How Do you "Lock On" to an altitude with Glass Cockpit

[Machinbird]

OK, I'm the first to admit I'm a steam gage trained pilot with zero experience with glass cockpits.

In a steam gage cockpit, it was extremely easy to visually "lock on" to an altitude that you needed to maintain. When flying manually, you simply kept the big altimeter needle pointing in the same direction once you arrived at your assigned altitude (to the best of your ability). Small deviations above and below your assigned altitude are easily seen without actually looking at the gage, you only need to scan your eyes across and note the needle angle.

Glass cockpits of both A and B aircraft appear to involve a simulated rotating drum display similar to a car's mechanical odometer, except that the drums can rotate either direction. ( I've seen the moving "tape" display behind the altitude indication on the "bus in operation.)

What is not clear to me is what technique a pilot would use to "lock on" to an assigned altitude if they wished to manually fly the aircraft. If you are holding FL370, for example, on or a little above altitude reads 37030 perhaps,
a little low and it reads 36950, but you must actually look at the indication to read it, don't you? Is there a simple way to scan it without looking at it?

I'm wondering whether current PFD altitude displays make it difficult to maintain altitude awareness while flying manually.

[Occy]

autopilot bro

[irishpilot1990]

It is just a new scan that you learn, same way it took you a while to learn your steam gauge scan.

You speak about locking the needle and being able to see small deviations, you still need to read the altitude (?) to see how big your deviation is and if you have corrected it.

Your just keeping the altimeter constant, not sure where you are getting confused?

[noip]

Gotta say that when I transitioned from the "Classic" to the "400" that I found it was (in the end) easier to maintain a more accurate altitude.

These days of course, the only places you can normally hand fly at 37,000 ft are in Afghanistan and Antarctica (on our company's routes).

The technique I tended to use (which is what you are searching for) was to keep the VSI pointer pointing at the Altimeter number you wanted.


hth.


N

[Cagedh]

Well, if I want to keep my altitude in an A320 during manual flight with FD off, I would indeed be monitoring the digital readout, but you could also look at the area I highlighted with an orange ellipse in the picture below.

You can see that the selected altitude in blue is slightly above the yellow reference line, so you know that your slightly low just by looking at that.

At FL370 in RVSM airspace, you would normally be flying on the autopilot of course, but there's nothing stopping the pilot from flying manually, FD off, during approach (unless you fly for some stupid company which doesn't allow their crew to do so)

http://i1261.photobucket.com/albums/ii582/Cagedh/A320pfd1.jpg

[John Citizen]

I find it rather instinctive to "lock onto an altitude", without the need to actually read any numbers.

Just set the desired altitude in the FCU (MCP), and then keep the altitude symbol/bug (the big blue square) aligned together with the altitude indication box (yellow box).

If the blue square (altitude bug) rises above the yellow square (altitude indication), it means you are descending, so just pitch upwards.
If the blue square (altitude bug) moves downwards, it means you are climbing, so pitch downwards.

[Capn Bloggs]

What a crappy-looking PFD. :}

FWIW, machinbird, I do have more difficulty interpreting IAS and Altitude on a tape-PFD. No problem of course, but it is more difficult. The saviour is, as mentioned above, the big bug adjacent to the desired IAS or altitude target that helps you spot an error.

A needle approaching the six or twelve o'clock was more intuitive. No space on the square "dial" for that these days though! I always laugh when remember during my EFIS conversion, the CBT man saying that the PFD was in the classic T instrument layout. Ha! :D

[olepilot]

Use the FPV/Bird

[alf5071h]

Perhaps many pilots have difficulty with raw instrument flying when using glass display formats.
A short, limited history of the development of ‘glass’ instruments (including LCD, LED, etc) indicate that there have been many weaknesses with tape formats in comparison with conventional dial instruments. The reappearance of dial formats on the larger ‘glass’ displays perhaps reinforce this point.

Instrument formats with glass were initially constrained by the size of the equipment and computing power. A simple ‘look-alike’ attitude display provided an easier certification route for an untested technology. There were also parallel needs in computation for reliability and avoidance of ‘hazardous / misleading’ information; techniques such as dual-dissimilar hardware and software were used. This latter area and display media have progressed rapidly in the last 20 yrs.

