Commissioning needs stage evidence, not a launch moment

Complex equipment rarely moves from shipment to full performance in one clean step. Ashley Balzer’s NASA Science mission update, published on 15 September 2026, documents the in-space activation of the Nancy Grace Roman Space Telescope’s Wide Field Instrument, a 300-megapixel infrared camera. Before activation, the instrument spent 10 days drying and decontaminating, then cooled from minus 65 Celsius to minus 143 Celsius so its 18 detectors could be switched on.

The team tested a sequence rather than a single outcome. Engineers activated the calibration system, sent test data to the ground, operated a wheel containing filters and other optics in weightlessness, and checked the focus mechanism. The first starlight image was deliberately far from best focus because the detector array remained in its launch position; it provides a baseline for alignment rather than a finished science product.

A second instrument, the coronagraph, went through a different level of checkout. Operators confirmed communication with its software, thermal controls, mechanisms, cameras and avionics, then returned it to a decontamination period. The update says the assessments performed as expected, while further calibration remains and first science images are planned for early 2027. Activation therefore marks verified progress, not completion.

Dunes Bureau interpretation: buyers commissioning a production line, scanning system or data-centre installation can use the same evidence structure without pretending the technologies are equivalent. Separate power-on, control response, data flow, environmental stabilisation, calibration and operational-performance tests; record the acceptance owner and evidence for each stage. Roman does not demonstrate commercial savings or supply an industrial benchmark, but it offers a clear example of why an early imperfect output may be valid evidence when its purpose is defined in advance.