As studi says, completely different type, completely different systems design. Notably the old A300/310 system had a very loud warning bell when the trim went too far out of position. Of course, those types were not FBW and as such were not designed to have a constantly-correcting automatic trim - a loud bell signifying trim movement would quickly become a nuisance with the FBW design.

With a traditional empennage design, and with all control systems functioning, it is never completely impossible to un-stall the aircraft due to aerodynamics. The limiting factor is purely time (i.e. how long you have before running out of altitude to recover), and if the PF had pushed the stick forward and held it there, the THS would have come forward and recovery would have been straightforward in a matter of a few seconds (presuming that the recovery was effected with enough altitude to spare).

By looking at the DFDR traces, it's evident that the aircraft did not stay nose-high all the way down. What happened was that the nose came down, the speed built up, but the elevator and THS position caused it to pitch back up again - just like a paper dart running out of speed.

The THS went to full-NU as quickly as it did because the PF was still pulling up as the aircraft ran out of the speed at apogee and continued pulling well into the stall. The aircraft only ever did what was demanded of it.

Remember that spoilers work as they do because of their relative closeness to the aircraft's centre of gravity. Spoilers wouldn't work on the tail because they'd cause the extreme pitch-up moment that was seen when the nose dropped in the AF447 sequence - due to being placed well aft of the aircraft's CoG.

@Organfreak - I don't know if you can see this, but I don't think I've ever even hinted that the Airbus systems design is perfect (hell, I'm a software guy and know full well there's no such thing!). All I've ever objected to is kneejerk assertions that such-and-such a change would make things better, when the people suggesting it don't fully understand the decisions behind that design - some of whom have never even tried to understand it and work from the basis that it's fundamentally wrong based on their own prejudice.

I think that is the reason I stuck with Boeing. Just fly it as pilots have flown for over 100 years. There are no laws, just airmanship. If you are very stupid you can stall so just don't be stupid. How hard is that?

bubbers

Why is it do you suppose that Autotrim is kept when the Law has degraded? What advantage does it have over Direct Law, (manual reversion), such that the a/c remains at a specific Pitch incidence to work from?

This does not apply to 447, as I have pointed out, though the aircraft should have had Autotrim, it did not, and when it Stalled it did have it?

Curioser

@bubbers44 - There have been a lot of changes to how pilots control aircraft in the last 100 years, most if not all of which were necessary to progress. The very first one was going from wing-warping to aileron panels - admittedly that was more like 102 years ago now, but I suspect you were including the Wright Bros. in your timeframe. Another big one was going from taildraggers to tricycle gear, which required a significant change in methodology.

In general most of these changes were made (as were those underpinning the Airbus FBW design) to make life easier for the pilot.

As many have pointed out, it's because it's what the crew will have been used to on that type for over 99% of their time flying the thing.

HazelNuts39

Howso? With elevators and THS deflected max NU, plus AoA, what value do you derive? Never mind, I must have had a touch of vertigo.......arse about, apologies.

Dozy:

Quote. "By looking at the DFDR traces, it's evident that the aircraft did not stay nose-high all the way down. What happened was that the nose came down, the speed built up, but the elevator and THS position caused it to pitch back up again - just like a paper dart running out of speed."

Yes, and why did the THS not trim ND?

And. Quote. "The THS went to full-NU as quickly as it did because the PF was still pulling up as the aircraft ran out of the speed at apogee and continued pulling well into the stall. The aircraft only ever did what was demanded of it."

That is just wrong. It spent 43 seconds 'parked' at three degrees NU. It did not trim NU as the climb began and continued. PF's commands were a combination of ND and NU, and some time was spent -1G. So it should have trimmed, what inhibited it?

And Quote: [B]"As many have pointed out, it's because it's what the crew will have been used to on that type for over 99% of their time flying the thing."

I was asking bubbers44 what he thought.

So why carry a 99 into the one percent? Manual should mean manual, not a 'blend'

Because it was never commanded to - none of the inputs of the PF and PNF were ever sustained long enough to make a difference, with one exception. The PNF's input at 02:13:42 would probably have been enough to move the THS - however it's a moot point for two reasons:

• The aircraft was already too low by that point.
• The PF was still pulling up - hence the "DUAL INPUT" warning.

Yes, because the elevators' authority was sufficient to meet the demand.

I've said this more times than I care to think about, but because the ND command has to be *sustained*. In the A320 sim, from THS full-NU to return to neutral required full forward sidestick input for about 8 seconds. The PF never applied corrective inputs for anywhere near long enough.

