This would be quite a confession.
The list of value-destroying and quality-destroying spin-offs in corporate America is legion. But a few insiders and I-bankers make guaranteed F***-you money off doing them every time regardless of the longer-term outcome.
So they remain very popular.
Folks raised in a single-pilot mindset (which included me) have quite a transition to make to properly operate a multi-pilot or even multi-crewmember airplane. Human nature does not find that a natural act.
Part of my get-acquainted briefing to FOs was
There’s always 5 things happening: What you think is happening, what I think is happening, what the automation thinks is happening, what the controller thinks is happening, and what the real stick-and-rudder airplane is actually doing. There’s no guarantee about who’s right, but the airplane wins every f***ing time.
Our job is to a) detect and correct any discrepancies between the 5 versions of reality before they get big enough to matter, and b) make the real airplane do the right thing. Once we’re F***ing sure we know what that is. Every time all the time. An undetected discrepancy means we are quite literally already starting to lose control of the situation and the airplane.
It took a lot of years for me to distill that into that few words. But IMO they’re words to live by.
yep … the “economies” (and boni) start pouring in next quarter … but if you are lucky it takes 5-10 years (or longer) until your reputation is shot forever.
e.g. HP, Sears, Brooks Brothers, …
The accident report (linked above, but here it is again) is really good at specifying these exact failures every step of the way, starting on page 122*:
“During this time, there was no FMA cross check and verbal confirmation, related to “AP-1+2, A/THR SPEED 250 kts, G/S LOC blue” and “THR IDLE, LOC*” by flight crew, indicative of lack of airmanship and adherence to SOP (Golden Rules for Pilot FCTM / AOP / 40 & Glide Slope Interception from above FCTM / PR / NP / SOP / 190 / GUI).”
Their initial mistake was a shared erroneous mental picture (your #1 and #2) of what the plane was physically doing – in this case, its altitude – (your #5) – because they’d both forgotten that the computed flight plan – in this case, a holding pattern – (your #4) would override the chosen autopilot setting (your #3).
Not a fatal mistake at all (those would be:1. Not going around from an unstabilized approach; and 2. Not communicating about the gear-up condition) – but indicative of the multiple “detect and correct” failures to come.
*Actually, it started earlier:
P. 120: “It was observed that no FMA callout or check (FL50 - check) were performed by either of flight crew, they were discussing various topics not relevant to this phase of flight. Indicative of lack of application of procedures / adherence to SOP (Golden Rules for Pilot FCTM / AOP / 40).”
P. 119: “After clearance of PIA 8303 to descent FL100 direct to MAKLI, there was neither FMA callout nor cross check of FMS setting by flight crew (“DIR TO MAKLI” and “DES Green – FL100 – Check”). As per the Airbus Golden Rules for pilots provided in the FCTM, understanding / monitoring the FMA is required at all time and it also includes cross-checking of all FMA changes.”
FMA refers to this:
(I love this caveat in the description of the fourth column, “approach annunciations”:
“These annunciations show the aircraft’s capabilities and are independent of the actual airport and runway capabilities.”)
Also explained well here, why the pilots must check the FMA so often, and the PF must verbally announce each change in it, and the PM must verbally say something (like “FMA?”) if she doesn’t, (Link is to 9-minute YouTube vid; key info is in the first three minutes.)
(Toward the end he well explains what “armed” settings are for – e.g., having the lateral mode pre-set to switch from “heading select” to “localizer” when the plane reaches a localizer beacon.)
(I meant “switch from ‘heading select’ to ‘glide slope’”).
I didn’t realize that those two airlines were that big.
They’re not; even if they had merged the resulting combined airline would still be smaller than the Big 4 US airlines.
I hope you got a chuckle.
Yesterday’s Last Week Tonight with John Oliver (March 3) had Boeing as the main story, and boy did Oliver excoriate them! Some of the things I knew, most I did not. As mentioned upthread, the merger with McDonnell Douglas was a major factor in Boeing’s decline. A few things that come to mind … Boeing’s move of its corporate headquarters from Seattle to Chicago in 2001 (later relocated again to Arlington, Virginia) was intentionally to get executives away from day-to-day involvement in production. Pervasive outsourcing was and remains a major contributor to quality problems. At one point Oliver posted comments from employees building a new airliner (I think it was the 787 Dreamliner) where two-thirds of them said they wouldn’t want to fly on the thing. And as quality control continued to deteriorate, Boeing on multiple occasions spent tens of billions at a time on stock buybacks to boost the stock price. There’s lots more – definitely worth a look if you have access to it.
