Why a Typical Home Solar Setup Does Not Work with the Grid Down
syonyk.blogspot.com
syonyk.blogspot.com
I built a 13 kW ground-mount system feeding a pair of SMA inverters. I have tested this feature by disconnecting from the grid and enabling the outlets (one per inverter). I didn't quite get to the 2,000 W rating SMA claims but got close. Which means that with this size of an array and two inverters I get somewhere between 3kW and 4kW of power to run various devices while the sun is up.
Considering that we might have a couple of power outages a year on average (if that), I felt this was a reasonable investment. Going with batteries is just too expensive and not justifiable at all given the reliability of the grid. One way to think about this is that the grid is your battery. A stretch, I know.
Funny that there's a picture of a gasoline generator towards the end of the article. My guess is that I am likely to invest in a 5 kW to 6 kW generator before I ever add batteries to this system. Again, it's a matter of ROI. Also, I would not go with a gasoline powered generator at all. Gasoline degrades with time and could be a nightmare to maintain the system with sporadic use. I think a propane fueled generator might be a better idea. The fuel does not degrade. So long as you don't have leaks it'll be there ready to go when you need it.
I know way too many people who have been mercilessly duped by these solar companies who come in, hook them on some kind of a lease, install inadequate systems and move on to the next victim. Lack of understanding on the side of consumers has created a situation where solar is equivalent to magic and unscrupulous actors can take advantage of them. That part is sad.
This is wrong and dangerous. Suddenly energising a section of grid which has been isolated from all known power sources could easily kill someone working on it. Yes, lineworkers will isolate and test lines before touching them. No, this won't protect them because your inverter could start outputting after they've tested. Yes, your inverter will probably overload and trip out before putting too much voltage into your local section of the grid, depending on where it's isolated. No, it's not OK to bet a random stranger's life on it.
A thing may be technically correct, but for the health and safety of the public at large we need to adhere to a consistent policy with what we say in the public sphere.
Mains voltage is dangerous and kills kills kills. Never connect anything to the mains power supply unless you are licensed to do so, or employ someone who is licensed to do so. Always assume a circuit is live unless you-personally have checked and locked out the breaker with your own personal lockout device that only you have the key to.
I'm being silly, but you are going on about how the wording of statements matters.
Pointing out "problems with the phrasing" doesn't make for very substantive discussions. Also, snark is deprecated here. Please see https://news.ycombinator.com/newsguidelines.html.
Albeit the thread didn't invite substantive discussion on it because of the snark.
And if you plugged in your TV yourself, and they find out, you just unplug it immediately and wait for the licensed union electrician to come plug it in for you again.
- move furniture that didn't have wheels (so no moving of desks, side-tables, or even some of the larger waiting-room-style chairs). We were allowed to install wheels on furniture ourselves and then move it however :)
- hang pictures or anything on the wall (except when using that blue tacky stuff; you'd be surprised what you can hang with massive quantities of tacky)
- plug in a new power-strip, surge-protector, or extension-cord (we quickly learned to get the big 20-outlet strips)
- set up any kind of networking equipment, even if it was air-gapped from the university's network (we violated this all the time)
- plug in anything that drew more than a certain number of amps (an amperage level that was quite close to what many tower desktop computers were drawing at the time; we always just risked that)
Union workers were always prompt and polite when we'd ask them to do these things and they always had a "ya we know it's dumb" attitude about these requests. One time I saw the bill for the services: around $50 to plug in a surge-protector (that we provided) and $100 to nail a framed picture into the wall. The fine for not getting the union to do these jobs was something like 10x the cost of the labor iirc.
On the bright side, at least you know your electricians weren't being exploited :)
Well, actually there's a logical reason why only the union workers were allowed to do the mentioned things (tl;dr: company/university wants to shield themselves from liability):
- move furniture: shield the university from injury claims in case something goes wrong. For example, consider you moving a yuge chair around wearing flip-flops and you accidentally crush your toe during lifting. The university can now deny your claim as you were not supposed to have done this. In the crazy US system with its even crazier damage payouts, I wonder why this isn't commonplace.
- hang stuff on walls: the older the building, the bigger the chance that you'll hit some live wire, a water pipe or whatever is buried in the wall. Plus, again, personal injury claim risk.
- set up networking equipment: all it takes to bring the network to a grinding halt is accidentally connecting that router you planned to use for playing Counterstrike to the university network and whoops, where's that extra DHCP daemon coming from?
- amperage limit for devices: older circuit breakers tend to get a bit... trigger happy with age. Which means what will work fine now may be too much (especially surge) load in a couple of years and these issues are tricky to debug. In addition, I have seen a wild variation of power cables and extension cords when it comes to their current capacity (e.g. there exist power cords with 3x 0.75mm2 wire, these are rated for 2.5A tops aka something around 550W - but what if the component uses 800W?). Having someone trained look at all involved components before plugging them in can and will prevent fires. Not only to check if they're safe from an amperage limit, but also to check if their isolation is broken, there are scarred areas from arcing, ...
RE the wall they also said that nails and tape were also damaging to the paint (and repainting the walls was another union's thing), so the wall-hanging guy would have to log the "damage" to the wall caused by the hung item, presumably for cost/accounting purposes when it came time to repaint the wall.
That CYA stuff is actually typical for any company that ever gets hit with a personal injury or other expensive lawsuit. It's pervasive in government/bigco's because they had to deal with decades of every imaginable stuff happening.
If you want to get rid of it, you gotta limit personal injury liability (which would border on being inhumane) or develop a decent social safety net with proper insurances...
This is important: big orgs—government and private—not only have more chance of having an incident (because they deal with more events because of their scale), they are often bigger litigation targets when an event occurs, because they are giant bags of money.
There are plenty of people that don't understand(and probably don't want to understand) how things are connected. I doubt they think that the light switch magically turns on the light, or that the water magically comes from the valve attached to the wall, but the method, or desire to know the method, in which those things happen is just not in their world.
Nope. There's a type of wire called "Stegleitung", essentially a flat cable. Often enough it's simply nailed into the wall, then a thin coat of mortar is applied on the wall, and then paint.
Source: zapped myself once by hitting such a cable.
In addition, the usual injury liability is a problem.
Distrust—usually well earned—of management is the main way. The incidental items that are mocked are not the issue, the issue is that without an inflexible blanket rule, management will bend any other rule to use people in nominally different job duties to replace union labor.