For which you'll get an answer in line with his way of thinking - i.e. that the traditional control setup is the best way to do it.

Since when do the elevators act alone? Trim is automatic, and so quote me where it is printed that trim won't follow? Include trend please, and g references.

Quote. "Because it was never commanded to - none of the inputs of the PF and PNF were ever sustained long enough to make a difference, with one exception. The PNF's input at 02:13:42 would probably have been enough to move the THS - however it's a moot point for two reasons:"

Eight seconds w/o trimming? Reference if you do not mind....

"For which you'll get an answer in line with his way of thinking - i.e. that the traditional control setup is the best way to do it."......priceless

Lyman,

If you take a look at figure 64 of the english version of the final report, you will see that
(a) the THS did not remain 'parked' during 43 seconds, and
(b) the THS is too slow to follow quick movements of the elevator.

Parked on the ground, you can hold the SS full forward for 8 seconds, 8 minutes, or 8 years and the THS will not move.....

.....because there is no change in the existing dynamic pressure.

HazelNuts39


HazelNuts39 Lyman,

If you take a look at figure 64 of the english version of the final report, you will see that
(a) the THS did not remain 'parked' during 43 seconds, and
(b) the THS is too slow to follow quick movements of the elevator.

According to the traces of THS in IR#3, it was parked for 43 seconds.
How is a graph different due language in the accompanying text?

Too slow? So it ignores trends over time? Pretty easy to circumvent its action; so then why retain it? Bump the stick many small times so as not to wake the beast?

I trust you, as always, but here I do not believe you. For me, I would need to see the pertinent data from Airbus, re: trimming.

WAIT.... So that is why all Bus drivers stir mayonnaise? Eureka, now we know they do not like Autotrim either. Small bumps and suss... Small bumps and suss, etc. etc. fear? Or loathing? Both?

Very true - 8 seconds (or thereabouts) was the value we got in the A320 sim with THS full NU, followed by sustained full-ND on the sidestick when we attempted the AF447 scenario. Approximately 8 seconds from full-NU to neutral, controlling roll with rudder rather than ailerons (due to stalled state).

NB. I'm well aware that this experience does not necessarily reflect the behaviour of the real thing in an aerodynamic sense, but the systems are the same - no reason to think THS return to neutral would have been much different...

Lyman, I cannot cite this b/c I'm an airhead with few remaining memory cells, but, I am sure that I have read in an AB publication on control laws that the ND input must be definite and prolonged before the AC will trim accordingly. I'll look for it in my mahssive collection of documents. I think you should give up on this line of speculative inquiry. Ppl been telling you repeatedly that Bonin's ND inputs weren't enough to do it. No hidden conspiracy there, I accept it as fact.

Very true - 8 seconds (or thereabouts) was the value we got in the A320 sim with THS full NU, followed by sustained full-ND on the sidestick when we attempted the AF447 scenario. Approximately 8 seconds from full-NU to neutral, controlling roll with rudder rather than ailerons (due to stalled state).

Dozy, please. Eight seconds to return to NEUTRAL..... When was it going to start its eight second journey? Or does it linger at FULL NU and then instantly return to NEUTRAL?

Skip the ailerons, why did you not think that your sim proved nothing? 447 s THS did not BUDGE, eight seconds or eight days, etc.....

More to the point, at what time did the THS start to move? Two seconds?

Are you making this up as we go along for the 'ride' ?

Hammond man


Organfreak
*

Lyman, I cannot cite this b/c I'm an airhead with few remaining memory cells, but, I am sure that I have read in an AB publication on control laws that the ND input must be definite and prolonged before the AC will trim accordingly. I'll look for it in my mahssive collection of documents. I think you should give up on this line of speculative inquiry. Ppl been telling you repeatedly that Bonin's ND inputs weren't enough to do it. No hidden conspiracy there, I accept it as fact.

Not at all, why is there no discussion on the THS as to precisely when it moves,

Why would it hesitate, to see if there is a established Pitch? What is "established pitch"?

This is not speculative, it is a demand in the wind for data to be supplied. BEA have left anyone who wants some technical back up hanging.....

Because Airbus owns the DATA? And that is important and needs to be protected exactly, WHY?

It didn't budge because it never received a command requiring it to. In the sim, the trim wheel movement and behaviour seemed akin to a curve - with full nose down it started moving after a second or two, and had returned to neutral 8 seconds later.