Wider displays enabled a tape format for airspeed; there was significant debate about the direction of the numbers, but the increased use of FD/autopilot/throttle was argued to balance any ‘human factors’ concerns (speed/pitch/control convention). Note that those aircraft which initially did not have autothrottle chose an alternative orientation – Gulfstream high numbers down; IIRC many manufacturers gave operators the choice, but none strayed from the norm.
Automation may hide these weaknesses, but they are still there; try a non-autothrottle, non FD night/IMC takeoff (as per the MEL), with an engine failure +- V2.
Research evaluated alternative formats; a nonlinear scale to give a wider view of the speed range, a curved scale to aid rate (acceleration) information, but none were really convincing.
The mandated Vmo/Mmo markings were supplemented with AOA computation to improved low speed awareness as did a range of bugs for ‘command’ or temporary limit speeds. A trend vector was added to aid acceleration (a crutch), and there were suggestions that it could be used as a ‘director’. Some formats might still be used in this way in lieu of AT. There were also issues about the direction of acceleration in a stall with trend vectors using an airspeed/inertial mix – aircraft speed decelerating but with inertial acceleration Earthwards! (cf AF447)

Many of these basic deficiencies are seen in early altimeter formats which became available with the larger (square) displays. Research again focused on dial formats including ‘egg’ shapes to improve the much deficient rate information deemed important in acquiring and maintaining altitude (“ … difficult to maintain altitude awareness while flying manually.”).
Note that glass VS (alt rate) formats also suffered; some VS formats change the indicator’s point of rotation at high VS to accentuate the value, they also use numerals and colour change to supplement the reduced analogue cues.

The key features in the difference between tape and dial formats are the lack of angular awareness of a moving pointer (including high speed = high, etc), the reduced length of the display scale, the swept area (dial airspeed, MD 80 anyone?).
Altitude formats need a counter pointer system; this required some of the new computation techniques such as dynamic widowing – larger displays and more computing power = cost, thus initially a compromise.
The altitude deficiencies have been slowly eroded with ‘crutches’, such as the shape of the ‘set bug’, with chevron/diagonal structures which aid detection of small movements, differing size of tape scale markings and colouring, and additional bugs (and even an alt trend vector?).

These appear to have improved the general performance of the displays to meet the industry’s requirements for monitoring automated flight.
However for manual flight, together with reducing experience in raw instrument flying, and perhaps without specific instrument flight training on tape displays (weaknesses in alt / attitude / speed scan), the height-keeping accuracy may be degraded, or at least a more demanding task which increases workload.

[172_driver]

Not that I have flown that much "lock on" with A/P and F/D offs. But if I compare it with flying a raw data ILS I would use the high resolution of the PFD to my benefit. I try to find a very exact attitude that results in maintaining altitude. I mean for instance "the top of the airplane symbol has a small margin to the bottom of the 2,5 degree pitch line on the PFD". From there, working with very small changes of yoke pressure / attitude changes. This of course does not relieve me from scanning the altitude tape/VSI, but I can always go back and "rest" my eyes on the PFD knowing what attitude I am looking for. Harder in bumpy air.

[Lightning Mate]

No idea. :\

However, if you wanna know abaht lock on to a target for missile launch,
let me know. :)

[Machinbird]

Wow. A lot of good inputs:ok:. A big thank you for your answers. The key secret of easy altitude maintenance appears to be the 'bug' on the tape and its relationship to the altitude window. Special thanks to alf5071h from one 'prune' to another for the historical/developmental perspective (hint: look at his bio). I'm going to have to look at some of the other techniques mentioned.

Lightning Mate. Been there, done that.:} The best save I ever made was when the target altitude didn't look right and I didn't pull the trigger.

I asked the question because I was wondering about a couple of AF situations wherein the crews were slow in recognizing they had left the assigned altitude.
Once you are >500 feet off, the bug is no longer visible and the scrolling altitude tape & rolling numbers didn't trigger a mental alarm bell for them.
When you have been cruising along for hours and the AP suddenly kicks off, it must be interesting getting your scan going, particularly at night in turbulent conditions.:eek:

[flyboyike]

This reminds me. When I first transitioned to glass (PA46-500TP), the instructor told me that my scan was about to become a stare. He was right.

