jcjeant

Hi,

Not an aswer to your question but in the BEA note:

Around fifteen seconds later "at + - 2 h 12 min 17" , the PF made pitch-down inputs.
For how many time ?
And
the speeds became valid again
the angle of attack decreased (but still over 35°)
So .. abnormal law or not ?
simultaneous inputs by both pilots on the sidesticks
Up or down inputs ?
As already commented before this BEA note don't give real clues about what are the pilots actions and what they tell ... it's just a riddle
Not big difference between this note or any press article in a newspaper like "Le Figaro"

Lonewolf_50

gums:
Doesn't that add another feedback/scan loop to any attitude change on instruments? Instead of setting a pitch attitude, you make a 'g' input and correct to desired attitude? Maybe I am overthinking this, and in practice it boils down to the same thing - set pitch by looking at the attitude indicator, then trim it in (Viper Coolie Hat) or AB auto trim?

gums

Your thinking, Wolf!!! We might all be "thinking" too much, but down deep I feel we're homing in on a quirk with the Airbus design.

To clarify my understanding, it seems the 'bus trims for one gee ( Nz) in "normal" modes. It is roll- compensated ( unlike our Viper), so the pilot need not back trim while in a turn. Makes it easier to maintain a level turn, but how about a descending turn or a climbing turn? It is also compensated for pitch (unlike our Viper).

So my feeling is the thing trimmed the THS for a nose up attitude, and things got worse with low or no speed inputs to the "confusers". Worse, the AoA protection seems to have gone on vacation, so we have this beast with almost full nose up pitch trim and no AoA input helping to get the nose down. Something ain't adding up.

Our jet had no elevator - the whole stabilator moved, both of them. They also moved differentially to help roll. Nevertheless, our jet trimmed somewhat like the Airbus to maintain a gee command. Gear down it blended AoA and pitch rate into the gee command to make it "feel" like a normal jet. We still had a little trouble trimming for an AoA or speed like "normal" jets, but what the hell.

Only time we had "autotrim" was with otto on in "altitude hold". So if we pulled back the throttle the thing would trim nose up as we slowed until reaching 13 or 14 degrees AoA, then we would slowly descend( had one CFIT doing this, BTW).

Bottomline: The jet didn't help the pilots. The pilots were nor sure of the exact control authority and "protections" that they had. There will be a sad finding in the end, I fear. And i pray that training and maybe some slight mods to displays and control laws will be implemented.

Lonewolf_50

gums, thanks (sorry about the CFIT, losing a brother pilot always hurts) for the amplification. From what the BEA has released, it appears that the THS commands to nose up were from the side stick commands ... but if that was in response to a low speed stability input, a previous trimmed condition, or anything else (or a combination) remains unclear to me. Will hope for clarity in due course.

I am still not sure how well the BEA analysis can suss out from the data available just how turbulent (or smooth) the airmass was just before, during, and just after the initial upset. Perhaps their next release will have some more clarification on that, as flying in turbulent air not infrequently adds to the number of control inputs (particularly if the auto trim in one channel or another is not in play). Working near cells that big, you are bound to have some vertical movement of the airmass, or run across some. The question is, at what magnitude?

GarageYears

This discussion merits a momentary distraction.

First the idea of driving the PNF stick in response to the PF inputs. The first issue I have is that the side-sticks are 'handed' - the left stick leans toward the cockpit center by perhaps 5-10 degrees, the right stick leans the opposite direction, presumably to 'sit' nicely in the appropriate hand. Does that have any implication as far as the PNF interpreting the stick inputs, as opposed to giving those stick inputs. I suspect it does? Secondly, my impression is that stick inputs are generally subtle (since the stick itself does not have much travel - not sure of distances). So the second concern is whether these stick inputs would be particularly noticeable, particularly when things are going to hell in a hand-basket? I believe gums noted in prior post that the PNF is much more likely to notice the change in 'g' due to control inputs (my butt is being pushed into the seat = we're pitching up...). Given that the sticks were motorized and driven, what happens when a dual-input occurs? Currently the system issues an audible warning "dual input" and I believe sums the inputs. Override can be achieved by pressing the "take control" button. With a driven system what happens?

The throw away comment above regarding stall warning is spurious and flippant in my book, you could equally say the same thing about a fire warning, etc. Nothing to do with what was being discussed.

DozyWannabe

My question is how, given the "confusers" - as you delightfully put them - were not in control of the aircraft after the loss of speed indication was detected, was the "thing" supposed to command a THS pitch-up? The autoflight system will have maintained a slight pitch up prior to disconnection, but nowhere near the angle the aircraft ended up at.

AlphaZuluRomeo

@ jcjeant
"So .. abnormal law or not ?"
Conditions are met for the abnormal attitude law. So, I guess yes, until proven wrong.

