Habsheim

[Clandestino]

Any scenario is possible.Not so in the real life. Gravity, inertia, thermodynamics and aerodynamics set limits which are all too cheerfully ignored around here.

Methinks it was Einstein who observed that thought experiments are always successful.

[awblain]

C'ture… surely at Habsheim there was no speed to trade for height? And no power to allow a climb… The option to fly a Pugachev Cobra over the trees was not available.

Could a pair of A320 test pilots have wrestled it closer to stalling, missing the trees, if handed the accident aircraft without its fly by wire protections five seconds away from the tree line? I'm skeptical. How many tries would typically be required in a simulator to miss the obstacle once full power was requested?

The hardware and software seemed to have done a fairly good job of giving the crew the chance to save the day from a bad position.

Did the throttle refuse to comply with their demands? I'm not convinced. The conspiracy-inclined documentaries I've seen seem to be a bit light on discussions of sound travel time and speed reconstruction from the video frames. I've seen no reason to doubt the official report.

Did the elevators refuse to allow them to pitch up into a stall? I would contend that the answer is yes… as designed.

I think a fairer question is whether the fly-by-wire 320 and its competitor are been lost in accidents at a different rate. A320s are at about 24:10,000, while 737NGs are at about 9:3500. It's about equal, although A320s have rather longer in service to compare. FBW is clearly doing no great harm across the spectrum.

[roulishollandais]

Congratulation to Asseline if it was HE who managed that beautiful flare on the wood.
Did he?

[DozyWannabe]

What strikes me is how in Habsheim, New-York or Bilbao, the pitch authority has been compromised by features built into the system, whenever the aerodynamic had still more to deliver.

You don't know that - you're just assuming it based on this theory you've made from whole cloth based on your interpretation of Alpha Max as 17.5 degrees dead, when in the real world the laws of physics will cause variance.

[Chris Scott]

Having returned from three days' absence, and now catching up, I've got brief questions for a couple of you guys:

Confit,
1) You didn't pick up on my query of your alpha-max figure for Flaps/Config FULL (15 deg). Where did you find that? Why 2.5 deg lower than Flaps 3, and why would the alpha-max be only 0.5 deg above alpha-floor (14.5 deg) in the Flaps FULL case, instead of the apparent 2.5 deg margin with Flaps 3 (17.5 deg and 15 deg respectively)?
Remember, this was an A320-100 in 1988.
2) By the way, what config do you think was set in the Gordon Corps video?

rudrudrat,
What is the source of info for your assertion that the stall test for certification was done in Pitch Alternate Law?

[CONF iture]

You didn't pick up on my query of your alpha-max figure for Flaps/Config FULL (15 deg).
Actually I did here (http://www.pprune.org/8247226-post235.html).
Also valid for your second question.

Why 2.5 deg lower than Flaps 3, and why would the alpha-max be only 0.5 deg above alpha-floor (14.5 deg) in the Flaps FULL case, instead of the apparent 2.5 deg margin with Flaps 3 (17.5 deg and 15 deg respectively)?
Good question, but it would be one for Toulouse.
My guess would be that, as CONF FULL is the normal configuration for landing, Airbus was not too excited to allow 17.5 deg of AoA that close from the ground with a higher risk for catastrophic tail strike ...

[CONF iture]

One question for OK465 maybe, as you seem to give some simulator training.
What is the capacity for a modern flight simulator to register a variety of data such as the AoA, attitude, flight command inputs, flight control movements ... something similar to a FDR or QAR ?

[C_Star]

Why 2.5 deg lower than Flaps 3, and why would the alpha-max be only 0.5 deg above alpha-floor (14.5 deg) in the Flaps FULL case, instead of the apparent 2.5 deg margin with Flaps 3 (17.5 deg and 15 deg respectively)?


I seem to remember from aerodynamics classes that higher flap settings yield lower critical AoA (but of course higher CL values) - that could explain why Conf Full has lower alphamax than Conf3.

Then, there's higher slat deflection for Conf Full, which should increase the stall AoA, but maybe the flaps effect offsets the slat effect?:confused:

[Chris Scott]

Quote from me, addressed to CONF_iture:
You didn't pick up on my query of your alpha-max figure for Flaps/Config FULL (15 deg). Where did you find that?

CONF_iture's reply:
Actually I did here (http://www.pprune.org/8247226-post235.html).

No: you didn't! Your post merely restates the figures. Please state a specific reference for the Flaps FULL figure, if you have one.

Quote from C_Star:
I seem to remember from aerodynamics classes that higher flap settings yield lower critical AoA (but of course higher CL values) - that could explain why Conf Full has lower alphamax than Conf3.
Then, there's higher slat deflection for Conf Full, which should increase the stall AoA, but maybe the flaps effect offsets the slat effect?

That seems to be born out to some extent by the figures of alpha-floor in my BCAL/BA Tech Manual of 17FEB1988, which I quoted in a previous post (http://www.pprune.org/tech-log/528034-habsheim-11.html#post8244380) four days ago:
"Alpha-floor is activated when:
- alpha > alpha-floor (9.5 deg in conf 0; 15 deg in conf 1, 2, 3; 14.5 deg in conf FULL, or [...]"

But, as you see, the difference is only half a degree. Confit's figure suggests:
(1) the difference is 2.5 deg;
(2) alpha-max with Flaps FULL was only 0.5 deg above alpha-floor, which also seems improbable.

Unlike the Airbus FCOM extracts I quoted in the same post, those figures from the BCAL/BA Technical Manual (Ch 09 Section 10) - 17FEB1988 revision - were still current at the A320's entry into service in April 1988, so may be the most reliable information available (so far) for the Habsheim a/c. They appear at the foot of the familiar curve of Cz-versus-alpha.

The equivalent page on the preceding version, dated 25AUG1987, is a duplicate of the Airbus FCOM Flight Controls chapter (1.09.10 P8, REV 03, SEQ 001) that I also quoted in that post. As you can see, there are no specific values of alpha for alpha-floor or the other two protection levels. The figures provided define each of the 3 protection levels simply as the alphas associated with factors of Vs in any weight/configuration at low Mach, but neglect to define Vs itself. It may be that, with certification still pending, the definition of Vs was still under discussion. (The A320 was type-certificated in February 1988.)

[CONF iture]

No: you didn't! Your post merely restates the figures. Please state a specific reference for the Flaps FULL figure, if you have one.
At time 1035 in the video, the Airbus test pilot clearly states :
"We're close to 15 deg AoA which is the maximum for this configuration"
I take it as a specific reference, why should I doubt such statement from a guy who seems pretty aware on what's going on ... ?

[noske]

If not yet read .. you can read this Airbus report
"The A320 Habcheim Accident
An Airbus Industrie response to allegations made in television programme and other medias"
Rapport Airbus.pdf - Petit Fichier (http://www.petit-fichier.fr/2013/12/18/rapport-airbus)

Thanks, I hadn't seen that before.

Btw., for anyone frustrated with the poor quality PDF of the BEA report, here's a HTML version of it: Habsheim F-GFKC (http://wnoeker.de/kc/)

[roulishollandais]

Thank you Noske, HTML and J.O.

[rudderrudderrat]

Hi Chris,
What is the source of info for your assertion that the stall test for certification was done in Pitch Alternate Law?
I can't find a document to support it, except on page 17 of http://lessonslearned.faa.gov/IndianAir605/A320%20Final%20Special%20Condition.pdf
"Several commenters note that the new VS1 usage has not been uniformly applied throughout all Part 25 subparts. As configured for operation, the A320 is prevented from stalling by the incorporation of an angle-of-attack limiting feature. This feature would then not allow demonstration of the stall speeds used for structural design. During development flight testing, Airbus deactivated the stall protection feature and demonstrated compliance with the existing rules as currently interpreted both in the clean and flaps down configurations. Therefore, the special condition was not proposed for Subparts C&D."

Besides forcing the aircraft into ALT Law, how else do you think they deactivated the stall protection feature?

[Chris Scott]

Quote from CONF_iture:
At time 1035 in the video, the Airbus test pilot clearly states :
"We're close to 15 deg AoA which is the maximum for this configuration"

Thanks for reminding me. I assumed you had a paper reference.

Interpreting GC's introductory speech, the video seems to date from mid-1989. I agree with you that, in 1988/9, the standard landing configuration was Flaps/Config FULL, even single-engined (subject to its satisfying the WAT figure for approach-climb performance). So the only issue is whether GC's figure of 15 deg was a reference to alpha-max. Taken on face value, it would appear to be so, but - as rudderrudderrat and I have previously discussed here - the AoA indicator later exceeds 15 for some time while GC is demonstrating the ability of the a/c to bank with full back-stick.

During the deceleration, alpha-floor appears on the FMA as GC points out an AoA of +14 (time 9:54), which is close to my Tech Manual figure of 14.5, and is quickly cancelled. The AoA seems to stabilise at +15 while the wings are level, but quickly increases to a peak of nearly 17 deg when the a/c is rolled to about 25 deg of left bank. This increase is quickly acknowledged by GC, who gestures (10:53) at the AoA indicator with his pointer, saying "... slightly increased angle of attack, it's gone to sixteen degrees... maintain that bank angle..." without any indication of surprise in his intonation.

So, despite having stated previously that 15 degrees AoA was the "...maximum for this configuration", he was not surprised or concerned that it exceeded that figure as soon as bank was applied. Unless we can find evidence that alpha-max is merely a target maximum for the EFCS when the stick is fully back, rather than an absolute maximum, I think we need to treat with caution the 15 deg figure of alpha-max for Config FULL.

Is it possible that the EFCS might retain a margin of a couple of degrees from alpha-max when the wings are level, in order to retain manoeuvre capability? Such a margin from alpha-max would cater for the case where the pilot needs to turn suddenly and unexpectedly after alpha-max has already been reached, in which case allowing bank with an associated increase in Nz (normal acceleration) might lead to the stall. If such a margin was employed, that might explain the apparent shortfall of AoA in the (wings-level) Habsheim case.

[Owain Glyndwr]

rrr

Besides forcing the aircraft into ALT Law, how else do you think they deactivated the stall protection feature?

It is not unusual for the software used on flight test airplanes to be more flexible than the production version. This is to allow gains etc to be tweaked easily for test purposes. So it may be that they simply inhibited the incidence protection software for stall tests.

[HazelNuts39]

The question is not whether the airplane was tested beyond the arbitrarily chosen alpha-max incorporated in the FCS software in normal law. Of course it was flight-tested beyond alpha-max, either in alternate law, direct law or with some 'experimental' development version of the FCS software as suggested by OwainG.

The surprise for me is that apparently the special conditions under which the A320 was certified specify operational speeds such as V2 and Vref referenced to a Vs1g that can not be demonstrated with its systems operating as designed, i.e. in normal law. In other words, alpha-max is not determining Vs1g as it would have been under later regulation, but was selected somewhat independently.

[Chris Scott]

Hi rudderrudderrat,

Thanks for that very interesting link, and I note Owain Glyndwr's quick response that reversion to Pitch Alternate may not have been necessary during the certification flights. Experience suggests he is usually right, in which case I infer that Normal Law and Alpha-Protection would have been retained, but alpha-max modified to a figure at or above Cz-max. Alpha-Floor would also be inhibited. If so, this would have the advantage that THS (pitch) up-trimming would cease at alpha-prot. That would make the exercise much safer than using Pitch Alternate, in which - as you have pointed out - the THS would continue trimming to the stall and beyond (as it did later on AF447).

I am told that there have traditionally been three alternative scenarios for establishing figures of Vs during the certification of aeroplanes. In each case the a/c is flown straight and level (Nz = 1.0g) for as long as possible with a deceleration of 1 kt/sec. The following are very rough, layman's descriptions for OG, HN39 and others to correct/amend/expand:
(1) If full back stick/column is reached before the stall, Vs is defined as the lowest speed achieved. (Normally, that definition applies only if elevator authority is insufficient to achieve a classic stall. However, it would have applied to the A320 in the Alpha-Protection mode of Normal Law for artificial reasons.)
(2) The a/c is maintained at 1g until the stall is denoted by a marked "break" (nose-drop), at which point the speed is nominated as the Vs. (This was the traditional British method for BCAR certification.)
(3) As for (2), but the a/c was allowed to execute a mild bunt, probably starting momentarily before the 1-g break, with continuing deceleration, until the Nz fell to a certain level. At that point the resulting speed was nominated as the Vs, sometimes described as the "minimum speed in the stall", provided it was not more than a certain percentage below the Vs1g. (This was the traditional American method for FAR certification.)

The FAR method (3), with its lower values of Vs, gave American a/c a commercial advantage over British a/c when field performance was being considered, because the different regulators applied the same factors to Vs for the calculations of V2 (1.2 Vs) and Vref (1.3 Vs). I always wondered why - on a calm day when the numbers were right - one could confidently close the throttles at 50 ft on a VC10 or BAC1-11, but wouldn't even consider it on a B707.

Sometime in the 1980s, there is said to have been a levelling of the playing field by agreement, but I notice that my stall graphs for the A310 - the last Airbus type to be certificated before the A320 - are headed "VS MINIMUM". So did the DGAC use FAR criteria for the A310 certification?

So to the A320. The version of the cryptic, ubiquitous curve of CL-versus-alpha that you posted (http://www.pprune.org/tech-log/528034-habsheim-11.html#post8244247) 4 days ago is annotated "1G stall" over the apogee associated with CL max (unlike the versions of Cz-versus-alpha in the early FCOMs, which omit it). Although John Farley pointed out (on seeing it earlier) that the caption was redundant and potentially misleading, AI was presumably trying to send a message that their Vs is taken (a) at CLmax and (b) at an Nz of 1G, not a lower figure.

[HazelNuts39]

Just for accuracy, a few comments in response to Chris Scott's post #270):

Regarding Chris' stall condition (1), I don't think it is such a 'rarity' even for conventional airplanes.
Regarding Chris' (2) I am somewhat sceptical but can't comment because I wasn't involved in BCAR certifications in those days. However, strictly speaking the so called g-break speed is close to but not equal to the 1-g stall speed, and both are not obtained at 1-g in the stall maneuver. The 1-g stall speed is the lowest speed at which wing lift alone can support the weight of the airplane in 1-g flight. It is defined by the point in the stall maneuver at which the lift coefficient cL is at a maximum. At that point the normal acceleration is usually not equal to 1, and the reference stall speed is then calculated as Vs1g = V(cLmax) / √(nz)
Regarding Chris' (3) I would observe that, although the minimum speed in the stall maneuver is usually at less than 1 g, it is not obtained at 0.94 g, nor at √0.94 g.

[Chris Scott]

Hello HN39,
Thanks for the useful corrections and clarifications to my attempted layman's guide to stall certification. particularly for pointing out my mistake in condition (3), in which my figure of minimum Nz appears to have been incorrect. I have amended my post, while noting that you will remain not entirely in agreement with it in principle.
Thanks also for reminding me that the British ARB's equivalent of FARs (and since replaced by European JARs), were the BCARs (not BARs, as I stated).

[DozyWannabe]

The surprise for me is that apparently the special conditions under which the A320 was certified specify operational speeds such as V2 and Vref referenced to a Vs1g that can not be demonstrated with its systems operating as designed, i.e. in normal law.

If my recollection is correct, the systems were designed in such a way that each layer of control "laws" was implemented as a "wrapper" around the control law below it.

What this means is that the Direct Law implementation (where PFC commands are directly translated to the surfaces) serves as the operational "core" of the EFCS. What came to be known as the Alternate Laws (C* functions and soft protections) were built on top of the Direct Law core, and Normal Law (hard protections and tweaked rate functions) built on top of those.

As OG hints at, the software used in development, which would have occurred both on the A300 testbed and A320 prototypes*, would likely have had logical "hooks" linked to switches on the test consoles and possibly in the flight deck allowing for easy transfer between the different laws, modes and the two distinct implementations - not just for testing and certification purposes, but also as a failsafe safety measure should there be a significant software problem with the 'higher' level implementations during testing. The software used in production would be identical and still have those logical hooks - but would have nothing connected to them on the production aircraft.

* - Along with the A330/340/380/350 prototypes later on.

[CONF iture]

Is it possible that the EFCS might retain a margin of a couple of degrees from alpha-max when the wings are level, in order to retain manoeuvre capability?
The margin is already in alpha max and that is why alpha max has not been set to alpha stall and V alpha max is not Vs1g.

A very temporary excursion over alpha max is acceptable and is part of the protection mode. The system will not try to achieve anything more than alpha max but it will tolerate an overshoot if turbulence or sudden turn happen when already established at alpha max and of course the system will correct to bring back the AoA where it should be, its alpha max value.

[roulishollandais]

Thanks for that very interesting link, and I note Owain Glyndwr's quick response that reversion to Pitch Alternate may not have been necessary during the certification flights. Experience suggests he is usually right, in which case I infer that Normal Law and Alpha-Protection would have been retained, but alpha-max modified to a figure at or above Cz-max. Alpha-Floor would also be inhibited. If so, this would have the advantage that THS (pitch) up-trimming would cease at alpha-prot. That would make the exercise much safer than using Pitch Alternate, in which - as you have pointed out - the THS would continue trimming to the stall and beyond (as it did later on AF447).To avoid a new AF447 and his autotrim problems, inhibit Alpha-Floor, and modify alpha-max at or above Cz-max...:p prototypes*, would likely have had logical "hooks" linked to switches on the test consoles and possibly in the flight deck allowing for easy transfer between the different laws, modes Why not put these switches on the consoles again to give observability and controllability :) to the "ordinary" pilots (and gums who is not an ordinary pilot !) ?

[DozyWannabe]

The margin is already in alpha max and that is why alpha max has not been set to alpha stall and V alpha max is not Vs1g.

A very temporary excursion over alpha max is acceptable and is part of the protection mode. The system will not try to achieve anything more than alpha max but it will tolerate an overshoot if turbulence or sudden turn happen when already established at alpha max...

Honest conjecture - please don't be offended, but that's your interpretation and yours only.

You mention the Airbus demonstrator saying/showing "the airplane will rapidly reach alpha max", but I bet the demonstrator said nothing about it reaching a precise 17.5 degrees of AoA with the same urgency. The manuals, both FCOM and FCTM, seem to shy away from giving precise quantitative values for limit and rate when describing the procedure and functions, and this makes sense because there will be occasions where to do so would be inadvisable (e.g. adverse weather and/or a decelerating aircraft).

I mentioned this before, but I'll point out again that the difference between Bilbao (where phugoid damping was an issue) and Habsheim (where it most likely wasn't) was a gusting tailwind causing a sudden (approx. 10kt in under a second) drop in IAS in the former case.

The demonstration video by Capt. Corps would likely have been seen by Capt. Asseline, and in all probability he'd have sat in on a similar demonstration using the prototype (with the AoA gauge) - therefore he'd likely have been aware that immediate response to full back-stick in the landing configurations with protection active would have been in the 15-16 degree range.

The absolute Alpha Max value of 17.5 degrees was mentioned (albeit obliquely) in the BEA report to which both Asseline and his legal team would have had access prior to the court cases. Given the myriad ways in which they tried to draw attention to Airbus and possible technical issues, one would think that the apparent discrepancy between that absolute value and the AoA achieved would have made a cornerstone of their case, but it wasn't even touched upon. The only reasons I can think of that this would be the case are either that they missed it (IMO unlikely), or that they knew such behaviour was normal and therefore not a technical fault or design flaw.

[EDIT

To avoid a new AF447 and his autotrim problems, inhibit Alpha-Floor, and modify alpha-max at or above Cz-max...:p
AF447 was in Alternate Law (2b) which does not have Alpha Floor (or for that matter any hard alpha protections) available. The autotrim behaviour was not a "problem" in the purely technical sense - it was simply trying to comply with the demands of the pilot.

Why not put these switches on the consoles again to give observability and controllability :) to the "ordinary" pilots?

Because the automatic switching mechanism has been proven sufficient and safe during over 2 and a half decades of operation. Also, apropos of nothing, the B777's bypass mode exists as we know, but in the very few incidents where it might have been useful (I'm thinking in particular of MH124), none of the crews used it because they were too busy trying to control the aircraft.

PS. Gums is certainly not an ordinary pilot - I'd love to see him get a go in the A320 prototype one day to help him understand how it all fits together and why.

]

[AlphaZuluRomeo]

CONF, (re: #274) agreed for turbulences, but we were discussing turns (which are not so "very temporary").

Dozy, as an attempt to illustrate your #273, here are two pictures of A400M cockpits :
- test aircraft: Photos: Airbus A400M Grizzly Aircraft Pictures | Airliners.net (http://www.airliners.net/photo/Airbus-Industrie/Airbus-A400M-Grizzly/2159557/L/)
- production aircraft: Photos: Airbus A400M Grizzly Aircraft Pictures | Airliners.net (http://www.airliners.net/photo/France---Air/Airbus-A400M-Grizzly/2275423/L/)
On the test aircraft, between the standby ADI and the gear indicator, there are two guarded switches. The right one is labelled "FLT CTL DIR-LAW", with 2 positions: DIR and NORM.
Thoses switches are absent from the production aircraft, the second delivered to the French Air Force.

[DozyWannabe]

Mine too.

...

From a practical implementation standpoint, without an AOA gauge, the pilot sees only that he has achieved and is able to hold Valphamax. He is then a happy camper.

Agreed, and I'm with you. The difference between your input (as I understand it) and CONF iture's is that he seems to interpret the term "alpha max" as adhering strictly to 17.5 degrees AoA and believes anything other than a swift acquisition and maintaining of that value to be a technical fault or anomaly.

[CONF iture]

Honest conjecture - please don't be offended, but that's your interpretation and yours only.
It appears that My interpretation and His demonstration match pretty well ...

[DozyWannabe]

It appears that My interpretation and His demonstration match pretty well ...

If so, at what point in the video does the AoA gauge hit and stay at 17.5 degrees precisely?

[CONF iture]

If so, at what point in the video does the AoA gauge hit and stay at 17.5 degrees precisely?
Why would it do that when the alpha max value for the demo was 15 ... ?

[DozyWannabe]

Why would it do that when the alpha max value for the demo was 15 ... ?

Because that value is briefly exceeded, and right now we only have your supposition that the immediate alpha max value at Habsheim should have been 17.5 degrees due to CONF setting...

Would you be willing to at least consider the possibility that there are many valid reasons why the aircraft won't attain 17.5deg in a hurry?

[CONF iture]

Because that value is briefly exceeded
Yes, alpha max has been exceeded, briefly ... and yet no stall no crash ...
The problem is ... ?

and right now we only have your supposition that the immediate alpha max value at Habsheim should have been 17.5 degrees due to CONF setting...
That alpha max was at 17.5 deg for Habsheim is a certitude.
Regarding the 'supposition' it is Yours not Mine ... mine is still here (http://www.pprune.org/8242443-post194.html).

[DozyWannabe]

That alpha max was at 17.5 deg for Habsheim is a certitude.

But the claim that it should have been reached immediately in that scenario is not.

You still have provided no evidence that the difference in Alpha Max is down to the CONF setting.

[CONF iture]

But the claim that it should have been reached immediately in that scenario is not.
Where is such claim ... ?

You still have provided no evidence that the difference in Alpha Max is down to the CONF setting.
The words of your Hero is not enough ?

[DozyWannabe]

Where is such claim ... ?

That's your claim - that the system prevented optimum performance by not getting to 17.5deg AoA quickly enough, is it not?

The words of your Hero is not enough ?

He didn't specify CONF 3 or CONF FULL for the demo - the assumption that there was a difference in the demo config versus that of AF296 has only come from you.

[CONF iture]

That's your claim - that the system prevented optimum performance by not getting to 17.5deg AoA quickly enough, is it not?
It is not - It is more specific - It still did not change (http://www.pprune.org/8242443-post194.html).

He didn't specify CONF 3 or CONF FULL for the demo - the assumption that there was a difference in the demo config versus that of AF296 has only come from you.
If CONF 3 was the setting for the video demo, why alpha max was at 15 when it was at 17.5 for Habsheim ... ?

[DozyWannabe]

It is not - It is more specific - It still did not change (http://www.pprune.org/8242443-post194.html).

OK, so let's have a look:

It certainly won't efficiently as the elevators are commanded nose-down as a start.
And if the elevators are commanded nose-down to arrest the pitch-up tendency from the engines, as submitted by dozy, it is even a poorer concept.
Remember, we are still 2.5 deg short of alpha max.

Elevators don't effect an AoA, they effect a pitch attitude.

Therefore, if the pitch attitude is limited by AoA protection and the pitch-up moment from the increase in thrust is expected to exceed that attitude, then the systems will be correct in counteracting that tendency through the elevators. As Capt. Bechet stated, this is normal behaviour, and something that Capt. Asseline should have been aware of.

Nothing in the documentation or the video states that 17.5 degrees AoA is guaranteed upon application of full back-stick. The demonstration in the video was not performed with the engines spooled down, and that would have made a massive difference to systems behaviour.


It does not have to be, what matters is the intention, the intention to efficiently deliver.

Which, in the circumstances, the systems could be argued to have done just that.

The systems were designed to keep the aircraft flying and stable even at attitudes which would be considered extreme for a transport aircraft. I'd be prepared to bet money that the combination of Asseline's inputs, had the aircraft been in Direct Law, plus the pitch-up moment from the thrust increase would almost certainly have stalled the aircraft. I'd also be prepared to bet that no human pilot would have been able to take in all the flight parameters and keep the aircraft more stable than the EFCS did.

but a contributory factor for the crash it has to be.

Can you prove that the extra 2.5 degrees of pitch would have avoided the crash? I know I certainly wouldn't be able to.

If CONF 3 was the setting for the video demo, why alpha max was at 15 when it was at 17.5 for Habsheim ... ?

We don't know what the CONF setting was for either.

[Pugilistic Animus]

DW, the elevators most definitely control AoA...along withthust setting obviously exempting gliders

[DozyWannabe]

DW, the elevators most definitely control AoA...along withthust setting obviously exempting gliders

Correct - perhaps I should have said that elevators *alone* do not effect an AoA.

The point I've been trying to make is that when the aircraft is decelerating and thrust is not yet sufficient either to arrest the deceleration or power out of the projected flightpath, then either the pilot or the EFCS must take that into account when trying to keep the aircraft aloft. Commanding a pitch attitude equivalent to 15.4 degrees of AoA at timestamp X may well be the equivalent of 17.5 degrees AoA or thereabouts at timestamp X+1s.

[Machinbird]

I haven't read the complete accident report and only wish to offer a quick comment regarding this accident. What I see gives some credence to Confiture’s point of view. For common reference, I'll provide a link to a particular video of the incident and mention some key events. Crash Airbus A320-Vol 296 Air France - YouTube (http://youtu.be/Yk-Hy83k2Nk).

