Dear experts! I am working in Human Factors area (not in Aviation) in decision supporting systems, and right now I am searching for the incidents reports about situations, when the pilot's intervention to autopilot's/automatics actions made the situation worse. Of course, we never can truly compare with IF situations, but I am looking for cases when relying on automatics (and not taking much actions) could lead to better outcomes by the expert's conclusions. May be there is a set of famous incidents of this type, which are well-known in the community. Will be very grateful for your help!

"There is no problem so bad you can't make it worse." - Chris Hadfield.



I don't know if these fit your requirements exactly, but certainly they are accidents where misunderstanding automatics was a major factor. In the first two, reverting to less automated modes would most likely have changed the outcomes.

- Asiana 214. https://en.wikipedia.org/wiki/Asiana_Ai ... Flight_214
"Over-reliance on automation and lack of systems understanding by the pilots were cited as major factors contributing to the accident."
"The NTSB further determined that the pilot's faulty mental model of the airplane's automation logic led to his inadvertent deactivation of automatic airspeed control."

- Emirates 521: https://en.wikipedia.org/wiki/Emirates_Flight_521
The flight crew did not effectively scan and monitor the primary flight instrumentation parameters during the landing and the attempted go-around. The flight crew were unaware that the autothrottle (A/T) had not responded to move the engine thrust levers to the takeoff/go-around switch (TO/GA) position after the commander pushed the TO/GA switch at the initiation of the FCOM ̶ go-around and missed approach procedure.[33]

- Air France 296Q. https://en.wikipedia.org/wiki/Air_France_Flight_296Q
Automatics were not a cause per se, but the pilots seemed to demonstrate a belief that Alpha Prot and Alpha Max would keep them out of trouble. These mechanisms are not magic and cannot break the laws of physics..






For flavour, here are some quotes from the Airbus FCTM chapter "Golden Rules for Pilots".

- Rule 1. Fly. Navigate. Communicate. In this order and with appropriate tasksharing.

- Rule 2: "Use the appropriate level of automation at all times".
"Determine and select the appropriate level of automation that can include manual flight."

- Rule 3. "Understand the FMA at all times."

- Rule 4, "Take action if things do not go as expected."
"If the aircraft does not follow the desired vertical or lateral flight path, or the selected targets, and if the flight crew does not have sufficient time to analyze and solve the situation, the flight crew must immediately take appropriate or required actions, as follows:
The PF should change the level of automation:
‐ From managed guidance to selected guidance, or
‐ From selected guidance to manual flying.
The PM should perform the following actions in sequence:
‐ Communicate with the PF
‐ Challenge the actions of the PF, when necessary
‐ Take over, when necessary."

I seem to remember some Airbus incident wayyyyy back in the 90s outside of Moscow, repeated stalls of the aircraft as the crew fought against the automation. I remember it being attributed to the crew not understand the aircraft was actually right and that they shouldn’t have fought the computers.

Edit, here it is.

https://aviation-safety.net/wikibase/147079

Then there is this TAROM incident outside Paris.

https://m.youtube.com/watch?v=JlMdMTy8JLg

AC320tech wrote:

I seem to remember some Airbus incident wayyyyy back in the 90s outside of Moscow, repeated stalls of the aircraft as the crew fought against the automation. I remember it being attributed to the crew not understand the aircraft was actually right and that they shouldn’t have fought the computers.

Edit, here it is.

https://aviation-safety.net/wikibase/147079

Then there is this TAROM incident outside Paris.

https://m.youtube.com/watch?v=JlMdMTy8JLg

I wouldn't say the aircraft is "right". I would say that it is operating as intended. In such cases, the crew is unable to follow the priority of the golden rules.

- Rule 3. "Understand the FMA at all times."
- Rule 4, "Take action if things do not go as expected."

Put another way, your actions may have unintended consequences if you do not understand the aircraft in the first place.

There were two China Airlines (Taiwan) A300 accidents that might apply here. IIRC, one was at Nagoya in the spring of 1994 and another at Taipei (TPE) somewhat later.

NGO: https://simpleflying.com/china-airlines ... ash-story/

TPE: https://en.wikipedia.org/wiki/China_Airlines_Flight_676

I began flying into NGO just days after that accident and was surprised at how small the impact area was. RIP to all.

Velocirapture wrote:

There were two China Airlines (Taiwan) A300 accidents that might apply here. IIRC, one was at Nagoya in the spring of 1994 and another at Taipei (TPE) somewhat later.

