L1 On-Board Solar Charging

Bottom line. On-board solar is worth fitting to the VTOL and not, as a range extender, to the quad. Wing-integrated cells refill about three quarters of the VTOL pack on a clear summer day and add 30–36% endurance in daylight flight. The quad's body panel returns 38% of its pack per clear summer day but costs 4–6% endurance in every flight. With the mounting and rules below, on-board solar is safe under every single and double fault simulated and never conflicts with the pod (simulation package, section 13).

No airframe data was supplied, so both airframes are representative assumptions. These are not figures for the L1 aircraft: the L1-Q and L1-W each carry about 194 Wh, and the L1-Q about 50 W of solar, so they have been rerun with their own values under Applying the design to the L1 aircraft. Every value sits in drone_solar.py for replacement.

Temperature: mounting decides it

The pack must not sit directly under the panel. The panel runs 35–50 K above ambient; directly under it the face cells sit about 20 K above ambient.

Direct mounting also fails a single-sensor fault: with the face sensor reading 10 K low, face cells charged above 45 °C while the core BMS still read in range. With the vented gap the panel shades the body, so the pack runs no hotter than on the same airframe with no panel. Even vented, the VTOL pack reaches 62 °C parked on a dark roof at 48 °C in still air: do not park on dark roofs above 45 °C.

Cold. No charge below 2 °C (charger) or 0 °C (BMS), and at most 0.1 C below 10 °C (lithium plating). The solar-fed heater lets the VTOL take 333 Wh over two −6 to 4 °C March days, against 190 Wh without it. At 55 °N in December neither airframe charges: the sun cannot power the heater.

Daily yield, day 1 from 20%, vented mounting

Flight: endurance and heating

In flight the panel output feeds the motor bus through its own gated output, limited to the load. It never charges the pack in flight.

The quad's pack temperature is set by flight, not the sun. An 18-minute hover at about 2 C raises the pack 30 K; the VTOL's 90-minute flight raises it 6.5 K. At 30 °C ambient the quad pack needs 84 minutes parked before solar can start (40 °C) and 133 minutes before it is presented to the pod (35 °C). High-power cells with half the resistance halve the rise; specify them for the quad.

Electrical design and pod hand-off rules

The on-board charger shares the pack's charge port with the pod rectifier. These rules keep the two chargers from conflicting; each was checked in simulation.

1. Same setpoint. CV 24.6 V (4.10 V per cell), ±0.5%, so at most 24.72 V: below the pod's hardware OV band (24.89 V minimum) even if both chargers are on.
2. Own protection. A latched OVP comparator at 25.05 V, below the BMS 4.25 V per cell. It resets when the panel goes dark at night; three trips lock the charger out. It refuses packs below 15 V and pre-charges at C/20 below 18 V, so it can bring a 15–17.6 V pack back into the pod's sense window.
3. Gated by the BMS. The charger's output switch closes only when the BMS allows charging. If solar could drive the rail while the BMS charge FET is open, the pod would read the rail as a full pack and never charge. The simulation shows that failure when the gating is removed.
4. Pod has priority. A hardware comparator on the contact-rail current shunt disables solar within 100 ms of pad current above 0.2 A. Solar re-enables only after 5 s below 0.05 A, so the pod's 20 ms re-sense pauses never re-enable it. With solar off, the pod's taper rule works as designed.
5. Two sensors at the two extremes. The charger's NTC sits on the cells nearest the panel; the BMS NTC sits in the core. Charging needs both inside their windows: solar 5–40 °C to start, 2–44 °C to run, derated above 40 °C; BMS 0–45 °C.
6. Pack presented to the pod only at 10–35 °C. Outside that range the drone holds its charge FET open, so the pod sees an empty pad and keeps sensing every 2 s. The drone also opens the FET if either sensor falls below 10 °C during a pod charge. This stops the pod's 1 C charge from plating lithium in cold packs and from overheating packs that land hot.
7. BMS overvoltage releases only after discharge (below 4.05 V per cell). Otherwise a 5S pack fitted by mistake is pushed to 4.25 V again on every pod sense cycle.
8. Heater fed from solar only. It is switched from the charger output with the charge switch open, so it never draws from the pack or the pod. It runs only when both sensors agree the pack is cold (core below 15 °C), so a drifted face sensor cannot overheat the pack.
9. In-flight output. A separate ideal-diode output onto the discharge bus, enabled only when armed, inside the charger's temperature window and not locked out. It is limited to the measured bus load, so it never net-charges the pack in flight.
10. Sleep on the pod. The VTOL's awake draw (0.39 A) exceeds the pod's 0.3 A avionics budget and the quad's (0.29 A) leaves no margin, so both must sleep during a pod charge (19–29 mA).

The MPPT input stays in range at both limits: 25.2 V maximum string Voc at −30 °C (deployable kit) against 80 V; at least 10.4 V Vmp at 70 °C and 200 W/m² against a 6 V minimum.

Verification of on-board solar

The model covers the sun (position, clear-sky beam and diffuse, hourly cloud), the cells (temperature coefficient, bypass diodes, MPPT input floor), a transient thermal network (panel sections, airframe interior, a two-node pack), the 6S pack with its BMS, the charger, the pod's charge logic and flight loads. Results:

• Integrated missions. Three days per airframe, 14 scripted events each, all pass: daylight flight with solar on the bus, warm landing, wilderness and dark-roof parking, a pod visit with the interlock, cloud, 50% shade and a cold front with the heater. On the pod the quad landed hot, was held until 35 °C, was interrupted once by the 45 °C limit in sun, then resumed and ended on taper.
• Stress campaign. 400 random two-day scenarios (−30 to 48 °C, latitudes to 60°, all seasons and ground types, random flights and pod visits) with 11 fault types injected singly or in pairs. Result: no safety violation with one, two or three faults, and no refusal without a physical reason. Without a CV fault, cells never exceeded 4.10 V while charging; with one, 4.248 V (the BMS cut-off).
• Named extremes. 32 cases, all as expected. These include a CV runaway with the charger's OVP failed (the BMS cuts at 4.25 V), an interlock failure through a 3-hour pod charge (both chargers at 24.6 V, no fault), and the two direct-mount sensor-fault cases that show why direct mounting is rejected.
• Independent review. It found 10 defects in the first model. All were corrected before the results above: shade was counted twice; the panel's electrical draw was charged at full MPP power even when the charger drew less; one lumped pack node hid the face cells under the panel; the heater ignored the interlock and lockout; refusals were judged by the controller's own flags rather than the true state; the pod could charge at 1 C below 10 °C; and the in-flight output had no defined path.
• Tests. 37 pass/fail checks in tests/test_drone_solar.py; the full suite is 165 checks.

Limits. The airframes, panel-to-body coupling coefficients, pack resistance, cell data and flight power are representative values. Replace them with measurements:

• a sun test of the real airframe with thermocouples on the face and core cells;
• a flight-heating test at the planned hover time;
• the cell maker's charge-temperature and charge-rate limits, which must permit the pod's 1 C.