-Inverter-> Battery Bank

Outputs:

- Battery Bank -> Washing Machine

- Battery Bank -> Lights

- Battery Bank -> Induction Stove

- Battery Bank -> USC-C to charge devices (my devices could charge so fast!)

Why do you think DC to DC fast charging will be better than AC to DC fast charging? Both involve voltage conversion and the conversion has about the same efficiency. The one advantage is that could choose higher DC voltage, like 480V, and send more power on same wires. But that won't get you DC fast charging, will still take a couple hours to charge.

A lot of the things you are talking about will require conversion. LEDs will still need conversion from line voltage. Any incandescent lights will need AC conversion. The USB-C adapter will still exist, and there would no effect on charging speeds.

You can get most of the benefit by having DC power cables between the battery, solar panels, and inverter.

https://www.noodoe.com/blog/what-is-dc-fast-charging-and-how...

Most certainly DC to DC is faster. It is only limited by the local infrastructure. i.e. want to charge faster? own a bigger / better battery bank, with bigger charging cables and a better car.

DC -> DC (with some PWM / battery management) can also be slowed down (or use reduced amps) to best suit the moment. So if versatility is better, then DC to DC is again "better".

This is the scaled up equivalent of an external laptop AC/DC power brick. It’s faster because it’s supplying much, much more power. Multiple homes worth.

Batteries need DC to charge and putting the transformer in the external fast charger means the one in the car itself only needs to be sized for lower power domestic charging.

You can currently use most appliances with switching power supplies on DC, but you need at least 100V.

100V doesn't cut it, we need them to work off USB-C. The current generation of appliances is a bit too old, and should soon be phased out. It has already begun, but the next generation of appliances all have built in battery and capacitor banks. Batteries are trickle charged through USB-C (i.e. like phones and laptops).

e.g. cordless vacuums / lawn equipment with swappable batteries.

e.g. water proof electric shavers, toothbrushes, blow dryers.

Likely soon induction stoves, air fryers (for instant pre-heat), heat pumps (including the fridge / freezer), blenders, etc. If it can run on USB-C, off the wall cool. If it can't add a battery.

I think you’re underestimating the resistive losses at lower voltage. For the same amount of power (volts x amps) resistive losses are inverse square to the voltage so losses are 100x greater at 10V x 10A than at 100V x 1A.

Resistance is inverse to the cross sectional area of the wire so to compensate you need thicker wire, 10x the diameter (100x the area) for 1/10th the voltage x 10x the current. That’s why the wire that connects to your car battery is so much thicker than the mains wire in your home.

You could have high voltage DC in your home but then you still need the transformer so I’m not sure it really buys you anything.

Have you ever had an issue with the resistive losses in your long ethernet cords?

If you're prefer you can think of it as: each room gets its own Tesla powerwall, and therefore the battery to wall socket resistive losses are negligible.

Power over Ethernet usually supplies milliamps of current. Resistive losses are proportional to the square of the amps so they’re not a big problem for most things you power over PoE.

Latest standards seem to max out at 90W. https://www.analog.com/en/technical-articles/ltpoe-extends-p...

Over the 100M max cable length voltage will drop 25% from 56V input to 48V output. http://poe-world.com/Calculator/

For the same power draw, losses are 4x as high as at 110V, and 16x as high as at 220V. DC/DC power adapters are more efficient at low power (a few watts), so PoE might conceivably be marginal efficiency win for powering a bunch of low power VoIP phones. But as soon as you want to do something like power a laptop or even fast charge a mobile phone, you’re going to be better off plugging it into the mains with a modern AC/DC adapter that can be 95% efficient.

It’s not that AC is better than DC, it’s that higher voltages are far more efficient than lower voltages as soon as you need even moderate amounts of power over moderate distances.

Health.

Caveat: Unsubstantiated statements follow. There would be some EMF around the wires. While 60 Hz is probably less detrimental, the harmonics (of higher frequencies) are. The effects are subtle and possibly not obvious for healthy people, but can make a difference with autoimmune issues.

Why do you think this is so?

You have to look at the experiments conducted by a guy called Trevor Marshall http://www.trevormarshall.com/

Note most of his experiments correspond to radio waves ( WiFi, AM FM etc). But I see no reason to not extend the concept to 50/60Hz and resulting harmonics.

So you want to replace the cost of an inverter with the cost of a battery?

That seems like a bad trade-off to me. Batteries live 5 years? Inverters live 20+.

Inverters are also cheaper than batteries. I'm not seeing any advantage to your idea.

LFP batteries have a much longer lifespan.

