TLB

OFF TOPIC.
Reminds me of one flight I had during my CF-101 Voodoo experience (2,000 hrs). I was wing man and while in IMC, lead called me into line-astern position for maneuvers. After a few minutes I remarked to my backseater that - gosh, did I ever have the leans - and I could swear that we are inverted right now. He responded: "Yep we are - this is the last half of three rolls" !
Back ON TOPIC

PJ2

Hello Bearfoil;

Understand what you mean but there is no indication in the BEA Update that the aircraft was not in stable flight, likely level, not climbing or descending. The QRH drill and checklist are clear. The meaning of "stability" in this context has to do with steady-state power settings and a level pitch attitude as required by the QRH checklist. Certainly there are going to be pitch variations around a "nodal point" in turbulence but the power isn't going to change once the AT is disconnected.

An autopilot disconnecting in cruise is not a big deal but one certainly, immediately, looks for the cause. I recall from discussion with people who knew the airplane well (when learning the airplane in 1999) that the EFCS is "way ahead" of all systems - and responds much faster than the pilot can to changes or system problems. That it took ten seconds or so between the AP disconnect and the annunciation of the pitot problem may be due to the extensive BITE test processes which ensure the accuracy of the fault, (correlation) and both communication (SDAC > FIDS > FWC1 > FWC2 > CMC, CMS & DMCs). Lack of correlation will cause a WRG fault to be included in the fault message. During that time we do not know what either crew member saw on his PFD in terms of failing AS. We cannot know what the #2 PFD displayed but I have argued elsewhere that it was about the same as the other two; no pilot would try to control high speed with such a pitch-up. The power would be brought back first, then possibly speedbrakes used. We will know soon enough.

Regardless, a loss of airspeed information is no time for swift action, especially in weather. It is time for deliberate, measured action that is slowed down and closely coordinated with the PNF, (announcement by the PF of the drill or checklist, PNF's participation in the checklist or memory item and/or concurrence if a change in flight path/direction is indicated, completion of the challenge-response checklist, securing the aircraft after the abnormal is over, communicating with ATC, Company, F/As).

It is no time for individual action which leaves the other crew member unaware of the course of action and therefore unable to assist. We do not know what occurred during this period except that which the BEA Update includes. Again, we will know soon enough.

The failure to log onto DAKAR still bothers me. In itself it is no big deal providing communications are established on HF (which they were) but they needed a clearance to deviate off-course and didn't get it. CPDLC makes such clearances easily and quickly obtainable. It didn't seem to be an emergency diversion around a build-up they didn't see. Anyway...another item for the BEA to think about.

RR_NDB

bearfoil Question: During their ordeal the System presented a transparent (an clear) picture to the poor guys? They were informed why the plane "degraded". They were reliably informed they were stalling? The stall warning worked as expected by a crew badly needing reliable information? IMHO this is "little bit" closer to "partner opacity".

The "persistent NU..." (from BEA) is coherent with "transparency" from the partner to the PF?

HarryMann

It has occurred to me, that when comparing the flyability, ergonomics, shock resistance & recovery ability of the Airbus FBW cockpit and control systems, as a flight control centre, it should be borne in mind that effectively what Airbus did in a decade or so was:

re-invent the cockpit form a pilot's perpective. Throttle levers that don't feedback, sidestick control, glass display including 'digital' and 'text' (albeit some strip) instruments replacing mostly analogue, etc.

A pretty monumental paradigm shift - but did everyone truly admit that this is what it represented, rather than perhaps Airbus & others leading the industry to describe (even excuse) it somewhat as 'not really different in principal'.

It is in fact very different, all-round, and in principal to what three-quarters of a century of flight up to that point had evolved as a modern cockpit...

It has been 'bedding down' ever since, into the aviation world's sub-conscious. Additional settling in is still taking place... the investigation of this accident, as so many before with more conventional cockpits had done, will be another step to fully accepting some of the imperfections that this man-machine interface represents and correcting them.

