Power
Battery life estimator
Estimate how long a duty cycled node runs on a battery, from what it draws asleep, what it draws awake, and how often it wakes. The split between the two is the part worth reading.
Measured, not from the datasheet. Datasheet figures assume every peripheral is off.
How long it stays awake each time.
How often it wakes.
Share of nameplate you actually get. 80 % is a reasonable default.
Runtime
7.78 months
Averaging 0.282 mA from 1600 mAh of usable capacity.
Average current
0.282mA
Runtime in days
236.7days
Runtime in months
7.8months
Where the energy goes
- Awake and working
- 95 % · 0.267 mA
- Asleep
- 5 % · 0.0149 mA
Note
Active time is the thing to fix
95 % of the budget goes to the awake period. Waking less often, or finishing the work faster, buys you runtime directly. Chasing the last microamp of sleep current will not.
Note
Self discharge and temperature are not modelled
A cell loses charge sitting on a shelf, and it loses capacity in the cold. On a node with a multi-year calculated runtime, self discharge can be the larger term. Alkaline cells also sag badly under the pulse load of a radio transmit and recover between pulses, so their usable capacity under a duty cycled load is well below nameplate. Lithium primary cells hold up far better under the same load.
Why this matters
People size a battery from the active current and get a number that is wrong by a factor of ten or more, in either direction. The duty cycle is the whole answer: a node that draws 80 mA for two seconds every ten minutes averages well under a milliamp, and almost all of its energy goes to sitting still. Until you know which half of the budget dominates, every optimisation is a guess. Halving the wake frequency on a node whose sleep current dominates changes nothing at all.
Assumptions and sources
- Average current is a straight time weighted average of the two states. Real hardware has a wake-up ramp, a radio transmit peak, and a settle time that are all higher than the figure you enter as active current.
- The derate is a flat percentage of nameplate capacity. Default is 80 %, which is a rule of thumb covering the fact that a cell does not deliver its rated capacity under a real load down to a real cut off voltage.
- Self discharge is not modelled. On a node with a calculated runtime measured in years, self discharge can be the larger term.
- Temperature is not modelled. Capacity falls in the cold, and a node that runs a year on a bench may not last a season outdoors.
- Alkaline cells sag hard under the pulse load of a radio transmit and recover between pulses, so their usable capacity under a duty cycled load is well below nameplate. Lithium primary cells hold up much better.