Induction Cooking

A hob is the cooktop burner itself — the heating coil and its power electronics, without an oven under it. An induction hob is a transformer with the pot as its secondary winding: a 20–100 kHz alternating field in the coil induces eddy currents in a ferromagnetic pot base, and the pot itself becomes the heating element. About 85% of electrical energy ends up in the food — against ~40% for LPG and 10–15% for open-fire charcoal cooking — and the hob surface stays cooler than the pot, which matters in a shelter full of children. SolarHearth's 1.8 kW single-coil hob is driven by a resonant stage directly from the 48 V DC bus (no inverter), costs about $65 at commodity scale, and gets a pot-locking rim for the vigorous stirring that ugali demands.

A family of five cooks two hot meals on 2.0 kWh/day. The same coil moonlights: it regenerates the optional water-from-air cartridge through its steel susceptor shell, drives the flow pasteurizer pot, and — at 400–500 °C in a retort pot — chars bones into fluoride-removal media (chapter 6).

In the camps this unit is sized for — Mahama in Rwanda, Kakuma and Dadaab in Kenya — the staples are long-simmer dishes: dried beans (2–3 hours), githeri, maize porridge, ugali. Three consequences are designed in. Timing: the main meal is cooked at dusk, after the sun — which is why the battery is sized for the evening meal and a daylight-only cooker was rejected. Long simmers: beans are brought to boil on the hob, then finished in the insulated retained-heat basket — without it, bean days would run ~3 kWh instead of 2. Pots: the ubiquitous East African sufuria is aluminum, which induction cannot heat — so the unit ships with two flat-bottomed stainless-clad sufuria-pattern pots. Cooking demonstrations and a familiar pot shape have mattered more to clean-cooking adoption in these camps than any efficiency number.

The retained-heat basket (a modern haybox: an insulated cavity that lets a boiled pot finish cooking on its own stored heat) is standard equipment, not an accessory — it trims ~15% of cooking energy, frees the hob for the next pot, and needs no power at all.

Camp households typically spend $15–40 a month on charcoal or firewood. The core cooker — 466 W of PV, 1.5 kWh of battery, the hob, enclosure and controls — prices at $655, or $8.50/month amortized over 8 years: it repays itself in displaced fuel in 1.5–3.5 years before counting anything else. What the fuel line doesn't show: indoor smoke from solid-fuel cooking is a leading child-health hazard, firewood collection is a documented protection risk for women and girls, and flameless cooking removes the ignition source behind catastrophic camp fires. Clean-cooking carbon credits — an established financing channel — can underwrite a meaningful slice of the capital cost.

Tea deserves its own line. Sweet milky chai is a several-times-daily ritual across Somali, Sudanese, and Congolese communities, and tea stalls are often the most numerous camp enterprise. A few 1–2 L boils cost 0.1–0.2 kWh/day, the pasteurizer's heat exchanger already delivers ~70 °C water to a hot-water tap that halves time-to-boil, and a tea-stall tier — hob, large kettle, verified-safe badge — is the smallest business system in the lineup at roughly $800.

The battery-plus-induction appliance is no longer hypothetical. Copper (formerly Channing Street Copper Company, made in Berkeley) builds Charlie, an induction range with a 34 kg LiFePO₄ battery under the oven that charges from an ordinary 120 V outlet — a gas-to-induction swap with no rewiring, which is SolarHearth's architecture wearing a grid plug instead of a PV array. In November 2025 Copper won New York's Induction Stove Challenge and signed a $32 million contract with the New York Power Authority, NYCHA, and NYSERDA: 100 prototype stoves are going into 100 public-housing apartments — cookware provided to each household, the same lesson as our sufurias — with a follow-on of up to 10,000 units gated on pilot results. The precedent matters twice over: it proves the integrated LFP-plus-induction appliance at housing-authority scale, and it establishes the group-pilot deployment model (institution buys, residents test, scale follows evidence) that a camp or settlement deployment of SolarHearth would reuse.