Only by being very open about the crew's cockpit experience, and their response - what was expected and what came naturally, can we further improve the modern 'cockpit flight centre' by integrating the best of both worlds.

Like the last UK Govts attempt at a 'truly joined-up government policy', the dots might all be there, but most are still quite some way from being even crudely connected, let alone integrated as a comprehensible, efficient and finely optimised cockpit for flight control.

It seems inevitable that mistakes were made in many areas of design as well as pilotage; even at this early stage it worries me that justifying them (individually) is quite possible, but never as a joined up whole for safely governing an aeroplane's flight under extreme and singular conditions.

jcjeant

Hi,

Why always repost this speculation ?
By myself I read (and it's not a speculation .. but a fact!) the aircraft was stable because he was in a cruise flight stabilized by the autopilot and auto throttle
When autopilot and auto throttle go out of the loop .. we don't know if the airplane stay "stable"
So maybe .. instead of touch nothing (hands off) the pilot had to act immediately for stabilize the plane
Remind .. the plane immediately banked to right (unfortunately no clue about the bank angle from BEA note ....)
So will the pilot wait the aircraft go in a 30° or more bank angle before react ?

bearfoil

Hi PJ2

Thanks for such a quick reply. I am happy to see you acknowledge that BITE may have forestalled a cockpit annunciation of the problem. Would this not mean that the PF directly after a/p loss would be flying with the assumption of NORMAL LAW? Otherwise you are condemning this man to a blunder: to wit, not being patient with the a/c post auto? At the least it establishes that there may have been confusion re: LAW.

As I have said, if the PF was performing as you state he should, then some control inputs would be most normal in NORMAL? His acknowledgment of "So, lost speeds", is an admission in fact that ALTERNATE LAW is a new development for the crew, occurring as it did eleven seconds after a/p loss. This is a very long time to be unaware in a task laden cockpit, with MASTER CAUTION, and a STALL STALL in the interim twixt a/p drop and observation of "Lost Speeds", No?

The Note states that the a/c passed through ten degrees NU and increasing "Before it started to climb", why so long if the a/c attitude was stable at 'Level'? Was this a/c starting from a ND? BEA doesn't give a clue.

However, if in any way established nose down, (descent), this would explain the extra time it took for her to reverse her nose transit to "UP".
It also would explain a lagging and exaggerated AoA at the initiation of the zoom. 7000fpm for 35 seconds is 3500 feet (+). Yet we are told that she had PF ND inputs which reduced the Vs to 700 fpm, and continued to 380.

Had she started to climb from 350, she would have flown to 385 and then coasted up to 390. This is not the case, and makes a good argument for her beginning this post a/p climb lower than we think.

I also appreciate your acknowledgement that mechanical issues may have been in play. A Speedbrake deployment is ripe for explaining the onset and chronic display of Right roll. A left Roll correction would drop the Outer (right) aileron, and complete a split surface drag that would be remarkable. If the spoiler was not retracted for any reason, the Starboard wing would continually want to horse the a/c down and right.

Kind regards, and deep respect.

HarryMann

jcjeant

But the a/p does 'give up' before the a/c is out of control or in an excessive or unusual attitude, surely. That is what it is programmed to do, and not wait until pitch or lateral control is as good as lost. It's a fair or reasonable assumption, though one could postulate a single upset that 'might' at one and the same time disconnect the a/p and fully upset the a/c...

Seems from the last BEA report that's not what they are thinking though...

bearfoil

Harry Mann

Hi. The a/p is required for flight into turbulence, and its limits are virtually identical to what is defined as "Upset". So, the a/p could easily have been bumping up against its mechanical limits for continued select. Technically, then, the PF was (may have been) in upset when he "I have the controls"....? LOC came with the STALL, and the initial STALLSTALL may have been quite valid, exceeding the AoA threshold for warning as the ship climbed (rotated) smartly at .81M. Comes the MASTER CAUTION, and the CAVALRY CHARGE, and the workload (and sensory load) get rather.....extreme?

cheers

PJ2

HarryMan;
Yes, I think so. Each paragraph in your post has something important to say about the change, or rather "the shift", but your last paragraph is exceptionally important, the justification being economic.

