Flight Director distraction during recovery from an unusual attitude
We train for the unknown, yet are killed by that already known
Yo gums, et al.
Beware of what is assumed. Would training every pilot for JT type of event assure success. Was JT actually a failure to recover; not directly. Similar to many high profile accidents where the initiators are in the certification process - design and regulation.
Why should operators allocate more time and money on training to add experience, reduce startle, aid awareness, and act appropriately in very rare technical system related situations. The regulators rushed to implement recovery training, which actually manages a weakness in their certification process (AF447, Westair CRJ, JT 737, Asiana 777).
The safety focus has changed; now to avoid situations, on common aspects of LoC - speed, thrust, attitude.
Simulators can be used to practice recovery, less so for avoidance - if successful, avoidance is a non event, often cited as a waste of simulator time. AF447 required reading a checklist and doing nothing, JT required a checklist, the CRJ information / software revision, Asiana information - knowledge.
So back to #1, why are simulator upset exercises being flown, … because the regulations mandate them?
How much time is spent on avoidance or mitigating the initial onset conditions?
What is the form of this training; ‘how to’, and to what effect?
Which deficient systems have been rectified, not 777, CRJ, (JT).
We train for the unknown, yet are killed by that already known - distraction and switching workload required to compensate for a system ‘feature’.
Beware of what is assumed. Would training every pilot for JT type of event assure success. Was JT actually a failure to recover; not directly. Similar to many high profile accidents where the initiators are in the certification process - design and regulation.
Why should operators allocate more time and money on training to add experience, reduce startle, aid awareness, and act appropriately in very rare technical system related situations. The regulators rushed to implement recovery training, which actually manages a weakness in their certification process (AF447, Westair CRJ, JT 737, Asiana 777).
The safety focus has changed; now to avoid situations, on common aspects of LoC - speed, thrust, attitude.
Simulators can be used to practice recovery, less so for avoidance - if successful, avoidance is a non event, often cited as a waste of simulator time. AF447 required reading a checklist and doing nothing, JT required a checklist, the CRJ information / software revision, Asiana information - knowledge.
So back to #1, why are simulator upset exercises being flown, … because the regulations mandate them?
How much time is spent on avoidance or mitigating the initial onset conditions?
What is the form of this training; ‘how to’, and to what effect?
Which deficient systems have been rectified, not 777, CRJ, (JT).
We train for the unknown, yet are killed by that already known - distraction and switching workload required to compensate for a system ‘feature’.
People with extensive GA experience without FD's sometimes have to be reminded to follow the FD in normal ops during early training transitioning on big jet types, they are used to other cues..
Thanks, Stall
A long time ago, with another philosophy in effect, I managed ( earned) to get my fighter assignment from the basic drill. I went to Air Defense Command as a pilot and was from the only training unit that still flew T-33 jets ( Craig AFB)
ADC had such a low opinion of the early "children of the magenta line" that they required 20 or so hours of "instrument" training in the T-33 for all the newbies from the T-38 places Reason was basic instrument flying skills and the T-38 folks had the flight director stuff for TACAN and ILS nav. They also had neat attitude indicators and such, while we had the basic J-8 thingie. So I flew two or three rides and they checked me off. I then flew "target missions" for rest of my "spare" time while going thru F-102 checkout. Loved it, and would fly a Deuce mission and then a target mission later in the day. Along the way I learned a lot about instrument approaches
using the basics that were invented by Doolittle 40 years before.
"Looking thru" to the real world background is easier for some than others, as Stall has implied. I see the biggest problem in that the FD symbology has more precedence on the displys than the actual attitude and airspeed/AoA. In the A-7D, many pilots flew the raw data due to the poor mechanization of the FD box ( loose dampening and no connect with the basic INS/Doppler nav system).
Thanks for helping me understand the concern with following the FD.
Gums sends...
A long time ago, with another philosophy in effect, I managed ( earned) to get my fighter assignment from the basic drill. I went to Air Defense Command as a pilot and was from the only training unit that still flew T-33 jets ( Craig AFB)
ADC had such a low opinion of the early "children of the magenta line" that they required 20 or so hours of "instrument" training in the T-33 for all the newbies from the T-38 places Reason was basic instrument flying skills and the T-38 folks had the flight director stuff for TACAN and ILS nav. They also had neat attitude indicators and such, while we had the basic J-8 thingie. So I flew two or three rides and they checked me off. I then flew "target missions" for rest of my "spare" time while going thru F-102 checkout. Loved it, and would fly a Deuce mission and then a target mission later in the day. Along the way I learned a lot about instrument approaches
using the basics that were invented by Doolittle 40 years before.
