Airbus hand-flying characteristics

[hikoushi]

Hello!

Since you folks here always seem to have the answers, I have a question for you. A few, really.

I've been flying the 717 for about 4 years and am just about to start training on the A330. Finally bit the bullet and decided to go take a crack at the big French whale. Looking forward to it immensely, and in the meantime enjoying the last few months of honest flying in an honest airplane. Which brings up the question.

What the hell does it feel like to physically fly the stupid thing? I mean, let's say you have the autothrust OFF and the autopilot OFF. You are level at say, 10,000 feet at 250 knots. You pull the throttles to IDLE without touching the sidestick. What does it do? In every aircraft I've flown up to this point, which includes the 717 as well as a smattering of various regional airline and GA planes prior, the plane will drop it's nose and ATTEMPT to capture it's trimmed airspeed, and enter a descending phugoid of some kind while trying to sort itself out. Instead of letting it bob around drunkenly, we apply a few gentle "guiding" nudges of pressure and maybe a small trim change will establish a flight-idle descent at that same speed.

I understand that in Normal Law the sidestick basically commands a load factor, or "G" request (I haven't hit ground school yet so stop me if I'm wildly wrong at any point). Does that mean that the Airbus, if I idled the throttles without touching the stick (A/P and A/THR OFF) that instead of dropping the nose, the plane would try to maintain 1 g by holding altitude and bleeding off speed until it hit Alpha Floor (or whatever the hell that is called), or would it pitch down like a "normal" airplane and just do so very gently and precisely while staying close to 250? What does it do? I am again talking all manual hand flying here. Is it maintaining flight path vector, or airspeed if you move the thrust levers around without moving the stick?

Also, again, stupid question here. I am hand flying the airplane in level flight with the autothrust off. I select (I believe you call it) open descent / FLCH at 250 knots. I pull the throttles to idle and pitch down for the flight director. Airplane stabilizes in the descent. I let go of the stick. I push the throttles forward about halfway without touching anything else. What happens? Does the airplane make any pitch adjustments to maintain the selected airspeed, or does it maintain it's attitude and flight path, and let the speed increase? Again, this is with everything OFF.

Just curious about how the airplane actually FLIES. Will get all the deep technical stuff in the schoolhouse when the time comes, just really curious how the airplane behaves itself when you actually try to physically fly it like an airplane. Don't care about behaviour with degraded control laws, etc; will get to that once I get my head unplugged re: normal law handling characteristics.

Specifically referring to the A330 here, but assuming it is the same for all Airbus models, so please chime in. Thank you in advance!

[TyroPicard]

If you are able not to think of the A330 as a "stupid thing" you might pass the course....

[Jonty]

The aircraft is pitch stable. Not speed stable. In all cases it will try to maintain the selected pitch until you reach one of the reversion limits (normal law).

[windshear-a-head]

Not sure if the 717 has it but it is a very similar feel to CWS on Boeing. The 330 was my first Airbus too, I'm not keen to be honest, at first its a very un-natural way to fly, eventually though like most things it becomes second nature and you realise how little you have to actually touch it (your best mates **** etc).

In gusty/windy conditions the 330 especially can become a handful.

The real joy comes with ECAM mixed with QRH. :ugh:

Good luck. :ok:

[Silvio Pettirossi]

Well, I can try to answer how the a320 would react and AFAIK, it should be the same on the a330;

First case: With A/THR and autopilot off in level flight, thrust at idle, airspeed will slowly bleed off while the plane maintains level flight/1g as you said. Reaching alpha floor AOA it will go into TOGA, speed will increase with the plane still in level flight until a few knots above Vmo/Mmo where it will start climbing in the overspeed protection, always remaining in TOGALOCK.

Second case: In open descent, again with every thing off an thrust levers at idle, you pitch down following the FD bar to maintain 250kt. If you let go off the stick it will maintain 1g in the descent you established. With no further corrections on the stick it will not maintain 250kt but only 1g. Now you increase thrust. This will increase your speed but still it will try to remain at the 1g descent untill again, you hit overspeed protection and it will pitch up.