In the case of AF447, the only time full-ND was held on the sidestick for more than a fraction of a second came right at the end by the PNF, and was counteracted by the PF pulling full-back on his.

Remember that 8 seconds to THS neutral has the elevators commanded to full ND throughout.

I don't know why this is so complicated. We have all spent a lot of our lives at FL350. How many of us would pull up into a 15 degree climb? None of us, right? It wouldn't matter what warnings were going off any competent pilot would have looked at attitude and power and been just fine. For some reason they couldn't.

In IR#3, look on page 42 of the english version or page 44 of the french version. That graph was erroneously included as Figure 64 in the english version of the final report, the intended graph can be downloaded separately from the BEA site.

Lyman,
You're not listening, not to me nor anyone else.

But worse, why in hell did you go to the trouble of posting my personal info? Was it to underline the fact that I'm not a pilot?

:confused:

I'm so sorry, I went to copy your post and it grabbed the whole deal.

This is not for that reason, but just because on the ground the Law in force is Direct, so no autotrim.
But nevertheless the reason you mention could very well apply to an aircraft that is totally stalled and therefore in state of neutral balance in which no dynamic forces apply anymore ... ?

The question could be, if the sidestick is maintained full fwd, so the request is for minus 1G, would the elevators still go to the full down position, and the THS to the most positive setting, in research to satisfy anything in that direction of minus 1G ?

BEA ... all that stuff should be already in your report.

I know you don't think much of it, but in our sim sessions it certainly did.

Why? The only reason it would need to be in the report would be if they tried it in a simulated session and it didn't work as it was supposed to.

Quote:
Originally Posted by Lyman
According to the traces of THS in IR#3, it was parked for 43 seconds. How is a graph different due language in the accompanying text?

HazelNuts reports:
In IR#3, look on page 42 of the english version or page 44 of the french version. That graph was erroneously included as Figure 64 in the english version of the final report, the intended graph can be downloaded separately from the BEA site.

Erroneously included? Or Erroneous?

What does the new graph depict? The reason I ask, is there have been several
Items, removed and replaced, prior, and I tend to go with the original information, unless there is a written and authenticated correction.

For instance, in the original thread on the "ARM 36 G" the included image showed no damage to the end of the strut, it was completely different from what the text of the explanation offered. With the final report, a different image is provided, one with "damage" that more closely matched the text offered from the beginning. Moreover, that image changed in the course of time, to include "better" resolution of the damaged area, whilst the rest of the "photo" was unchanged.

I do not wish to imply any thing at all, but I'll cite the original, when I can.

Rgds.

Firstly - going from "interim" (which is another way of saying draft, preliminary etc.) to "final" implies that the "final" version is the one with correct data. That's the same if you're talking about aviation, legislation, engineering specifications or even creative writing.

Don't think for a second that the data in the interim report went straight from the flight recorders into the statistical/graphing software (which looks to me to have been MS Excel) - it will have been copy/pasted and moved around several times in both cases. The likelihood is that there was a mistake transferring the data that ended up in the interim report that was not picked up in the proof-reading, then later corrected in the final report and annexes/appendices.

In fact the apparent low resolution of the appended DFDR traces in both cases looks to me like the PDF conversion was performed by someone who wasn't aware that the resolution would be degraded to such an extent (or the correct settings to compensate).

The correct Figure 64 is actually quite interesting. It shows the longitudinal, lateral and vertical components of wind during the first minute, the simulation including the effect of those winds (IR#3 did not include those) and the simulation with wind but without pilot action.

Noted. That leaves the two different images of the "Arm 36". Actually three, counting the morphed fasten seat. Interesting...This photo was available from the beginning, since it recovered to the warehouse as part of the VS/Rudder.

Are there different images, and why?

I'll search the other computers, I have all three pictures......

Thank you to Dozy and HN for patience and data....

Which is in itself a significant achievement - the computations involved would have required some very heavy number-crunching.

But in mine it didn't for the THS.
Also, my full fwd action on the stick did not change the attitude, were the elevators moving in the full down position, I would expect so but cannot guaranty it as I didn't check the flight controls page.

Why would they try in a simulated session as a sim is not representative of real life.
Our sim experiences were for fun only, we did report what we saw, that's it.

The tremendous amount of messages and interest around the THS show that adequat information is clearly needed in the official documentation or through the BEA reports as last ressort.
You have demonstrated already your misconception.
I, as an Airbus pilot, would need to know much more.