The coil ceiling. The winding literature — led by the Universidad de Zaragoza appliance group, with Chinese contributions on twisted-litz AC-resistance computation — shows optimized litz-wire designs cutting winding losses 18–36% against conventional single-bundle windings (litz wire is a bundle of many thin, individually insulated copper strands woven together so that high-frequency current uses the full copper cross-section instead of crowding to the conductor surface — the skin effect), by tuning strand count and diameter against skin depth and turn placement against the local field; GaN and SiC drivers push switching frequencies above the traditional 20–40 kHz, shrinking coil copper and improving coupling. The honest read: total cooktop efficiency is ~84% (the LBNL/DOE measurement, against ~40% for gas), winding loss is one slice of the remaining 16%, so a state-of-the-art coil buys a 1–3% system gain — worth taking as commodity boards improve, but an order less than the retained-heat basket's 15%. The larger levers stay unglamorous: pot-base coupling (the quality of the ferromagnetic layer in the clad sufuria) and driver topology.

China's deeper contribution is not the research but the manufacturing base: the Guangdong and Zhejiang induction-cooker industry (Midea, Supor, Joyoung and their contract factories) ships portable single-burner cookers in the tens of millions a year, which is why a $65 hob line-item exists at all. Inside every one is essentially the same board: a mains rectifier and filter; a single-switch quasi-resonant inverter built around one 1200 V-class IGBT; the resonant tank — a litz-wire pancake coil, typically 140–200 mm across and a few tens of microhenries, with its parallel resonant capacitor; ferrite bars under the coil to guide flux into the pot; pot-detection, an NTC temperature sensor, a fan, and an 8-bit microcontroller. Produced at that volume, the whole power stage costs a few dollars — it is the natural donor hardware for a camp hob.

Adapting it to the 48 V bus is the one real engineering task, because the commodity board expects rectified mains (~170–325 V DC link), not 48 V. Two routes. Route A — boost and reuse: a 48→300 V boost front-end feeds the commodity power stage with its rectifier deleted; minimal redesign, but the boost stage adds cost and a conversion loss. Route B — rewind and switch low: keep the coil geometry, ferrites, and controls, but rewind the pancake with fewer turns of heavier litz and drive it with paralleled low-voltage MOSFETs in a half-bridge — at 48 V, 1.8 kW means ~38 A, well within cheap automotive-class FETs, and the deleted rectifier and boost stages claw back their losses. Route B is the better endpoint (fewer stages, higher efficiency, cheaper at volume); Route A is the faster pilot. Either way the bill of materials rides on a supply chain that already exists at the tens-of-millions scale — the camp hob is a variant, not an invention.

What the pressure-cooker pilots teach. The MECS / Bayes Consulting DC-EPC battery-swap pilot (Got Ngur, northern Uganda, 2026; 50 households, 3 months) is the best-documented field test of electric pressure cooking off-grid, and it is candid. The cooker itself performed — 96% reliability, 0.755 kWh per cooking event, beans from ~4 hours on firewood to 30–45 minutes — but 40% of households had the 1.4 kWh swap battery die mid-meal, revenue covered 1.04% of operating costs, charging demand pushed the host mini-grid's load up 38.7% until 11 cookers were withdrawn, and no household abandoned firewood: all kept it for tea, posho, and the dishes a sealed pot cannot make. Three lessons land directly here: energy budgets must be sized to measured demand; centralized charging rations itself on a constrained grid (per-household PV does not); and a point-solution appliance leaves the fire lit.

Adopt the physics, not the appliance. Pressure is the right pot, not necessarily the right machine: a stovetop pressure cooker with an induction-compatible clad base ($25–40) reaches 110–117 °C and cuts bean cooking to ~35 minutes at ~0.4–0.5 kWh — versus 2.0 kWh on the open hob — with no added electronics, and it stacks with the retained-heat basket. Since beans are the long-simmer core of the 2.0 kWh/day budget, this is the largest single cooking-energy lever in the whole design.

Frying is a requirement, not an afterthought. Mandazi, sambusa, and bhajia are daily foods and a large share of the street-vendor economy, and no pressure vessel or hotplate serves them well. Induction does: a heavy karai holds thermal mass, the coil responds in seconds, and a temperature-limited frying mode caps oil below flashpoint — making the hob the safest deep-frying platform available in a dense settlement, where oil fires are a real hazard. For vendors, the tea-stall and restaurant tiers carry the duty cycle. The claim the stack argument earns: boil, pressure, simmer, fry, ugali, tea — one energy bus, every pot, no dish left to the fire.