Also a Li-Ion with a good BMS lives far longer than 5 years

Anything with a motor will get a lot more expensive. AC motors are dirt cheap.

Also safety is harder on DC. For one the relays must be made much larger to withstand the nasty arcing. Arc extinguishing is much harder too.

Efficiency is a toss up. On one hand, rectifiers are very efficient and you could get by with one rectifier instead of many. On the other, they're not 100% efficient and the voltages are much lower -> more copper.

Reliability wont be as good. Rectifiers go all the time (think PC's CPU), there's no comparable failure component in the AC home.

Overall the idea has merit, mainly because the most home loads are DC nowadays anyway. But your AC, washer, fridge, garage opener, pool pump, furnace blower, refrigerator, dryer and dishwasher will not be pleased.

The problem is that there is no standard for house-scale DC power. It has to be higher voltage than 12V or 48V that are used off-grid. There are no standard plugs or sockets, and it has to be different than AC.

The big problem is that everyone has AC sockets, and it isn't worth the trouble to switch. It is going to be huge expense to buy new appliances for potential small efficiency gain. The in-wall voltage is going to mean that need adapters for other DC uses.

A better approach is to have a standard for DC power between AC side, batteries, and solar panels. Then have a single big inverter between AC and DC. It also makes possible to have DC-only in places like cabins that aren't connected to the grid.

Finally, many houses and apartments can't add batteries or solar panel. Not everyone has space like a basement. Like there is efficiency advantage like with utility solar, it might make sense to put batteries in substations where can use containers.

USB-C might be that DC standard. It is easy to install aftermarket AC outlet modules with an embedded transformer feeding a pair of DC sockets. One could easily imagine a house wired with such USB power outlets, lacking the AC sockets - they'd presumably be supplied by buck converters from some higher bus voltage, likely 48v.

There are outlets with USB-A and USB-C ports. I don’t think there are any with USB-PD because chips are expensive for that. It would be possible to have multiple ports and no AC ports.

But problem is places with lots of devices, then would need power strip and might as well use AC one.

The problem is appliances that draw too much power for USB-C. They also draw too much power for 48V DC. There isn’t any advantage to switching the wall voltage.

> I don’t think there are any with USB-PD because chips are expensive for that.

Leviton makes a 60W unit now that supports PD. Be warned it's a little janky in that it tries to be "smart" and re-negotiate (well, outright drop to 5V) the port if your device stops drawing substantially less current than requested. This can confuse some devices - especially those without batteries. Otherwise I can say the 20V/3A profile works fine on my macbook.

I installed a few outlets last year which claimed to offer USB-PD at 18 watts. They were certainly more expensive than any other kind of receptacle, but not so much that I remember what I spent on them.

I don't think anyone would switch the wall voltage altogether; not in one generation, at least. But a majority of the devices we plug into AC outlets these days convert that power to DC before doing anything with it, and a majority of those devices do so using USB; so one could imagine that a house which had solar panels and batteries could also have DC power wiring and dedicated USB-PD receptacles, skipping the inverter and all the rectifiers.

The majority of devices might be USB powered, but the majority of energy use is from large AC appliances. USB is solved by adapter or special outlets.

Also, the DC voltage for house wires and batteries need to be much higher than USB voltage so will need conversion. Might as well have it be AC to DC converter.

There is a problem connecting solar panels to batteries and needing two inverters. But that could be solved with higher voltage DC between them.

A 48V dc circuit over a 12 AWG wire would provide less than 1/2 the wattage (power) available in a typical 120V AC circuit. You’re still going to need some circuits that can deliver 1500W to power appliances, space heaters, window AC, etc.

Yeah I feel like this is very tractable problem. Most of my stuff runs on USB-C anyways, I just need USB power outlets everywhere. My lights, motor blinds, bathroom fans can run on POE, including data.

My oven should already have its own battery installed, or a capacitor bank for very fast preheat.

My fridge can do the same. It’s rarely actually compressing.

I’m mostly just trying to avoid needing to buy very expensive inverters.

Those types of appliances are coming:

Induction stoves with batteries built in, and why they matter

A conversation with Sam Calisch of Channing Street Copper Company and Wyatt Merrill of DOE.

https://www.volts.wtf/p/induction-stoves-with-batteries-buil...

And so we come full circle. Using a DC battery to power an inherently AC induction process. These are not your mother's home appliances.

Induction DC stove tops are also popular for off the grid cooking…when you just have solar panels and a car battery.

USB is too low a voltage to be useful. You will always need an adaptor to get from mains voltage to something useful for devices. There is nothing wrong with AC for mains, and the ease of transformers is not to be ignored.