THS time response
 

In a limited defense of DOZE, I submit:

Other than the aft cee gee problem that we had with the Viper, I agree. Holding stick forward for a few seconds should see the "auto trim" kick in to allow reducing stick fwd/back pressure, just as if you rolled the trim wheel. Secondly, you don't want a short time constant on the auto trim. A few seconds, like 3 or 4, should be about right. Otherwise, you could have a "computer-assisted PIO", as if the auto-trim reacted in a tenth of a second, heh heh. Can also average the stick inputs for a few seconds and not require a constant 1.3 gee or -0.85 gee command. Ahhhh, the wonders of a FBW system.

I agree with several here that the auto-trim did not help, but I agree with those that maintain it was not the major contributing factor. No doubt trimming nose down using the wheel would have helped, but that didn't seem high on the priority of the PF. And the nearly constant nose up command seems to me to be the most damning crew action.

I am becoming upset with the "personal" stuff I have seen lately.

If anyone wants to attack me, then I already have my kevlar flak vest on and NOMEX shirt.

Otherwise, I would prefer to see a more "professional" dialogue, as the name of this forum implies.

How long did you hold it for?

In this case it really needs to be a positive and sustained movement (just as it would be on the yoke in a traditional setup). I don't know the circumstances of your experiment, but in ours I was fixated upon trying to follow the ADI and the trim wheel simultaneously - almost as if I had my right eye and left eye trained in different directions. On the second attempt I followed the ADI and the standby altimeter and V/S (as the only way to simulate similar conditions involved failing the ADIRU on my side).


It can't realistically simulate aircraft behaviour outside of the tested envelope, but it will always be faithful in terms of system behaviour, which ultimately was what we were trying to prove.

Maybe, but as far as I can tell the "tremendous amount" of messages seem to be coming repeatedly from the same 4 or 5 posters.

The BEA's remit (like any other accident investigation body) is to report on facts derived from recorded data applying to the incident. In order to establish those facts, experiments are performed to test the circumstances in which the recorded data makes sense.

If those experiments don't apply to the data from the incident then the information derived from those experiments doesn't end up in the report.

As I said earlier, prior to a lucky break in which a faulty B737 rudder PCU reversed itself following thermal shock with no fatalities, the US NTSB was forced to publish a report on UA535 as having no conclusive data - they were about to do the same with USAir427. Investigative agencies can only publish relevant data which can be proved scientifically.

No matter what your personal opinion on autotrim is, the fact is that it did not perform contrary to the way it is supposed to and as such has no place in the report.

Misconception of what? (and I sincerely want to know). You tell me I'm wrong, yet there are plenty of posters (including line pilots) who say I'm not.

Hi,

What about ......
Incidence more of 30°
Unusual position (law)
THS frozen

DW
In first place I dunno why a BEA report (if I take in account your criteria) as the BEA already tell in a interim report that the plane acted as per design ..
So nothing who din't work as it was supposed to ...
Seem's the BEA report is only pointing and investigating the pilots actions (hence the human factor group created especially for)
paradoxically.. many recommendations concern the plane .....

If I may add to DW's reply: It should not be assumed that the simulation was done in a training simulator. More likely it was done in an engineering simulator, or perhaps just in a computer.

Not sure what you're getting at
Abnormal Attitude Law was never triggered
No evidence to suggest that because no input was ever made that would have moved it.

The BEA's remit concerns the events of the accident (technical, human and procedural). They may have confirmed that the aircraft behaved as per spec by the time of IR#3, but the Human Factors team required a lot longer to attempt to draw positive conclusions about the crew's behaviour.

The BEA are "pointing" at nothing (nor have they ever done so) - they have simply produced a report which states that based on the information they have, this was the likely sequence of events. Where conclusive evidence exists, they have outlined it - and where evidence is inconclusive, they have explored the possible reasons for it.

One of the unfortunate side-effects of the Continental approach to accident investigation is that the BEA - who are responsible for a purely factual approach to the investigation - are frequently misattributed to be behind the later legal and criminal findings.

My guess would be that the systems behaviour testing was performed at a Level D sim in Toulouse, and the aerodynamic behaviour was simulated on a "big iron" setup normally used for design work.

HN39 post 1233
Hazlenuts, I think you mis-understood the intent of my post. It was a hypothetical what would have happened if the aircraft dropped to Direct Law instead of switching to Alt2B? My conclusion was that the aircraft would not have been easier to fly, but that the likelihood of disaster was significantly less.