The traditional controls (even the stick) are a concession to history, but in truth, because it is all digital, the thrust levers, (for example) could have been four push-buttons (for each detent) or sliding dimmer switches. The "cultural" collision has been between the necessary ultra-conservative, (as in extremely slow to change/alter familiar controls which we operate on an entirely sub-conscious level) and the entirely new developments in control (with a bow towards ergonomics) made possible by reliable, light-weight chips and wires. (I do not mean to downplay here the complexity and vision of the concepts which underlie the design).

The failure hasn't been in the concept and resulting design but in the lack of anticipation of the philosophical change in "world view" from the cockpit with regard to very specific technical tasks, right down to the practical "how do it know?" puzzlements of crews brand new to the airplane. Your comment regarding "being open" means (and meant) to me that if one didn't like what it was doing, one disconnected it and flew it like a DC9 or a B737. My 15-year experience reified that view completely.

One just simply didn't accept that this airplane, any airplane, couldn't be stalled and so flew it with the same respect and knowledge one had for the boundaries of controlled flight and certification limits as one always had with one's aircraft. It doesn't take much observation to see that that understanding has changed over the years.

In early years of "the shift" however, it wasn't "more of the same" in the sense that the shift from piston-to-jets, straight-wing to swept wing, mid-altitude to high altitude flight took beginning in the late 50's was still with a conventional cockpit with conventional controls. The accidents were due to slower acceleration (engine and therefore airplane rates - no big 'wash' over the wing from the props, - higher weights, higher altitudes where yaw dampers and Mach number meant something, an extremely clean profile, and the effects of buffet boundaries never encountered in piston flight.

Indeed, if one is entirely open it can be successfully done and to a very large extent, has, as the accident data has borne out. The other side of "the shift" is reading into the design and the aircraft, more complexity than is there which leads to second-guessing and out-guessing the solutions to the problems of flight, which, as gums consistently discusses so well apropos his fighter experience, still are new to transport aircraft. Despite the complexity underlying the relative simplicity of the cockpit, the airplane remains an airplane when one "looks through" the levels of automation. Despite minor disagreements, BOAC's and NoD's comments, as I have read them, reinforce this important notion. In this accident, I think that there is very rich learning for those well beyond and upstream of the cockpit.

jcjeant

Hi,

Well at today technology available a plane like a Airbus can be piloted with some tactil (touch) screens .. a keyboard and a mouse as bonus ...
All is available for that ....
It's already at work in some industries ......
And all this can be remote controlled if necessary ..
The ingenuity of humans and progress never stop ...
We go on the Moon ... after all ....

john_tullamarine

Ah, yes.

However, I think PJ2 is more concerned about the problem of "when it's working it's real fine but when it falls in a heap and dumps the lot in the pilot's lap ... what have we done to ensure that the average line pilot - and, in this case, the average pair of line F/Os - is up to the task on the day (or the dark and stormy night, as the case may be) " ?

We should all acknowledge the overall fine record of the Airbus approach to life and the general statistics of the safety record but, when I look at the postulated attitude presentation a couple of pages ago and consider that that looks not unlike an early jet A/H (think, say, a 733 with the FD selected OFF) - which, together with a handful of throttles, we could all use to pilot the aeroplane quite nicely .. I am left wondering just a little whether the Industry has been seduced by the siren of Gee-Whizzery and, perhaps, lost the basic pilotage competence plot .. at least just a bit ?

gums

TNX for nice words, PJ.

I only jumped in here because I am now a SLF and it's really hard to "ride" when I see things happening and wonder what in the hell the troop up front is doing, or is capable of doing. I just want to feel safe!

If we all wish to have the remote operator back in an air-conditioned room, with all kindsa displays and such, who "steers" the jet, then I will quit riding.