Also, sometimes, the rules are negotiated in part by unions other than those doing the work, to prevent their covered staff from being compelled to do other kinds of work (and selected against for inability to do so: if lifting heavy items is part of the job duties of, say, someone in a clerical union, then inability to do that well can be used as a hiring, promotional, or termination consideration for such an employee.)
Even in places (as sometimes happens) where orgabized labor and management have good day-to-day relations, there usually is a feeling—with good historical reasons behind it—that the foundation of that good relationship is a “good fences make good neighbors” style solid set of baseline rules around which, where it is important (and the little stuff that is fun to mock isn't that) negotiated exceptions can be made.
"Do not rest your tongue on the plug as you insert it into the wall."
I get that electricity is dangerous, but are there really functioning adults hired by a company that can't be trusted to plug in a simple appliance without a safety briefing first? It seems insulting to their intelligence. Or do they make everybody do this so the one or two absolute morons don't feel bad about being forced to take the "completely basic life skills" course before they're allowed to function at even a very low adult level?
Yes, it’s meant to say something like us unlicensed folk aren’t permitted to hardwire anything to the mains supply.
In Australia you’re not even supposed to replace an existing power point outlet thingo unlicensed. If course, the hardware stores sell them, and we do. But there ya go.
In fact you're not allowed to install any fixed wiring without an electrical license. It's ridiculous - technically you aren't even allowed to install Ethernet outlets in your own house because it counts as "fixed wiring". Other trades have just as bad a grip on legislature, you're technically not allowed to replace your own tap seals either.
> Of course, the hardware stores sell them, and we do. But there ya go.
Of course, in theory there's no difference between theory and practice... ;)
In Victoria for example, you're not allowed to do any work on your own home if it exceeds a certain monetary value without first running through a Kafkaesque bureaucratic process to obtain a certificate.
This doesn't sound unreasonable until you realize that they don't care if you have any qualifications related to the work and don't even want to know what the work is. They simply want to confirm that you haven't done any other work (on your own property) in the last five years.
If you have, sorry, need to hire a registered builder. This prevents anyone from legally buying a home, renovating, and selling it, or even doing any significant work on your own property more than once in five years.
And if you want any other tradesperson to come to assist for anything while your doing it (installing a power point for example), then you need in person occupational health and safety training at a "registered training institute".
You only need to apply to be an owner-builder if you are doing your own work instead of hiring someone who is already permitted to do the work. This is no different from the US practice of having licensed contractors/electricians carrying out (or signing off on) large projects.
You only need to apply to be an owner-builder if you are doing your own work instead of hiring someone who is already permitted to do the work.
That is precisely what we are discussing. The restrictions on doing work on your own house yourself.
And this applies to any work at at, not just work that requires licensing (plumbing, electrical etc.). Even if you're doing the overall work yourself as an owner builder, you still need to hire licensed professionals for those tasks.
So the difference between "electrician installed" and "mysterious, unknown person installed" is only in the finished product.
I'm not an electrician so correct me if I'm wrong, but for those unfamiliar with residential wiring in US homes, ordinarily homes are supplied with two out of phase 110VAC lines. Heavy appliances like electric clothes driers and electric car charging points use the 220VAC that I available between the two "hot" lines. Everything else in the house use either one of the "hot" lines (black wires according to the national electrical code). There are actually 3 wires going to each outlet, light, etc. though the house: one black (belonging to one of the two sources of 110VAC), one white (called the neutral line, approximately 0V above ground, and one green a separate green wire that is connected to ground around the point where power enters the house.
Fundamentally, the power grid is based on three phase power. Typically, the average telephone pole will have four power wires on it, and one or two thick bundles of phone line or maybe cable TV, I dunno. Three of those four power lines are the three phase electricity, and the fourth is the common neutral line that is ideally near 0V, but typically fluctuates a bit.
Only three out of those four lines enter your house. One is is the neutral line, the other two are just two of the three power lines. Those two power lines are not 180 degrees out of phase, they are 120 degrees out of phase. If they were 180 degrees out of phase, the math is simple: it's obviously just 240V. Since it's 120 degrees out of phase, you end up with a sine wave that looks something like 120\(sin(x)-sin(x+pi\2/3)). (only you need to adjust it for frequency) If you're better at your trig identities than I am, you can calculate that out exactly, but for simpletons like me I just plug it into my calculator and find a local maximum of 207.846 something.
(note: In the US, we use neutral and ground to refer to the common center of the three phase power, and the wire that is buried in the ground near your house, respectively. In the UK, I think they call these two wires ground and earth, respectively. Note that the term "ground" has a very different meaning (read: one of them can kill you, the other one keeps you safe) depending on the context.)
Two of the phases DO NOT run to the house, a typical US household service is split-phase. The primary winding in the transformer at the pole connects between a single phase and ground, and the secondary winding is center-tapped to provide a 'neutral', with 120v on either side of the neutral. The voltages are indeed 180 degrees out of phase.
Your test would show 220-240v in the vast majority of US-based residential situations.
The 120/208 you are referring to is when all 3 phases are fed in to a WYE transformer that has 3 120v secodary windings fed from a center point, which becomes the neutral. With a WYE transformer, each phase gets 120v to ground/neutral, and the phase-to-phase voltage indeed peaks at 208v due to the 120 degree rotation between phases. This is typically only found in large residential(600+ amps), and medium to large commercial and industrial buildings with 3-phase service.
Another method would be using a 'Delta' transformer, in which one of the phase-to-phase windings are center-tapped, similar but a little different than the single-phase residential service. The two phases on either side of that center-tap will be 120v to neutral, but the 3rd phase will be 208v to netural/ground. This is usually called the 'wild' leg and is often used to power lighting loads.
Source: I've been doing this for a long time.
Also https://www.doityourself.com/forum/electrical-ac-dc/549775-l...
And https://ctlsys.com/support/electrical_service_types_and_volt...
The few houses that I was involved in wiring of in the states were definitely "split phase" systems, which work like the GP describes. In those, three current-carrying wires come from the transformer - two ends and a centre tap from one winding. Those are wired in the breaker box so that the house has two circuits that are 180* off from each other, giving 110V on either circuit, or 220V between the two "hot" legs. There's a better description at https://en.wikipedia.org/wiki/Split-phase_electric_power.