[Centaurus]

This reminds me. When I first transitioned to glass (PA46-500TP), the instructor told me that my scan was about to become a stare. He was right.

No argument with that. Which is the main reason when this Ppruner first introduces engine failure on take off during the type rating on the 737 simulator, the flight director is switched off before taking off and the whole exercise is done raw data and hand flying. Attitudes are easy to see rather than having to `look behind` FD bars and once they get the hang of steady rudder application and avoid too much aileron, it is a pleasant surprise how quickly candidates pick up the correct attitude.

On the other hand I have frequently watched those starting with the flight director on for take off (which is normal procedure during type rating) having great trouble trying to pin the needles together. They chase the needles and because of their intense concentration on pinning both needles into the miniature square their instrument scan is lost as is their situational awareness for a short period of time.

In the old days, pilots under training were criticised for chasing the ILS needles rather than pick a heading to keep on the ILS. Yet nowadays during engine failure after take off the pilot is exhorted to concentrate and FLY the needles of the flight director. No wonder some have continued difficulty using flight director indications during one engine inop climb out. In fact some pilots get seriously spooked while chasing the FD needles and it doesn't work out for them.

The instructional trick is to switch the FD off for initial engine out after take off training and once the pilot has that beaten then introduce the FD.

[alf5071h]

Mach, the recent incidents / accidents (AF447) might be a problem of general awareness due to weaknesses in instrument scanning – training and currency. In addition, with multiple ADC failures the pressure error correction to the altimeter may be removed and thus there is an apparent jump in the display which could encourage an overreaction. Also, possible bias from the dominating need for accurate height keeping in normal operations; thus crew’s not being familiar with the need to change thinking / behaviour patterns between normal and abnormal normal operations.

OK465, IIRC most of the US designed mechanical tapes were mounted vertical, whereas the UK Lightning had a combined Mach / Airspeed tape orientated left to right (high numbers right). Lightning Mate may have flown this version, but I don’t recall teaching him about missile launches (the other side of the hanger and ‘at camp’).
Although HUDs use similar formats and scales to those in EFIS, they may have different characteristics. A HUD ‘real world’ conformal display may not have the same characteristics as the FD overlaying ADI attitude. The pitch scaling can differ by as much as 5:1, perhaps modifying the perception of pitch rate between HUD and an ADI.

Any training for new technology requires modified techniques; Centaurus gives a good example with the FD, but I suspect that tape altimeters, HUD, etc, also require some ‘tricks’ in training.
A potential problem is that the industry accepts many aspects of normal operation of new systems without sufficient thought of abnormal ops where seeming minor differences can have significant effects on workload and situation awareness. Unfortunately these may only surface in incidents or accidents.

[Machinbird]

Mach, the recent incidents / accidents (AF447) might be a problem of general awareness due to weaknesses in instrument scanning – training and currency. In addition, with multiple ADC failures the pressure error correction to the altimeter may be removed and thus there is an apparent jump in the display which could encourage an overreaction. Also, possible bias from the dominating need for accurate height keeping in normal operations; thus crew’s not being familiar with the need to change thinking / behaviour patterns between normal and abnormal normal operations.
ALF, I think currency is the key. A crew may be engaged in the navigational problem throughout the flight, but typically the aircraft handling problem belongs to OTTO for flight after flight after flight and the crew is almost completely disengaged from aircraft handling.. The few seconds on approach are in a different environment than during cruise. The bit of sim work every 6 months or so is also usually down low, hitting the mandatory low altitude critical problems.

Throw in the surprise factor, mental overload, adverse Circadian rhythms, night, turbulence and a different flight control law such as the half fish, half fowl, Alt2(b) law and a rusty pilot will be in a world of difficulty getting any scan going.

IMHO pilots are going to have to fly more to maintain handling skills, or else AF447 may well mark the reaching of the tipping point where handling skills can no longer be trusted to save the day. The RVSM airspace reason for not hand flying must be overcome. Think of the training opportunities going to waste!