@ DozyWannabe
IMO, the pitch up THS came from the pitch up pilot inputs.
Basically: autotrim.
Pictorial: "My pilot wants nose up", thought the plane, "let's help him with the trim, so he doesn't have to hold the sidestick anymore".

DozyWannabe

AZR, that's certainly a possibility, but I don't want to say one way or the other until we know for sure. It's all too easy in this kind of forum atmosphere to equate the theory that the THS movement was pilot-induced with "blaming the pilot". I know they're not the same thing, you know they're not the same thing, but there are some who will get indignant about it.

gums

Well, Doze, this is the big question.

How did the THS get trimmed to near the limit?

What happened to AoA protection? And that looks to be a biggie according to the manuals' description of the "protections'.

How did the crew not see attitude steadily increasing? I thought attitude sensors and displays were independent of the pitot-static system? Same for AoA protection.

I can tell you exactly how my little jet trimmed the stab to full nose DOWN as it entered a stable, fairly smooth deep stall. We ran outta nose down pitch moment. Already posted the link and enclosed snippets from the test pilots. So we were briefed, watched movies, had a new switch installed to help us, etc. Better yet, our designers and military management were not so damned proud of their design that they refused to change a few things. They didn't continue claims that "the system did exactly what it was suppose to do" ( even tho it did EXACTLY what it was supposed to do, it wasn't designed for the condition we clever pilots could get to).

But now we see a similar phenomena in the Airbus, and due to a combination stabilizer and elevator, it is possible that the elevator doesn't have the control authority to get the nose down at full forward stick due to the disconnected/locked THS.

Machinbird

It would seem logical to have a slight time delay in activating autotrim.

When you make a demand on the elevator, that needs to be in the context of the aircraft's current trim.

If the aircraft were to quickly trim while you were making a pitch input, the aircraft wouldn't be in the same trim when you relaxed the input.

The idea must be, lets wait and see how serious the pilot is about this control demand. If he keeps it up for say 5 seconds, he must be serious, so we can trim some of the back stick out.

But in the case when the nose down control input returned airspeeds to functional in the deep stall. The stall warning return probably discouraged the nose down input inside of the delay time. Therefore the THS did not move toward nose down.

The actual aircraft control logic is no doubt proprietary to Airbus. We will only get applicable snippets of it in the control analysis when BEA finally publishes. There are probably clues in other AB accident reports how the thing works.

GarageYears

Gums: The initial 'zoom'climb' (for want of a better description), occurred prior to significant THS movement, presumably as a result of normal crew input. Only after "the PF maintained nose-up inputs", following TO/GA selection did the THS move significantly away from what might be considered a normal attitude of 3 degrees.

I believe AoA protection is diminished as follows in ALT LAW:

My reading of the BEA note is that the THS did exactly what was commanded by the PF. The aircraft was already in trouble quite a while prior to the THS reaching 13 degrees.

gums

Right on the spot, Doze!

I realize that the THS didn't start moving until later in the sequence of events.

Your AoA protection quote bugs most of we pilots here, as AoA is the primary concern with a stall or approach to a stall, not airspeed. I look at the wimpy AoA bars on the display and it's hard to understand how an important indication is not "expanded" vertically when it becomes an active player in aircraft control.

I also note confusing inputs to the "confusers" re: speed versus AoA when either or both becomes primary for "protecting" the pilots and the jet. You would think that the jet would revert to AoA if speed was deemed unreliable and SCREW THE SPEED! And then quit giving warnings until the speed was not only deemed reliable, but well above the computed stall speed ( but AoA still determines stall, not speed).

BTW, we didn't call our stuff "protection" as much as we called them "limiters". About the only "protection" control law we had was added during FSD, and it was a yaw command by the system if AoA was above the limiter value. It was one of the things that made our deep stall stable and kept us from entering a spin. It didn't work inverted, so it was possible to enter an inverted spin when in an inverted deep stall, kinda scary, huh? Oh well, a small thing we had to put up with, heh heh.

bearfoil

For whatever the cause of a/p loss, the need for PF to 'correct' ('climb', 'roll'), and THS behaviour, the time of accident begins at 15 seconds prior to, and perhaps 15 seconds post, a/p loss. Upset prior to loss of a/p? Quite likely, for whatever the reason it dropped out, the definition for upset was concurrent. If in a climb that cannot be corrected, that is LOC. Whether electro or Weather, the definition for mechanical failure also is met, for loss of authority qualifies as loss of mechanical controls. Airframe damage is also quite possibly in there, the g string will be interesting.

DozyWannabe

You still hoping the tail came off, bear? Yeesh!

@gums - GarageYears was the one talking about angle of attack - I'm a logic man, trig and physics aren't my strong points!

gums

@ Doze: Sorry to cross-reference posts. I apologize.

@ the Bear: It's possible that parts came off the jet or that some systems had mechanical failures ( not computer failures), but I doubt it. The CVR snippets we have so far are not filled with "whoa!", "Holy crapola!", or other indications that the ride was very rough.

