SolarHearth

Cooking. A 1.8 kW induction hob runs directly off the 48 V DC bus — no AC inverter stage — delivering ~85% of electrical energy into the pot. A family of five cooks two hot meals on 2.0 kWh/day; an insulated retained-heat basket finishes rice and beans off-coil, holding real-world use near that figure. Displacing charcoal or firewood saves $15–40 a month and removes indoor smoke and the protection risks of fuel collection.

Cooking must fit the kitchen, not the other way around. 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 (and ugali's vigorous stirring is why the hob gets a pot-locking rim). Cooking demonstrations and a familiar pot shape have mattered more to clean-cooking adoption in these camps than any efficiency number.

Biologically safe water — two barriers, because the failure modes differ. Camps mostly have water quantity; the recurring failure is quality at the household storage bucket. The core therefore treats the family's ~20 L/day drinking-and-cooking supply twice:

Barrier one: chlorine, generated on-site. A brine-electrolysis cell (WATA-class, MMO-titanium electrodes) turns table salt into sodium hypochlorite at ~4.5 Wh per gram of chlorine. Dosing at 2 mg/L costs 0.009 Wh/L — and chlorine is the only barrier that leaves a residual, protecting water for the next day inside the storage container. The same cell makes outbreak-response surface disinfectant. Ships with hydrogen venting and DPD residual test strips.

Barrier two: heat, for what chlorine misses. Chlorine fails against Cryptosporidium and loses potency in turbid water. The counter is a flow pasteurizer: a coil-pot on the induction hob holding 65–70 °C for six minutes (full pasteurization — boiling is unnecessary), with a counterflow heat exchanger recovering 80% of the heat. Cost: 12.3 Wh/L, versus 71 Wh/L for a rolling boil. A UV-C flow cell (2 Wh/L) gives a third, instant option for clear water.

Making chlorine is the easy half; a decade of field evidence says the design decision that determines whether water actually gets treated is where the dose is introduced. Manual household chlorination — a bottle of hypochlorite and a measuring cap — works in supervised trials and then collapses in sustained use, because it demands a measured daily action from every household. The work of Amy Pickering's group (now at UC Berkeley) showed the alternative: move the dose upstream, into the water's path, so treatment happens without anyone doing anything.

In urban Dhaka, Bangladesh, her team ran a double-blind cluster-randomized trial across 100 shared water points fitted with passive in-line dosers that chlorinate automatically, without electricity, as water flows to the tap. Households collected water exactly as they always had. Treated taps carried a detectable free-chlorine residual 83% of the time (mean 0.37 ppm) against 0% at control taps, and child diarrhea fell by nearly a quarter. In Kisumu County, Kenya, the group field-tested an electricity-free Venturi doser at commercial water kiosks: flowing water pulls liquid chlorine into itself through the Venturi effect — no moving parts, 97% dosing consistency, 98% of all water sold carrying a residual and 86% inside the 0.2–1.2 mg/L target band. Just as telling: kiosk owners paid for it — over two-thirds bought the device outright by the end of the six-month lease. The lesson generalizes: chlorination succeeds as ambient infrastructure and fails as a daily chore.

SolarHearth adopts this wholesale. The electrolysis cell never asks a household to measure anything: its hypochlorite fills a small reservoir feeding an in-line Venturi doser on the unit's treated-water tap, so every liter drawn is dosed on its way into the storage container. At the cluster kiosk the configuration is exactly Kisumu's — a Venturi at the point of collection — with one upgrade: the on-site cell replaces the purchased-hypochlorite supply chain, so the consumable is table salt. The dose is tuned to the same 0.2–1.2 mg/L residual band — enough to protect stored water through the next day, below the taste threshold that drives quiet abandonment — verified by the test module's DPD strips, and preserved by a narrow-mouth, tapped storage container (wide-mouth buckets are the classic recontamination pathway). Manual cap-dosing remains only as the fallback mode.

