How to Size an Off-Grid Battery Bank

Sizing a battery bank is the most consequential arithmetic in an off-grid build. Get it wrong low and you sit in the dark in February. Get it wrong high and you have spent thousands of dollars on capacity you never use, on a component that ages whether you use it or not.

The method is straightforward. The judgement is in the inputs.

Step 1: daily energy use

Everything starts from how many watt-hours you use in a day.

List each appliance with its running watts and realistic hours of use per day, and multiply. The two things people get wrong:

Thermostatic loads cycle. A fridge rated 150 W does not draw 150 W for 24 hours. It runs perhaps 30 to 50% of the time depending on ambient temperature and how often the door opens. Entering 24 hours at full draw overstates it by two or three times, and that error then inflates every other component in the system.

Phantom loads add up. Anything with a standby light, a clock, or a remote receiver draws power continuously. Individually trivial, collectively often 10% of a small system’s budget.

Suppose the answer is 2,000 Wh per day. That is a realistic figure for a small cabin with LED lighting, a modest fridge, a laptop, and a water pump.

Step 2: days of autonomy

How long must the system run with no meaningful solar input?

Situation	Typical autonomy
Weekend cabin, generator available	1 to 2 days
Full-time cabin, generator backup	2 to 3 days
Full-time home, no generator	3 to 5 days
Cloudy northern climate, no generator	4 to 5 days

This is a budget decision as much as a technical one. Each additional day multiplies the largest cost in the system. A generator is very often cheaper than the two extra days of storage it substitutes for, which is why so many off-grid homes have one.

Take 2 days: 2,000 × 2 = 4,000 Wh of energy you must be able to take out of the bank.

Step 3: depth of discharge

This is where nameplate capacity and usable capacity separate.

Chemistry	Usable DoD	Nameplate needed for 4,000 Wh usable
Flooded lead-acid	50%	8,000 Wh
AGM	50%	8,000 Wh
LiFePO4	80%	5,000 Wh

Divide your required usable energy by the depth of discharge. Lead-acid needs roughly 8,000 Wh of nameplate capacity; LiFePO4 needs about 5,000 Wh.

This is why comparing batteries on price per rated amp-hour is misleading. Compare cost per usable watt-hour, and then compare cost per usable watt-hour across the expected cycle life. See Off-Grid Battery Basics for that comparison.

Step 4: losses and margins

The number from step 3 is a floor. Three things push it up.

Inverter efficiency. AC loads pay a conversion tax of roughly 10 to 15%. If most of your consumption is AC, your 2,000 Wh of appliance demand is closer to 2,300 Wh drawn from the bank. Either build this into your daily figure or add 15% here, but do it once, not twice.

Temperature. Batteries deliver less when cold. Lead-acid can lose 20 to 30% of its capacity near freezing. If the bank lives in an unheated shed in a cold climate, size for the temperature it will actually be at. This is a large correction and it is routinely ignored.

Aging. A battery at the end of its warranted life is typically down to 80% of its original capacity. A bank sized with no headroom is undersized for most of its service life.

Applying a combined margin of roughly 20 to 30% to the step 3 figure is reasonable. For the LiFePO4 example, 5,000 Wh becomes about 6,000 to 6,500 Wh.

Step 5: convert to amp-hours

Batteries are usually sold in amp-hours at a nominal voltage, so divide watt-hours by system voltage:

System voltage	6,000 Wh becomes
12 V	500 Ah
24 V	250 Ah
48 V	125 Ah

Same energy, very different current and therefore very different cable. A 500 Ah bank at 12 V feeding a 2,000 W inverter is pushing nearly 200 A, which needs seriously heavy conductors. See wiring and safety.

Step 6: check the array can actually charge it

A bank the array cannot refill is worse than no bank at all, particularly for lead-acid, where chronic undercharging causes sulfation and kills it early.

As a sanity check, your array should be able to replace a day’s consumption plus recharge losses during the peak sun hours of your worst month, not your annual average. If it cannot, either the array is too small or the bank is too large for it.

Lead-acid also has a charge-rate requirement: it generally wants a charge current of at least 10% of capacity to reach a proper full charge. A 500 Ah lead-acid bank wants around 50 A of charging available. LiFePO4 is far more tolerant and will accept whatever you can give it.

Worked example

A cabin using 2,000 Wh/day, wanting 2 days autonomy, on 24 V, using LiFePO4:

1. Daily use: 2,000 Wh
2. Autonomy: 2,000 × 2 = 4,000 Wh usable required
3. DoD 80%: 4,000 ÷ 0.8 = 5,000 Wh nameplate
4. Margin 25% for inverter losses, cold, and aging: about 6,250 Wh
5. At 24 V: 6,250 ÷ 24 = about 260 Ah

In practice that is three 100 Ah 24 V batteries, or a comparable arrangement in series and parallel.

The same cabin on flooded lead-acid would need roughly 10,000 Wh nameplate, or about 420 Ah at 24 V, and would weigh several times as much.

Common mistakes

• Sizing on average consumption rather than a realistic bad week.
• Forgetting the inverter tax, then wondering why runtime is 15% short.
• Ignoring temperature for a bank that lives outside.
• Buying capacity the array cannot refill, which slowly destroys lead-acid.
• Mixing old and new batteries to expand a bank. The weakest cell governs the whole thing, so this wastes the new ones.
• Planning to expand later. Batteries should be bought as a matched set at the same time. Adding capacity in a year usually means replacing the bank instead.