PF would still have rocked his wings, and he would have pulled his nose up as he did before, but as the aircraft slowed, he would have had to pull much harder to hold the nose attitude, or he would have had to trim.

But trimming nose up means intentionally selecting a lower speed. That would have required a conscious decision on his part. The aircraft would have actually told him it was getting slow by the amount of nose up effort required to hold the nose up.

The aircraft crashed because PF did not understand what was going on, not because it was difficult to fly. Autotrim masked the loss of speed.

Using a lowest common denominator design strategy, dropping to Direct Law instead of Alt 2 B would actually improve the pilots understanding of where the aircraft was in its flight envelope. I'm betting he would not have rolled in more nose up trim if it was under his direct control.

MachinBird, I almost agree. I totally agree with the intent of your post and you did understand my earlier post. If we're talking about any other transport category in my knowledge, their speed stability would have required more elevator force to hold the nose up. We both understand what that means I think. The Bus, however, offers no control feedback, therefore the pilot would never know how much effort was required to maintain the stall.

That is my gripe with the Bus, not the fact that it auto trims, or how it auto trims, but that it isn't speed stable in trim. Multiple tens of thousands of pilots around the world fly multiple tens of thousands of airplanes and the vast, vast majority of those airplanes are speed stable.

What I find surprising is that the only crew that couldn't deal with the Airbus and UAS was a Bus only crew. You might have thought that some of us old steam gauge cable control dinosaur pilots with more time trimming on short final than some ab initio pilot would have been the ones to mess it up. Cheers.

Something to think about.

A line pilot almost never applies any significant amount of nose down elevator. I fly narrowbody Bus's, not widebodies, but I spent some time observing my own last five legs and even though I flew one leg from sea level all the way to FL290 and leveled off there, I only pushed three times. Once to smooth out the flight path upon flap retraction, once to accelerate at FL100, and once to level off at cruise. Not one of those nose down inputs required more than a minimal forward SS input. Each input require approx 1/8th the available forward stick travel.

If the pilot is a gear up , autopilot on, kind of guy: or a minimums autopilot off, kind of guy. It is very likely that any significant amount of forward SS is a very foreign situation.

PS, no I don't stir mayonasse, nor make quick hits on the SS. I apply continuous pressure and I hand fly extensively.

Edit, after further thought, I did make many minor corrections to pitch during the climb and each was smaller in SS deflection than the three I recount above. Be that as it may, any forward SS movement make in normal flight is extremely minor.

Machinbird,

It wasn't clear to me that you were discussing direct law. Thanks for clarifying.

EDIT::
Agreed, but he didn't need to trim to stall the airplane, the stickforce SS on stop would have been the same, and that did not deter him.

As EMIT points out on the other thread, in the end the human factor is stronger than the technology.

I simply can't believe what I am reading.

From whatever cause the airframe gets so badly trimmed that it goes into a slow speed fast sink condition.

We are calmly told that the way out is to apply a total of ten seconds full nose down before any beneficial effect is seen.

Not exactly intuitive, and unlikely to be discovered amongst all the other similarly improbable actions.

The aircraft simply should not permit this to happen. There is no justification for the THS to move so far so quickly. If it had been rate limited, at an appropriately slow rate, it would not have put the airframe in such an impossible state.

Or can anyone think of a reason why the autotrim would have to respond so dramatically in steady high altitude cruise?

CVR revisited
 

On the one hand, the sterile, "What's it doing now?" of the CVR may accurately reflect the nature of the conversation in the cockpit on this fateful flight. Furthermore, it may actually shield the pilots from greater criticism than that to which they have already been subjected. On the other hand, as an interested observer since Thread 1, I'd like to know if the pilots conducted themselves in the sterile manner depicted, or if they screamed profanitiy-laden abuse at each. My desire is not prurient, rather, the tone of the conversation will, in my opinion, yield a degree of context that is very much lacking in the report as presented.

Hi mike-wsm,
The stab trim simply supported the elevator control as they responded to the real reduction in airspeed. The PF attempted to fly outside the envelope. If he had maintained at FL 350, the real airspeed would not have changed much, and the stab trim would not have had to move.

Ratty, friend,

Yes, understood. But that does not answer my question. I suggested rate restriction and asked for reasons why rate restriction would be unacceptable.

Final report, 1.1 History of the flight: