Habsheim

[Owain Glyndwr]

Another nice piece of disinformation. Who should I credit it to, OG or you ... ?

Let me first say that I strongly object to it being suggested that I would present disinformation!
Now let me repeat what I have said previously - the AOA was held down by the phugoid damping term in the Alphaprotect law. I said nothing about

waits a fraction of a second to stabilise before determining whether the remaining AoA is to be used for extra pitch or for bank input.

I think that is . The EFCS doesn't decide whether to add pitch (AOA?) or bank; the pilot does that, the EFCS merely decides whether it is safe to fulfill his commands. The system was designed to facilitate this by adding thrust if the energy management needed such thrust to maintain safety. By switching off alphafloor Asseline transferred that responsibility for safety to himself.
Apart from the fact that I have never seen an Airbus statement "advertising"* any specific value of alphamax, they would certainly not have made any claim that this could be safely achieved by simple action on the sidestick with only part of the system working.
Although Airbus test pilots routinely made high AOA demonstrations with alphafloor switched off they invariably compensated for that by selecting a level of thrust commensurate with maintaining speed and altitude.
When test flying to establish Vs1g, which must be done with idle thrust they were (are) careful to do this at a safe altitude so that the aircraft can be allowed to sink.

Edit:

* "advertising" as in writing for publicity or sales purposes intended for a general audience as against writing in a technical publication intended for knowledgeable persons

[DozyWannabe]

Sorry OG - *extremely* poor phrasing on my part (on top of an initial howler to start with - I'm a bit at sea when it comes to grasping the damping part). I did go back and try to correct it.

[Owain Glyndwr]

HN39

The graph shows the CG height based on the pressure altitude for both values of QNH.

There is also an unknown effect coming from any difference between static pressure position errors established in free air (up to alphamax) and those which would apply at the same AOAs in ground effect. So far as I am aware, no company has been zany enough to try to establish PEs near stall conditions at 30ft AGL!

[HazelNuts39]

Owain Glyndwr,

I agree with that. The last five seconds seem to indicate a change of PEC as the airplane gets nearer to the ground at increasing AoA.

However, in the preceding 20 seconds we see the opposite trend: the airplane seems to descend steeper based on pressure altitude than indicated by the radio altimeter + terrain profile. Do you have an explanation for that?

[Chris Scott]

Quote from roullishollandais:
What is still unknown is the altitude read on the altimeter on the cockpit

That is correct, strictly speaking, for two reasons:
(1) there is no video of the cockpit altimeters;
(2) we only have the testimony of the pilots that they had set their altimeter sub-scale settings to the transmitted QFE of 984.

However, as you know, it is a simple process to convert the DFDR readings of pressure altitude ("FINF") to the reading of an altimeter set to 984 hPa, simply by subtracting 808 ft.
For those who have not already done that, here are the readings at one-second intervals,
starting at TGEN 310 (t -24), and finishing at TGEN 334 (t -0):
126, 115, 111, 104, 101, 89, 85, 75, 70, 69, 68, 61, 60, 60, 60, 54, 54, 56, 53, 51, 52, 55, 56, 60, 56.

As I have stated in an earlier post, it appears that the a/c levelled off at 61 ft on the QFE at about t -13, gently lost indicated height from about t -10, and then recovered most of it by t -1. It is likely, however, that the accuracy of the readings available to the crew would have suffered from increasing position-error in the last seconds, because of the high AoA. Also, the static ports are between the cockpit and the CG, so the increasing pitch-attitude means the CG is increasingly lower than the static ports (whereas the pilot eye -height is increasingly higher).

[CONF iture]

Originally Posted by Dozy

If you say it was CONF 3 and the report says it was CONF 3, then it was CONF 3. I've been aware of that for a while.

... but never have acknowledged it since JAN 8th when you pretended otherwise.

Your argument seems to be that Airbus - to use your own word - "advertised" that in CONF 3, 17.5 degrees AoA would definitely be achieved under any circumstance with the application of full back-stick, when the documentation says no such thing.

I have not such argument - But if I had one it could be how the Airbus documentation describes the normal high AoA protection functioning but not when under given circumstances further additional restrictions apply.

There seems to be a definite short delay between acquiring 15 degrees AoA and going further.

Where do you see such definite short delay ... ?

How so? Please elaborate.

Under similar deceleration Bechet had obviously no delay to go straight to alpha max.

For one thing, you yourself said that Bechet's work in the simulator and real-world testing did not involve a precise reconstruction of the event - so I'm puzzled as to how they could disprove an assertion if that was the case.

Until you publicly acknowledge how the BEA work is questionable, I won't miss the opportunity to make you face their and your own contradictions ...

