AF 447 Thread No. 12

[Owain Glyndwr]

think I get what the "report" was getting at, but I also think it does not apply in direct law.

Gums, I agree – in direct law the conditions that gave rise to the special condition do not apply so the aircraft has to meet JAR25.173 and 25.175 as written. It is certificated, so it must meet those requirements, but this is not strictly relevant to the present discussion.
I must admit I struggle to get my head around statements like:

This absence specifically results in the fact that it is not necessary to make or increase a nose-up input to compensate for a loss of speed while maintaining aeroplane altitude

I think I understood it eventually, but at first it seems counterproductive to apply a nose up command to compensate a speed loss.

But this is the first I have seen for the jet not having positive static stability. I mean basic aerodynamic stability, unfettered by computers using rate and gee and air data sensors

I think the problem in part arises from the labelling of the FAR/JAR paragraph. When “static stability” comes up my mind (and maybe yours gums?) goes to considerations of short period response and return to the trimmed AOA after a disturbance. I don’t think there is any doubt that the ‘bus is stable or strongly stable in this respect in ANY law.
The FAR/JAR wording however relates solely to speed variations seen through stick movements and which involve excitation and damping of the long period (phugoid) dynamic characteristics. But the ‘g’ feedback in normal and alternate laws controls the aircraft to a steady flight path which takes away one degree of freedom and the phugoid disappears to be replaced (in speed stability terms) by a one degree of freedom convergence or divergence depending on where the aircraft sits on the (thrust – drag) curve. The conventional stick/speed relationships then do not apply; hence the SC, which says that:

longitudinal static stability characteristics will need to be determined on the basis of the aeroplane's response to disturbances

Because of compressibility effects the (T-D) profile of a modern airliner in cruise is more like a bathtub than the classic parabola. At the high end of the speed range, where the aircraft is hovering on the edge of drag rise, the response to a speed disturbance will be stable. From there down to the minimum operational speed (say 1.3Vs) the aircraft will be marginally stable. It would be not until one gets below that that instability creeps in.
This might be seen as weasel wording, but given the probability of dropping to ALTN law (most pilots will never see it throughout their career we are told), a marginally stable aircraft which meets the book (just) in normal operating speed range might be considered acceptable even without protections.
Below 1.3Vs the aircraft is speed unstable and this is where the BEA comment to the effect that the aircraft might drift to stall even with zero stick input becomes valid. However, if one puts some real numbers into the equations one finds that in level flight it would take about twenty minutes to decelerate from 1.3Vs down to Vs. Put some mild climb demand into the sums so that the initial thrust deficit is greater and a very different story emerges – that deceleration could take twenty seconds rather than minutes. [My numbers might be challengeable, but I think the principle holds – if you start from a speed yielding nearly neutral stability the initial divergence rate will be very low]
So the ‘hole’ in the certification would appear to be that the static speed stability in a combination of alternate law and flight outside the normal flight operational envelope was not checked. [But note that FAR/JAR 25.175 does not apply to speeds less than 1.3VSR1 so if this philosophy was carried over to the SC there would be no hole].
The point I am trying to make is that an aircraft may be neutrally speed stable in response to stick inputs but might still be speed stable in response to atmospheric disturbances and it is this latter that is used to define whether the requirements are met.
Gysbreght, I agree that according to JAR principles the full handling requirements might not be applicable in a failure case (and jcjeant, ALTN is a consequence of some sort of failure so the probability principle applies) but you cannot abandon the handling requirements altogether, so there must be something there appropriate to the failure probability level. Since there are no specific requirements relating to ALTN, my guess would be that the normal requirements were applied, albeit with some less stringent interpretation. Others may wish to comment on the probity of more relaxed interpretation of the rules at a time when things are getting difficult!
I have some difficulty with the logic of the final BEA sentence in:

The specific consequence is that in this control law the aeroplane, placed in a configuration where the thrust is not sufficient to maintain speed on the flight path, would end up by stalling without any inputs on the sidestick. It appears that this absence of positive static stability could have contributed to the PF not identifying the approach to stall."

He might have been unaware that the aircraft could be driven towards stall without any input from him, but why would he not be aware that maintaining up elevator for a prolonged period would drive the aircraft towards stall?