NGO: https://simpleflying.com/china-airlines ... ash-story/

TPE: https://en.wikipedia.org/wiki/China_Airlines_Flight_676

I began flying into NGO just days after that accident and was surprised at how small the impact area was. RIP to all.

China Airlines 676 seems like a classic case of not understanding the FMA (or equivalent) at all times. The pilot flying thought the autopilot was engaged when it was not.

These days, an overeliance on automatics seems to be much more of an issue than an under reliance. Asiana 214 and two Emirates incidents have demonstrated this. AF 447 also involved poor manual flying skill in a crisis situation.

N1120A wrote:

These days, an overeliance on automatics seems to be much more of an issue than an under reliance. Asiana 214 and two Emirates incidents have demonstrated this. AF 447 also involved poor manual flying skill in a crisis situation.

I would categorize Asiana 214 and Emirates 521 more as "incomplete and/or faulty understanding of automation" than overreliance. At least, the level of automation that the pilots thought was active did not see inappropriate to me.

AF447 showed many deficiencies, of which poor manual flying skill was one. In their defense, the pilots did not have any real exposure or training to high altitude upsets. That sort of training became standard because of AF447.

Go backseat a few IFR flight lessons at your local flight school and you'll see it in action. Student gets a bit flustered and punches AP, alt, and heading, not realizing that the bug is pointing somewhere else....

That's why the sequence needs to be:
1. FD on
2. Select modes
3. Point and read out modes on the scoreboard to validate: https://code7700.com/pointing_and_calling.htm
4. Engage AP and hover over the disconnect button
5. Confirm it's doing what you think you asked it to do, else disconnect and start over

I think the challenge with Airbus' rule 3 is that people sometimes don't know that they don't know something, hence rule 4, which in this case implements rule 2 (don't get what the box is doing, shut it off and hand fly).

American 965 is another good case study. I teach this as a caution to using direct-enter-enter and validating inputs, especially when you're in a rush.

HDG knob has “push to sync” there for a reason. Push, then select HDG mode.

GalaxyFlyer wrote:

HDG knob has “push to sync” there for a reason. Push, then select HDG mode.

That, or keep it in roll/level mode until validating the input. It's fun to teach AP use in the sim because you can let the situation develop further without it becoming a safety problem. Law of intensity at work.

It’s fun watching Embraer pilots adjust from a nonstandard FCU to a standard FCU on non Embraer planes.

For some reason Embraer decided to configure their FCU opposite to every other aircraft manufacturer so sometimes when they’re not thinking they’ll turn the heading knob when they meant to turn the speed knob.

Same thing when you go from a not Embraer plane to an Embraer you have to watch you don’t turn the wrong knob in high task saturation moments and verify those FMAs.

Starlionblue wrote:

N1120A wrote:

These days, an overeliance on automatics seems to be much more of an issue than an under reliance. Asiana 214 and two Emirates incidents have demonstrated this. AF 447 also involved poor manual flying skill in a crisis situation.

I would categorize Asiana 214 and Emirates 521 more as "incomplete and/or faulty understanding of automation" than overreliance. At least, the level of automation that the pilots thought was active did not see inappropriate to me.

AF447 showed many deficiencies, of which poor manual flying skill was one. In their defense, the pilots did not have any real exposure or training to high altitude upsets. That sort of training became standard because of AF447.

Fair, though incomplete and/or faulty understanding can also partially be mitigated by focusing on basic manual flying of the airplane and backing up the automation with physical confirmation and being ready to assure the airplane does what you want.

LH707330 wrote:

GalaxyFlyer wrote:

HDG knob has “push to sync” there for a reason. Push, then select HDG mode.

That, or keep it in roll/level mode until validating the input. It's fun to teach AP use in the sim because you can let the situation develop further without it becoming a safety problem. Law of intensity at work.

Fair enough, as PF, I always had the FD mode selected, FMA checked and cues lined up before asking for A/P ON. Never recall seeing ROLL/PTCH. Different types

GalaxyFlyer wrote:

HDG knob has “push to sync” there for a reason. Push, then select HDG mode.

I guess all planes have their traps, but I never got that "feature" on Boeing. Is there any reason the heading bug isn't automatically to the current heading while in LNAV?

On Airbus, the heading selection in the FCU (glareshield) is blank in in NAV (LNAV). If you pull the heading knob for HDG mode, you simply get the current heading. This seems less prone to error.

