Ask The Pilot

[tgh]

Have any substantial changes been made by Airbus as a result of this incident ?
I note that Airbus says that the pilot always has ultimate control ; permitting , one assumes, the capacity to opt to fly visual and manual.
The article describing this incident seems to contradict this, reading it took me back to Kubrick and HAL….. :-(

The untold story of QF72: What happens when 'psycho' automation leaves pilots powerless?


Interesting reading, and pretty close to the truth. This event was much more frightening than an oxy bottle going bang, or an engine failure in a 4 engined aircraft.

 

[jb747]

I was reading about ETOPS certs and I was wondering what Qantas 789 would fall under having to operate PER-LHR ?

I'm not sure just how many minutes it can operate to, but basically the big twins can go anywhere now, pretty much without restriction.

 

[jb747]

Can you explain the philosophy behind "letting go of the controls"

If the issue ends as suddenly as it started (which it did), and you have full aft stick, the aircraft will suddenly pull a massive amount of positive G. The system should limit that to 2.5g, but that's the same system that's running amok. If it failed to limit the g, you could do untold damage.

 

[jb747]

Great Circle route would come under ETOPS120 Great Circle Mapper

The aircraft is certified by the FAA to 330 minutes. It is effectively unlimited. It would not operate under ETOPs 120 or 180.

 

[jb747]

Have any substantial changes been made by Airbus as a result of this incident ?

To the hardware, no. The software glitch that caused this has presumably been erased, but in millions of lines of code, I'm sure it has some relatives.

I note that Airbus says that the pilot always has ultimate control ; permitting , one assumes, the capacity to opt to fly visual and manual.
The article describing this incident seems to contradict this, reading it took me back to Kubrick and HAL….. :-(

Airbus says...that's called marketing. It has other names.

If the aircraft is in anything other than direct law, then the pilots do not have ultimate control. That has nothing to do with flying visual or 'manual'. It means that the aircraft obeys the joystick if it feels like it. Normally, of course it does. This was an instance of a physical failure showing up a software issue. There is no instant switch that gets you to direct law. To get there you need the aircraft to detect enough failures to switch to it by itself, or you need to turn off multiple systems to force the change.

 

[SFSC]

.. or you need to turn off multiple systems to force the change.

Does a checklist exist to attain this? How long would it take to carry out the actions that would force the change to direct law?

 

[AviatorInsight]

Interesting article. What scares me is that what control pilots didn't have in stopping the second dive. How are the automations different in say 777 compared to the Airbus. I assume Boeings method of control is that it can always override the automatic systems ?

While I've never flown Airbus, I can't really comment on their automation philosophies, however, on the B777 your assumptions are correct. Should we ever need to override the autopilot/autothrottle, we can.

But it's not always straight forward. For whatever reason that we would need to execute a steep turn (greater than 45º), even with the autopilot off, the aircraft has a bank protection of 25º+15º over bank. This means that at 40º angle of bank, the "autopilot" protection systems are still in place to roll us wings level. It takes a greater amount of force to try and hold it at 45º as you're constantly fighting the controls but it's still doable.

The same thing can also be said for the autothrottle system. Where, if a certain thrust level is required (lets say idle), we can physically hold the thrust levers in the idle position. Once we let go, then the thrust levers will resume the commanded position.

 

[Quickstatus]

If the issue ends as suddenly as it started (which it did), and you have full aft stick, the aircraft will suddenly pull a massive amount of positive G. The system should limit that to 2.5g, but that's the same system that's running amok. If it failed to limit the g, you could do untold damage.

https://www.atsb.gov.au/media/3532398/ao2008070.pdf

Interesting graphs from page39++ showing the "data spikes" and appears the aircraft descended by 200 feet at most, -0.8G And pitch down -8.4deg.

In a sudden pitch down like this where is the centre of rotation?.

Why did the rear of the aircraft experience the greatest vertical acceleration? (Fig A2 Pg 226)

Do aircraft like these display which Flight control law it is operating in?

The flight crew used a Satphone to communicate with monitors in Syd. Presumably sat phones of the types used here do not interfere with aircraft systems?

Interestingly AF477 A330 used a different ADIRU manufacturer (no data spikes) whereas this data spikes have occurred previously on this same aircraft same ADIRU unit and another A330 VH-QPG same ADIRU manufacturer . Do data spikes still occur?

 

[jb747]

Does a checklist exist to attain this? How long would it take to carry out the actions that would force the change to direct law?

You won't be doing any checklists in this sort of event. There are a number of ways to force the law reversion. Turn off all three ADRs (air data computers), select the alternate electrical configuration, or, the one that I'd use, turn off PRIMs 1, 2, 3.

I've heard, but never been able to confirm, that at least one military operator of the 330 tanker demanded a single switch to force the change, to be mounted on the coaming. When Airbus refused, they were informed that their aircraft was no longer under consideration...so they rethought their position.

 

[jb747]

While I've never flown Airbus, I can't really comment on their automation philosophies, however, on the B777 your assumptions are correct. Should we ever need to override the autopilot/autothrottle, we can.

The Airbus only displays this sort of behaviour when the 'protections' are invoked. Normally they behave pretty much exactly as AVI has described the 777. Protections are there because there should be no reason to operate beyond them. For instance if the stall protection activates (alpha protection), it takes you in the only safe direction. There is no reason to ever want to go deeper into the stall (unless you want to duplicate AF447). Overspeed protection is the same...it will let you go a bit past the normal limit, but there is no reason whatsoever to want to be substantially over it.

On the other hand, the protections give the ability to have what is effectively 'care free' manoeuvring. I expect that there has been more that one aircraft saved by the ability of the aircraft to sit right on the maximum angle of attack.

