Your LiFePO4 battery is not flat and it is not faulty. The kettle did not drain it. What tripped the house was the discharge current limit — the C-rate — and the battery management system did exactly what it is designed to do when the current climbs past the pack's continuous rating. State of charge tells you how much energy is sitting in the cells. It says nothing about how fast you are allowed to pull that energy out, and those two figures live in separate columns on the datasheet.
I was standing at my own kitchen counter at ten past six when the kettle hit its boil and everything died. Not a flicker. A clean, total silence — fridge compressor gone, TV gone, the small green light on the fibre box gone — and then, two seconds later, the click and whirr of the inverter restarting itself in the garage and the house coming back like nothing had happened.
The display was lying to me, gently
The first thing I did was walk to the garage and look at the battery. Indicator ring solid, no red, no fault light. I opened the app and it said 78%.
Seventy-eight percent. That number is a coulomb count or a voltage estimate depending on the pack, and in both cases it answers one question: how much energy is left. It does not answer the question the inverter was asking, which was "can you give me 3 000 W right now, this second, and keep giving it for the next three minutes?"
I had been treating the battery like a fuel tank. It is not a fuel tank. It is a fuel tank with a maximum flow rate stamped on the side of the tap.
What C-rate actually means in amps
C-rate is the discharge current expressed as a multiple of the pack's capacity. One C means the pack can be drained in an hour. Half C means two hours. For a lithium pack, the datasheet states a continuous discharge figure in amps, sometimes alongside a shorter peak figure, and sometimes as a C number you have to convert yourself.
Take a nominal 51.2 V pack with 5 kWh of energy. That is roughly 100 Ah, because amp-hours times volts gives watt-hours — 100 × 51.2 = 5 120 Wh. Now apply the ratings:
- 0.3C on a 100 Ah pack is 30 A, about 1.5 kW at nominal voltage.
- 0.5C is 50 A, roughly 2.5 kW.
- 1C is 100 A, roughly 5 kW.
- 2C would be 200 A, around 10 kW — a rating you mostly see on high-power cells rather than on home storage packs.
My kettle is a 3 kW one. Add the 300 to 500 W of background load — fridge, router, fibre box, a few LED lights, the TV — and the moment that element switches on I am asking the pack for somewhere between 3.3 and 3.5 kW. At 51.2 V that is about 65 A, and if the pack is sitting nearer 50 V under load, closer to 70 A. On a battery with a 50 A continuous limit, that is a 40% overshoot. The BMS holds it for a few seconds — some boards allow a brief overcurrent window — and then opens the contactor.
That is the four-second kettle trip. If yours dies almost instantly, the overshoot is bigger. If it dies after about ten seconds, you are sitting just over the line.
The inverter's rating and the battery's rating are different promises
My inverter will happily advertise a 5 kW continuous output and a 10 kW peak for a few seconds. People read that and assume the battery must match it. It does not have to, and often it does not.
The inverter's number describes what the electronics can convert. The battery's number describes what the cells and the BMS will allow through. They are two separate bottlenecks in series, and the system is only as strong as the tighter one. An inverter rated for 5 kW fed by a pack with a 50 A continuous limit has an effective ceiling of roughly 2.5 kW from the battery — the rest comes from solar or the grid, or it does not come at all.
This is also why a hybrid system behaves so differently from a pure backup setup. If the grid is up and the inverter is configured to supply the heavy loads from the grid, the kettle never touches the battery and nothing trips. The moment load shedding starts and the battery is the only source, that ceiling becomes real.
Surge loads are a different animal to a kettle
A kettle is a resistive load. There is no inrush worth mentioning — the element goes from zero to full power in a step. The trap with a kettle is simply simultaneity and size.
Motors are the ones that bite in the first half-second. A fridge or freezer compressor, a pool pump, a borehole pump, an older washing machine — these draw several times their running current for a fraction of a second while the rotor comes up to speed. A pump running at 750 W might pull 3 000 W or more for 200 milliseconds at start. That is fast enough that the inverter may ride through it, but the BMS may not, because the current sensor sees a spike well above the continuous limit, and some boards trip on the rate of change as much as on the absolute value.
The nastiest combinations are the ones that overlap. Kettle on, fridge compressor kicks in, someone opens the microwave — three steps stacked inside two seconds. That is the trip people describe as "random", and it is almost never random.
A bigger battery does not automatically fix this
This is the part that costs people money. Capacity and current capability are separate specifications. A 10 kWh pack is not automatically twice as powerful as a 5 kWh pack. It has twice the energy.