Also, if you enter a Direct To in the FM and insert, and you're in HDG, NAV will automatically activate, and as before the heading selection will go blank on the FCU. This does lead to the call "ready to pull for heading?" if you're programming an intercept, but it at least spares you from forgetting to engage NAV.

LH707330 wrote:

GalaxyFlyer wrote:

HDG knob has “push to sync” there for a reason. Push, then select HDG mode.

That, or keep it in roll/level mode until validating the input. It's fun to teach AP use in the sim because you can let the situation develop further without it becoming a safety problem. Law of intensity at work.

Always know what you're about to do before doing it. If in doubt, stop for a second and think about it. If still in doubt, ask the other guy.

When entering stuff in the FM, I try to type, pause, look again, and only then send the input to the appropriate box. Especially for performance data!

GalaxyFlyer wrote:

LH707330 wrote:

GalaxyFlyer wrote:

HDG knob has “push to sync” there for a reason. Push, then select HDG mode.

That, or keep it in roll/level mode until validating the input. It's fun to teach AP use in the sim because you can let the situation develop further without it becoming a safety problem. Law of intensity at work.

Fair enough, as PF, I always had the FD mode selected, FMA checked and cues lined up before asking for A/P ON. Never recall seeing ROLL/PTCH. Different types

This is especially important after something like windshear, TCAS, or an upset. Veryify the selections before pressing that magic AP button, or you might find yourself in a new fun situation just after you exctricated yourself from the previous one.

Starlionblue wrote:

I guess all planes have their traps, but I never got that "feature" on Boeing. Is there any reason the heading bug isn't automatically to the current heading while in LNAV?

Sometimes it is useful to be able to "pre-select" a heading when still in LNAV. At my home airport a common instruction I get when on a particular arrival is "Depart (fix) heading 160, vectors ILS..." This is different from the published arrival course but is nonetheless ATC's usual method of sequencing traffic. As we usually get the instruction some distance before the fix in question, it is nice to be able to set the heading bug to 160 as soon as we get the instruction, then simply push HDG SEL as we approach the fix. As I've never flown an Airbus I can't speak to the differences but I would think the ability to do what I described would be useful.

That being said I can't for the life of me understand why Boeing doesn't include some kind of one-push heading sync (or for that matter, a backcourse mode) on the 737.

Starlionblue wrote:

GalaxyFlyer wrote:
HDG knob has “push to sync” there for a reason. Push, then select HDG mode.



I guess all planes have their traps, but I never got that "feature" on Boeing. Is there any reason the heading bug isn't automatically to the current heading while in LNAV?

On Airbus, the heading selection in the FCU (glareshield) is blank in in NAV (LNAV). If you pull the heading knob for HDG mode, you simply get the current heading. This seems less prone to error.

Also, if you enter a Direct To in the FM and insert, and you're in HDG, NAV will automatically activate, and as before the heading selection will go blank on the FCU. This does lead to the call "ready to pull for heading?" if you're programming an intercept, but it at least spares you from forgetting to engage NAV.

The Airbus setup does seem much less error prone with that design. The situations where I'd want to see a heading select when in nav mode are few, so I'd rather have the idiot proofing and deal with that occasional frustration. Having a heading and CDI needle can also be distracting for some, so having it disappear probably also helps avoid fixation on the wrong instrument.

GalaxyFlyer wrote:

LH707330 wrote:

GalaxyFlyer wrote:

HDG knob has “push to sync” there for a reason. Push, then select HDG mode.

That, or keep it in roll/level mode until validating the input. It's fun to teach AP use in the sim because you can let the situation develop further without it becoming a safety problem. Law of intensity at work.

Fair enough, as PF, I always had the FD mode selected, FMA checked and cues lined up before asking for A/P ON. Never recall seeing ROLL/PTCH. Different types

That makes the most sense, that's how I do it when I'm using the AP. I teach students on single-pilot G1000 aircraft. The default AP modes there are roll and pitch level, then you select another mode and it does that. Part of the learning process with those is double checking the inputs before engaging, otherwise the nasty surprises await. Better to learn that skill set and develop a thorough respect for the AP on button early in IFR training than later on IMHO.

ArcticFlyer wrote:

Starlionblue wrote:

I guess all planes have their traps, but I never got that "feature" on Boeing. Is there any reason the heading bug isn't automatically to the current heading while in LNAV?

