Watts, kilowatt-hours, inverters and clipping: how a solar and battery system actually works

This is the distinction everything else rests on.

A 2 kW heater running for three hours uses 6 kWh. Change either number — a bigger heater, or a longer run — and the energy changes. Your electricity bill charges you for kWh, because that is what you consumed. Every piece of equipment on your roof and wall is rated in kW, because that is what it can do at any given instant.

Two consequences worth holding onto:

• A system with a big kW rating is not the same as a system that produces a lot of energy. A large array on a shaded east-facing roof can produce less than a smaller one facing north.
• A battery with a big kWh rating is not necessarily one that can run your house. That depends on its kW rating, which is a separate specification — see below.

A modern solar-and-battery installation carries four ratings that people routinely mix up. They are measured at different points, in different units, and each is limited by something different.

A quote that gives you only one of these has not told you what you are buying.

A panel labelled 440 W will produce 440 W under Standard Test Conditions: 1000 W/m² of light, a cell temperature of 25 °C, and a defined light spectrum. Those conditions exist in a laboratory. They do not exist on a roof in Sydney.

Array size is simply the panels added up. Thirty 440 W panels is a 13,200 W array — 13.2 kW DC. That figure is what small-scale technology certificates are calculated on, and it is the number most people quote when they say “I have a 13 kW system”.

What actually arrives at your switchboard is less, for reasons that are all normal:

Seeing about three-quarters of nameplate at the best moment of a clear day is a healthy system, not a faulty one. This is the single most common “my solar is underperforming” call, and the answer is usually arithmetic rather than a fault.

The inverter converts DC from the roof into AC your house and the grid can use. Its rating — 5 kW, 10 kW — is the maximum AC power it will ever deliver. Whatever the array is capable of, this is the ceiling.

Three things about that number that matter in practice:

• It is the number the network cares about. Connection approvals and export limits are written against inverter capacity, not against how many panels are on the roof.
• It is continuous, not peak. An inverter is designed to sit at its rating all day in summer heat. Some derate slightly at high ambient temperatures, which is a reason not to mount one in full afternoon sun.
• It has its own DC input limits. Maximum DC input power, maximum current per MPPT and a maximum input voltage all appear on the datasheet, and string design has to respect all of them. Array voltage peaks on the coldest clear morning, not the hottest afternoon — a string sized only for summer can exceed the inverter's window in July.

An MPPT — maximum power point tracker — is the circuit that continuously hunts for the voltage at which a string produces the most power. Most residential inverters have two. Each roof plane with a different orientation generally wants its own, which is why the number of MPPTs, not just the kW rating, decides whether a particular inverter suits a particular roof.

This is the question behind most of the others, and it usually arrives in a form that already contains the mistake: “my bill says I use 20 kilowatt-hours a day, so I need 20 kilowatts of solar.”

Those are different units. 20 kWh a day is energy over time; 20 kW is power at an instant. To turn one into the other you need to know how much energy a kilowatt of panels produces in a day — and that is a number you can carry in your head.

So the 20 kWh-a-day household needs about 20 ÷ 4.5 = 4.4 kW of panels to generate as much energy across a year as it consumes — not 20 kW. Twenty kilowatts of panels on that roof would produce around 90 kWh a day, four and a half times what the house uses.

Use 4.5 for a good roof. Adjust it down when the roof is working against you:

Note the last three rows. The 4.5 figure is an average: a system sized to match consumption on average is generating a surplus every summer day and falling short every winter day. That asymmetry is one of the reasons the answer is rarely the number the division gives you.

Will it fit on the roof?

A 440 W panel is roughly 2 m², so each kilowatt is about 4.5 m² of panel — call it 5 to 6 m² of roof per kW once you allow for setbacks from edges and ridges, access paths and the odd vent or aerial. A 10 kW system therefore wants somewhere around 50 to 60 m² of usable roof, ideally on one or two planes rather than five.

