A detector rail boundary scan holds measured, dark, gain, and offset fixed (giving a fixed unclipped output of 25) and checks three rail values: rail 20 clips the reported output down to 20, exactly matching chapter two's own single checked example; rail 25 lands exactly at the unclipped value; rail 30 stays clear of it entirely, reporting the full 25 exactly for the stated toy calibration rule. Chapter two checked only the clipped case once; this scan shows the same fixed affine output landing in all three possible relationships to a rail, not just the clipped one.
The same fixed affine output is checked against three rail values, from clipped through exactly-at to clear above. Exact arithmetic here means exact results for the stated model inputs; measured inputs still carry uncertainty and significant-figure limits.
highlighted = computed this step
Chapter two's own clipped case is this scan's first row
The affine output before any rail is 25. Chapter two checked exactly one rail value, 20, which clips the reported output down to 20.
25>20⇒Y=20
Three rail values bracket the same fixed affine output
The measured, dark, gain, and offset sources all stay fixed, so the unclipped output stays the same in every row. Only the rail changes, moving the reported output from clipped, to exactly-at-rail, to fully clear.
rail202530Yunclipped252525Y202525
A rail clear above the output never clips it
With rail 30, well above the unclipped output of 25, the reported output is the full unclipped value, 25.