The pilot had two problems. He needed to accelerate his engines and needed to clear the trees. Jet pilots cannot forget the spool up time of their engines but at Habsheim that day, use of automation seems to have gotten in the way and created a surprise and a delay in adding power.

After rollout on heading, there are two small pitch adjustments visible. One at T=11 sec, and one at T=14 sec. These are possibly pulls on the stick intended to trigger Alpha Floor (which was not available).
The aircraft began to impact trees around T=26, or about 12 seconds after the second pitch adjustment.
If the pilot had been able to activate Alpha Floor at the second nose twitch, it appears (to me) that he would have cleared the trees.

But let’s look at the problem a little differently, from the viewpoint of what a small increment of g acceleration would have done for the flight path.

If .01 additional g was available at the second nose twitch and for the next 12 seconds, the aircraft cg would be about 23 feet higher-probably not enough to avoid an accident.
If .02 additional g was available at the second nose twitch for the next 12 seconds, the aircraft cg would be about 46 feet higher and would probably have escaped since the engines were spooling up nicely by then. (actually slightly less high since he would have converted some of his kinetic energy to potential energy and it probably would not be possible to maintain all the additional g for the duration.)

Perhaps you suspect that even that that small amount of additional g was not available, but consider that the video shows an alignment turn seconds before the nose twitches up. Was he decelerating that fast that there was absolutely nothing left to maneuver at the end?

In other words, if a measly .02 additional g had been available, the Habsheim event likely would not have been an accident. Would an additional.02 g have been available if the maneuver was flown in direct law and the pilot had pulled just slightly past optimum AOA and then eased off to optimum? That is definitely faster than an asymptotic approach to applying g and is the way most experienced manual pilots would lay on rapid g.

Looking at the Gordon Corps video, it seems there might well be some additional g potential/ AOA in the A320 Normal law being held back as a “safety margin”.

[DozyWannabe]

...but at Habsheim that day, use of automation seems to have gotten in the way and created a surprise and a delay in adding power.

How so? Bearing in mind they were not using automation at any point during the approach or flypast, nor were they intending to.

If the pilot had been able to activate Alpha Floor at the second nose twitch, it appears (to me) that he would have cleared the trees.

The flight plan included complete disabling of A/THR (and thus A. Floor) - that this step of the procedure was forgotten in the event doesn't alter that.

That is definitely faster than an asymptotic approach to applying g and is the way most experienced manual pilots would lay on rapid g.

While taking into account deceleration and the pitch-up movement from slamming thrust back on? I very much doubt a human pilot could have done better.

[Machinbird]

Originally Posted by Machinbird http://images.ibsrv.net/ibsrv/res/src:www.pprune.org/get/images/buttons/viewpost.gif (http://www.pprune.org/tech-log/528034-habsheim-post8254491.html#post8254491)
...but at Habsheim that day, use of automation seems to have gotten in the way and created a surprise and a delay in adding power.
How so? Bearing in mind they were not using automation at any point during the approach or flypast, nor were they intending to.I think I need to clarify that there was apparently a mental 'expectation' that Alpha Floor was going to kick in, thus even though he may have forgotten his specific autothrottle configuration, the expectation that it would operate caused confusion and a delay in actually adding power.

The flight plan included complete disabling of A/THR (and thus A. Floor) - that this step of the procedure was forgotten in the event doesn't alter that. The point is that his point of attempted Alpha Floor activation was not entirely unreasonable from a maneuvering standpoint. It completely failed because he failed to allow for the unexpected non-response.
When maneuvering near the ground, there really isn't much time for goof ups and when doing atypical flying, there is a need for an exceptional level of planning and practice. That was not done in preparation for the Habsheim flight.

While taking into account deceleration and the pitch-up movement from slamming thrust back on? I very much doubt a human pilot could have done better. Dozy, I'm going to have to tell you to check your 6 'cause you are dealing with a fighter pilot who could lay on extreme g during violent maneuvers without departing the aircraft:}.

Consider basic servo theory for how quickly the output can match a step input. Remember that a human is a variable gain computer, and although not error free, can error correct once trained in error correction.

A computer must use conservative control strategies since it is supposed to operate independently, thus the asymptotic (essentially an overdamped servo approach) control of AOA.
A human can use a variable gain underdamped approach to a step input and even though they may overshoot the ideal AOA they will correct and arrive at the desired AOA sooner, particularly if there is some sort of feedback mechanism for them to observe (such as mild airflow separation causing a slight vibration.) The key is training and practice.

[CONF iture]

Therefore, if the pitch attitude is limited by AoA protection and the pitch-up moment from the increase in thrust is expected to exceed that attitude, then the systems will be correct in counteracting that tendency through the elevators.
If thrust alone is able to pitch up the airplane, why the system should waste such a potential when 2.5 deg are still to be grabbed ?

Can you prove that the extra 2.5 degrees of pitch would have avoided the crash?
No and I have no desire to, but I certainly can prove that preventing the aircraft from getting the available 2.5 deg of AoA made the crash inevitable.

We don't know what the CONF setting was for either.
What about you read the Habsheim report ... ?

You're getting some signs of confusion now. Maybe time to take a rest.

[Goldenrivett]

Originally Posted by Machinbird
I think I need to clarify that there was apparently a mental 'expectation' that Alpha Floor was going to kick in, thus even though he may have forgotten his specific autothrottle configuration, the expectation that it would operate caused confusion and a delay in actually adding power.

Page 35 of the French report:
At 12h 46' 26'' (Crash - 13 seconds)
CDB "Bon Je vais bien la debrayer l'auto-manette."
(Google translate = "Good I'm fine to disengage the auto-throttle").
Radio Altimeter shows 40 / 50 / 40 feet.

I agree. It also seems to me as though he was going to "demonstrate" how clever the aircraft was and that Alpha Floor was going to prevent a crash, despite the thrust levers being closed.
It would explain the crews insistence that they didn't deliberately fly below 100 radio and did not hear the radio call outs because they had both ears covered by their headsets. (From a crash investigator TV program)

[noske]

I think I need to clarify that there was apparently a mental 'expectation' that Alpha Floor was going to kick in, thus even though he may have forgotten his specific autothrottle configuration, the expectation that it would operate caused confusion and a delay in actually adding power.


This seems to be a good moment to mention Asseline's book. :E It has his own annotated version of the final 60 seconds of the CVR transcript, how he himself remembers the accident flight. Although he seems to be wrong about at least one detail, it may still give an idea about what he was planning to do and what he was expecting. Mentionable items:


When PNF said "Bon, t'arrives à 100 pieds là fais, fais", this was well above 100 ft, and Asseline is certain that after he levelled off, his pressure altimeter indicated 100 ft exactly.
For the fly by, he was mainly focused on keeping the vertical speed at zero.
In the moment when he said "Bon, j'y vais bien là, débrayer l'auto-manette", he pushed the auto-thrust disconnect buttons and advanced the thrust levers to some intermediate thrust setting. (Here the DFDR disagrees: No movement of the thrust levers yet, both still at idle.)
His original plan for the fly by was to continuously adjust thrust to maintain his intended speed and to apply go-around thrust when passing the control tower. But since his speed was still 160 kts after levelling off, there was initially no need to add thrust.
Just before passing the control tower, he becomes aware that the engines haven't spooled up from idle. Also he realizes that what looked like just some vegetation of a darker shade of green at the edge of the airstrip is actually a forest.
This is when he pulled the thrust levers all the way back and pushed them all the way forward again. (The CVR records this as three clacks of the levers moving across their detents.)

Paraphrased from pp. 5-8 in the book: http://www.crashdehabsheim.net/livre%20le%20pilote%20est%20il%20coupable.pdf . Annotations in () are mine.

What I make of this:


If you're surprised by a forest in your intended flight path, there's something seriously wrong with your flight preparation. Duh.
It looks like Asseline made one serious omission (meant to adjust thrust right after levelling off, but didn't), but could never admit it.

[Machinbird]

This is when he pulled the thrust levers all the way back and pushed them all the way forward again. (The CVR records this as three clacks of the levers moving across their detents.)Why would he pull the throttles all the way back? I'm not an Airbus guy so I'm curious as to the reasoning.

Jet engines decelerate very rapidly, and if there was any RPM above idle prior, that was undoubtedly lost and that would cause/contribute to a long engine accelerate time.

Time to get the engines spooled up was one of the critical problems on that flight.

[DozyWannabe]

I think I need to clarify that there was apparently a mental 'expectation' that Alpha Floor was going to kick in, thus even though he may have forgotten his specific autothrottle configuration, the expectation that it would operate caused confusion and a delay in actually adding power.

But disabling A/THR (and thus Alpha Floor) was central to the entirety of the attempted demonstration - there would have been no expectation of Alpha Floor operating.

Dozy, I'm going to have to tell you to check your 6 'cause you are dealing with a fighter pilot who could lay on extreme g during violent maneuvers without departing the aircraft:}.

Understood, sir. But we're not talking about a fighter here, we're talking about a jetliner which, as you know, has a considerably lesser thrust to weight ratio and considerably greater inertia. The reason the EFCS makes the A320 an exceptionally nimble jetliner when handflown is because the systems are designed to take that into account and have been precisely calibrated to do so.

EDIT : In fact, the whole reason Airbus (as opposed to AF, who were responsible for this flight) used the high alpha demonstration to show off and sell the A320 was because no other airliner at the time could perform that manoeuvre precisely enough to be safe! In terms of risk, barrel-rolling the Dash 80 - while a superlative example of aircraft control - was a Sunday stroll compared to a low-speed high-AoA manoeuvre without EFCS backup.

@noske: While the Captain undoubtedly made mistakes as a result of the improvised approach, at least part of the reason the forest was "surprising" can be laid at the door of AF's operations department - they only had one copy of the airfield chart, so the chart supplied to the crew was a B&W photocopy. Unfortunately, the graphics indicating trees at the far end of the grass strip were of a light grey which did not transfer to the photocopy.

[Machinbird]

Dozy,
In the right hands and without all the "protections," your genteel A320 could be made to do a low g airshow on the order of what Bob Hoover did with his Aero Commander.

To make it fly in the genteel mode, they have provided some safety margin from the performance boundaries. What people are not talking about is how big the safety margins are (that is closely held data because in the wrong legal hands, it could be said that AI took away a pilot's maneuvering margins and thus caused an accident.

But I'll bet that there was .02 g available from the airframe that the computers were not letting the pilot obtain in the Habsheim accident.

Surely you must recognize the need for some safety margin from an engineering standpoint. There is variability between aircraft after all. Surface contamination of the wings does happen.

What a properly trained pilot can do is to operate an aircraft right up to the boundaries and if the boundaries are slightly exceeded, he will maintain control and correct back toward the actual boundary. (Unless an intervening computer limits this performance.:E)

[Machinbird]

[QUOTE=DozyBut disabling A/THR (and thus Alpha Floor) was central to the entirety of the attempted demonstration - there would have been no expectation of Alpha Floor operating.[/QUOTE]
I might suggest that you modify 'would' into 'should' in the above sentence. It is probably closer to the truth.

[DozyWannabe]

In the right hands and without all the "protections," your genteel A320 could be made to do a low g airshow on the order of what Bob Hoover did with his Aero Commander.

That's as may be - however not every pilot is on the order of a Bob Hoover, Roly Beamont or "Tex" Johnson. In this case we're talking about an AF line pilot - technically capable, certainly - but nowhere near on the order of a Bob Hoover.

To make it fly in the genteel mode, they have provided some safety margin from the performance boundaries.

The A320 is not particularly "genteel", particularly in airliner terms!

What people are not talking about is how big the safety margins are (that is closely held data because in the wrong legal hands, it could be said that AI took away a pilot's maneuvering margins and thus caused an accident.

That's a rather cynical way of thinking about it

But I'll bet that there was .02 g available from the airframe that the computers were not letting the pilot obtain in the Habsheim accident.

Popping back to HN39's graph near the beginning of the thread:
AlphaCmd in the graph represents the sidestick position PROFON_CAP, assuming a linear relation between AlphaProt=14.5 for SS=0 and AlphaMax=17.5 for SS=-16.

http://i.imgur.com/xDSJu8l.gif

I've got to say I'm not entirely convinced. Admittedly I'm unversed in the knowledge you're using to come to that conclusion, particularly with regard to CoG height and relationship to lift...

Surely you must recognize the need for some safety margin from an engineering standpoint. There is variability between aircraft after all. Surface contamination of the wings does happen.

Of course - however the margins are not particularly restrictive, again, especially for an airliner. Ultimately, though, the reason the aircraft crashed has nothing to do with the technology. Human factors aside, the overarching reason the aircraft crashed was because where it was going, there was a forest in the way.

EDIT - re:would/should, I think the wording is reasonable as it stands - the crew were briefed to have Alpha Floor disabled in order to perform the manoeuvre, and while the approach was improvised and the Captain's physical SA seems to have been compromised by that, he seemed to be pretty sure that A.Floor was out of the equation.

[gums]

While I disagree with 'bird about using "direct law", I find a flaw in your logic, Doze.

and while the approach was improvised and the Captain's physical SA seems to have been compromised by that, he seemed to be pretty sure that A.Floor was out of the equation. You point out two factors that are essential to understanding the final outcome.

- "improvised" ???? Huh? Flying a multi-zillion $$$ jet without having done a dozen practice profiles? BEAM ME UP!!!!

- thinking that the magic FBW system with all the "protections" and limits and such would "save" him and all the folks on-board? Give me a break.

- Not using manual throttle throughout the flyby? Sheesh. All the automation and such can only compensate for so much.

- blaming pictures of the field? Good grief!!! Just two or three practice missions would have made a huge difference.

I enjoyed flying my first fully-FBW jet, and we always had one place to hang our hat upon - max AoA was always there if we pulled full back on the sidestick. No auto-throttle, but that was up to us. The jet gave you everything it could.

[roulishollandais]

Since Gen 317 seconds RA<52 Ft, the longitudinal law is pitch law until Gen 327 (RA <32 Ft), and then incidence law with derotation. Both are nearly direct and no more Nz law.
There is a problem between Gen 312 and 315 due the trees BEFORE the runway. (These trees have been cut very quickly after the crash!) The expert Max Venet said at the first trial they did not exactly knew how much time was needed to switch the law to "flare" (incidence law with derotation).
It seems descending 2 Ft is enough to switch the law. But Machinbird is talking about 36 Ft which is much more due to loss of. 02 g. We know that stall is achieved on the A320 with g<1 (plane descending!) due to beeing "protected against stall" (rudderruddererrat link post 269). But AF447 learned us that the A330 (and probably the A320) DOES STALL. Using the stall definition with g<1 (.94?, .97?) is transfering ≈50 ft or more from the pilot's margin toward system margin, as Machinbird was arguing.
Contradiction accepted but wih proofs.:O

[CONF iture]

max AoA was always there if we pulled full back on the sidestick. No auto-throttle, but that was up to us. The jet gave you everything it could.
In theory, the Airbus is not different, as the procedure is to rely on the protections to get the most out of it, but as I wrote earlier :
What strikes me is how in Habsheim, New-York or Bilbao, the pitch authority has been compromised by features built into the system, whenever the aerodynamic had still more to deliver.

[CONF iture]

Why would he pull the throttles all the way back? I'm not an Airbus guy so I'm curious as to the reasoning.
Here is what Asseline writes in his book on page 105 :
When I pushed initially the thrust levers to an intermediate position when arriving above the Habsheim field, the N1 command was elaborated with the previously demonstrated delay, which I interpreted as a non response from the engines. Instinctively, I recycled the levers to the IDLE then TOGA position.

[noske]

And pushing the thrust levers to said intermediate position would have taken place about six seconds before TOGA thrust was commanded. Unfortunately there is no evidence of it on the flight recorders. As the Airbus presentation on Habsheim phrases it:

Only one thrust lever movement is shown after flight idle is selected at the commencement of the descent, and concurrently with the clicks being heard, at 4.5 seconds before the aircraft enters the trees.Possible explanations for this discrepancy (ignoring the extreme assumptions that Asseline ist just a liar or that the whole investigation was rigged to protect Airbus):


Asseline honestly believes that he commanded intermediate thrust at the transition from descent to level flight, but he never really did.
He did, but the thrust levers failed in such a way that not even thrust lever angle as recorded by the DFDR reflected the change.

And no, I don't think that both scenarios are equally likely.

[gums]

@conf

WRT to my comment about getting max performance when commanding max performance in the Viper....

In theory, the Airbus is not different, as the procedure is to rely on the protections to get the most out of it, but as I wrote earlier :
What strikes me is how in Habsheim, New-York or Bilbao, the pitch authority has been compromised by features built into the system, whenever the aerodynamic had still more to deliver

Which features compromise the pilot's command for max performance?

I also wonder if there were connections with the throttle implementation ( think Asiana) that the pilot did not expect, hence trying to "re-cycle" the system?

Secondly, our primitive system did not require full stick forward or back to command max AoA. As with the 'bus, it was a gee command, and of approx one gee per 4 pounds of stick pressure ( no movement, just pressure sensors), limited by AoA, so at max AoA all we could command was one gee even with 34 pounds of back stick pressure! In short, we could have duplicated the crash into the trees if power-limited and too slow. Hence my question about the throttle implementation.

Thanks ahead of time.

[CONF iture]

Which features compromise the pilot's command for max performance?

Hudson - Phugoid damping
Bilbao - Logic in the AOA protection in case of turbulent conditions
Habsheim - Radio silence ...

I also wonder if there were connections with the throttle implementation ( think Asiana) that the pilot did not expect, hence trying to "re-cycle" the system?
Two different things related to the thrust system happened earlier :

During the take off run, the system bizarrely requested to set the thrust levers to the CLB position.
At 1000 AGL when it was time to set the thrust levers to the CLB position, the N1 command did not follow the request and remained to the previous setting. Asseline had to retard the levers behind the CLB position, only then the N1 command was elaborated but not without delay.


Nothing of this is mentioned in the BEA report.
The final report does not even include the FDR data for that flight portion.

[gums]

Thanks, Conf

I'll have to dig out my FCOM stuff from the AF447 files and see how that throttle works.

Of course, a few practice flyby runs may have shown any problems with the throttle mechanization.

[Goldenrivett]

Originally posted by CONF iture..At 1000 AGL when it was time to set the thrust levers to the CLB position, the N1 command did not follow the request and remained to the previous setting. Asseline had to retard the levers behind the CLB position, only then the N1 command was elaborated but not without delay.
You would have thought then, that Capt. Asseline would have made doubly sure that the engine thrust was following his TL position, by flying at alpha MAX earlier in the demonstration (say 500 ft agl) and descending towards the runway using a controlled amount of manual thrust. To do a flight idle fly pass and hope the engines spooled up when requested (with the knowledge of the known thrust problem on take off), does not seem to be prudent to me.

[CONF iture]

To do a flight idle fly pass and hope the engines spooled up when requested (with the knowledge of the known thrust problem on take off), does not seem to be prudent to me.
They did interpret the initial situation as a temporary bug, but I do agree with you, it should have been a signal to not go and play at very low altitude ...

Second observation is :
Why the BEA is hiding the malfunction ?
What else are they hiding ?

[HazelNuts39]

Why the BEA is hiding the malfunction ?It did not, it's all in the CVR transcript. At 12:41:19 the captain comments that he knows about that 'bug' and at 12:42:12 that it is because he has reduced below a certain N1.

[Machinbird]

It did not, it's all in the CVR transcript. At 12:41:19 the captain comments that he knows about that 'bug' and at 12:42:12 that it is because he has reduced below a certain N1. Was this discussed in the Body of the BEA report?

I'd think they would want to address how that specific feature works and what influence it may have had with respect to the delayed power response.

[CONF iture]

It did not, it's all in the CVR transcript. At 12:41:19 the captain comments that he knows about that 'bug' and at 12:42:12 that it is because he has reduced below a certain N1.
The comment at 12:42:12 is not related to the thrust 'bug' but to the landing gear status.
Was this discussed in the Body of the BEA report?
Nowhere - Totally ignored ...

[roulishollandais]

It did not, it's all in the CVR transcript. At 12:41:19 the captain comments that he knows about that 'bug' and at 12:42:12 that it is because he has reduced below a certain N1.

The comment at 12:42:12 is not related to the thrust 'bug' but to the landing gear status.

Was this discussed in the Body of the BEA report?

Nowhere - Totally ignored ...
As often when the BEA does not want to comment anything they only write without discussion they did not think that is linked to the accident!
:
[…]
4.3 Aéronef
4.3.1. La commission a constaté que certaines anomalies de fonctionnement n'ayant joué aucun rôle dans l'accident (défauts mentionnées dans les compte rendus mécaniques) avaient été corrigées depuis sur les appareils en service. Elle n'estime donc pas avoir à formuler de recommandation a leur sujet. (sic)
[…]

[roulishollandais]

Report approved with unanimity by the members of the enquiry commission :
Claude Bechet, François Gonin, Bernard Farthouat, Robert Auffret, Philippe Gourguechon, Paul Arslanian, jean-Paul de Villeneuve:suspect:

[Chris Scott]

Quote from CONF_iture (my clarification in square brackets):
"Two different things related to the thrust system happened earlier :
(1) During the take off run, the system bizarrely requested to set the thrust levers to the CLB position.
(2) At 1000 AGL [roughly top of climb] when it was time to set the thrust levers to the CLB position, the N1 command did not follow the request and remained to the previous setting. Asseline had to retard the levers behind the CLB position, only then the N1 command was elaborated but not without delay."

Before I read the captain's memoir, I may be able to explain (1), and I've some observations on (2).

(1) In the early months of A320 airline operation, we had quite a lot of minor glitches on the FMGS. The PERF page includes separate figures for thrust-reduction and acceleration altitudes, the defaults of which are either 1000 ft or 1500 ft above the departure aerodrome, but each can be amended manually before T/O. Basle (LFSB/BSL) is 883 ft amsl, so the default settings may have been 1900/1900. AFAIK, the only function of the THR RED function is to provide an aide-memoire for the crew to reduce to climb thrust (if both donks are still running), which it does by a flashing "CLB" in amber on both pilots' thrust FMA.
The glitch was that the THR RED figure would be lost just after take-off thrust was commanded on T/O, leading to an immediate flashing CLB prompt. This was distracting, but had no effect on engine performance.

(2) After T/O the L/G was retracted in the normal way. After climbing only 1100 ft, the short cruise was conducted at 2000 ft QNH, i.e., roughly 2000ft amsl, maintaining the T/O Flap setting of 1+F (Slats 18 deg, Flaps 10), so ALT* (altitude-capture mode) would have been triggered well below it, pre-empting the acceleration phase for the FDs, which seem to have been in use. "Alt Star" was called by the copilot at 12:41:47, about 17 seconds after airborne, and his call may have been some seconds after the mode change.
At ALT*, the thrust would have needed to be reduced immediately and substantially below CLB thrust, to avoid exceeding the flap limiting speed. (With Habsheim so close, an IAS of about 185 kt would have been reasonable.) In fact, the CVR recorded sounds of thrust reduction starting 5 sec after reaching 2000 ft. At that point, the A/THR was still engaged, and would have changed to SPD (speed) mode because of ALT*. So the thrust reduction would have been achieved by resetting the speed to a figure at or below the current IAS, and/or by retarding the thrust levers well behind the CLB gate, thereby introducing an upper thrust limit less than CLB. (Partially retarding the thrust levers to the desired N1 would in either case have been desirable before changing to manual thrust.)
A/THR was in fact disengaged about 8 sec after start of thrust reduction, and happened to coincide with a L/G CRC warning. The latter would have been caused by a combination of N1 < 75% and radio height < 1000ft, (as the captain intimated, although my 1988 FCOM states < 750 ft). There were some low hills to their left at this point.

So, prior to reading Capt Asseline's account, I see no evidence of anomalies of engine performance in the T/O, climb or cruise phases. I agree with CONF_iture, however, that it would be good to see the DFDR for the whole flight.

[Machinbird]

Secondly, our primitive system did not require full stick forward or back to command max AoA. As with the 'bus, it was a gee command, and of approx one gee per 4 pounds of stick pressure ( no movement, just pressure sensors), limited by AoA, so at max AoA all we could command was one gee even with 34 pounds of back stick pressure! In short, we could have duplicated the crash into the trees if power-limited and too slow. Hence my question about the throttle implementation.
Gums-re, the underlined portion:
That would only apply to level flight. I would hope you could still pull a fractional g load if headed straight up and needing to ease the nose over-
A related question however:
Did your system allow you to pull right into buffet, or did it keep you out of the buffet boundary? How much of your ultimate potential performance did it let you have? Did some Vipers turn a little bit better than others?
I ask this from the vantage point of a Phantom Pflyer where if the outer wing panels weren't shaking, you were not asking for maximum performance!:}

I'd think that the A320 would not be allowed to enter the buffet boundary to keep from scaring the cargo, but I could be wrong. (Normal Law)

Buffet would more likely be encountered in clean configuration if it is possible to achieve the buffet boundary.
With flaps/slats down, there may not be a buffet boundary or a very minimal one.

[A33Zab]

Extracted from VFW614 ATTAS Hi AoA protection tests: LINK (http://ftp.rta.nato.int/public/PubFulltext/AGARD/CP/AGARD-CP-593/31chap26.pdf)


High AOA Protection (dynamic)

Fig. 18 shows that the high AOA protection operates precisely even in dynamic maneuvers, such as the side stick full back step input at VCAS = 180kts.
The aircraft reacts with a ∆nZ =0.8g pitch up movement up to 22° pitch attitude.
The high AOA protection counters this dynamic nose-up maneuver with a 4° nose down elevator command.
The aircraft stabilizes at α= αlim and VCAS = Vαlim.
In neither of the high AOA protection tests the maximum angle-of-attack αlim has been exceeded.


http://i474.photobucket.com/albums/rr101/Zab999/DynamicAoA_AoAmax_zpscd279ac2.jpg

[roulishollandais]

I just edited my post 304.
I replaced (Chris Scott link) to "rudderrudderrat link post 269".
Apologize to both.
That link with the 3 definitions of stall was :
http://lessonslearned.faa.gov/Indian...0Condition.pdf (http://lessonslearned.faa.gov/IndianAir605/A320%20Final%20Special%20Condition.pdf)
... new edit : after rereading all that, I find NO definitive definition of stall, only rumours

[HazelNuts39]

Hi A33Zab,

Long time no see! Thanks for the graph amd the link to the DASA/NLR report on the ATTAS-SAFIR research project. A valuable contribution to the never ending Habsheim debate and the intriguing questions still posed 25 years later.

[gums]

@ 'bird: Sorry, I needed to make it clear that I could have duplicated the crash, as even with max back stick, my jet would only hang at 27 deg AoA and 1 gee. No way to command a higher AoA or gee. Biggest difference was no autothrottle modes and such. Push forward, motor spools up. Pull back, motor spools down. Unlike the GE J-79, both the fans I flew ( A-7D and F-16A) spooled up very slowly, in comparison. At max AoA and low altitude, I had plenty of thrusties left over, so a go around was easy. Aftr reading the India 605 discussion, I got a better understanding of the autothottle, and I see why the Captain under scrutiny here pulled back and then advanced to ensure he was out of the autothrottle system. Too late. He had plenty of excess thrust available, but "pressed" too hard and too long.