That last line is the real subject of the sizing conversation. The house cannot absorb what it is not awake for. Every kilowatt-hour above the middle bar leaves the property at the feed-in rate, which is a fraction of what you pay to import — so beyond a certain point, extra panels stop reducing your bill and start selling cheap electricity to your retailer.

Having established that 20 kWh a day implies roughly 4.4 kW, the honest next admission is that we frequently recommend 6.6 kW or 10 kW to that same household. That is not a contradiction, and the reasons are worth being able to give without hesitating.

Reasons to go bigger than the arithmetic

• Winter. A system that matches consumption on the annual average produces roughly half of that in midwinter, when a heated house is using more. Sizing above the average is how you keep winter respectable.
• A battery, now or later. Storage only earns its keep if there is surplus to store. In the third bar above, the same array covers 18 of the 20 kWh once a battery moves the daytime surplus into the evening — but the surplus has to exist first.
• An EV. A car adds roughly 10–15 kWh on a driving day, and it is a load you can schedule into the middle of the day. That is the single biggest legitimate reason to add panels.
• Electrification. Replacing gas hot water, cooking or heating moves energy onto the electricity bill — usually onto loads that can be timed for daylight.
• Cost per kW falls as size rises. Scaffold, switchboard work, the inverter, travel, design and paperwork barely change between 6.6 and 10 kW. The extra panels are close to the cheapest energy you will ever buy, which is also why installing once beats installing twice.

Reasons the answer is capped anyway

Can panels be added later?

Sometimes, and it is worth asking before the first system is designed rather than after. What decides it: whether the inverter has DC headroom left inside the 133% rule and its own input limits, whether the new panels can form a sensible string with the existing ones (matched orientation, matched module characteristics), and whether the network will vary the existing approval. Certificates are created on the added capacity at the rate applying when it is installed, not the original rate. In practice, a system designed with the next stage in mind — inverter headroom, spare switchboard space, cable capacity — costs very little extra now and a great deal less later.

What sizing will not do

No array size takes a bill to zero. There is a daily supply charge regardless of what you generate, and unless you have storage, everything consumed after dark is imported at the full rate. Sizing decides how much of your daytime load you cover and how much surplus you have to work with; storage, load shifting and tariff choice decide most of what happens to the rest.

This is the question customers ask most often, and it sounds like someone is cutting a corner. It is the opposite: it is the design working properly.

Go back to the derating above. A 13.2 kW array reaches 13.2 kW essentially never. On a good Sydney day it peaks somewhere around 9–10 kW for a short window either side of noon, and spends the rest of the day well below that. If you matched a 13.2 kW inverter to it, you would have paid for inverter capacity that is idle from dawn to dusk, every day, for twenty years.

Oversizing the array against the inverter fills in the shoulders of the day instead.

Two further limits sit alongside that rule, and both are checked at design:

1. The manufacturer's maximum DC input for that specific inverter, which can be more or less generous than 133%.
2. The network's approval, which is granted against inverter capacity and export limit. Adding panels under an approved inverter does not change what you may export.

Clipping is what happens in the moments when the array could produce more than the inverter can pass. The inverter holds its output at the ceiling and moves the array off its maximum power point to make it produce less.

The important nuance, and the one that reassures customers: the extra energy is never generated. It is not dumped as heat, nothing overheats, and nothing wears out faster. The panels simply operate at a different point on their curve for a few minutes.

The cost is real but small, and it is dwarfed by what the extra panels earn the rest of the time. On a typical Sydney roof at a 1.3 ratio, modelled clipping usually lands in the low single digits as a percentage of annual output. It is modelled per site rather than assumed, because the things that drive it vary:

Said plainly to a customer: clipping is the price of getting more energy in the morning, the afternoon and all winter. On almost every roof we design, that trade is worth making — and if it is not, the model will show it.

These two get conflated constantly, including by people selling systems. They are different mechanisms with different consequences.