The current for the same wattage will be roughly double in North America. So lamps, etc. intended for use in Europe will have wiring that is undersized for the current draw if used in North America.
We found exactly that, lamp cords and extension cords, when we started our remodel a month ago. The former owner did some renovation ten years ago, and was an architect so we always assumed he did a decent job. That was one of a few questionable choices he made, so we're now re-wiring the entire house.
An organisation called TechSafe[1] is charged with carrying out random inspections.
Everything thinks they're smart, they can do things safely, etc., while only for a much smaller subset of people is that true.
Generally, playing with voltages above 50V is something that should be done with care and a professional/adult present.
Why should everybody have to pay a continuing overhead simply so people with small children don't have to install outlet covers?
If you're actually concerned with electrical safety, you're much better off installing solid "commercial-grade" receptacles than pretending that garbage-grade ("residential") has been rendered safe by being more difficult to use.
[1] https://ae01.alicdn.com/kf/HTB1Cya9KVXXXXXoXpXXq6xXFXXXK/202...
Here in the US we did change our plugs, by introducing polarized plugs. We don't have to be change-averse.
I still occasionally see an old switchboard with ceramic wire fuses with the fuse wire replaced with a bit of coat hanger wire, or the correct fuse wire wrapped around the terminals ten times. They often seem to be the same houses that give you a tingle if you touch any of the metal plumbing fixtures. Yikes.
Hardware store still carry traditional outlets as well.
The issue is exposed live voltage when you plug it in. Schuko and Europlugs generally don't allow this (even the flat plug is designed such that the contact aren't exposed anymore when they make contact)
Any time I see them working directly on high-voltage lines, they connect a device that looks like jumper cables in-between the phase(s) and ground. They also put bright orange flags on this device so that it is easily visible to anyone nearby. It seems to be a great practice and seems to make things nearly fool-proof because this effectively shorts out any power that may occur on those lines.
I don't think I've ever seen a lineman work on a power line in a manner that would cause him to be electrocuted if surprise power were to occur on the line. I'd bet it's in policies and procedures somewhere, possibly even in OSHA requirements.
The requirement of not backfeeding is another layer in this multi-layer protection scheme.
If a system is successfully backfeeding the grid, something is badly wrong. If a linemen is then electrocuted by this backfeed, something is almost impossibly wrong.
It's unreasonable to assume that a lineman can identify which sections of a downed line are isolated before they get to work repairing the power lines. There are circumstances where the line might even be isolated without being physically disconnected because a fuse blew downstream of where the line was grounded before the lines actually came down. Fuses will have a visual indication that they are blown but it's just a pin that sticks out the bottom and it's easy to miss when you think the only problem is a downed power line. There's an army of fools who know just enough to be dangerous and it's easy to see this on YouTube. Here's a great example of what linemen have to worry about when repairing downed lines.
https://www.youtube.com/watch?v=KiByz1E4IBU
Wouldn't you be worried working with downed lines not knowing if that guy just so happened to live across the street?
Not doing so is putting an awful lot of trust into a potentially large number of people that you've never even met.
Because you just don't know what some idiot has done...like the guy in the video.
That completely changed my perspective on all those quite lines dancing in the wind.
I cannot think of a time that I saw a lineman working on a power line that was not shorted, without protective gear.
Working on a hot wire is a whole different story because every thing around you becomes a hazard.
I do work on towers, sometimes in hot-RF environments (TV and radio antennas transmit with up to ~100,000 watts). The antennas will literally cook(like a microwave oven does) you if you get too close while they are live. For lower power situations, you can even hold your hand near the antenna -- it feels like a heat-lamp. Even at reduced levels, it can be harmful to areas of your body with poor circulation and cooling(such as your eyes), often causing damage before you're even aware of it. We wear an RF monitor when in those situations that alerts us if the RF field becomes stronger than is safe. I wonder if linemen wear a similar device that can alert them if an EM field becomes stronger than expected?
If I worked on wires where I was not in control of incoming power, then I can not be sure the wire isn’t hot just because I measured once. This is for example the case when working with stuff with big capacitors.
So my point is: can line workers really assume a wire isn’t hot just because they tested it? They don’t control both ends, since as you pointed out, anyone could have a) a power source or b) a faulty connection, whether it’s allowed or not. So as a line worker I’d basically just not touch anything even in supposedly dead grid sections.
Used on electrical circuits as well as dangerous machines if you need to climb inside a compactor for maintenance or something.
https://www.osha.gov/Publications/3120.html
http://www.deenergize.com/?target=Universal_Circuit_Breaker_...
Locks on the breaker itself are more rare (usually for low voltage repairs where the common person might frequent the area)
The lineworkers I've seen that aren't working with live wires use a special tool to ground the wires (it's a set of three clamps connected by a thick wire to a grounding rod). Once the wire's grounded, it's going to be hard to put any voltage on it.
The obviously missing third top line is the one that broke. It simply snapped around the middle of the span between that pole and the pole that is outside the photo on the right. They went up in the cherry picker, cut the dangling line, and then put on those clamps.
They then covered the lower arms of the pole with orange sheets: https://i.imgur.com/zZ3DFSd.jpg
Does anyone happen to know what the orange sheets are for? They did not use them anywhere else where they grounded the wires.
The place where they put these orange sheets in your photo looks like a disconnect switch between the wires on either side of the pole. So the wires coming from the left side of your photo are still live, and the disconnect switch itself is still live. Since they would be working near the disconnect switch (to hang the new wire), they put these insulating orange covers for extra protection.
I don’t know if the workers make a habit of physically disconnecting all of the houses that are connected.
For this reason, I'd be most surprised if there's a single backfeed system in existence which doesn't isolate your house from the grid once the grid goes down.
Edit: Finally got around to reading the blog post to the end and found that the author makes this exact point in the paragraph following the one quoted in the comment above.
> If you have a typical grid tied system (microinverters or normal string inverters, so easily 95+% of installed rooftop solar), the system is technically incapable of running off grid (without additional hardware). There's no waveform to sync with,
A residential no-break has no waveform to sync with as well.
Something capable of syncing to the grid and then more or less keeping pace even if the main grid goes down should cost very little today. (And when the grid goes back it shouldn't have drifted too much unless something ridiculously big happened)
Only running at a fraction of their max available output allow enough headroom for high peak startup draw from loads, and headroom for clouds and planes passing overhead.
