Why is automation dependency encouraged in modern aviation ?

[safetypee]

vilas, interesting points, #114.
The revised question is challenging. It could be divided further; what standard of training is required to fly (operate) an aircraft in normal circumstances, and then in abnormal situations which generally involve manual flight, but not always in a normal aircraft.

This thread, like many others degrade to hardened extremes; fixed solutions without first reconsidering assumptions, alternatives - the essence of critical thinking. There may not be any one solution, only middle-ground compromise, small improvements in many areas; all requiring the skill of thinking.
Modern aviation suffers the worldly effects of modern education, instant access solutions, unverified web answers, even electronic training - yes no answers.
Pilots still need the skill of thinking, judgement of risk, managing uncertainty, all refreshed every flight. Yet we appear to be happy to let the autos do that (auto dependency), without realising that autos do not think.

KP, the latter part of your post #117 provides the answer to your original question. If you are unable to fly raw data without assistance - modified displays, then you are automation dependent, whether this is encouraged or not.

[Uplinker]

Hi vilas, and safetypee no not really; my suggestion of a logarithmic or expanded centre scale instrument - like the V/S - is simply a different presentation of raw data. It is not a FD - it will not tell you which way to point the aircraft, nor is it automation in the sense we are using it in this thread.
A machmeter for example does internal calculations and presents a reading, but we would not call a machmeter "automation". Our IAS readout arithmetically subtracts static pressure from pitot pressure but that is not what we mean here by "automation", it is still raw data.

In my example of a PA28, the instrumentation makes it a difficult job to track an NDB accurately and consistently. If all the information were superimposed onto one instrument, it becomes very easy. Still raw data, just presented better.

A flight director tells you which way to point the airplane to centre the raw data needles. With raw data alone you have to be able to work it out by yourself. The current ILS displays make it hard to notice initial deviations - especially if one is hand-flying as well. Kaypam suggests a trend arrow, which would require some computer processing, but even that is not a flight director. (Note: He is not suggesting the arrow points which way to roll to correct the deviation, it indicates the trend of the marker movement; the rate and direction of change of the marker and so makes the raw data easier for the pilot to see and assimilate).

The thrust levers - I am talking about a method to physically move thrust levers by small, precise amounts - in any aircraft - for example increasing N1 from 53% to 55%, or to 'null out' the speed trend arrow: just tiny increments. I think vilas is talking about matching the doughnuts to the present thrust setting before disconnecting the A/THR - a completely different subject.

[Vessbot]

Originally Posted by vilas

vessbot, we aren't much in disagreement. But about the above sentence, perhaps the only thing the crew(in both the cases) was doing was keeping the aircraft on the correct descent path without the necessary thrust, manual or auto immaterial. That's why the speed was dropping.

Somewhat true for Asiana, who was within a fuselage length of pavement, but not even close for Indian, who was 2300 feet short (or 3300 from aimpoint, that’s over half a mile.) Dropping airspeed pays for some amount of flight path, but not necessarily all of it; Indian was low on both.

And when Airbus crew at 400ft realised the wrong mode the only solution was to try another time.

Fully agreed, a go around was the only sensible choice at that point. I meant “fly the plane” generically to include whatever the correct flight is, I didn’t mean it to press on (and I see how I could have come across that way, especially with previous examples)

[vilas]

I think vilas is talking about matching the doughnuts to the present thrust setting before disconnecting the A/THR - a completely different subject.

No! Not at all. Airbus thrust levers have no friction, no back pressure. Whatever thrust you want just gently push both the levers to place the doughnuts there. That's all.

[FullWings]

Originally Posted by Uplinker

Kaypam, from what you say about control movements, don't worry, you will get used to it.

The control movements can seem weirdly small until you get used to it. For taxiing, take the brakes off and you will normally start to move. If not you advance the thrust levers say a quarter of the available arc and as soon as you see the start of movement, you put them back to idle. Now you will taxi.
If you ever need to make fine adjustments to thrust levers, "walk" them against each other; twist your hand left and right to move each lever at a time by a small amount.

Reducing the pitch if the AP drops out 800' below the level. Gentle pressure on a conventional control column, not a push or even a movement, just gentle pressure.

On Airbus FBW on my initial type rating NOBODY could tell me how to use the side-stick. Even though I asked my TRE and several others, neither they or anybody could tell me. I eventually taught myself after seeing a film of a pilot operating the joystick in a Tornado (military fast jet):

On Airbus FBW, the attitude will stay where it was until a further input is received. So in your level-off scenario push the side-stick forwards against the centring spring one very brief forward push of about 20% of the full travel and immediately let the side-stick return to central. So just a small nudge or jab or bump against the spring, lasting half a second and then centre. So the action is nudge-release in half the time it takes to say that phrase. This will lower the nose a small amount and the FBW will hold the aircraft at this new attitude. If it was not enough, nudge-release again. Same applies in roll. That is how to make very small fine adjustments to the attitude of an Airbus FBW.

