The flare detector was missing from the flare-detector schematic

Rev A has a payload computer, a regulated supply, protection, storage, two context sensors, a comparator and both the development and spacecraft interfaces. Reading the sheet, it looks like a finished design. Read against the fourteen top-level requirements, the first five — measure soft-X-ray irradiance, detect onset and peak and end, separate real increases from artefacts, cover the target irradiance range, sample at one second or faster — have no hardware behind them at all.

I stopped treating the missing pieces as components to add later and wrote them out as a chain that has to be designed as a unit: Sun-facing aperture, X-ray-transmitting and visible-blocking filters, the detector itself, a low-noise transimpedance or charge-sensitive amplifier, an ADC path, and a temperature sensor at the detector with calibration behind it. Every one of those is a required addition, and the first four items of the Rev B change list are exactly that chain.

The review method was what surfaced it. Going requirement by requirement rather than looking at the drawing meant the question was never "is this schematic complete" but "which requirement does this part serve" — and six of them had no answer.

Support electronics and the instrument are different things. What I had designed was the half of the payload that supports the measurement, and none of the half that makes it.

Design Rev B starting at the aperture and working inwards, rather than adding a detector onto the existing sheet.