Battery Runtime Calculator

The calculator turns your battery into watt-hours, takes the share you are willing to use, subtracts the losses between battery and load, and divides by your load. On LiFePO4 that is the whole of it.

Runtime (hours) = (Capacity x Nominal Voltage x DoD x Efficiency) / Load

A 100Ah 12V LiFePO4 battery at 80% DoD and 90% efficiency powering a 100W load: (100 x 12.8 x 0.80 x 0.90) / 100 = 9.22 hours. The 12.8V is not a typo — Victron prints "Nominal voltage 12,8 V" and "1280 Wh" for its 100Ah LiFePO4 block, so feeding a flat 12 understates the pack by 6%. That same battery on a 50W LED worklight lasts 18.4 hours, enough for two full nights. The 90% efficiency default is Victron's own printed maximum for the Phoenix VE.Direct 500VA and 800VA inverters at 12V; its 250VA and 375VA models print 87% and 89%.

Lead-acid does not work that way, and the calculator no longer pretends it does. Pick flooded, AGM or gel and the runtime is computed from a discharge-rate correction fitted to figures the makers print: Trojan's T-105 sheet, Lifeline's GPL-31T minutes-of-discharge rows, and Victron's AGM and gel derate table. Across those printed points the fit reproduces flooded within about 10%, AGM within about 6% and gel within about 2%.

The formula assumes a constant load, which rarely happens in practice. A refrigerator cycles on and off, pulling 150W during the compressor run and near-zero in between. For cycling loads, use the average power draw over a full cycle. Our battery runtime guide walks through more worked examples across different battery types and load scenarios.

Depth of discharge (DoD) is the variable most people leave at whatever the box says. It decides both how long you run and how many times you get to do it. Victron prints the trade for its own AGM line: 400 cycles at 80% discharge, 600 at 50%, and 1500 at 30%. Run the pack harder and you buy hours today with cycles later.

LiFePO4 changes the arithmetic. Battle Born rates its 100Ah block at "100% Depth of Discharge" outright, and Victron's LiFePO4 sheet prints 2500 cycles at 80% DoD, 3000 at 70% and 5000 at 50% — at the deepest setting it still outlasts a lead-acid pack run gently.

AGM sits between the two, and its maker is specific: Concorde's manual for the Lifeline line says the average depth of discharge should be "no greater than 50% of the battery's 20 hour rating". Gel is the same story with slightly better numbers, Victron printing 750 cycles at 50% against 500 at 80%.

The practical takeaway has changed with the engine. A 200Ah flooded bank at 50% DoD holds 1,200 Wh; a 100Ah LiFePO4 pack at 100% holds 1,280 Wh. Run both at 100W on this calculator and the flooded bank returns 10.3 hours against the lithium pack's 11.5, because the flooded correction takes its cut and the lithium one does not. Add that lead-acid is far heavier for the same usable energy and the cost-per-cycle maths usually favours lithium. Our depth of discharge guide breaks down the chemistry-by-chemistry DoD limits and their lifespan impact.

Switch the chemistry and you switch the method, not just a number.

On LiFePO4 the calculator uses the plain energy identity. That is honest here because the label is measured at a 1C discharge: Victron footnotes its nominal-capacity row "Discharge current 1C", so a 100Ah pack really does give 100Ah at loads up to 100A. The only correction is the nominal voltage, 12.8V per 12V block.

On flooded, AGM and gel the rated amp-hours are measured at the 20-hour rate, and the pack delivers less than that at any heavier load. Trojan's own T-105 sheet prints the spread on one row: 250Ah at the 100-hour rate, 225Ah at 20 hours, 207Ah at 10 hours, 185Ah at 5 hours. The calculator works out the current your load actually draws, expresses it as a fraction of the 20-hour rating, and reads the runtime off a curve fitted to those printed points.

Two things follow. Runtimes on lead-acid are shorter than the bare formula gives, by more as the load grows. And when your load falls outside the range of points the makers print, the result says so instead of extrapolating quietly.