This is writen out of memory without manuals at hand so everyone correct me if I am wrong.
The Bus, in normal law, will probably not stall, spin, overspeed or overstress itself if left unatended but it will happily fly in to the ground.

[Stuck_in_an_ATR]

One correction. If you set the thrust to idle, the Bus will maintain flightpath until it reaches Valpfaprot. THEN it will pitch down to maintain Valphaprot. Alpha floor won't kick in until you pull through Valphaprot towards Valphamax (or so I was taught, no manuals at hand presently).

As to the conversion to the 'bus. Having read all the stuff about FBW control laws, etc., I expected the bus to handle totally differently to whatever I've flown before. However, once in the sim, I've found it quite pleasant to handfly and, surprisingly, quite conventional.

[hikoushi]

Thank you all for your responses. I think I get the idea. As an aside, if we were to fly the same FLCH / open descent at 250 knots, but this time the autoPILOT was still on, but the autoTHRUST was off (A/P engaged while manually moving the throttles), would the autopilot / FD still adjust pitch to maintain airspeed as I adjust the power? In other words, it sounds like while hand-flying with everything OFF, if you let everything go the airplane is basically always in a "speed-on-thrust" kind of CWS-ish mode. So while flying with the autopilot ON but using MANUAL thrust, does it behave basically like a Boeing or McD would and pitch up and down to maintain the selected speed as I move the throttles around? I know most people do not typically go flying around with A/THR off and A/P on, this is just for educational purposes (although we do it in the 717 quite a bit, but that is just because the autothrottles are slow, and tend to get somewhat "out of phase" and can fairly dramatically overcompensate in the typically very gusty conditions of our area of the world. On final it seems to work much better and give a much more stable and comfortable approach and landing just to leave the power set around 1.09-ish EPR and accept the minor speed variations, manually doing what needs to be done when the changes start to get large).

Again, thank you in advance. And TyroPicard, take it easy, buddy;)! Please understand that I mean "stupid thing" as an affectionate, sarcastic American-style term of endearment. I'm sure the Airbus is much smarter than me!:ok:

[NigelOnDraft]

If you set the thrust to idle, the Bus will maintain flightpath until...This is the theory. In practice it will roughly maintain the attitude, so you will need to increase pitch to maintain level flight.

The theory completely falls down when you start taking flaps when left to itself it will fail to maintain attitude or level, and the AP is not much better in the short term. So some skill / satisfaction can be had in flying to a better standard than the AP ;)

NoD

[TyroPicard]

No worries, hikoushi, I was trying to be Ironic. Now that I know you are from the USA I won't try again...
TP

[Microburst2002]

In OP DES, AP or FD will command pitch to maintain airspeed. Thrust changes will only affect rate, if in AP or if you follow the FD.

But let me explain you the concept of airbus fbw. The system is a flight path stable loop system. It will tend to maintain a straight path when the stick is released to neutral (it is springloaded to neutral and its feel is totally independent of speed or any other variable other than the stick deflection angle itself). It is not speed stable as a conventional system. In your 717 you need speed related feel in the stick because you need to trim off stick forces in order to have the airplane tend to fly in the phugoid you desire. When pushing or pulling it, you feel the aerodynamic effect that your input is likely going to have in the airplane. It is not the movement inof the yoke but the force on it which tells you if you will overstress the airframe. At low speeds you will instinctively push or pull to obtain a given pitch rate, at higher speeds you will modulate stick forces to obtain a g load. Stick forces help you to modulate your inputs.

