Quoting hivue (Reply 50): Quoting BoeingGuy (Reply 47): I'm willing to bet you have no understanding of the Asiana accident and automation on the 777. In the immediate aftermath of the accident I learned (on these forums) more about the so called "flight level change mode trap" than most laymen would probably be interested in learning. I will certainly plead guilty to not being an expert, though.

So I need to be careful what I post, and I had no involvement at all in any investigation. However, I'm well familiar with all Boeing autoflight modes; when they should be used; and what exactly they do; and the interactions between them. I don't personally agree it's a "trap", because it's only a trap if you clearly use the mode incorrectly. My cruise control analogy in a previous post is still valid.

So which was it? A post about the incident or your "USair check pilot buddy?"

Quoting reltney (Reply 13): The PILOTS did NOT pull the throttles back. About a week after the crash, a post was up about the incident that explained what happened. Here is a quick outline .

Quoting reltney (Reply 36): Wow, totally different from what USair check pilot buddy said a week after. I should know better and wait for the NTSB report only. My bad.

Quoting hivue (Reply 46): Right. Ask the Asiana crew at SFO how much good back driven throttles did them. They might also have a comment or two on the usefulness of 777 automation

Funny that you referenced the Asiana crew. When I started reading this thread, my thoughts were something like: so if this would have been an Asia crew doing this we would read right here how poor Asia crew is, their poor CRM, and that stuff like this would never happen to properly trained crews from airlines operating in Europe / North America with oversight from FAA / EASA / local CAA's . . .

Quoting trent772 (Reply 27): flight control inputs are "algebraically" added so if I push forward half way and you do too the elevator will reach full deflection.

Quoting longhauler (Reply 41): No. It adds the inputs algebraicly, up to the maximum allowed.

Interesting. What's the rationale behind this? Setting aside the CRM component of dual inputs, if both sticks are pushed forward halfway, neither PF nor FO are commanding full deflection. Why would the computer not interpret this as both inputs commanding half deflection? Is there a scenario in which adding the inputs together would be necessary or helpful?

Quoting caoimhin (Reply 54): What's the rationale behind this?

Here is my guess (I am not an engineer -- aeronautical or otherwise): Take the very simple example of yokes cabled up to the flight controls (discard hydraulic boost as well; pulley and lever arm moments can be ignored without loss of generality). If the pilot and copilot both push forward on their yokes with a force of, say, 80 lbs each, the elevator is going to "feel" a force of 160lbs nose down. This concept has just been carried over to electronic controls.

From some old threads I understand the sidestick is a "roll and pitch rate selector." The control surface deflection and the force needed to implement the request is computed by the electronics.

From that model, I guess the reason they use addition (or "algebraic combination") of movements is based on the idea that each pilot is constantly making an instantaneous judgement of the current rates of roll and pitch. Looking at my instruments I see the aircraft is rolling left, but not enough - I command some more left roll. Or, I see it's nosing down, so I command to pitch upwards.

In this model, it seems like the *cause* of the aircraft's current roll and pitch state is supposed to not be relevant (it may be due to outside factors and/or the other pilot's inputs). All that matters is that I think the roll or pitch state should be different, so I command a change. If that's the idea it puzzles me. Because depending on WHY the aircraft is in a certain attitude, the "right" inputs to restore "normal" flight might differ (for instance, spins/stalls).

Seems to me this whole model would break down where the two pilots seriously disagree about the WHY, then they make different inputs and the result is the inputs cancel each other.

But Airbus must have thought it through so what am I missing?

Quoting seat55a (Reply 56): From that model, I guess the reason they use addition (or "algebraic combination") of movements is based on the idea that each pilot is constantly making an instantaneous judgement of the current rates of roll and pitch. Looking at my instruments I see the aircraft is rolling left, but not enough - I command some more left roll. Or, I see it's nosing down, so I command to pitch upwards.