When I look at all the "protections" such as bank angle, AoA limits, auto-throttles that depend upon various flight phases, etc. etc. ad nauseum. I get scared. Don't get me wrong, as I used my autopilot a lot in the lites I flew. I used them to reduce workload, and not to accomplish the mission. But this 'bus implementation seems too much of a pinball machine or video game for 90% of the mission. And the philosophy is reflected in all the "laws" and "alt laws" and sub laws.

The plane we're discussing seems to be a very fine jet, with great aero performance and stability and...... It does not appear to have any terrible or unusual flight characteristics that a basic pilot could not handle. The big "BUT" for me is the plethora of "protections" and laws and sub-laws and on and on that do not seem necessary. I fully understand autopilot functions that ease the workload. So my big "BUT" is the extent to which those AP functions are integrated into the jet's flight control laws. And IMHO, they need not be so closely coupled.

Sorry to rant...

Svarin

syseng68k,

thank you for your response regarding "certification after initial" for computers and other parts. Will look it up.

Do you mean subsystems inside one given computer/unit, inside one PRIM or inside one ADR for example ?

Or could we have interaction between systems (one PRIM talking with one ADR), such interaction being along a multiplexing line shared by many other units ? How is this last one dealt with in "recurrent-next-version-testing" ?

I understand you might wish to distinguish between these two issues, although from the perspective of the "user", this still appears as a "bug" (and not as a feature ). I mean "bug" in the sense that a computerized system acted in ways unexpected both by its user and its designer. It remains explainable and tracing the lines of code one would understand why it did that. But the main issue is that behaviour of the system is not expected by anyone.

As a native speaker told me : "I know it's not what you meant, but it's what you said." This about sums up a workable definition of a "bug" for me, both from the programmer's point of view, and the user's.

Graybeard

Back to the PFD illustration. Is there any evidence the altitude display had been removed as shown?

CONF iture

gums,

I do agree, the 320 and followers would have been much better aircrafts without such complexity.

FBW was needed to save weight and attract the clients, fine, were the protections their request too ... I don't think so.

Starting with the autotrim under manual flying ... who needs anything like it ?
To have to trim under manual flying has always been part of the pleasure and understanding of what it's all about piloting.
Raise your hand if you thought that an autotrim under manual flying was a necessity (Remember your time on the other types ...)

AF447 has been autotrimmed to the max up position under STALL WRN ... something is awfully wrong here !
At least, leave to the pilots that responsabillity if such was really their wish ...

BOAC

- John - why not come over to the Safety forum and join my thread - the water is lovely and warm

piperp2

Unfortunately I am not a pilot and this is my first post on this tragedy, but from my limited understanding of the trim function, it is to take the load of the control column or stick to reduce or maintain a balanced feel with no loads felt.

Obviously in the Airbus with the stick having no feedback on control surface loads, I therefore assume the trim is there to reduce hydro-mechanical /electro mechanical loads on the actuation systems during ascent, descent and level flight modes. If this is the case how would the pilot manual trim in alternate law?

What would be his indications to make the corrections? Why with the stick back would auto trim go to fully max with such a low horizontal speed (i.e. elevator control surface loading would have been at a minimum if the aircraft was at or below stall speed , does vertical speed have an influence here?)

BOAC

May I rewind to thread #1? As we approach 900 posts on this thread (4), we are really no further forward, 'experts' having 'dissected' the computer logic, argued with each other over what 'latches', 'defaults', 'degrades' or 'votes' etc etc - and we are all ignoring the basic puzzle.

Why did they climb so steeply? I still remain suspicious of the limited CVR release from BEA. I cannot accept that ANY pilot, out of line training, would sit idly by while the other pitches up to a large angle and the altimeter (and IVSI?) scream skywards. Why is this part missing? Let me make it clear, I do not wish to hear 'the voices', but where is the crew inter-action?

So, either PNF saw this climb or he didn't. If he saw it, what was said? You just do not zoom up that level at that weight in the ITCZ, if anywhere.