Secondly --- No real buffet in the Viper at max AoA. Kinda a low freq buzz. This was due to the leading edge flaps moving all the time, eventually reaching full down above a certain AoA. Provided excellent directional and lateral stability - like full aileron command and 250 - 300 deg/sec roll rate all without departing.

Last answer for the 'bird -- - could always push to a lower gee unless in a deep stall. The lack of pronounced buffet and the loss of the nose down pitch authority we had in a deep stall was the main reason I joined the AF447 crowd. Although the 'bus still has nose down authority when deeply stalled, our contributors here have shown it takes a long time to finally get flying again. Unlike the 'bus, we trimmed for a gee with either a small wheel or the coolie hat. We did not have the 'bus implementation. If I let go of the stick, the sucker tried to get to whatever gee I had it trimmed for. Did this at any attitude.

Didn't mean to repeat what I had posted on AF447, but simply answered the 'bird's question and added an opinion of the Captain's actions/judgement.

[Owain Glyndwr]

A33zab

Yes, thanks for that link and graph. I see that for quite a time the elevator was moving downwards and the AOA dropping even though the sidestick was being held fully back. Hmmmm!

[CONF iture]

Before I read the captain's memoir, I may be able to explain (1), and I've some observations on (2).
Thanks for such first hand information, I am satisfied with point (1) - It is the minimum the BEA should have mentioned.
Point (2) is more delicate - Your information seems accurate to me, but the correlation with the CVR data, and also the Asseline's account, would need that crucial missing piece of info that BEA kept out of our sight ...

Take your time to go through MA account, but we'll need to look deeper as there is so much more to be discussed ...

[Machinbird]

Yes, thanks for that link and graph. I see that for quite a time the elevator was moving downwards and the AOA dropping even though the sidestick was being held fully back.I read the alpha graph to indicate that the aircraft quickly approached its 22 degree pitch limit and the down elevator bunt maneuver was to keep it from overshooting. The time ticks are relatively large 20 second increments.

Much more maneuvering g available in the example than at Habsheim.

I'm wondering whether these systems calculate limiting pitch angle based on excess thrust available at the instant? I'll have to read the report more carefully.

A33Zab, thank you for your post, which, although not directly relating to AI products, is still highly relevant.

[Chris Scott]

Quotes from OK465:
"But the current [Airbus] software is probably many FCS software iteration revisions from 1988."

Yes! I wish all our contributors would bear that in mind, particularly when interpreting data from later accident reports... Why would it not have changed? The guys at St Martin-du-Touch would have been constantly striving for improvement, as well as accommodating the new birds.

"I was young then."

Really? :rolleyes:

[HazelNuts39]

The guys at St Martin-du-Touch would have been constantly striving for improvement, ... ... and more than a runway separated them from the guys at 1 Rond Point Maurice Bellonte.

[Linktrained]

The flight had been briefed to be at 100 ft along R/W 34 L.
Perhaps due to the realignment with grass R/W 34 R the aircraft used some of its potential energy (now only height and airspeed, with flight idle having been set at 12.44.14) the descent to 46 ft became inevitable. This required more of a climb than had been planned - or an earlier selection of TOGA by a few seconds ( 10 ?).


Many pilots must have learned a valuable lesson since seeing what happened.


PS. Some could be younger, too !

[HazelNuts39]

Vs1g was demonstrated in flight test in ALT LAW.In your post #266 you quoted from the preamble of the FAA Special Conditions for the US type certification of the A320. Development testing is not certification testing. In the Special Conditions it is evident that certification tests to determine Vs1g are conducted in normal law, the AoA protection system operating normally, except that alpha-floor must be disengaged.

Special Condition 6. Flight Envelope protection, (b) Angle-of-Attack Limiting, under (1)(i), states: Conditions associated with the determination of the stalling speed are those provided in § 25.103 of the FAR.

Relevant sections of FAR § 25.103 as amended by the special conditions read:

(a) VS is the reference stalling speed with--
(1) Stalling speed determined at not greater than IDLE thrust (NOTE: automatic go-around thrust application feature must be disengaged);
(2) (...)
(b) The stalling speed VS is the minimum speed obtained as follows:
(1) From a stabilized straight flight condition at any speed not less than 1.16 Vs (or speed at AOA protection onset, if greater) nor more than 1,30 Vs, apply elevator control to decelerate the airplane so that the speed reduction at the stall does not exceed one knot per second.
(2) (...)

In short: Vs is the speed obtained in normal law when the sidestick reaches the aft stop, i.e. Vs = V(alpha-max)

[CONF iture]

Yes, thanks for that link and graph. I see that for quite a time the elevator was moving downwards and the AOA dropping even though the sidestick was being held fully back. Hmmmm!
Is it surprising to you ?

[Chris Scott]

Me: "The guys at St Martin-du-Touch would have been constantly striving for improvement, ..."
HN39: "... and more than a runway separated them from the guys at 1 Rond Point Maurice Bellonte."

Yes, the rather uninspiring Airbus premises - did they even exist in 1988? As you know, the south side of la ville de Blagnac has changed almost beyond recognition since the late 1980s, partly because of the creation of the rocade. (And, more recently, the north side of l'aeroport de Blagnac went completely bonkers - but that's another story.)

[HazelNuts39]

For me St Martin-du-Touch were the premises of Aerospatiale Engineering and Production facilities, and I was assuming that Airbus was responsible for FCOM, FCTM, pilot training and Customer Services in general, and of course, marketing and sales.

[Owain Glyndwr]

Originally posted by Confiture
Is it surprising to you ?

No not at all.

[CONF iture]

http://i55.servimg.com/u/f55/11/75/17/84/image10.jpg (http://www.servimg.com/image_preview.php?i=158&u=11751784)

If I wanted to confuse and discourage a reader, I would probably do that :

Publish the photocopy of a copy of a copy ... of the original
Not include the seconds of the GMT
Add a TGEN for which I would not specify the origin
Publish a limited amount of FDR data
Publish those data for only a short period of that short flight

[HazelNuts39]

Hi Chris,

In your post #133 you wrote:
My interpretation is that 329.0s = 12:45:33 (see explanation offered in my previous post).I don't know if that issue was resolved. Recently I stumbled by chance upon a document entitled: The A320 Habsheim accident - An Airbus Industrie response to allegations ... .

On page 17 under "The CVR timing" Airbus writes:
CVR time 12H 44'31'' equated to DFDR time TGEN 266
TGEN is the timing in seconds from the start of the DFDR reading made by the investigators.

Does that help?

[Chris Scott]

Hi HN39,

"On page 17 under "The CVR timing" Airbus writes:
CVR time 12H 44'31'' equated to DFDR time TGEN 266"

Thanks for drawing that CVR timing reference to my attention, which I had not noticed in the 1991 Airbus document. It certainly confirms your interpretation of the co-ordination between the GMT references and the "TGEN" time frames which you had used in your graph (http://www.pprune.org/tech-log/528034-habsheim-6.html#post8212705), and contradicts (by one second) my own as used in my post (http://www.pprune.org/tech-log/528034-habsheim-7.html#post8214459) #122 of Dec 18, which I also used in #130 and #133.

So the BEA definition of t=0 at 1245:39 clearly occurs at DFDR time-frame 334.0 secs. IIRC, that is the last time frame in which none of the (sequentially recorded?) data was corrupted.

And the GMT minute is not updated until the time frame (TGEN) 1 second after the beginning of the minute, i.e., the first 1245 GMT (at TGEN 296.0) represents 1245:01, and the last 1244 GMT (at TGEN 295.0) represents 1244:60, which would be 1245:00.

[HazelNuts39]

RE:: Sequentially recorded?

The same document on page 12 has following Note:
The graphical representations of the DFDR data were prepared by Airbus Industrie from the data in the appendices of the Final Report. Recording times are adjusted to reflect the exact time at which each parameter is recorded. There remains a small potential error (up to 0.3 seconds) due to the variable delay between reading and recording a parameter.

[CONF iture]

And the GMT minute is not updated until the time frame (TGEN) 1 second after the beginning of the minute, i.e., the first 1245 GMT (at TGEN 296.0) represents 1245:01, and the last 1244 GMT (at TGEN 295.0) represents 1244:60, which would be 1245:00.
Why suddenly changing the norm ... ?
Why simply not publishing the extensive FDR time reference ... seconds included ?

"TGEN is the timing in seconds from the start of the DFDR reading made by the investigators."
Is it really too much asking than expecting the investigators to specify which GMT H:M:S has been selected ?

Yes, and I'm still inclined to the theory that TGEN 000.0 seconds is most likely to be the instant that TOGA was selected on T/O.
It would be surprising that TOGA or FLEX has been selected before any takeoff clearance, but again, FDR data would tell more ... if only published.

[rudderrudderrat]

Hi Chris Scott,
Sorry for the delay in finding this info.
Post #270: I note Owain Glyndwr's quick response that reversion to Pitch Alternate may not have been necessary during the certification flights. Experience suggests he is usually right,...
You and Owain are correct. Please see http://sfte2013.com/files/78988645.pdf, slide 35

"Test Conduct - Stalls for certification
in FBW aircraft
• Initially performed at forward CG
• Must be performed in Normal Law
– In case a stall occurs despite the protections
– Airbus - Normal Law modified with higher limiting AOA"

edit.
Hi Owain Glyndwr,
I see that for quite a time the elevator was moving downwards and the AOA dropping even though the sidestick was being held fully back. Hmmmm!
Could than be explained by the FBW computers sensing the rate of speed decay, knowing the rate of pitch change available, and anticipating the change in attitude necessary to avoid exceeding aMax?

Hi HN39,
In short: Vs is the speed obtained when the sidestick reaches the aft stop, i.e. Vs = V(alpha-max)
The above presentation suggests the aircraft is actually taken beyond CL Max, and its behaviour is observed in both Normal law and "Additional stalls performed with degraded control laws" (slide 35) to determine VS1g.

[HazelNuts39]

Hi rudderrudderrat,

Thanks for the link to an interesting paper. For the purpose of our discussion I think a few points are worth mentioning:

-- The presentation is dated 29 October 2013
-- It presents the views of Boeing and Airbus FTP's and FTE's, not necessarily that of the airworthiness authorities
-- The first bullet is "Stall Definition and Requirements", but the paper does not contain a discussion of that subject
-- The paper does not distinguish the tests conducted to determine the stall speeds from those to verify the stall characteristics or to identify aerodynamic characteristics such as the CLalpha curve

The A320 Special Conditions are dated 1988 and state clearly that Vs1g must be determined in normal law with alpha-floor disengaged. Consequently, Vs1g for the A320 is the speed obtained when the airplane is decelerated at 1 kt/s, the sidestick reaches the aft stop and is held there a few seconds. The published stall speeds and the cl-alpha observed in the accident (t-6 122 kt alpha=13.5) suggest that Vs1g corresponds to alpha-max of 17.5 degrees.

[Owain Glyndwr]

Hi rrr

Could than be explained by the FBW computers sensing the rate of speed decay, knowing the rate of pitch change available, and anticipating the change in attitude necessary to avoid exceeding aMax?Possibly, but it could be any number of reasons, and with absolutely zero knowledge of the VFW laws I would rather not speculate.

I would add though that the time history as shown is not IMO a very good response to a step elevator command. It certainly shows an effective limitation to alphamax, but as a response to a demand for maximum lift NOW! it leaves a lot to be desired. Sure it gives alphamax fairly quickly (as it should at 180 kts) but it only stays there for a very short time and that is followed by a very large response reversal and a slow climb up to maximum lift. In fact it takes about 20 seconds to get to a sustained alphamax.

[Winnerhofer]

May we go back to AF447 instead of pantomiming?
Order of French tragedies:
1) Rio
2) Perpignan
3) Toulouse
4) St Odile
5) Habsheim

[DozyWannabe]

Possibly, but it could be any number of reasons, and with absolutely zero knowledge of the VFW laws I would rather not speculate.

OK, so coming back to this after a brief absence I just have to check - we're talking about the VFW-Fokker 614 used in the ATTAS programme - *not* the A320 - right?

From DLR Portal - Research Aircraft - VFW 614 / ATTAS - out of operation (http://www.dlr.de/dlr/en/desktopdefault.aspx/tabid-10203/1127_read-266#gallery/112) :

ATTAS was primarily designed as a 'flying simulator', to simulate the flying behaviour of other - real (existing) or virtual - aircraft

Or in other words, a European version of NASA's "General Purpose Airborne Simulator" JetStar. So I think it is at least safe to assume that the specification of that EFCS would have had significantly different behaviour goals than that of the A320.

I agree with CONF_iture, however, that it would be good to see the DFDR for the whole flight.

Would you also say that while it may have been "good", it was nevertheless reasonable to truncate the published DFDR output to the time period relevant to the accident by the standards of the time?

Thanks for such first hand information, I am satisfied with point (1) - It is the minimum the BEA should have mentioned.
Point (2) is more delicate - Your information seems accurate to me, but the correlation with the CVR data, and also the Asseline's account, would need that crucial missing piece of info that BEA kept out of our sight ...

The BEA are only charged with publishing material proven relevant to the accident. Neither of those issues on the face of it seem to be relevant to the accident itself. The A320 at the time had some technical issues to be ironed out - this was and is true of every new type one cares to mention (and the truism has been demonstrated recently with the A380 Trent issues highlighted by QF32 as well as the B787 battery issues demonstrated several times).

That the report didn't mention those particular issues is not a case of "hiding" anything, it's a case of those issues being irrelevant to the accident sequence in the first instance, and stating the bleedin' obvious in the second!

[roulishollandais]

@Winnerhofer
And Concorde,
And Toronto
And unknown incidents

@rudderruddererrat
Thanks for the link. I hope A and B will continue these meetings with ICAO

@Hazelnuts39
I agree with your doubts.Who decided but Airbus themselves the special 1988 conditions replacing the regulators.


Does anybody know the certification date of Airbus in Vietnam (ref for USA...)?

[DozyWannabe]

@Hazelnuts39
I agree with your doubts.Who decided but Airbus themselves the special 1988 conditions replacing the regulators.

In 1988 Capt. Gordon Corps (of the Air Registration Board and its successor, the Design and Manufacturing Group of the Airworthiness Division of the Civil Aviation Authority) was in charge of that work. At that time he was still an independent consultant and not an Airbus employee.

[HazelNuts39]

Who decided but Airbus themselves the special 1988 conditions replacing the regulators.Another poster and I have been quoting from the FAA Special Conditions for the A320. These have followed the normal FAA process of rulemaking. A Notice of Proposed Special Conditions was published in the Federal Register on October 19, 1987. The comments received from the public in response to that notice are discussed in the preamble of the final rule. Regarding the reference stall speed the A320 Special Conditions are largely identical to the current standard of the regulation, proposed in 1995 and adopted in 2002 with FAR Amendment 25-108.

Many people from airworthiness authorities, manufacturers, operators and pilot associations in the U.S., Canada and Europe participated in the development of those changes in the regulations. In fact it would be difficult to identify a manufacturer of transport category airplanes or its regulatory agency that was not involved in that process. I doubt that any of those individuals would have claimed to be 'in charge of that work'.

[Chris Scott]

Quote from me:
"...I'm still inclined to the theory that TGEN 000.0 seconds is most likely to be the instant that TOGA was selected on T/O."
Quote from CONF_iture:
"It would be surprising that TOGA or FLEX has been selected before any takeoff clearance..."

Yes, I should have done the arithmetic! That puts TGEN 000.0 seconds at 1240:05z, and T/O clearance was not issued until 1240:40z.

We know that movement of the throttle levers through the various thrust-limit gates, as necessary for the selection of TOGA from idle, is audible on the CVR. My assumption was that TGEN 000.0 would have marked a point of synchronisation between the CVR and the DFDR, but there is nothing in the CVR transcript - such as a lever movement, or the start of a transmission on the R/T - to suggest any audible event at 1240:05z that would have been simultaneously recorded on the DFDR.

So it may be that TGEN 000.0 is a purely arbitrary point. The Airbus 1991 document explains that the CVR is not time-stamped, but that the CVR records at a steady speed. I infer that a minimum of two synchronisation points some minutes apart would be necessary, to enable the CVR and DFDR recordings to be synchronised for the whole flight. The selection of TOGA a few seconds before impact would appear to be one ideal example. **

BTW, I'm still reading Capt Asseline's account. My PDF is missing the Annexes. Can anyone steer me towards obtaining them, please?

** The selection of T/O thrust on take-off would be another.

[CONF iture]

I infer that a minimum of two synchronisation points some minutes apart would be necessary, to enable the CVR and DFDR recordings to be synchronised for the whole flight.
The BEA has made every effort to confuse the situation and make sure that what should be simple and straightforward is ... NOT.
Clac Clic Cloc and even Crac ! are happily mixed all along ...

You can find the Annexes for Asseline book here :
Crashdehabsheim.net (http://www.crashdehabsheim.net/)

Go to :

Bibliotheque
Le temoignage du pilote (telechargeable ici)
Telecharger les annexes

[DozyWannabe]

I doubt that any of those individuals would have claimed to be 'in charge of that work'.

Poor phrasing on my part - I put it down to a combination of exhaustion and jetlag. :zzz:

Of course more people and organisations were involved, and I shouldn't have been so sloppy with my wording - but what I was trying to get at was that in Capt. Corps case at least, he was still primarily answerable to his CAA role in determining certification criteria and his role as a consultant would be driven by that as opposed to Airbus's corporate benefit (as rh seemed to be insinuating).

The BEA has made every effort to confuse the situation and make sure that what should be simple and straightforward is ... NOT.

Er - how are they doing that? If I've got this right, you're talking about an appendix with the raw data over which the BEA have very limited editorial control. Remember that DFDRs were a relatively new technology at the time, and as such best practices were still being determined*.

Apropos of nothing, I find it somewhat ironic that you are criticising the BEA for supplying raw data (which can be opaque and confusing) in this case, yet you were demanding the release of raw data on the AF447 threads and criticising the BEA for not doing so. It seems to me that they can never do right as far as you're concerned.

EDIT :

* - For one thing, the computerised data manipulation tools and techniques which we take for granted today weren't fit for purpose in 1988. These days microcomputers are powerful enough to import vast amounts of raw data into a spreadsheet, and aligning the data visually is as simple as clicking and dragging a mouse. In 1988 there was still a significant gulf between the mainframes and minicomputers used in industry and the microcomputers used in offices and homes. The storage of data in a binary format was far from standardised - and even if that hurdle could be overcome, spreadsheets were only used on microcomputers, the most powerful examples of which in 1988 didn't have the RAM to effectively manipulate that much data.

[Chris Scott]

Hi ruderrudderrat,
Thanks for the link to that recent presentation on stall testing. Very interesting.

Quote from me:
"I agree with CONF_iture, however, that it would be good to see the DFDR for the whole flight."
Reply from Dozy Wannabe:
"Would you also say that while it may have been "good", it was nevertheless reasonable to truncate the published DFDR output to the time period relevant to the accident by the standards of the time?"

Yes. If not, I would have said "...would have been good..."

It would be useful, however, for the purpose of our present discussions, not least to improve our understanding of the data. I would like to see the T/O and climb, if only to observe the acceleration of the engines from idle to T/O thrust (presumably TOGA, but I haven't seen a reference to that, and the copilot's FMA call of "THR" is apparently anomalous, unless the PF had selected Flex Thrust with no "assumed temperature" entered in the FMGS), the height at which the right turn was initiated, and the changing status of A/THR.

Hello Confit,
Thanks for that link to the annexes in Capt Asseline's account.

[DozyWannabe]

It would be useful, however, for the purpose of our present discussions, not least to improve our understanding of the data. I would like to see the T/O and climb, if only to observe the acceleration of the engines from idle to T/O thrust...

Agreed, if only for the sake of completeness. However, the BEA's response to Capt. Asseline's correct assertion that the elevators briefly deflected down while he had the stick back was a thorough real-world reconstruction of the aircraft's commanded flightpath - an experiment entailing significant effort and expense*. Is it not therefore reasonable to infer that the BEA compared engine performance data at take-off and during the accident sequence against both each other and an optimum set of test data - and found no significant anomalies?

One thing that has always puzzled me about Capt. Asseline's responses in the aftermath of the accident is that the BEA's report, far from assigning the causes purely to the actions of the flight crew, actually implicitly hauls AF's operations practices over the coals - not just in terms of highlighting the woefully inadequate briefing materials, but also highlighting contradictory rules and regulations pertaining to conduct of display flights. Reading between the lines the report more-or-less implies a reasonable argument that the crew were almost set up to fail. Capt. Asseline and his lawyers could likely have turned this into the most clear-cut example of mitigating circumstances in history, and yet they ignored that angle in favour of going after Airbus. To me this makes so little sense that I can't begin to fathom out why.

EDIT :

* - To put the question from another angle, if the BEA were being pressured to go easy on AI, why go to all that trouble and expense when a verbal response with no experiment would have carried the same weight in an official sense?

[CONF iture]

I find it somewhat ironic that you are criticising the BEA for supplying raw data
Quote please ... ?

I would like to see the T/O and climb, if only to observe the acceleration of the engines from idle to T/O thrust (presumably TOGA, but I haven't seen a reference to that, and the copilot's FMA call of "THR" is apparently anomalous, unless the PF had selected Flex Thrust with no "assumed temperature" entered in the FMGS)
Except from the "Clac Clac" mentioned in the CVR transcript at time 12 41 04 which could possibly mean that the FLEX detent was the one used for the take off, I cannot remember anything explicitly mentioned in the report.

Also from Asseline's account on page 103 :
Pour décoller de Bâle-Mulhouse, j'avais positionné les manettes des gaz dans un cran intermédiaire (FLX MCT), déterminant une poussée réduite au décollage, de manière à économiser les moteurs tout en garantissant une poussée suffisante compte tenu de la masse réelle de l'avion.

[DozyWannabe]

Quote please ... ?

Your post #336:



Publish the photocopy of a copy of a copy ... of the original
Not include the seconds of the GMT
Add a TGEN for which I would not specify the origin
Publish a limited amount of FDR data
Publish those data for only a short period of that short flight



Points 2 and 3 in particular. You're looking at raw output right there - even if the technology existed to do so at the time, the BEA could not add GMT seconds or TGEN origin without by definition altering the data.

Additionally, your point 1 is likely just a consequence of the age of the document. The original digital files are likely locked away in an archive - the online version is purely for casual use, so digging the original up would likely not be cost-effective. Your point 5 is essentially the same as your point 4 and therefore redundant. Remember that the BEA is only charged with publishing data relevant to the accident. As I said above, if the engine performance at take-off and during the accident sequence was shown to be within normal limits, then there's no requirement for the BEA to publish it.

[Chris Scott]

Quote from Dozy Wannabe:
"Is it not therefore reasonable to infer that the BEA compared engine performance data at take-off and during the accident sequence against both each other and an optimum set of test data - and found no significant anomalies?"

If one believed beyond any doubt that the BEA carried out a wide-ranging, impartial, and fully-informed investigation, that would indeed be a reasonable inference. But it would cut no ice with the BEA's detractors, and the bureau itself is unlikely - if you'll forgive the understatement - to enter into discussion on any alleged deficiencies.

Alternatively, we can all accept the findings and suspend further discussion...

[Chris Scott]

Salut Confit,

I have not reached page 103 yet...

Your quote from Capt Asseline:
"Pour décoller de Bâle-Mulhouse, j'avais positionné les manettes des gaz dans un cran intermédiaire (FLX MCT), déterminant une poussée réduite au décollage [...]"

The CVR quotes the copilot's FMA call as "Thrust...", which happens to be abbreviated (as you know) to "THR" on the thrust FMA.

I therefore infer that the throttle levers were placed in the FLX/MCT gate (i.e., Flexible T/O thrust was commanded for take-off), but that no assumed temperature was presemt in the FMGS.

For other readers, the flexible-thrust "assumed temperature" is a device to order the engines to develop the lower T/O thrust associated with a higher ambient air-temperature than that pertaining on the day. It is calculated before departure as part of the T/O performance calculation, and entered manually into the PERF / Take-Off page of the FMGS by the flight crew. When the throttle levers are placed in the FLX/MCT gate for take-off, the thrust FMA should read "FLX nn", where "nn" is the assumed temperature, e.g., "FLX 53".

[DozyWannabe]

If one believed beyond any doubt that the BEA carried out a wide-ranging, impartial, and fully-informed investigation, that would indeed be a reasonable inference. But it would cut no ice with the BEA's detractors, and the bureau itself is unlikely - if you'll forgive the understatement - to enter into discussion on any alleged deficiencies.

The issue as I see it is that most if not all of those "detractors" resort to obfuscating and, on occasion, outright misrepresenting the contents of the report in order to support their arguments. At the time, the lawyers representing Capt. Asseline and the SNPL union embarked on an almost unprecedented campaign to muddy the waters, briefing journalists with any allegation they could muster.

The ultimate outcome of this scorched-earth approach is that even people who are knowledgeable on the subject are frequently misinformed. I still hear talk of "the computer thought it was landing", "the autopilot was in command when it crashed" and "the report blamed the pilot" among other things - most of which, like those examples, are completely and demonstrably untrue.

Alternatively, we can all accept the findings and suspend further discussion...

If only...! :sad:

Seriously though, there's no such thing as a perfect report in the real world - but in my opinion the BEA performed an investigation that was at least as "wide-ranging, impartial, and fully-informed" as could be expected.

[HazelNuts39]

Publish the photocopy of a copy of a copy ... of the original
Not include the seconds of the GMT


RE 1: In those days the BEA didn't have facilities to read DFDR tapes and decode the data. The tape was read at the CEV (Centre d'essais en vol), the french military flight test centre, which also does civil certification test flying and houses the french national test pilot school. The print looks to me like it was produced on a IBM typewriter or daisy-wheel printer and published in the report as it was received from the CEV.
RE 2: The DFDR GMT parameter probably records only hours and minutes (remember the timestamp of the AF447 ACARS messages?). The seconds have to be determined by counting the frames, and fractions of a second from the parameter position within the frame word sequence. Who sets the GMT clock in the airplane? The wording of the note in the Airbus memo suggests that TGEN is not a recorded parameter but a frame count added in the CEV printout of DFDR data.

[Chris Scott]

Quote from HN39,
"Who sets the GMT clock in the airplane?"

I reckon that's a rhetorical question, but I can confirm that it's anyone who is inclined to do so. We never did it in flight. According to the FCOM, its timings are "sent to the CFDIU, FDIU and FMGC." The FDIU, of course, feeds the DFDR.

