Sizing a solar battery bank is arithmetic, not intuition — and most failed solar installations failed at this desk, not in the field. This guide works the complete method with the XD deep-cycling and solar gel lines as the worked material. Bring a pencil: by the end you will have a bank size, a voltage, and a block count you can defend.
Step 1: the load audit, at night
List what the battery alone must power between sundown and useful morning sun: appliance watts × hours each, summed into nightly watt-hours. Be honest about the two loads everyone underestimates — refrigeration (it cycles all night) and standby draws (they never sleep). The daily figure that matters is the worst-season figure, not the annual average: size for the month the system must survive, or consciously accept generator support in that month.
Step 2: depth of discharge — buy your cycle life
Divide the nightly watt-hours by your chosen depth of discharge to get the nameplate capacity the bank must carry. For lead acid, 50% is the standard engineering bargain; 30% buys notably more cycles at notably more battery; 70% spends the bank’s life for a smaller invoice. Gel’s deep-discharge tolerance makes the 50% point comfortable rather than brave. There is no free direction on this dial — only a conscious one.
Step 3: autonomy days
Multiply by the consecutive sunless days the system must bridge. Homes commonly choose one to two days plus a generator; critical loads choose three to five and pay in cells. Autonomy is the single most expensive word in off-grid design, so say it as a number and price it as a line item — our quotations do exactly that.
Step 4: system voltage
Convert to amp-hours at a system voltage: 12V for small systems, 24V as banks pass a few kilowatt-hours, 48V beyond — because current, cable cost and parallel-string count all punish low voltage at scale. Then choose the unit: 12V blocks in series-parallel for the smaller banks, 2V OPzV/OPzS cells in a single series string as the bank grows past roughly 10 kWh, and 51.2V storage systems where inverter integration and modularity lead. The 2V vs 12V guide prices that crossover honestly.
Step 5: the array must be able to refill it
A bank is only as good as its refill. The array and controller must return the daily draw plus system losses on an average day of the design month — otherwise the bank lives in chronic deficit, sulfates, and dies young with the datasheet blameless. Controller profile must match chemistry (gel’s lower absorption voltage is the classic trap). If the array cannot grow, shrink the bank or the load; an honestly small system outlives an ambitious starved one.
A worked miniature
Night load 2.4 kWh; design depth 50%; autonomy 2 days; voltage 24V. Nameplate energy needed: 2.4 ÷ 0.5 × 2 = 9.6 kWh → 400Ah at 24V. In blocks: two parallel strings of two 200Ah-class blocks (four blocks). In cells: a 24V string is twelve 2V cells — at 400Ah class, the OPzS/OPzV families. In storage: two 5.12 kWh-class units’ worth of capacity in the 51.2V line, inverter-paired. Three legitimate answers; the site, the staffing and the budget pick one. Send us your five numbers and we return all three, priced.
Can I size from my panel wattage instead?
No — panels describe supply on a good day; batteries must serve demand on bad ones. Size the bank from the load, then verify the array can refill it.
What if my loads grow next year?
Banks are one age by discipline. Either size for the grown load now, or plan the second bank as a separate system. Blending ages is the expensive middle.
Step 6: write the bank’s rules before you buy it
A sized bank still needs its operating constitution, decided at design time and taped to the battery box: the low-voltage disconnect setting that enforces the design depth of discharge (so a careless week cannot spend next year’s cycles); the generator start rule for systems that have one (autonomy is cheaper as fuel than as cells, past a point); and the equalisation or boost policy appropriate to the chemistry — which for gel banks means a firm "never at flooded voltages". Banks do not usually die of one bad night; they die of a hundred small exceptions. The constitution is what makes the exceptions visible.
Should the bank be sized for winter or the annual average?
For the worst month the system must survive unaided — or state, in writing, the support (generator, load shedding) that covers the gap. Sizing to the average is how systems fail every December.
Common sizing failures, named
The failures repeat so faithfully they deserve names. The Fair-Weather Size: array and bank sized on summer yield, dying every winter. The Nameplate Bank: capacity read off the label at the fastest rate, so the "100Ah" bank delivers 70 in service. The Phantom Load: the inverter’s standby and the always-on router forgotten, a fifth of the bank spent on nothing. The Growth Surprise: the freezer, pump or second building added in year two against a bank sized with no margin story. The Profile Mismatch: a correctly sized gel bank charged on a flooded profile, dying of thirst under a charger’s good intentions. Every one is preventable at the sizing table with this guide’s six steps — and every one is expensive by the time it announces itself, because batteries report their grievances in seasons, not in error messages.
The sizing conversation, summarised
Loads with their hours; days of autonomy chosen consciously; usable capacity divided by the design depth; the result corrected for temperature and charge efficiency; the charger and array confirmed able to refill what the arithmetic promises; and the bank’s operating rules written down before the order is placed. Six steps, none optional, none difficult. Solar systems fail at the battery more than anywhere else not because the step is hard, but because it is the one most often skipped in favour of buying a round number. Do the arithmetic once, and the bank stops being the system’s question mark.
Finally, a word about who should check the arithmetic: anyone competent, twice, independently. Sizing errors compound silently for years and announce themselves expensively; a second pair of eyes on the six steps — ideally the supplier’s, against your load sheet — is the cheapest insurance in the entire system. Send us the sheet; checking it is part of the quotation, not a favour.