Now to the second option, he didn't. Way back on thread #1 I asked whether the ADIRU in this failure situation would degrade to IRU or lose all, and I was assured that attitude and inertial info would 'remain'. Suppose it did not.

What do we have?

We are expected to accept a silence in the cockpit while this manoeuvre takes place (and indeed much of the to and fro on these threads ignores this) following

PNF "so, we’ve lost the speeds" then "alternate law […]" followed by

PF "I don’t have any more indications", and the PNF said "we have
no valid indications".

None of this fits with Mr Occam.

If we are looking at the former option, huge and worrying questions on the philosophy of operation of this aircraft in general and AF training arise.

If the second...........

Instead of all the pilots here proving how clever they are at deciphering the IF....THEN and WHILE....WENDs why do they not address the basic PILOTING illogicality of all this? After the first 2 stall warnings, none further during the climb. PF makes 'nose-down' inputs, the climb continues but levels out. NO sign up to now of any more stall or overspeed, indeed no sign of the extant stall recovery actions of stick back and TOGA - until a new stall warning. So why climb? I just do not understand. Is it to be believed that two adult, trained pilots would react this way?

cwatters

It might be called "auto trim" but "trimming" seems to be only part of it's function. From descriptions here and in the reports it sounds more like a form of "power assisted steering". It sounds like when you call for a lot of up/down elevator it's the auto-trim that (eventually) delivers a large part of it?

With that in mind I'm unclear why the auto trim function is lost in direct law? Let me re-phrase that.. I can see why automatic control might have to disconnect but not why the pilot should be left with reduced maximium control throws and have to use manual trim to supliment them?

The available throw from the trim system seems to be very large? Under what conditions do you need so much up/down trim? Landing only?

or am I missing something?

DozyWannabe

But you can't infer from that information that the PNF was unaware of what was going on because he wasn't getting any feel through his sidestick. Perhaps he felt that the other pilot was doing what was necessary, perhaps he didn't feel confident enough to intervene (see Birgenair again), perhaps he got absorbed in the "PNF" role to the extent that he was more focused on the ECAM messages than his PFD or - and I know this will be controversial - perhaps he'd fallen asleep and was woken by the alarms. All of which are possible, but we don't know - we'll find out with the arrival of the report.

I disagree - saying that the system behaved as designed and that design is logical is categorically *not the same* as blaming the pilot. You're taking a technical aspect of the discussion and trying to make it emotional.

Also, confusion re: Law is (IMO) not the issue. At its most basic (and this is something I've been trying to get through to gums on as well), all you have to know about Law in a pressure situation like the one these guys were in is that you can't rely on hard protections once you're out of Normal Law - the very existence of the Alternate laws is to make sure that the aircraft handles much as it does in Normal Law, and we've had several posts from current FBW Airbus pilots stating that this is indeed the case.

As I said, I expect this aspect of the case will be causing a very thorough human factors investigation on the part of the BEA - if they didn't release it last month, I suspect it's because they're unsure of the significance of what (if anything) was said.

I think you'll find that it's been us techies trying to decipher the logic flow rather than the pilots "trying to look clever". And the fact is that occasionally pilots do illogical things. All this back-and-forth isn't really getting us anywhere and is simply giving those with an axe to grind an opportunity to hammer the square peg of this case into the round hole of their preconceptions.

(Also, you're definitely showing your age with the WHILE/WEND comment, as am I by acknowledging that I know what it means... (I haven't written any Pascal since I finished my A-Levels, and it was pretty old hat even then!))

Well - kind of but not really, especially from the late 1960s onwards. The cockpits may have *appeared* conventional but were in fact masking a complex fully-hydraulic system with no direct connection to the flight surfaces in the traditional sense, as some early DC-10 pilots and the crew of JAL123 sadly learned the hard way.

I know you know this by the way, just putting it down in the record.

And the aforementioned "loss of hydraulics" accidents in the third generation jets.