Autopilot and auto thrust are not part of this discussion...like the Boeings, they can be eliminated at any time. I assume the 777/787 also have some form of protections built in. There is no point in allowing such an aircraft to be stalled, so I'd be interest to hear just what the Boeing laws do. The Asiana SFO accident would have been difficult to emulate in an AB.

The issue with QF72, was that the protections did activate, and that was what caused the pitch down. To my tiny mind, it really sounded like poor software for it to allow spiking to have any effect at all. If the aircraft is not stalled, or anywhere near that situation, one second ago, then it most certainly isn't now. The data should have been effectively smoothed and filtered. From what I understand of the event, the timing of the spikes also served to defeat the system.

 

[tgh]

I suspect that the average passenger would also endorse the availability of a one switch, instant manual control reversion.

Of course most of us hope that we have top gun up front when we may , in fact, have popgun at the joystick …..with hal riding shotgun ( to continue the gun theme… )

 

[jb747]

https://www.atsb.gov.au/media/3532398/ao2008070.pdf

Interesting graphs from page39++ showing the "data spikes" and appears the aircraft descended by 200 feet at most, -0.8G And pitch down -8.4deg.

In a sudden pitch down like this where is the centre of rotation?.

Normally an aircraft moves around the CofG, so the rotation would have been around the centre of the aircraft. But, once the wing got to a negative angle of attack, it would just fall away, and I expect the rotation would be more towards the rear. The pitch pulses were very short, but very violent.

Why did the rear of the aircraft experience the greatest vertical acceleration? (Fig A2 Pg 226)

Geometry...moment arm.

Do aircraft like these display which Flight control law it is operating in?

Perhaps not as clearly as you might imagine. It will show up on the ECAMs, but the indication on the PFD, especially for alternate law, is somewhat obscure. Their thinking is obviously that the ECAM will alert you...which it will. I have a feeling that aircraft are now showing so much data, that it can be quite easy to lose important information.

The flight crew used a Satphone to communicate with monitors in Syd. Presumably sat phones of the types used here do not interfere with aircraft systems?

It is part of the aircraft systems. The A380 has two. We access in exactly the same way we would any other radio.

Interestingly AF477 A330 used a different ADIRU manufacturer (no data spikes) whereas this data spikes have occurred previously on this same aircraft same ADIRU unit and another A330 VH-QPG same ADIRU manufacturer . Do data spikes still occur?

AF447 was NOTHING like this event. In that case, icing led to a temporary loss of air data. Quite literally all that they had to do was disconnect the autopilot/flight director and auto thrust. Put the power to where it had been for the past hour or so (maybe plus a couple of percent). Hold the attitude at 2.5º (also exactly where it had been). Nothing would have happened, and the air data would have eventually returned, at which point you tidy up and continue on your way. In NO WORLD, does holding full aft stick at altitude give you anything other than a deep stall, and it would never be a recovery that any pilot should consider viable. Stalls are something we play with in smaller aircraft. What was done was the aviation equivalent of trying to make your car turn right, by moving the steering wheel anti clockwise. Not a likely winner.

 

[jb747]

I suspect that the average passenger would also endorse the availability of a one switch, instant manual control reversion.

Of course most of us hope that we have top gun up front when we may , in fact, have popgun at the joystick …..with hal riding shotgun ( to continue the gun theme… )

When I first flew the Airbus, I would probably have agreed with you. But, I see the point of the way they have done the protections, and whilst there are some things I would change, I can see a reason for not doing so. The popguns you allude to are common enough, and you really don't want them playing with the laws. The military had totally different reasons for wanting the ability to control the laws....plus they have a much more standardised product than do the disparate airlines of the world.

 

[A100]

The untold story of QF72: What happens when 'psycho' automation leaves pilots powerless?


Interesting reading, and pretty close to the truth. This event was much more frightening than an oxy bottle going bang, or an engine failure in a 4 engined aircraft.

whats the practical difference between a PAN and a MAYDAY call? Obviously Mayday is more serious / catastrophic, but does ATC treat the 2 any different,y in terms of the priority you get? In the US, isn't it just an emergency ?

 

[petercr]

I've heard, but never been able to confirm, that at least one military operator of the 330 tanker demanded a single switch to force the change, to be mounted on the coaming. When Airbus refused, they were informed that their aircraft was no longer under consideration...so they rethought their position.

That's certainly one way of making sure 'what the customer wants is what the customer gets'... Given that there's only 4x Air Force that operate the MRTT (and few more recent sign-ups), it would be interesting to know who...

 

[Quickstatus]

Do data spikes still occur?

 

[Boca68]

When I first flew the Airbus, I would probably have agreed with you. But, I see the point of the way they have done the protections, and whilst there are some things I would change, I can see a reason for not doing so. The popguns you allude to are common enough, and you really don't want them playing with the laws. The military had totally different reasons for wanting the ability to control the laws....plus they have a much more standardised product than do the disparate airlines of the world.

So, if I'm understanding correctly, its like the lesser of two potential evils. The protections/automations which correct and may potentially caused a relatively minor incident against relying on humans hands to correct, which they may not. Or even correct erroneously.

The ability to override automatic systems has to be countered with, if they can be overridden correctly? Applying the KISS principle?

 

[eastwest101]

Or is this situation one of the options that is done in simulator training for Airbus flight crew?

Seeing as the software engineers can't supply any evidence as to what caused QF72 it would make sense to train for it happening again...

 

[jb747]

Do data spikes still occur?

Not something we would know....but if they do, they aren't having any effect.

 

[Quickstatus]

has QF72 been replicated in the sim?