🏠 Homelab
Battery runtime, solar sizing, power consumption and Docker RAM calculators for self-hosted setups
Calculate how long a battery will power your device — input Ah, voltage and watts
Open tool →Size your solar panel and battery for off-grid Raspberry Pi and homelab setups
Open tool →Calculate voltage, current, resistance, or power from any two known values.
Open tool →Plan container memory — pick services from presets and check against available RAM
Open tool →Sizing tools for self-hosted infrastructure — the calculations that decide whether a build actually survives contact with reality. How long a battery holds a server up, how much solar you need to run one continuously, what a rack draws from the wall, and how much RAM to give a stack of containers.
These came out of building the things they describe. The numbers they produce are the ones worth checking before you buy hardware, because the failure mode of guessing is either an expensive over-build or a server that dies every time the sun goes behind a cloud.
Which tool do I want?
| If you… | Use |
|---|---|
| You want to know how long a UPS or battery lasts | Battery Runtime |
| You are planning an off-grid or solar-backed setup | Solar Panel Sizing |
| You need total draw and running cost for a rack | Power Calculator (V/I/R/W) |
| You are deciding how much RAM a container stack needs | Docker RAM Planner |
Battery capacity is not the number on the label
A 100 Ah lead-acid battery does not give you 100 Ah. Discharging lead-acid below about 50% wrecks its cycle life, so the usable figure is nearer 50 Ah, and that is before inverter losses of 10–15% and the capacity that Peukert's law eats at higher discharge rates. A rule of thumb that has never let me down: take the label figure, halve it for lead-acid, then remove another 15%.
LiFePO4 changes the arithmetic. You can take it down to 20% without meaningful damage, and it holds voltage far more steadily as it drains, so the usable fraction is more like 80%. It costs more per nominal amp-hour and less per usable one, which is the comparison that actually matters.
The other number people miss is idle draw. An inverter running with no load still burns 10–20 W, which is around 0.4 kWh a day doing nothing at all. On a small off-grid system that overhead can exceed what the server itself uses.