A 10 kW inverter export-limited to 5 kW still runs 10 kW of load inside your own house at midday. It is only the surplus heading out to the street that is capped. That is also why a battery is the standard answer to a tight export limit: it converts energy you are not allowed to sell into energy you use after dark at full retail value. The network side of this is covered in our export limits guide.

Whether your property has one phase or three changes several answers at once, and it is worth checking before any other decision is made.

If you are on single phase and planning solar, a battery and an EV over the next few years, have the upgrade conversation once at the start rather than three times.

Batteries are advertised by capacity, which tells you only half of what you need. Every battery has an energy rating in kWh and a power rating in kW, and they answer different questions.

This is where the second half of the customer's question lands: a 13.5 kWh battery has enough energy for my evening, so why did it not run the house? Because the evening total and the instantaneous demand are different problems.

Nothing failed there. The battery supplied its 5 kW and the grid quietly covered the remainder, which is exactly how a grid-connected system is supposed to behave. The same evening during a blackout would look very different, because there is no grid to cover the gap — which is the next section.

Two more things that surprise people:

• Adding kWh does not always add kW. Stacking a second battery module usually adds energy; whether it raises the power rating depends on the product. Check the datasheet rather than assuming.
• AC-coupled and DC-coupled are different retrofits. An AC-coupled battery has its own inverter and sits alongside your existing solar, which is the usual path when adding storage to a working system. A DC-coupled battery shares the solar inverter, which is generally more efficient but means replacing that inverter if it is not already a hybrid.

Sizing a battery from your actual interval data, rather than from your solar system size, is a subject of its own — see the battery sizing method.

Start with the fact that surprises nearly every new customer: a normal grid-connected solar system shuts down in a blackout, even at noon in perfect sunshine.

This is deliberate and required. Australian inverter standards require anti-islanding protection: on losing grid supply, the inverter must disconnect within two seconds. If it did not, your roof could keep a section of dead network energised while somebody is working on it. The protection is not optional and cannot be defeated.

Keeping a house running through an outage therefore needs three things that must be installed on purpose:

1. A battery. Solar alone cannot do it — there is nothing to stabilise a supply that varies with every passing cloud.
2. An inverter with a backup output, capable of forming its own grid rather than following one.
3. A changeover device and an essential-circuits sub-board, wired at installation, so the property is safely isolated from the network before the inverter re-energises anything.

Transfer is not always seamless. Some systems switch over in tens of milliseconds and most equipment never notices; others take a few seconds, which is long enough for computers to restart. If a seamless changeover matters to you, it is a datasheet question to ask before you buy, not after.

Backup is genuinely useful and routinely oversold. Everything below is a normal characteristic of a well-designed system, not a fault, and all of it is easier to explain before installation than during an outage.

How long backup lasts is straightforward arithmetic: reserve energy divided by backed-up load. Essential circuits typically average 0.3–0.6 kW, so 10 kWh available at the moment the power fails runs them for the better part of a day — longer if the system can recharge from solar, considerably shorter if someone puts the kettle on repeatedly.

A system designed to run indefinitely without a grid at all is a different specification again — see hybrid and off-grid systems.

Whether you are a customer comparing proposals or a new team member checking one, these five lines tell you most of what you need to know.

1. Panel count × panel watts. Does it equal the array size printed at the top? Thirty × 440 W must be 13.2 kW, not “about 13”.
2. Inverter model and AC rating. Divide the array by it. Under 1.33 is normal; over it is not eligible for certificates.
3. Battery usable kWh and continuous kW. If only one appears, ask for the other before comparing prices.
4. The export limit assumed. A proposal modelled on unlimited export at an address that will be approved for 5 kW is a forecast that will not survive connection.
5. Backup: included or not, and which circuits. “Battery included” is not the same as “works in a blackout”. The essential-circuits list should be in writing.

Then the incentive lines: the STC discount should show system size × zone rating × deeming period, and any battery incentive should be itemised separately. Both are set out in the STC guide and the battery rebate guide.