Make partner subsystems work together

NASA’s integration guidance treats interfaces and subsystem interactions as work to engineer throughout development. Bringing together parts is more than attaching them. Electrical, data, mechanical and thermal interfaces can create problems that were absent in an isolated test. AfDev-Sat’s shared development makes that coordination central to the mission’s engineering story.

Consider an invented sensor that reports temperature as an integer in tenths of a degree. A receiving computer interprets the same number as whole degrees. The value 250 means 25 degrees at the sender but 250 at the receiver. Both components can pass tests using their own conventions while the combined system fails to communicate the intended measurement.

An invented integration test begins with a computer commanding the power board to enable a payload. The payload starts, records an observation and returns data with a timestamp. The test checks the command, power response, measurement and received record. Its purpose is to expose failures across boundaries, not merely to watch separate components light up.

If a startup transient resets the computer, a payload-only test may miss it. Reproduce the relevant combined load and configuration, then retain the observed trace. A successful room-temperature functional sequence also has a limited scope: environmental behaviour and other required conditions need appropriate evidence of their own.

Test the meaning across the interface

For a fictional partner delivery, record the hardware revision, software version, interface document, test conditions, expected outcome and actual observations. An unresolved mismatch gets an owner and a retest condition. The next team should be able to distinguish a reviewed deviation from an undocumented change.

Imagine that version A passes a command sequence, but version B changes the message format. Reusing the old pass result without checking the changed interface would detach the evidence from the delivered system. Strong collaboration preserves useful results while identifying what must be repeated. A subsystem milestone deserves recognition without being inflated into proof of launch or an operational observing service.