Put the numbers side by side. A 5 kWh pack rated at 1C delivers about 100 A. A 10 kWh pack — roughly 195 Ah at 51.2 V — rated at 0.5C delivers about 97 A. Same current, double the stored energy. If the larger pack's BMS is limited to 50 A continuous regardless of what the cells can do, you get 50 A and a bigger bill. If the bigger pack uses a cell configuration or a BMS board with a lower C-rating, upsizing the energy can actually leave you with less peak current than you had before.
Where upsizing does help is when the current limits add. Two identical packs in parallel split the load between them, so the available continuous current roughly doubles. But that is a different change from "I bought a bigger battery". Before you spend anything, find the continuous discharge amps on the datasheet and the BMS current limit, and check whether that figure is per pack or per system.
What the BMS does when you push it, and why the lights go out
There are two behaviours, and both look like a blackout from the kitchen.
The gentler one is a throttle. If the battery and inverter are talking over a communications cable — a closed-loop setup — the BMS can tell the inverter to reduce its discharge demand. You will see the load drop, or the inverter quietly pulling the difference from the grid. Some people never realise their battery is being limited, because the system simply works and the grid pays.
The blunt one is a hard disconnect. The contactor opens while the inverter is still pulling 70 A, and all that current in the DC cables and the inverter's input capacitors has nowhere to go. The voltage spikes, and the inverter usually logs a DC bus high or overvoltage fault before it reboots. Do that often enough and you are putting wear on the contactor and the pre-charge circuit that nobody designed for.
There is a third possibility worth checking before you blame the battery at all. Under a heavy pull, pack voltage sags — a couple of volts, sometimes more if the DC cables are undersized or the run to the inverter is long. If the inverter's low DC cutoff is set generously high, the inverter itself can fault on low DC voltage before the BMS has done anything at all. Same symptom, completely different cause, and the fix is a setting rather than a new battery.
Reading your own house before you read another quote
I stopped guessing by writing down actual numbers. On a piece of paper: every appliance that could plausibly run on the backup circuit, and the wattage on its label. Kettle, microwave, air fryer, toaster, iron, hairdryer, vacuum, the geyser element if it is on the essential list. Then I marked which ones could be on at the same time — realistically, not theoretically. It is never just the kettle. It is the kettle plus whatever was already running.
Then convert. Watts divided by the pack's nominal voltage gives amps. Compare that figure against the continuous discharge amps on the battery datasheet, not against the inverter's output rating and not against the marketing kilowatt figure on the box. If your simultaneous load in amps is anywhere near the continuous rating, expect trouble, because the pack will not be at nominal voltage under that load, and the amps will be higher than the paper sum.
There is an early warning worth listening for. Before anything trips, watch the lights when a heavy appliance switches on. If they dip for a heartbeat, or the inverter beeps without shutting down, you are already operating near the ceiling. That dip is voltage sag, and it is the system telling you the current is higher than it wants to give.
The fixes that actually work
The cheapest one is sequencing. Not running the kettle, the microwave and the air fryer at the same time sounds trivially obvious until you realise that in most homes they sit on the same 16 A kitchen circuit and nobody gives it a thought. During load shedding, they cannot. One at a time.
The next is making the heavy loads grid loads. If the grid is present, set the inverter so that high-demand appliance circuits are supplied from the grid rather than from the battery, and let solar charge the battery instead. This costs you nothing in battery capability and removes the trip entirely while the grid is up.
Smaller appliances help more than people expect. A 1 500 W kettle instead of a 3 000 W one halves the worst step in the house. A gas hob or an instant hot water tap removes it completely. For a battery that is otherwise doing exactly what it should, that is a far cheaper answer than another pack.
If you do add battery capacity, add it for the right reason and check the right column. Buy it to run the house longer through the night or through a longer outage. If you buy it to run more appliances at once, confirm that the continuous discharge current — in amps — actually goes up in the configuration you are being quoted.
And accept that none of this gives you a guaranteed runtime for a specific appliance. It depends on the pack, the BMS, the inverter, the cable run, the ambient temperature and what else is running. Cold LiFePO4 packs deliver a little less current capability than warm ones. Two houses with identical batteries will behave differently. The point is not to memorise a number — it is to know which column on the datasheet decides whether the kettle finishes boiling.
Mine boils now, most of the time. When it does not, I know it is the four seconds of current that did it, not the 78% on the screen.