Re instruments, I agree, indications are too small for a given deviation or are badly designed, and this is why raw data flying is such a challenge. I remember finding NDB tracking in a PA28 to be very difficult because the NDB needle was on one dial and the heading bug was on another, and there was no bug for the NDB track - you had to remember what it was. One had to continually compare the needle with the heading instrument, and parallax and misreading could occur. Also, you were flying an aircraft that never stayed where you put it, so you were busy hand flying and continually correcting the aircraft and tracking a non bugged NDB needle, it was quite common for the NDB to drift more than 5 degrees out. When I later flew the Dash 8, you could overlay the NDB needle, the NDB track bug and the compass rose all on one instrument and suddenly NDB tracking was a piece of piss ! Instead of having to remember the NDB track and look at a different instrument to read what the heading was and then go back to the NDB instrument, all you had to do was glance at the one composite overlay. You did not even have to read any values, you could see at a glance if the NDB needle was under the track bug, and if it was just one degree out one side, you clicked the AP heading bug towards it by one degree. Really easy.

I have always found LOC and G/S displays to be too limited. By the time you can see a deviation, it is quite a large error. With my engineer's hat on, I would redesign the display so the markers were in two halves. One half would move as they currently do, the other half would move over its whole range of the display for say 1/4 of a degree LOC or 50' G/S - much more sensitive and a large movement for a small deviation - so you would be able to see a deviation before it became too big. The other half would display as it does now. A lot of the time the sensitive marker would be pegged on one extreme or the other, but when you had captured the LOC and G/S they would come off the stops, and a perfect ILS would see all the bugs centred.

And there I rest my case about User Interfaces (no insult to the author!)

It just seems to be the last thing on the minds of those who design this stuff that it might be used by humans. Flying is easy - it’s all the essential workarounds of other peoples’ bad ideas that are the problem...

[Vessbot]

Originally Posted by vilas

not required in airbus, just look at the gauge and move the doughnuts to the requirement thrust will catch-up.

This doesn’t solve the problem that walking the throttles solves. The problem is of a physical nature, when you want to make a very small and precise movement but can’t due to a rough friction nut or what have you. The solution is to change the lever arm from originating at your shoulder, (a few feet) to originating at the non-moving throttle, which acts as an anchor. The much shorter lever arm allows for the precise input needed. (Just think, can you better manipulate a light switch with a yardstick, or your finger directly?) Make a tiny movement, then the moved throttle becomes the anchor for you to move the previously fixed one, repeat.

I do this all the time in a modern FADEC jet even with smooth throttle friction in cruise, to adjust the N1 by a few tenths of a %. And I think sometimes in taxi, to increase the thrust in very small amounts above my initial “yay much” if the plane isn’t moving enough.

[Vessbot]

Originally Posted by vilas

No! Not at all. Airbus thrust levers have no friction, no back pressure. Whatever thrust you want just gently push both the levers to place the doughnuts there. That's all.

To an accuracy of 0.1%? It seems to me that airbus throttles, that are known to have a shorter travel arc (therefore more output change per movement distance) than most planes, would be even tougher to control these small amounts. (lack of necessity due to AT aside)

[Check Airman]

Originally Posted by vilas

Check airman


NTSB Chairman on Asiana SFO

I know that it wasn’t my brightest moment. I was just pointing out how easy it is to slip into complacency, and the dangers that arise. What if the BA 777 crew had been complacent that morning going into LHR?

[vilas]

Originally Posted by Vessbot

To an accuracy of 0.1%? It seems to me that airbus throttles, that are known to have a shorter travel arc (therefore more output change per movement distance) than most planes, would be even tougher to control these small amounts. (lack of necessity due to AT aside)

Yes! It's miniaturisation. But not an issue when done consciously. We all get used to it. May be it was easier for me because I came to A320 from B747 classic where everything was huge and four of them which were rarely aligned. But I flew other twins A300 and A310 never did walking. It helps when there is friction which could be unequal.

[Uplinker]

Apologies, vilas for my incorrect assumption.

However, All thrust levers have some friction, (otherwise they wouldn't stay where you put them!), and walking the thrust levers is a very useful technique for small adjustments - especially on Airbus FBW, where as has been noted, the lever travel and the levers are relatively short.


