Two ratings, two different jobs
An inverter carries two numbers and they answer different questions. Continuous rating is what it can hold indefinitely at some stated ambient temperature. Surge rating is what it can hold for a second or two while a motor gets moving. A well pump drawing 900 watts running can want three times that for the fraction of a second the rotor is stationary, and an inverter that cannot supply it simply shuts down — the pump does not start, the inverter faults, and nothing about the running load explains why.
So the sizing is two calculations. Continuous is the coincident load plus headroom. Surge is the coincident load plus the largest single starting increment, on the assumption that motors do not all start in the same instant. The second number usually lands well above the first, and the inverter has to satisfy both.
The coincidence factor matters more than people expect. A load list summed at face value assumes the microwave, the washing machine, the well pump and every light run simultaneously, which is not a household, it is a stress test. Sixty percent is a reasonable planning figure for a mixed domestic list; if the list is short and everything on it genuinely runs together, use 100.
Surge, and the loads that cause it
| Load | Starting behaviour |
|---|---|
| Lights, chargers, kettles, resistive heaters | No surge worth modelling. Running equals starting. |
| Fridges and freezers | Compressor motor. Commonly three to six times running for well under a second. |
| Well pumps, submersibles, air compressors | The worst case in most systems. Three times running is ordinary, more under load. |
| Furnace blowers, shop tools with induction motors | Two to three times, and repeated frequently. |
| Anything on an inverter drive or a soft starter | Little or no surge, which is why a soft starter is often cheaper than the next inverter up. |
The datasheet figures are the ones to use where you have them. Where you do not, a nameplate horsepower figure and a multiplier is a starting estimate, but it is only that, and the difference between three times and six times is the difference between two inverter sizes.
MPPT and PWM are not two grades of the same thing
A PWM controller is a switch. It connects the string to the battery and modulates the connection, which drags the panels down to whatever the battery voltage happens to be. A modern 60-cell or 72-cell module has a maximum power voltage somewhere in the thirties or low forties; connect one to a 12 volt battery through PWM and it operates at about 13 volts. The current is roughly unchanged, so the power delivered is roughly current times 13 rather than current times 41. The rest is not stored anywhere and it is not wasted as heat — it simply never leaves the panel.
An MPPT controller is a DC-to-DC converter with a tracking algorithm. It holds the array at its maximum power point and converts the surplus voltage into extra current at the battery, so the watts carry through less conversion losses. That is why the two current calculations look so different: MPPT output current is array watts divided by battery volts, while PWM output current is array watts divided by string Vmp.
Worked, for a 2,000 watt array on a 48 volt bank. Through MPPT it is 2,000 divided by 48, which is 41.7 amps, and with a 25 percent factor that wants a 60 amp controller, delivering the full 2,000 watts. Through PWM you first need the string to sit above 48 volts at all, so call it two 41 volt modules in series for a string Vmp of 82. That string carries 2,000 divided by 82, about 24.4 amps, and PWM delivers 24.4 amps at 48 volts, which is roughly 1,170 watts. The harvest ratio is simply battery volts over string Vmp — 48 over 82, about 59 percent. PWM earns its place in exactly one situation: small systems where the string Vmp is already close to the battery voltage, which in practice means nominal 12 volt panels on a 12 volt bank.
Voltage windows, and the cold morning that kills controllers
Controller current is the easy constraint. The one that destroys equipment is maximum DC input voltage. Module open-circuit voltage rises as temperature falls, at a rate the datasheet gives as a temperature coefficient, typically a few tenths of a percent per degree. A string sized against its summer operating voltage can exceed the controller's absolute maximum on a clear, cold, still morning — the worst case is bright sun at the record low temperature for the site, before any load is drawn. String length is decided against that case, not against the comfortable one, and the same logic applies to grid-tied inverter input windows.
The part that is not arithmetic
Everything above sizes equipment. It does not design a system. The conductor between the battery and the inverter carries hundreds of amps in normal operation and vastly more into a fault, and it needs overcurrent protection rated for the fault current the bank can actually deliver, positioned close to the battery, with a disconnect a person can reach in a hurry. A battery bank is not a beginner wiring job. A modest lithium bank can push thousands of amps into a shorted spanner without blinking, and DC arcs do not self-extinguish the way AC arcs do. Every battery and every string needs correctly rated overcurrent protection close to the source, a disconnect you can reach, and conductors sized for the fault current rather than the running current. Anything that ties a solar system to a building supply or to the grid is permit and inspection territory in most places. Interconnection agreements, rapid shutdown provisions, labelling, disconnect placement and who is allowed to do the work vary by jurisdiction and by utility, and the code edition your Authority Having Jurisdiction has adopted is what governs the installation, not a web page. Use these numbers to plan and to price, then have the design reviewed by someone who knows what your utility and your inspector expect.
Questions people ask
What size inverter do I need for my house?
For a whole house, size it on the coincident load rather than the service size, because a 200 amp service is a capacity limit, not a consumption figure. Most homes run a few hundred watts most of the time and spike into the low thousands when cooking, pumping or drying. The practical approach is to list what genuinely has to run, apply a coincidence factor, add headroom, and then check the surge separately against the largest motor. Where people go wrong is the other direction: buying a 12 kW inverter to run a 1.5 kW load, which costs more, has higher idle consumption, and does nothing for the surge problem if the offending motor is still too much for it.
Should I use one big inverter or several smaller ones?
One larger unit is usually simpler, cheaper per watt and easier to wire, and it means one thing to fault-find. Several smaller units, or stackable units of the same family, buy you redundancy and better light-load efficiency, since a large inverter running a 100 watt night load wastes a meaningful share of it just being switched on. Stacking also lets you add capacity later without replacing what you have. The catch is that units generally have to be the same model and firmware to stack, so this is a decision to make at purchase rather than to discover later. Idle draw is worth checking on any unit that will be on around the clock; over a year it is not a small number.
Can my charge controller be smaller than the array current?
Some controllers explicitly permit it and will simply clip output at their rating, and some manufacturers state a maximum array wattage above the current rating for exactly this reason. Whether it is acceptable is a per-product question answered by the manual, not a general rule, and doing it with a controller that does not support it is how controllers die. Where it is supported, the trade is that you lose the peaks on the best days in exchange for a cheaper controller, which can be a sensible economic choice on an array that only reaches nameplate for a few hours a year. Read the manual for the specific unit.
Why does my inverter shut down when the pump starts even though it is big enough?
Usually the battery, not the inverter. A starting surge pulls an enormous DC current for a moment, and if the bank has high internal resistance, or the cable between it and the inverter is undersized or long, the voltage at the inverter terminals sags below its low-voltage cut-out and it trips. The inverter is protecting itself from an input problem, and its wattage rating had nothing to do with it. Check the cable size and length first, then the terminations, then the state of the bank, and measure the voltage at the inverter terminals during a start rather than at the battery. Cold lead-acid is a common culprit, since its internal resistance rises as it gets colder.