Can anyone think of a reason why this won't work? (I believe the institutional problems can be overcome.)
Let us think outside the box for a moment. Supposing we added in a manual flying training mode into our aircraft.

How did you learn to hold altitude? When you began to get a little off altitude, didn't your instructor cluck at you and if you continued to diverge, didn't he give the stick a nudge to put you back in the tolerance band. Why couldn't our aircraft with their sophisticated autopilots do something similar? No need to breech RVSM limits yet we can still get actual handling time and build a scan. Any pilot worth his salt hates having the stick nudged. When you get tired, you can let OTTO have the complete aircraft. Think of the potential to actually monitor a pilot's handling skill and to automatically record actual handling time! True, you won't get as much time to balance your checkbook or to read a magazine, but remember why you are sitting in the front seats!

[alf5071h]

Mach, I disagree with the majority of the arguments calling for greater focus on manual flying skills; not just the view in #17.
Recent accidents indicate that the human problems were predominantly due to the choice of action, and not that the crew couldn’t fly the aircraft. In many cases the aircraft was flown (accurately) into the wrong part of the flight envelope. I exclude AF447 because there may be other system and trim issues, and we don’t have all of the details yet.

Choice of action requires good situation awareness. With inaccurate situation assessment, then flight control inputs for a chosen action might be hesitant, or exploratory – getting a feel for the situation. The resultant aircraft motion may not match the predictions from the pilot’s understanding.
With good situation assessment, choosing an appropriate action should result in a high probability of success even if the flying is not particularly accurate. Furthermore this action should change the situation to something which was expected, thus confirming the choice – thus a need to continuously reassess the situation.
Training for this should also involve how to combat surprise and 'getting back into the loop'; changing from passive monitoring (automation) to active awareness and control (manual).

Flight instruments provide a major contribution to awareness. A danger with modern displays is that single cue or automatic flight may result in an expectation of ‘single-source’ recovery in abnormal conditions (instant answer).
Good instrument flying skill (awareness) involves scanning a range of instruments, and the construction a time based model of what has happened, what is happening, and what might happen; past, present, future. These snapshots are combined with previous knowledge to provide a mental model – the situation.

Whilst modern displays provide very good inputs to this process for navigation (includes time), it might be questionable if the aircraft handling parameters (speed, attitude, altitude) are as good as conventional ‘dial’ instruments. In particular, note the weaknesses in rate information, this is a critical timeline input to the mental model.
Poor situation awareness due to these aspects may also result in poor timing of the control inputs – sharp / harsh movements – getting the feel for or generating a new understanding of the situation as much as for the aircraft response. This might be interpreted as poor flying skills, but the cause is elsewhere.

[Natstrackalpha]

autopilot bro

It SO is !!

[misd-agin]

Initial transition to tapes from round dials does present some instrument scan issues. I was surprised at the difference in effort. Our transition sim had us fly a handflown level off at 2000' immediately after takeoff. Simple... except everyone overspeed 250 kts because they were struggling with their scan transisitioning to the tape altimeter/VSI.

Looking at the altimeter and VSI you can instantly see the altimeter needle pointing straight up and the VSI needle at 9 o'clock and you KNOW you're at a thousand foot with zero sink. It's not that obvious on tapes, especially when transitioning from years of round dial flying. You adjust and it becomes much easier. It eventually becomes second nature, just like round dials.

[pilotmike]

How do you manage to tell the time on a digital clock?

[Fly3]

Machinbird.
On all the tapes on the PFD if the target goes "off scale" as you put it, it becomes a number at the relative end of that scale and so is always there to remind the crew. The altitude deviation warning might also wake them up.

[roulishollandais]

A simple ‘look-alike’ attitude display provided an easier certification route for an untested technology
(bolds are mine)

Let us forget ECAM and the counterfeit 'look-alike' Airbus PFD ! :ugh:

Let us start with a real glass cockpit : the KLOFSTEIN Head-Up-Display :O

USA and MIT honored him :
OCR Document (http://www.headupflight.net/rapports%20de%20vol/MIT.htm) :ok: (en)
( headupflight.net/rapports%20de%20vol/MIT (http://www.headupflight.net/rapports%20de%20vol/MIT.htm) )
ICARE N°64BIS (http://www.headupflight.net/articles2/tourneeUSA.htm) (fr)
but Airbus and French-German airspace community (Zieglers and co mafia) did not accepted that a french Jewish man would be a better test-pilot and test-engineer as them.

