AF 447 Thread No. 5

[Lonewolf_50]

@ A33Zab: Flight Mode

In pitch, when an input is made on the sidestick, the flight control computers interpret this input as a “g” demand/pitch rate. Consequently, elevator deflection is not directly related to sidestick input.
The aircraft responds to a sidestick order with a pitch rate at low speed and a flight path rate or “g” at high speed. When no input is made on the sidestick, the computers maintain a 1g flight path.
Pitch changes due to changes in speed, thrust and/or configuration, which in a conventional aircraft would require the pilot to re-trim the aircraft, are
compensated for by the computers repositioning the THS.
The pitch trim wheel moves as the control law compensates for these changes. Sometimes, changes of trim due to changes in thrust may be too large for the system to compensate, and the aircraft may respond to them in pitch in the conventional sense and then hold the new attitude at which it has stabilised after the trim change.

When I look at the typical pattern of flying, it would seem that most "hands on" flying by crews in the AB330 would be in the take off, departure, climb, and then terminal phases, to include landing. While the SS does not have artificial feedback per se, the SS (joystick) positions needed "to get the plane to do X" will over time become part of the pilot's tactile memory. (fingers, wrist, forearm, brain all having "remembered" this pattern).

I find it of interest that the "responds with 'g' " versus "responds with pitch" zones are so markedly different. If hand flying at altitude is not practiced, then the only muscle/touch/brain memory to rely on is the patterns from the low end and low speed flying regime.

This makes me wonder: was the aircraft responding as the pilot flying expected it to? If it wasn't (see some previous posters comments on the delicacy of hand flying at high altitudes) this would explain to me an early onset of pilot frustration and confusion. (Some people have used the term panic, which I see no evidence of. There is ample evidence of frustration, and of confusion).

The PNF is telling him to do this or that with the nose, he makes inputs, he acknowledges what he needs to do, but he can't get the aircraft to do it to his own satisfaction. (Nor the PNF's, nor later the Captain's).

I had previously suggested that Loss of Control is roughly defined as "your flight controls will not do as you command them to." AF447 is not just an upset scenario. From what CVR excerpts have shown so far, there seems to have been a mild case of loss of control (as defined above) even before the stall. LOC in this case meaning "he couldn't get it to do what he wanted it to do using his flight controls."

In sports terms -- you tend to play the way you practice.

Old Carthusian:

The pitot tube issue despite being the trigger is actually irrelevant to what caused the accident ... snip ... One has to look at training and responses to unusual situations. As many others have noted - on the information we have this was not initially a serious incident.

We are actually in rough agreement. I won't beat the dead horse about necessary and sufficient conditions, which pitot failure classifies as, but yes, as I've said on numerous occasions, malfunction not emergency by itself.

I understand your point, and I have been addressing and discussing training and systemic human factors issues for some weeks.

We are in the realms of psychology and human reactions here, not physics or computer science.

And program management.

But I will disagree somewhat with your summary here, or maybe just flesh it out a bit.



We are dealing specifically with

1. the man / machine interface (which means machine design is a player)
2. both training and proficiency (and currency/recency)
3. and
4. with the overall system's influences on behavior, both formal and informal. The aircraft manufacturer and company management are both part of "the overall system" in this regard. So too the regulatory realm, and airworthiness.

I'll take this criticism a step further: if you take a min/max approach to systems performance, you will frequently find decisions made that optimize an area of key interest (say, fuel consumption) which action sub optimizes another area. (My understanding on this is informed somewhat by retail, and by aircraft operations and maintenance).

I'll leave to the reader to puzzle out whose min/max priorities are a root cause here, in terms of how one runs an operation or an industry.