Quote from Dozy Wannabe:
"One thing that has always puzzled me about Capt. Asseline's responses in the aftermath of the accident is that the BEA's report, far from assigning the causes purely to the actions of the flight crew, actually implicitly hauls AF's operations practices over the coals - not just in terms of highlighting the woefully inadequate briefing materials, but also highlighting contradictory rules and regulations pertaining to conduct of display flights. Reading between the lines the report more-or-less implies a reasonable argument that the crew were almost set up to fail. Capt. Asseline and his lawyers could likely have turned this into the most clear-cut example of mitigating circumstances in history, and yet they ignored that angle in favour of going after Airbus. To me this makes so little sense that I can't begin to fathom out why."

You make a very reasoned argument, and it may well be that Capt Asseline was badly advised. Union politics are far too complex for the likes of me.

[HazelNuts39]

The above presentation suggests the aircraft is actually taken beyond CL Max, and its behaviour is observed in both Normal law and "Additional stalls performed with degraded control laws" (slide 35) to determine VS1g. The presentation describes recent experience during testing of A380 and B787 and is looking toward the future. Slide 35 doesn't mention Vs1g.

Nothing in the presentation is in conflict with my position that the certified Vs1g for the A320 was determined in accordance with the Special Conditions for type certification of the A320: in normal law as designed for the production airplane except with alpha-prot disengaged. Additional tests were conducted with modified or degraded control laws to demonstrate stall characteristics and to obtain aerodynamic data beyond alpha-max (and probably beyond cLmax) for design purposes and for the flight simulator data package.

P.S.
This effectively means that the 1987 FCOM (Chris Scott #213) was in error when it said "alpha-max = alpha (1.06 Vs)".

[Owain Glyndwr]

@Hn39

Nothing in the presentation is in conflict with my position that the certified Vs1g for the A320 was determined in accordance with the Special Conditions for type certification of the A320: in normal law as designed for the production airplane except with alpha-prot disengaged. Additional tests were conducted with modified or degraded control laws to demonstrate stall characteristics and to obtain aerodynamic data beyond alpha-max (and probably beyond cLmax) for design purposes and for the flight simulator data package.

That makes sense. Those additional tests (with an increased alphamax so that it didn't interfere with the testing) would have come first to establish what alphamax might be and the certification tests then completed up to that alphamax/Vs1g as you suggest.

[CONF iture]

The CVR quotes the copilot's FMA call as "THR...". That is presumably the transcriber's shorthand for the spoken word "Thrust...", which happens to be abbreviated (as you know) to "THR" on the thrust FMA.
Actually the CVR transcript is clear :
"Thrust SRS ! RUNWAY" at time 12 40 59

and the copilot's FMA call of "THR" is apparently anomalous, unless the PF had selected Flex Thrust with no "assumed temperature" entered in the FMGS
Anomalous I agree, as the call should have been either MAN FLX or MAN TOGA, but if no flex temp had been previously entered, and Asseline had set the thrust levers to the FLX detent, the ECAM ENG THR LEVERS NOT SET would have been triggered ... Was it different in 88 ?

[CONF iture]

Your post #336
And where my post #336 (http://www.pprune.org/8267463-post336.html) is "criticising the BEA for supplying raw data" ?
All I can see is demanding for more ... not less.

[Chris Scott]

Quotes from CONF_iture:

"Actually the CVR transcript is clear :
"Thrust SRS ! RUNWAY" at time 12 40 59"

Yes, so we'll have to hope that my recollection of 25 years ago is better than that of what I read 25 hours ago! Thanks, I've amended my post.

"Anomalous I agree, as the call should have been either MAN FLX or MAN TOGA, but if no flex temp had been previously entered, and Asseline had set the thrust levers to the FLX detent, the ECAM ENG THR LEVERS NOT SET would have been triggered ... Was it different in 88 ?"

Yes, it was certainly different on the A320 in 1988 (I'm not au-fait with current A320 FMAs, which may or may not have changed). Your "MAN FLX" and "MAN TOGA" were simply "FLX nn" and "TOGA". In the absence of an assumed temperature in the FMGS (as inevitable on a touch and go), and the throttle levers at FLX/MCT, the FMA would indicate "THR". I don't recall any associated ECAM mesage, but am not sure. IIRC, the FADEC would deliver TOGA in that configuration, but it would seem sensible to advance the throttles to make sure. I've yet to find any reference to this situation in my contemporary FCOMs.

[CONF iture]

RE 1: In those days the BEA didn't have facilities to read DFDR tapes and decode the data. The tape was read at the CEV (Centre d'essais en vol), the french military flight test centre, which also does civil certification test flying and houses the french national test pilot school. The print looks to me like it was produced on a IBM typewriter or daisy-wheel printer and published in the report as it was received from the CEV.
Coming from you, I am usually expecting more solid argumentation ...

RE 2: The DFDR GMT parameter probably records only hours and minutes (remember the timestamp of the AF447 ACARS messages?). The seconds have to be determined by counting the frames, and fractions of a second from the parameter position within the frame word sequence. Who sets the GMT clock in the airplane? The wording of the note in the Airbus memo suggests that TGEN is not a recorded parameter but a frame count added in the CEV printout of DFDR data.
the DFDR has a very accurate elapsed time-base and however erroneous the airplane clock could be set, what matters is the full DFDR time for the duration of the flight and the correlation with the CVR that can be established for the total duration of that flight.

[AlphaZuluRomeo]

And where my post #336 (http://www.pprune.org/8267463-post336.html) is "criticising the BEA for supplying raw data" ?
All I can see is demanding for more ... not less.
Conf,
Like you, I would feel more confortable with a precise and clear timing.

But when you demand "more" (i.e. precise GMT timing), then by definition you are demanding something that is not raw data.

Indeed I think Dozy'point was that raw data should not be altered ; that means nothing added nor removed.
If not, then they are no raw data anymore, but a transcription.

Cheers.

[HazelNuts39]

the DFDR has a very accurate elapsed time-baseThat may be so (and probably that makes the frame count an accurate measure of elapsed time), but apparently this DFDR records GMT (or UTC) obtained from an aircraft system clock. Large aircraft in the public transport category are equipped with flight data acquisition units (FDAUs). (*). The FDAU (or equivalent) orders data and sends them to the DFDR, whose function is limited to data recording.

Data acquisition systems output a binary file sequenced in four-second frames.
Each frame is divided into four one-second-subframes.
Each subframe is divided into 64, 128, 256 or 512 “words” of 12 bits each, depending on the FDR’s technology.
The bit is the basic binary unit whose value is either 0 or 1.

A 12 bit word can have values from 0 to 4095. The time parameter is recorded in the first 'word' of a data frame.

That 12 bit word can contain the 24*60=1440 minutes in a 24 hour day, but 24*60*60=86400 seconds requires at least 5 bits more.

According to the Habsheim accident report the DFDR recorded 200 parameters, including 141 'discretes' (yes/no parameters). The attachment shows that this number of parameters required reducing the recording frequency of many parameters to once every 4 seconds. Wouldn't it be a waste of valuable capacity to record GMT with a resolution of one second when 1 minute is sufficient because the seconds can be derived from the frame count?



(*) This and other quotes are from a BEA report "Flight Data Recorder Read-Out - Technical and Regulatory Aspects". The quotes are edited for continuity.

P.S.
BTW, subtracting 141 discretes from 200 recorded leaves 59 parameters. Counting the parameters in Tome 1 through 6 that readout must be pretty complete for the number of parameters.

[DozyWannabe]

But when you demand "more" (i.e. precise GMT timing), then by definition you are demanding something that is not raw data.

Indeed I think Dozy'point was that raw data should not be altered ; that means nothing added nor removed.
If not, then they are no raw data anymore, but a transcription.

Got it in one. Also, if HN39 is correct that they had to go to an outside agency to extract the data in a format they could use (apparently hard copy in this case), then they would not have been able to easily rework it into a digital format again. I reiterate - in computing/information science terms the state of the art in 1988 would by today's standards be considered not just incredibly primitive, but also awkward to use, time-consuming and expensive. Particularly of note is that not only did microcomputer paradigms differ from mainframe/minicomputers in general, but even among microcomputers there was no universal standard for representing data - the norm was a mishmash of proprietary formats, all of them incompatible with each other. Transferring data was theoretically possible, but every byte would have to be individually hand-checked to ensure the data was identical at both ends.

It is therefore entirely understandable that the BEA simply tacked a copy of the raw output to the report as an appendix. I'd also be inclined to suggest that because the output is clearly raw data dumped to hard copy, it rather undermines the "tampered data" allegations. Ironically, according to Airbus's response it was the outside investigator's lack of experience with this new data format that caused him to misinterpret the synchronisation.

EDIT (as HN39's post is a useful illustration!) :

Data acquisition systems output a binary file sequenced in four-second frames.
Each frame is divided into four one-second-subframes.
Each subframe is divided into 64, 128, 256 or 512 “words” of 12 bits each, depending on the FDR’s technology.
The bit is the basic binary unit whose value is either 0 or 1.
A 12 bit word can have values from 0 to 4095.

And if that description sounds complicated enough to begin with, I suspect that in 1988 one would require specialist hardware and software to translate that binary data into a human-readable format (as in the hard copy of the appendix). To transfer it digitally in the human readable form would still require specialist, bespoke software to make sure the values were not corrupted in translation (ASCII or EBCDIC? MSB or LSB?), and the internal RAM capacity of even high-end business machines may not have been enough to process the data in one pass, let alone manipulate it.

Wouldn't it be a waste of valuable capacity to record GMT with a resolution of one second when 1 minute is sufficient because the seconds can be derived from the frame count?

Indeed, and I think with today's level of technology in everyday use it's easy to lose sight of the fact that the growth in storage capacity and processing power is not linear, but practically exponential. The example I use is of my current cellphone, which is two years old and even back then was considered only modestly powerful, yet it has more raw processing power and storage capacity than the state-of-the-art desktop PC I built for my own use in 2002 - i.e only a little more than a decade ago, and almost a decade and a half after Habsheim. Extending that analogy, a decade before Habsheim computers were the size of a room, and reel-to-reel tape was the preferred storage medium if you were lucky - some were still using punch cards!

[Linktrained]

Doxy,
The Wirek Wire recorder issued to me as a 20 year old in 1948 recorded data DIGITALLY on a long steel wire. It couldn't do anything else, just ON or NOT ON..... It was the size of a small suitcase - but heavier.


Some 30 years later, my employers had their B707 Flight plans prepared at JFK ( or was it Idlewild ?) and telexed, digitally to the UK.


My very own ZX81 battery powered computer from 1982, recorded digitally on cassette tapes. (I learned just a little binary, too !) I saw one used with a digital camera (B&W) a year or two later.


Once data is recorded digitally, surely the use of a checksum tended to prevent or at least identify possible errors.

[jcjeant]

Got it in one. Also, if HN39 is correct that they had to go to an outside agency to extract the data in a format they could use (apparently hard copy in this case), then they would not have been able to easily rework it into a digital format again. I reiterate - in computing/information science terms the state of the art in 1988 would by today's standards be considered not just incredibly primitive, but also awkward to use, time-consuming and expensive. Particularly of note is that not only did microcomputer paradigms differ from mainframe/minicomputers in general, but even among microcomputers there was no universal standard for representing data - the norm was a mishmash of proprietary formats, all of them incompatible with each other. Transferring data was theoretically possible, but every byte would have to be individually hand-checked to ensure the data was identical at both ends.

It is therefore entirely understandable that the BEA simply tacked a copy of the raw output to the report as an appendix. I'd also be inclined to suggest that because the output is clearly raw data dumped to hard copy, it rather undermines the "tampered data" allegations. Ironically, according to Airbus's response it was the outside investigator's lack of experience with this new data format that caused him to misinterpret the synchronisation.Read it all .. it begs the question of who authorized and certified such a system .. which is one of the elements that can (should) participate in the improvement of safety ... how to record music can have multiple incompatible formats .. but in this case it is not for fun it is for safety
These competent authorities ... are they were really competent at the time?
Certified a system that has no common standards .. the results become uncertain because of the difficulty to analyze .. is this serious?

[DozyWannabe]

My very own ZX81 battery powered computer from 1982, recorded digitally on cassette tapes.

Actually, the recordings were analogue, and Sinclair machines used DA/AD conversion to translate between the two. This was why my Spectrum-owning friends all had tape recorders with a spot of Tipp-Ex on the volume and mic pots marking the "sweet spot" where that particular Speccy could make sense of what was on the tape.

In reference to what you're saying though, all the examples you give are of digital transmissions where the source and target spoke the same digital "language". Going back to your ZX81 example, you could not take the tape with your program on it and load that program into, say, a Spectrum - which shared the same processor but the ROM kernel and architecture was different - or a Vic-20, which had a different CPU as well as a very different architecture.

Checksums only work if the algorithms at each end are identical, and even then you still have the issue of how the data is processed at each end. It is the interpretation, processing and storage of the data between different architectures rather than the transmission that is the key problem here.

@jcj : Whoa there - the system used by the CEV on behalf of the BEA had common standards, *was* compatible and *was* certified. It seems the BEA wisely chose to avoid any further processing of the data beyond using the certified systems to provide a human-readable hard copy for the very reasons I've suggested. These days, the technologies used to process, store and render the data in different forms are several orders of magnitude more capable than they were back then.

Let me make this clear - the post extract of mine you just quoted does not relate to how the data was actually handled, it relates purely to the difficulty level, technically speaking, that would have been involved in attempting to transfer the data to another machine to reformat it, a process which is commonplace today. I was trying to explain why CONF iture's complaint about the GMT time formatting etc. was somewhat unreasonable given the technical state of the art at the time.

Here's a copy of the FAA regs pertaining to DFDRs : 14 CFR 121.344 - Digital flight data recorders for transport category airplanes. | LII / Legal Information Institute (http://www.law.cornell.edu/cfr/text/14/121.344)

[Chris Scott]

Hi jcjeant,

I think you'll find that in 1988 many airliners (particularly American ones) were still using scratch-foil (analogue) flight recorders, monitoring about 7 parameters. The DFDR on the A320 was advanced for its time.

BTW, it was unfortunate that the QAR cassette in the electronics bay was (presumably) destroyed in the ensuing fire, because that would have revealed even more data than the DFDR. ***


PS [by EDIT on Jan24]

Now that DozyW has provided that useful link to the relevant FARs for the present day, it may be worth having a look at the FAA regulations pertaining, at the time of Habsheim (1988), to turbine-powered airplane types that had been type-certificated before October 1969.

14 CFR 121.343 - Flight data recorders. | LII / Legal Information Institute (http://www.law.cornell.edu/cfr/text/14/121.343)

Types still enjoying the leniency of the above FARs in 1988 would have included all derivative models of B737 and DC-9 (and presumably the MD-80 series, which was certificated by amendment to the DC-9 type-certificate).

To summarise the relevant parts:
aircraft in that category (type-certificated before Oct 1969) needed only 6 parameters (one of which was time), although by 25MAY1989 the 6 parameters had to be recorded digitally (so the scratch-foil recorder was no longer acceptable), and by 25MAY1995 the minimum number of parameters rose to 11;
and individual aircraft manufactured after 26MAY1989 needed a DFDR with a minimum of 17 parameters.

So in the case of a B737-400, which was competing for sales head-on with the A320, I infer that hulls manufactured at the time of the Habsheim accident could have been registered in the U.S. with a 6-parameter DFDR, or even a scratch-foil recorder initially.

*** [by EDIT on Jan27]
Re the A320 QAR, further enquiry suggests that - on the contrary - it only records roughly the same number of parameters (and the same sampling rates) as the DFDR, although the precise specification may vary from airline to airline. A print-out of a 1988 BCAL/BA A320 QAR confirms this, and is presented in 5 "passes" - roughly corresponding to the 5 "tomes" shown in the BEA report for Habsheim.

[CONF iture]

But when you demand "more" (i.e. precise GMT timing), then by definition you are demanding something that is not raw data.
You are misinformed - It is raw data.
The time is recorded continuously on the DFDR.

Wouldn't it be a waste of valuable capacity to record GMT with a resolution of one second when 1 minute is sufficient because the seconds can be derived from the frame count?
That is why one sampling every 4 seconds is practiced and is enough... Where is it on the Habsheim listing ?

Bad luck, as the pilots did not die during the crash, they had their own testimony to produce, and it appears to be slightly different from the BEA account.
Now, if you want to quickly discredit the pilots version, all you need to do is opening the data on the table, ALL the data, not the very limited amount the BEA chose to share ...

[HazelNuts39]

Where is it on the Habsheim listing ?You could look, for example, at the fragment you copied in your post #336 to see that FL and MACH are recorded at intervals of 4 seconds.

all you need to do is opening the data on the table, ALL the data, not the very limited amount the BEA chose to share ... As I said, I believe the DFDR readout published in the accident report contains all parameters recorded on the DFDR for the relevant part of the accident flight.

[HazelNuts39]

From the FAA Regulatory and Guidance Library (http://rgl.faa.gov/Regulatory_and_Guidance_Library%5CrgFAR.nsf/0/F4232BB0F43C93E8852566EF006E20F6?OpenDocument):

Code of Federal Regulations

Part 121 OPERATING REQUIREMENTS: DOMESTIC, FLAG, AND SUPPLEMENTAL OPERATIONS
Appendix B--Aircraft Flight Recorder Specifications

Sec. B121.1

Appendix B--Airplane Flight Recorder Specification.


Parameter: Time (GMT or Frame Counter) (range 0 to 4095, sampled 1 per frame).
Range: 24 hours------------
Accuracy sensor input to DFDR readout: ± 0.125% Per Hour-------------------
Sampling interval (per second): 0.25 (1 per 4 seconds).
Resolution readout4: 1 sec.


4This column applies to aircraft manufactured after October 11, 1991.

Amdt. 121-197, Eff. 10/11/88

[CONF iture]

You could look, for example, at the fragment you copied in your post #336 to see that FL and MACH are recorded at intervals of 4 seconds.
We were talking about the time ... remember ?
Where are the seconds ?

As I said, I believe the DFDR readout published in the accident report contains all parameters recorded on the DFDR for the relevant part of the accident flight.
The crew has questioned the engines response and thrust operation for the latest part of that flight but also for the earliest part as well as the middle one, so the "relevant part" of that flight has to be the all flight starting from engines light up ... and we're not talking about a transatlantic flight here.

[HazelNuts39]

The time is recorded continuously on the DFDR.Nothing is recorded 'continuously' on a DFDR. All parameters, including time, are sampled by the FDAU at specific intervals, with a certain accuracy and a certain resolution, and sent to the DFDR to be recorded. The DFDR does not generate a 'time' parameter as such. I don't know if the 'very accurate elapsed time-base' resides in the FDAU or in the DFDR, but it only serves to write subframes at intervals of very accurately one second.

Where are the seconds ?In the Frame Counter, (range 0 to 4095, sampled 1 per subframe of 1 second and labelled TGEN in the printout).

[DozyWannabe]

Bad luck, as the pilots did not die during the crash, they had their own testimony to produce, and it appears to be slightly different from the BEA account.

You've got it backwards. For one thing, the allegations regarding aircraft performance (such as elevator behaviour and engine performance) were made by Capt. Asseline *during* the investigation, *before* the BEA report was released. We have already ascertained that the BEA's response was to take those allegations seriously, checked the DFDR data and flew an identical profile to ascertain why the aircraft behaved as it did. The engine performance was shown to be within normal limits, and the elevator behaviour shown to be a result of alpha protection. All this is included in the report. There was no need to "discredit" anyone, merely prove scientifically and by example that the aircraft behaved normally - a requirement satisfied by the report.

Notably, once things had settled, it was only Asseline who kept insisting that the aircraft had failed him - nothing was heard from his first officers since. If the crew as a whole had such faith in Asseline's position, why did they not put their names to his book? Why did his lawyers never pursue the fact - laid out clearly in the BEA's report - that the airline's standard of operations were not fit for purpose in this case? I reckon SNPL union politics have a lot to answer for here - I don't think they ever had his interests at heart, and instead saw him as a tool in their ongoing vendetta with Airbus.

[NeoFit]

Notably, once things had settled, it was only Asseline who kept insisting that the aircraft had failed him - nothing was heard from his first officers since. If the crew as a whole had such faith in Asseline's position, why did they not put their names to his book? Why did his lawyers never pursue the fact - laid out clearly in the BEA's report - that the airline's standard of operations were not fit for purpose in this case? I reckon SNPL union politics have a lot to answer for here - I don't think they ever had his interests at heart, and instead saw him as a tool in their ongoing vendetta with Airbus.

However certain details are surprising (http://denidejustice.wordpress.com/2010/12/17/2004-le-jugement-du-crash-de-lairbus-a320-a-habsheim/)

translate.google.com
The two flight recorders are recovered intact but disappear the same night, transported by the director of the DGAC, a Daniel Tenenbaum without counsel Mulhouse Jean WOLF considered it appropriate to affix court sealed !

The following evening, Monday, June 27, all authorities will conclude the innocence of the accident airplane charging multiple pilot errors. The plane was flying too slowly, too low and the gas were delivered too late. All control systems of the aircraft worked perfectly. A survey at least expeditious ...
As per design

[DozyWannabe]

@NeoFit

The "tampered flight recorder" argument was based on a misreading of the data by the independent investigator hired by Asseline's SNPL team, which amateur investigators tied to a photo in which a Swiss photo expert said the two sets of flight recorder casings were different.

The problem with the latter is that no-one from the investigation ever said that the cases in the photo were those recovered from the aircraft - they may have been "dummy" units shown to the people working the site so they knew what they were looking for.

[awblain]

The photos of the "wrong recorders" don't even look convincingly "wrong".

It really does show how ambitious doing FBW was at the start in the 1980s, when the recorder technology is compared.

[CONF iture]

In the Frame Counter, labelled TGEN.
Those are not the seconds that would allow us to validate the correlation FDR CVR for the total duration of that short flight and definitely establish that the pilots were wrongly questioning the aircraft functioning.

[CONF iture]

they may have been "dummy" units shown to the people working the site so they knew what they were looking for.
I love you too Dozy.

[DozyWannabe]

CONF, every bit of footage of a search briefing involving police and volunteers I've ever seen includes a display of examples of what they will be searching for, whether that be a photo or a near-identical object.

The report states that engine behaviour was normal, meaning that the investigators compared the full FDR data to expected behaviour. They flew the aircraft for real to determine why the elevators briefly deflected nose-down. In other words, they acted on the allegations made by Asseline and in the latter case backed up what he said, though that behaviour also turned out to be expected.

[CONF iture]

The DFDR does not generate a 'time' parameter as such.
Yes it does, the CVR doesn't.
Where does the labelled GMT H MN on the Habsheim listing come from then ?

Nothing is recorded 'continuously' on a DFDR.
If the DFDR is able to restitute every second and confirm them every 4 seconds, that's pretty 'continuous' to me.


However certain details are surprising (http://denidejustice.wordpress.com/2010/12/17/2004-le-jugement-du-crash-de-lairbus-a320-a-habsheim/)
I didn't know that one, but it seems to be a pretty good resume on the Justice side ...

[DozyWannabe]

Where does the labelled GMT H MN on the Habsheim listing come from then ?

He already told you:

RE 2: The DFDR GMT parameter probably records only hours and minutes (remember the timestamp of the AF447 ACARS messages?). The seconds have to be determined by counting the frames, and fractions of a second from the parameter position within the frame word sequence. Who sets the GMT clock in the airplane? The wording of the note in the Airbus memo suggests that TGEN is not a recorded parameter but a frame count added in the CEV printout of DFDR data.

Regarding the blog post, not only does it get significant allegations wrong (translated):

Monday, June 27, all authorities will conclude the innocence of the accident airplane charging multiple pilot errors.

We've already established that the report does not conclude pilot error to be the only factor, and I've explained that it would be practically impossible for anyone to have manipulated the data. Plus the fact that the only extraction employed and performed by the CEV was a direct dump to paper hard copy indicates there was no intermediate process where data could have been changed (at least not without being glaringly obvious to the naked eye).

The latter, under the pressure of complaints by passengers and the Union of Airline Pilots (SNPL) will handle the case in an emergency such as law requires.

There's those four letters again - they can throw around as many allegations about the judicial handling of the case as they like, but it doesn't prove malfeasance on the judicial side and it certainly doesn't have any bearing on the technical investigation.

[rudderrudderrat]

Hi DW,
The report states that engine behaviour was normal,

If we all agree on the end of recording (impact) at time 339 seconds, then is there any point about discussing when the recording started?

If you look at data on TOME 4, at time 240 secs, the engine N1s (1,2 %)follow the TL positions (Manettes 1,2 DA (angle)). As the TLs are closed (+014 to -0), the engine revs decay slowly from 68% to about 30% - which appears normal.

Between time 328 to 330, the TLs are advanced to the GA position (+45) and the engine revs increase from 29% to 83% by time 334.
The TLs are then closed briefly (-0) at time 336, and then advanced to +25 position at time 338. Meanwhile the engine revs decay from 83% to 56% then back to 86/87% - then end of recording.

It appears that the engines did respond normally (they achieved 83% within 4 to 6 seconds of moving the TLs from idle), but Capt. Asseline thought they had not and briefly closed them seconds before impact.

[DozyWannabe]

Between time 328 to 330, the TLs are advanced to the GA position (+45) and the engine revs increase from 29% to 83% by time 334.
The TLs are then closed briefly (-0) at time 336, and then advanced to +25 position at time 338. Meanwhile the engine revs decay from 83% to 56% then back to 86/87% - then end of recording.

It appears that the engines did respond normally (they achieved 83% within 4 to 6 seconds of moving the TLs from idle), but Capt. Asseline thought they had not and briefly closed them seconds before impact.

Interesting. Am I right in thinking that procedure of going back to the idle detent and thence forward to climb or TOGA only really applied to situations where A/THR is active? If so, it might indicate a last-ditch reversion to trained behaviour on the part of the Captain, even though A/THR was supposed to be disengaged?

[awblain]

That reduction to 56% certainly wasn't going to help matters in front of the trees; it would also have set the computers an unusual - if not difficult - task, to hold the aircraft stable close to maximum AoA as the power cycled up, half- way down and then up again.

I wonder if that power reduction about 5 seconds before impact agrees with the much-discussed-on-documentaries sound recording timing from the crowd's video? Perhaps the conspiracy-minded are hearing and timing the second rise in N1 just before impact with the trees, when in fact they should be timing the first rise in N1 after which the throttles were cycled. I remember "5s" is always quoted as the claimed engine delay.

I also wonder whether the perceived lack of apparent response from the engines that led to Asseline cycling the power levers was reinforced by odd cues from being lower and slower than expected?