Here you will find one of Klopfstein's message (PPRunE did not exist !)Etude Rationnelle du Pilotage des Avions (http://www.headupflight.net/articles2/Etude%20rationnelle%20du%20pilotage%20des%20avions.htm) (fr)

[alf5071h]

Roul, thank you for reviving an interesting part of the history of aircraft flight displays. By chance I am familiar with Klopfstein’s work, but I did not fly the system.
The RAE in the UK followed the research aspects (BLEU); British Aerospace and Smiths industries (Elliot Automation?) represented the manufacturers (John Wilson / Keith Dougan), who also considered operational aspects, particularly for aircraft to replace the HS 121 Trident. This involved some airline technical departments (Atlas group?).
AFAIR the main debate revolved around manual flight in Cat 3 with HUD vs autopilot/autoland; basically French – British viewpoints, determined initially by the availability of equipment. The French had HUD, the British, triplex autoland; the British HUDs were used by the military and not readily available for civil research. Also, HUD was not favoured by the operators at that time (BEA/BA?); there were certification concerns (less so in France) and installation issues, and an autopilot would still be required, which provided greater potential.

Klopfstein’s work on perspective displays with runway outline, have been used in many HUDs. There was some research in the UK which evaluated the format head-down (Smiths display, BAE Advanced Flight Deck simulator?)
This work is now reappearing in the new wide screen ‘GA’ displays, where perhaps some of the underlying concerns of these and advanced concepts are re-emerging.

The choice of many ‘simple look-alike’ attitude format in early EFIS was driven by certification concerns, particularly at a time when the FAA were requiring ‘proof-of-concept’ testing. Obviously same-type-rating and training aspects were considered; money drove the decision.

The HUD may have been the better technical display solution, but raised debates about crew consistency (flight technical error), cross monitoring, and ultimate capability in low vis. There still appears to be training issues with some of the advanced features – FPV, potential energy, etc.
Obviously this is debateable; but IMHO on a world-wide scale, there are those pilots who understand it, those who think they do but might not if tested in demanding conditions, and those who cannot comprehend.
Now is this similar to some of the issues in modern aviation?
Many of these problems were seen in early display research (RAE / NASA), now there are hints of similar issues with the new wide screen displays particularly when attempting to replicate the real world background.

The fundamental problems appear to relate to how pilots determine their mental model of the outside world. The industry has evolved with dial instruments and scanning techniques to assemble a coherent ‘picture’ – the mental model of what the aircraft is doing.
EFIS and HUD use similar but often different ‘mechanisms’ to create what is actually an illusion of the real world (synthetic vision), this is fraught with all of the problems of mental perception, situation awareness, human bias, and variable performance.

I was fortunate to be within the debates of the era, not that research or what is best technically drove the decisions; – need and money always prevailed.

A bit more history 1968 | 1 - 0233 | Flight Archive (http://www.flightglobal.com/pdfarchive/view/1968/1968-1%20-%200233.html)

[Machinbird]

Fly3, Thank you for the point about the bug moving to the end of the scale and remaining visible.

Mach, I disagree with the majority of the arguments calling for greater focus on manual flying skills; not just the view in #17.
Recent accidents indicate that the human problems were predominantly due to the choice of action, and not that the crew couldn’t fly the aircraft. In many cases the aircraft was flown (accurately) into the wrong part of the flight envelope. I exclude AF447 because there may be other system and trim issues, and we don’t have all of the details yet.
Choice of action requires good situation awareness. With inaccurate situation assessment, then flight control inputs for a chosen action might be hesitant, or exploratory – getting a feel for the situation. The resultant aircraft motion may not match the predictions from the pilot’s understanding.
With good situation assessment, choosing an appropriate action should result in a high probability of success even if the flying is not particularly accurate. Furthermore this action should change the situation to something which was expected, thus confirming the choice – thus a need to continuously reassess the situation.
Training for this should also involve how to combat surprise and 'getting back into the loop'; changing from passive monitoring (automation) to active awareness and control (manual).

