Originally Posted by
MechEngr
I came across one where two AoA sensors froze and, because they agreed, voted the correct one off the island. As other conditions changed, the aircraft began applying AND trim, causing the plane to pitch down. The pilot response was to pull the breakers for computers.
Of interest is recent news that Airbus has made a change to detect exactly this condition - two AoA sensors in close agreement, but not changing in reading, and to deselect them and use the remaining sensor instead.
The problem wasn’t the trim though, the trimming was just a result. The problem was the envelope protection applying nose down flight control inputs. So there was no runaway trim, the trim worked as intended.