Choosing a satellite bus
Selecting a satellite bus is a systems decision, not a procurement one. Get it wrong and the payload underperforms regardless of its own quality. This page sets out the engineering logic that should drive the choice.
The payload writes the requirement; the bus answers it
Every bus decision starts in the same place: what does the payload need? That means four numbers before anything else. Mass, because the bus must carry it without exceeding the launch vehicle's dispenser limits. Average power draw, because solar-array area and battery capacity are fixed by the bus class. Peak power, because imaging radars and high-throughput transmitters spike hard and short, and a bus sized for average draw will brown out at the worst moment. And pointing stability, because a synthetic-aperture radar that needs 0.1-degree attitude knowledge has no use for a bus designed around a 1-degree star-tracker.
Only once those four are on the table does bus class become a meaningful conversation. A 6U cubesat bus delivers roughly 10 to 20 watts to the payload and carries perhaps 1 to 2 kilograms of payload mass. A dedicated small-satellite bus in the 150 to 500 kg range can offer 200 to 500 watts and carry payloads of 80 to 200 kg. The gap between those two worlds is not bridged by clever integration. It is a structural choice made at the start of the programme.
Heritage trades schedule against ambition
A bus with documented flight heritage means the thermal model has been validated on orbit, the attitude-control firmware has survived eclipse cycling, and the failure modes are at least partially mapped. That is worth a great deal when a national programme has one launch and one chance. Surrey Satellite Technology's SSTL-150 and SSTL-300 platforms, for instance, carry decades of accumulated anomaly data. Planet's Dove bus is arguably the most flight-proven small imaging platform in history, with hundreds of units flown.
New or lightly flown designs offer more flexibility, particularly when the payload has unusual form-factor or power requirements that heritage buses were never sized for. The honest trade is this: a new bus design adds twelve to thirty-six months of qualification risk to a programme schedule, and that risk is not evenly distributed. Thermal-vacuum testing catches some problems. On-orbit commissioning catches the rest, often expensively.
For a government buyer on a first programme, heritage is almost always the right answer. The exception is when the mission requirement genuinely cannot be met by any existing platform, at which point the programme must budget for the qualification campaign explicitly, not optimistically.
Procurement model shapes who owns the risk
There are two broad procurement models, and they distribute risk very differently. In a buy-bus-then-integrate approach, the government or prime contractor purchases a bus from one vendor and a payload from another, then manages the integration itself. This can reduce unit cost and preserves flexibility, but it creates an interface gap: when the system underperforms, the bus vendor points at the payload and the payload vendor points at the bus. Someone must own that interface technically and contractually. If that someone is not clearly named before signature, it will be named during the anomaly review, which is a worse time.
A turnkey procurement places a single contractor accountable for the integrated satellite. This costs more per unit in most cases, but it compresses the interface risk into one contractual relationship. For programmes where the government team is building its first satellite, the turnkey model also preserves the option for genuine technology transfer: the prime can be required to train national engineers on integration and test procedures as a contract deliverable, not as an afterthought.
What the bus vendor's datasheet does not tell you
Published bus specifications describe best-case performance. They do not describe the margin available after the payload is integrated. A bus rated at 300 watts generated does not deliver 300 watts to the payload; it delivers 300 watts minus housekeeping, minus thermal control heaters, minus attitude actuators, minus the margin the bus vendor has already consumed in their own design. Ask for the payload power budget with margin explicitly called out, not the headline generation figure.
Pointing specifications deserve similar scrutiny. Attitude determination and control accuracy is frequently quoted as a steady-state figure under benign conditions. Ask for the slew-and-settle time to the quoted accuracy, and ask what happens during eclipse transitions when thermal gradients disturb the structure. For optical payloads, ask for the jitter power spectral density, not just the RMS pointing error. A bus that points to 0.01 degrees on average but vibrates at 50 millidegrees per second during reaction-wheel zero-crossing is not useful for sub-metre imaging.
Finally, ask about data interfaces. A payload that outputs data at 400 megabits per second is useless on a bus whose onboard computer accepts 100 megabits per second. These mismatches are common and expensive to fix after contract signature.
Where bus choice genuinely goes wrong
The most common failure mode is not technical. It is schedule-driven under-specification. A programme with a fixed launch window selects the bus that can be delivered in time rather than the bus that fits the payload. The result is a satellite that launches on schedule and underperforms for its entire operational life. A three-month delay to select the right bus is almost always cheaper than five years of degraded mission return.
A second failure mode is orbit-bus mismatch. A bus qualified for low Earth orbit at 500 kilometres will face a significantly different radiation environment at 600 kilometres, where the South Atlantic Anomaly exposure increases sharply. Total ionising dose requirements, single-event upset rates and component screening levels all change. A bus vendor who has not flown at your target altitude cannot simply extrapolate from their heritage data. Ask for the radiation analysis specific to your orbit, inclination and altitude, not a generic LEO qualification.
Thermal design is the third common surprise. A bus designed for a sun-synchronous dawn-dusk orbit, where solar illumination is nearly continuous, will have a very different thermal architecture from one designed for a mid-inclination orbit with long eclipses. Switching orbit type after bus selection is not a paperwork change.
The questions worth asking before the proposal arrives
Before any vendor submits a proposal, four questions clarify more than any datasheet. First: what is the payload power available at end of life, with margin, after all bus housekeeping is accounted for? Second: what is the jitter environment at the payload interface, measured, not modelled? Third: what anomalies has this bus experienced on orbit, and how were they resolved? A vendor who cannot answer the third question has not flown the bus enough to know. Fourth: what are the source-access and audit rights for the flight software? For a government programme, running national infrastructure on firmware that cannot be independently inspected is a sovereignty question, not just a technical one.
Satellize structures its bus procurement around these questions before any integration contract is placed. The Tonga sovereign-communications restoration and the subsequent crop-estimation analytics programme both required bus and payload combinations that could be handed over to national operators, which meant software access terms were agreed before hardware was ordered, not negotiated after delivery.
Engineering parameters
| Cubesat bus (6U) payload power | 10 to 20 W available to payload (end of life, typical) |
| Microsatellite bus (50–150 kg) payload power | 50 to 150 W available to payload (end of life, typical) |
| Small satellite bus (150–500 kg) payload power | 200 to 500 W available to payload (end of life, typical) |
| Attitude determination accuracy, heritage platforms | 0.01° to 1° depending on sensor suite (star tracker vs. magnetometer) |
| Onboard data storage, small satellite class | 128 GB to 2 TB solid-state, mission-dependent |
| Design lifetime, LEO | 3 to 7 years for cubesat class; 5 to 15 years for small satellite class |
| Bus qualification lead time (heritage) | 12 to 24 months from contract to delivery, integration not included |
| Bus qualification lead time (new design) | 24 to 48 months; add 6 to 18 months qualification campaign risk |
| Total ionising dose tolerance, screened components | 20 to 100 krad(Si) depending on component grade and shielding |
One contract, one accountable engineer
Commissioned as one programme, not a stack of contracts: spacecraft, launch, ground segment, mission control, training and handover are priced together. Source-access terms and audit rights are agreed in writing before signature. Request a payload-to-bus compatibility review.