1 · PV rear-duct preheat (PVT). Modules run 25–30 °C above ambient; a sheet-metal duct behind the array captures that waste heat to preheat the pasteurizer's feed water by ~10 °C — cutting its load about 20% — while the airflow cools the modules a few degrees for a 2–3% PV yield gain. 2 · Counterflow recovery. The pasteurizer's outgoing hot water preheats incoming feed at 80% effectiveness; this single exchanger is what turns 71 Wh/L boiling into 12.3 Wh/L pasteurization. 3 · Retained-heat cooking. An insulated basket finishes simmer dishes off-coil, trimming ~15% of cooking energy and freeing hob time. 4 · Electronics waste heat. The MPPT and driver electronics warm a 35 °C incubator pocket used by the water-test module for overnight coliform presence/absence tests. The optional water-from-air module adds a fifth coupling: a glazed solar-thermal box (55% efficient) does primary sorbent desorption, with the induction coil topping up through the cartridge's steel susceptor shell.

Why a battery and not a flywheel: a flywheel sized for the evening meal keeps only ~54% of its charge overnight (at an optimistic 5%/hr standby loss) and costs ~12× LiFePO₄ per kWh. Battery, decisively.

Against a cooking-sized power system these loads are noise — under 0.5 kWh/day combined — so each module's true cost is its hardware plus a sliver of PV and battery. Mesh Wi-Fi/LoRa makes every unit a network tile: camp-wide alerts, distribution announcements, and the telemetry that tells an operator which of 10,000 units needs a visit — the maintenance economics of the whole fleet ride on that channel. Phone charging restores family links, mobile money, and agency contact. Sensing pairs the TDS/turbidity electronics with the incubator pocket; households verify their own storage hygiene, and aggregated over the mesh the camp gets a live water-quality map.

Camp water is already a market. Kiosk vendors, cart sellers, and tap operators supply most of what families actually use each day — and a device that arrives as "free replacement water" makes organized enemies of them, which is how good water technology fails politically. SolarHearth's household output is deliberately narrow: ~20 L/day of treated drinking and cooking water. The 60–100 L a family uses for washing, bathing, and laundry still comes from the existing market. The unit is a quality layer over the volume economy, not a volume competitor — and the network is designed so vendors experience it as new business.

The quality map grows a price layer. The sensing module's camp-wide water-quality map generalizes naturally into market infrastructure: any vendor who joins lists their water point with its location, its live price, and a verified-safe badge fed by the same instruments households already use — DPD chlorine residual, turbidity, the overnight coliform test. Real-time price visibility does the ordinary market work: households route to value, transparency disciplines shortage gouging without an administrator setting prices. But the badge is the recruitment tool. The Kisumu result in §2 — kiosk owners buying chlorination hardware with their own money once it won them customers — is the demonstrated behavior this network scales.

What joining gets a vendor: placement and demand routing on the map; the verified-safe badge as a marketing asset agencies and households both trust; chlorine as a service — cluster-kitchen electrolysis cells wholesale hypochlorite refills for vendor Venturi dosers, replacing purchased bleach and tablet supply chains with camp-made chlorine at table-salt prices; and a recorded quality-and-price history that functions as a business credential when agencies procure water locally during distributions. The vendor's consumable cost falls, their product becomes provably safe, and their stall becomes findable.

Three system tiers, one network. Household units self-supply their 20 L/day. Household-group cluster kitchens treat ~375 L/day, host the fridge, and run the chlorine wholesale hub. Restaurant systems serve the real food economy of camps like Kakuma and Dadaab — a wider hob array and 100–200 L/day of treatment capacity, at roughly cluster-kitchen cost (~$3k) — and become the network's anchor nodes: a food stall visibly cooking on clean power with a verified-safe badge normalizes the standard for every customer who eats there. The result is a deployment that enters as infrastructure the existing market plugs into, rather than as a subsidized competitor that must first defeat it.