You are on the right track, in that it is not just control input that effects a change, but also the rate at which that control was input. So the FBW computers "know" the difference between slowly rolling the aircraft to a 45 degree banked attitude and swiftly moving the sidestick to one side, "sensing" the urgency. So when hand flying the aircraft, the pilot quickly learns how the aircraft will react to various control inputs. The pilot also quickly learns that fast control inputs will result in a very quick aircraft movement. (It is very impressive for such a large aircraft).

Now ... this is where it gets hard to explain, if you haven't flown an Airbus with a sidestick. In Normal Law, (which is 99% of the time as take-off and landing are slightly different), within certain limits, the aircraft attitude will remain the same. So you set at attitude with the sidestick, then stop input. The aircraft will remain in that attitude until you use the sidestick to change it. It is very different from say a Boeing 737, by design. Oddly enough, I remember my first time flying the A320, and it all became very intuitive very quickly. It just seemed to make sense, and felt very natural.

So, with this style of flight (all of these examples with the autopilot off), it is very unusual to have to put a lot of input into flying, therefore even more unusual that two pilots would feel the need to input. But, by design, if two did ... the FBW computers would "sense" the urgency. But remember, the algebraic result of two inputs can never be more than full input on one side.

And, if the need arose, one pilot can "lock out" the other pilot's sidestick. Either short or long term. It is very clearly annunciated, both aurally and visually which sidestick is working. In over 6000 hours on the A320 series aircraft, not once did I ever have to do that, and not once did more than one pilot have to "adjust" what the other pilot was doing.

I really don't understand what went on in the heads of the pilots, but the most surprising is that after having decided to continue the take off anyway, knowing something wasn't exactly right, why did they not take into account the alarms blaring, especially if they didn't understand them ? Sure that would have meant a RTO, but they were still at a safe speed to do so. Were they thinking of their careers ?

But I guess when you consider the first of the WTF decisions they made, all bets are off after that.

Quoting seat55a (Reply 56): Seems to me this whole model would break down where the two pilots seriously disagree about the WHY, then they make different inputs and the result is the inputs cancel each other.

Agree, but then the model would also break down on airplanes with interconnected control columns as well, say I want to make a right turn but you want to make a left and we both move our set of controls, the result will be one the same on any kind of aircraft, we're not going to go anywhere.
If you think about it, the system is very well thought out and all it really does is follow the same path of what would happen on a "conventional" aircraft.

Quoting longhauler (Reply 57): Oddly enough, I remember my first time flying the A320, and it all became very intuitive very quickly. It just seemed to make sense, and felt very natural.

Same here, it took me about a minute to get the hang of it, "natural" as you put it is the best way to describe it, despite of what all the naysayers might come up with, I think the Airbus is great machine, I've had the pleasure of flying both A and B, in both of their narrow and wide offerings, IMHO the Boeing is more of a hands on experience, it's a great airplane if you want beat the pattern and do touch and goes for three hours straight, the Airbus on an everyday basis is a much more comfortable "office".

Don't get me wrong, I love Boeings, I just think the Yoke needs to die.

Quoting Aesma (Reply 58): I really don't understand what went on in the heads of the pilots, but the most surprising is that after having decided to continue the take off anyway, knowing something wasn't exactly right, why did they not take into account the alarms blaring, especially if they didn't understand them ? Sure that would have meant a RTO, but they were still at a safe speed to do so. Were they thinking of their careers ?

No master warning. The RETARD warning is not well understood (they interviewed several other pilots and none of them knew RETARD could sound during takeoff) and did not happen until 80 kts. It was company policy to limit high speed aborts (above 80 kts) for minor threats.

BTW all of you talking control summing etc, the PNF (copilot) did not touch her controls per the DFDR, so the topic is not relevant.

Did the plane ever really got airborne? At no time did all three landing gear WOW switches show AIR. The nose gear did. Then none of them. Then the two main gear. Then none.