"Looking thru" to the real world background is easier for some than others, as Stall has implied. I see the biggest problem in that the FD symbology has more precedence on the displys than the actual attitude and airspeed/AoA. In the A-7D, many pilots flew the raw data due to the poor mechanization of the FD box ( loose dampening and no connect with the basic INS/Doppler nav system).
Thanks for helping me understand the concern with following the FD.
Gums sends...
Interesting discussion. I'd like to add another perspective if I may.
Attitude. Not just the attitude of the airplane, but of the pilot. I started off with basic flight instruments like most pilots of my civilian general aviation background. Later when I started flying jets, we had the Collins FD109. It worked well enough most of the time, but didn't always keep up when capturing VOR radials, altitudes, ILS localizers or glideslopes. So one had to anticipate a bit and try to stay "slightly ahead" of the FD command bars. That meant scanning the raw data and "predicting" when the command bars should react. If the bars moved when they should, follow them. If not, fly "through" them.
The first line Captain I flew with regularly called it the "flight advisor". He cautioned against allowing it to act as an "electronic flight instructor". Though he was joking, there was a serious side to his characterizations. For various reasons, depending upon the flight guidance computer to provide timely and correct pitch and roll commands could sometimes lead to some pretty rough or oscillatory intercepts. The A/P didn't make very smooth intercepts either because it relied upon the same guidance. So if you wanted fly smooth and accurate course or altitude captures, you hand flew it just slightly ahead of the FD command. You could help the automation to work better by making very shallow intercepts, but this is not always practical in the approach environment.
Of course, most of the analogue electromechanical FD issues are all but gone in later generation digital avionics, but the analogue mindset will still serve a pilot well when surprises occur.
Returning to the subject of pilot attitude, I think Captain E. had it right. The flight director is best considered to be an advisor, not an instructor! And if a pilot maintains that mental "attitude" in all subsequent flying, the reaction to an uncommanded change of attitude is more likely to be automatic for that pilot. Click-Click, crosscheck, flight controls and thrust as appropriate. Get it back to where it's supposed to be, (course altitude, speed) THEN communicate, analyze, troubleshoot, checklists etc...
Someone brought up the subject of training earlier. Initial training on two of my types included both unusual attitude recovery and a couple of opportunities to put that training to work in the form of unannounced surprise upsets. Sim instructors the world over have a sadistic streak that can be put to good advantage! During a complex RNAV DP while on A/P, a distraction or two can be introduced along with an uncommanded attitude deviation. Pitch trim runaway is a favorite of course, but wake vortex and several other problems are equally demonstrative and provide a good opportunity to put briefing room theory into practice. I once saw a guy throw up his hands and scream when the airplane rolled over. (the instructor had secretly asked him to) I was supposed to take over like a good FO should. Cardiac stress test passed!
My point regarding pilot attitude is that the pilot must believe in their own ability to fly the airplane on raw attitude and performance data. This belief can only be earned by doing it. In normal line operations these days, various levels of automation will be engaged most of the time. So the basic attitude instrument skills must be revisited often, and re-enforced by realistic simulator training scenarios requiring their use.
Of course, all of the above is just my own take on what I see as a problem of some pilots attitude towards their role as a pilot. They didn't acquire this attitude problem all on their own either. Manufacturers, operators and regulatory authorities had their roles in promoting an "automation first" mindset across the industry. That's what needs to change. Automation is fantastic in so many ways. Until it isn't.
Attitude. Not just the attitude of the airplane, but of the pilot. I started off with basic flight instruments like most pilots of my civilian general aviation background. Later when I started flying jets, we had the Collins FD109. It worked well enough most of the time, but didn't always keep up when capturing VOR radials, altitudes, ILS localizers or glideslopes. So one had to anticipate a bit and try to stay "slightly ahead" of the FD command bars. That meant scanning the raw data and "predicting" when the command bars should react. If the bars moved when they should, follow them. If not, fly "through" them.
The first line Captain I flew with regularly called it the "flight advisor". He cautioned against allowing it to act as an "electronic flight instructor". Though he was joking, there was a serious side to his characterizations. For various reasons, depending upon the flight guidance computer to provide timely and correct pitch and roll commands could sometimes lead to some pretty rough or oscillatory intercepts. The A/P didn't make very smooth intercepts either because it relied upon the same guidance. So if you wanted fly smooth and accurate course or altitude captures, you hand flew it just slightly ahead of the FD command. You could help the automation to work better by making very shallow intercepts, but this is not always practical in the approach environment.