[rudderrudderrat]

Hi DW,
Am I right in thinking that procedure of going back to the idle detent and thence forward to climb or TOGA only really applied to situations where A/THR is active?
If you select TOGA, you should get TOGA power irrespective of whether the A/T is engaged or not.
It appears to me, that had he maintained the TLs in the TOGA position, then he would have had TOGA thrust by time 335 (engines accelerate very quickly at high power settings).

[Chris Scott]

Hi rrr,

"If we all agree on the end of recording (impact) at time 339 seconds, ..."

Afraid not... HN39 and I have discussed this at some length. The last time frame whose data is totally uncorrupted is TGEN 334.0, and is treated, IIRC, as t-zero in the reports.

During TGEN 335.0, the data start to unhinge. That, following what appear to be uncorrupted signs of the initial impact, is particularly noticeable in the 4 sequential values of x and y accelerations (longitudinal and lateral respectively), and the 8 values of z (normal) acceleration presented in Tome 3. In Tome 1, the single sample of GS glitches to 648 kt.

[DozyWannabe]

The last time frame whose data is totally uncorrupted is TGEN 334.0, and is treated, IIRC, as t-zero in the reports.

During TGEN 335.0, the data start to unhinge. That, following what appear to be uncorrupted signs of the initial impact, is particularly noticeable in the 4 sequential values of x and y accelerations (longitudinal and lateral respectively), and the 8 values of z (normal) acceleration presented in Tome 3. In Tome 1, the single sample of GS glitches to 648 kt.

Just to be clear, this should not be seen as anything particularly untoward in terms of process. As it turned out, the Loral-branded F800 DFDRs fitted to early A320s had issues with data corruption when subjected to shocks over a given threshold:

G-KMAM Incident (http://www.rvs.uni-bielefeld.de/publications/Incidents/DOCS/ComAndRep/Gatwick/pm-excalibur.html)

The DFDR was a Loral F800. This type of recorder has been found to give poor performance on the A320, and the investigators discovered that the DFDRs on all four of Excalibur's A320s had a history of problems, including random track changing, incorrect Built-In Test Equipment (BITE) indications, and corruption of data (AAIB 2/95 Sections 1.11 and 2.11). The track changing and BITE problems were cured by the replacement of a certain Electrically Erasable Programmable Read Only Memory (EEPROM) unit, a modification which Loral made mandatory only in response to considerable pressure from industry and regulatory agencies.

The data corruption appeared to be due to vibration. The anti-vibration tray on the A320 had not been tested to the requirements of Radio Technical Commission for Aeronautics (RTCA) document DO-160 "Environmental Conditions and Test Procedures for Airborne Equipment". Trials after the incident showed that the recording quality was improved by mounting the DFDR on a tray conforming to the DO-160 standard.

(AAIB 2/95 specifies the DFDR as "Loral F800". The DFDR on F-GFKC which crashed at Habsheim in 1988 was referred to as a "Fairchild F800" and that on F-GGED which crashed at Strasbourg in 1992 as a "LORAL-Fairchild F800" in the appropriate reports. Presumably these are all references to the same make and model of DFDR.)

Since 1989, Airbus Industrie have used a different make of recorder for all test and certification flights. It is stated (AAIB 2/95 Section 1.11, p 17) that the Loral recorder is fitted only on delivery to the customer. Precisely what the status of the airworthiness certificate is following this modification is not clear!

A320 Spoiler Disconnection (http://www.rvs.uni-bielefeld.de/publications/Incidents/DOCS/ComAndRep/Gatwick/A320-Excalibur-report.html)

Investigation of this revealed that this was a common problem and that although Airbus had made a number of approaches to Loral regarding the unit, the situation had not improved.

[CONF iture]

The TLs are then closed briefly (-0) at time 336, and then advanced to +25 position at time 338.
Interesting
That reduction to 56% certainly wasn't going to help matters in front of the trees
And all at 10 46 GMT when the crash was at 12 45 ...
You are all very confused, and the BEA made sure you would be with such a 'pitiful' report.
Never wonder why the English version was not on the BEA site ... ?

[Chris Scott]

Gentlemen,

Although the DFDR itself has the advantage of being in a protective compartment at the tail of the a/c, the scores of sensors that supply the data to it, not to mention the FDIU/FDAU itself, are not afforded that luxury. They wouldn't be much use if they were...

[rudderrudderrat]

Hi Chris,
The last time frame whose data is totally uncorrupted is TGEN 334.0, and is treated, IIRC, as t-zero in the reports.
Thanks for the correction.
In which case, Capt. Asseline didn't select TOGA until sometime after 328 and before 330.
That was only 4 to 6 seconds before impact with the tree tops, and the engines had accelerated to 83% & 84%.

Edit: I don't see what was unusual about the engines response.

http://www.blackholes.org.uk/PP/TypicalEngineResponseChart.jpg

[AlphaZuluRomeo]

You are misinformed
Am I? How convenient :)

[Bidule]

@ CONF, post 383

If the crew questioned the aircraft functioning BEFORE their presentation at Habsheim, they have been even more stupid to start such an unusual flight, very close to the limits of any operation particularly with passengers on board!

[CONF iture]

@ Bidule, post 398

It is not forbidden to read the thread first.
Habsheim #312 (http://www.pprune.org/8258581-post312.html)

[DozyWannabe]

@CONF:

The BEA aren't hiding the earlier (known) malfunction, it simply isn't relevant to the incident. They are hiding nothing else. People who are still tilting at windmills, like yourself and your chums at the SNPL however, are very much on a hiding to nothing if this thread is anything to go by.

[CONF iture]

They are hiding nothing else.
Not even the English version of the report or the annexe VII ... but why am I asking that to a guy who even don't know the CONF setting used in Habsheim (http://www.pprune.org/8254085-post289.html) but cannot stop commenting on things he simply has no knowledge on ... !?

[DozyWannabe]

Not even the English version of the report or the annexe VII ...

If you use the report search facility on the BEA's website, you'll note that even today the BEA do not routinely publish a report in a language other than French unless one of the major parties involved comes from - or unless the accident itself occurred - outside of France. No "hiding" there.

Annexe VII covers the period directly concerning the accident only - it was, and in many cases still is, standard practice for most agencies to do this. Any data from other phases of the flight is quoted in short form during the narrative section of the report - again, nothing "hidden" and nothing untoward.


but why am I asking that to a guy who even don't know the CONF setting used in Habsheim (http://www.pprune.org/8254085-post289.html) but cannot stop commenting on things he simply has no knowledge on ... !?

I was tempted to report your post for this unnecessary personal swipe, but haven't done so as I don't want to seem petty.

For what its worth I've deliberately restricted my input to the computer systems design and specification as much as possible - any comment of mine outside of those bounds only refers to aspects that can be understood with a little basic aero knowledge plus a decent level of reading comprehension (or explicitly quotes people with the requisite qualification when responding to them).

Far more qualified minds than mine have been picking apart the technical and piloting aspects - not to mention the operations angle in which, as the BEA correctly stated in their report, AF are deservedly found significantly wanting (I found gums' sense of disbelief very appropriate on that point). You've got A320 line pilots both current and former - as well as at least one A320 TRE - saying that the aircraft behaviour seems normal.

From my point of view, and I've said this before, I take back my initial reservations about going over this again because I've learned a whole new raft of information for which I am very grateful. I came into this new discussion ready to defend my existing position that the BEA report was fit for purpose and the aircraft's behaviour expected, but prepared to be open to change that opinion if evidence showed otherwise - I have to say that the new or clarified information uncovered has tended to strengthen that opinion (for what little my opinion is worth) rather than undermine it.

As far as the non-technical aspects of the aftermath go, it seems to me that it was the parties acting for Capt. Asseline and the union who showed more interest in pursuing a predetermined agenda than the judiciary, Airbus or the BEA. Of the latter three parties, it is the BEA who in retrospect come across as by far the most even-handed, going to considerable lengths to take evidence, both primary and secondary (from all parties) into consideration when trying to determine what happened. The oft-repeated myths that the authorities sought to solely blame the flight crew, ignored allegations made by Asseline and the SNPL and tampered with flight data simply do not stand up to scrutiny when even a cursory review of the evidence is undertaken.

I do have some sympathy for Capt. Asseline, because while his judgement skills regarding conduct of the flight undoubtedly fell short of the standards required, he and his flight crew were poorly prepared and briefed by the airline prior to the flight (and the airline subsequently had no qualms about using a contradictory rule regarding minimum altitude for display flights to throw him under the bus). Not only that, but as Chris Scott suggests, I'm beginning to think he was poorly advised by those representing him - leading him away from what may have been a reasonable argument against the airline and instead convincing him to pursue allegations of technical failure against the manufacturer, a much more difficult proposition legally, but more in keeping with their long-held desire to score points against Airbus.

That it can be inferred such advice was given to an undoubtedly traumatised man wrestling with his own sense of responsiblity - of which a natural psychological factor is denial - does little to endear them in my opinion.

[Bidule]

@ CONF

It had been read but as you again went this way, the repetition seems to be useful!

[Chris Scott]

Quote from (http://www.pprune.org/tech-log/528034-habsheim-17.html#post8262036) Linktrained (January 13th):
"The flight had been briefed to be at 100 ft along R/W 34 L.
Perhaps due to the realignment with grass R/W 34 R the aircraft used some of its potential energy (now only height and airspeed, with flight idle having been set at 12.44.14) the descent to 46 ft became inevitable. This required more of a climb than hadbeen planned - or an earlier selection of TOGA by a few seconds ( 10 ?)."

Sorry to respond so late. First, perhaps it's worth reminding ourselves and others again that the fly-past was ill-conceived by the airline, with no provision made for a recce-flight or a ground inspection, and minimal preparation by both the airline and the crew. The notion of making an approach with a load of commercial passengers to a platform of 100ft QFE on an unknown non-runway (in A320 terms) - followed by a level-off, during which the a/c would be decelerated to within a couple of degrees of the stall AoA; followed by a high-alpha go-around - must rate as one of the most irresponsible in public-transport ops since the flying circuses of the 1920s and 30s.

Once committed, there would be little room for even one technical failure or significant crew-misjudgement. This crew seems to have made at least three serious errors in the execution of the briefed game-plan.

(1) The a/c arrived at the airfield boundary with far too much energy - kinetic and potential - to achieve the briefed game-plan of establishing stable flight at high AoA, using thrust to maintain speed in level flight during the transit of the airfield.

(2) The a/c levelled off at an indicated barometric height of about 60 ft above the reference altitude of the airfield - on both the pilots’ altimeters. (This assumes they had been correctly set to the QFE of 984 hPa, as announced in the CVR transcript). During the next 14 seconds before first impact, the barometric height fell slightly to about 50 ft, finally recovering to about 60 ft. Allowing for pressure-altimeter tolerances, together with the forward position of the static-pressure port and the aircraft’s high pitch-attitude, as much as 20 feet could be subtracted from the above values to “guess-timate” the actual heights.
It seems clear that, however foolish the airline and the captain were to plan the flypast at 100 ft on the QFE, the a/c would in all probability have cleared the trees had it been flown thus.

(3) The go-around was initiated too late to achieve any significant climb before the tree line was reached. In the event, insufficient time was allowed for the certificated time of engine acceleration from idle thrust, and the a/c was carrying a negligible surplus of kinetic energy to convert to any increase in altitude.
Prior to selecting TOGA thrust, the crew may have neither been aware of the horizontal distance to the tree line; nor the fact that the engine nacelles, landing-gear, and rear fuselage were lower than the tree tops. Their opportunities to recognise the hazard of the trees may have been limited: due initially to the rushed approach, and later to the visibility restrictions of the high pitch-attitude.

[roulishollandais]

Is it usual to see such mistakes punished by nine months firm jail?

[DozyWannabe]

Is it usual to see such mistakes punished by nine months firm jail?

I for one don't know - this is an aviation forum, not a legal forum.

Speaking for myself, I have grave reservations with the practice of treating every aviation accident as a criminal case regardless of circumstance, as seems to be the norm on the continent (and has happened here with similarly uncomfortable results).

What is certainly true is that the technical investigation and report by the BEA had no significant bearing on the outcome of the prosecution, and indeed the report practically handed Capt. Asseline a lifeline by pointing out the airline's slapdash approach to flight preparation and briefing. Why his legal team failed to take advantage of this remains baffling to me.

[awblain]

@DozyW

I agree with you entirely on the criminalization of accidents.

The potential problem I see with the legal investigations into accidents in some places - Italy especially comes to mind - is that there's no parallel to the English inquest process. If there's a death, then the legal investigation into the circumstances involves the same sort of prosecutors who'd eventually be involved in filing charges against suspects.

The professional accident investigation could thus potentially be hindered and disrupted. It could make it difficult to get people to testify freely, since there's a clearer lingering threat of prosecution in the process. If inquiries into the causes and facts proceed before and without relevance to potential prosecutions, then I suspect there's more chance of getting to the bottom of the circumstances, and reducing the opportunities for repeats.

[DozyWannabe]

If inquiries into the causes and facts proceed before and without relevance to potential prosecutions, then I suspect there's more chance of getting to the bottom of the circumstances, and reducing the opportunities for repeats.

In France, that's exactly what the BEA are charged with. They will, as I said, take the primary evidence and secondary evidence from all parties involved and attempt to make sense of it from a purely technical and operational standpoint. Their remit, as with most civil service agencies charged with aviation incident investigation (including our AAIB), forbids them from explicitly apportioning responsibility. That is left to the criminal and civil judicial apparatus.

Investigators may testify during the legal proceedings, but their testimony is strictly limited to the content of the report.

Most of the myths that grew up around this case are a result of deliberate obfuscation on the part of legal representatives via press briefings in particular. This is why a lot of commonly understood aspects are in fact incorrect.

[Chris Scott]

Quote from roulishollandais:
"Is it usual to see such mistakes punished by nine months firm jail?"

Probably not in the UK, and AFAIK not in other Anglophone countries, but "usual" is not a word that can easily be applied to this accident. I know nothing of the judicial process that led to the sentence.

I tried my best to base my assessment as much as possible on my interpretation of the published DFDR, rather than the narrative of the reports, and have yet to finish reading the captain's account. If the published DFDR is good - and so far I haven't read anything that proves otherwise - I doubt that the three main errors I describe will need much amendment. They are harder on the crew than the BEA report, which - in the tradition of air accident findings - carefully avoided direct criticism.

My opinion is that no one should be imprisoned unless (s)he is found guilty of a criminal act, and I see nothing to suggest that at Habsheim. There were misjudgments that led to fatalities, and a degree of negligence may have been involved. But as long as sane pilots are sitting lawfully at the sharp end of an aircraft, it is difficult for a prosecutor to demonstrate a casual disregard for the safety of passengers and crew.

To pick up on Dozy Wannabe's point, the tendency in some countries to prosecute personnel who are responsible for the safe conduct of flights, and who are thought to have made errors of commission or omission, positively discourages self-reporting of an error that might cause an accident next time. On the other hand, there can be no authority without responsibility. So the captain must ultimately carry the can, but what sanction should be imposed when he drops it?

The reality is that, even when exonerated from any blame for a big, publicised accident, more often than not aspects of a captain's performance will be questioned overtly or covertly within the airline, effectively ruining his or her career. And, if the situation becomes intolerable, the very fact of having been in the public eye will frighten-off any prospective employer.

There may be worse punishments than nine months at the president's pleasure. Nowadays, senior politicians seem to redeem themselves from their indiscretions in a year or two. Pilots don't.

[Linktrained]

Chris Scott, Hi,


This may have been a display which had been practiced a number of times successfully, probably along a "standard R/W" somewhere. To have to alter the R/W on "short finals", to fly over the grass R/W at Habsheim, whilst still at Flight Idle, probably meant a loss of some of the intended or planned altitude.


The practice flights would not have had trees, I would suppose... Nor a requirement to alter to a parallel R/W when low. So just IF there had been a slow build up of thrust, or even a failure of one engine (and THIS DID NOT HAPPEN at Habsheim) an adequate climb at V2 ought to be enough ... If there were no trees along the extended centre line...


I can only guess that almost any other transport aircraft, if it attempted to follow the same trajectory, would have ended in a similar result.

[awblain]

@DW,

I fully agree about France: BEA matches the role of AAIB/NTSB.

In Italy though, the investigating magistrate seems to sometimes need to impound wreckage and potentially disrupt the equivalent of BEA.

Of course, there's a step from having a vigorous ongoing judicial process to interfering with the dedicated crash investigation. And, the criminalization is subject to the choices of the magistrate in question.

[Chris Scott]

Quote from Linktrained:
To have to alter the R/W on "short finals", to fly over the grass R/W at Habsheim, whilst still at Flight Idle, probably meant a loss of some of the intended or planned altitude.

Why would the a/c have to lose height when the PF was trying to reduce speed? And on an A320, the FBW specifically counteracts the tendency of an a/c to lose height in a turn.

[Linktrained]

I should have used the phrase "potential energy ..." again.


And I had forgotten/neglected to think, of THAT feature of A320's FBW.


I'm sorry.

After " Flight Idle " had been selected the aircraft was rather like a slightly under-powered glider, at least until TOGA was selected ( perhaps too late ).

[RAT 5]

Yesterday I saw Air Crash Investigation about this accident. It was the 1st time I'd seen it. I am not an Airbus pilot, but Boeing hydraulics. From those in the know was the TV program accurate? It did include many new aspects that I had not heard before. It also seems the captain is still trying to clear his name from blame.
Regarding the choice & change of RWY, it stated that the only data given by AF was from the tarmac RW02, but on arrival they saw the crowd aligned along RW34 and so deviated at the last moment causing them to be high and hence at idle thrust.
It did quote the captain as calling for A/T disconnect at low level. Why would he do that? It did confirm the brief was for a flypast >100' but they descended and for a moment were <30'. This is quite obvious from the video.

But back to the question: was the program accurate?

[DozyWannabe]

But back to the question: was the program accurate?

Some of the dramatisation was a little contrived (the 'pilot or plane' briefing in particular, especially given that the final report considered the airline's practices to be a significant factor). It had to condense and skip through some aspects of the case faster than others, but to the best of my knowledge it is a fairly accurate summary. Regarding the A/THR disconnect (which apparently did not happen as briefed), there is a possibility it was called for and not subsequently performed due to task saturation from the last-minute approach change.

A/THR disconnect (in fact disabling of A/THR by holding the disconnect buttons down) is required for the manoeuvre as briefed (i.e. at or above 100ft RA), because otherwise Alpha Floor would activate and the aircraft would climb out. In this case apparently the disabling of A/THR did not occur, but Alpha Floor was instead inhibited by the aircraft remaining below 100ft RA.

The programme interviewed my old Software Engineering Professor (Dr. Peter Mellor), and I believe it was the last interview he did as he sadly passed away before the programme aired.

Regarding the archive interview with Ray Davis regarding time discrepancy, page 16 of this document (Airbus's 1995 follow-up report):

http://www.crashdehabsheim.net/Rapport%20Airbus.pdf

provides a reasonable response. In addition, I should point out that Capt. Davis was an excellent and thorough investigator and this should not be considered a mark against him, he was simply inexperienced with digital recorders, having retired before their introduction.

[Chris Scott]

Quote from RAT 5:
"From those in the know was the TV program accurate? [...]
"Regarding the choice & change of RWY, it stated that the only data given by AF was from the tarmac RW02, but on arrival they saw the crowd aligned along RW34 and so deviated at the last moment causing them to be high and hence at idle thrust."

I haven't seen the documentary, but that argument doesn't make much sense to me.

Approaching from the south, the direct distance to the threshold of Rwy 02 would have been several hundred metres LESS than to the thresholds of the grass strips 34L and 34R, which are several hundred metres further north.

According to the CVR, the first pilot positively to identify the airfield was the PF (captain) at about 1244:00, just as the PNF was busy initiating his first R/T call to Habsheim. At 1244:01, the PF announced (in French): "There's the field. It's there... you have it, eh?"

That was 98 seconds before first impact. The published DFDR print-out starts 4 seconds later, at 1244:05, with a time-frame notated as TGEN 240 seconds. (The last uncorrupted time frame is TGEN 334 seconds, and is considered to be the point of first impact with the trees on the north boundary of the airfield, defined as "t". So TGEN 240 is t -94 seconds.)

Habsheim ATC did not suggest that they fly along Rwy 34, and it seems doubtful that the crowd line along it would have been visible at this point.

At t -94, IAS and GS were 202 kt, passing height just over 1000 ft above the airfield elevation, and position about 3.5 nm SSE of the Rwy 02 threshold on a HDG of 336 (probably magnetic, but variation was only 2W). Had they decided at that point to position for Rwy 02, a rapid swerve to the left would have been necessary, shortly followed by a right turn to line up. That would have only increased the track distance-to-run to about 4 nm, a negligible difference from going direct to the thresholds of the grass strips.

In either straight-in scenario, there simply was not enough distance to get the height and speed off in accordance with the game plan. They needed an orbit or, preferably, a close circuit of the airfield while they slowed down and configured for the flypast.

In the event, they continued more or less straight towards the airfield, perhaps assessing the situation. There were no significant heading changes until t -50, with a 5-degree left turn; and t -25, when the a/c started a gentle right turn of about 15 deg on to its final heading of about 345 for the flypast of Rwy 34.

My interpretation of the DFDR suggests that the a/c was still descending as it passed over the threshold of Rwy 02. At t -13 it levelled off, at 60 ft baro on the QFE, just before the threshold of the shorter grass strip 34R. IAS was 132 kt - still more than 15 kt above the planned display speed. Hence the maintenance of idle thrust.

[DozyWannabe]

Had they decided at that point to position for Rwy 02, a rapid swerve to the left would have been necessary, shortly followed by a right turn to line up. That would have only increased the track distance-to-run to about 4 nm, a negligible difference from going direct to the thresholds of the grass strips.

Agreed, but given their fast-and-high state, wouldn't those turns have provided at least some opportunity to slow down and thus manage the descent more precisely, whereas a straight-in to the 34 grass strip make slowing down and managing altitude precisely more difficult?

There is no dispute that the original AF briefing assumed (as did Capt. Asseline) 02 would be the runway they would be using. As to precisely what time they realised that it was in fact 34, there is no clear evidence.

[jcjeant]

DW
Speaking for myself, I have grave reservations with the practice of treating every aviation accident as a criminal case regardless of circumstance, as seems to be the norm on the continent (and has happened here with similarly uncomfortable results).In France (as in many other countries) you (pilot .. driver .. company .. employer .. etc ..) can be tried on charges of "responsibility to have killed or injured people without premeditation and intent to kill or injure"
It is a charge quite normal when victims are the result of an accident or an event involving injury death of people
The choice of the appropriate court of justice (correctionelle or assises) will be primarily on the fact of presence or abscence of premeditation and intent
The difference between "assises" (jury) or "correctionelle" (no jury) is the penalty
This judgment is required (always) to determine who will be compensated and by who !

[gums]

Thanks, Chris.

Way I see it as well, with more to add.

Fer chrissakes, if you are trying to impress the folks watching, least you could do was to have made a few practice runs ahead of time. Sheesh. And then there's "pride".

Abandoning the planned power and altitude parameters by improvising also disturbs me deeply.

Despite practice, there's always a way to screw up. Witness the Thuderbird crash at Mountain Home ten years ago. Those folks have one or two practice missions before the main show. And then the nugget uses a bad altimeter setting and doesn't realize that it looks different than practice! Says on radio, "don't think I'm gonna make it". No kidding, dude.

As far as the law goes, in the U.S., any of the slf folks could file a civil lawsuit re: negligence. No criminal actions possible.

[DozyWannabe]

Habsheim ATC did not suggest that they fly along Rwy 34...

Looking at the airfield layout, it's unlikely they'd have felt they needed to. Based on the track as indicated on page 19/37 of the 1995 report and the location of the control tower at Habsheim, it would have appeared to the controllers that the A320 was already more-or-less lined up with the strip.

@gums : Legal matters regarding aviation accidents in the UK usually remain in the civil arena, though there has been the odd exception - one of which had a particularly tragic ending. That incident also demonstrated why regulators tend to make poor incident investigators.

[CONF iture]

But back to the question: was the program (http://www.dailymotion.com/video/x18uqai_mayday-s09e03-pilot-vs-plane_shortfilms) accurate?
Far from it ... The cockpit action and dialogue are especially rubbish and even don't fit the official version.
On the other hand, all the interviews are of interest.

Regarding the archive interview with Ray Davis regarding time discrepancy, page 16 of this document (Airbus's 1995 follow-up report (http://www.crashdehabsheim.net/Rapport%20Airbus.pdf)) provides a reasonable response. In addition, I should point out that Capt. Davis was an excellent and thorough investigator and this should not be considered a mark against him, he was simply inexperienced with digital recorders, having retired before their introduction.

What make you state it's 95 ... ?
You don't need experience with digital recorders to read a listing.
And what make you state he did not have experience with DFDR ?
BTW when did he retire from the AAIB ?

If you use the report search facility on the BEA's website, you'll note that even today the BEA do not routinely publish a report in a language other than French unless one of the major parties involved comes from - or unless the accident itself occurred - outside of France. No "hiding" there.
If no hiding why the BEA did not put a link for the English translation of the accident Final Report that Airbus is quoting ?
Annexe VII covers the period directly concerning the accident only - it was, and in many cases still is, standard practice for most agencies to do this. Any data from other phases of the flight is quoted in short form during the narrative section of the report - again, nothing "hidden" and nothing untoward.
Tell me more about Annexe VII ... now you're getting interesting ...

[Linktrained]

Habsheim was 25 years ago but still some of the lessons are there to be learned by newbies (as well as by one 86 year old !)


As a trainee at Hamble in 1949 we heard only a limited amount of the information about some of the many more frequent accidents which occurred. ( Someone had taken off in a Tudor with some of the external controls locked... But was able to land... Because he was brilliant... or exceptionally lucky ?)


The York was my first large aircraft, and it was based on the Lanchester bomber but with a different fuselage. Of the 259 built there were recorded 87 " hull losses" in the twenty years of service. In too many cases, if the pilot(s) were killed, the accident could be attributed to PILOT ERROR, without really trying to find what were the factors which contributed. An early study by David Beaty was called "The Human Factors in Aircraft Accidents".


The technology, radio communication and navigation aids together with their reliability and redundancies are now built in to an aircraft at relatively small cost in terms of weight.


A round trip from Stansted- Keflavik- Gander- N/S -Bermuda- Kingston- 2 N/S- Bermuda-Gander -N/S -Stansted took 52 hours flying, using about 10400 gallons of 100 octane petrol to carry 36 troops, each way. (Capt, F/O, N/O, R/O, G/E and Sss). Astro and weather ships for fixes, no LORAN, or ILS. ( We had SBA !)



Today might be different !

