View Full Version : Flight Director Indications in an unbalanced turn?

Alex Whittingham
16th Mar 2011, 12:47
My question is, in a theoretical turn, if the aircraft was turning to the left with too much (or any) left rudder applied, ie skidding, and less bank than a balanced turn would require but still turning at the 'anticipated' rate because of the rudder input, would the flight director indicate a satisfied bank command or would it demand more bank?

In other words will it look like this


or this?


I imagine this depends on whether the FD demands a bank angle or turn rate.

de facto
17th Mar 2011, 11:20
I THINK the picture nr 2 would be correct if you suddenly increase the rudder and the rate is momentarily decreased.
An increase of bank would then be required.(higher bank for same turn radius and speed).
Bank angles are computed depending on fmc speed .If the actual speed is increased(increased tailwind during turn),the roll guidance will increase to max(30 deg in LNAV).
Roll modes using heading SEL allow pilot bank to set a bank limit(high altitude,turbulence,turn into a tailwind..).

Ready to stand corrected on this one:E

17th Mar 2011, 11:58
As the flight director has to calculate the performance, in order to direct you to comply with it - it would calculate a bank angle, and assume balanced flight.

Alex Whittingham
17th Mar 2011, 14:04
Yes, that was my initial interpretation, but then I thought that your statement would be equally true if the FD computed a required turn rate. The fact that the demand is speed dependant (which I didn't know) suggests that to me as well.

There's clearly a graded demand for roll in and roll out, it doesn't just whap to a 30 bank demand, that could be either a graded demand based on 'normal' roll rates or a graded demand for a low turn rate initially, it would display the same in a balanced turn. In essence, I'm wondering what the exact meaning of 'calculate the performance' is?

This is to get the illustrations in a manual correct, in case you're wondering, and my gut feeling is also that the second diagram is correct.

17th Mar 2011, 23:25
Well to correct that number two picture you would need to push right rudder to push the ball into the middle first and then apply more left bank as demanded by the FD...The bank is applied to meet the demands of the FD and rudder is applied to keep the ball in the middle.

17th Mar 2011, 23:40
I somehow doubt that it's of little consequence... the "directed" bank angle is that required to meet the performance and that skidding or slipping doesn't enter into the equation. The fact that the controls are unbalanced is a "control issue"... the director is doing the job that's being asked of it.

There again... I could be talking absolute rubbish!


21st Mar 2011, 19:28

This style of illustration looks familiar...

Is there a new Instrument Flying Handbook in the works?

22nd Mar 2011, 11:09

My understanding is that the FD is a RATE demander, ie the computer computes a rate of climb or rate of turn to achieve the required solution. I believe that the bank angle required is predicated on a balanced turn.

In the example you have portrayed, if the angle of bank was that required before the rudder was introduced, my expectation is that the FD would indicate a reduced rate of turn is required. In other words, the FD lateral bar would be on the other side - the third option not illustrated. Why? Because the angle of bank previously gave the required turn rate, the introduction of pro-turn rudder would increase the actual rate beyond that required to achieve the computed solution.

Stay Alive,

de facto
22nd Mar 2011, 11:15
I will check it in the sim very soon:E

Notso Fantastic
22nd Mar 2011, 11:36
Skid or slip is very relevant in FD indications. If you are on engine out and untrimmed and unbalanced, following FD exactly will not be very helpful. Therefore the second image will be the one you see.

Piltdown Man
22nd Mar 2011, 13:23
I believe the turn command in a flight director requests a turn based on current heading vs the desired heading, from whatever source is selected. No more and no less. It will normally be limited to 25˚ fr bank for large heading displacements and will reduce to zero as the "error" reduces. However, unless the ball is in the middle (or out in the appropriate direction when N-1), the anticipated/desired track will not be made good nor will the calculated climb performance be met. So answering your initial question, I believe the top one would be correct.


Alex Whittingham
22nd Mar 2011, 17:21
Thanks for your input all. It makes you think, doesn't it? I'd be interested to see what the sim shows. The scenario is that the same rate of turn is achieved as in a balanced turn, but by a reduced bank angle and a bootful of rudder rather than a 'normal' bank angle and no rudder.

22nd Mar 2011, 18:19
Hi Alex

Although it's going against the grain, I would be inclined to say picture 1 is correct. From what I remember flying the 757, in a SE case, if your heading was constant but you had too much/too little rudder, then the FD bars would command a bank to maintain the selected heading. Therefore it seems to me that the FD bars in this aircraft will command a heading change rather than a specific bank.*
Applying this to your scenario, as long as the aircraft was turning as desired then the FD would be 'happy' with the result and be centred on the bank angle which resulted in this. IE picture 1
Please someone correct me if I'm wrong or not understanding the question correctly as I don't want to give duff info.


23rd Mar 2011, 00:46

From the responses, this may be platform specific...

make sure to ref several platforms in the sim....

23rd Mar 2011, 02:29

The scenario is that the same rate of turn is achieved as in a balanced turn, but by a reduced bank angle and a bootful of rudder rather than a 'normal' bank angle and no rudder.

In that case it is Figure 1.

The logic is as I described previously. My understanding of the FD logic is that it senses an error between present state and desired state and then, through a smoothing algorithm, demands a rate of change to achieve the null position (zero error) over a fixed time interval. It allows for the time taken to achieve a steady state change and to recover to the new state as well as the time in the turn/climb/descent. It is this behaviour that pilots learn by observation and then decide when to lead or lag FD indications in certain circumstances.

The classic example of this is when you do not initiate the demanded change on cue, so the FD demands a greater and greater change (beyond anything sensible) but only until such time as you do commence the change, whereupon the FD "comes to meet you" as it recomputes the situation.

23rd Mar 2011, 07:54
The answer lies in the algorithm that the FMS implements to calculate the roll command.

The systems I have worked with are not driven directly by a rate, but by an error. To calculate the commanded bank angle the turn points are positioned to obtain the necessary transitions between legs. Once achieved, the turn begins and a roll command is calculated proportional to the difference between the current track and the desired track. The proportion can be dependent on a few things, such as distance to go to the calculated points etc. in order to achieve smooth transitions. In addition speed is used to limit the bank angle to give a rate one turn (or other value as set) or the limit is set to an absolute max bank angle. This command would then be sent to an autopilot. The autopilot would then aim to achieve this bank angle and maintain a co-ordinated turn. A comparison is made between the actual bank anlge and that commanded to give an error, which is then scaled and fed to the flight director bar. In an uncoupled mode the calculation remains the same, for consistency, and so the flight director bars would show the error between the current bank angle and that required for the desired, co-ordinated turn.

As I say this is only true for the systems I have worked with, other implementations are available. I would be interested to see what the sim tests produce.

24th Mar 2011, 18:55
I have yet to meet a FD system that does anything but command a bank angle. The commanded bank angle might be modified by mode, speed or heading error, but once commanded bank is met the bar is centred regardless of whether you are actually turning or not.

Note to IGh : You should seriously consider what your formatting does to readability.

26th Mar 2011, 16:42

I have yet to meet a FD system that does anything but command a bank angle.

Well done! That's probably because bank angle is what the instrument displays when you are generating a rate of turn in balanced flight.... :ugh: :ugh:

And of course, if you fly around with shetloads of rudder counteracting the effect of bank,

whether you are actually turning or not

Spare me,

26th Mar 2011, 20:09
Don't forget that when you move the sidestick in an airbus you are asking it for a rate not a bank angle in the first place, before you even think about the FDs.