In the airbus you don't need such forces as the system, when the stick is released, will move surfaces on its own to maintain flight path. To change flight path you will push or pull e stick. At low speeds you will move it more or less to achieve a pitch rate. At higher speeds you will move it more or less to achieve a g load. Same as in conventional airplanes. That is why it is very natural to fly. The control law does command g load blended with pitch rate at low speeds and pure g load at higher speeds. Once you achieve desired flight path you just release gently the stick and then you have trimmed it. Releasing the stick equates to trimming. No need for artificial forces. Even with a jammed stabiliser the elevators will deal with it and the airplanestill has autotrim. You still fly with a loop system demanding g load (pure in this case as younwill be in alternate law).

Boeing's fbw is similar in the control law, which demands a blend of pitch rate and g load dependent on speed. But they added a trim speed induced pitch input such that the airplane will behave as speed stable. It still has trimming switches and artificial feel.

Don't resist airbus fbw. Reject wrong notions that many have about it, even many airbus pilots. The protections and the more or less authority they give or take are certainly stuff for debate, yes, but the way it flies is absolutely natural. It is not an AP nor a CWS. You fly the bloody thing, pitch and thrust, just like the 717s., only easier because you don't have to trim

[hikoushi]

Outstanding information; really appreciate it. Definitely looking forward to flying it now. Happy New Year!

[Stuck_in_an_ATR]

@ Microburst

The control law does command g load blended with pitch rate at low speeds and pure g load at higher speeds.

Could you please provide any reference for that? The FCOM states they the pitch law is pure "g load", without any blending.

My understanding is that the "pitch rate" law at low airspeeds is only apparent - the sidestick still commands "g", but at low speed you need significant pitch rate to achieve "g" - so the pilot mostly sees and feels the pitch rate. At higher speeds you need little pitch rate to attain desired "g-load", so the pilot feels mostly "g"...

[Microburst2002]

stuck in

I cannot, since that is totally "transparent" to the pilots in the FCOMs. No reference at all. However as it happens I researched a little about fbw control laws and I found that there is pitch factor blended at low speeds.

It is called a C* law. B777 has a C*U, the U being the trim speed thing. C* is a pitch rate and g load law, the proportion of both depending on speed.

Surf the internet, you will find plenty of stuff.

cheers

[Gryphon]

FCTM-OP-020 -OPERATIONAL PHILOSOPHY-FLIGHT CONTROLS

NORMAL LAW

CHARACTERISTICS IN PITCH

IN FLIGHT

When the PF performs sidestick inputs, a constant G-load maneuver is ordered, and the aircraft responds with a G-Load/Pitch rate. Therefore, the PF’s order is consistent with the response that is "naturally" expected from the aircraft: Pitch rate at low speed; Flight Path Rate or G, at high speed.

[Microburst2002]

Excellent, it had totally gone unnoticed by me till now.

as a note, when they designed these control laws, they first researched on what do pilots expect when we introduce inputs in the flight controls and the conclusion was that at low speed, we expect a pitch rate as the output, while we expect a g load as the output, at high speeds. I think they were right. when at high speed I modulate my inputs looking at the effect in vertical speed which is immediate. If the increase is too sharp, I relax the stick, if it is insufficient, I pull or push more. At low speed, I modulate looking at the effect in pitch rate, too fast, I relax, too slow, I push or pull more.

[Stuck_in_an_ATR]

Actually, if you read the passage from the FCTM carefully,it says nothing about blending:

When the PF performs sidestick inputs, a constant G-load maneuver is ordered, and the aircraft responds with a G-Load/Pitch rate. Therefore, the PF’s order is consistent with the response that is "naturally" expected from the aircraft: Pitch rate at low speed; Flight Path Rate or G, at high speed.

The wy I read this, is that pilot commands constant "g". Aircraft's response is g-load/pitch rate (it always is, in any aircraft!). However, at low speed one needs higher pitch rate to obtain the same g, than at hi speed.

So, for any commanded "g":

Lo speed - pitch rate is large, so the pilot sees mainly pitch rate
Hi speed - pitch rate small, so the pilot feels "g"...

Besides, I guess there's no speed input into the pitch law (alternate pitch law is the same as normal and it works even if all ADR's are lost).How would this g/pitch blending work? :confused:

[gums]

Salute!