It seems to be a recurring theme these days with the automation available. Much as Sully Sullenberger was concerned with, many pilots these days have come to rely very heavily on the automation. That's fine as long as everything works normally but as soon as something out of the ordinary happens they don't seem to know how to handle the basics of flying the airplane. AF447, US1702, and OZ214 are examples of that. The problem lies not just with the increased automation but the lack of training. I used those three examples to illustrate that its not just a national or regional problem but a global problem. I also used those three because its not just an Airbus or Boeing problem. Its both.

Quoting seat55a (Reply 56): Seems to me this whole model would break down where the two pilots seriously disagree about the WHY, then they make different inputs and the result is the inputs cancel each other. But Airbus must have thought it through so what am I missing?

You could have a system where the strongest pilot wins. I don't know if that is any better. Or one just knocks out the other.

Quoting RickNRoll (Reply 62): You could have a system where the strongest pilot wins. I don't know if that is any better. Or one just knocks out the other.

I surely agree, once it comes down to duelling inputs instead of agreed roles, then only luck or external factors are going to save the situation.

But here's Airbus with the new idea that strength doesn't win.

So I wonder if they had thought beyond the priority button to some sort of scenarios for its use. The bare pushing of the button (or compensating input) just seems like a passive-aggressive way of declaring "My aircraft" without saying it. Was that supposed to be empowering for a FO who can see the CA screwing up but wouldn't dare say so?

Of course these situations are so rare that probably most pilots go through a whole career without facing one - we all hope so.

Quoting ADent (Reply 60): BTW all of you talking control summing etc, the PNF (copilot) did not touch her controls per the DFDR, so the topic is not relevant.

Duly noted.

Quoting longhauler (Reply 57): So the FBW computers "know" the difference between slowly rolling the aircraft to a 45 degree banked attitude and swiftly moving the sidestick to one side, "sensing" the urgency.

Quoting longhauler (Reply 57): The pilot also quickly learns that fast control inputs will result in a very quick aircraft movement. (It is very impressive for such a large aircraft).

I thought the rate of roll is determined by how far the stick is deflected, not how rapidly.

Quoting ADent (Reply 60):

Thanks for the excellent summary.

Quoting ADent (Reply 60): No master warning. The RETARD warning is not well understood (they interviewed several other pilots and none of them knew RETARD could sound during takeoff) and did not happen until 80 kts. It was company policy to limit high speed aborts (above 80 kts) for minor threats.

Lets use sound judgement: If you hear a warning coming out that you never heard before... isn't it safer to RTO and investigate what is happening, fill out paper-work and give the chief pilot a call than it is to ignore a clear warning (even though not heard before) and risk a crash....

You are not required to be a certified airline pilot to use common sense. If you are next to a machine you have never seen before that is beeping and alarming, would you put your hand in it?


Just my two cents...

Quoting 744lover (Reply 65):

If they've entered the high speed regime (it seems AA/US uses 80 knots) you typically whittle down what you'll reject for. Fires, engine failures, unsafe or unable to fly. A high speed reject for what could be a minor problem can create a much more unsafe situation than taking a minor problem airborne. A master caution for an AC pack tripping off is a non event to take into the air compared to a possible brake fire or running off the end. In the high speed regime if it's something weird (in this case getting the Retard aural) you have to decide if it falls into the unsafe to fly category and fast.

Quoting hivue (Reply 64): I thought the rate of roll is determined by how far the stick is deflected, not how rapidly.

Just like a normal control column directly connected to actual flight controls, both are being "sensed" ... movement and rate of movement. Namely moving the stick slowly to a desired attitude will have a very different result than moving the stick quickly.

Quoting longhauler (Reply 67): moving the stick slowly to a desired attitude will have a very different result than moving the stick quickly.

Not trying to be argumentative -- I'm not qualified to argue on this subject.   But FBW is an interesting subject to me.

