Vacuum removes convection—not waste heat
Terrestrial facilities can move heat into air or water and then discharge it through cooling towers, dry coolers or other plant. In vacuum there is no surrounding fluid for convection. Heat must move through the spacecraft by conduction or a closed transport loop before an external surface can reject it as electromagnetic radiation.
Orbital cooling is a heat-transport and radiator problem, not free access to a cold environment.
The thermal design must include processors, memory, power electronics, batteries, radios and conversion losses. A payload that budgets only the nominal compute load understates the rejection requirement.
Capture, transport, buffer, reject and control
Interfaces and cold plates move heat away from high-flux components while keeping junction temperatures inside verified limits.
Conductive straps, heat pipes or pumped loops carry heat to radiator surfaces without destabilising the host spacecraft.
Thermal mass or phase-change material may absorb short peaks, but stored heat must still be rejected later.
Surface area, temperature, emissivity, orientation and environmental flux determine useful rejection.
Power scheduling is part of the thermal system. Workloads can be throttled, shifted between nodes or deferred during eclipse and hot-attitude conditions. That can lower peak demand, but it also changes delivered compute and must appear in performance and economics models.
Radiator area follows the declared heat load
Q̇ = εσA(T⁴rad − T⁴environment)Useful sizing also accounts for absorbed solar and planetary flux, view factor, surface degradation, conductive leakage and design margin.
The fourth-power temperature relationship can make higher radiator temperature attractive, but component temperature limits, efficiency, lifetime and material compatibility constrain the operating point. A decision-grade result must publish the assumed heat load, temperatures, emissivity, orbital attitude, eclipse fraction, degradation, radiator mass and uncertainty—not only a favourable area number.
Thermal capacity and useful compute must be modelled together
Sunlight, eclipse, Earth infrared flux and spacecraft attitude change across an orbit. The system therefore needs hot-case, cold-case and transition analysis. A storage-oriented node with brief authenticated access windows can have a different thermal profile from continuous high-performance compute.
- Define the workload and electrical power trace, including conversion losses.
- Model sunlit, eclipse, safe-mode and degraded-radiator cases.
- Declare throttling and the amount of service still delivered.
- Test a failed transport loop, sensor drift and loss of attitude control.
- Carry thermal limits into scheduling, replacement and cost models.
A thermal architecture becomes evidence through test
- Correlate a versioned thermal model against component and bench measurements.
- Measure interface resistance, transport capacity and transient storage behaviour.
- Run thermal-vacuum cycles at representative loads and boundary conditions.
- Verify performance after vibration, radiation exposure and surface-ageing assumptions.
- Publish margins, sensor uncertainty, failed cases and corrective changes.
- Require integrated host-payload thermal acceptance before flight.
Proof of Pulse currently presents thermal concepts and engineering targets. Flight-qualified performance and sustained orbital operation remain future evidence gates.