therefore i try to summarize the issue about stability in my own understanding (feel free to correct).
The airframe itself is, no, must be natural positive stable, otherwise it would not be flyable in direct law.
Quite so - but when people say that they are usually talking about the short period motion. CG ahead of aerodynamic centre in the classic sense for stability - that qualification says nothing about the long period (phugoid) motion
In Nz-Law, which applies in all Laws except Direct Law Ground Law and Flare Law (did i forget one?) the flight control system creates an artificial neutral stability by use of elevators and THS trim, leading to a stable flightpath without SS input. Not the SS drives the elevators, but the flight control system using present flightpath and the demanded change from that flightpath.
C*, by itself, is a system to shape the short period response into a form that pilots will like. When implemented as Airbus do it acts, in the absence of any further pilot input, to maintain the flight path constant as Franzl says.
But if the flight path is constrained, whether that be level flight or an approach path, the natural phugoid is suppressed because you have taken away one of its degrees of freedom. What you are left with (for the 'natural aircraft at least and apart from a stable short period) is a simple subsidence or divergence in airspeed, i.e.it becomes a speed stability issue and one where what matters is how (Thrust - Drag) varies with airspeed with throttles held constant.
Something like the A330 is speed stable for excursions above the (trimmed) cruise speed because you very soon run into drag rise. Below trimmed speed however it becomes a bit more problematical.
When one takes Mach number effects as well as the classical drag 'bucket' into account modern airliners have a drag vs speed curve that looks more like a flat-bottomed 'bathtub' than the classic parabolic shape. Depending on how thrust varies with airspeed (again at constant throttle) the speed stability is just about neutral over a fairly wide range of airspeed. It varies a bit with altitude, being slightly less stable at say FL370 than at FL350.
It is because of this neutral stability on the low speed side that 'buses' use A/T in cruise to sharpen up the aircraft response to disturbances.
Before anyone starts to yell "I told you so - the aircraft is unstable without A/T" just remember that if the aircraft is only slightly speed unstable (as seems to be the case) it will take a very long time for small initial diversions in speed to develop into serious excursions. Typically, from cruise down to Vs1g might take anything from 5 to 10 minutes but the rate of divergence escalates rapidly as one gets near that point and most of the speed divergence will be in the last say 30 secs. Once you get beyond 'stall' the rate of divergence rockets up. Hence the need for prompt recovery action.
It must also be apparent that modest speed instability, by itself, could not explain this accident. To reach the timescale recorded the aircraft divergence would have to have been 'helped' by pilot action - as it was. But once the aircraft was down to speeds at or below Vs1g the aircraft's natural speed instability would have made a significant contribution to the rate at which speed was lost.
When those protections are degraded or lost, we still have an airframe with artificial neutral stability, as the flight control system still works along the demands of Nz Law, but the protections to keep the aircraft within the flight envelope are gone and have to be replaced by the skill of the crew.
This NU SS had not much influence after the stall happened, as the flight control system was beyond it´s capability to maintain the flightpath (descending already instead of continuing on the ordered climbing trajectory).
I'm not so sure about this. The flight control system would still have been trying to maintain the flight path, but the flight path it was trying to maintain was that being demanded by the pilot as seen by SS movements - an increased 'g' command. So far as I can see, the continued NU SS and accompanying THS movement would have made things worse for recovery because they trimmed the aircraft into a higher AoA than it would otherwise have achieved.
Neutral SS wouldn´t have cured the problem, only ND SS would have changed the trajectory.
Last edited by Owain Glyndwr; 18th Jul 2012 at 15:14.
If Bonin was the speaker, then the words "allo? oui allo?" etcetera would have been placed in his column, labeled "Copilote place droite".
As it is, the words appear in the column labeled, among others, "Autres voix".
The mystery of the voice appearing on Track 1 (Piste1) is not so mysterious - if she picks up her phone, she is on the line, i.e. will be recorded. Compare it to the situation where a pilot pulls his headset cables from the plug and keeps his speaker off - what is received via radio will still be on the recorder track, even though that pilot will not hear it.
I guess we have to define terms, and "flight path" is one term that may be causing us to debate a few things about the control laws and the natural aero characteristics of the 'bus.
A very good explanation of some things, Retired. But I expand...
Courtesy of some of the pilots here, I have a decent set of the manuals. And from what I see, the FBW system is strongly biased for a one gee flight, and not an "attitude hold" mode as we had/have in many jets with autopilots. It also appears to have little AoA bias, but only some "limits" to stay outta the stall regime. Body rates and gains are contributing inputs, as was also the case in my primitive system. The positive longitudinal stability most of us were/are used to is not based on AoA in this system. Let go of the stick and the sucker shoots for one gee corrected for pitch attitude. If that ain't the way the thing works or is supposed to work, please correct me.
I iterate this point because it's exactly the way our primitive system worked. The biggest difference was we could set the desired trimmed gee, and our system wasn't trying for one gee all the time, nor did it correct for pitch attitude. We had zero static stability "feel" for AoA as most of us are/were used to. Only reason the nose went down when relaxing pressure on the stick was to achieve the trimmed gee ( nominally one gee according our trim wheel, which allowed about 3.5 positive and a bit over one gee negative). If we trimmed full back, and let go, the thing would do a beautiful 3.5 gee loop, and the gee would decrease as we hit the AoA limit until we got to one gee or a bit less. AoA would be pegged at 27 degrees. Coming down the back side, and speed building, the thing would get back to the 3.5 gee trim command and finish the loop.