Frequently cutting in and out as the load regularly exceeds available supply.
Keeping the frequency sync is the easiest problem to solve, the article even covers what would happen if you tried to use a little generator to produce a sync signal.
Most solar dealers have a one-size-fits-all product that they fit to nearly all homes without much modification. The homeowner thinks they are buying something that will save them a fortune and eliminate their need for grid power. In reality they discoverer that solar is persnickety and they will get nowhere near eliminating their grid reliance. And when you do the math you realise that it might take 10+ years to recoup the high cost of installation and by that point your batteries will need to be replaced and your solar panels will have lost some of their efficiency.
I know people who have gone fully off-grid in Ireland, but they don't just rely on solar. They supplement with wind turbines and in some case hydro power from streams.
Rooftop solar is popular because the payback period is short enough for homeowners - well under 10 years here for a system that is built to last at least 20. Modules built now degrade about 0.25% per year. Solar farms built now are typically financed as 25-30 year projects.
Source: I am a data analyst at a solar engineering firm.
Do you have any suggestions or recommendations on things to read or companies to look at? i.e. Tesla power?
PS: Look at taxes for Farm use vehicles fuel to get a better picture of the actual 'subsidy' vs 'tax'. The tipping point is very much a subsidy. https://www.irs.gov/businesses/small-businesses-self-employe...
As you said:
> Keeping the frequency sync is the easiest problem to solve
When the grid comes back from a blackout, chances are that it browned out beforehand so your sync is to a low frequency and coming back it'll be high frequency because the grid wants to compensate for loads jumping back on.
Additionally the components to generate your own waveform are cheap, yes, but not that cheap, adding them to the microinverter would increase cost quite a bit.
And you'd still need a transfer switch because if you happen to be 180 degrees out of phase, which CAN HAPPEN then your panel will behave like a dead short at double grid voltage. The current flow will definitely exceed the maximum tolerances and the magic smoke goes out.
You will absolutely need a transfer switch just so you don't fry all your devices the moment the grid comes back. Even then, syncing to the grid is a rather delicate maneuver since the grid will be constantly changing phase and it'll be simpler to shut down all inverters, connect back and have it all run back up on the grid itself.
It's called a line-interactive UPS. You can buy it for $200.
A line interactive UPS does not sync necessarily sync you to the grid, though the electronic will usually try to keep it in sync. It's not necessary here.
Not really, but depends on the state of the infrastructure, if it was really because of a power overload, yes, but most likely "your circuit" (which could be your street or your neighborhood) got shut off, in this case there shouldn't be much difference
Yes, a phase difference of 180 can definitely happen but I guess most electronics can survive a 1/60s (I'd say even 1/10) switching time, which is probably enough to have the grid take over.
(Or of course you could have the grid and solar charging batteries then your own high power inverter for your house but that of course would mean $$$$$$)
And you'd still have to sync the inverter, having the inverter simply continue to run until it's back in sync with the grid, then just reconnect (as a previous comment wanted) is likely not an option for most consumers.
For that it would likely be cheaper to have a full DC stage as you mentioned.
Just shut it down, flip the switch and restart. Everything else will just be prohibitively expensive because it needs to be very safe.
If you get the phase wrong then you'll either reduce the lifetime of your components or the components explode after the nearest power plant tries to pump all available power into your poor inverter.
Only if you design it to. The phase locked loop would "listen" to the mains frequency and slew to match. Slew speed is simply a design parameter you can set to any value.
I’m pretty sure companies like Victron Energy already make these kinds of systems - combination solar inverters and battery chargers that have transfer switches to be able to seamlessly switch to UPS mode when the grid drops out but can still export excess energy when it’s up.
And that's for what you need to allow your inverter to handle this automatically (you might need another voltage sensing channel to sense the grid-side of this breaker).
Your UPS handles an order of magnitude (or more) less power than a whole-home solar installation.
None of these problems are intractable of course, but you are oversimplifying the problem a little.
One can get a UPS affordably that will power 1500 watts of continuous power. Being able to supply just that much, or twice that much, from solar panels in a grid-down scenario would be tremendously useful, even if it's not enough power to fire up my welder.
When the two are aligned within a good-enough tolerance the system will switch back to it's regular state of mains + solar ( + batteries + wind + generator + etc / whatever).
These are all solved problems with commercial off the shelf components.
It looks to me like microinverters are a commercial off the shelf product?
Will a microinverter do all of the things a multi-component phase-syncing system with automatic transfer switches do? I don't see why a microinverter can't be built with these components integrated. I can't tell you if such a unit exists as I'm not well versed in the product range.
As far as a price comparison goes, I guess it only makes sense to compare a like-for-like system?
Could it be made cheaper? Probably. If you get it wrong, the grid probably doesn't care but you'll briefly pump about 500W into the device that is supposed to have 500W going out of it. The reason these are big and expensive is that it requires significant safety gear so nothing explodes even in the worst case. And that safety gear is expensive. So you sell it to people who not only can afford it but also really really need it (ie, 1MW and upwards where you enter the domain of "can fry small section of grid")
The specialist devices for large installations you’re talking about are only expensive because you need more expensive parts for the far larger amounts of current you’re handling (and probably because they’re made in lower volumes than commodity inverters), not because they’re doing anything particularly difficult.
There’s no technical reason the requisite electronics can’t be built on a much smaller scale.
Home grid-tie solar inverters are clearly capable of syncing, so the electronics are already present.
Additionally producing the clean sinewave that you'd need for this is not that easy, atleast not at the quality levels you want for this (if your DAC that produces the wave is off by 1% then at a 2kW load you're going to burn up 20W somewhere that doesn't like 20W being burned up)
How is being off by 1% going to hurt anything? I'm quite sure my mains voltage is already more than 1% off.
If I get solar it had better not explode every time my air conditioner kicks on.
However, if you have your own generator it matters a lot.
If your phase is off by 1 degree then that 1 degree will burn roughly .2% of the incoming power of the grid at the inverter (which is unlikely designed to handle this). If you're off by 1% you burn 20 Watts on a device not designed for it.