Originally Posted by Owain Glyndwr

Although Airbus test pilots routinely made high AOA demonstrations with alphafloor switched off they invariably compensated for that by selecting a level of thrust commensurate with maintaining speed and altitude.

... with maintaining altitude, speed not being dependent on the level of thrust.

[DozyWannabe]

Originally Posted by CONF iture

... but never have acknowledged it since JAN 8th when you pretended otherwise.

No, I was asking where you were getting the evidence for your claim that the demonstration flown by GC on the video was at CONF FULL - I never disputed the contents of the report that I can recall.

I have not such argument - But if I had one it could be how the Airbus documentation describes the normal high AoA protection functioning but not when under given circumstances further additional restrictions apply.

It's not a restriction, but a delay - thanks to a timely explanation offline (given slowly and carefully so that even I could understand), the damping effect goes some way to explaining that delay, and I won't embarrass myself by trying to reword it - OG and HN39 seem to have covered it.

Under similar deceleration Bechet had obviously no delay to go straight to alpha max.

On which graph do you see that? I must confess that I can't follow the hand annotations at the current resolution.

Until you publicly acknowledge how the BEA work is questionable, I won't miss the opportunity to make you face their and your own contradictions ...

Ultimatums - really?

Look, I'm not qualified to comment in any detail on the BEA report, but from a layman's perspective, and especially taking into account that this was the first public investigation of a commercial airliner crash that used digital technology in flight controls as well as data recording, the report seems to be at least fit for purpose and certainly as good as reports of the same vintage from other investigative authorities.

I'm sure there are some areas that could have been improved, but on the whole it seems solid.

What would you have had them do differently?

[Owain Glyndwr]

Originally posted by Confiture

speed not being dependent on the level of thrust.

An interesting interpretation of the laws of aerodynamics; but I was taught that to maintain constant speed in level flight (constant altitude) thrust has to be set equal to drag and since drag varies with speed it follows that speed and thrust are interlinked.

[HazelNuts39]

Originally Posted by CONF iture

Under similar deceleration Bechet had obviously no delay to go straight to alpha max.

I was under the expression that you had accepted the explanation I offered in post#605:

The difference is in pulling the sidestick back - Bechet at the same instant and Asseline 3 seconds later. So Bechet was still in pitch control law when he pulled the stick back while Asseline was in alpha-prot.

There was also a difference in the way the sidestick was pulled back - Bechet in one quick move and Asseline more hesitating.

[Owain Glyndwr]

Originally posted by HN39

However, in the preceding 20 seconds we see the opposite trend: the airplane seems to descend steeper based on pressure altitude than indicated by the radio altimeter + terrain profile. Do you have an explanation for that?

The short answer is no - or at least nothing really convincing, just a couple of thoughts.

There may be two effects with opposing trends involved; the effect of height at more or less constant AOA and the effect of AOA at constant height. This will be complicated by the fact that the AOA effect is very possibly more powerful at low heights than far off the ground. The motion seems to consist of one part with reducing height at more or less constant AOA and the other part increasing AOA at more or less constant height. OTOH, ground effect is rather weak above, say, 50ft so one would not expect any significant effect from the first trend over much of the relevant sequence.

[roulishollandais]

Classical vs digital design reliability

Originally Posted by Chris Scott

Originally Posted by roullishollandais

What is still unknown is the altitude read on the altimeter on the cockpit

That is correct, strictly speaking, for two reasons:
(1) there is no video of the cockpit altimeters;
(2) we only have the testimony of the pilots that they had set their altimeter sub-scale settings to the transmitted QFE of 984.

However, as you know, it is a simple process to convert the DFDR readings of pressure altitude ("FINF") to the reading of an altimeter set to 984 hPa, simply by subtracting 808 ft.k
For those who have not already done that, here are the readings at one-second intervals,
starting at TGEN 310 (t -24), and finishing at TGEN 334 (t -0):
126, 115, 111, 104, 101, 89, 85, 75, 70, 69, 68, 61, 60, 60, 60, 54, 54, 56, 53, 51, 52, 55, 56, 60, 56.

As I have stated in an earlier post, it appears that the a/c levelled off at 61 ft on the QFE at about t -13, gently lost indicated height from about t -10, and then recovered most of it by t -1. It is likely, however, that the accuracy of the readings available to the crew would have suffered from increasing position-error in the last seconds, because of the high AoA. Also, the static ports are between the cockpit and the CG, so the increasing pitch-attitude means the CG is increasingly lower than the static ports (whereas the pilot eye -height is increasingly higher).