Of course, most of the analogue electromechanical FD issues are all but gone in later generation digital avionics, but the analogue mindset will still serve a pilot well when surprises occur.
Returning to the subject of pilot attitude, I think Captain E. had it right. The flight director is best considered to be an advisor, not an instructor! And if a pilot maintains that mental "attitude" in all subsequent flying, the reaction to an uncommanded change of attitude is more likely to be automatic for that pilot. Click-Click, crosscheck, flight controls and thrust as appropriate. Get it back to where it's supposed to be, (course altitude, speed) THEN communicate, analyze, troubleshoot, checklists etc...
Someone brought up the subject of training earlier. Initial training on two of my types included both unusual attitude recovery and a couple of opportunities to put that training to work in the form of unannounced surprise upsets. Sim instructors the world over have a sadistic streak that can be put to good advantage! During a complex RNAV DP while on A/P, a distraction or two can be introduced along with an uncommanded attitude deviation. Pitch trim runaway is a favorite of course, but wake vortex and several other problems are equally demonstrative and provide a good opportunity to put briefing room theory into practice. I once saw a guy throw up his hands and scream when the airplane rolled over. (the instructor had secretly asked him to) I was supposed to take over like a good FO should. Cardiac stress test passed!
My point regarding pilot attitude is that the pilot must believe in their own ability to fly the airplane on raw attitude and performance data. This belief can only be earned by doing it. In normal line operations these days, various levels of automation will be engaged most of the time. So the basic attitude instrument skills must be revisited often, and re-enforced by realistic simulator training scenarios requiring their use.
Of course, all of the above is just my own take on what I see as a problem of some pilots attitude towards their role as a pilot. They didn't acquire this attitude problem all on their own either. Manufacturers, operators and regulatory authorities had their roles in promoting an "automation first" mindset across the industry. That's what needs to change. Automation is fantastic in so many ways. Until it isn't.
Last edited by westhawk; 25th Jan 2019 at 07:29.
So the basic attitude instrument skills must be revisited often, and re-enforced by realistic simulator training scenarios requiring their use.
That said, it is common to see (as the simulator instructor) children of the magenta line pilots going heads down before engine start into the PFM box dialing in the runway, and even way points around the circuit and even the destination. None of this info is needed to fly a simple circuit pattern. It is only a simulator, right? Yet we see pilots sneaking in a pitch bar on the FD to help maintain correct circuit height.
When told to switch off the pitch bar and simply fly the aeroplane, flying becomes inaccurate and plus or minus 200 feet from circuit height becomes the norm.
We see some pilots getting quite irritated with themselves with their inability to fly a circuit, and can't wait to plug in the FD and AT and hdg mode. Other types enjoy the opportunity to increase their pure flying skills and cannot wait for a friendly instructor to say "feel like a few touch and go circuits?" Horses for courses as the saying goes.
Besides, it's the most fun you'll probably have all day in a sim. If someone was willing to pay me to do it, I'd be happy to do nothing but circuits and bumps! (and maybe some raw data ILSs)
Join Date: Mar 2001
Location: I wouldn't know.
Posts: 4,497
Likes: 0
Received 0 Likes
on
0 Posts
Really? I rather do the raw data stuff on the line. Reality is always better than the simulator.
Join Date: Feb 2012
Location: USA
Posts: 241
Likes: 0
Received 0 Likes
on
0 Posts
Because this is a professional pilots' forum, it is sensible that the attention ought to be first on things that pilots can directly control (piloting technique) and secondly on things that pilots can influence (airline SOP). But to my eye there's a huge point that's being missed here; one that ought to get the attention of engineers and system designers:
The automation is supposed to help, i.e. by reducing pilot workload. Automation that misleads, for example by a FD giving bad advice, is not only not helpful, it is potentially fatally dangerous.
So, beyond deciding whether or not to shut off the FD during unusual attitude recovery, pilots ought to be looking at the engineers and shouting, "what the actual hell?" and demanding automation that does its job. In principle there is no technical reason why the aircraft automation cannot detect unusual attitudes and give correct FD indications, or , at a bare minimum automatically suppress the FD display when the FD display is going to be misleading.
