A 5 kW label on an inverter describes one operating point, and a borehole pump asks an entirely different question. Inverter size is chosen from two numbers: the total load running at the same instant, and the largest start-up surge among the appliances that switch on, which in a South African home is almost always a motor. Panel count belongs to a separate sum about daily energy, and it has no bearing on whether a submersible pump will turn over at 06:20 on a Saturday.
I was standing on wet paving next to the borehole tank with a coffee going cold in one hand, watching the inverter app on my phone throw the same warning it had thrown three mornings in a row. The pump was making its low grinding noise, the one that means it has not actually started. Inside, on the garage wall, the inverter's fan had changed note. Sixty seconds later the pump gave up and the app cleared itself, as if nothing had happened. The inverter has a 5 kW label on it. The pump is rated at 0.75 kW. On paper, that is a joke.
That gap between two numbers on paper and what actually happens at the tank is where most home solar sizing goes wrong.
The two numbers that actually pick the machine
The first is simultaneous load. Not the total of everything in the house. Not the sum of every appliance nameplate stuck to the DB board with tape. The load that is running at the same moment the worst appliance starts.
Walk around your own house with a pen at 07:00 on a weekday and again at 18:00. You will find the geyser element recovering, the kettle, a microwave, the fridge and freezer compressors, possibly a pool pump, a pressure pump, the tumble dryer if it is an old one, an induction plate. A sensible house keeps that stack to 2 to 4 kW. A Saturday morning with everyone home can push it to 6 or 8 kW without anybody doing anything unreasonable.
The second number is the largest start-up surge. Any appliance with a motor in it draws several times its running current for a moment when it starts. Air conditioners and fridges do it. Pool pumps do it. Pressure pumps do it. Borehole pumps are the worst of the lot, because a submersible pump starts against a head of water, sometimes 60 or 80 metres down a borehole, through a long cable, with no soft start. That first second is brutal, and it is the number that decides whether your inverter holds or trips.
Why the pump is worse than the fridge
I noticed the pattern only after living with it for a while. The fridge and freezer cycling caused nothing. The pool pump occasionally made the lights flicker for half a heartbeat. The borehole pump is the one that decides the day.
Two reasons.
First, the surge. Running current for a small submersible pump might sit somewhere near 4 amps. Starting current can be several times that for a second or two. Multiply by roughly 230 volts and the arithmetic stops being comfortable, and it lands on top of whatever else is running at that moment: the geyser element, the kettle, whatever the children have switched on.
Second, and this is the part people miss, a pump gets harder over its life. Worn impellers, silt, a water table that has dropped since the borehole was drilled. A system that started the pump comfortably in year one can be tripping on overload in year three, and the inverter gets blamed for a pump fault.
Where the surge has to come from when the grid is off
This is the section I would write on a card and hand to every installer's client.
During a normal day, with the grid up, the inverter is passing power through and the surge has somewhere to go. During load shedding it does not. The surge has to come out of the battery, through the inverter, in a fraction of a second.
So the inverter's surge rating is not the only ceiling. There is a second ceiling, lower and far less advertised: the battery's continuous discharge current.
Here is an illustrative example. A 5 kWh lithium battery at roughly 51 volts, limited to 50 amps continuous discharge, can deliver around 2.5 kW. Not 5. Two houses can own identical inverters with identical labels and only one of them will start the borehole pump, because one battery can push 100 amps and the other cannot.
When the battery's management system trips on overcurrent, the whole house goes dark, the inverter shows a fault code you then have to look up, and nothing on the label explains what happened. I have watched someone replace a perfectly good inverter because nobody thought to open the battery datasheet.
Continuous, peak, and the fine print between them
A 5 kW inverter is not 5 kW of everything. There are usually three or four numbers hiding behind that label.
Continuous output. This is the honest one, and it is measured at a stated ambient temperature, often 25 degrees Celsius. A garage in Gauteng in February is not 25 degrees Celsius. Manufacturers publish derating curves, and a machine that makes 5 kW at 25 might make noticeably less at 45. Read the curve, not the cover of the brochure.