If your voltage is off relative to the grid by 1% then you burn the difference, at 2kW that's about 20 Watts. And that's per volt. You'd be burning somewhere around 300 W if you happen to have the grid on the higher end of the tolerance and yours on the lower.
A grid-tie inverter gets around this by simply following along the sine wave of the grid, this can be done relatively cheaply and safely with analog components so the error can be much smaller than 1% and deep into random noise territory.
If you generate your own sine wave and compare it to an existing one it's much more difficult since you have to match amplitude and phase almost perfectly.
So with grid-tie nothing will explode. With an autotransfer nothing explodes either. Wanting to seamlessly couple back in requires a lot of care and expensive components.
Next to your offline-capable sine wave generator, put a copy of that cheap safe analog circuit.
Once you get almost in sync, crossfade over to the analog circuit. (If that's even necessary. It might be just fine to swap to it at a zero-crossing.)
Now you're completely in sync. Reconnect to the mains.
The cheapest is VFD which is basically a battery parellel to the mains which in case of a power failure interrupts mains and inserts it's own voltage. Usually labelled as "offline" or "standby" UPS since they're not active most of the time. The output frequency and voltage is the mains output and voltage until switched over, something to keep in mind if devices are sensitive to that.
These can simply switch back to mains when it's back since they usually use a simple transfer relay.
VI (Line interactive, Delta Conversion) uses the mains frequency as orientation. They don't have a transfer and can basically just compensate whatever the mains is doing to output a 230V signal. Internally they have a inverter with AC input and AC ouput which means they measure if mains is coming back from there and adapt the signal on the internal inverter for the battery.
If mains comes back on a VI they usually change frequency very abruptly which is not ideal from some devices.
VDI is completely independent of both voltage and frequency as it first converts mains AC to internal DC, simply plugs in the battery into the DC and then converts DC to AC. They don't need to synchronize at all and are the more common for datacenters since they isolate the input fairly well from output and don't have to switch anything to go from mains to backup, DC voltage is fairly good for dealing with this. They are also most expensive.
If mains comes back on a VDI they don#t do anything of notable interest other than switching the battery charger on.
Manufacturers of solar hardware love to charge you an extra $1000 for that extra firmware though...
I live in an off grid bus. My whole electrical system cost in the US$9-10000 range with only 810W of solar. The other components cost a lot more:
* ~$900 for 3x 270W Renogy solar panels
* $1500 for 3.6kWh of LiFePO4 batteries (12 100A 3.6V cells)
* $1600 for a Victron 3kW inverter / charger
* $2000 for a Honda EU3000is generator to run my AC
The rest was the Victron solar controller, color control panel (that runs Linux!), breakers, relays, transfer switch, fuses, cable, and all the other stuff you need to hook it together.
This isn't at house scale either. The cost would go up significantly to support a house scale system.
OTOH not having a combustion-based backup is likely not very wise if you live off-grid. I wonder how often do you need to run it? That is, what part of your energy budget is covered by the solar energy?
The generator is part of the whole power system. In most implementations solar is not reliable enough on its own. You need a backup. In my case that's the generator and the bus' alternator.
Right now I'm running the generator all day to power my AC. I'm not running it most nights. It could be cycled on and off but I haven't setup an auto start system for it yet.
Especially the batteries.
I didn't itemize the rest because I don't have the time to list it all. The roof rack and hardware cost around $1000. There are a lot of individual items that cost around $100, (eg starter battery isolator relay, breaker panel + breakers). Heavy gauge cables are necessary and expensive. All the small things add up.
a couple also made a solar powered van, but as people say, it's probably very short on mileage (basically move a bit and rest while it charges.
It's not ideal, but this is so much better than what most installs are built with (nothing).
It is possible to get inverter that will work independently, and batteries to buffer the load, but the off-grid system is a lot more expensive than normal rooftop solar system.
There are a few ways that anti-islanding systems can work, but it's reasonable to imagine that the inverter is measuring the impedance between its output and a nominal AC source: the power grid. If that impedance is too high - and it will be unless you've got a rather large generator - then the inverter will not source power.
What about just using batteries?
Typical auto transfer switches are fast enough that computer equipment etc won't notice.
Edited to add: if you want truly uninterruptible power then you should do what data centres do: everything runs off the UPS all the time (well, all the computers, maybe not all the cooling system), the only thing that changes is the supply: either mains, batteries, solar, or generator. Or some combination of batteries + the other three.
Back in the 00s I visited a DC that two very large flywheel UPS that were being trialed to replace the battery room.
Though I imagine eventually battery tech will take over.
I don’t know if anyone is building systems like that though?
PG&E had a transformer fire which caused large voltage swings. The flywheels kept the generators turning, and the ten Diesel engines started OK. Then outside power came back up, and after a time delay, the Diesels shut down. Then outside power failed again, and because of a timing incompatibility between the power control and engine control, some of the engines would not restart that soon. There was a minimum "off" time on the engine; it had to stop before it could be restarted.
The engines that did start were not enough to handle the load; they overloaded and their controllers shut them down, shutting down the whole data center.
[1] https://www.datacenterknowledge.com/archives/2007/07/24/gene...
What's your source for the first part of this phrase?
Dave Jones has a good video on it (well series to a degree): https://www.youtube.com/watch?v=SoeVNaz7KGc
Magic smoke is injected into electronic components at the factory, in order to make them work. If you let the smoke escape the package, your part has failed permanently. But as long as the smoke remains inside, there is still a chance you can repair any failure by cooling the part off for a while.
Huh, where can I read about this?
1: https://www.almanacnews.com/news/2018/07/25/pge-may-cut-powe...
https://www.sfchronicle.com/business/article/For-PG-E-cuttin...
https://sacramento.cbslocal.com/2018/06/25/pge-cutting-power...
A couple links:
https://yubanet.com/regional/pge-shutting-off-power-during-h...
https://www.mercurynews.com/2018/06/12/pge-could-shut-off-po...
I have a Magnum 3000w Hybrid inverter in my RV with 320w of solar with a MPPT solar controller. The solar controller feeds the batteries and the inverter draws from batteries or charges them with UPS grade change over. Magnum hybrids can do interesting things like if the sun is out and I am running the air conditioner off generator or shore power it will pull batteries down to float voltage and invert excess power to lower draw, I usually get about 1 amp reduction with 320w of panels. You can also dial in your shore amps and if a load needs more the inverter will step in and provide the overage from the batteries. Nice if you driveway surfing and only have a 15 amp outlet but want to run the microwave or whatever. The RV is basically a giant UPS with generator.