Current voltage AC is still ideal for powering a house '

If USB PD were used, you would have up to 48V (240W) available, which should cover all but the largest appliances.

48 volts is not enough to cover a house without very large and expensive wires. And you still need to get higher voltage from the street to the house.

My fridge can do the same. It’s rarely actually compressing

Some new fridges seem to have much smalker compressors, which run all the time.

Their secret is variable speed compressors, advertised as "Inverter technology".

The compressor run continuously, it just changes speed so there is no noisy and inefficient start-up surge current.

Also common in heat=pumps, Air-conditioners, etc.

Didn’t utilities in NYC offer DC service for a long time? See https://archive.nytimes.com/cityroom.blogs.nytimes.com/2007/...

Old buildings had elevators, lights, and so on, that operated on DC.

Yes, that was edisons original system.

i could easily see USB 4 Power Delivery Extend Power Delivery and its successor become a de-facto standard. Many electrical outlets are already coming with usb ports nect to standard AC outlets. once that becomes standard on all new outlets you will see many more devices take advantage of that using USB for power then. after that you will see more demand for even higher voltage in future versions of the USB standard. once it reaches a tipping point it will make since to run DC throughout houses and instead of converting from AC to DC for the USB ports it will be reversed converting to ac for the legacy outlet next to the USB power ports

USB is too flimsy and awkward. A power profile negotiation using existing power sockets would be better.

Low-power 12V or 24V DC is very doable around the house, for small-scale LED lighting, etc. But if you limit yourself to small connectors and thus reasonably low currents (like 10A, as I did), it's 120W or 240W per connector at the very top.

Serious power, in the kilowatt range, would require much higher voltage, and still much bulkier connectors and thicker conductors. OTOH 400V DC should be comparable in this regard with 220V AC, which has a 380V amplitude. The 400V DC standard is relatively widespread and well-supported by existing industrial equipment. Connectors are comparable in size to the (grounded) Euro plugs.

Certain things get a lot more complicated when you go from AC to DC. Two of the most significant issues for home use are connects that can be disconnected under load (arcing is a serious issue here) and turning DC power on/off with mechanical switches (arcing is also the issue here).

Industrial DC devices get around these issues with stricter rules and bulkier/more complex devices. E.g. electric cars use contactors to connect batteries rather than a simpler relay/switch.

This is indeed so.

But the standard clearance for air gaps is 1 mm / kV, and arching distance is below 0.4 mm / kV. More advanced and faster mechanical switches should work fine at 400V DC.

Also IGBTs may be a preferred way, especially with various smart switches.

Sockets and plugs seem to be the hardest part to make cheap and reliable. But industrial solutions can likely be adapted.

I have no idea what the average household load looks like.

I could imagine a scenario where every outlet also has a USB port. It could run your laptop and your lights. Obviously it can't run a fridge or a stove; maybe future fridges would run off the same current as the stove?

What I'm most unclear on is the middle range. I know you can't run a hair dryer or microwave that way. What about a food processor? Maybe a toaster?

I'm trying to figure out how much that middle range makes this a deal breaker.

I wonder what kind of onsite loads they could have to use extra capacity? Is there any high power industrial process doesn't require too much space, setup cost, or transport infrastructure for materials in and out, and can run when there's sun available?

Maybe carbon capture, if that's a thing yet?

This is the idea behind Terraform Industries: https://terraformindustries.com/

They're capturing carbon to turn into a useful product (fossil fuels) though.

That website--1990s text-only style--is a strong statement. I am not sure if it's a statement that "we're going to take over the world, get on board now" or "in hindsight, we'll say we should have spent more on marketing," but it does get one's attention.

I thought it was about not wasting bytes on unnecessary stuff, but do they load shittons of JS extra marketing tracking analytics or not is the real question.

Brutalism has been back for a while on webdev

The footer:

> "Why does our website look like this? At TI we believe we can change the world by displacing fossil hydrocarbon production at global scale. Like our website, our machines are simple so we can build millions of them as quickly as possible. Our website embodies our cultural commitment to allocating resources where they solve the most important problems."

It's quite rich to self-refer as "TI". I feel like Texas Instruments has claimed that.

desalinization (does require transport infra, but mostly steady state)

DC-to-DC EV fast charging

Optics coating

Carbon capture is just like plastic recycling, it doesn't actually work at scale and it would be best to just not produce that much carbon in the first place. We should only be using as little fossil fuels as possible to replace the grid with clean/nuclear power and batteries.

I think H2 production with electrolisys would be such a process to store energy.

One could also make hot water for their own DHW use or charge their electric car.

Reply on news.ycombinator.com