Batteries

Array size answers "how much do I make". Battery size answers a different question: how long can I run when I make nothing?

bank kWh = daily load × days of autonomy ÷ depth of discharge ÷ round-trip efficiency

Those last two terms are why nameplate capacity is not comparable across chemistries. A 22 kWh lithium bank and a 40 kWh lead-acid bank do the same job, because lead-acid can only give up half its capacity without shortening its life dramatically.

Depth of discharge is a lifespan setting, not a limit

Compare banks by cost per delivered kWh — purchase price ÷ (usable kWh × rated cycles). It is the only comparison that survives the chemistry differences, and it usually reverses the answer that price-per-nameplate-kWh gives.

Days of autonomy

Two days is the usual off-grid figure with a generator as backstop. Past about three, a generator is cheaper than the equivalent battery and only burns fuel on the days it runs.

Size the bank for the worst month, not the annual average. Off-grid systems fail in January.

Charge controllers

MPPT converts surplus panel voltage into extra current; PWM discards it. The gap is 20 to 30%, and it is widest exactly when you need the energy most — cold clear days, when panel voltage is highest and the battery is lowest.

Size the controller's maximum input voltage against the array's cold open-circuit voltage, not its rated Voc. Voc rises as temperature falls, and this is the most common way controllers are destroyed.

Temperature

Lead-acid loses about 20% of capacity at 0 °C. Lithium must not be charged below freezing without a heater or a low-temperature cutoff. An unheated shed in a northern climate is not a battery location without allowing for both.