Sometimes it is useful to be able to "pre-select" a heading when still in LNAV. At my home airport a common instruction I get when on a particular arrival is "Depart (fix) heading 160, vectors ILS..." This is different from the published arrival course but is nonetheless ATC's usual method of sequencing traffic. As we usually get the instruction some distance before the fix in question, it is nice to be able to set the heading bug to 160 as soon as we get the instruction, then simply push HDG SEL as we approach the fix. As I've never flown an Airbus I can't speak to the differences but I would think the ability to do what I described would be useful.

That being said I can't for the life of me understand why Boeing doesn't include some kind of one-push heading sync (or for that matter, a backcourse mode) on the 737.

I can absolutely see the need for preselecting, and you can also do that on Airbus, but without the need for a constantly on heading bug that needs to be remembered. Turn the heading knob to the desired heading without pulling and it will keep that number for you. Once ready for the turn, pull the knob to go to HDG mode. This is also used for "maintain runway heading after takeoff" instructions. Before setting takeoff thrust, turn the heading knob until it displays the runway QDM, and RWY TRK will automatically activate once airborne.

Yet another way this preselection can be used is when you want to fly non-precision approaches at a constant descent angle (CDA). Preselect -3.0 degrees (or whatever it is) one or two miles before the descent point. Once you reach 0.3nm to the descent point, pull for FPA. You have to pull a bit ahead because it takes a bit of time for the flight path to change.


I guess there's no way to make a plane completely idiot-proof, but I am fascinated by the amount of thought that goes into these things to make them as intuitive and clear as possible, and that goes for all manufacturers.

Dear all, thank you for the interesting discussion and great examples, I am going through them now!

Starlionblue wrote:

GalaxyFlyer wrote:

HDG knob has “push to sync” there for a reason. Push, then select HDG mode.

I guess all planes have their traps, but I never got that "feature" on Boeing. Is there any reason the heading bug isn't automatically to the current heading while in LNAV?

On Airbus, the heading selection in the FCU (glareshield) is blank in in NAV (LNAV). If you pull the heading knob for HDG mode, you simply get the current heading. This seems less prone to error.

Actually the 777-9 has an optional feature that does this. It’s called Enhanced Heading Select. The Heading window blanks when LNAV is engaged. If you push the HDG Switch it syncs to the current heading.

Most customers have taken this feature. One or two have not due to commonality issues with the 787. It would take an MCP hardware change to implement this feature on the 787. It hope it’s added on the 787 eventually.

Woodreau wrote:

It’s fun watching Embraer pilots adjust from a nonstandard FCU to a standard FCU on non Embraer planes.

For some reason Embraer decided to configure their FCU opposite to every other aircraft manufacturer so sometimes when they’re not thinking they’ll turn the heading knob when they meant to turn the speed knob.

Same thing when you go from a not Embraer plane to an Embraer you have to watch you don’t turn the wrong knob in high task saturation moments and verify those FMAs.

Glad they fixed this for E2.

Starlionblue wrote:

Also, if you enter a Direct To in the FM and insert, and you're in HDG, NAV will automatically activate, and as before the heading selection will go blank on the FCU. This does lead to the call "ready to pull for heading?" if you're programming an intercept, but it at least spares you from forgetting to engage NAV.

Couldn't you use direct to radial in on the MCDU, at least when you're intercepting an approach?

xl0hr wrote:

Starlionblue wrote:

Also, if you enter a Direct To in the FM and insert, and you're in HDG, NAV will automatically activate, and as before the heading selection will go blank on the FCU. This does lead to the call "ready to pull for heading?" if you're programming an intercept, but it at least spares you from forgetting to engage NAV.

Couldn't you use direct to radial in on the MCDU, at least when you're intercepting an approach?

Yes, you can do this also. And in that case NAV will arm, then activate once you actually intercept.

Starlionblue wrote:

Yes, you can do this also. And in that case NAV will arm, then activate once you actually intercept.

thanks!

An ancient incident regarding relying on the autopilot that resulted in tragedy was KAL 007 which strayed over Russia and was shot down.

This was not an FMS aircraft but had one switch position which determined the course guidance..heading mode or INS track mode (and others). If the heading mode was guiding the aircraft and on an intercept to the INS track, then you could switch it to 'INS' and when the aircraft was within a certain distance would capture the course.

Trouble was, the KAL crew switched it to "INS" but the aircraft's current heading would not have met the capture requirements. They didn't notice that error and the aircraft flew a constant heading into Russian airspace. Wiki has a pretty good writeup.