[roulishollandais]

In too many cases, if the pilot(s) were killed, theaccident could be attributed to PILOT ERROR, without really trying to find what were the factors which contributed. An early study by David Beaty was called "The Human Factors in Aircraft Accidents". It was quite clear during Asselline/Mazières Habsheim trial, that some would have preferd they be dead instead asking questions about the A320 design and working that day. Asseline said it to the Court. Hipocrisy could be seen from some too.

[noske]

If no hiding why the BEA did not put a link for the English translation of the accident Final Report that Airbus is quoting ?

Here's a link for you (ha ha): Commission of Inquiry Into the Accident on 26 June 1988 at Mulhouse-Habsheim ... - France. Ministry of Equipment, Housing, Transport and the Sea - Google Books (http://books.google.com/books/about/Commission_of_Inquiry_Into_the_Accident.html?id=qiUxHQAACAAJ&redir_esc=y)

More seriously: I find it also frustrating to know that an English translation of the report was made, but was never published on the web. But I wouldn't call that malice, just laziness. The BEA seem to underestimate the continued worldwide interest in this particular report.

Tell me more about Annexe VII ... Good one. @Dozy: Page 23 in the report has a list of annexes, including "7. Corrélation DFDR-CVR", which never actually appears anywhere in the PDF. That is indeed a bit strange, but looks like an honest mistake to me. For some reason, the "Rapport final (additif)" (starting at page 55) consists of odd-numbered pages only, with blank pages in between. So annexe 7 may just have been on one of those even-numbered pages.

That should be easy to verify with a printed copy of the Journal Officiel, volume 1990. (A brief search across libraries in Germany, for example, already seems to indicate seven such copies.) Did anyone ever bother? :8

[DozyWannabe]

More seriously: I find it also frustrating to know that an English translation of the report was made, but was never published on the web.

I don't think it was - I suspect that's Airbus's own translation. They certainly have no shortage of multilingual people working for them!

@Dozy: Page 23 in the report has a list of annexes, including "7. Corrélation DFDR-CVR", which never actually appears anywhere in the PDF. That is indeed a bit strange, but looks like an honest mistake to me. For some reason, the "Rapport final (additif)" (starting at page 55) consists of odd-numbered pages only, with blank pages in between. So annexe 7 may just have been on one of those even-numbered pages.

That's weird, on this PDF (the one I've been using):
http://www.bea.aero/docspa/1988/f-kc880626/pdf/f-kc880626.pdf

Annexe VII starts, clearly headed, on page 36.

Possibly there's an earlier version doing the rounds where some hapless intern loaded the scanning hopper incorrectly?

...that some would have preferd they be dead instead asking questions about the A320 design and working that day.

Who? I love all this daft conspiracy talk, but spit it out - who do you think would have preferred them dead?

[noske]

Well, that's the same PDF that I've got, downloaded from the same location.

And you're right, it clearly says "Annexe VII", on the first page of the DFDR listing. But then, on page 45, halfway through the DFDR data, there is "Annexe VIII". :confused:

That is weird. That cannot be what they had intended, and it certainly does not match the table of appendices. But at least it would explain why the real appendix 7 (supposedly describing the CVR/DFDR synchronization) failed to get published.

Regarding the English version, all I know is that Asseline, in a rant about what he only calls "the Béchet report", complains that this piece of crap was even translated into English and then distributed among potential Airbus customers worldwide. Or something like that. Oh well.

[DozyWannabe]

But at least it would explain why the real appendix 7 (supposedly describing the CVR/DFDR synchronization) failed to get published.

Nope - look closely. The formatting of the data is different between Annexes 7 and 8. I suspect Annexe 7 is the *output used to synchronise* the complete DFDR data params (in Annexe 8) with the CVR.

Annexes/Appendices are usually used for supporting data, not narrative. An overview of how the flight data recorder was retrieved is in the BEA report, and an even more detailed description is in the Airbus "Executive Summary" report, along with the inadvertent error made by Capt. Davis in reading the output. Notably, the ACI programme got this wrong too - it was not ATC's discrepancy with the CVR, it was a misinterpretation of the DFDR. Capt. Davis seems to have treated the DFDR's transmit/receive flag as referring to the call from ATC (which would be almost 5 seconds adrift from the ATC recording and unacceptable). In fact the DFDR does not flag incoming, only outgoing transmissions, and the DFDR referred to the crew's response (just shy of 1 second adrift from the ATC timestamp, and within the margin of error).

All this and more is in the Airbus document - I don't understand why CONF is asking me to repeat tons of it.

Capt. Davis appeared in the Black Box series which was broadcast in 1996 and was listed as a former AAIB investigator. The UK TV programme about Asseline in which Capt. Davis attempted to decipher the recordings was broadcast in about 1989/1990 if I recall correctly. The last document I can find referring to his being active with AAIB was his work on the Air India bombing in 1985.

Regarding the English version, all I know is that Asseline, in a rant about what he only calls "the Béchet report", complains that this piece of crap was even translated into English and then distributed among potential Airbus customers worldwide. Or something like that. Oh well.

Well, if we disregard Capt. Asseline's somewhat salty language, that it was translated into English for Airbus cutomers' perusal could imply that the translation was performed by Airbus (and therefore unofficial).

Incidentally, according to Béchet himself in the ACI/Mayday programme, Capt. Asseline was initially very co-operative. It was only when it appeared the investigation was disproving his claims that the aircraft malfunctioned that he turned so savagely on them.

[Chris Scott]

Noske and Dozy,

Whoever did that PDF made a real pig's-ear of l'assemblage, as the page numbers of the Annexes are clearly not even close to being sequential. However, some of the missing ones in "Additif" could be because the drawings/photos might have been one-sided, and the copier didn't even notice (s)he'd copied the blank sides.

Then there's the original itself...

The contents of several of the Annexes do not correspond to the index. The chaos includes the fact that (as you say, Noske) Annexe VII is listed in the index as "Correlation...", presumably a narrative justification of the method used to synchronise the timings of CVR, DFDR, Basel ATC, and Habsheim ATC. It is missing from that PDF.

Instead, the first 3 "Tomes" of the DFDR print-out are erroneously headed Appendix VII, and Tomes 4 - 6 (correctly) as Appendix VIII.

Quote from DozyW:
"The formatting of the data is different between Annexes 7 and 8. I suspect Annexe 7 is the *output used to synchronise* the complete DFDR data params (in Annexe 8) with the CVR."

I think not. In fact the most valuable point of correlation between DFDR and CVR probably occurs at TGEN 266 in Tome 5, which (as HN39 pointed out (http://www.pprune.org/tech-log/528034-habsheim-17.html#post8268105) some time ago) is defined in the Airbus 1991 report as being 1244:31 GMT. That point (and hopefully at least one other) would have been used to insert a time stamp on the CVR transcript, as the CVR does not record any time stamp of its own, but (fortunately) runs at a steady speed. The correlation is the DFDR's discrete of the co-pilot's brief transmission on VHF1 (see also the CVR). So Tome 5 is essential in the correlation process.

BTW, I wasn't aware that Ray Davis, formerly of the AIB/AAIB, was a pilot, as you seem to be implying. At Farnborough in the 1980s, he was known simply as Ray Davis.

[DozyWannabe]

BTW, I wasn't aware that Ray Davis, formerly of the AIB/AAIB, was a pilot, as you seem to be implying. At Farnborough in the 1980s, he was known simply as Ray Davis.

He referred to himself as a "Trident man" during his interview for Black Box, which I suppose doesn't necessarily mean he was a pilot, but given that and the the fact the AAIB has always had a habit of hiring line pilots I thought it was a reasonable assumption. If I was incorrect, my apologies.

Regarding the PDF, it looks like the "additif" section consisted of double-sided pages, and either a blank page was left in situ on the side with the even numbered pages, or the scanner was incorrectly set up. This only seems to apply to that section however.

EDIT : Ouch - it looks like another scan hopper mistake was made in the original report Annexes - the pages between 26 and 32 seem to have been loaded initially in reverse, then almost random order. It does return to sequential order after that and no pages appear to be missing.

As I do regarding Mr. Davis above, we all need to be careful when "presuming" things about which we lack firm supporting evidence - I guess the only way to confirm the status of Annexe VII (which I'm still not convinced is erroneous) would be to write to the BEA themselves and ask.

[HazelNuts39]

Instead, the first 3 "Tomes" of the DFDR print-out are erroneously headed Appendix VII, and Tomes 4 - 6 (correctly) as Appendix VIII.In the html version of the Journal Officiel de la Republique Tomes 1 through 6 are in Appendix VIII, but Appendix no.7 is still missing.

[roulishollandais]

French aviation considers they don't have to respect the Law.

[AlphaZuluRomeo]

I will repost the link provided by noske:
Btw., for anyone frustrated with the poor quality PDF of the BEA report, here's a HTML version of it: Habsheim F-GFKC (http://wnoeker.de/kc/)

About the correlation, the plain text of the report explains how the CVR was exploited (§1.11.1.1):
Pour obtenir une chronologie précise, la vitesse de défilement de la bande a d'abord été recalée à l'aide de la fréquence 400 Hz de l'alimentation électrique de bord qui apparaît dans le spectre enregistré. Ensuite, les enregistrements des émissions radio ont été corrélés avec ceux des fréquences du contrôle de la circulation aérienne (ces derniers comportent une piste horaire).
But no explanation of how CVR and DFDR were correlated; there is OTOH a mention saying it was good (§ 1.11.2):
Enfin, la cohérence des données issues de l'exploitation du CVR, du DFDR et de certains moyens externes (photographies, bandes vidéo...) peut être qualifiée d'excellente.

[DozyWannabe]

In the html version of the Journal Officiel de la Republique Tomes 1 through 6 are in Appendix VIII, but Appendix no.7 is still missing.

The HTML version is an unofficial transcription though, so it's probably a transcription error.

roulishollandais - This is getting a bit tiresome. The truth is that in all likelihood it is Capt. Asseline who is mistaken about there being a conspiracy/cover-up. He'd have probably realised this himself had certain parties not convinced him to pursue the angle for their own interests.

AZR - As HN39/Chris Scott pointed out earlier, in the Airbus supplementary report, page 17/33 the example given uses two VHF 1 transmissions to synchronise the timings:

CVR time 12H 44' 17" equated to DFDR time TGEN 252
CVR time 12H 44' 31" equated to DFDR time TGEN 266

Other events may have been used earlier in the sequence though.

[Chris Scott]

The co-pilot's VHF 1 transmission at 1244:31 (TGEN 266) is particularly good, IMO, because it is so brief and easily defined: "Recu."

They probably needed another as close to TGEN 000 as possible, in which case there should be a good opportunity at the beginning of the T/O. The CVR transcript shows "Ca roule 296 Q" (probably spoken as "Ca roule deux-cent quatre-vingt-seize quebec"?) with a retro time-stamp of 1240:44, which presumably coincides with a DFDR discrete starting at about TGEN 039.

Brief, knee-jerk acknowledgements are unlikely to be preceded by a pause with the Tx key already open. Transmissions requiring thought might be.

However, what the CVR transcript does not reveal is that (IIRC) the CVR has separate channels for each pilot's Tx, and each pilot's Rx, as well as the area microphone. Therefore, I would expect the co-pilot's Tx channel to pick up the sound of the transmit key being opened and closed. If so, the CVR keying event would be well within the one-second accuracy provided by the DFDR discretes.

Quote from HN39:
"In the html version of the Journal Officiel de la Republique Tomes 1 through 6 are in Appendix VIII, but Appendix no.7 is still missing."

Yes, the HTML version is so much better. Apart from the missing Annexe VIII, the only gripes are the lack of page numbering, and the Habsheim airfield plate being just as blurred as the PDF. The shambolic "Additif" section is not needed. I'm wondering if the PDF might have been created from an early draft of the Report.

[Chris Scott]

Quote from Dozy Wannabe (http://www.pprune.org/tech-log/528034-habsheim-21.html#post8288704):
"Agreed, but given their fast-and-high state, wouldn't those turns [to adjust track for Rwy 02] have provided at least some opportunity to slow down and thus manage the descent more precisely, whereas a straight-in to the 34 grass strip make slowing down and managing altitude precisely more difficult?"

As I wrote, there was little scope to increase the track distance from 3.5 nm. A sharp S-turn starting at 200 kts would have been impossible. So, for reasons I have already explained, the switch of runway made no significant difference. Ultimately, the essence of a descent for a direct approach is energy management, particularly on a jet, and energy was too high throughout.

The A320 was/is a conventional, medium-sized jet, perhaps marginally more slippery than a B737. The FBW computerisation makes no difference whatsoever to the mechanics of manoeuvrability or descent management. Ideally (and with a degree of hindsight), the PF should not have accelerated to 200 kt in transit. Having sensibly retained the T/O flap setting of 1+F, he could have maintained 'S' speed (about 182 kt at that weight). In need of drag, the L/G could have been extended about 10 seconds earlier, as soon as the field was in sight, and Flaps 2 simultaneously. But their workload was high, and continuing straight in (rather than joining the airfield circuit) was one of the decisions that they may have been pondering in the years since.

[AlphaZuluRomeo]

Dozy (re: #433) Thanks :) Grouches may point (and be right) that it's still missing in the "official" report (at least in the PDF version). I suggest the idea to go to a library to check a (really) official paper copy would do, for those interested.
(and remember: don't feed...)

The CVR transcript shows "Ca roule 296 Q" (probably spoken as "Ca roule deux-cent quatre-vingt-seize quebec"?)
Yes, that sounds right for a native :)

[CONF iture]

The contents of several of the Annexes do not all correspond to the index. The chaos includes the fact that (as you say, Noske) Annexe VII is listed in the index as "Correlation...", presumably a narrative justification of the method used to synchronise the timings of CVR, DFDR, Basel ATC, and Habsheim ATC. It is missing from that PDF.
Chaos is the appropriate designation, engineered chaos. Confusion all the way. They don't want you to have it easy and simple.

Annexe 5 - Every single transmission between Basel ATC and ACF296Q has a timestamp but no FDR listing is provided for the period.
Annexe 5 - Only a few transmission between Habsheim ATC and ACF296Q have a timestamp for the precise period a FDR listing is published.

Everything is done to NOT properly correlate ATC CVR FDR for the all flight.
No wonder Annexe 7 there is none.
But take a part of Annexe 8 to label it 7 ... why not.

[Chris Scott]

Linktrained,
"Habsheim was 25 years ago but still some of the lessons are there to be learned by newbies (as well as by one 86 year old !)"

Yes, it's said that when you stop learning, it's time to stop. Some of us are still learning, even though we already did...

"Today might be different !"

A little, but the basics are the same - only people's perceptions and expectations have changed. The air remains as fickle and fascinating as ever.
Check your PMs?

[Natstrackalpha]

rudderrudderrat


Quote
It appears to me that the aircraft was deliberately flown below 50 RA with the speed allowed to wash off, apparently waiting for Alpha Floor to "save the day". Unquote


I think that was precisely it. Everyone seems to have missed it - no mention from Chris Scott about it either.


He forgot below 100 feet.


He most likely applied full back press on the stick, waiting, as you said for Alpha Floor to kick in - it did not, aircraft THOUGHT IT WAS LANDING - too.


No alpha floor, no toga to boot. = No noisy climb out.


Had he have done all the above at >100` RA then we would not be having this thread.

[roulishollandais]

the aircraft was deliberately flown below 50 RA with the speed allowed to wash off,
He was flown below 50 RA when he was over some unexpected trees BEFORE the runway.
That has been pointed by the expert Max Venet at the trial,. He asked too to cut these trees computing with a patented Venet software the path (one addition or one substraction for each tree ($$$$...).
But these trees were unknown from the pilots who previewed to fly over the other runway. I studied the VAL chart and discovered that the grass ruway was already illegal for light single engine airclub planes... Today that runway no more exists...
Question is : did Asseline DELIBERATELY ovefly these seldom trees? Really I don't think it.

[Chris Scott]

Bon soir roulis,

The copse ("bosquet") that they overflew, causing the two-second ramp of radio altimeter readings, was just outside the airfield boundary, and roughly south-east of the Rwy 02 threshold. At that time, they were in a gentle right turn (about 10 deg of bank) to line up with the grass Rwy 34R.

The current satellite map shows a fairly large wooded area south-east of the airfield, and (assuming it already existed in 1988) they probably overflew the west corner of it. As they had not been briefed of the existence of the wood at the north boundary, it seems unlikely that they would have known of the wood to the south-east.

The two readings indicating the "copse" were 32 ft at TGEN 313 (t -21 seconds) and 24 ft at TGEN 314 (t -20 sec). The reading recorded at the next second was 60 ft, and the a/c did not level off until about t -13 sec, by which time I estimate it had passed abeam the Rwy 34L grass threshold, and was just short of the Rwy 34R grass threshold.

At level-off (TGEN 321, or t -13 sec), the rad alt showed a height of 40 ft. At the same time, the pressure altitude was +869 ft, from which I calculate the pilots' altimeters (set to the QFE of 984 hPa) would have been indicating about 60 ft.

[DozyWannabe]

Quote
It appears to me that the aircraft was deliberately flown below 50 RA with the speed allowed to wash off, apparently waiting for Alpha Floor to "save the day". Unquote

I think that was precisely it. Everyone seems to have missed it - no mention from Chris Scott about it either.

It wasn't missed. It was rebutted because *taking Alpha Floor out of the equation was a central aspect of the manoeuvre being performed*. That the action to do so (holding down A/THR disconnect) was erroneously omitted - probably as a result of task saturation - doesn't alter the fact that the Captain should not have been expecting Alpha Floor at any point.

Regarding the online version of the report, I made the following observations:


Looking at the PDF properties, the online PDF version was created on the 10th June 2008 and last modified (presumably uploaded) on the 12th
The PDF Producer field lists "Acrobat Distiller 8" for Windows. Acrobat Distiller is usually used for converting existing digital documents in the PostScript format to PDF
This therefore implies that the document was scanned some time before it was converted for the web. It's entirely possible that the original PostScript file was scanned years prior to conversion and uploading
A random sampling of documents from the site of a similar vintage shows that the conversion and upload was performed in bulk batches
The same random sampling indicates that the documents of similar vintage are of similar fidelity to the Habsheim report, including a similar number of scanning errors
It should also be noted that (for example) online NTSB reports of a similar vintage are of similar fidelity, being image-centric PDFs and equally troublesome to extract textual data from electronically


The BEA's search form indicates that there are currently 2,256 individual documents in their web-accessible archive. It seems reasonable to infer that to proof-read and correct each of these documents individually would be a mammoth task in terms of man-hours and expense. That the accessible reports from the NTSB and UK AAIB of a similar vintage are also of a similar quality indicates that these archives are supplied "as-is", and some scanning/conversion errors are to be expected.

Furthermore, the appendices/"annexes" of such reports are intended for use only as reference data for the reports themselves. They are not intended for use as the basis of amateur investigations, nor should their content be criticised by amateur investigators for being unable to suit their specific needs.

It must be borne in mind that when this version of the report was published in April 1990, the web existed only as a paper proposal at CERN. The report itself is obviously aimed purely at the print medium, and the CVR transcript could be condensed via word processing in a way that the DFDR output could not. That the DFDR output included in the appendices deals with the immediate precursors of the accident sequence only, and earlier analysis is provided in the narrative makes sense if one takes this into account.

There is absolutely no evidence to suggest that this particular report, either in general or with specific regard to the archival version available online, was intentionally rendered difficult to interpret - it is simply an artifact of its time.

@roulishollandais - in addition to Chris Scott's answer above, it's worth reiterating that overflying the copse of trees made no difference to the behaviour of the aircraft.

[Chris Scott]

Quote from rudderrudderrat:
"It appears to me that the aircraft was deliberately flown below 50 RA with the speed allowed to wash off, apparently waiting for Alpha Floor to 'save the day'. "
Response from Natstrackalpha:
"I think that was precisely it. Everyone seems to have missed it - no mention from Chris Scott about it either."

Welcome, Natstrackalpha. For Airbus FBW pilots and long-term readers of Habsheim discussions, Alpha-Floor inhibition below 100R (100 ft on the radio-altimeter) is a given. The trigger for this thread occurred on AF447 Thread 11, in which we were briefly discussing the inhibition criteria. I rashly cited Habsheim, and CONF_iture had to remind me that the captain had clearly briefed that he would inhibit (disable) Alpha Floor manually to cater for his plan to achieve a stabilised Alpha-max at 100ft on the QFE - at which the height on the rad-alt might be higher or lower than 100R. My reply (http://www.pprune.org/tech-log/528034-habsheim.html#post8159637) was copied to form the first post of this Habsheim thread, which was instigated by CONF_iture (see his post #2). Re Alpha-Floor inhibition, I refer you to my first and last paragraphs.

Regarding the flypast as executed, you will have noted that it was flown well below 100R, and (as Dozy Wannabe writes above) it is not entirely clear whether the captain ever carried out his briefed manual inhibition of Alpha Floor.

It is not unreasonable to speculate that, notwithstanding the briefing recorded on the CVR, Capt Asseline may have planned to fly below 100ft all along. This was certainly not incompetence, because his height-keeping after level-off at t -13 (see my post to roulishollandais above) was remarkably accurate – whether you study the baro readings, or the heights recorded from the rad-alt. Although he did lose height, it amounted to no more than 10 ft on his baro-altimeter, and the same on the rad-alt. He has stated that he was using only his baro altimeter throughout, and claims that it must have been giving false readings, but – other than his own testimony – I’m not aware of any evidence to support that.

You argue that, if they had flown the flypast above 100R, Alpha Floor would have prevented the accident. But at that height, they would have cleared the treetops anyway - provided TOGA had been selected manually in time to stop the inevitable sink with flight-idle at Alpha-max.

BTW, how's the conversion going?

[DozyWannabe]

It is not unreasonable to speculate that, notwithstanding the briefing recorded on the CVR, Capt Asseline may have planned to fly below 100ft all along.

I'd say that's probably unlikely, given that such intent would have constituted a deliberate breach of AF's rules regarding display flights - specifically a "hard deck" of 100ft. Notwithstanding the somewhat contradictory DGAC rule specifying 500ft, Asseline would have rightly expected to be hung out to dry if it could be proven he intended to go any lower than the regs allowed for.

Regarding competency, I think it would have been impossible for a pilot to be selected for that role without being a very competent hand-flyer, and you're right about his holding the altitude fairly stable. However on this occasion it was his decision-making that fell short - a consequence of which was substandard thrust and speed management.

As you've said before, there was ample opportunity to circle, descend and then achieve a more stable short final - I believe that the old "press-on-itis" psychology had a hand in their not doing so.

[gums]

From the "light" peanut gallery @ Doze and Chris....

Yep, Doze, "pride" gets in the way one or two times, but usually only once. Our aviation mistress is harsh in that regard.

The "press on" attitude may apply in combat, but otherwise it is not a smart thing regardless of how good your "hands" are. I know about the combat reference from one mission, but I digress.

Let's face it, the dude had a poor setup and tried to "save" it. In any other plane at that time, he would likely have stalled and crashed right in front of everybody there. The "magic" protections and limiters can only do so much, and flying at the "limits" when carrying SLF and such is not very "professional" to this old pilot.

[Chris Scott]

Dozy,

Are you arguing that he levelled off 40 ft lower than he intended, despite that being a potentially-hangable offence, and did absolutely nothing about it for nearly 10 seconds?

Or are you suggesting his barometric altimeter was over-reading by at least 40 ft, due to an error in an ADC or DMC?

[noske]

At level-off (TGEN 321, or t -13 sec), the rad alt showed a height of 40 ft. At the same time, the pressure altitude was +869 ft, from which I calculate the pilots' altimeters (set to the QFE of 984 hPa) would have been indicating about 60 ft.
It's kind of annoying that the report leaves those calculations as an exercise to the reader (I used this Pressure Altitude Calculator (http://www.srh.noaa.gov/epz/?n=wxcalc_pressurealtitude)), but yes, it shows that the QFE value received from Habsheim must have been correct.

Capt Asseline may have planned to fly below 100ft all along. This was certainly not incompetence, because his height-keeping after level-off at t -13 (see my post to roulishollandais above) was remarkably accurate – whether you study the baro readings, or the heights recorded from the rad-alt. Although he did lose height, it amounted to no more than 10 ft on his baro-altimeter, and the same on the rad-alt. He has stated that he was using only his baro altimeter throughout, and claims that it must have been giving false readings, but – other than his own testimony – I’m not aware of any evidence to support that.But Asseline seems honestly upset about the investigators not believing him ("Do they think that both of us are too stupid to read an altimeter?").

To reconcile the DFDR values with an indication of 100 ft on the captain's PFD, one could assume that the crew had erroneously entered a QFE value of 985.5. Now that seems totally unlikely. On the CVR, they repeat the correct value three times (and "cinq" and "quatre" don't even sound similar).

Asseline has another explanation, that the reference value for his altimeter must have shifted somehow, and he has indeed a hair-raising story to tell, where this happened to him on a flight from Paris to Geneva, and he ended up 1000 ft too low, almost (as in "too low, terrain") striking some mountains.

While I have some sympathy for this particular theory of his, it still leads back to the basic problem: the flight crew being unfamiliar with the airfield. Had they been aware that Habsheim control tower is not a good 100-ft-reference, and had they known about these trees that are just about as tall as the tower, it wouldn't have mattered if their altimeter reading was off by a lousy 40 ft.

[DozyWannabe]

From the "light" peanut gallery...

Hah. Don't be daft, you're always welcome and your input is always thought-provoking in a good way.

..."pride" gets in the way one or two times, but usually only once. Our aviation mistress is harsh in that regard.

No doubt, though on occasion it can be about other things - for example, a tendency to stick to the plan even though some aspects have already gone awry.

The "press on" attitude may apply in combat, but otherwise it is not a smart thing ... I know about the combat reference from one mission, but I digress.

An interesting point and one with which I'd agree, with the caveat that in today's cut-throat business world I'd be surprised if most airline captains haven't been concerned at one time or another about the potential fallout from disgruntled pax complaining to the firm when deciding whether to go-around and/or divert. With airline management consisting mostly of MBAs, many of whom don't appreciate the risks involved, the worry that making a safe call could harm or even end your career might be palpable.

In this case though, I'd be surprised if the schedule wasn't more flexible than the majority of airline ops.

Let's face it, the dude had a poor setup and tried to "save" it.

Yup, and the 'setup' was poor all the way back to the airline's briefing. However, at the same time there were opportunities to make a more stable approach which were not taken, so some responsibility for the need to "save" the aircraft remains with the captain.

The "magic" protections and limiters can only do so much, and flying at the "limits" when carrying SLF and such is not very "professional" to this old pilot.

As you well know from your FBW fighter experience, such systems are a result of hard-nosed engineering practice - far from "magic" on closer inspection! Whatever one's views on the technology, the laws of physics remain immutable and if you can't avoid solid objects in your path you are going to hit them...

One of the outcomes was a move to ban pax from the general public on display flights.