Everything below rides on hardware already specified — the bus, the pasteurizer, the chlorinator, the fridge, the mesh — plus at most one extra PV module. The discipline stays the same: each extension is priced by its marginal energy and hardware, and nothing here changes the core bill of materials.

Tea. 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, and the pasteurizer's heat exchanger already delivers ~70 °C water to a hot-water tap, halving time-to-boil. A tea-stall tier — hob, large kettle, verified-safe badge — is the smallest restaurant system, roughly $800.

Goat milk. Pastoralist milk economies ring Kakuma, and spoilage is a chronic loss for the women who dominate the trade. The flow pasteurizer is a milk pasteurizer at different settings (63 °C for 30 minutes, in a cleanable foul-tolerant coil variant), the cluster fridge is the chiller, and the incubator pocket runs milk hygiene tests. A cluster milk bay — pasteurize, chill, test, badge — brings dairy vendors onto the map exactly as water vendors join.

Grain milling. Posho mills are a real vendor economy currently run on diesel. Electric milling costs ~0.1 kWh/kg, so a 10 kg/day service fits one added PV module at a cluster kitchen.

E-bike battery swap. Standard e-bike packs are 48 V — the bus voltage. Nowhere does this matter more than in hilly settlements like Kibera, where the water economy is literally carried uphill: a full jerrycan weighs 20 kg, the water point is at the bottom of the valley, and the household is at the top, so every day of water costs a woman or a child repeated loaded climbs. One 0.5–0.7 kWh swap pack hauls dozens of jerrycans up those same hills on an e-cargo trike — the battery does the climbing. A swap shelf at the cluster kitchen (≈ one extra PV module per few daily swaps, +$150–300) creates a charging business, a paid delivery trade for the trike owner, and vendor water that arrives at the door at a price that competes with the customer's own carrying time. Swapped packs double as home backup storage.

Fire detection. Dense camps burn catastrophically — the 2021 Cox's Bazar fire destroyed some 10,000 shelters. SolarHearth attacks the root cause (flameless cooking) and the mesh turns $3 smoke/heat sensors on every unit into a camp-wide early-warning grid: cluster sirens, fire location on the map for responders, negligible energy.

Diarrhea surveillance. Households report cases with one button on the unit (or a phone short-code); the map overlays case clusters on the chlorine residual and coliform data it already carries. Clustered failures plus rising cases are exactly the early-warning signature agencies want, days before clinic data shows it. Reporting is aggregated by block, never displayed per household. Cluster kitchens stock oral rehydration solution mixed with verified-safe water.

Cholera — anomaly detection, not diagnosis. Specific Vibrio identification stays with clinic labs. The network's honest role is the surveillance above plus the response already built in: cells surge to outbreak dosing and produce 0.2–0.5% surface disinfectant on demand.

Theft-registered solar. Shelter roofs (tarp, sheet metal) cannot carry glass modules, and theft is the documented failure mode of camp PV. Panels stay on the lockable ground rack, but the mesh is the real lock: every module is network-registered, a panel leaving its geofence or dropping off-mesh raises an alarm, and MPPTs accept only registered serials. The network is what makes household solar insurable.

The mosque. Wudu — ablution before each of the five daily prayers — is a substantial, non-negotiable clean-water demand, and the mosque is among the most trusted institutions in most camps. A mosque tier pairs a storage tank with Venturi-dosed taps and a wudu station (dosed at the low end of the residual band, where chlorine is imperceptible), with wudu graywater reused for gardens. A trusted institution running visibly on the badge normalizes it for the whole community, and Friday announcements are the adoption channel money can't buy. The same logic extends to churches and schools.

Newborns. Hygiene at birth and in the first weeks is a leading lever on neonatal sepsis. The unit contributes warm washing water from the pasteurizer's mixing tap, chlorine for surface and hand hygiene, safe water for oral rehydration — and, where formula is unavoidable (breastfeeding remains the promoted standard), the pasteurizer delivers the WHO-specified 70 °C preparation water exactly. The detachable lamp becomes the night-feeding light.

Hand-washing stations. Stations fail by running dry or soapless, not by missing. Cluster and latrine stands run from the treated supply through a mini-Venturi holding a low residual, on foot pumps; a flow sensor puts each station on the map, so an unused station is a visible red flag; soap moves through the vendor network. Students: the detachable lamps double as homework lights — 15 Wh buys three hours of study, the cheapest education intervention on this page — and a motion-sensed 20 W mast light at each cluster kitchen (~240 Wh/day) covers paths and latrine routes, a standing protection request in every camp assessment.

Latrine operators as network nodes. Sanergy's Fresh Life network in Nairobi's informal settlements runs over 6,000 container-based urine-diverting dry toilets — pay-per-use franchises operated by local entrepreneurs, installed by landlords on residential plots, by schools, by markets, and by bars — each required to keep a handwashing station with soap and water, with sealed waste cartridges collected for conversion to fertilizer and insect protein. A dry toilet needs no flush water; its entire water demand is exactly what SolarHearth's smallest node provides: a flow-sensored, chlorine-dosed handwash stand and a cleaning tap. Add an LED light and the latrine works — and earns — after dark, which is precisely when women most need it and when operators currently lose their custom. On the map, the operator's stall carries a verified: clean, lit, open badge that routes users; the node pays for itself out of extended per-use revenue. A bar hosting a Fresh Life unit on its plot gets the same package on the same spur.

What slum businesses buy: margin, not hardware. For the small hotels (in the East African sense — eateries and tea rooms — as well as lodgings), bars, kiosks, and salons of a settlement, each SolarHearth service converts directly into revenue. Evening lighting is the single biggest change — it extends business hours past dark, when wage-earners actually have time and money. The fridge sells cold drinks at bar margins and makes safe ice from verified water — a genuine premium product where ice is otherwise frozen from whatever water was nearest. The badge markets the tea, the food, and the water; phone charging is a paid amenity; the hot-water tap speeds every kitchen. These businesses are the natural paying anchor customers of a settlement deployment — leased service packages priced against measured extra revenue, exactly the lease-to-own behavior the Kisumu kiosk owners demonstrated — and their payments are what cross-subsidize household units, the same direction the slum-scale economics below already point.

Clean clothes are employment infrastructure. Schools require uniforms — often a single owned uniform, washed at night to be worn clean at dawn — and anyone commuting out of the settlement to work needs spotless clothes as the price of being taken seriously. The result is that hand-washing laundry consumes hours of most days for women in these settlements: hauling wash water, scrubbing, wringing, and guarding line-drying clothes against dust and rain. It is among the largest unpaid time costs in a slum household, and it is almost entirely automatable with hardware this page has already priced.

The laundry bay. A cluster-kitchen bay adds two compact machines (3–5 kg twin-tub class, ~0.3 kWh per cold cycle — motor energy only) and, critically, spin dryers (~200 W for five minutes, under 20 Wh a load): spinning removes most of the water, so the overnight-washed school uniform is dry by morning and line-time in the dust drops from hours to minutes. Ten loads a day costs ~3 kWh — two added PV modules — so the whole bay prices near $800–1,000 and sells pay-per-load service against hours of a customer's own labor. Wash graywater is captured for latrine cleaning.

Cold or warm? The detergents already dominant in these markets — bar soap and standard powders — are used at ambient temperature today, and enzyme cold-water formulations are arriving through the same retail channels, so machine cold-washing needs no behavior change. But warmth still earns its keep: enzymes and surfactants work markedly better at 30–40 °C than at tap temperature, which is what gets a white uniform actually white — and the PV rear-duct already delivers that 10–15 °C lift to the wash fill for free, with the hot-water tap available for stain treatment. No heating element in the machine, no new energy budget: warm-wash quality at cold-wash cost.

