Originally Posted by
flyingchanges
...good to know approximate thrust lever angles (sorry Airbus pilots)...
I'm a little confused by that - As I understand it, if autothrust is disengaged the FBW Airbus thrust levers work the "old-fashioned" way - I think the only difference between TLs on FBW Airbii and other types is that the Airbus A/THR doesn't physically move the TLs to match its settings, and other types' autothrottles do.
@sudden twang:
(Caveat : I'm a software engineer, not a professional pilot - but I've been aviation-mad from the crib, and have been around this forum long enough to pick up a few things. Apologies if this becomes a mini-essay, but I hope you'll find it interesting... )
So - Unreliable airspeed (or UAS). If I've understood things correctly from what I've read and been told over the years it's undoubtedly a potentially hazardous condition, but if handled correctly it's more of an annoyance than a major threat. If it occurs in-flight, the piloting mantra "Aviate, Navigate, Communicate" (in that order) is key - before doing anything else (including diagnosing the problem), you have to make sure your flightpath is relatively stable. In the case of the Airbus FBW types, the flight data computers will detect a discrepancy in airspeed readings and automatically disconnect autopilot.
If the flightpath was straight, level and more-or-less stable (e.g. in the cruise phase) at the time, then there's a good chance that little or no manual intervention will be necessary. However if a pilot is not sure, one thing you'll see repeatedly through the AF447 thread is a reference to the term "Pitch and Power". What this means is that in the event of losing airspeed information, the manufacturers provide a lookup table in the reference handbook which says that if the pilot sets a certain pitch attitude and thrust/power setting, then the aircraft will tend to maintain altitude and sufficient airspeed. The settings from the A330 handbook are below (note that the higher you go, the lesser the required pitch attitude):
Once the aircraft is as straight, level and as stable as possible (external factors like turbulence notwithstanding), then the crew can take the time to troubleshoot the issue. Most modern airliners will have flight data computers which are capable of determining which of the airspeed data sources is incorrect, otherwise it's a case of the crew comparing the airspeed instrument readings on the Captain's side, the FO's side and the standby and rejecting the one which is reading differently from the other two. In the event of ice crystal blockage, the problem will usually go away on its own in a short space of time.
The technical background to the issues leading up to AF447 was that the A330 was originally built and certified using pitot tubes (airspeed sensors) made by Goodrich, and in fact the standard fit was still using the Goodrich models at the time of AF447. I think it was in the late 1990s that Airbus gave customers the option of having pitot tubes made by Thales fitted instead, and AF was one of the customers who took up that option. While the Thales "AA" sensors met all the certification requirements, it became apparent on the line that they were more susceptible to ice crystal blockage than the Goodrich sensors. In fact it became apparent that in certain conditions it was possible for two, or even all three of the sensors to ice over, which was a very serious issue, as both automatic and manual UAS diagnostic procedures start with the assumption that only one reading is faulty. When this came to light, Thales revised the design and produced a new "AB" model sensor - and Airbus began a service programme intended to replace all AA models. I believe that the accident aircraft was due to have the fix performed around a month after the crash.
In any case, what's important to remember is that the UAS problem itself was not a major causal factor in the crash of AF447. In fact there were a little over 30 UAS incidents involving the A330 between the introduction of the Thales option and AF447 and all of them were resolved safely.
The above relates solely to a UAS condition once the aircraft is already in-flight. If you watch an flight deck video of an airliner taking off, you'll hear the non-handling pilot refer to the airspeed indicator coming alive, and making airspeed callouts as the aircraft accelerates. As I understand it, these callouts are cues for the handling pilot to glance down from the windscreen to check their own airspeed indicator and see if it matches. If it doesn't, then the takeoff should be rejected and the aircraft won't leave the ground.
That *should* have been the case with Birgenair Flight 301, a B757 chartered at short notice to replace an aircraft with technical problems - but for whatever reason, the Captain (as handling pilot) elected to continue the take-off and become airborne. Being a slightly earlier generation of aircraft, the B757's air data setup does not automatically compare the three inputs. The autopilot was configured to take air data readings from the Captain's side, and unfortunately it turned out to be the side suffering from faulty readings. By the time the crew realised that the autopilot was behaving strangely, the confusion and workload had increased to a point where they were at an unusually steep pitch attitude, and they made the incorrect call that neither of the airspeed indicators were working, and they eventually lost control and crashed.
The reason I bring that accident up is to illustrate what I said at the beginning - namely that UAS in and of itself isn't necessarily dangerous. In most cases that I know of, it is the response to the UAS - and/or not paying attention to the precautions intended to prevent going into the air with unreliable airspeed readings which are the real danger. However, it's worth bearing in mind that UAS incidents are pretty rare to start with, and improper responses even rarer.