Calculate battery runtime

Available energy is E = C · U · (SoC_start − SoC_min) · f_temp · f_age; after the discharge path through cabling and inverter, E · η_discharge arrives at the load, and runtime is t = E / P. The inverter's idle draw is added to the load — 15 to 30 W of standing consumption is typical in off-grid systems, and with small loads it becomes the dominant item.

No. Round-trip efficiency belongs on the charging side, in array sizing and charge-time calculations. Applying it again to a battery that is already full double-counts it and makes the result 8 to 15 % too pessimistic. On the discharge side only the cabling and, for the AC share of the load, the inverter matter.

No — it assumes a constant load and a battery at rated temperature. Strongly varying loads, voltage sag during motor start, self-discharge over long idle periods and ageing during the discharge are not covered. For lead chemistries the Peukert relation replaces the linear division, since a linear calculation would be too optimistic there; for LiFePO4 the effect is practically zero.

Because the inverter idle keeps running: at 100 W load and 25 W idle the system draws 125 W; halving the load to 50 W leaves 75 W — the runtime grows by two thirds, not by half. The three scenario bars show this directly, and the clock label turns it into an actionable statement: “lasts until 07:40 tomorrow” instead of “13.6 hours”.

The linear C/I calculation assumes capacity is independent of discharge current. For lead that is false (Peukert exponent ~1.25): higher currents drain disproportionately more capacity. The calculator shows both values — naive and corrected — and deliberately applies no correction for LiFePO₄, where the effect is practically zero.

It describes charge plus discharge losses together and belongs on the charging side. On the discharge side only cable and inverter losses count (proportional to the AC share). Applying round trip on top double-counts losses — the engine even checks consistency with the usable-capacity calculator as an invariant test.

From the chosen chemistry's characteristic curve at the temperature of the installation site. Previously it was fixed at 1.0 — room temperature, regardless of where the battery sits — and not even reachable through the form. If the battery sits outdoors or in a vehicle, the monthly mean at your location now applies: in a Berlin January that is 0.6 °C and a factor of 86 % for LiFePO4. The example runtime drops from 7.4 to 6.3 hours.

Because the question differs from bank sizing. Calculator 3 designs for the worst case and therefore uses the 1 % percentile of daily minima. Here the question is the expectable runtime in a chosen month — for that the monthly mean is the honest figure. If you want the coldest case, choose January; if you want to be more conservative, enter the cold factor yourself via the switch.

From the appliance catalogue, for the device class your load falls into — at 100 W that is the 1 kW class at 10 W idle, not the blanket 20 W previously preset. With a pure DC load (AC share 0) the item disappears entirely because no inverter runs. The value from your unit's datasheet remains the better number via the switch.

Yes — since 5 August 2026 via the switch in the advanced assumptions. The value existed in the calculation schema before but was unreachable through the form; the chemistry's curve midpoint was always used, and the calculator flagged that as the dominating assumption. The datasheet value makes lead runtimes considerably more solid: for AGM the range spans 1.05 to 1.15.

It shifts the usable window: a battery entering an outage at only 80 % has a fifth less energy than the full-start calculation assumes. The field was previously unreachable and fixed at 100 % — now it can be set in the advanced assumptions, for storage that rarely fills up in winter, for example.