Are you arguing that he levelled off 40 ft lower than he intended, despite that being a potentially-hangable offence, and did absolutely nothing about it for nearly 10 seconds?

Or are you suggesting his barometric altimeter was over-reading by at least 40 ft, due to an error in an ADC or DMC?

I'm not arguing or suggesting anything, because it's outside my area of expertise!

I think it's not a simplistic either/or proposition though. It could be that he moved to level off at 100ft, but the inertia of the rapid, improvised descent profile coupled with an abnormally low thrust setting caused the aircraft to continue to sink while he was heads-up, looking out of the windscreen. The rapid increase in workload due in part to continued improvisation could have lead to confirmation bias and misreading of the altimeter when he looked heads-down again. I'm not saying that's what happened as much as asking you if you think it plausible.

But Asseline seems honestly upset about the investigators not believing him ("Do they think that both of us are too stupid to read an altimeter?").

And in my personal view, that - namely taking things personally and reflexively defending with counter-attacks - has always been his Achilles' heel when it comes to the credibility of some of his claims. For their part, the investigators certainly believed that *he* believed the baro alt was showing 100ft - but every bit of hard evidence staring them in the face made it unlikely.

Interesting use of "both of us" as well. To the best of my knowledge, the FO has never spoken up publicly on the subject. I may try hunting for material there.

Asseline has another explanation, that the reference value for his altimeter must have shifted somehow, and he has indeed a hair-raising story to tell, where this happened to him on a flight from Paris to Geneva, and he ended up 1000 ft too low, almost (as in "too low, terrain") striking some mountains.

Which rather begs the question of why, that issue being known to and experienced by him personally, he neither used the RA display as a backup himself, nor - given his claim that the RA display was difficult to read - asked his FO, as pilot monitoring, to do his best to cross-check both baro and radalt.

While I have some sympathy for this particular theory of his, it still leads back to the basic problem: the flight crew being unfamiliar with the airfield.

It's one fundamental issue certainly, but there seems to be a general sense of creeping complacency about the operation as a whole.

20/20 hindsight is a wonderful thing.

[roulishollandais]

So we do agree about the passage of more than 50 FT RA to the unexpected less than 30 FT RA before the runway threshold.

Being particularly concerned in flying systems as a former manager of mathematics and computational methods, the comment of the expert Max Venet indicates to me a lack of computational methods. From memory, he said the following sentence: "We do not really know why the aircraft did not set in flare passing below thirty FT RA, or how long he should stay less than thirty feet. "
This sentence shows that Venet did not found in Airbus a person who could tell him that or a document with this information. He could not access either to a history of the evolution of the software before its certification, and he did not find the information in the certification documents.

I believe that these serious gaps do not correspond to the state of the art computer at that time (or today!). If indeed at that time certain documents required to decode a specific 0 and 1 know-how it was even more to accompany these documents and comments in everyday language. In any case it is so that AS or AB (or any other designer of the system) should proceed.

The absence of such documents can only be accompanied by operational documents being incomplete and unsafe to the user (airlines, pilots, mechanics).

I say that as a computer specialist who had to put up this kind of working methods among researchers and engineers. And it had nothing difficult, but it is of course a constraint.

I do not see the point either to keep secret the software description. The fight against the copy is much easier to exercise if everyone puts their workon the table in daylight in a reasonable time. Only that can serve as engineers who wish to sell their know-how and pilots who wish to return alive from flight when they flew as we taught them

[DozyWannabe]

Being particularly concerned in flying systems as a former manager of mathematics and computational methods, the comment of the expert Max Venet indicates to me a lack of computational methods. From memory, he said the following sentence: "We do not really know why the aircraft did not set in flare passing below thirty FT RA, or how long he should stay less than thirty feet. "

We've covered a lot of this ground before, and I believe you had a look at my old Prof's overview from his 1993 visit, where he saw use of advanced software engineering and reliability techniques in person.

Would this be the same Max Venet who was President of the SNPL from 1979 to 1982? If so, in which area of expertise was he qualified to comment?

This sentence shows that Venet did not found in Airbus a person who could tell him that or a document with this information.

No it doesn't, it shows that he didn't have the information - whether this was because he didn't try hard enough to get it or because Airbus were wary of supplying (at the time) very commercially-sensitive data to a representative of a hostile organisation is unknown.

The absence of such documents can only be accompanied by operational documents being incomplete and unsafe to the user (airlines, pilots, mechanics).

M. Venet stating that he didn't have those documents is not the same thing as those documents not existing. Again, we're back to Prof. Mellor's visit where he saw contemporary best practice in action.

I say that as a computer specialist who had to put up this kind of working methods among researchers and engineers.
...
I do not see the point either to keep secret the software description. The fight against the copy is much easier to exercise if everyone puts their workon the table in daylight in a reasonable time.

And I respond as a professional software engineer and aviation nerd who had the good fortune to be educated by someone who saw behind the curtain.

These days it's clear to those with an interest that some of the techniques pioneered with the Airbus FBW systems have percolated down into more mainstream software practice (test-driven development and regression testing in particular). The kind of source-code sharing you describe has also become well-known, spawning the Open Source movement. However in 1988 the legalities surrounding software copyright and ownership were still something of a grey area, and the people dictating commercial policy guarded all aspects ferociously.

Getting back to the point, we've already discussed that Capt. Bechet organised an accurate real-world replication (minus the trees that were hit) of the flight at Toulouse. Flare mode was not activated, but Alpha Prot was.

Based on my understanding of how the systems were designed, it's likely the reason Flare Mode was not activated was because the brief RA reductions were not consistent with the overall trend of the values. M. Venet appears to have been working on the assumption that a value of 30ft RA will immediately trigger Flare Mode under all circumstances. This in turn suggests an overly simplistic understanding of software in general and indicates that his expertise was not in software, real-time or otherwise.

If Flare Mode had been (completely - h/t HN39) activated by overflying the copse, the DFDR would have indicated the elevators down shortly after crossing the strip threshold. Instead, the EFCS complied with the sidestick commands until shortly before impact, where Alpha Prot mediated the pitch-up demand to avoid stall.

And as rudderrudderrat said all the way back in post #45, Flare Mode can be overridden at any point:
The baulked landing procedure (Select TOGA power and rotate to 10 degs pitch) allows you to fly away from any attempted landing - even after touch down.

[noske]

Interesting use of "both of us" as well. To the best of my knowledge, the FO has never spoken up publicly on the subject. I may try hunting for material there.
I was referring to Asseline's book, p. 35/36. Scattered all over that book, there are a few passages where I can imagine Mazières putting down his copy, muttering "Will you please just speak for yourself, Michel?", and this is one of them. But all in all, Asseline was reasonably cautious about that issue.

[HazelNuts39]

From the FCTM: "When reaching 50 ft, auto-trim ceases and the pitch law is modified to flare law.
(...) The system memorizes the attitude at 50 ft, and that attitude becomes the initial reference for pitch attitude control. As the aircraft descends through 30 ft, the system begins to reduce the pitch attitude at a predetermined rate of 2 ° down in 8 s."

This aligns with paragraph 1.16.1.2 of the accident report where it says:
"Pendant la phase de descente entre 50 et 30 pieds, la loi de pilotage est modifiée et prend progressivement en compte (1), au lieu du terme en facteur de charge, un terme en assiette longitudinale (écart entre l'assiette constatée au passage à 50 pieds dite assiette de consigne et l'assiette réelle instantanée) ; Au passage à 30 pieds, un ordre supplémentaire destiné à simuler un effet de sol conventionnel est introduit en complément à la loi de pilotage en assiette précédente : cet ordre ramène progressivement l'assiette de consigne (constatée à 50 pieds) à - 2° en 8 secondes, créant un moment à piquer que le pilote doit contrer (« dérotation »)".


According to the footnote to this paragraph, the transition to flare law is progressive and takes one second to complete. In the same paragraph the report states that flare law (loi de pilotage en assiette) was active between t-21s and t-20s (i.e. when the airplane was overflying the 'bosquet').

RA, sidestick and elevator position are recorded once per second. Even if 'loi de dérotation' was momentarily activated, the time below 30 ft was probably too short (considering the ramps of 1 second and 8 seconds) to result in a movement of the elevator that would have been observable in the DFDR data.

[DozyWannabe]

@HN39 - That's interesting. So the transition began, but the 'drop' below 30ft was probably too brief to complete it? OK - that makes sense (and is in keeping with the 'trends and deltas' nature of the systems as I was taught).

Sometimes I really wish I'd taken French alongside German - it'd make this so much easier. It also means that the question of Flare Mode was in fact asked and answered in the report, which somewhat dents Asseline's assertion that they didn't check for it/someone was hiding something.

[Chris Scott]

Hi HN39,

Thanks for recapping the information from the post-1988 FCTM.

My English-language FCOM from 1988 is in agreement that the first stage of what it calls Landing Mode involves two things passing 50R (50 ft on the radio altimeter): memorising of the pitch attitude; and freezing of the THS setting.

The BEA report indicates that this first stage would have been engaged as a result of the ramp of height to 32R at t -21 (TGEN 313), as the a/c passed over the copse ("bosquet") before the airfield boundary.

As you know, the second stage of what is now called Flare Mode, which begins crossing 30R, introduces a slow bias of pitch that modifies the zero-stick pitch progressively from what it was at 50R to minus-2 degrees - this during a period of 8 seconds. That sub-mode would probably have been engaged by the recorded height of 24R at t -20.

The BEA states that, because the rad-alt height in the following 2 seconds was over 50 ft, flare mode was disengaged back to Normal Law for that period. AFAIK, there is nothing in the recorded data to suggest otherwise.

By t -17, however, the height had once again fallen below 50R, which would have re-engaged the first stage of Flare Mode, freezing the THS at its current setting of U04 (4 deg nose-up trim). The second stage of Flare Mode would never have engaged, because TOGA thrust had been selected by t -4 (see Tome 4), at which point the height had not crossed 30R. (A second later, the AoA exceeded 14.5 deg, and Alpha-Prot Law would have engaged.)

In summary, the only adverse effect of the regime of the first stage of Flare Mode between t -17 and t -5 would have been the freezing of the THS at 4 deg nose-up trim. After levelling off at t -13, the PF had to pull back slightly and increasingly on the sidestick as the a/c decelerated. However, in my experience, pilots of a/c with conventional flight controls flying deliberate approaches to the stall do not continue pitch-trimming below Vref. At level-off, the IAS was already below the Vref for Flaps 3.

Bon soir, roulishollandais,

Does the above interpretation address any of your concerns?

[roulishollandais]

Sometimes I really wish I'd taken French alongside German - it'd make this so much easier. Sometimes Iself really wish I'd taken English alongside German as first foreign language - it'd make this so much easier.
Specialy here, where we are walking on eggs.
I shall take some time to write my answer in English.
Already Thank you to Hazelnuts39 and Chris Scott for analysing these sequence over the bosquet, to gums for global analyse of the flight, to Dozy for the interest he finds now in that forum who could never take place since 25 years, for any reason. Thank you to Conf_iture who shows us that not retired pilots are not all ectoplasms, to John who allowed that forum to exis, and to all the excellent contributors, and to the followers.

My last posts don't want to comment the crash itself (I removed my pilot's cap) but what did me wonder as computer guy, as it may be a chapter of factors in the lessons to learn from Habsheim. But we need a good analysis of the facts.

Just a word about Venet :
He did the mentioned declaration during the first Trial, answering to the President of the Court (dec 1997) . When he says "We" it means himself and Belotti.
Retired AF airline pilot, and former (elected by the members) President of the SNPL the most important pilots Union in France. That does not make him the Devil! Pilots are never on the ground... and sleep (after night flight) when ground People are working hard in Office, etc. We know that old story between Navigants and Rampants So pilots need to be represented in Unions when they are not on the ground, despite someone don't like existence of Unions.

Four enquiries have been done :
1. 27. june 1988 BEA : President Bechet, Gonin, Auffret, Farthouat, Arslanian, Gourguechon, Villeneuve
2. Judiciary expertise from the Juge d'instruction Mrs Marchiani : Auffray and Bourgeois
3. 11.january1990, from Mrs Marchiani, two supplementary experts : Venet and Belotti
4. 28. june 1988, Air France investigation Commission :
Richer the President, Vuillard, Gauthier, Merloz, Bourbon, Conrozier and prof. Org. (Chatelain, Scattolin, Thiébaut, Brédy, Edelbloude)

[DozyWannabe]

Just a word about Venet :
...
Retired AF airline pilot, and former (elected by the members) President of the SNPL the most important pilots Union in France. That does not make him the Devil!

Didn't say it did. However he is not an expert in the field of computing, and the statement showed a lack of understanding of how such a feature is implemented in terms of real-time safety-critical software disciplines.

The issue with the SNPL regarding any action against Airbus is that the union has held a long-standing animosity against the company ever since the A300 went into service as the first long-haul widebody to do away with the Flight Engineer's station. An evolution of design which their members fought tooth-and-nail, and seemingly never forgave.

As such, there will be an understandable suspicion of bias against Airbus in such representations. In Asseline's case, it seems his union-supplied legal representatives steered him in the direction of pursuing claims of technical issues against the aircraft despite a lack of supporting evidence, when he would have been better served by pursuing the airline's poor operations practices which put him in a bad position in the first place, and for which there was significant evidence and potentially a much stronger case when it came to mitigating factors.

I have to say my personal opinion is that all this political nonsense has served only to hinder understanding of the incident and the issues involved, and furthermore has led to a significant number of people developing entrenched opinions that are, in some cases, still held with an almost religious fervour. This in turn means that there's a degree of personal animosity and ad hominem to the discussion that saddens me immensely.

[gums]

Well, we get to replay the incident over and over......

Before I return to the "monitor mode", I think Doze summarized the whole thing, and it wasn't "flare mode" or altimeter settings or anything else except:

I think it's not a simplistic either/or proposition though. It could be that he moved to level off at 100ft, but the inertia of the rapid, improvised descent profile coupled with an abnormally low thrust setting caused the aircraft to continue to sink while he was heads-up, looking out of the windscreen. The rapid increase in workload due in part to continued improvisation could have lead to confirmation bias and misreading of the altimeter when he looked heads-down again. I'm not saying that's what happened as much as asking you if you think it plausible.I have personally lost friends that did stupid things like "pressing" when they should have done a "go around" or broken off the attack ( fighter stuff). Pride is a player, and "get the job done" is also there on most of the losses. Then there is loss of situational awareness and basic failure to use all the displayed data of your flight path and such. On the civilian side, I knew the Cali pilot and the guy at Little Rock ( not like Cali, but still sad to lose a student of mine). Didn't cross paths with Sully, but might have, and he did it right. I got tired of going to memorial services my last two years in the "business" - 4 pilots that were of my age and experience that screwed up, and were all pilot error.

Sorry for the personal stuff.

Let's face it, the accident/crash happened due to poor judgement, and the jet did everything it was designed to do. Not enough to compensate for a sorry execution of a simple flyby, but you play the cards you are dealt.

I am not sure about the purpose of a low, slow flyby at a ridiculous altitude for a big jet and at a high AoA. But the captain tried it. Poor planning and practice. Then getting behind the jet's capabilities. Sad, sad.


I relish the info on the flight control laws and such. Also like the insight I see from professional "heavy" pilots and a few engineers.

[DozyWannabe]

Sorry for the personal stuff.

Don't be - it's all good. There's no shortage of examples where otherwise excellent pilots paid dearly for a single call that turned out to be poor with 20/20 hindsight, and it goes without saying I'm sorry for your loss.

I am not sure about the purpose of a low, slow flyby at a ridiculous altitude for a big jet and at a high AoA.

Because no other jet of that kind at the time could do it safely. I have little doubt that had the crew been properly briefed on the terrain surrounding the airfield we wouldn't be having this discussion.

[Machinbird]

Quote:
I am not sure about the purpose of a low, slow flyby at a ridiculous altitude for a big jet and at a high AoA.
Because no other jet of that kind at the time could do it safely. I have little doubt that had the crew been properly briefed on the terrain surrounding the airfield we wouldn't be having this discussion. Without A/T engaged and Alpha Floor, the A320 couldn't do it safely either.
So what were they thinking?

[DozyWannabe]

Without A/T engaged and Alpha Floor, the A320 couldn't do it safely either.

Surely it could do that perfectly safely without either as long as there were no obstacles in the way! A caveat which was firmly in the hands of the operator and crew...

(referring strictly to the aircraft capabilities, not the advisability of the sortie itself)

[HazelNuts39]

Surely it could do that perfectly safely without either as long as there were no obstacles in the way!I concur with Machinbird. The demonstration as planned: a fly-past at Valphamax and alpha-max, therefore with thrust control the only means for controlling the height of the flight path, could not be done safely at 100 ft above terrain, let alone 40 ft. In any case it could not have been done with the level of safety required for a passenger-carrying flight.

Just imagine: To achieve alpha-max you have to pull the sidestick to the rear stop and keep it there. To maintain level flight at alpha-max you have to slow down to Valphamax. The only way to 'pull up' in that situation is to increase thrust to accelerate to a speed greater than Valphamax.

[gums]

Thanks 'bird and Nuts

We pilots must remember that rate of climb is the delta between thrust available and thrust required for the AoA and weight and...... ( power required)

It's why the nasal radiators tend to use throttle for glide path rate of descent to land on a boat. They trim for an AoA/speed and control the flight path with power. Just look as the Asiana fiasco, much less this one we are re-hashing.

Besides going well below the planned altitude, and a screwed up approach, the power side of the equation was not even close to what was required to "save" the screwed up flyby. Sheesh. I don't like being harsh about a fellow pilot, but this one really bugs me, and I am thankful that so many folks walked away from it.

Lastly, who in the world on the sidelines would have appreciated a high AoA, slow speed, extremely low altitude flyby? I'll bet 99% of the witnesses didn't have a clue about what the pilot was trying to demonstrate.

[HazelNuts39]

Lastly, who in the world on the sidelines would have appreciated a high AoA, slow speed, extremely low altitude flyby? I'll bet 99% of the witnesses didn't have a clue about what the pilot was trying to demonstrate.I still have a vivid memory of the demonstration at the Farnborough airshow of the high angle-of-attack protection by Gordon Corps on the A300 serving as a flying testbed for the FBW system developed for the A320. The speaker didn't inform the audience about the altitude observed in the demonstration but, from memory, it must have been closer to 1000 ft than to 100 ft.

[DozyWannabe]

As I said,

(referring strictly to the aircraft capabilities, not the advisability of the sortie itself)

Theoretically speaking, at 100ft over a flat landscape and with the power set correctly the A320 would be capable of safely maintaining Alpha Max indefinitely until it ran out of fuel.

With 20/20 hindsight the sortie shaved safety margins to the bone, but the responsibility for that didn't lay with the aircraft.

[Linktrained]

Farnborough had NO trees exceeding 500 ft. ( QFE ) along the extended centre line of the R/Ws ( IIRC ).
THAT factor alone, would have made a lot of difference for Gordon Corps' demonstration, if done at 1000ft.

[HazelNuts39]

... with the power set correctly the A320 would be capable of safely maintaining Alpha Max indefinitelySure, but with the longitudinal control at the stop it doesn't have the controllability required to maintain 100 ft terrain clearance safely in anything but an absolutely 'dead' atmosphere.

[Chris Scott]

Quote from Dozy Wannabe (my emphasis):
"Theoretically speaking, at 100ft over a flat landscape and with the power set correctly the A320 would be capable of safely maintaining Alpha Max indefinitely until it ran out of fuel."

Theoretically, perhaps... Let's try to examine that in practice.

Quote from gums:
"...the nasal radiators [naval aviators] tend to use throttle for glide path rate of descent to land on a boat. They trim for an AoA/speed and control the flight path with power."

That's rather like most of us are taught to do before our first ever solo-flight on a light, propeller-driven aeroplane - albeit without an AoA gauge. We maintain the correct approach speed by adjusting the pitch-attitude, and adjust rate/angle of descent with throttle/power. (With a tractor-propeller, admittedly, the varying slipstream over the wing itself affects the wing lift slightly.)

Later, in jets, we are taught to adjust the rate/angle of descent with pitch-attitude, while maintaining the correct approach speed with power (thrust) - again without an AoA gauge. The two techniques are, however, essentially the same - achieving a suitable, stable AoA, while varying the flight-path angle (FPA) through the air. Although I've cited the approach case, it's worth pointing out that - in either technique - the FPA does not have to be negative.

How does this relate to Habsheim?

Quote from HN39:
"To achieve alpha-max you have to pull the sidestick to the rear stop and keep it there. To maintain level flight at alpha-max you have to slow down to Valphamax. The only way to 'pull up' in that situation is to increase thrust to accelerate to a speed greater than Valphamax."

So it's clear that the Habsheim scenario - once the stick is fully back, and the AoA is at alpha-max - is critically different from an approach at Vref. In fact, any change of the FPA through the air, as HN39 points out, becomes completely dependent on IAS (strictly speaking, CAS). Any change of FPA requires a change in "normal" acceleration (Nz). In a steady airstream, if that FPA change is to be upwards, the necessary increase in lift can only be achieved by an increase of IAS. That, in turn, can only be achieved by increasing the thrust. On the other hand, descent is not a problem...

Clearly, low flight at alpha-max is hazardous, for at least one reason: wind-shear, which is inevitable over and near trees and buildings, unless the wind at all levels is flat-calm (a rare event). Any loss of headwind or increase of tailwind leads to a loss of IAS. Recovery of IAS requires an increase in GS, i.e., kinetic energy. With no surplus of potential energy to convert, that increase in kinetic can only be supplied by an increase in thrust.

A jet engine will respond fairly quickly when already at a medium-to-high thrust setting (hopefully both engines, symmetrically...), but the pilot's reaction time also has to be taken into account. Arresting the inevitable descent caused by any loss of IAS would take time. There is little or no height to trade for IAS recovery.

To quote again from HN39:
"...with the longitudinal control at the stop it doesn't have the controllability required to maintain 100 ft terrain clearance safely."

[gums]

C'mon, Chris.

The potential rate of climb is a simple function of power available versus power required for your configuration.

My light plane training was the pitch attitude and then power for speed. In later planes with constant speed props, this was even better. My first jets were straight wings and the same applied. Then I went to the high aspect ratio bent wings and things were different. No FBW, so trim was still AoA, or speed for those not used to the great indication of your actual lift capabilities and stall.

So the dude at the flyby needed more power to get the jet climbing, even at max AoA that the control laws allowed. A "conventional" system would have allowed him to pull further back on the stick and then get way behind the "power curve". You know, the area of reverse command.

Let's face it. Wasn't flare mode or Alpha protect or any of the limiters. It was low power setting and a poorly executed demo. And then the 6 "pees".

[Chris Scott]

Hello gums,
"The potential rate of climb is a simple function of power available versus power required for your configuration."

Agreed! I was addressing the problems of flying level and VERY low, stabilised at alpha-max - as they intended, but never achieved. I think that's what HN39 also had in mind.

The TOGA go-around is another matter, although they didn't initiate it in time. Pity (the seventh P)...

[gums]

Yeah, Chris, we are on the same page.

I did not mean to challenge techniques and such, just point out that "power" was the problem at the flyby, and not the flight control laws and such.

[Machinbird]

Back at the dawn of the commercial jet age, when piston pilots were transitioning to jets, there was a rather horrific accident caused by a poorly planned approach and failure to allow sufficient time to spool the engines up. United Airlines Flight 227 - Wikipedia, the free encyclopedia (http://en.wikipedia.org/wiki/United_Airlines_Flight_227) The Captain had 334 hours of jet time at the time of the accident having transitioned from the DC-6. Again, this was early in the jet age, and many of the safety initiatives that you might consider standard had not yet been developed.

Since I was then undergoing basic jet training and was soon to go to the boat, I understood the reasons for the accident and it made an impression on me that still remains. You do not ever want to be caught near the ground on the wrong side of the jet engine acceleration curve. Too bad that lesson was forgotten/not known by the Habsheim crew.

More recently we have had a couple of accidents that resulted from pilots expecting A/T to manage their thrust while they flew the nose of their aircraft. When the A/T did something unexpected, they were caught with insufficient time to spool up before smiting the ground.

By this point in the jet age, pilots should universally understand that jet engines have a significant spool up time from idle (although it is much better now than when I was carrier qualifying in aircraft powered by a centrifugal flow jet engines and early axial flow engines.)

Since we seem to design modern cockpits where you must stare at a PFD to get your critical flight data, do we need some simple thrust indications on the PFD? Do we need a "bitching Betty" when your thrust vs altitude doesn't make sense? (Something that would not be disabled by the Radalt locking on zero as happened in Amsterdam.) I think we still have room for improvements in thrust awareness.

[CONF iture]

The only way to 'pull up' in that situation is to increase thrust to accelerate to a speed greater than Valphamax.
No need for acceleration, sufficient thrust alone will do.

That the action to do so (holding down A/THR disconnect) was erroneously omitted
It was not erroneously omitted, it was simply not necessary.

Regarding the flypast as executed, you will have noted that it was flown well below 100R, and (as Dozy Wannabe writes above) it is not entirely clear whether the captain ever carried out his briefed manual inhibition of Alpha Floor.
If he had, I do suspect the BEA would not have kept quiet about it.

In any other plane at that time, he would likely have stalled and crashed right in front of everybody there.
Or he could have got what was still in the aerodynamics to attempt to survive the tree tops and head back to BSL ...

[awblain]

Yes, if you're flying a fighter, yes it will, although it might not be the fastest way to achieve the desired height gain.

With an A320 full of people, weighing something like 130,000lb and with about 50,000lb of thrust, it'd be hard to climb without getting more speed to add some lift to help.

[HazelNuts39]

No need for acceleration, sufficient thrust alone will do.Some basic aerodynamics and flight mechanics: Changing the direction of the flightpath upwards requires that the lift force plus the vertical component of thrust is greater than the weight. The only way to increase lift at alpha-max is to increase airspeed. At alpha-max of 17.5 degrees roughly 30% of the gross thrust force contributes to the lift, while 95% contributes to acceleration. So yes, increasing thrust sufficiently to cause the airplane to accelerate will change the flight path upwards.

[gums]

Guess I had a different aero course and pilot training courses than those here.

I can't find a single reference to the effect that we must increase lift to climb, with the notable exception of being so far behind the "power curve" that we are virtually stalling.

See: http://www.dept.aoe.vt.edu/~lutze/AOE3104/climb.pdf

See also discussions right here on pPrune forums:

BEST ANGLE vs BEST RATE of climb [Archive] - PPRuNe Forums (http://www.pprune.org/archive/index.php/t-517953.html)

So I stand behind the Vermont aero department statement referenced above:

13. Climbing Flight
In order to increase altitude, we must add energy to the aircraft. We can do this by increasing the thrust or power available. If we do that, one of three things can happen:
1. We will increase kinetic energy (accelerate).
2. We will increase potential energy (climb).
3. We will do both, accelerate and climb.
If we desire to climb, we should hold the airspeed constant and use all excess power to increase our potential energy. Further, as Nuts pointed out, at an attitude of 17.5 degrees +/-, a significant amount of thrust should contribute to a climb. No need to increase the trimmed AoA or even the negligible amount of gee command to the 'bus computers.