The automation is supposed to help, i.e. by reducing pilot workload. Automation that misleads, for example by a FD giving bad advice, is not only not helpful, it is potentially fatally dangerous.
So, beyond deciding whether or not to shut off the FD during unusual attitude recovery, pilots ought to be looking at the engineers and shouting, "what the actual hell?" and demanding automation that does its job. In principle there is no technical reason why the aircraft automation cannot detect unusual attitudes and give correct FD indications, or , at a bare minimum automatically suppress the FD display when the FD display is going to be misleading.
Join Date: Jan 2014
Location: N5109.2W10.5
Posts: 720
Likes: 0
Received 0 Likes
on
0 Posts
pilots ought to be looking at the engineers and shouting, "what the actual hell?" and demanding automation that does its job.
https://www.bea.aero/docspa/2009/f-c...p090601.en.pdf
"However, analysis of previous events shows that the AP and the A/THR always disconnected automatically(25). The flight directors always disappeared (at least temporarily) when the A/THR disconnected, but reappeared automatically when the operating conditions were regained, whereas the re-engagement of the AP or of the A/THR required action by the crew. This difference in behaviour between the AP and the A/THR on the one hand, and the FDs on the other, probably played a role in the accident as a result of the conjunction of several effects: ˆ The credibility of the cross bars is strengthened by their disappearance followed by their re-appearance: if they appear, it implies that the indications that they display are valid; ˆ Since they attract the crew’s attention (green colour and presentation in the centre of the PFD), the presence of the cross bars could have influenced the actions of the PF, notably in respect to his reaction to the stall warning; ˆ It is only possible to be aware of the changes in active modes (when the cross bars reappear) by reading the FMA, which is probably difficult to do in a high workload situation induced by piloting or failure management tasks. One may therefore question the suitability of the automatic reappearance of the flight directors once they have disappeared."
& Page 211. "Consequently, the BEA recommends that: € EASA require a review of the re-display and reconnection logic of the flight directors after their disappearance, in particular to review the conditions in which an action by the crew would be necessary to re-engage them; [Recommendation FRAN-2012-047]"
Join Date: Sep 2002
Location: La Belle Province
Posts: 2,179
Likes: 0
Received 0 Likes
on
0 Posts
For the FAA, AC 23.131-1C para 17.4 and AC25-11b (more tangentially in a couple of places) both discuss "decluttering" displays when in unusual attitudes (I think there's another AC which parallels AC23.1311 for part 25, but I can't find it right now). Suffice it to say, for those calling upon the "engineers and designers" to address this, there's already a bunch of guidance out there which does influence current designs. Indeed, one item for debate is what needs to be decluttered and what needs to be kept; something that is distracting clutter in one case might be the vital piece that completes the mental picture for the crew in another....
In principle there is no technical reason why the aircraft automation cannot detect unusual attitudes and give correct FD indications,
A FD recovery (if it is invented) would not prioritize wings level first. It would presumably only display direction of turn required and direction of pitch up (aircraft in a spiral dive for example). The pilot could be sucked into simultaneously rolling and pulling top satisfy FD commands. Not a wise move.
And conversely, an extreme nose high / high roll scenario - broadly speaking - you want the opposite, nose down then roll level. Another variable in the program!
I did specify that I was referring to the sim.
Originally Posted by westhawk
Besides, it's the most fun you'll probably have all day in a sim.
Join Date: Oct 2010
Location: 5° above the Equator, 75° left of Greenwich
Posts: 410
Likes: 0
Received 0 Likes
on
0 Posts
Someone asked how do you teach a pilot through the FDs. Well, for starters, teach them to be in command, not a proactive passenger. You simply cannot fully, 100%, blindly trust everything the machine is telling you. It makes mistakes too, it fails, no matter how modern and techy it is.
I fly an A320 for a regular job and I like to get a kick out of it by manually flying it a bit more than most pilots I’ve seen. How to look through the FD? Well, since I’m flying it, I make my scan and my inputs are based off it. Take a look at the FD, does it match my own “mental commands”? Yes? Follow them. No? Diagnose why it doesn’t (diagnose includes my own commands, I can of course be wrong and the machine right).
I liked the concept mentioned earlier: Flight Advisory. Most of the time it will be right, but then one time it may not... what will be your course of action?