But as one who is considering residential solar, I guess there’s more research to do as home solar apparently doesn’t work that way at all (though it probably will when I do it).
Side note: thanks for the inverter mention. That’s next on the list, and the Magnum sounds to be worth a look.
Magnum on the other hand has excellent documentation and a somewhat simpler product line that integrates well together, remote, inverter, battery monitor. Their comm bus is just RS-485 and documentation can be found, I wrote a node server to display info in realtime along with my Morningstar Tristar solar controller which uses modes over TCP.
My understanding is Victrons load support feature does not go down as far as Magnum, 10 amps vs 5 if thats important .They have a very good reputation as does Magnum, different league than cheap inverters.
Here is a video I did of the Magnums search watt function as some wanted to see, also shows how it deals with large loads and the web interface I wrote for it: https://www.youtube.com/watch?v=l_jqzY1wNDU
It doesn't seem like a battery or capacitor to handle blips like an airplane flying overhead would need to be very large?
But surely such events (i.e. grid power outages) are pretty rare. In Australia it's rare for the power to be out more than once a year.
Whenever your battery gets super discharged, picture this happening inside it: https://www.youtube.com/watch?v=r-YbQN_twpw
Notice how running the current the other way will break up the formed crystals, but it won't actually re-dissolve them. That's what it means for a battery to wear out. (Battery electrolytes are chosen specifically to be resistant to this, so the effect is mainly visible as a thin "rust" of deposited crystals on the battery's anode, rather than the full-scale crystals seen here.)
A similar but distinct process occurs when overcharging a battery, where, instead of splitting the electrolyte molecules apart, you're reacting and bonding them together to form new molecules or molecular complexes, with the reaction usually being one-way rather than an equilibrium reaction. (For lithium batteries this process is exothermic and catalyzed by the presence of the product—thus lithium battery explosions.)
It's much better for the life of the battery if it basically just hovers around 40-60% charge for its whole useful life, since then you're just generating tiny seed crystals (on discharge) and then re-dissolving them (on charge), where those crystals are small enough that they can be fully re-absorbed.
And this is true even in battery-cell technologies that require a "deep charge cycle" to erase their "battery memory." Battery memory is basically the electrolyte causing enough crystalline rust specifically on the anode to increase its resistance. A deep discharge can capture and erase this rust—but it still shortens the battery's lifespan, because you're still producing non-reabsorbable large crystals within the electrolyte.
To my knowledge, the best and safest battery system is to have each battery individually monitored by a charge controller with the capability to fully disconnect the battery at will.
The downside is that this will be expensive so people usually settle for just eating the rare chance of a dead short battery.
Additionally, batteries that you just leave around doing nothing will probably die at some point too. You don't want it sitting around you want to use it for efficiency or else you swap your battery and find out it was dead too.
To reduce burst loads on batteries you could use supercaps if you find ones that can handle the voltage and current (that will be very expensive).
This wouldn't happen if batteries were designed to act like nuclear reactors, where the "fuel" (the fuel rods; the electrolyte) can be completely removed from the substrate that makes it react (the neutron medium; the anode+cathode.)
But a battery that can withdraw its anode+cathode from the solution would be damned expensive. It'd make more sense as an architecture if you had just a few, super-large battery cells, e.g. giant vats of lead-acid.
It might be possible to design regular battery cells such that they wouldn't start degrading until they were first exposed to a voltage load, though. (I think the "50 year" Duracell NiCd batteries have this property—they probably have an antifuse oxide layer between the anode/cathode and the electrolyte, that gets broken down when you put load on the circuit.)
These would be kinda neat as off-grid generators and you can take the nuclear element out (and it would last longer but still be radioactive).
If you need 100Ah from your batteries you’re better off provisioning as many times more up to some cost-benefit analysis.
I know of some railway train setup, which is deployed in Saudi Arabia by Siemens, that uses a combination of battery and super caps to power the train. While the train starts it takes the sparking current from the super caps. While riding it will take the energy slowly from the batteries. While deceleration it will recuperate and will charge up the super caps, because they can be charged easily with a high current. So the super caps are used as a high available energy buffer with very fast charge and discharge capabilities. While the battery is only used during the times of less energy consumption. Also at the stations the super caps are super charged very quickly. With that concept, they can run the train the whole day with only 40% discharge of the battery. The battery is actually way smaller then in a Tesla.
This concept can be applied also locally. The problem: these super capacitors are way expansive so you need to design them to your specific requirements.
AFAIK, that should be possible.
Here is a presentation about the topic. The interesting information is in last pages. https://www.siemens.com/press/pool/de/events/2015/mobility/2...
I wouldn't want to ride a battery pack like that all day and every day, but it should be fine for emergencies.
If you use a small supercapacitor, which commonly has a minimum discharge time of about 2~20 seconds (similar to Lithium-ion pouch cells "LiPo", which are available with different "C" ratings, and negatively correlated gravimetric/volumetric energy density and maximum average discharge power (i.e., optimizing discharge current vs time to maximize the ratio of total power extracted divided by time until empty). You need more metal in the electrodes and the plate/foil that aggregates the current of the individual electrodes, leaving less space/weight for the parts that actually store energy.), you can use much more advanced power reduction techniques for electronic devices, as you have time for something approaching a proper shutdown. Compare e.g. the time your Laptop takes until it is off from the moment you close the lid.
Oh, and yes, these capacitors are cheap. One about 1kg / 1 liter size for about 100$ can provide 20kW average power, for 2 seconds. The main difference is that they don't care (much) if you cycle them at 100% depth of charge, apart from internal losses heating it up and thereby causing damage.
What do people actually use?
My ideal situation would be a microgrid with 10-20 homes (in a compound/neighborhood or condo building) and some commercial/office, which pushes all the electronics costs down to something more reasonable.
That said, I've started seeing lithium batteries slowly catch on in amateur radio as the price on the LiFePo4 cells go down. Especially for portable setups.
Does anyone know, what's a good resource for learning more about power generation, grid regulation, etc.? The relationship between frequency/voltage and available vs. demanded current, frequency sync, grid maintenance, etc.…
Electrical Machines, Drives and Power Systems by Theordore Wildi
So, your example of energy options other than solar is solar with hydro storage? Or did you mean “other than photovoltaic” when you said “othet than solar”.