The gee and other aero data from the recorders will provide some clues as far as parts falling off or actuators jammed or.......

@ all here:

I only use my own experience and understanding of another FBW system, with it's "laws" and logic, so folks won't think all FBW systems are the same. Additionally, I hoped to let folks know that despite some claims/assertions that the stick moves the control surfaces the same way, the same rate, the same amount, etc. all the time, even in alternate laws - THEY DO NOT! We're not in Kansas any more, Toto.

I also wanted to show folks that there are flight conditions that the designers never figured on when developing the system logic and gains and "forward-lagged rate" inputs to the commanded gee or roll rates. Same for the reverse. For example, the roll rate command we had used a lot less onset rate when rolling into a bank than rolling out. Next time you see a T-bird performance, note that the jet stops the roll on a dime. Contrast this with the Blues, with their pseudo-FBW system. I can see the slight hesitation when a Blue stops his roll ( the solo guys/gals). The Viper looks like a video game and the roll stops instantly.

bratschewurst

From reading the many posts on this and previous threads, as well as some of the linked material, it appears to me that the Airbus design philosophy is biased against what might be called "tactile feedback" to the pilots. This bias appears, in turn, to stem from a philosophy that, in regular operations, it is best that the pilots to tell the computers what they want the airplane to do, rather than command the controls more or less directly to achieve aircraft performance.

In a sense, AB aircraft have two autopilots; the traditional one that would, in a conventional aircraft, command the controls to do what the pilots command via knobs and switches, and the flight control system, where the pilots move what appear to be more-or-less conventional controls (stick, pedals and throttle) that actually only command the computer to move the controls with considerable computer intermediation to achieve, for instance, a stable bank angle or a max climb angle without stalling the aircraft.

I can see how the idea that the pilots never actually command the controls would lead to AB believing that tactile feedback through the controls (or control position, as in moving throttles) is not necessary. But that philosophy is flawed on the rare occasions when the pilots actually are commanding the controls directly, as was the case for 447 in roll in ALT2 law. The PF had no feedback from the stick other than a/c performance, the PNF had no physical feedback from his controls re what the PF was doing, and (according to an earlier poster who had flown the 330 in ALT2 law) the aircraft performed quite differently in response to stick inputs than did the simulator, which of course is almost certainly the only time the PF experienced ALT2 law prior to the incident.

So, in essence, the PF was flying that particular configuration for the very first time that night, in the soup with contradictory speed indications, and without the other crew being able to fully follow just what he was doing.

This was compounded by AB using the same interface, with the same lack of tactile feedback, for two very different tasks: telling the computers to place the a/c in a certain attitude and hold it there in Normal Law vs directly commanding the ailerons in ALT2 law, which runs the serious risk of the PF not making the appropriate modifications to his inputs until after a short period of actual experience with the differences.

Microburst2002

Could it be an erroneusly low airspeed in all ADRs such that the airplane was actually deep in the overspeed and then the APs disconnected and the pilot carried out unreliable speed procedure memory item when above 10,000 ft: pitch 5º, climb thrust.

What was the pitch and AoA at the time of the AP disconnection? I think that is a critical information. is it available?

if these were low, that scenario would explain the pitch up input, and such a sudden and large pitch change (say almost 5º) at that level and speed (overspeed) would make possible a very large lift increment and climb rate. but I am not sure that a stall could result from this.

we will have to wait.

rgbrock1

Forgive me for asking the following question which I have not seen answered anywhere else but, if it has, mea culpa.

From the information provided by BEA so far, it seems that the aircraft in question descended at a great rate of speed. Would not this "great rate of speed" descent be felt by anyone? Wouldn't PF and PNF, as well as the captain, not actually feel this descent.

I know as a passenger when an "air pocket" is encountered and the aircraft loses
a bit of altitude that is certainly felt. So I can only imagine the descent experienced by
AF44 would be felt quite a bit more. No?

deSitter

A constant descent rate = no vertical acceleration = 1 g. However the out-of-vertical part would be sensible if the attitude of the plane was more than 15 degrees nose up. The plane would appear tilted upward to passengers with good sense of balance.

Tex Johnston's famous roll of the 707 family prototype was done at 1 g relative to the fuselage. A glass of water could be set on the dash and would not empty during such a roll, if well exectued.

lomapaseo

rgbrock1

From the information provided by BEA so far, it seems that the aircraft in question descended at a great rate of speed. Would not this "great rate of speed" descent be felt by anyone? Wouldn't PF and PNF, as well as the captain, not actually feel this descent.

I know as a passenger when an "air pocket" is encountered and the aircraft loses
a bit of altitude that is certainly felt. So I can only imagine the descent experienced by
AF44 would be felt quite a bit more. No? [/quote]

Take your bike, coast it off a cliff with your eyes closed. Tell us if you felt anything before you hit the ground.

Forgive me if you've already done this