I know all that Chris Scott ! I am able to do substractions... but please accept that we don't know how the pressure (total and static) are 'handled' by the software to appear as a graduated scale on the PFD on the glass cockpit.
AEROSPATIALE/AIRBUS INDUSTRY chosed to keep the A320 and next types software secret/private/hiden.
Not a post from Dozy Wanabee assessing that option as legitim, not a post from Iself assessing the opposite option that responsibility of the pilots need they may have access to ALL algorithms used on the aircraft.
Dozy thought it was in the sense of Open Source Software, free of copyright ; no, it is with protected copyright, as you may protect your rights when you are an author of published text or music that you would not hide !!!
What is different with digital design has been well pointed by the JACQUES-LOUIS LIONS report after the ARIANESPACE rocket ARIANE 501 crash. That crash and that report came much later than HABSHEIM, eight years later, but just before the HABSHEIM trial.
Since the beginning of the Airbus project Aérospatiale was growing on a paranoid way of writing software (ie the A320 simulator with the 'locked software' used to train the AF pilots) and some typical method failures were done by that community .
These many many method failures have been studied with a great accuracy (and diplomacy) in the report but read between the lines as a software professional I was in a highly challenging science ' the conclusions were just terrifying, and some sentences are absolute rules we just NEVER can jump over after that référence report.
Here is the link to the english traduction of the 12 pages report, followed by the most important sentences that no line of code may ignore.
Owain G already said me that Aérospatiale teams in the both projects were not the same, but from what I know from that history of software methods, both were involved in similar software design ... temerity. That temerity is a natural result of the fact that human brain does not work so logicaly as it seems, engineers -happily- incluses. They are many ways to do the altimeter figures and scale wrong via hiden software... Alas

ARIANE 5 Failure - Full Report

the Board wishes to point out that software is an expression of a highly detailed design and does not fail in the same sense as a mechanical system. Furthermore software is flexible and expressive and thus encourages highly demanding requirements, which in turn lead to complex implementations which are difficult to assess.

An underlying theme in the development of Ariane 5 is the bias towards the mitigation of random failure. The supplier of the SRI was only following the specification given to it, which stipulated that in the event of any detected exception the processor was to be stopped. The exception which occurred was not due to random failure but a design error. The exception was detected, but inappropriately handled because the view had been taken that software should be considered correct until it is shown to be at fault. The Board has reason to believe that this view is also accepted in other areas of Ariane 5 software design. The Board is in favour of the opposite view, that software should be assumed to be faulty until applying the currently accepted best practice methods can demonstrate that it is correct.

[CONF iture]

Originally Posted by Owain Glyndwr

An interesting interpretation of the laws of aerodynamics; but I was taught that to maintain constant speed in level flight (constant altitude) thrust has to be set equal to drag and since drag varies with speed it follows that speed and thrust are interlinked.

Not in the high AoA demonstration world to which you made reference where Airbus test pilots routinely fly alpha max and use thrust to control altitude but not speed ...

Originally Posted by Dozy

I never disputed the contents of the report that I can recall.

You did not dispute, you pretended the config for Habsheim was unknown.

No, I was asking where you were getting the evidence for your claim that the demonstration flown by GC on the video was at CONF FULL.

Already answered here and here.

It's not a restriction, but a delay

What a delay if not a restriction ?

On which graph do you see that?

Same graph mentioned by HN39 earlier.

I'm sure there are some areas that could have been improved, but on the whole it seems solid.

For someone who wanna stay blind, surely.

Originally Posted by NH39

I was under the expression that you had accepted the explanation I offered in post#605

If you're satisfied with it, then it shows how the concept is poor and inefficient.

[Owain Glyndwr]

Originally posted by Confiture

Not in the high AoA demonstration world to which you made reference where Airbus test pilots routinely fly alpha max and use thrust to control altitude but not speed ...

You are wrong there. They know that to make a correct demonstration of alphamax it is necessary to have enough thrust not only to maintain altitude but also to maintain speed. If that is not so the EFCS will apply a nose down corrective pitch command that will prevent achievement of alphamax.

That is why I wrote:

invariably compensated for that by selecting a level of thrust commensurate with maintaining speed and altitude.

[HazelNuts39]

Originally Posted by CONF iture

Not in the high AoA demonstration world to which you made reference where Airbus test pilots routinely fly alpha max and use thrust to control altitude but not speed ...

It depends. If you do the demonstration at low altitude you have to set the thrust that maintains level flight.

If you do the demonstration in free air there is no need to maintain altitude, but maintaining alphamax implies constant speed and pitch attitude. Vs1g is demonstrated with engines idle, and the sidestick held on the aft stop for 2 seconds or until there is no further increase of AoA.

So in both situations it is constant speed and attitude, therefore no phugoid damping.

[noske]

Originally Posted by HazelNuts39

The difference is in pulling the sidestick back - Bechet at the same instant and Asseline 3 seconds later. So Bechet was still in pitch control law when he pulled the stick back while Asseline was in alpha-prot.