Surprised to see the question and the responses.

The AF447 threads went over this and over this and ....

Way I understand it:

- A gee command for pitch that is corrected for pitch attitude, but basically a one gee command with stick neutral. So at 30 deg of climb, the FBW attempts to hold about 0.87 gee. No regard for speed until AoA gets to some limit, and then the system relaxes the gee to keep from stalling. The basic jet seems very benign, but the FBW control laws can get in the way under a few situations/conditions.

- The FBW control laws in "normal" or even a back-up mode continue with this gee implementation except some "protections are lost for max roll and pitch and other "limits".

- The hozontail tail will trim to maintain the gee command and reduce the need for back stick so the elevators can return to "neutral".

Some folks from the AF447 thread can add/correct to this description.

[HazelNuts39]

When the PF performs sidestick inputs, a constant G-load maneuver is ordered, and the aircraft responds with a G-Load/Pitch rate.
Actually, if you read the passage from the FCTM carefully,it says nothing about blending. The wy I read this, is that pilot commands constant "g". Aircraft's response is g-load/pitch rate (it always is, in any aircraft!). However, at low speed one needs higher pitch rate to obtain the same g, than at hi speed.
That is correct for the steady-state constant "g" that is eventually achieved. AIUI the 'blending' of G-Load and Pitch rate refers to the way the flight control computer controls the transition from the initial "1G" steady state to the commanded "x G" steady state.

The FCC initially orders a movement of the elevator. It monitors the response of the airplane and adjusts the elevator movement accordingly. The 'response of the airplane' is expressed in a 'feed back parameter' C* which is formed by adding the sensed pitch rate multiplied by a constant to the sensed G-Load. Thus the value of the constant determines the 'blend' between the pitch rate and G-Load terms that together form the feed back.

[RetiredF4]

Sometimes it comes handy to go way back....to 1986
A320: fly by wire airliner (http://www.flightglobal.com/pdfarchive/view/1986/1986%20-%202148.html)

It even answers an old discussion concerning the profile of the tailplane we had in one of the AF447 threads:

The A320 tailplane is really a small wing, says Bord, and has a sophisticated inverse profile, the result of a similar optimisation process to that used for the wing itself.

[mm43]

It even answers an old discussion concerning the profile of the tailplane we had in one of the AF447 threadsIt certainly does, and I recall we agreed on the negative camber (http://www.pprune.org/tech-log/447730-af447-wreckage-found-89.html#post6522412) at the time.

[Iver]

Most Airbus pilots I talk to either love it or hate it. Learning the French FMS is evidently difficult if you are used to the Boeing way of doing things. In this economic climate, flying any airliner (keeping your job) should be appreciated. Enjoy flying to the US Mainland, Australia and Asia in the A330! Mahalo!

Best video of using the sidestick is this one:

AMAZING Views of an Airbus A319 Cockpit - YouTube

[Microburst2002]

Besides, I guess there's no speed input into the pitch law (alternate pitch law is the same as normal and it works even if all ADR's are lost).How would this g/pitch blending work?


In alternate law, pitch law is purely G load demand, with no pitch rate, but I can't give you an specific document with that. It seems logical, however. Without any speed input, how to blend pitch and g load according to speed?

[HazelNuts39]

@Microburst2002:

The blend of pitch rate and G load is determined by the constant factor of the pitch rate term in the C* feed back parameter. It does not require a speed input.

The airplane aerodynamic characteristics (not the FBW control law) determine the airplane response to elevator movement in terms of pitch rate and G-load. These are speed-dependent, of course.

The 'gains' used in the elevator control loop are normally speed-dependent. However, in alternate 2B law, the final report on AF447 states in paragraph 2.2.5:
In the specific case of alternate 2B law, some coefficients used in the longitudinal flight control law become speed-independent and are set for te maximum speed for the aeroplane configuration (330 knots in clean configuration)

[Microburst2002]

hazel

to be honest, I don't understand that factor you mention in the C* law.

