1. The test that either qualifies or stops the mission

Jim Lovell and Jeffrey Kluger’s Lost Moon returns one last time for this arc because the Apollo 13 crew’s improvised tests had no second chance. A ground test, by contrast, is allowed to fail. In fact, a ground test that fails is often more valuable than one that passes, because it finds a problem that can still be fixed. The important thing is to know what failure means and what must happen next.

This entry defines success and failure for the payload integration and operation test.

2. Success criteria

The test is a success if the desktop can host and operate all three representative payloads without violating safety, interface, or policy constraints.

Specifically:

  • Mechanical integration: each payload mounts cleanly, aligns within tolerance, and does not interfere with deployment or other subsystems.
  • Electrical integration: each payload powers on and off within voltage, current, and inrush limits; no bus collapse or back-feeding occurs.
  • Thermal integration: each payload and the surrounding desktop structure remain within predicted temperature limits during operation and survival.
  • Data integration: command and telemetry paths work for all payloads; command validation rejects unsafe sequences.
  • Scheduled operation: the autonomous scheduler accepts valid requests, defers or modifies constrained requests, and rejects impossible requests with a reason.
  • Multi-payload operation: two payloads can operate concurrently without power, thermal, or pointing violations.
  • Fault containment: a payload fault is detected, the payload is safed, the bus remains healthy, and ground is notified.
  • Autonomous override: a safety-critical condition overrides payload operations and restores the platform to a safe state.

3. Failure categories

Failure is not a single outcome. The test can fail in ways that demand different responses:

  • Critical failure: the payload integration or operation damages the desktop or creates an unsafe condition. Response: halt payload hosting for that design, fix the interface or payload, and retest.
  • Major failure: the payload operates but violates a requirement such as power margin, thermal limit, or command safety. Response: redesign the interface constraint, modify the payload, or update policy, and retest.
  • Minor failure: the payload operates within requirements but the scheduler response is unclear, slow, or poorly explained to the customer. Response: improve scheduler logic, logging, or customer notification, and retest the affected scenario.
  • Test artifact failure: the test setup, proxy payload, or measurement is wrong, not the desktop. Response: fix the test and rerun before drawing conclusions.

4. What to measure

Objective measurements make success and failure defensible:

  • peak and average power draw;
  • inrush current at turn-on;
  • payload and host temperatures during operation;
  • command latency and success rate;
  • telemetry update rate and integrity;
  • scheduler decision time and explanation quality;
  • time from fault injection to safe state;
  • data throughput and storage use;
  • number of autonomous overrides and their triggers.

5. What this changes

  • Success means the desktop can safely host and autonomously operate representative payloads.
  • Failure is categorized by severity and response, not treated as a single stop signal.
  • The next entry will close the payload integration and operation test arc.