Module 7: Anatomy of a Satellite

A guided teardown: the bus, the payload, power and propulsion, antennas and lasers — and what happens from factory floor to fiery deorbit.

The Two Halves: Bus and Payload

Every satellite splits into the bus (keeps it alive) and the payload (does the job it was launched for).

The Bus: The Satellite's Body

Strip away what makes a communications satellite different from a weather satellite or a spy satellite, and what remains is the bus — the standardized chassis that keeps any satellite alive. The bus provides structure (the frame everything bolts to), electrical power, propulsion, thermal control, attitude determination (knowing which way it points), and the command link that lets ground operators talk to it. Bus platforms are often reused across missions: Airbus's OneSat, Boeing's 702, and Maxar's 1300 series have each carried dozens of very different payloads. Manufacturers like K2 Space are now building high-power buses (their GRAVITAS platform) as a product in itself, betting that standardized, mass-produced buses will do for satellites what standard PC architecture did for computing.

The Payload: The Reason It Exists

The payload is whatever the satellite was actually launched to do. On a communications satellite that means transponders, modems, routing processors, and antennas. On an Earth-observation satellite it's cameras or synthetic-aperture radar (like ICEYE's SAR instruments). On a navigation satellite it's atomic clocks. Payload and bus are designed against a strict budget of three resources: mass (every kilogram costs launch money), power (solar arrays only make so many kilowatts), and thermal (every watt consumed becomes heat that must be radiated away in a vacuum).

A typical breakdown for a communications satellite: the payload is roughly 30-40% of the dry mass but drives nearly all of the value. The other 60-70% — bus, propellant, solar arrays — exists to keep that payload pointed, powered, and cool.

Bent-Pipe vs Regenerative: How Smart Is the Payload?

Communications payloads span a spectrum of intelligence. A bent-pipe (transparent) payload simply receives, amplifies, frequency-shifts, and retransmits — all the intelligence lives on the ground. A regenerative payload demodulates and processes signals in orbit: it can route packets between beams, run onboard compute, or even host a full 5G base station (gNB), which 3GPP Release 19 now standardizes. Starlink's newer generations process and route traffic between satellites entirely in space; the most advanced designs are effectively flying routers with radios attached.

Size Extremes

The range is enormous. A 1U CubeSat is a 10 cm cube massing about 1.3 kg — enough for basic IoT store-and-forward. A Starlink V2 Mini masses ~800 kg. A large GEO communications satellite is a bus-sized 6,000+ kg machine. And AST SpaceMobile's BlueBirds unfold a 223 m² phased array — the size of a squash court — because physics demands enormous apertures to talk directly to an unmodified smartphone from orbit.

Why it matters: Bus-versus-payload is the vocabulary every satellite story is told in — once you see the split, specs, costs, and failures all make sense.

Power, Propulsion, and Staying Alive

Solar wings, batteries for eclipse, electric thrusters, and the unglamorous art of not overheating.

Power: Solar Wings and Eclipse Batteries

Nearly every satellite is solar-powered. Deployable solar arrays unfold after launch and track the Sun, generating anywhere from a few hundred watts (smallsats) to 20+ kW (large GEO platforms and high-power LEO buses). But every orbit includes eclipse — passes through Earth's shadow — so lithium-ion batteries carry the load for up to 35 minutes per LEO orbit (up to 72 minutes in GEO during eclipse seasons). Battery degradation over thousands of charge cycles is a primary factor limiting satellite lifetime.

Propulsion: Mostly Electric Now

Chemical thrusters deliver strong, fast impulses but drink propellant. The industry has largely shifted to electric propulsion — Hall-effect thrusters that ionize a noble gas (xenon traditionally; Starlink pioneered cheaper krypton, then argon) and accelerate it electrically. Electric thrust is feather-light (a sheet of paper pushes harder) but 5-10x more propellant-efficient, which is why a Starlink satellite can spend months gently raising its orbit, dodge debris on command, and still keep years of station-keeping reserve. Propellant, not hardware failure, usually defines end-of-life for a healthy satellite.

A satellite with empty tanks is a satellite at end-of-life, no matter how healthy its electronics are. This is why refueling demos and on-orbit servicing (Northrop's MEV missions already extended two Intelsat GEO birds) are such a big deal.

Attitude: Knowing and Controlling Where You Point

Antennas must point at Earth, solar panels at the Sun, and sensors away from both — simultaneously, forever. Star trackers (cameras that recognize star fields), gyroscopes, and GPS receivers determine orientation; reaction wheels — spinning flywheels whose speed changes rotate the spacecraft — control it without spending propellant. When wheels saturate, magnetorquers (in LEO) or small thruster pulses unload them. Reaction-wheel failure is one of the most common satellite-killing faults.

Thermal: The Vacuum Is a Terrible Radiator

In vacuum there is no air to carry heat away — everything must be radiated as infrared. A satellite simultaneously roasts on its Sun-facing side (+120°C) and freezes on its shadow side (-150°C). Multi-layer insulation (the gold foil in every satellite photo), heat pipes, louvers, and dedicated radiator panels move heat from electronics to space. Thermal design quietly constrains everything: transmit power, battery placement, even how long a high-power payload can run.