ALF
Despite the fact that you start out with disagreement with more emphasis on manual flying skills, IMO we are very close in our evaluation of the problem.

In the example of the AF447, If you or I had taken control of the aircraft, you would have set the wings and nose for level flight and then scanned the altimeter to see how well you had done and what correction was necessary to get back to assigned altitude. This never happened with the AF447 crew. PF was oblivious to the climb he had generated. PM was not much better in that although he eventually perceived a climb, he never voiced the magnitude of the climb above assigned altitude, implying that he was not aware of it.

This is indicative of major scan problems with both AF447 pilots.

I was not convinced of this possibility until I saw the BEA report on the Air France A343 incident near Guadeloupe on Jul 22nd 2011 that similarly pitched up to a near stall. The report details how long it took the pilots to recognize the key performance parameters. This again was a scan problem with additional special circumstances of warnings being overridden by higher priority warnings.

The problem seems clear. Some pilots are becoming excessively disengaged from the actual handling process. They are not monitoring the flight parameters sufficiently to promptly catch deviations, and they are starting cold from a disengaged mental viewpoint when required to take over manual control. The process of using the eyes effectively to form the mental picture of what the aircraft is doing is taking an excessive period of time to develop and this is contributing to loss of situational awareness.

What is not clear is the full breadth of the problem. What percentage of pilots are susceptible? What airlines are particularly susceptible? What type of flight operations promote scan problems? Are there aircraft specific conditions that promote scan problems?

I started this thread because I was not sure how easy the glass cockpit was to fly manually. The consensus seems to be that it might be a bit harder to scan, but it can be more accurate.

Perhaps someone can tell me why a scrolling altimeter does not always trigger an Uh OH exclamation when the aircraft is supposed to be at an assigned altitude? That should have been a big red flag to the crews of the two aforementioned jets.

[alf5071h]

Mach re “… we are very close in our evaluation of the problem.”
Yes, :ok:
Also, with the views on the actions required for recovery.
The root of this is knowing what actions to take, which implies knowledge of the situation – an understanding. Thus in the example, it’s not necessarily a failure in the act of scanning, but one of obtaining an overall awareness and understanding – the bigger picture (e.g. altitude up, attitude up, probably speed down, possibility of stall).

It's suggested that previously (the old days) pilots knew what was important. Nowadays this knowledge should originate from training, or is it insufficient and being eroded by the complexity and demands of modern flying. This knowledge is part of airmanship (professionalism) and must be developed with experience.

The breath of the problem? It’s human nature. We might not be able to shape or change this, but we can constrain and guide it.

Altimeter conundrum; its digits vs dial, single cue focus vs scanning, monitoring (see http://www.pprune.org/safety-crm-qa-emergency-response-planning/478368-monitoring-intervention.html), reliance on alerting systems, overconfidence in technology, etc, etc.
It involves how we do things and what we pay attention to, it’s being aware of the bigger picture. This requires discipline in mental behaviour – the process at the root of professional flying.
Industry, regulators, organisations, and individuals fail to appreciate all of the implications of change.
In complex systems, small steps in experiment (design, regulation, operation, behaviour) control very little, but influence everything.

misd-agin “It eventually becomes second nature,”
From the rules of evaluation – if something is hard at the first try, then there is a deficiency (“Oh it’s me, I will learn”).
During the second try you should attempt to understand what is deficient. Its not you – look at the interface, the system; don’t just learn the tricks to overcome the problems.
The third try is easy because you have learnt some of the tricks, but you must question if have you understood why these tricks are necessary and what they imply in operation and non-normal situations.

[roulishollandais]

Aircraft is not a complex system. Otherwise you cannot survive with fire on board and major flight system failures, IMC, by night,etc.:ugh:
If you are able to fly in these condition, why don't you fly the same manner in normal flight ?
KISS !
headupflight.net/articles2/Securiteoumarketing

(GILBERT KLOPFSTEIN)