Consider the following 2 hypothetical scenarios: (1) the pilot moves the stick left at X deg/sec (stick movement here, not roll rate) for 25% of its full travel then holds the stick there indefinitely (does not return it to neutral); (2) the pilot moves the stick left at the same X deg/sec for 100% of its full travel then holds the stick there indefinitely. In both instances the airplane will eventually wind up inverted, right? And since the stick was moved at exactly the same speed you're saying the roll rate will be the same in both instances? If that's correct, what's the difference between partial stick deflection and full stick deflection?

I have read that the Space Shuttle's digital FBW was what was called "rate control attitude hold," meaning stick movement commanded roll rate about the pertinent axis rather than attitude. I have always assumed Airbus uses the same basic FBW philosophy (don't know about Boeing).

Quoting hivue (Reply 68): the airplane will eventually wind up inverted, right?

I should have said "at maximum bank angle."

Quoting hivue (Reply 68): And since the stick was moved at exactly the same speed you're saying the roll rate will be the same in both instances? If that's correct, what's the difference between partial stick deflection and full stick deflection?

We'll start with this one first.

When you hold the stick anywhere but neutral, attitude will change. That is because the sidestick is setting an attitude and not a movement .... but the FBW system is using a movement to achieve that attitude. If you release the stick, the attitude will remain the same, and the aircraft will do it's level best to maintain that attitude. (Understand though, there are limitations with regard to aileron, roll spoiler and elevator/stab capability ... not to mention power).

If you slowly eased the stick to one side, the aircraft will slowly respond. If you kept the stick deflected, the aircraft will continue to change attitudes, release the stick .. THAT attitude is maintained. Move it slowly, the rate of attitude change is slow, increase the stick deflection and the rate increases. Or start it quickly, and the rate is fast from the start ... the end result is the same, and that gets me to your next question.

Quoting hivue (Reply 68): (1) the pilot moves the stick left at X deg/sec (stick movement here, not roll rate) for 25% of its full travel then holds the stick there indefinitely (does not return it to neutral); (2) the pilot moves the stick left at the same X deg/sec for 100% of its full travel then holds the stick there indefinitely. In both instances the airplane will eventually wind up inverted, right?

This is more a question about limits in case you were wondering. In Normal Law, the aircraft has set pitch and bank limits. Once those limits are reached, the aircraft will not exceed them, no matter how far you deflect the sidestick. And, as you would expect, the sidestick actually "pressures" back against the input, the closer you get to those limits.

This artificial feel is for two purposes. Firstly to make it feel like a normally rigged aircraft, ie. non FBW. Secondly, to remind the pilot you are getting close to a limit that you likely didn't intend to near.

Quoting hivue (Reply 68): Not trying to be argumentative

Not at all, it IS a very interesting concept. Using FBW, Airbus is able to make the A318 to the A321 all "feel" the same. To me, that is a feat of engineering. One can jump between all narrow body Airbuses and even though they are very very different aircraft, they fly more or less the same!

Also, you can have various levels of failure and control surface loss, that you may not even know about, because the FBW system is compensating using other control surfaces.

What I found the most amazing though, is that it seems "right". It feels very natural and it doesn't take long to really like the way the aircraft flies. Honestly .... I found it hard to go back to an old Boeing cockpit, (B767) after having flown many hours in the A320 series.

I meant to post this update a week or so ago:

The aircraft has been moved to the far east end of Atlantic Aviation's ramp. They've created a dolly to act in place of the nose gear. It is still wrapped in Blue and it has been re-secured to concrete blocks on the ground.

Last I heard they were going to get it to the river somehow and put it on a barge. I'm not sure if that has just been put off or if the plan has changed.

Quoting caoimhin (Reply 54): What's the rationale behind this? Setting aside the CRM component of dual inputs, if both sticks are pushed forward halfway, neither PF nor FO are commanding full deflection. Why would the computer not interpret this as both inputs commanding half deflection? Is there a scenario in which adding the inputs together would be necessary or helpful?

To understand the rationale, you have to understand when would you actually have two inputs on both sidesticks. As you would imagine, it would be very very rare that both pilots would feel the need to "fly" the aircraft.