I know that the airplane will respond aerodynamically to elevator changes, but the elevators will constantly move to meet the demand of the sidestick as interpreted by the computers, which is different depending on the active law, normal or alternate and flight condition.

I don´t know exactly how the "blending" works, but the idea that I had is that at low speeds we are demanding not only g load but also pitch rate, while at high speeds we are demanding g load alone, and that speed is somewhere in the algorithm.

What you say is that normal law simply demands pure g load, as it is stated in the FCOM, with no "blended" demand of anything else?

[DozyWannabe]

Here's some detail from an e-mail correspondence I had some time ago (it provides some info on the differences with the Boeing system too):

Wrt 'stability' as provided by Airbus FBW laws etc, they are actually using the space shuttle C* law.

Aircraft natural longitudinal stability can be regarded as having two modes: the short period which is essentially an interchange of pitch and vertical speed at constant forward speed (and where of course vertical speed variations at constant forward speed are effectively AoA changes), and the so called 'phugoid' mode which is effectively an interchange of forward speed and altitude. This latter shows itself to the pilots as an ability to hold trimmed speed without any pilot input.

With a C* law in place this changes. The short period motion is obviously modified because stick movement no longer commands pitch acceleration but pitch rate or normal acceleration and the damping is adjusted to give an optimum response as seen by the pilot; that is a reasonable response rate commensurate with the size of the aircraft combined with limited overshoot (about 5% used to be a good value)

The bigger change is that with zero stick deflection the system (as used by Airbus) maintains 1g level flight. This means that the phugoid is suppressed since no interchange of potential and kinetic energy is possible. This is turn means that trim stability is wholly dependent on how (T-D)/W varies with airspeed. Most of the time the aircraft flies above minimum drag and the 'speed stability' is positive, but maybe not exciting. For this reason Airbus FBW aircraft are flown with autothrust engaged most of the time. If A/T is not available then one is thrown back on the (T-D)/W variation.

Boeing, in their FBW version chose to modify C* by adding a reference speed term and calling the result C*U. With this system they are effectively producing a stable 'phugoid' of the classical type, but of course they must be accepting small altitude variations to maintain airspeed if there is no A/T in operation.

Effectively the Airbus FBW designs are neutrally stable in pitch and may, or may not be, speed stable depending on flight condition.

[HazelNuts39]

Hi Microburst2002,

Like you, I don't know exactly how the "blending" works. I'm just piecing together various bits of information that have been posted or linked in various threads on this forum. At the moment I'm slightly handicapped writing from memory because I'm away from the computer where I stored those bits.

IIRC there was a Cranfield paper that defined C* as: C* = n + A*q,
where n is incremental normal load factor, q is pitch rate, and A is a constant.

There was also a paper from a mr. Fabre of Airbus that described the implementation of the C* law on Airbus airplanes and showed a diagram of the longitudinal control logic in an inner and an outer control loop. The outer control loop compared the feed-back value of C* calculated from the actual values of n and q to the commanded value of C*. The commanded movement of the elevator was then proportional to the difference between feed-back C* and commanded C*.

Standing by for corrections.

[Stuck_in_an_ATR]

Turns out it's much mor complex than I've ever suspected... :{

@Hazelnuts - can you please explain what exactly does the term C* is? Is it some physical value? What does it represent?

I am starting to regret not paying attention on Automatics 101 at my uni... :ugh:

[HazelNuts39]

Hi Stuck,

What can I add to C* = n + A*q ? It represents a 'blend' of G-load and pitch rate that is fed into the control loop to command the position of the elevator.

[mm43]

@HazelNuts39

This AF447 Thread No8, page 89 (http://www.pprune.org/tech-log/482356-af-447-thread-no-8-a-68.html#post7251922) link will take you back to the original discussion re the C. Favre paper, and other C* references on the same page.