I’m surprised the episode is online but this appears to it (John Oliver’s coverage of the problems at Boeing):
True. I hadn’t delved into the detail at the time and defaulted to “they” because it was ultimately the crew who ballsed it up. It should be noted that the investigators don’t know who selected the gear up, but it’s a reasonable assumption that the FO did it.
On The Reading of FMAs
This procedure is not standardised and how much it is applied varies from airline to airline and manufacturer to manufacturer. @LSLGuy is a Boeing guy and, he can correct me if wrong, Boeing do not require reading every little FMA change. Airbus, on the other hand, would like us to “talk the aeroplane in to the air”, as some put it.
I’ve flown only three types of transport aircraft, the Dash 8, BAe146, and A320. Each has been for a different company, though the Dash 8 and BAe146 were for the same parent company and the procedures had a common ancestor. Each has had different levels of announcing and checking mode changes. My current employer also has Dash 8s but I haven’t flown them here and their procedures are quite different.
In the Dash 8 we announced mode changes we made and we did this by reading the FMA from the EADI (electronic attitude direction indicator). We did not announce and check the actual buttons we pushed. So we would push the vertical speed mode button silently, then read “VS” from the EADI.
In the BAe146 we started out announcing the button we pushed after silently verifying from the FMA that we had selected the correct mode. Some pilots would get lazy with this and just announce the button they thought they’d pushed and never verify from the FMA.
“Air speed hold selected”
“Ahh, no, you’ve got vertical speed there.”
“Oh yes, so I do, thank you, now it’s air speed hold.”
“Checked.”
If both pilots are having a bad day and the error isn’t trapped, this can lead to mode confusion and bad things happening, such as altitude busts.
Then we started operating the Avro RJ version of the 146 which had a more modern autopilot and we employed some fresh people in the training department with a lot of European experience flying the RJ. They brought with them some new ideas about mode awareness. We would now announce the button we pushed followed by announcing the actual mode from the FMA. “VS selected - VS”, “Checked”.
We would also announce mode changes that were not initiated by the pilot. Previously we would have waited for the autopilot to capture the selected altitude and say, for example, “maintaining flight level two six zero”. Now we would observe the FMA change to ALT* and announce “ALT star” then once the capture was complete the mode would change to ALT and we would announce “ALT - maintaining flight level two six zero”. This resulted in an excessive amount of talking in my view. I didn’t have a problem with reading the FMAs but I thought that announcing the button pushed as well as the FMA was overkill. But I was a good boy, so I did as I was told.
Airbus are quite clear with how they want you to do it. You should announce every mode change, whether pilot initiated or automatic, from the FMA. They make this easy to do by highlighting every new FMA with a white box for several seconds so your attention is drawn to it and a downgrade to a basic mode is also accompanied by a loud triple click through the cockpit speakers. You are also to announce armed modes as “blue”, eg “Heading, nav blue”, means heading is the active mode and nav is the armed mode.
The image below shows the sequence of FMAs for arming the ILS then intercepting and capturing the localiser. Note the white highlight boxes around the new active modes (the yellow circles around G/S and LOC aren’t part of the display).
Steps 2 and 3 are combined because they happen pretty much simultaneously.
- “CAT 3 dual, A P 1 plus 2, glide-slope, loc blue” / “checked”.
- “LOC star” / “checked”.
- “LOC” / “checked”.
So all of this is happening throughout the flight, from taxi (“climb, nav, 34000 blue” / “checked”) to after landing (“spoilers, reverse green, brakes medium, decel”). It can seem cumbersome at times and a good crew working well together don’t need it to be safe but it does go a long way towards ensuring a common shared mental model between pilots and aircraft. Also a good crew working well together and a crew who has slightly lost the plot and aren’t working well together can easily be the same two people at different stages of the same flight. The “good crew working well together” isn’t a static entity that either exists or doesn’t, it’s something that must be nurtured and maintained throughout the working day.
It seems the Pakistani crew didn’t do any of this, and they needed to.
By contrast to all of the above, my Boeing B777 manual, as operated by my employer, simply says, “The crew must verify manually selected or automatic AFDS changes.” No mention of announcing mode changes, they must just be “verified”. It then goes on to say “Announcing changes on the FMA and thrust mode display when they occur is a good CRM practice”. So it’s a good idea, but not mandatory. This same paragraph is reproduced in my employer’s Dash 8 manual, so I can only assume it was written by the airline and not Boeing or Bombardier / De Havilland. Which is another topic in itself. I operate an Airbus in accordance with an Airbus flight crew manual, but I can’t tell which bits of the manual were written by Airbus and which bits are company specific.
