Photovoltaics & Storage
Everything on the unit runs from one small solar array and one battery on a 48 V DC bus. This chapter gives the full electrical design: why the array is sized to 640 W, why the bus is DC end-to-end, and why the battery is lithium iron phosphate and not a flywheel.
Sizing from the load, not the panel
The design load is 2.73 kWh/day: 2.0 kWh of induction cooking, 0.26 kWh of water treatment (pasteurizer, UV-C, chlorinator), and under 0.5 kWh of accessories (lights, phones, mesh node, radio, fan, sensors). At a design solar resource of 5.5 kWh/m²/day — typical for East Africa, the Middle East, and South Asia — with a 0.78 system derate for dust, temperature, and wiring, each watt of PV delivers about 4.3 Wh/day. The load therefore needs ~635 W of array, which two standard 320 Wp modules supply with a 1% margin: 640 W × 4.29 = 2.75 kWh/day.
- Modules2 × 320 Wp monocrystalline PERC, 60-cell, tempered glass, IP65 junction box
- Dimensions1665 × 1002 × 35 mm per module · 18 kg each · array area 3.4 m²
- ElectricalVmp 33.6 V, Imp 9.5 A per module; wired 2-series: Vmp 67 V, Voc 82 V (90 V cold)
- Charge controlMPPT, 100 V max input, 20 A, into the 48 V DC bus
- BatteryLiFePO4 51.2 V / 40 Ah = 2.05 kWh usable · 6,000 cycles to 80% · ~18 kg
- MountingGround or roof rack, tilt 5–15° (site latitude), equator-facing; wind-rated frame; weekly dust wipe assumed in the 0.78 derate
- ExpansionRack and MPPT accept a third module + 1 kWh battery for the fridge, laundry bay, or water-from-air options
- CostArray $286 ($0.45/W installed small-scale) · battery $320 ($160/kWh usable)
Why the bus is DC end-to-end
There is no AC inverter anywhere in the unit. The induction hob is driven by a resonant DC-fed stage directly from the 48 V bus; a 48→12 V/5 V converter feeds lights, comms, sensors, the UV-C cell, and the chlorinator. Every conversion stage removed is 3–8% of energy and one field-failure mode fewer — and 48 V is below the 60 V DC touch-safe threshold, so household maintenance does not require an electrician. The choice pays a bonus later: standard e-bike packs are 48 V, which is what makes the battery-swap economy of the extensions chapter possible without additional electronics.
Battery sizing follows the evening: 1.2 kWh for the dusk meal (the main meal in the camps this unit serves — a daylight-only cooker was rejected on cultural grounds, not technical ones), ~0.5 kWh of overnight lights, comms, and fans, and buffer. LiFePO4 at 6,000 cycles is 16 years of daily cycling — twice the 8-year design life — and is the safest common lithium chemistry in a hot, dusty enclosure.
The flywheel verdict
The design brief asked for battery or centrifugal storage, so the flywheel got a fair hearing and lost on arithmetic. A flywheel sized for the evening meal (1.5 kWh at dusk) retains only ~54% of its charge after 12 hours even at an optimistic 5%/hour standby loss — low-cost bearings do worse — and small steel flywheels cost roughly $2,000/kWh against $160/kWh for LiFePO4: 12× the price to lose half the dinner by dawn. Flywheels make sense for seconds-to-minutes surge buffering, but the induction driver's surge is already handled by $20 of extra battery headroom. Physics, not fashion, picked the battery.
Theft, the real enemy of camp solar
Shelter roofs — tarp and sheet metal — cannot carry 18 kg glass modules, and flexible thin-film laminates cost more per watt and degrade fast in equatorial heat. So modules live on a lockable ground rack, and the mesh network is the real lock: every module is network-registered, a panel that leaves its geofence or drops off-mesh raises an alarm, and MPPTs accept only registered serials. A panel that cannot be quietly resold is a panel not worth stealing — the network is what makes household solar insurable.