Air Asia Indonesia Lost Contact from Surabaya to Singapore

[Ian W]

Originally Posted by Machinbird

Take HEALY's points on Airbus computer re-set philosophy and consider that Airbus engineers gave the computers dominance over the pilots in their operational concept. That you and I think this is foolish has no bearing on what has transpired already. We may only be able to influence what direction future aircraft designs take.

I suspected early on that QZ8501 was going to be a man/machine authority horror story and so far, I am not dissuaded from that viewpoint.

I'd like to point out that without the FDR data, we readers do not know when the FAC shutdown attempt occurred relative to the flight control problems. The sooner they can publish the FDR readouts, the sooner we can begin to properly wrap our minds around what caused this accident.

We can already surmise a lot from the lack of reaction from Airbus who already know all there is to know of the DFDR/CVR recordings. There have been no urgent ADs issued by Airbus as one would expect if something had 'failed'. So we know that in Airbus' view the aircraft worked as expected.

So I tend to agree with Machinbird that although Airbus thinks that the aircraft worked as expected, the crew may not have had that opinion at the time. Were they trying to fix a zoom climb by pulling the FACs - is that the action for frozen AOA vanes? If they dropped into Alternate Law - from the AFR447 thread - the roll is far more sensitive than pitch - is that the case in the A-320? In trying to fix a problem did they put themselves into a test of their hand-flying the A-320 IMC, at height, in severe turbulence and in alternate law?

[vilas]

If AOA sensors had frozen at lower speed(higher AOA) then they would trigger Valpha protection(not alpha floor) at higher speed/M as the AOA threshold for the protection is lowered and the speed tape will show valpha prot/alpha max masking VLS and since it is a latching condition normally would cause a descent at Valpha prot. But if there is an updraft and speed tends to increase the aircraft in trying to maintain Valphaprot will climb. If AOA vanes were normal you can get out of alpha prot by simply pushing the stick forward. In QZ8501 case the aircraft was cruising at .78 for a while without a problem so AOA vanes should be at the correct threshold so I am not convinced that is the cause. Also to override protections you need to switch off two ADRs and not FACs unless they did it by mistake. After evoking abnormal attitude law, after recovery the aircraft goes in alternate law and stays in alternate even with gear down. If you want direct law for landing you need to reset ELACs

[matkat]

Sikpilot, just for clarity purposes Airbus have no authority to issue any AD only a national authority has that power.

[cwatters]

Many issues found on the ground are fixed by just resetting the system...

Makes me cringe to hear that. The problem might go away but it's not "fixed" unless you find out why it needed resetting. Repeatedly resetting a system is a very bad habit to get into. It's a bit like repeatedly replacing a fuse that keeps blowing.

[Derfred]

If it latches on alpha-prot and starts an uncommanded climb, wouldn't the pilot just push forward on the stick?

The leaks thus far have indicated that the pilot in the RHS pulled back on the stick.

[vilas]

Gysbreght
Alpha prot and alpha max are AOAs but shown on PFD as speeds hence the V. Aircraft when latches on to that it maintains by doing whatever it takes. Since reduced bank angle protection to 45 degrees is not relevant to the topic I didn't mention it. Also from alpha to alpha max side stick commands alpha and not load factor and it maintains the alpha prot(AOA) and not G.

[rideforever]

We can already surmise a lot from the lack of reaction from Airbus who already know all there is to know of the DFDR/CVR recordings. There have been no urgent ADs issued by Airbus as one would expect if something had 'failed'. So we know that in Airbus' view the aircraft worked as expected.

They must conclude that an AD would not prevent recurrence of the event. Which would be the case if for instance the human / computer interface needs "refinement".

I do wonder what is occurring at Airbus right now. They must be quite frantic.

"You can never make the same mistake twice; because the second time you make it, it's not a mistake, it's a choice."

[lapp]

I do wonder what is occurring at Airbus right now. They must be quite frantic.


Why they should be. Millions of their airplanes fly everyday perfectly fine. We don't even know what has happened beside unofficial leaks that do not even positively indicate a defect.

[lomapaseo]

They must conclude that an AD would not prevent recurrence of the event. Which would be the case if for instance the human / computer interface needs "refinement".

I do wonder what is occurring at Airbus right now. They must be quite frantic.

"You can never make the same mistake twice; because the second time you make it, it's not a mistake, it's a choice."

Yes

But who made the mistake under what regulation ?

Design/manufacture, Maintenance or Operations?

[vilas]

Derfred
In the level bust of A330 I think with over speed the pilot disconnected AP and when the aircraft climbed initially due to high speed protection and then due to alpha protection he didn't do any thing. It was strange because pushing the stick is a normal reaction to prevent climb and would have that cancelled the protection. In QZ case we do not know the whole story. The problem with the protections is that they were designed to prevent pilot indiscretions and not severe environmental changes. They work fine and may saved hundreds of incidents which we do not come to know. But the protections get triggered even when the environment is the cause like Bilbao, where aircraft was sinking in a severe down draft and pilot was struggling to pull but alpha prot was triggered by the high AOA and pitched the aircraft down as it was programmed to. In the level bust it maintained the alpha prot but due the prevailing AOA due to wind changes caused the aircraft to climb instead pitch down as it would have normally.