I don't remember seeing any articles that blame the power company for home solar not working off-grid... I do remember seeing articles saying that most home solar installations won't work if the power grid is down, but nothing that implied that the power companies were behind it.
I'm glad I finally understand why because all that I'd seen reported was the solar systems installed must be turned off. I'd always assumed all these systems (many of which are meant to be used in a storm + lower electricity costs) had a transfer switch.
Like a sibling commented here, the costs would be enormous to convert an AC grid into a DC grid. AFAIK are there two reasons that the grid is AC: generators produce AC, and AC is so easy to transform up and down. That transforming is essential since you want to transport energy at high voltages [2] but you don't want 380kV in someone's house.
[1]: https://en.wikipedia.org/wiki/High-voltage_direct_current [2]: https://en.wikipedia.org/wiki/Electric_power_transmission#Ad...
They’re frequently called dynamos and you find them on old bikes to power the lights.
With modern electronics doing DC-DC transforming is pretty easy (because modern electronics is basically magic). And the reason why high voltage DC transmission is a thing, is because you can get way more power down the same wire with DC compared to AC.
Many high voltage DC lines used to be high voltage AC that was converted to increase the transmission capacity without building more cables.
It wouldn't work. The needed wires would be like garden hoses.
Might not work for everything (eg, refrigerator? Microwave?) but it could be worthwhile to run a secondary DC network.
I think there are some issues on longer lines with things like inductance, cable size requirements, etc as well.
In hot climates you have some power use when the sun is up (cooling, pool heating,) but where I live I’d need my power most when the sun is absent in night and winter (power use is probably 80% heating). Any amount of solar power would be nearly useless if only available when the sun is shining.
Not a subject matter expert by any means but depending on how much sun you get, there are probably better options than solar for you (solar might not be the right choice for you).
I learned from the article that lower temps helps efficiency, so perhaps our climate with short days and subzero temps for 5-7 months is still viable because of higher efficiency?
At this temperature you could even dump waste heat of a computer (well, CPU and GPU, maybe DRAM, not the rest) into this thermal storage.
I built an audio amplifier in college that had a couple of huge capacitors on the power supply (about the size of a can of soda). It would keep the amp running a few seconds after I shut off the power.
One of those "1 Farad" car stereo caps (not the classiest example) can store 72 joules (watt-seconds). So to handle a 1kW excursion for 5 seconds, would take over 70 of them. Possible, yes. But probably not worthwhile.
The frustrating thing about this article is that it keeps conflating fundamental physical problems (like needing to store energy to deal with peaks), with the limitations of currently-productized circuit topologies. It's not hard to imagine an inverter that would do something similar to MPPT without needing to pull/dump the maximum power possible, would manage a small bit of local storage (a combination of caps and abused batteries), and could intelligently load shed (with "kind" brownouts) on multiple outlets.
It's also not hard to imagine an inverter (brushless motor) air conditioner that would be a lot kinder of a load, and could even operate at variable power.
But of course with the current state of homebuilding, these are bespoke ideas better filed as "off the grid". Which is the entire larger point - an "off the grid" system fundamentally costs more in materials and also presently application engineering time, than the on-grid systems being sold to homeowners.
(Although I do have to wonder about adding a piece to the current puzzle that would generate/manage a local grid frequency, and shed excess power using off-the-shelf electric heaters.)
This doesn't sound that bad to me. Size isn't much of an issue, houses have room for them. Prices for them seem all over the place:
https://www.alibaba.com/showroom/1-farad-capacitor.html https://www.amazon.com/dp/B00381UKD0/ref=psdc_2230642011_t2_...
The idea with a capacitor is they shouldn't degrade like batteries do.
The same with an inductor, which is just a coil of wire.
Electrolytics do wear out. Lifetime hours versus temperature/ripple current is a key design parameter.
[0] And assuming stuff designed for the car stereo market actually meets its advertised specs. I just went to Youtube to verify that said devices weren't just a can with other components inside of them. A video with a freehanded jigsaw set to hip hop confirmed.
As I was saying before, the main difference from your amplifier was the amount of power involved. In general, audio amplifiers have large capacitors on the power rails to smooth out the ripple voltage, and will coast for quite some time as they don't actually need that much power for typical music and volumes.
Here's a similar example - high current USB charging bricks, with a power indicator LED. If you have no devices connected and unplug it from the wall, the LED will stay on for quite some time. The LED just doesn't draw that much current (say 10mA) compared to the capacitance, which has been designed for smoothing the full charging current of 10A. Also the LED will happily ride the voltage down from 5V to its junction voltage of 2.5V or whatever, using even less current. And due to the logarithmic nature of your eyes (ears), you don't fully perceive that dimming.
I do know that caps decay over time, but have no idea how that compares with batteries. I expect, though, that long life caps could be built.
I also wonder about using inductors instead. They are, after all, just a coil of wire.
https://electronics.stackexchange.com/questions/42084/can-yo...
https://electronics.stackexchange.com/questions/326177/energ...
Perhaps in the future we'll have supercooled inductors providing power storage, and not requiring periodic changing the way say batteries do. Alas, it's just not right now.
As an aside, it's amazing how little power it actually takes to produce worthwhile sound out of modern speakers. I once wanted to test some tower speakers without a proper receiver around, so I wired up a breadboard with some stupid opamps around poorly-heatsunk transistors, powered by a lab supply. The speakers turned out to be disappointing, but not because of the amplifier!
There's also inverter drive microwaves, and fridges even. My friend's new condo has an inverter drive fridge
Electric motors, especially the simple ones found in compressors for cooling, are running from AC. You can't put a capacitor into the AC path. You can only convert it to DC, charge the capacitor and then convert it back to AC. Where we are back to the original problem.
The output of solar panels is DC, too. The capacitors would be placed in the DC side.
No TV, no Washer/Dryer (you will hand wash), No regular fridge, get a small 12v fridge. No microwave, use stove to heat up food. No A/C, design the space to not get hot, high windows, under the shade, facing away from the sun. Maybe a small fan. The biggest optimization begins with your appliances.
Lots of homebrew 'you can plug this magick cable I made' methods to supply volts across the fence to neighbours houses when the power goes down are really bad: behind the meter is just as lethal as in front of the meter, if you get it wrong.
TL;DR if you need to read a website like this for clue how to avoid losing power, you are putting your life at risk if you don't follow code, and don't know what you're doing.
You have to read the specs carefully to make sure they are a good match for your use.
So.. "home" solar setup.. home.. like fridges, aircon and washing machines...
Do you see where this is going?
With some carefully selected appliances you would probably be fine. This stuff used to be super pricey, but now even the cheaper brands are putting out inverter drive ACs for example. Not window units, but mini splits and home systems
If you are a prepper, with your existing lifestyle and think a genset will give you tide-over-until-the-zombies-die, I think you need to think harder: a genny won't work, if you don't pre-design the load for it. Lots of things you own right now probably won't be very happy with unstable volts, and lots of things you own right now will suck hard on the generator, making it surge.
So yes. you can design for it. But a typical home solar setup (which is what the article was about) is not going to just work, offline, with your generator. It takes more work.
An off-grid capable inverter needs to handle several state transitions smoothly, and some of those are hard. For instance, starting up can be a lot harder when there's no mains voltage. An inverter will have a maximum output current, which is generally not too far from the maximum power output of the inverter divided by minimum mains voltage. If you have a load that presents as a constant resistance, like a water heater, then as the inverter's output voltage goes from 0 to 240V (120V for areas with weenie power) the current will smoothly go from nothing up to whatever the load normally draws, along with the voltage, and there's no problem. But! A lot of loads aren't resistive - things like computers, battery chargers, and inverter-driven motors (as newer appliances often have) will tend to look more like constant-power loads. So, once they decide it's OK to turn on, if the voltage happens to be half of nominal at that point, they will try to pull twice their nominal current. If that causes the total draw to exceed what the inverter can provide, it stops and maybe tries again. It's practically impossible to measure what the load characteristic of something like a house, from the perspective of an inverter, so the best you can do is something like guess-and-check, and that's not a great solution.
Then, how to handle reconnection with the grid; you either need to re-synchronise before reconnecting, or ensure that the intrepid homeowner first shuts off the house before turning the switch to reconnect.
Switches that can handle that job may need to be specified per-installation, and there are all sorts of regulations around how they work and where they can go. They are not cheap, as a rule. Some regions have the solar generation stuff on a separate export power meter from the import one. This means that the switch needs to be on he supply side of both meters, and that could be regulatorily complicated.
What about multiple inverters? These things all need to play nice with each other when solo, or installed with others. Perhaps you didn't design those, and they might try to start up too, then fire up their anti islanding which should promptly trip out again.
Remember this all has to work with variable input power available, and it needs to be packaged up so that installation is straightforward, and operation is basically hands-off. Oh, and inexpensive too!
[0] https://en.wikipedia.org/wiki/Pumped-storage_hydroelectricit...
The end part that basically says, 'solar panels need batteries and a more expensive inverter to run without the grid" needs to be at the beginning of the article.'
The details of power curves then come for those who want more details about why.
modern bench top power supplies can limit current
The obvious approach is a propane generator. Propane rather than gasoline because we only get a small number of outages a year, and most are fairly short, but there is occasionally one that is 10+ hours. I'd want to store enough fuel for at least 10 hours of generator use, but most years would only actually use a fraction of that, and so I'd have to be constantly rotating the fuel (probably into my car, as the only other gas powered thing I have, my lawn mower, only uses about a gallon a year) to keep it fresh enough to still work for the generator. Propane does not go bad, so storing 10 hours worth of propane for years is no problem.
But both of the things I want to power (fridge and pump) are very intermittent loads. I don't want the generator running all the time during an outage, wasting fuel, just to be ready when the fridge or pump needs to actually run. Even idling the generator will be consuming fuel. I could leave the generator off except when I need to actually have the fridge or well run, but that would be quite annoying as I'd have to go outside to start and stop the generator--and power failures are usually when it is quite unpleasant outside.
So what I've been thinking of is having a dedicated battery and inverter for the fridge, and a dedicated battery and inverter for the well pump.
When power fails, I can start watching the fridge temperature (it has a battery powered wireless temperature monitor, so I can see the temperature without opening the door). When it warms up to the high end of its normal variation, I can unplug it from the wall outlet, plug it into the inverter, and start the inverter running off the battery. When the fridge cools back down, turn off the inverter. Repeat throughout the outage.
Similar for well pump. When I run out of water, go run the pump from the inverter and battery until the tank is full.
I'd still have the generator, but now it would be used to recharge batteries, not power any house circuits or devices. Over time, I'd probably add batteries, eventually reaching a point where I have enough that I can get through most outages without needing the generator.
(Actually, I'd probably do the well directly from the generator instead of inverter and battery. When I go out to turn on the generator to recharge batteries, I'd also plug the well pump into the generator to refill the well).
I could start with one inverter and a battery or two, without a generator. This would probably get me through a large fraction of the outages. Then over time I could add more batteries, more inverters, and the generator to extend the number of things I can use during an outage and how long an outage I can get through, in a nice budget friendly way.
[1] I probably don't actually need the well pump--as soon as an outage starts I fill a couple water bottles to have drinking water, and fill the bath tub to have water for flushing the toilet, and that will almost always be good enough. Why immediately fill things instead of just leaving the water in the well tank and hot water heater until I actually need it? Because I have no idea how much water I have. The outage could have happened right after the pump finished filling the tank, or it could be just before the pump was going to start and fill a nearly empty tank. Filling the water bottles and bathtub right away lets me know what I've got.
Duh! Wtf do I want AC for? People think theyre all going to have a stupid fridge which opens from the side to let as much heat in as possible and with a highly inneficient compressor which is tucked behind it where the heat can only escape through it. Or a giant motor to shake their clothes in (electrically heated) water! Ahahah! I swear i keep telling people the KWh is going to cost 150-300 mg of gold (0.10 - 0.20 gold $) and their eyes just glaze over; they think it sounds nuts? If they had off grid power, they might be forced to agree. So the average person will be able to afford a couple or three KWh per month. Two low-powered computers and some lights. We light the whole place with less than 5w. The mere fact that we so often use measures like kilo watts is a good tell tale of the current power glut. The powergrid rate where we are is about 0.0015 gold $ per KWh, or 100 times below sensible prices.