Join the crewMy track

Classes / Systems class / Systems Department

Systems Track: Spacecraft Links and Flight Software Basics · Module 3 of 4 · about 30 min

The flight computer: commands, telemetry and fault protection

We study the onboard command and data subsystem described in NASA's Basics of Space Flight. Missions may use other names. It keeps time, interprets commands from Earth, collects and formats telemetry for the trip home, and runs high-level fault protection. Avionics is the broader term for the electronics supporting these functions. Jobs labeled avionics or embedded can involve building, testing or operating that hardware and its software.

Most spacecraft activity runs from stored command sequences. Operators build a timeline on the ground, check it in simulation, uplink it, and the flight computer executes it on schedule. With minutes of light delay, that is the only practical way to work, and it is why ground software, sequencing and simulation are careers of their own.

Space is hard on electronics. Energetic particles can flip a bit in memory, an event called a single event upset. Designers respond with radiation-tolerant parts, memory that detects and corrects errors, and watchdog timers that reset a processor if the software stops checking in. When something goes wrong that the software cannot handle, fault protection puts the spacecraft into a safe mode: it stops the planned activity, points its power and antennas where they need to be, and waits for instructions from the ground.

A lot of flight software is built on shared frameworks rather than written from scratch. NASA Goddard's core Flight System (cFS) is one example. It is open source, it splits flight software into reusable apps, and you can download it from GitHub and run it on a laptop. If you want something concrete to put on a resume, building and running cFS and writing a small app for it is a real project.

NASA also sets rules for how its software is engineered. NASA Procedural Requirements 7150.2 sorts software into classes by how much is riding on it, with the most critical software held to the strictest process. You do not need to memorize the classes, but knowing that safety-critical flight code is reviewed, tested and documented far more heavily than an app is the kind of context interviewers notice.

We separate a command's acceptance, execution and resulting state. In a fictional simulator, an accepted status means only that a request passed the simulator's stated input checks; a completed status is separate, and a sensor reading supplies another observation. These definitions belong to the exercise, not a universal mission protocol. If an acknowledgement is absent, we have uncertainty about delivery or reporting, not proof that an action never happened. Blind repetition could create an unintended second action.

Worked example, fictional fault review: a stored sequence expects a camera observation after a power check. A fault-protection response interrupts that sequence and the latest science packet predates the interruption. We do not infer that the observation completed from an earlier plan or an old image. We record the time of the fault, the sequence's last confirmed step and the available health evidence. Ground recovery remains the authorized mission team's decision; this lesson does not prescribe disabling safing or restarting a real spacecraft.

NASA's onboard-systems chapter describes fault protection as autonomous because both light time and limited tracking prevent immediate ground intervention. Safing can preserve the mission even while interrupting useful work. We evaluate a proposed response against protection of the spacecraft and the ability to regain contact, not only against whether science continued. A test that checks the happy path alone cannot demonstrate what happens when the assumed contact or component fails.

The cFS architecture separates reusable services from mission-specific behavior. NASA describes the platform support package, operating system abstraction layer and core flight executive as its main components, with applications and interfaces above that foundation. Portability helps reuse; it does not prove that a particular app meets another mission's timing, hardware or safety requirements. A laptop demonstration supplies evidence about the demonstrated setup, not automatic flight approval. NASA's engineering requirements and assurance processes remain distinct from an educational certificate.

Practice, in a paper model or isolated simulator: we define three states and a test table containing normal input, missing acknowledgement, delayed telemetry and a fault interruption. We state expected observations before looking at the result and keep requirements linked to test evidence. Our self-check identifies what each test proves and what it leaves untested. We make no connection to mission infrastructure, send no live commands and describe the result as study or sandbox practice rather than operational authorization.

Sources

Module quiz

10 questions, 8 correct to pass. You'll see what you missed and why, and you can retake it.

Next: Packets and protocols: how the data is packaged →