I fly an A320 for a regular job and I like to get a kick out of it by manually flying it a bit more than most pilots I’ve seen. How to look through the FD? Well, since I’m flying it, I make my scan and my inputs are based off it. Take a look at the FD, does it match my own “mental commands”? Yes? Follow them. No? Diagnose why it doesn’t (diagnose includes my own commands, I can of course be wrong and the machine right).
I liked the concept mentioned earlier: Flight Advisory. Most of the time it will be right, but then one time it may not... what will be your course of action?
Join Date: Feb 2012
Location: USA
Posts: 241
Likes: 0
Received 0 Likes
on
0 Posts
The general rule for recovery on instruments from (say) a spiral dive, is first level the wings. There may be other things including simultaneously selecting idle thrust to prevent speed build up and selecting speed brakes out. Once the wings are level, then apply elevator to raise the nose. The important thing is not to pull back on the stick while the aircraft is still in a spiral. All this is ab-initio stuff.
A FD recovery (if it is invented) would not prioritize wings level first.
It would presumably only display direction of turn required and direction of pitch up (aircraft in a spiral dive for example). The pilot could be sucked into simultaneously rolling and pulling top satisfy FD commands. Not a wise move.
My argument is that if you, or I, or Wolfgang Langewiesche, or your average CFI is capable of writing down in words what to do in a given recovery situation, then a FD can be programmed to recommend exactly that.
Join Date: Mar 2001
Location: I wouldn't know.
Posts: 4,497
Likes: 0
Received 0 Likes
on
0 Posts
The general rule for recovery on instruments from (say) a spiral dive, is first level the wings. There may be other things including simultaneously selecting idle thrust to prevent speed build up and selecting speed brakes out. Once the wings are level, then apply elevator to raise the nose. The important thing is not to pull back on the stick while the aircraft is still in a spiral. All this is ab-initio stuff.
A FD recovery (if it is invented) would not prioritize wings level first. It would presumably only display direction of turn required and direction of pitch up (aircraft in a spiral dive for example). The pilot could be sucked into simultaneously rolling and pulling top satisfy FD commands. Not a wise move.
A FD recovery (if it is invented) would not prioritize wings level first. It would presumably only display direction of turn required and direction of pitch up (aircraft in a spiral dive for example). The pilot could be sucked into simultaneously rolling and pulling top satisfy FD commands. Not a wise move.
Join Date: Jan 2006
Location: US
Posts: 2,205
Likes: 0
Received 0 Likes
on
0 Posts
How do you ignore, ‘look through’, the FD’s? By using basic, raw data, ‘attitude then performance’ instrument flying skills.
We get paid to realize when the FD’s aren’t providing the correct information. Revert to a lower level of automation, as low as raw data (ignoring the FD), and turn the FD’s off or reestablish the FD’s to give proper guidance.
We get paid to realize when the FD’s aren’t providing the correct information. Revert to a lower level of automation, as low as raw data (ignoring the FD), and turn the FD’s off or reestablish the FD’s to give proper guidance.
misd - you appear to assume in stating ’ignore’, that the process is conscious, rational, easy, can be taught, etc. This might apply in less stressful circumstances, but with shock, surprise, we tend to ‘tunnel’, focus on one cue as a solution; using the FD right or wrong.
In this state we are unable to ‘realize’, the need to change the course of action; thus we don’t ‘see’ some instruments or consider alternative actions.
Many discussions overlook how the upset was encountered, we forget about previous knowledge and events which contribute to the current state - why we all have different views of the same situation. Many problems in upset conditions stem from not knowing how you got there; if you did, perhaps there would be a simple mental undo button.
The focus should be awareness, before and during the event; these are often bypassed in simulator training. The exercise might be announced - no surprise, or canned recovery actions briefed, whereas what is required is to understand the situation before acting.
Understanding requires thought, opposed to moving the controls, thus often considered a waste of simulator time. We need a simulator for the mind.
In this state we are unable to ‘realize’, the need to change the course of action; thus we don’t ‘see’ some instruments or consider alternative actions.
Many discussions overlook how the upset was encountered, we forget about previous knowledge and events which contribute to the current state - why we all have different views of the same situation. Many problems in upset conditions stem from not knowing how you got there; if you did, perhaps there would be a simple mental undo button.
The focus should be awareness, before and during the event; these are often bypassed in simulator training. The exercise might be announced - no surprise, or canned recovery actions briefed, whereas what is required is to understand the situation before acting.
Understanding requires thought, opposed to moving the controls, thus often considered a waste of simulator time. We need a simulator for the mind.