There was also a difference in the way the sidestick was pulled back - Bechet in one quick move and Asseline more hesitating.

Asseline is well aware of this hesitation, and in his book (in his timeline of the accident in the first chapter) he volunteers an explanation (my translation):

"The mental image I had made for the fly by falls apart. Instinctively, in the fraction of a second, I move the thrust levers to idle (*), then to full thrust. I maintain level flight by visual reference. The forest approaches. Still no thrust. [...] I must not pull the stick all the way back too soon, for, if the thrust doesn't come, we'd only fall from a greater height."

(*) Because, as he recalls the events, he had already added some thrust six seconds earlier, when levelling off after the descent.

Nevertheless, Bechet would still have hit the trees if he had not been in a simulator.

Asseline should be relieved to hear that.

[CONF iture]

Originally Posted by Owain Glyndwr

You are wrong there.

Am I ?

• Thrust deficit you go down
• Thrust surplus you go up
• But alpha remains at alpha max
• And speed remains at Valphamax
• Only elevator controls speed through alpha
• Thrust has no say on speed

[Owain Glyndwr]

Originally posted by Confiture

Am I ?

Yes because if you generate a situation where speed decays whilst in alphaprotect and trying to fly at alphamax you will not get to or remain at alphamax so your equalities break down.

Only elevator controls speed through alpha

So when he applied up elevator at the end of the flight he was actually trying to slow down?

[CONF iture]

Originally Posted by Owain Glyndwr

Yes because if you generate a situation where speed decays whilst in alphaprotect and trying to fly at alphamax you will not get to or remain at alphamax so your equalities break down.

Now you come up with the "if" situations ...
So "if" the speed decays below Valphamax it would mean you're over alpha max and that's the elevator job to correct back to alpha max and its corresponding Valphamax, not the thrust job.

So when he applied up elevator at the end of the flight he was actually trying to slow down?

Unfortunately Asseline had not the power to apply up elevator, only to apply back stick.
Hopefully a cooperative FCS would have commanded the required up elevator necessary to increase alpha up to alpha max and decrease the speed to the corresponding Valphamax.

[Owain Glyndwr]

Now you come up with the "if" situations ...
So "if" the speed decays below Valphamax it would mean you're over alpha max and that's the elevator job to correct back to alpha max and its corresponding Valphamax, not the thrust job.

If you don't like if, change it to when.

As for the rest, you are distorting my words to suit your own purposes. I made no mention of Valphamax, those are your words. I made no mention of flight at more than alphamax, in fact I said that you could not achieve alphamax, so in my world the situation you describe could not happen. I could perhaps have been more specific, and stated that the restriction to alpha can occur at any AOA between alphaprot and alphamax. You choose to interpret it as being at and only at alphamax.

I made no mention of speed falling below Valphamax, I simply spoke of speed falling - again at any AOA between alphaprot and alphamax.

[Chris Scott]

Quotes from roulishollandais:
"I know all that Chris Scott ! I am able to do substractions..."

Forgive me... But your posts are usually so cryptic that it is sometimes difficult for a simple soul like me to estimate your experience and knowledge of aeronautical matters in general, and glass cockpits in particular. You talk of an MD80 pilot course, and then you say you are a software professional. If you read my posts, you know exactly my experience in the context of Habsheim, but I know little of yours. So I have no alternative but to take your questions on face-value, and answer them as well as I can. I have also to consider other readers who may have less knowledge of altimetry than you. (By the way, I admire your courage in posting on an English-language forum, because I would not want to try a French-language one.)

"...please accept that we don't know how the pressure (total and static) are 'handled' by the software to appear as a graduated scale on the PFD on the glass cockpit."

That is true, but how many pilots of a/c with mechanical or electro-mechanical altimeters know the precise mechanism that drives the needles, or the chances of them misreading? In my case, altimeters had already become more and more sophisticated as my career gradually progressed from the DC-3/C-47 to the A310 and DC10. Pilots have to take many things on trust, unless and until they find a problem. The A310 of 1983 had a PFD and ND, on which we relied for most of our flight parameters, except altitude and VS. So the DMC and other links in the display system were not unprecedented in airline service. I was flying A320s before this accident, and experienced no altimeter indication problems that I was aware of.

We know that Capt Asseline had had a problem on a previous flight, in a descent over the Jura mountains, when the selected sub-scale settings had changed spontaneously. That later resulted in the a/c being lower than he intended. But, as far as I know, there is no suggestion or likelihood that the same fault happened at Habsheim, and the pilots had both set the QFE of 984 less than a minute before they levelled off at 61 ft. Therefore, it is most likely that the indications on the PFDs were precisely 808 ft lower than the pressure altitudes recorded by the DFDR.