[DozyWannabe]

Turns out it's much mor complex than I've ever suspected... :{

Not to worry, you don't need to know all the equations behind the technology - whether that be cables and counterweights, advanced FBW or anything inbetween - to fly the aircraft well.

[Stuck_in_an_ATR]

Not to worry, you don't need to know all the equations behind the technology - whether that be cables and counterweights, advanced FBW or anything inbetween - to fly the aircraft well.

I agree. That's a wonderful trait of human mind - it can successfully cope with whatever control arrangement it has, without even noticing the difference...

However, considering myself a bit of a techie type (un-justifiably, it seems:}), I'd really like to now how it all works...

What can I add to C* = n + A*q ? It represents a 'blend' of G-load and pitch rate that is fed into the control loop to command the position of the elevator.

I kind of understand this represents the blend of pitch rate and load factor (even found out in another paper that "A" is actually Vcrossover/g - that's where the airspeed comes in, even if not directly fed into the computer). However...

I have always thought that a "control law" would be represented as:

output = function of (input, other variables),

How does the C* term (which represents the nature of the control law) translate into the above?

[gums]

Salute!

Thank you Doze and Nuts to add the technical and aero stuff.

Our small jet used "stby gains" when we lost the air data. The stby gain for total pressure was basically sea level 14.7 psi, dynamic pressure was about 300 knots gear up and maybe 160 or 180 knots with gear down. The ratio was what HAL used to determine control surface deflection and rate. So like the 'bus, we had a "heavy stick" unless going the speed of stink.

Remember that the FBW control input never commands absolute surface movement rate nor position unless in the 'bus "direct" mode. We didn't have that option. So the biggie is the gains determine the rate of movement of the control surfaces. Body rates are blended with existing gee until reaching the commanded gee. This prevents a "rough" ride and helps to prevent overshooting your desired AoA/attitude. Once getting to your desired pitch attitude, relaxing the stick allows the system to maintain the gee ( one gee corrected for pitch attitude in the 'bus, and trimmed gee in our little jet).

Still seems like many here do not fully understand the control law implementation, and the plethora of reversion modes in the 'bus is still confusing to this old fart.

[HazelNuts39]

output=function of (input,other variables)As I understand it (until I get a better explanation):
Output is a signal to the actuator to move the elevator up or down
Function of - is the gain"
Input is delta (C*) , i.e. the difference between commanded and actual C*, where commanded C* is a function of side stick longitudinal angle and actual C* is calculated from the G and pitch rate sensed by the relevant accelerometer and gyro sensors.

[Stuck_in_an_ATR]

As I understand it (until I get a better explanation):
Output is a signal to the actuator to move the elevator up or down
Function of - is the gain"
Input is delta (C*) , i.e. the difference between commanded and actual C*, where commanded C* is a function of side stick longitudinal angle and actual C* is calculated from the G and pitch rate sensed by the relevant accelerometer and gyro sensors.


Thanks Hazelnuts! Now, it is starting to make sense! :ok:

Just to make sure I understand it correctly:

-Each SS longitudal angle actually corresponds a "g-load" AND a pitch rate

-Then, "commanded C*" is calculated (which is solely a function of SS angle).

- "actual C*" is computed from nz and pitchrate sensors and then subtracted from the "commanded C*) to obtain delta (C*).

- The computed delta(C*) x "gain" = elevator deflection.

- elevator deflection then changes nz and pitchrate. New "actual C*" is computed and fed back into the loop...

Right?

[Craggenmore]

Thanks Hazelnuts! Now, it is starting to make sense!

Just to make sure I understand it correctly:

-Each SS longitudal angle actually corresponds a "g-load" AND a pitch rate

-Then, "commanded C*" is calculated (which is solely a function of SS angle).

- "actual C*" is computed from nz and pitchrate sensors and then subtracted from the "commanded C*) to obtain delta (C*).

- The computed delta(C*) x "gain" = elevator deflection.

- elevator deflection then changes nz and pitchrate. New "actual C*" is computed and fed back into the loop...