Why it matters: Power, propellant, and heat budgets — not electronics — set the real limits on what a satellite can do and how long it lives.

Antennas: Dishes, Phased Arrays, and Lasers

From fixed parabolic dishes to electronically-steered arrays and inter-satellite laser links.

Parabolic Dishes: The Classics

The traditional satellite antenna is a parabolic reflector — a curved dish that focuses radio energy the way a telescope mirror focuses light. Bigger dish, tighter beam, more gain. GEO satellites carry precisely-shaped reflectors that paint fixed coverage patterns over continents. Dishes are simple and efficient, but they point where they're built to point; redirecting them means physically slewing the spacecraft or the reflector.

Phased Arrays: Beams That Move at Software Speed

A phased array replaces the dish with hundreds or thousands of small flat elements. By adjusting the phase of the signal at each element, the array steers its beam electronically — in microseconds, with no moving parts, and it can form *many* beams at once. This is the technology that makes LEO broadband possible: a Starlink satellite sweeps across the sky in minutes, yet its beams stay locked on user cells below, while user terminals (themselves phased arrays) track the satellite the same way. AST SpaceMobile's 223 m² arrays push the concept to its extreme — an aperture large enough to close the link with an ordinary smartphone antenna.

The rule of thumb of D2D: the less capable the device on the ground, the more monstrous the antenna in space must be. Your phone can't grow a dish — so the satellite grows one for both of you.

Inter-Satellite Links: The Backbone in the Sky

Modern LEO constellations connect satellites to each other with optical inter-satellite links (ISLs) — laser terminals moving data at up to 100-200 Gbps between spacecraft thousands of kilometers apart, each aiming at a moving target with sub-microradian precision. ISLs let traffic enter the constellation over an ocean and exit at a gateway on another continent without touching the ground in between — critical for coverage where no gateway can exist (oceans, poles, denied areas) and for latency: light in vacuum outruns light in fiber by ~45%.

Feeder vs User Links

Every communications satellite juggles two kinds of radio link: user links to the terminals it serves (phones, dishes, IoT sensors) and feeder links to its gateway ground stations, which aggregate all user traffic to and from the internet. Feeder links run at higher frequencies (Ka-band, V-band, and increasingly E-band or optical) because gateways can afford big precision antennas — spectrum that would never survive a consumer terminal's rain-soaked view of the sky.

Why it matters: Antenna architecture is the real dividing line between satellite generations — it decides who a satellite can talk to and how much data it can move.

A Satellite's Life: Factory to Fireball

Build and test, ride to orbit, commission, operate for years, then get out of the way.

Build and Test: Shake, Bake, and Blast

A satellite spends months — historically years — in assembly cleanrooms, then endures a brutal test campaign designed to break weak hardware on the ground rather than in orbit: vibration tables that replicate launch shaking, acoustic chambers that blast it with rocket-engine sound, and thermal-vacuum (TVAC) chambers that cycle it between deep-space cold and full-Sun heat in vacuum for weeks. Mass production changed the tempo: Starlink's factory reportedly builds multiple satellites per day — a cadence unthinkable in the one-off GEO era.

Launch and the Ride Uphill

On launch day the satellite is bolted inside the rocket fairing, folded like origami — arrays stowed, antennas latched. After deployment comes the quiet drama of commissioning: arrays unfold, the satellite detumbles, points its arrays at the Sun, phones home, and runs weeks of checkout. LEO satellites launched to a low insertion orbit then spend weeks or months raising themselves to their operational shell on electric thrust. In GEO, satellites drift to their assigned longitude slot and stop — 35,786 km up, parked over one spot on the equator.

Operations: Years of Quiet Routine

Operational life is station-keeping burns, battery conditioning, software updates, debris-avoidance maneuvers (Starlink satellites perform thousands of small dodges per year across the fleet), and — for the constellation era — constant fleet-level traffic management handled largely by software. Design lifetimes run ~5 years for LEO broadband satellites (deliberately short; the constellation refreshes with better hardware) and 15-20 years for GEO platforms.

The 5-year LEO lifetime is a feature, not a bug: it forces continuous fleet replacement, which means the average satellite overhead is always near the state of the art — and it caps how long any dead satellite can clutter the orbit.

End of Life: Getting Out of the Way

A responsible ending depends on the orbit. LEO satellites use their last propellant to lower their orbit and burn up on reentry — modern designs are "demisable," engineered to fully disintegrate. (Regulators now expect deorbit within 5 years of end-of-mission, a sharp tightening from the old 25-year guideline.) GEO satellites can't practically come home from 36,000 km; instead they climb ~300 km higher into a graveyard orbit and are passivated — tanks vented, batteries drained — so nothing can explode. The growing alternative: on-orbit servicing vehicles that dock with aging satellites to extend their lives, and active-debris-removal missions (like ClearSpace-1) for the ones that died without cleaning up.

Why it matters: Understanding the lifecycle explains satellite economics, the 5-year LEO refresh rhythm, and why debris rules shape constellation design.