Originally Posted by
FGD135
Quite commonplace, really, for critical airborne systems to use only one sensor, and only raw data at that. Examples:
Turkish Airlines flight 1951, 25 Feb, 2009. B737-800 where one radar altimeter was malfunctioning. The data from the sensor went to zero, the computers thought the plane was on the ground, so they reduced the engine power to idle. The result was a stall at low altitude where many occupants were killed. The idiotic thing was that not only were the computers using just one radar altimeter sensor, they were making no effort to inspect it for reasonableness or filter it against spikes. The data was showing valid heights but then instantly started showing zeros!
Qantas flight 72, 7 Oct, 2008. Airbus A330. Pilot's side air data computer had a momentary spike in the angle of attack data. Silly computers took this as indicating the aircraft was suddenly stalling, and at a speed of about 450 knots, pushed the nose down. Passengers were thrown into the ceiling and many were seriously injured. The idiotic thing was that the computers were using completely raw data and could therefore believe that the angle of attack could, in the space of one second, change from sensible values to a stalling angle. Also idiotic that the computers would happily perform a manouever of such violence.
Well I've only been an aviation professional for 52 years, so what do I know.
Seriously though, the Turkish Airlines case illustrated poor failure monitoring within the RADALT (Not to MENTION complete crew unawareness and monitoring) of IAS. RAD ALT inputs to AFCS typically is used for gain gearing and throttle retard and usually have good self monitoring (But not in this case of course).
The QANTAS A330 case has yet to be completely and adequately explained ADIRU fault, the suspect was a data labelling issue. (GREAT airmanship by the QANTAS crew however). The ADIRU functions certainly are monitored and compared.
For ANY flight control system, ALPHA as well as Ps & Pt inputs are invariably monitored.. The absence of an Alpha disagree function of MCAS is at the very best negligent to the extreme, as has been the whole frantically rushed throwing together of this appalling system. Pilots and passengers deserve FAR better than this.