Like Vessbot, I use this technique to adjust taxi thrust and flying thrust if I want to make small changes, (and have done so on turbo-props and the BAe146 for example as well as Airbus). To adjust thrust from, say, 53% to 55% N1, I can easily over shoot if I do it 'freehand' , but if I walk the levers up I can get my target thrust very precisely.
Note: when I say walk, I am taking steps of a few millimetres at a time, no more.

[Check Airman]

The walking thing seems to be very much a leftover technique. It seemed to be more of a way to avoid inadvertent large power changes.

[vilas]

Since Airbus thrust levers do not move in auto mode there's no spring, any mechanism or cables that can cause any undue friction. They slide evenly when pushed and stay where left. And since it is FBW their position just signals the required thrust and their rate of movement selects an acceleration schedule. If you slam them it selects the fastest schedule. The problem is if coming from other aircraft you are used to large movements. It can happen even with the stick but since the arm is rested only the wrist comes in play. That's Why Airbus insists on proper adjustment of the active arm rest. Compared to yoke movements the stick moves an inch or so. Again takes getting used to.

[tubby linton]

The only time I have seen a problem with athr on the bus is when I used to fly an A320 with different mod engines. The problem was with a cfm-56 dac engine that accelerated much slower than the standard cfm. You really had to wait until they were both at 50% before advancing any further on the take off roll if you wanted to stay on the runway.

[Vessbot]

Originally Posted by safetypee

Vessbot, do find time to read the previous links in https://www.pprune.org/tech-log/636976-why-automation-dependency-encouraged-modern-aviation.html#post10937097

OK, I’ve read them both now. They’re very general, and I can’t immediately find a point that slots into this thread’s discussion. But if you’ve gleaned such a thing, I suppose I’m ready to talk about it now.

Except one sentence, that jumped out in a huge way: “Under stress, decision makers fall back on their most familiar responses, which may not be appropriate to the current situation.”

P.S. re 777 example - you blame the crew and create a workaround for a weak aircraft system (double button press), but might the event be something to do with the aircraft:- https://www.pprune.org/rumours-news/636908-southern-air-777-stall-recovery-after-takeoff-nov-15th.html#post10932069

I’m not saying anything about this event - I have no clue what happened and blame no one. I’m only referring to the double-push technique, that came up in the off-shoot discussion, as an example of a common and immediately understood sequence (even with a catchy name) where I think it’s OK to use the automation solution. (To be clear I’ve never flown a Boeing, I’m just going off of that thread’s posts.) It’s a counterexample to my more general point where if you don’t immediately recognize the sequence and solution, it’s necessary to fly the plane, like all the other accidents we’ve been talking about.

[KayPam]

Originally Posted by vilas

This is what the FD does. A raw data is raw data. You are suggesting a emergency or limited function FD. The technology may provide you an emergency FD using different inputs. As it is A350 gives you AP even with double engine flameout. It does auto TCAS, auto TCAP(TCAS prevention), auto EMER DES. They aren't going to develop anything to help you fly raw data. They will develop emergency AP. You don't need to practice everything raw. How many times FD are lost during RNAV departure and you have no alternative? A direct to or radar vector is always possible. As long as you can climb, descent execute an approach raw data within acceptable accuracy it is good enough.

Yes, the FD does that (or rather, uses that), and actually much more. It also uses a notion of distance to the runway to modify its own sensibility. It uses a computed geographical position and real time turning radius capacity to anticipate the LOC interception.. There are many features embedded in the airbus flight guidance system.
Are you suggesting that DME speed (if you recall my speech about DME arcs using the old system vs the new system) on old bendix king DME receivers is not raw data ?

I am not suggesting a limited function FD at all.
I am suggesting that the aircraft stops hiding useful data that it has, such as LOC trend, crosstrack deviation for RNAV departures and approaches..

It is not about mitigating the consequences of FD failure, you lost the global objective of this topic.
Mitigating the consequences of FD failure would just be a "collateral benefit".
The whole point of the discussion is to allow pilots to fly more comfortably raw data (by presenting useful information), or to merely allow them to do so (by presenting required information)

If pilots have the possibility to fly raw data, they will do it more than if they have no possibility of doing it. Sorry for saying this sentence, which was completely obvious, even tautological.
The whole point of this is to encourage raw data flying, not because it is better, more precise than automatic flying, on the contrary. But because it develops the basic skill that we need to do our job.

All of this is about ergonomics.
You could imagine a PFD displaying the entire possible range of angles of pitch (from -90° to +90°) and two or three times smaller than it is now.
Would that make controlling the pitch easier ?
You surely agree that it would make no sense to do such a modification, because it would make the flying very impractical ?
So why stop at not deteriorating the ergonomics ? The logical thing to do is to improve the ergonomics, each time that it's possible.

