Spacecraft Systems

Prerequisites: The Space Environment, Rocket Propulsion Feeds problems: surviving space without a mechanic, in-space refueling

Practitioner

Every satellite divides into two parts. The payload is why the mission exists — the telescope, the camera, the communications transponders. The bus is everything that keeps the payload alive and pointed: typically half or more of the spacecraft’s mass and cost goes to the servant, not the star. Walk the bus one subsystem at a time and you’ve walked the whole discipline.

Power. Solar arrays feed the spacecraft and charge the batteries that carry it through eclipse — which in low orbit arrives every 90 minutes, so the batteries endure tens of thousands of charge cycles. The sizing rule that surprises newcomers: arrays are sized for the end of the mission, after years of radiation degradation, in the worst-case season — so a new satellite has surplus power and a design that looks overbuilt. It isn’t; it’s built for year twelve.

Thermal. The environment topic explained why heat can only radiate away; this subsystem does the plumbing — conducting heat from hot boxes to radiator panels, wrapping the rest in insulation blankets, running heaters in the cold cases. Thermal engineers hold veto power in layout debates: a component placed where its heat can’t reach a radiator is a dead component; it just doesn’t know it yet.

Attitude (ADCS). Star trackers and gyros determine which way the craft points; reaction wheels change it — spin a flywheel one way, the body turns the other, no propellant spent. But wheels saturate: absorbed disturbance torques accumulate until a wheel hits maximum speed and must be unloaded, either by thrusters (spending propellant) or, in low orbit, by magnetorquers pushing against Earth’s magnetic field for free. Pointing precision is a steep cost curve — a satellite that points to a tenth of a degree and one that points to an arcsecond are different classes of machine.

Communications. The radio link home obeys an unforgiving budget: data rate falls with distance squared and rises with antenna size and power. For imaging satellites the downlink, not the camera, is usually the bottleneck — a satellite can photograph far more than it can transmit in its few minutes per pass over a ground station. Hence ground-station networks, and increasingly, processing data on board and sending only answers.

Command and data handling. The onboard computer — radiation-tolerant, per the last topic — runs the mission mostly alone: a LEO satellite talks to its operators minutes per day. The load-bearing design pattern is safe mode: when anything confuses the spacecraft, it autonomously drops to the simplest survivable configuration — solar panels at the sun, radio listening — and waits for instructions. Safe mode is the mechanic you can send.

Propulsion and structure. Thrusters for station-keeping, wheel desaturation, collision avoidance, and the end-of-life disposal burn — chemical when maneuvers must be quick, electric when efficiency wins (and on constellations, it usually wins). The structure carries everything through launch vibration, which — along with the fairing — is why launch vehicles shape satellite design even after separation.

How does this get designed? As budgets — the mental model from section 2 made procedural. A mass budget, a power budget (worst case: end-of-life, in eclipse), a link budget, a propellant budget — each with explicit margin, 20–30% early, shrinking as the design matures. The budgets are the negotiation table where subsystems trade; a camera upgrade that fits mass but breaks power gets rejected by a spreadsheet long before it reaches a review board.

And the era’s defining fork: the exquisite spacecraft (redundant everything, fifteen-year life, cost measured in billions — the flagship telescope) versus the constellation unit (commercial parts, five-year life, a few percent annual failure rate accepted and covered by spares in orbit). Neither is wrong; they answer different questions — as the iterate-or-exquisite model predicts. The CubeSat standard sits at the far end of that spectrum: satellites built in 10 cm units from catalog parts, cheap enough that a university can afford flight heritage. That standardization, more than any single technology, is what unleashed the smallsat era.

Expert pointers

Watch onboard autonomy (processing imagery at the sensor and downlinking answers), optical inter-satellite links (constellations becoming mesh networks that bypass ground stations), software-defined payloads (retargetable after launch), and servicing interfaces — grapple fixtures and refueling ports standardized before they’re needed, so the next generation is repairable. That last one loops back to problem #7.

Misconceptions

• “The payload is most of the satellite.” The bus typically dominates mass and cost. The camera is the easy part; keeping it powered, cooled, pointed, and connected for a decade is the engineering.
• “More redundancy means more reliability.” Redundancy adds parts, and parts add failure modes — including common-mode failures that take out both copies. Past a point, simplicity is the more reliable design; constellations push that logic to its limit by putting the redundancy in orbit as whole spare satellites.
• “Satellites are flown by joystick.” They fly themselves. Operators uplink schedules during brief passes and read telemetry; between contacts the spacecraft is alone, which is why autonomy and safe modes — not pilot skill — are what save missions.

Check yourself

1. An imaging satellite works flawlessly but delivers 1% of the images it captures. Which subsystem was undersized, and what are two fixes on opposite ends of the spacecraft?
2. Why do reaction wheels saturate, and why does the free fix (magnetorquers) work in LEO but not in deep space?
3. A constellation operator flies satellites with zero internal redundancy and accepts 5% annual failures. Why is this rational for them and irrational for a flagship telescope program?
4. Your satellite’s solar arrays produce 40% more power than the payload needs at launch. Waste, or design intent? Argue it.

Apply it

Draft a one-page satellite concept for the mission you assessed in the environment topic: name the payload, then allocate a mass budget using rough smallsat fractions (payload ~30%, power ~20%, structure ~20%, ADCS ~10%, thermal ~5%, C&DH+comms ~10%, propulsion ~5%) and a power budget for sunlight and eclipse. Add 25% margin to both and note what no longer fits — that argument you just had with yourself is spacecraft engineering, and the page is the core of a mission-proposal capstone. (~45–60 minutes)