[HEALY]

cwatters

Agree entirely that continual resets does not fix the underlying problem, as I backed my idea that it's effectively a "get out of jail card". Some of the more common issues that may come up on a post flight maintenance report such as those related to FWC or Zone Controllers are generally sorted by engineers interrogating the system and resetting it, especially at an outport during a turnaround.

It takes judgement and big picture awareness by adding things up and determining if it's best course of action and any previous history is actually showing a more worrying trend that it's not just a brain fart with the computer but a serious issue with the system.

If the rumours are true that this acft had a history relating to possible FAC issues and a "quick fix" was all that was needed to get it dispatched then a larger issue with maintenance control and culture will come into play. Wether this becomes public will be another story.

[toffeez]

What Airbus may issue is an AOT (Alert Operator Transmission). They could issue this if there is a technical problem that may be repeated. Furthermore they have in the past issued AOTs to remind operators to follow correct operational procedures. That they have not done so in this case is equally telling.
All manufacturers are acutely aware of product liability responsibilities (and lawyers).

[darobstacraw]

reset is not a fix, but a return to a (hopefully) good known condition

"Makes me cringe to hear that. The problem might go away but it's not "fixed" unless you find out why it needed resetting. Repeatedly resetting a system is a very bad habit to get into. It's a bit like repeatedly replacing a fuse that keeps blowing."

This is a valid point, resetting to "fix" a problem is just dead wrong. However these systems are complex tools, and trying to tear down the system to understand it's current state, and determine what is actually wrong is a massive undertaking with the tools that are currently available.

The reason we go to a reset \ reboot is to clear all current conditions, and return to a good, known, configuration where the tools should begin to behave once again in a expected manner.

It is not a proper fix for problems, but it is a way to get a tool back into operation, and doing what we need it to do in a reasonable timeframe, and that is a good thing too.

[jcjeant]

This is a valid point, resetting to "fix" a problem is just dead wrong. However these systems are complex tools, and trying to tear down the system to understand it's current state, and determine what is actually wrong is a massive undertaking with the tools that are currently available.

If you manufacture something ( not disposable) you must be able to fix it in case of problem
If you can't fix it in a reasonable time for the customer ( commercial custom) your product is a bad design

[737er]

I'm guessing normal law overspeed protection put them in a high altitude uncommanded climb whilst in a cell and then couldn't keep up with the weather dynamics when they hit the backside of the updraft.

[jcjeant]

using normal attitude sensors, when a stall is detected

It's a nice idea ..
The problem is that the "sensors" have proven repeatedly they are not always reliable in spite of the tests that have allowed to grant them certification

[roulishollandais]

Originally Posted by Gysbreght

You observe correctly that sidestick commands alpha and not G. However, at any given alpha the G is constant unless the speed changes.

Steady state is different from transient state. We need to know the actual algorithms, the tempo, initial conditions, aso. We are seeing that Airbus never published them as far as I know.

[Lost in Saigon]

Originally Posted by 737er

I'm guessing normal law overspeed protection put them in a high altitude uncommanded climb whilst in a cell and then couldn't keep up with the weather dynamics when they hit the backside of the updraft.

I am guessing it is much more than that. This can't be the first time an A320 has inadvertently entered a thunderstorm.

[xcitation]

Quote:
Originally Posted by 737er View Post
I'm guessing normal law overspeed protection put them in a high altitude uncommanded climb whilst in a cell and then couldn't keep up with the weather dynamics when they hit the backside of the updraft.
I am guessing it is much more than that. This can't be the first time an A320 has inadvertently entered a thunderstorm.

The weather avoidance is just one hole in the swiss cheese. Also they might have penetrated an exceptional CB: super cooled water blocking sensors, huge ice particles, lightning flash, extreme turb, 100kt+ updrafts?

[Coagie]

What Airbus may issue is an AOT (Alert Operator Transmission). They could issue this if there is a technical problem that may be repeated. Furthermore they have in the past issued AOTs to remind operators to follow correct operational procedures. That they have not done so in this case is equally telling.
All manufacturers are acutely aware of product liability responsibilities (and lawyers).

Doubt Airbus would issue any document that could be seen as an admission to any culpability in the crash any time soon, unless it was something that was likely to happen again very soon (like some obvious problem in a recent upgrade to the aircraft). Lawyers for the manufacturer of the aircraft and the airline are contacted before the flight is even reported "Overdue". The lawyers work with actuaries and PR people to figure out every move their client makes, when it comes to the crash. Even though an AOT might be the ethical thing to do, that only enters into it if the PR people convince the lawyers, that it will help mitigate the damage to their client, and that would only be the case, if there is no doubt, that their client bares most of the blame.
This may sound cynical, but it's reality.