The COVID lesson was that dense settlements needed exactly four things this unit already carries: hand-washing infrastructure, bulk surface disinfectant (the cell's outbreak mode), a trusted local information channel (the mesh, against rumor), and safe water without crowded touchpoints (dosed taps rather than shared dosing). None of it assumes a refugee camp. Urban informal settlements — Kibera, Mathare, Dharavi, urban Bangladesh — have the same shape: unreliable grid, vendor water, charcoal cooking, no residual chlorine. The flagship evidence on this page already comes from there: Pickering's Dhaka trial ran in low-income urban communities, not camps. The scale path runs the same direction — roughly a billion people live in informal settlements against ~35 million in camps, and slum-market volume is what would drive the $1,533 down and cross-subsidize camp deployment. Carbon finance points the same way: clean-cooking credits are an established channel that could underwrite a meaningful slice of every unit.

Every load this page has priced in kilowatt-hours is currently paid in a different currency: women's and girls' hours. Fuel is collected on foot, fires are built and tended, water is queued for and carried, laundry is scrubbed by hand. Time-use studies across East African camps and settlements put this combined burden at four to six hours of a typical woman's day — the single largest claim on her waking labor. Because SolarHearth's functions map one-to-one onto exactly these tasks, its primary economic output is not electricity or water. It is time.

The arithmetic. Four-plus hours returned daily is roughly 1,600 hours a year — a full-time job's worth of labor capacity per household. Valued even at informal-sector shadow wages of $0.25–0.60/hour, that is $400–950 a year, against the unit's all-in cost of $264–360 a year. The unit pays for itself in returned time alone, before counting a shilling of fuel savings or vendor revenue. And the two ledgers compound: time converts to income only where there is work to sell it into, and the network is itself the venue — the tea stall, the milk bay, the laundry bay, the latrine franchise, the kiosk, the trike delivery route are all operator businesses, and in these settlements they are overwhelmingly women's businesses. SolarHearth frees the hours and supplies the market that buys them.

The girls' share. Water hauling and laundry fall heavily on daughters, and hours returned to a girl are measured in school attendance — compounding over a lifetime in a way no annual figure captures. The homework lamp is the visible education intervention on this page; the hours are the invisible one, and they are larger.

Honesty. These are mid-range planning figures, not measurements; a deployment should measure its own baseline (the time-use survey belongs in the pilot protocol) and let the map's usage data report the realized change. The claim that survives any reasonable parameter choice: the largest single benefit of this unit is not energy, water, or health — it is the working day it returns to the woman who runs the household.

Chemical contaminants are regional, not universal — fluoride belts (the Rift Valley, parts of South Asia) and arsenic belts (the Bengal basin) are well-mapped — so these columns ship selectively by hydrogeology, plumbed inline ahead of the disinfection barriers. The chemistry sends the two problems to different media: arsenic bonds to iron oxide; fluoride bonds to aluminum phases (iron oxide does little for fluoride).

Arsenic — the easy one. Well arsenic is measured in µg/L, so a granular ferric hydroxide (FeOOH) bed treating 250→10 µg/L at 20 L/day consumes only ~1.8 kg of media a family-year (~$11). The chlorinator sharpens it: pre-oxidizing As(III)→As(V) markedly improves capture. Spent media is hazardous — stabilized in concrete or swapped as a return-to-supplier cartridge.

