A drone that flies for 32 minutes in March can fly 19 in August, throttle back on its own on the fourth flight of the day, and shut a motor down at altitude in Colorado. Nothing broke. The aircraft got hot, and heat is the constraint that quietly sets the performance envelope of almost every electric aircraft.
Thermal design on a drone is unusual because the machine generates its own cooling airflow — then loses it at the moment it needs it most.
Where the Heat Comes From
Four sources dominate, and they are not equally important.
- Motors. A brushless motor at hover runs 80 to 88 percent efficient, so one drawing 300 W dumps 35 to 60 W into its windings and magnets — times four, six, or eight. Their limit is the magnet temperature at which they begin to demagnetize, typically 100°C to 150°C by grade.
- ESCs. Switching and conduction losses. A good ESC may dissipate only 5 to 15 W, but concentrated in a few small packages on a board with nowhere to send it.
- The battery. Internal resistance times current squared. Unlike the motors, a pack's efficiency falls as it warms, and its safety margin with it.
- Payload and compute. A vision computer draws 10 to 30 W continuously in a sealed pod with no airflow — often the hottest thing on the aircraft per unit volume.
Two of these have a nasty feedback property: motor resistance rises with temperature, so a hot motor draws more current for the same torque and heats further, and battery resistance behaves similarly. That slow upward drift is why the last five minutes of a mission are the risky ones.
Prop Wash Is the Heat Sink
A multirotor sits in its own downwash: air moving at 15 to 30 mph below each rotor. Any component in that column gets forced convection an order of magnitude better than still air, for free.
That drives the layout of every well-cooled drone. Put the ESCs under the motors, heat-sinked face toward the flow; put fins where air moves; put nothing hot in a dead-air pocket unless it has a conduction path out.
The catch is that the wash disappears on the ground. A drone that just landed after a hard flight has no airflow and soaks upward. Peak temperature on many aircraft occurs 30 to 90 seconds after landing, so test data covering only flight misses the worst case.
Forced convection also fights sealing head-on, since every vent is an ingress path — which is why weather-sealed aircraft run hotter and derate earlier. The general toolkit is in thermal management in electronic products.
Conduction Into the Frame
When airflow is unavailable, heat has to be conducted somewhere with mass and surface area. On a drone that somewhere is the frame, and the frame material decides how well it works.
Here many teams get an unpleasant surprise: carbon fiber is a poor thermal conductor across the laminate. It conducts along the fibers and badly through the thickness, so a carbon plate is nearly an insulator for pulling heat out of an ESC bolted to it. Switching an arm from aluminum to carbon for weight often raises motor and ESC temperatures 15°C to 25°C. The structural case for carbon fiber frames is still good, but the thermal path has to be designed back in.
The usual solutions: an aluminum motor mount or ESC plate bonded to the carbon as a spreader; a thin metal strap from the hot component to a surface that sees airflow; thermal pads with specified conductivity; and on larger aircraft an aluminum extrusion arm that is both structure and heat sink. Thermal interface material only works under pressure, so design the clamping, not just the material.
Hover Is the Worst Case
Forward flight helps twice: the airframe gets ram air, and a multirotor in translation needs less power for the same lift. Hover gives neither — maximum power, minimum airflow.
| Flight condition | Motor load | Airframe airflow | Thermal risk |
|---|---|---|---|
| Just landed after heavy flight | None | None | Highest soak temperature |
| Hover with payload | High | Prop wash only | High and sustained |
| Forward cruise | Moderate | Prop wash plus ram air | Low |
| Aggressive climb | Very high | Good | Short duration, high peak |
So the thermal case for an inspection drone hovering over a bridge for 20 minutes is far more demanding than for a mapping drone flying transects at 30 mph, even though the mapping drone flies longer. Set the design condition from the mission. Tethered aircraft are the extreme case — unlimited energy, unlimited heat, no duty-cycle relief — and their thermal design is closer to industrial equipment than to a battery drone.
Hot and High
Density altitude is where thermal design and aerodynamics collide. Thin air means rotors produce less thrust for the same RPM, so motors work harder, and it also carries away less heat. Both effects move the wrong way at once, and a 100°F ambient pushes an aircraft that was comfortable at sea level to its limits.
Practical consequences for anyone selling into the mountain West or summer agricultural work:
- Publish payload and endurance against density altitude and ambient temperature, not a single sea-level number. Customers flying at 6,000 ft will find out anyway.
- Size motors so hover throttle at the hot-and-high condition still sits below the continuous rating — run the logic in choosing drone motors and propellers at the worst case, not the nominal one.
- Put temperature in the flight controller's protection logic: derate gradually with a warning, rather than cutting a motor without notice.
- The battery is thermally limited too, and high current in high ambient shortens pack life quickly — pack thermal behavior is part of battery system design.
Measuring It
Thermal problems are invisible until you look, and looking is cheap. Two methods, used together.
Instrumented logging. Most ESCs and flight controllers already report temperature; log them at 1 Hz with current and throttle, and add sensors on the motor bells, battery, and payload compute. Fly the worst case — maximum payload, hover, maximum ambient — and watch whether the curves plateau. A curve still rising when the battery runs out means the design has no thermal steady state; it is being rescued by a short flight time.
Thermal imaging. A handheld thermal camera pointed at the aircraft right after landing shows hot spots in seconds, and they are almost never where the team expected: one ESC hotter than its siblings from an airflow shadow, a connector glowing from a poor crimp, a battery lead warmer than the pack. Repeat after 60 seconds of soak to see where heat migrates. The payload version of the technology is in thermal camera drone payloads.
Set explicit limits and test against them — motor bell below the magnet grade's limit, ESC junction under its rating, battery surface within the cell window, compute silicon below throttling — and treat exceeding them as a failure, not a data point.
Design Rules That Hold Up
- Put heat sources in the prop wash whenever sealing allows.
- Give every sealed component a conduction path to structure with airflow; assume no convection inside the box.
- Do not assume carbon fiber moves heat — add an aluminum spreader where it matters.
- Design the post-landing soak case, not just the flight case.
- Derate deliberately and tell the pilot.
- Set the design point at maximum payload, hover, maximum ambient, maximum density altitude — the mission you sold, not the one you tested on a nice day.
Projects House develops drones and electronics where heat is the limiting factor — layout for airflow, conduction paths and heat sinks, thermal instrumentation, and hot-and-high testing that finds the limit before a customer does. Tell us about your aircraft through our contact form.