Originally Posted by Judd

Fascinating. "Walking the throttles" was a DC3 technique used if the the throttle friction nut was poorly maintained making the throttles stiff to operate causing a jerky operation. I have seen pilots who unconsciously walk the thrust levers of a 737. It becomes a gimmicky habit. The result can be a slight but annoying change of noise or engine synchronisation. Old habits die hard. ...

I learnt very soon that this technique is very useful for fine adjustments.
As I said, because the useful distance traveled by the thrust levers is very very limited.

Originally Posted by vilas

not required in airbus, just look at the gauge and move the doughnuts to the requirement thrust will catch-up.

Are you reading the conversation ? If you want to correct the thrust for 800fpm to 700fpm, you have to move the thrust levers by a distance as short as an ant's bottom.
It sounds like you're working for Airbus. "The aircraft is perfect as is."

[Vessbot]

Not related to anything to do in an airliner, but the discussion on potential examples of what I’ll call “expanded raw data,” especially the logarithmic localizer, reminded me of a very elegant solution to a problem, that I encountered in cropdusting school. In modern times, you’re guided down the swath by a GPS lightbar, which works the same as a localizer: you’ve got the center mark, and a light (horizontal column)that moves left or right, showing which way the swath is relative to you. And, it’s logarithmic! On the swath, one mark of deviation (it’s a row of LED’s) is something like a lateral foot, while if you’re significantly off (including 90 degrees away on the turnaround prior to going down for the swath) you still get a meaningful indication. (If it was 1 light per foot all the way, it would be begged if you’re anything but near-perfect already)

So, that’s beautiful guidance when you’re on the swath and not following a row crop, but how do you provide guidance in the turn-on? With what I’ve described so far, it’s still impossible to estimate your closure rate with the centerline, and you’d have to join it long prior to entering the field. Maybe 10, 20, 30 seconds to adjust? Who knows. This would be highly uneconomical when repeated hundreds or thousands of times, and could not match the historical solution of a guy with a flag that you visually roll in on.

Enter the second row of lights, under the first one. It shows the derivative of your lateral position, or the cross-track rate. It’s also logarithmic, and it’s calibrated so that on the ideal turn-on, the top bar (localizer) and bottom bar (cross rate) shrink together! So, you watch the ends, and play the bottom vs. top. If the bottom bar lags (is longer than) the top, your cross rate is too large for where you are along the turn, so you’re gonna blow past, so you increase (or slow the decrease of) your bank a bit. In the opposite situation, if you’re set to roll out too soon, the cross rate is too small (bottom bar shorter than the top) so you unbank a bit more to get your profile closer to the centerline. Of course, while flying it, you don’t think about the if-this-then-that like I just described, you just watch the end of the bottom bar and adjust your bank so as to manipulate it in the direction you want, using the end of the top bar as the reference.

I only went to school for this for a few weeks and never got a job, but to this day I’m tickled by the elegance of this setup, and how with some practice, it delivers you perfectly centered down the swath and going straight the instant you roll out.

(Of course this can all be done with a much simpler flight director, but there’s the obvious problem of looking intently at the instrument panel as you’re making sure you don’t hit the ground or the wires as you’re pulling out of a rolling dive a few feet AGL.. While this setup is on top of the nose, far away from you, not just in your field of view but also in focus, while you’re looking at the ground and obstacles.

Having said this, they could also design the bottom bar to just show a FD roll command, the goal being to keep it centered all the time. For all I know, this might even exist. Pretty sure the the regulations in that field pose a smaller obstacle to tech progress than those in the transport sector, combined with the lifecycle times of airliner development)

[blind pew]

Another perspective

https://www.flyingthings.org/flying-...sgvqg9otljtgru

Mike Riley touches on the subject in this article in his blog on solo flying..

[pineteam]

Originally Posted by KayPam

Are you reading the conversation ? If you want to correct the thrust for 800fpm to 700fpm, you have to move the thrust levers by a distance as short as an ant's bottom."

On A320 during raw data approach approach you adjust the vertical speed with the pitch and the speed with the thrust levers.
I agree on the walking the levers technique . I do it all the time during approach for fine tuning. Quite difficult otherwise.

[Vessbot]

Originally Posted by pineteam

On A320 during raw data approach approach you adjust the vertical speed with the pitch and the speed with the thrust levers.

So just to make sure I haven’t lost the plot, if you change vertical speed you have to also change thrust right?

[pineteam]

Haha yes of course! But we first change the pitch to adjust the vertical speed then adjust the thrust as required to maintain the speed. We don’t adjust thrust to achieve a targeted vertical speed. Increasing or decreasing thrust on A320 will not affect or barely the vertical speed directly as it’s pitch stable.