The proven upgrade path: ECAR. Iron-based arsenic removal has a two-decade, two-continent track record under Ashok Gadgil's group at UC Berkeley and Lawrence Berkeley National Lab. Their ElectroChemical Arsenic Remediation (ECAR) replaces the imported sorbent cartridge with a plate of ordinary mild steel: a small current dissolves the iron in place, and the freshly formed iron-hydroxide flocs bind arsenic and settle out — the consumable becomes scrap-grade steel and a few watt-hours per liter, with no media supply chain at all. In West Bengal, India, a 10,000 L/day ECAR plant commissioned at Dhapdhapi High School near Kolkata has supplied arsenic-safe water (roughly 250 µg/L in, under 10 and typically under 5 out) to a school community of ~3,000 since 2016, free to students — and, handed to a local operator in 2017, it covers its costs by selling surplus safe water to the surrounding village at well under a US cent per liter. In Allensworth, California — a historically Black-founded Tulare County town on arsenic-contaminated groundwater — the same technology came home: a 2019 farm field trial grew into an EPA-funded small-community system with remote monitoring, holding output consistently below the WHO 10 µg/L guideline, with current work on zero-liquid-discharge operation. Gadgil received the U.S. National Medal of Technology and Innovation in 2023 for this body of work — which also includes UV Waterworks and the Berkeley-Darfur Stove, the two direct ancestors of this unit's disinfection and cooking cores. For SolarHearth, ECAR is the natural cluster-scale arsenic bay where wells run hot: steel plates and a trickle of DC from the same 48 V bus, with the household FeOOH cartridge as the entry-level option.

Fluoride — the bulk-media one. At 8→1.5 mg/L the same 20 L/day consumes 24–32 kg of media a year. Activated alumina (~$95/yr, regenerable with NaOH then alum) is standard; bone char (~$19/yr) can be made on-site by charring bones in a retort pot on the induction hob at 400–500 °C — media without a supply chain. Units ship with either cartridge.

A note on bone char. It is not an improvisation: bone char is the longest-proven defluoridation medium in East Africa — the Catholic Diocese of Nakuru has run bone-char defluoridation in Kenya's Rift Valley for decades, exactly the fluoride belt these camps sit in. Char quality matters (under-charred bone tastes and smells; over-charred loses capacity — the retort recipe is temperature-controlled by the induction hob itself, one advantage over open-fire charring). And acceptability is a per-community question, not a footnote: bone origin carries religious and cultural weight — communities that keep halal practice may accept only verified cattle or goat bone, and some households will refuse bone-derived media entirely. The deployment rule: offer bone char where the community accepts it after consultation, default to activated alumina where it does not, and never make bone char the only option.

Where aquifers are saline or contested and trucking is failing, a MOF (metal–organic framework) module harvests drinking water from night air: the sorbent bed adsorbs at night, desorbs at midday — glazed solar-thermal box first, induction top-up through the cartridge's steel susceptor shell second — and condenses on an ambient radiator. Physics sets the price: ~3.1 MJ of desorption heat per liter; with 40% recovery, 0.62 kWh of electricity per liter — 310× the cost of purifying water you already have. 20 L/day per household is feasible; 100 L/day exists only as a shared 5-household cluster unit (~17 kW of PV, roughly 80 m² of panels).

Deployment logic: every household gets power + core + modules ($1,423); the arsenic and fluoride columns ship by hydrogeology; the fridge and camp-scale chlorinator live at the cluster kitchen; the water-from-air module goes only where source water genuinely fails.

All figures derive from a parametric model, not vendor quotes: solar resource 5.5 kWh/m²/day with a 0.78 derate; PV at $0.45/W installed small-scale; LiFePO4 at $160/kWh usable; pasteurization at 188 kJ/L with 80% recovery; electrochlorination at 4.5 Wh/g; MOF working capacity 0.22–0.42 L/kg/cycle by humidity with desorption at 3.1 MJ/L thermal and 40% recovery in active mode. The figures most likely to move: MOF sorbent price ($30–80/kg today; ~$10/kg at scale would cut water-module capex 20–40%) and night humidity, which sets sorbent mass 2:1 across climates. Nothing here is a product; it is a feasibility envelope for one.

Each chapter takes one technology from this page and goes deeper: the physics, the specifications, the field evidence, and the open questions.