Peak or surge output. Sometimes 1.2 times continuous for a few seconds, sometimes double for one second, sometimes listed in VA rather than watts. The duration matters as much as the figure. A pump that takes two seconds to spin up is not helped by a spec that only holds for one.
Battery mode output. On a hybrid inverter, some units deliver a lower AC output when they are running on battery alone than when the grid or the panels are assisting. That is the number that governs your evenings and your load shedding mornings, and it is often on a different page of the datasheet.
If the unit is grid-tied with no battery at all, none of this helps during an outage. A grid-tied inverter has to shut down the moment the grid disappears, whatever its kW rating, because it is required not to keep energising a cable the municipality believes is dead. That is a topology problem, not a sizing problem, and no amount of money spent on kilowatts fixes it.
Oversizing on purpose, and what it quietly costs
I have chosen to oversize, and I would again. But I want to be honest about why, because "buy bigger" is lazy advice.
The reasons that hold up.
Headroom. The pump starts on top of the geyser and the kettle. Give the machine somewhere to go, and the surge stops being a daily coin toss.
Array room. A larger inverter usually accepts a larger array and a higher string voltage on each MPPT, which means the panels start producing a little earlier in the morning and a little later in the afternoon, and the system clips less on a clear day. Low-light behaviour depends heavily on the string configuration, and that is a design decision rather than a shopping decision.
Gentler life. Running at 40 percent instead of 85 percent keeps the fans quiet and the case cooler, and heat is what ages electronics.
The costs that hold up.
Standby draw. Every inverter eats something overnight. Twenty watts is small; sixty watts is not. Run that over a year and it turns into a real line on a prepaid meter, and on a battery it is the reason you arrive home to 68 percent instead of 80. On a bigger machine, the number is usually bigger.
Purchase price, and the DC array needed to justify it. An oversized inverter fed by a small array spends most of its life under-worked, and you paid for the capacity anyway.
The battery ceiling again. Oversizing the inverter without checking the battery's discharge limit buys you a bigger number on the wall and exactly the same trip at the tank.
The cheaper fixes that are not a bigger inverter
A soft starter on the pump motor reduces the inrush, and a pump controller or variable-speed drive does something similar on a three-phase borehole pump. That is an electrician's work and it changes the problem at the source rather than buying kilowatts to absorb it.
Staggering starts costs nothing. Geyser on a timer so it recovers before 06:00. Pool pump in the middle of the day. Borehole pump scheduled for the hours when the sun is doing its work. Four appliances that never start together need a much smaller inverter than four appliances that do.
And the unglamorous one: put the pump on a circuit that only runs from the grid. Keep the backup side of the board for the fridge, the lights, the router, the TV and a couple of plugs. If the borehole is filling a tank anyway, it does not need to chase you through load shedding.
The part I do not do myself
String configuration, maximum DC input voltage, cold morning open-circuit voltage, maximum current per MPPT, changeover wiring, neutral and earth bonding while the inverter is islanding, and everything SANS 10142-1 has to say about it. I do not touch any of it, and neither should anyone reading this. Those figures are checked by a qualified person against the panel datasheet, the inverter's input limits and the actual cable run, and they belong on paper before anything is mounted.
What I do is ask. Who is doing the string design. What is the maximum DC voltage on the coldest morning. Where is it written down. If the answer is vague, that is the answer.
What I ask before anything gets signed
I ask for the continuous AC output at 40 degrees Celsius, not at 25. I ask for the surge figure with its duration in seconds. I ask for the battery's continuous discharge current and its management system limit, in amps. I ask what the AC output is in battery-only mode. I ask whether the inverter model family carries the grid-connection certification my municipality wants, because that is a separate requirement from how big the machine is. And I ask which circuits land on the backup side of the board, by name.
Then I listen for the pump. That is the only test that matters, and it happens at 06:20, with the geyser still on and the grid still off.