Yes, the aircraft knows what it can physically do but it doesn’t know if the airport and flight crew are approved. To do an automatic landing in an Airbus is really easy, you just push the buttons for a normal ILS and leave the autopilots engaged. An automatic landing is the default and it will only downgrade if it loses internal capability, for example an autopilot fails or is intentionally disconnected. It will happily attempt to autoland at any airport from an ILS approach if you let it.
Indeed. One of the hardest things I found, when transitioning to multi-crew flying, was to put into words what I wanted the other pilot to do for me when my workload was high. If I was hand flying an NDB approach in the simulator with a failed engine, I knew I wanted the heading bug to be just about there, but that’s no good as an instruction to the pilot in the other seat, I needed to be able to say “set heading 310” but I found it a lot easier to just reach up for the heading knob and set it myself. This would prompt a comment of “I could’ve done that for you”. Yes you could have, except I couldn’t find the words to ask you to do it!
Part of improving that area was to get good enough at the basic flying tasks so that the talking part of the brain was given enough blood flow to be able operate the mouth, voice box, and lungs in a way that something intelligible came out. When single pilot you can hold your mental model inside your brain, there’s no need to share it or make use of external sources other than ATC.
That’s good. One of my early multi-crew instructors was big on resolving discrepancies. Never let a discrepancy just sit there without resolving it. If ATC clear us to an altitude, I read back “7000 feet” and the other pilot sets 6000’, I will confirm the cleared altitude with ATC so we can both hear it and agree we have the right number, even if I’m 100% sure that what I originally heard and read back was correct and the other pilot was wrong. If I didn’t do that, I’d be getting an uncomfortable feeling in my stomach, it’s become a compulsion to clarify any possible confusion.
Thanks, @Richard_Pearse – all good to know.
This is all true.
At a very high level, historically each airline pretty well developed their own procedures over the decades since the company was founded. And tried to maintain continuity over time, and as much commonality between fleets as possible to ease the inevitable transition as each pilot goes from newbie to retiree through 30 years of planes, promotions, and moves.
Which often led to wry comments like "why do we do [this] like [that]? It seems silly."to which the answer was “Because we did it that way on the DC-6”. There’s always a cost to pay to change procedures and habits. Which can lead to stasis htat eventually becomes obsolete stasis. See also “backwards compatibility” in IT.
What has been happening over the last 15-20 years is two things: Each manufacturer is trying as hard as their legacy designs permit to make every airplane in their product line as similar to operate as possible. As well, each manufacturer is pushing that all their airline customers fly the jet the manufacturer’s way, NOT the legacy airline way. And the carriers are slowly coming along.
It helps that many airlines are all Airbus or all Boeing. And in fact this manufacturer-level procedural standardization is part of their open secret plot to trap “their” carriers as captive customers unwilling to buy from the Other Guy.
Airlines that already operate both types have a bit of a conundrum. Or when dissimilarly equipped carriers merge. There are a myriad of ways the philosophy of the two manufacturers differs. FMA change coordination is but one. There is no single unequivocally better way to solve the “We have both Airbus and Boeing equipment and we want all our crews doing the same things the same way and we want to fly each airplane per its manufacturer’s philosophy.” There’s just no squaring that circle.
Turning from generalities to specifics …
The whole autoflight & FMA system on Boeings is similar enough that anyone familiar with only one type could easily ride along and kibitz a crew operating the other type. It’s a bit more than a matter of dialect, but that’s a decent analogy. And anything that seemed unfamiliar or weird about the systems or crew procedures on first observation would quickly be figured out as you watch more mode transitions occur.
IME, some Boeing carriers have imported more of the “verbalize everything” method into their procedures. Others still prefer the “silent except for exceptions” approach. All this ties in a bit with who’s doing what. If the PF is asking the PM to push buttons, that inherently coordinates & synchronizes their common intent. In and of itself it doesn’t do anything to confirm the asked-for button got pushed or that it elicited the expected reaction from the computers. But it is priming both parties to know to look at what they got.
Informally, many of us talk (or talked) about what we’re planning, doing, and observing more than others did.
That is soooo true!
When you’re a little behind, or task saturated or whatever, you want that change accomplished instantly as you think of it. Holding the need to say it in your mind, saying it, watching them do it, and observe the reaction is burning up a lot more short term memory and time than you have.
As PF you need to pace what you’re asking them to do; You can’t ask for 5 things in 3 seconds. And their results of each request aren’t instant. That takes a lot of getting used to as well. Getting good enough at your particular jet and its particular tasks such that you have spare “brain bytes” to do all that takes a bunch of practice.
That was great. I’m sending to a pilot friend of mine. Thanks!
Canadian flight with 5 on board crashed killing all 5 on board. Tried to land on a highway but was unsuccessful.