The most common reason might be not flaring enough on landing and the other pilots gently adding a bit more flare. (This would be most common in training, and the training Captain would also verbalize what he is doing). If the control inputs were averaged, instead of algebraicly added, then less flare would result ... and that would not be the intent. Not to mention a hard landing!

A less common reason would be some sort of urgency where both pilots jumped on the sidesticks at the same time. Likely, (hopefully) their input would be the same, and in the same direction. But if not, it would be like a Boeing and no movement would result ... ie. in a Boeing if both yokes are moved but in opposite directions then the control clutch would disconnect and one elevator would go up and the other down! (Or both ailerons/roll spoilers would move in the same direction instead of the normal opposite direction).

Yes, it is well thought out.

Quoting hufftheweevil (Reply 71): Last I heard they were going to get it to the river somehow and put it on a barge. I'm not sure if that has just been put off or if the plan has changed.

Sounds about right. The WN 737 that was written off at LGA was srinkwrapped in the same blue stuff before it was barged up to Albany for dismantling.

Quoting Spacepope (Reply 73): Sounds about right. The WN 737 that was written off at LGA was srinkwrapped in the same blue stuff before it was barged up to Albany for dismantling.

Why can't they just be dismantled on site?

Quoting hufftheweevil (Reply 74): Why can't they just be dismantled on site?

It depends, but more than likely AA isn't going to do the scrapping themselves. So usually it's up to either the owner/lessor or the scrapping company. There's always the whole not doing it at one of your major airports for millions of your passengers to see aspect too.

Quoting longhauler (Reply 70): What I found the most amazing though, is that it seems "right". It feels very natural and it doesn't take long to really like the way the aircraft flies.

I found the same thing the first time that I flew an aircraft with a side stick and then an aircraft with rams horn yolks.

Quoting hivue (Reply 68): I have always assumed Airbus uses the same basic FBW philosophy (don't know about Boeing).

The best way to look at the Airbus FBW (from how it has been described to me) is that stick movement isn't actually flying the aircraft. Rather it is telling the FBW computers what you want the aircraft to do and how fast you want it done. The computer then determines using all of the information available to it what is the best way to achieve those commands. As opposed to traditional control systems where when you move the stick you are commanding the specific control surfaces to deflect and the aircraft to change attitude as a result.

Quoting hufftheweevil (Reply 74): Why can't they just be dismantled on site?

In addition to the above mentioned reasons there are also environmental concerns associated with scrapping an airplane and those requirements may not have been implemented at that Airport and getting them implemented (SWPPP for example) may not be cost effective.

Quoting apfpilot (Reply 76): The best way to look at the Airbus FBW (from how it has been described to me) is that stick movement isn't actually flying the aircraft.

... as it is the case with all aircraft equipped with an EFCS ( electr(on)ic flight control system ), be they Airbus, Boeing, Dassault, Bombardier, Gulfstream, Sukhoi... F-22, F-16 and so on...
It would be better to learn rather than been told.
That's the onlmy simple way to avoid misconceptions / misunderstandings and eventually bias.

Quoting apfpilot (Reply 76): an aircraft with rams horn yolks.

Never heard of eggs for controlling flight. Please enlighten me.

Quoting Pihero (Reply 77): Never heard of eggs for controlling flight. Please enlighten me.

Cluck Cluck

Quoting Pihero (Reply 77): ... as it is the case with all aircraft equipped with an EFCS ( electr(on)ic flight control system ), be they Airbus, Boeing, Dassault, Bombardier, Gulfstream, Sukhoi... F-22, F-16 and so on... It would be better to learn rather than been told. That's the onlmy simple way to avoid misconceptions / misunderstandings and eventually bias.

Agreed, My point is that the way that it is implemented on the Airbus system is a bit different from the Boeing method. Not that one is better or worse, in fact as described I think after a couple of Hours the Airbus system sounds more intuitive.