I bring all this to the table to help folks understand the difference between FBW systems and the conventional control systems we flew that were based on a trimmed AoA. Sure, we could build a great FBW system that was strongly biased for AoA, just like the "old days". In fact, our system did that with gear down. We biased the gee command with AoA to give us the same old feeling we had in the days of yore ( no auto throttle in the Viper, unlike the Hornet, so we could trim for an AoA and use throttle for vertical velocity). In a FBW system you can overcome all kindsa nasty aero characteristics until the confusers go to la la land. The 'bus seems to be a very benign, stable platform. So even in "direct law" it can be flown by humans. Ours was not stable until above 0.95 M or so, and if the control surfaces went to neutral due to computer failure, the thing went nose down ( 22 negative gees on one of our first computer shutdowns due to power failure).
We could build a FBW system that only used the electrons to command control surface position, just like the old days except the hydraulics didn't use a valve at the base of the control wheel/yoke/stick. We could even build one that used electrons to power the actuators for each control surface.
But we haven't done that. We use the neat computers to reduce workload and overcome aero characteristics of the platform. So we see a lot of "autopilot" functions embedded in the system. Those are the things that bother me. Those are the things that gradually destroy basic airmanship and understanding how our jet flies.
I have no idea what Lyman has been smoking. His suggestion that the female voice was Bonin's is ludicrous. According to Otelli, the four transmissions in a female voice recorded on the CVR came through on the intercom speaker, were all made by cabin crew, and were all questioning or exclamatory (he transcribes the call at 2:10:59.4 as "Oui?" and the one at 2:11:24.9 as "Stéphanie!"). Even more ridiculous is Lyman's theory that the aircraft was not responding to Bonin's control inputs. It reacted exactly as expected. At 2:11:34.7 (ca. 1.5 min after Bonin had pulled back on the sidestick following A/P disconnect), at around FL360 and TOGA, Bonin announced that he had lost control of the plane and proceeded to pull the stick even further back and to the left and hold it there for 40 seconds. Equally silly is the suggestion that Dubois, when he re-entered the cockpit, did not want change places with Robert for fear that all control of the aircraft might be lost. They were still at 35,372 ft, surely to God you'd think you'd still have a fighting chance to rescue the situation. Finally, let's put at rest, once and for all, the impression that all three pilots were experienced veterans of long-haul flying. Bonin definitely was not. Before leaving the FD to take his rest and designate Bonin as PF, Dubois had to assure himself that Bonin possessed an ATPL.
My suggestion is you consider that without a complete record, posts like my last one cannot be contested without either full and naive faith in the report, or possession of substantiation that is not in the public record.
We can continue for months to discuss technical specific aspects of the Airbus A330 and continue to speculate about what would happen if and if ..
Nevertheless .. all of these discussions will not bring anything more than what is in the BEA report and even if one member of PPRuNe find the Holy Grail to solve some problems .. the final report will remain as it is .. that is to say engraved in stone
The next episode (for those who are still fortunate to be there ..) will be the trial court
It may be that short of it .. other experts (judicial) have other discoveries to argue ... and maybe some explanations will be sought from BEA and regulators and airline(s)
The debate (such as currently conducted in PPRuNe) will certainly be very interesting from all point of view .....
As I understand the C* control law: C* = q + a*nz, where q = pitch rate, nz = incremental loadfactor, and a determines the 'blend' between nz and q feedback. For side stick neutral C* = 0, hence q = -a*nz
OK, thanks. I really should have paid more attention in math classes, back then. Would that make the Alternate law stick free behavior different from that of the Normal law? q limits is different, yes, but?
the A/P would need to function differently in either law
I believe I've read somewhere that the different laws only affect the flight control system interpretation of side stick inputs. The autopilot commands are processed outside of the laws. Protections override side stick as well as A/P commands.
Last edited by HazelNuts39; 18th Jul 2012 at 21:00.
There are differences in Pitch law between ALt1 and Alt2b
@ Owain Glyndwr, mm43,HazelNuts39, AlphaZuluRomeo and all
I read the following out of the A330 technical training manual that AF447 had no high speed or low speed stability protection left (as it would have been iN Alt1), but only protection still available being the load factor protection.
ALTERNATE LAW WITH REDUCED PROTECTIONS In pitch normal laws, all the protections are available. In case of loss of surface actuation or of inputs, some protections can be lost. The pitch alternate laws with reduced protections are activated in the FCPCs in case of loss of :
- THS actuation (B+Y hydraulic circuit or 3 motors lost, ...) or - 2 Inertial Reference (IR) (second self detected) or 2 Air Data Reference (ADR) o
- All spoilers or all inboard ailerons or
- Slat position or pedal position or
- One elevator (B+G or Y+G hydraulic circuit lost, ...). The angle of attack (AOA) protection is lost and replaced by static stability Vc prot. (see Pitch Laws module). The pitch attitude protection is lost. The high speed protection (VMO1) is lost and replaced by high speed stability VMO2. (see Pitch Laws module). In case of loss of slat or flap position or aircraft weight computation, the static stability is also lost. If a double IR failure occurs, the second being self detected, specific accelerometers are used to consolidate the load factor information and pitch attitude rate information from last IR.
ALTERNATE LAW WITHOUT PROTECTION In this case, the pitch protections are lost except the load factor protection. This alternate law without protection is activated in the FCPCs after a triple ADR failure.