:D :D:D In gusty conditions at 300ft AAL, this is the last thing you are thinking about..!

[Chris Scott]

hikoushi,

You will have a lot of fun on the A330, and there's a lot of good stuff here. I don't know what Iver's problem is with the FMS part of the FMGS (or even if he is a user of it), but I must correct any false impressions you may have gleaned from the YouTube video he recommends.

The video itself is great; handling of the sidestick poor by both PFs. It's a tribute to the FBW system that stirring the pudding to that extent does not produce PIO.

In the case of the L/H seat PF, three faults are evident.
1) The stick is never stationary in the neutral position (which it should be much of the time, even when manoeuvring). It's always on the move.
2) He constantly holds the stick in his hand, instead of using fingers and thumb.
3) His armrest is so badly positioned (too low, and wrong rake-angle) that he is unable to rest his forearm on it. (See the landing approach at the two-minute point on the video.) As a result of this and (2), his whole forearm moves back and forth, unsupported, during pitch commands.

It's very important first to adjust the armrest to the best graduations of height and angle that suit your arm. These can be quickly established during type conversion. The best way to operate the sidestick is to rest your elbow stationary on the armrest so that your open hand surrounds the top of the stick without touching it. Your wrist should be clear of the armrest at all times while you are controlling the stick, so that your hand movements are unimpeded. Your fingers will be on the outside of the stick; your thumb on the inside of it.

The following description of how to control the stick assumes you are in the L/H seat, the R/H seat obviously being a mirror-image of the left (I had to maintain currency in both). Remember: your elbow should not move.

Use your thumb to move the stick to the left, and to move it forwards. The transition from one action to the other is achieved by twisting your wrist through about 90 degrees. Diagonal movements (forward-left) are achieved with the wrist somewhere between the two extremes.

Use your fingers (as many as you like) to move the stick to the right, and to move it backwards. The transition from one action to the other does not require as much of a twist of the wrist as with the thumb action.

Someone said recently on the Airbus FBW thread that stick commands should be small, but long-lasting. I partly disagree. The only long-lasting commands I can think of are to rotate the a/c on take-off or go-around, and rolling it from a turn one way into a turn the other way. Landing flare is comparatively brief. *** There are no hard and fast rules, but normal commands are small-to-medium jabs, their effects limited by short duration. Between each jab, the stick should be allowed to return to and remain in neutral while - with your hand relaxed - you observe the effect of the command.

So how to fly the Airbus FBW manually? I'm not going to try and discuss C* in this post (and others are currently doing that). But I can offer a flavour of what it's like, from a pilot's perspective. Five years ago, during PPRuNe discussions on an A320 crosswind incident at Hamburg, I wrote and amended this, which may be of some (unofficial) interest:

http://www.pprune.org/tech-log/316096-lh-a320-rough-landing-hamburg-12.html#post3956007

On the general subject of how the sidestick is often misused, I offered this:

http://www.pprune.org/tech-log/316096-lh-a320-rough-landing-hamburg-24.html#post3979423

Disclaimer: Although I flew the A320 for 14 years, I retired at the end of 2001; so am way out of date. Anyway, I hope this helps.

*** [EDIT] This is correct in terms of the time taken to rotate the a/c by only about 3 degrees. However, it overlooks the effect of Landing Mode, which - below a certain height (30ft in my 1988 FCOM) - requires the PF to pull the stick progressively backwards: even to maintain a steady pitch-attitude. This feature goes some way to simulate the stick load on landing with traditional controls. Without Landing Mode, it would be easy to over-flare.

[HazelNuts39]

Hi Stuck iaA,

That is my understanding too. But it is perhaps not the whole story. Some PPRuNers will recall my failed attempt to derive the C* coefficients from the FDR data of AF447.