Doggone it, the guy was behind "the curve" at his power setting - no climb capability without more thrust, and it appears he had plenty but got it applied way too late.

[HazelNuts39]

Hi gums,

strange that you never felt the 'gee' when you pulled up to increase your rate of climb. The point is that with the stick at the stop and the airplane at alpha-max in level flight you can't pull up.

[Chris Scott]

gums,

To repeat myself, the tricky scenario HN39 and I were discussing was not the go-around, but the planned fly-past at an AoA of alpha-max, using manual thrust to maintain height by keeping the IAS at Valpha-max. This would involve a delicate balancing act of thrust that doesn't apply in the case of the go-around.

As HN39 first pointed out: with the stick already fully back, the only way to arrest a descent (perhaps due to a tailwind shear) would be to add thrust. You are wrong to say that no extra lift is required to initiate recovery from a descent (or, for that matter, start a climb from level flight). As I stated in a previous post, any increase of FPA requires a temporary increase in Nz (normal acceleration), which can only be achieved by increasing the lift. With the AoA already at the maximum permitted, the extra lift is dependent on extra IAS or an increased vertical component of thrust.

Once the new FPA is established, the lift requirement reduces somewhat, but the higher pitch-attitude reduces the vertical component of the wing lift.

Hope that makes sense! When compared with what we airline guys are used to, it's worth remembering that in this case we are dealing with the very edge of the flight anvelope.

[awblain]

Gums,

Are you thinking of having rather more thrust to weight in hand to gain height?

I think you're right in general, you can use increased thrust to climb without changing speed, but perhaps not usefully in the special situation that the A320 was holding itself at near the maximum AoA, and lift, that it could achieve at that speed. I think that adding height from more thrust in that state would be possible, but gradual.

[roulishollandais]

Potatoes are better with salt, tea is worse with salt !
So is Any SOP's design, depending on the case, better or worse with automation.

Designers of a new aircraft must know the limit between the two domains in SOPs of that aircraft.

Human activities use mostly closed loops and wait the feedback .
Go around is nearly the only SOP that pilots learn to do in direct loops. That is taught before the first solo flight. Instructors and pilots know that it ALWAYS works. But automation and systems never claim probability=1.

Reading Machinbird's, gums' and others' experience, Go Around are much safer, better understood, better realized when it is strictly handflown rather with magic A/T TOGA button, despite we must learn the lesson from late thrust decision from Capt Gale C. Kehmeier and Michel Asseline , in United Airlines flight 227 and Habsheim.

Pilots must respect what their old and wise instructors taught hundreds times.
Then the SOP will be safe and not confusing. Designers must respect pilots' brain training.

TOGA button provides royalties on every sold aircraft l, would it crash or not.

Another issue for Asseline was that he has been surprised TWO times by the landing gear warnings. He was not waiting these alarms, he was behind his aircraft already at that moment. He was totaly in trouble, did not undestand the systems, the runways, the landing gear, the warning, and surely was no more able to resolve the energy, drag and
lift reckoning, angle and Vs climb, he was living in a schyzophrenic -Classic and Airbus - system. So was Mazières too : during the trial 9 years after Habsheim, he still said to the Court he thougt the plane could not stall, and seemed to still think it. Pierre Baud laughed and said the energy decided.

[gums]

Thank you, Okie.

From Okie's posts previously, I assume he had experience in a FBW system before the AB320 came on line. The Block 10 comment is the clue, and I flew Bk1,5,10 and 15.

The problem I have seen in the 'bus laws was the implementation of an AoA input, primarily with the gear down. In short, the only way to establish a constant AoA gear up or down is to use the Okie technique and pull full back, achieve whatever AoA the FCS allows and go from there. Don't try this at home, but if you have sim time, then go for it.

Our primitive FBW law with the gear down had an AoA bias mixed with the gee command. And recall that ours used a gee command like the 'bus, but we could trim for a certain gee if we wished. So with gear down we had the "feel" of what all of us were used to - speed stability due to AoA. On my leading edge flap failure approach, I was able to control flight path using power versus increasing or decreasing AoA/speed. This was important because the guy before me pranged by increasing the AoA in a "flare".

@ Chris , et al.....As far as feeling a Nz change when using power to increase or decrease rate of descent, it's negligible. 100 or 200 feet per minute is about 0.07 gee for a second or two until stabilized. up or down. Even my sensitive butt could not feel that.

So one more time: The 'bus FCS control law is for a one gee command corrected for pitch attitude. Without being at one of the AoA limits, you won't see any change in rate of climb with a power change. You just keep going faster or slower on the existing flight path. But once at an aero limit like AoA, you perform just like all the books say.

[HazelNuts39]

Having the SS full aft doesn't necessarily equate to full ANU elevator or any specific elevator position at all. That is entirely correct but not relevant. With the SS full aft the pilot has no control over the elevator position, the FCS in alpha-protection mode controls it. Holding the SS full aft sufficiently long results in AoA equal to alpha-max. With the thrust correctly set the airplane can fly around all day at Valphamax with the SS on the aft stop, provided it stays clear of terra firma. To stay clear of the ground it is not sufficient to 'set the correct thrust'. The airplane must be actively controlled to counteract disturbances. It requires a certain 'room to maneuver', depending on the magnitude of disturbances and how easily and accurately it can be controlled.

Sure, thrust changes will produce changes in vertical speed, but in my judgement this mode of flight path control is less responsive, less accurate, and less intuitive than pitch control.

As to the magnitude of disturbances, the accident occurred on a sunny summer afternoon. The wind was northerly at 5 - 10 knots. The presence of 1/8 cumulus at 780 m indicates that the lower atmosphere was slightly unstable, and surface wind therefore likely to be variable in magnitude and direction.

A change in headwind component results in a change of IAS. A loss of airspeed results in a loss of lift at constant AoA. But we do not know how accurately the FCS maintains a constant AoA because there is also the phugoid damping feature to consider.

So given these factors, what would have been a safe height above terrain to conduct the planned demonstration?

[Chris Scott]

Quotes from OK465:

"Having the SS full aft doesn't necessarily equate to full ANU elevator or any specific elevator position at all."

I don't recall anyone arguing here that it does.

"Though I don't have the luxury of an AOA gage, as long as there exists some residual elevator authority (i.e. elevator not pegged full ANU), the FCS can 'move' the nose resulting in a flight path change commensurate with thrust applied."

To "move the nose" up = instantaneous increase of AoA.

"The range of AOA that's available to the FCS to generate a pitch rate, I can't see on a cockpit gage unfortunately..."

Yes. BTW, is there a read-out on the sim-instructor panel?

"...but observed on FFS flight package diagnostics, ostensibly from flight test data, overshoots and undershoots from the stabilized Alphamax value appear to occur during vertical transitions due to thrust changes....but no more than a couple knots IAS variation depending on rate of thrust application or reduction."

Now that is interesting... And reminiscent of the Gordon Corps video (http://www.youtube.com/watch?v=9fqy8uPzW90&feature=youtu.be), in which the AoA does vary noticeably - despite his steady application of full back-stick (as discussed on this thread from January 2nd).

It suggests that the EFCS may keep a degree or two in hand - short of alpha-max - to allow the pilot a small increase in pitch to deliver an increase in AoA. That would provide an increase in lift to deliver the increase in Nz involved in any flare manoeuvre.

[CONF iture]

The point is that with the stick at the stop and the airplane at alpha-max in level flight you can't pull up.
If by pulling up you mean climbing, why couldn't you ?
Or are you already full thrust to barely maintain the altitude ?

The only way to increase lift at alpha-max is to increase airspeed.
If you increase speed, you're not at alpha max any more.
If you maintain full back stick you maintain alpha max, therefore you maintain Valphamax - No speed increase.
What makes you climb in this case is a sufficient thrust increase alone.

It suggests that the EFCS may keep a degree or two in hand - short of alpha-max - to allow the pilot a small increase in pitch to deliver an increase in AoA. That would provide an increase in lift to deliver the increase in Nz involved in any flare manoeuvre.
How would you want the pilot to go and get 'a small increase in pitch' as he's already full back stick in order to obtain and maintain alpha max ... ?

[vilas]

CONF iture
With full backstick flight path change due to elevators is taken out of equation so you can only climb due to thrust vector but it has to develop first then overcome the momentum so there will be noticeable lag. If you are doing this at 100ft or 50ft above ground I don't think you have a chance. If thrust was idle then it takes almost 3 to 4 seconds for the thrust to come out of idle.

[HazelNuts39]

If you increase speed, you're not at alpha max any more. If you maintain full back stick you maintain alpha max, therefore you maintain Valphamax - No speed increase. What makes you climb in this case is a sufficient thrust increase alone." If you maintain full back stick you maintain alpha max" - correct. The lift at alpha max is then proportional to airspeed squared. If the speed drops(*) you get nz < 1 and start to descend, if the speed increases you get nz > 1 and start to climb.

(*) Due to a change in headwind component, for example.

(To keep it simple I left out the thrust vector and phugoid damping).

[Chris Scott]

Quotes from CONF_iture:

"If you increase speed, you're not at alpha max any more."

That's only true if you maintain the same Nz. If you need to flare, you have to increase Nz (load factor) for the duration of the flare manoeuvre. As you know very well, Valpha-max rises with the load factor.

"If you maintain full back stick you maintain alpha max, therefore you maintain Valphamax - No speed increase."

That's only true after the flare manoeuvre has been completed, and the new FPA has been established. Then the Nz can return to (roughly) 1g. During the flare (i.e., the period when the FPA is increasing), the increased "g" causes an increase in Valpha-max. (See above.)

"What makes you climb in this case is a sufficient thrust increase alone."

The increase in thrust does help a bit with the flare, because of its vertcal component (at these high pitch attitudes). But its contribution is minor, as HN39 has explained here (http://www.pprune.org/tech-log/528034-habsheim-24.html#post8321844).

"How would you want the pilot to go and get 'a small increase in pitch' as he's already full back stick in order to obtain and maintain alpha max ... ?"

If the EFCS does leave a small margin below alpha-max in 1g flight, as I am hypothesising, the quickest way to increase lift is to rotate the pitch and obtain (temporarily) a higher AoA.

This brings us back to your main problem - why was the published Flaps-3 alpha-max of 17.5 deg not achieved at Habsheim? I'm wondering if the EFCS only permits alpha-max when the a/c is in a turn, at a specific bank-angle/load-factor. If so, that might enable a sudden emergency turn without automatic de-rotation.

[HazelNuts39]

This brings us back to your main problem - why was the published Flaps-3 alpha-max of 17.5 deg not achieved at Habsheim?

Hi Chris,

I think I answered this earlier.

From the report of the Aircraft Performance Group Chairman in the NTSB's investigation of the A320 ditching in the Hudson river:

(...) However, in α-protection mode, the flight control system incorporates a phugoid-damping feedback term in addition to side stick commands when computing the commanded elevator position (which in turn determines the pitch angle response). As described by Airbus,

"… the aircraft was in angle-of attack (AoA) protection from about 150 ft RA.
When in AoA protection law, stick command is AoA objective. Stick neutral commands alpha-prot and
full back stick commands alpha-max.
However, AoA protection shall take care of the A/C trajectory and, thus, looks after phugoid damping
as well as AoA control: there are feedbacks within the AoA protection law aiming at damping the
phugoid mode (low frequency mode). The feedbacks are CAS and pitch attitude variations. Without
these feedbacks, an aircraft upset from its stabilized flight point up to constant high AoA would enter
a phugoid (which is, by definition, a constant AoA oscillation) without possibility to stabilize the
trajectory. As a consequence, commanded AoA is modulated as a function of speed and attitude
variations: for instance, if A/C speed is decreasing and/or pitch attitude is increasing, pilot's
commanded AoA is lowered in order to avoid such a situation to degrade.
On the last 10 sec of the "Hudson" event, it is confirmed that pitch attitude is increasing and CAS decreasing. Then, the phugoid damping terms are non nul and are acting in the sense to decrease
the finally commanded AoA vs. the stick command, in order to prevent the aircraft from increasing the
phugoid features."

Based on this explanation, it appears that on the accident flight, the nose-up side stick commands from 15:30:36 to 15:30:43 were offset somewhat by the phugoid-damping feedback term, thereby limiting the pitch angle and α increase below 150 ft radio altitude.
(End of quote from the group chairman's report)

[DozyWannabe]

And let's not forget that the aircraft was still slowing down for several seconds after TOGA thrust was commanded while the engines were spooling up - the EFCS would likely behave more conservatively in such a scenario as I understand it.

I believe there may be some more concrete work on the EFCS forthcoming...

[gums]

I thot I saw a reference to a 'bus driver that tried to duplicate the crash. At a safe altitude, of course, and with lottsa prior planning. Reference requested.

I must also point out that commanded "gee" ( Nz) may not result in actual gee. Once the jet gets slow enough, it just can't do it and the AoA laws kick in until you have max or prot or whatever all those "limits" are. You can reach some of the AoA limits at "speed" well above what we saw in the crash profile. You know, the basic "high speed stall" conditions.

I also continue to see comments that the jet maintains one gee. I don't think this is accurate, as all the manuals I have seen and the video that Chris referenced make it clear that the jet maintains a one gee Nz " command" corrected for attitude if the stick is in neutral ( hands off). So after takeoff at a 15 degree or so climb attitude, you won't have a one gee seat of the pants, but will be slightly less. The video also implies that in a steady 30 degree bank angle that hands-off gee command is more than one gee Nz.

Oh well, think I'll go back to the lurk mode to see what all the "heavy" folks contribute.

[CONF iture]

It is not exactly clear why you guys need to introduce notions of "flare maneuver" or other "headwind component change" when the equation is simple.
As alpha max is maintained in level flight at Valphamax :

Any increase of thrust will produce a climb as alpha max is maintained at Valphamax
Any decrease of thrust will produce a descent as alpha max is maintained at Valphamax


Thrust controls the V/S
Thrust does not control the speed which stays at Valphamax

[HazelNuts39]

Thrust does not control the speed which stays at ValphamaxWhy is that? Please explain.

This is what I think will happen if, starting from steady level flight at Valphamax, thrust is rapidly increased with the sidestick maintained at the aft stop:

The airplane will start a phugoid trajectory, accelerating, climbing, decelerating and (possibly) descending, always at (close to) alphamax. The phugoid damping feature will attenuate the phugoid oscillation and eventually the airplane will end up in a steady climb at Valphamax, provided it stays clear of the ground and of structural limits.

[awblain]

HN,

Isn't the phugoid period of order 10-20s, and so near the ground, at close to maximum angle of attack, and with a big change in thrust, would it be perceptible in light of other changes?

Am I right that the reason to damp the phugoid is for comfort when making subtle changes when cruising, and thus that the FBW system applying damping within 10s of the ground might be a complication that the pilot does not benefit from?

I understand that the A330/40 was later found to suffer from an unusually pronounced phugoid mode and needed reshaping during testing, so active control alone might anyway not be sufficient to arrest its development when flight conditions are changing quickly.

[awblain]

Gums,

I agree that 1g can't be right.

If the aircraft is accelerating, turning or slowing, then the seat-of-the-pants force will change in magnitude or direction from its value sitting on the ground.
It has to, as the aircraft needs to push your straps/trousers to change your velocity so that you move along with it.

In a steady climb, descent or turn the seat-of-pants force is constant, but different from 1g. In a steady climb or descent it's 1g directed a little forward or back; in a steady banked turn it's about 1/cos(bank angle) directly up out of the seat. At 30 degrees that would make you feel about 15% heavy.

Perhaps "maintains 1g" is shorthand for "Taking your hands off signals the FBW system to keep constant the seat-of-pants forces that are currently being experienced."

[HazelNuts39]

Isn't the phugoid period of order 10-20s, and so near the ground, at close to maximum angle of attack, and with a big change in thrust, would it be perceptible in light of other changes? I really don't know that much about phugoids, and nothing about that of the A320. I merely stumbled on the phugoid when calculating cruise upset trajectories of the A330/340. In that case the period was of the order of 60 - 80s. The point I was trying to make is that increasing thrust doesn't instantly put the airplane in a steady climb at Valphamax.

Am I right that the reason to damp the phugoid is for comfort when making subtle changes when cruisingThe two accident/incident reports (Hudson and Bilbao) mention phugoid damping in the alpha protection mode only.

[FCSoverride]

On the subject of phugoids and attempting to clarify them… firstly, I must stress that I am NOT an Airbus man and am only familiar with publicly available information on the a320 FCS.

My understanding is that the control laws, when in normal law, maintain an attitude compensated “g” (within a limited bank angle range). To a pilot, this means that at centre stick, the aeroplane keeps going where it was pointed. The phugoid mode is basically an exchange of potential and kinetic energy at constant(roughly) angle of attack. So, if the a320 is flying along at a given airspeed at centre stick, an increase in thrust will result in an increase in airspeed along the current flight path. To achieve this, the control laws will pitch the nose down (reducing angle of attack) as airspeed increases. Eventually, drag will balance the thrust and the result is a new trimmed speed along the original flight path. The phugoid motion is, effectively, removed.

In a conventional aeroplane, the increase in thrust will cause an increase in airspeed and, thus, an increase in lift (remember that a stable aeroplane likes to remain at its trimmed angle of attack and there is no control law to modify this). The increase in lift will cause the flight path angle to rise. The increase in flight path angle eventually causes a decrease in airspeed which results in a decrease in lift (still at constant angle of attack) and the aeroplane noses over and begins to increase speed and lift again. Depending on the stability of the phugoid mode, a cyclic variation in airspeed and flight path angle could develop. If the phugoid is stable, the aeroplane would stabilise at a new flight path angle and airspeed, at the original angle of attack.

Now, if we go back to the a320 and consider the angle of attack protection control law. What this does is try to maintain the commanded angle of attack. I imagine that this (without phugoid damping) would return the phugoid behaviour of the a320 to more like that of a conventional aeroplane. Imagine slamming the stick to fully back with the thrust kept constant. The aeroplane would rapidly attain alpha max. Drag and flight path angle would increase, the airspeed would decay, the lift would reduce and the flight path angle would come back down again and off we could go into a potentially undamped phugoid.

What the phugoid damping terms in the alpha control laws seem to be designed to do is keep things nice and stable in terms of flight path, by looking at airspeed trend too. My guess is, at the expense of instant high angle of attack availability, flight path is considered (perhaps optimised) instead. In the case of a rapid pull from a low energy state, "firewalling" the throttles, the control laws make sure flight path doesn’t rapidly come back down again whilst the engines spool up. I also guess that they completely stabilise the phugoid to give the a320 characteristics that I have seen described, i.e. that it stabilises at airspeed for alpha max at full back stick (the eventual flightpath angle depending on the thrust set).

I would hazard a guess that Airbus engineers have done lots of simulation and modelling, with pilots too, to optimise flight path response in full back stick “avoidance” type manoeuvres.

[HazelNuts39]

Hi FCSoverride,

welcome to this thread and thanks for the great explanation!

However, I wonder if you have looked at the two documented alpha-prot encounters of A340's in cruise?

[Chris Scott]

HN39,
Thanks for grasping the nettle and reminding me and others that the designers have to make provisions for the complications of phugoid oscillations, as confirmed by the Airbus input to the NTSB report on the Hudson River accident. That seems to be the explanation of the apparent shortfall of achieved AoA at Habsheim. My tentative suggestion that it might be a measure to allow sudden bank applications without the need to drop the nose can probably be discounted.

Hi FCSoverride,
That strikes me as a very elegant description of phugoid (some of us have been there in conventional a/c). :}
I think you've explained why it doesn't happen in Normal (C*) Law, and why provision has to be made for it in Alpha-Protection mode.

Quotes from gums

"I thot I saw a reference to a 'bus driver that tried to duplicate the crash. At a safe altitude, of course, and with lottsa prior planning. Reference requested."

Yes, see the BEA report, first part of Annexe X (text in French):
Habsheim F-GFKC (http://wnoeker.de/kc/)
The flight test was performed at around 2000 ft agl, and the traces of the main parameters are superimposed on the equivalent Habsheim ones. A bit challenging to interpret.

"...the jet maintains a one gee Nz " command" corrected for attitude if the stick is in neutral ( hands off)."

In Normal Law, yes (also compensated for bank, BTW). But we are discussing Alpha-Protection mode here. The EFCS targets an AoA of alpha-prot with neutral stick, and alpha-max with the stick fully aft. Nz doesn't rule.

"So after takeoff at a 15 degree or so climb attitude, you won't have a one gee seat of the pants, but will be slightly less."

Correct, which is why, to avoid tripping over the trigonometry, I wrote "roughly 1g" in the following:
"...after the flare manoeuvre has been completed, and the new FPA has been established. Then the Nz can return to (roughly) 1g. During the flare (i.e., the period when the FPA is increasing), the increased "g" causes an increase in Valpha-max."

You can't change the trajectory of any moving object without changing the balance of forces - in this case, with a delta of lift, generating a delta of Nz.

[gums]

TNX for the reference. I agree, hard to accurately interpret, but look like the duplication follows the actual parameters well. So the crash was inevitable, huh?

Secondly:
"...the jet maintains a one gee Nz " command" corrected for attitude if the stick is in neutral ( hands off)."

In Normal Law, yes (also compensated for bank, BTW). But we are discussing Alpha-Protection mode here. The EFCS targets an AoA of alpha-prot with neutral stick, and alpha-max with the stick fully aft. Nz doesn't rule.

From looking at the FCOM stuff you folks provided two years ago, it looks like the "alpha prot" and such is simply the far right of the control law for pitch. In other words, AoA is blended into the Nz commands, and that resembles the law I had in the Viper. We could command 9 friggin' gees but the system would not let you get above a certain AoA versus gee. By the time you reached max AoA, all you could get was one gee positive Nz ( negative still available).

I don't see the alpha prot as an "alternate law" as such. Secondly, if not pulling, I can see why the system would continue to operate in an "alpha" mode unless speed was so slow and power reduced so much that the jet would stay there until thrust was increased ( and speed). Am I understanding that? Of course, with altitude, we could avoid that by rolling and getting the nose down while still holding full back stick ( not recommended for a large jet, heh heh).

Lastly, your description may explain some of what we saw with AF447. Once slow enough, pulling back on the stick would not help, although nose down authority was still available.

[HazelNuts39]

Hi OK465,

Thanks for your interesting observations in what I assume was a simulator exercise. When I wrote about the phugoid I had no idea of the amplitude of the oscillations, and I understand they were so small that most pilots would not recognize it as a phugoid.

Just to complete my picture, since the amplitude may be expected to be a function of the magnitude of the 'disturbance' that initiated it, do you recall the gradient or rate of climb achieved in the stabilized climb?

[DozyWannabe]

Thanks for grasping the nettle and reminding me and others that the designers have to make provisions for the complications of phugoid oscillations, as confirmed by the Airbus input to the NTSB report on the Hudson River accident. That seems to be the explanation of the apparent shortfall of achieved AoA at Habsheim. My tentative suggestion that it might be a measure to allow sudden bank applications without the need to drop the nose can probably be discounted.

Not necessarily. If I understand the concept correctly, the design will carry a certain amount of leeway for any potential change in dynamic loads (or "all of the above"). The graph posted earlier provides a qualitative view of how the system should behave, but that graph is not annotated with specific values - an omission which I'm convinced is deliberate.

I can't say for certain, but if I were designing such a system, I'd be inclined to make the behaviour consistent - but make specific attainable values dependent on the prevailing conditions. In particular, if the aircraft has airspeed or thrust in reserve, then reaching the optimum values should happen sooner than if the aircraft is already on the edge of the envelope.

While the deviation below 100ft RA made Alpha Floor moot in this case, its existence indicates an implicit understanding at the design stage that AoA protection can only get you so far. Without sufficient airspeed and/or thrust in reserve, attaining optimum AoA is likely to be doubtful.

[FCSoverride]

Hi HazelNuts39. I recall reading about those two events a while ago, though I have lost familiarity with them now. The impression I got (I might be wrong), from one of those events, was that the alpha protection control law in the Airbus seems like a separate “bolt on”feature on top of the C* law. I was surprised to learn that the alpha protection law can “latch” on.

The control law approach I am most familiar with is to blend smoothly between the “auto trim” and “alpha demand” modes based on pitch stick position and sensed alpha. It wouldn’t run into the problem of an alpha spike latching it in a mode in which the computed and commanded centre stick alpha value wasn’t appropriate.

An interesting point to note is that we have no phugoid damping features in the alpha demand path on our aeroplane. Full back stick gives unmodified maximum alpha within moments (unless g limited). Our phugoid is more damped naturally.

[DozyWannabe]

The impression I got (I might be wrong), from one of those events, was that the alpha protection control law in the Airbus seems like a separate “bolt on”feature on top of the C* law. I was surprised to learn that the alpha protection law can “latch” on.

In Airbus speak, "latching" is a case whereby a failure degrades the flight control law to the point where it cannot regain a higher-level control law. AoA Protection law/mode can be exited as soon as the conditions which triggered it no longer apply.

EDIT : Therefore, because AoA Protection mode is triggered by a commanded condition rather than a systems failure (as in Alternate/Direct/Abnormal Attitude laws), it does not "latch" in that sense

[HazelNuts39]

Alpha-protection is considered "Normal Law", distinct from Alternate and Direct laws. Normal law changes from C* mode to alpha-command mode when the phase-advanced angle-of-attack exceeds alpha-prot.

[gums]

Great inputs and insight from Nuts and Doze.

As far as Okie's comments about flying at the edge of the envelope.....

It seems to me that a functional 'bus could be flown to the "limiters" at a safe altitude to demonstrate the characteristics of the jet and the FCS. If I were flying one of the things, then I would like to do it, not in the sim. I would probably notice the "feel" of the jet and any burble or other indication that I was approaching stall.

Any other pilots here agree with that?

[CONF iture]

Why is that? Please explain.
That's how the alpha protection has been designed.
At full back stick the elevators are all for the AoA control to maintain it at alpha max.

The airplane will start a phugoid trajectory, accelerating, climbing, decelerating and (possibly) descending, always at (close to) alphamax.
If you have seen those Airbus alpha max demo during air show, you probably have noticed the engine thrust variations, aimed at maintaining the vertical speed at zero. With full back stick the FCS is keeping the AoA at alpha max, and at the end of the pass the pilot applies full thrust still maintaining full back stick and the airplanes climbs away at alpha max. I haven't seen anyone 'descending' ... even temporarily.

"The only way to 'pull up' in that situation is to increase thrust to accelerate to a speed greater than Valphamax."
Your statement is erroneous.