What puzzles me in particular is the FCOM and other Airbus documents always saying that the side stick commands a load factor, and Figure 4 in C. Favre's paper (copied in the post linked in mm43's post above) that seems to indicate a special treatment of (commanded - actual) vertical load factor. That's why I wrote that the 'blend' with pitch rate is mainly important for the build-up phase of load factor, and perhaps less so for maintaining the commanded load factor in a steady-state pull-up maneuver.

[Stuck_in_an_ATR]

Hi Stuck iaA,

That is my understanding too. But it is perhaps not the whole story. What puzzles me in particular is the FCOM and other Airbus documents saying that the side stick commands a load factor, and Figure 4 in C. Favre's paper (copied in the post linked in mm43's post above) that seems to indicate a special treatment of (commanded - actual) vertical load factor. That's why I wrote that the 'blend' with pitch rate is mainly important for the build-up phase of load factor, and perhaps less so for maintaining the commanded load factor in a steady-state pull-up maneuver.
4th Jan 2013 22:29


I'm glad I managed to understand at least a bit of this stuff... :8 Big thanks!

While I couldn't fully comrehend the figure from Favre's paper (and can't get access to the paper itself), I understand that in steady state the load-factor may be dominant. That, however, is different from the basic C* algorithm (or at least my understanding of it). Isn't it?

Anyway, I guess some book-digging is on order... :)

[hikoushi]

Really great stuff here, although I'll probably have to come back to some of the deep technical stuff both here and in the AF447 threads after getting checked out and having some time to get familliar in type. And Chris, great practical info on the handling techniques, will definitely reference that again before going to the sim.

Thankfully plenty of others who did the same transition, most of whom are smarter than me, to lean on for help and therefore be obligated to purchase libations for in the coming months.

And the "stick-handling" (ahem) in that video, looked more like the last 30 feet of a DC-9 landing in a 35 knot gusting wind than anything any of our Airbus converts have described!:}.

[Stuck_in_an_ATR]

And the "stick-handling" (ahem) in that video, looked more like the last 30 feet of a DC-9 landing in a 35 knot gusting wind than anything any of our Airbus converts have described!.

I used to criticize this guy's stick handling too. I thought my piloting technique is much better - until I installed a camera in the cockpit...:}

[Chris Scott]

Stuck_in_an_ATR, quote:
"I used to criticize this guy's stick handling too. I thought my piloting technique is much better - until I installed a camera in the cockpit..."

Love it...! (English-style modesty will get you nowhere in the dog-eat-dog world of PPRuNe, but I admire your self-deprecation!)

And, as I’ve recently commented on the Airbus FBW thread, the PNF has absolutely no idea what the PF is doing with his stick.

For hikoushi: not only are they not connected, but even looking across is futile. The only viewpoint is from the P3 seat, where I spent a proportion of my years.

[A33Zab]

Below a basic scheme of C* just to get an idea,
In real - pure C* - control is not applied in aviation, gains and filters are added to enhance the control law for certain flight phases.


http://i474.photobucket.com/albums/rr101/Zab999/C_star_zps795d1325.jpg

[Clandestino]

So how to fly the Airbus FBW manually?

Personally, I have never given much thought to it. I just went out and flew. C*U blending? Degraded laws? Whatever. If the aroplane's attitude doesn't follow your sustained input, you are in much, much bigger trouble than simple FC mode change.

Based on my limited experience, I can certify everything Chris Scott has written here on the subject of manual Airbus control is sound operational advice.

Mind you, there were two caveats on single-aisle Airbi regarding dual hydraulic system loss. It's been four years since my last flight on the minibus so I cannot remember anymore which was which but in one case you'd be left with elevators only and you had to drop the wheels at exact speed (which was "remembered" as neutral reference by elevators) otherwise approach feels normal but pitch-up at go around would be very hard to contain. The other was loss of one elevator leading to other being reduced to half its travel not to load the tail with too much twist, basically leaving you with just 1/4 of normal elevator authority - sufficient but the aeroplane felt sluggish so landing at airport with minimum crosswind and turbulence was advised.

I'm pretty sure widebodies have different hyd and FC architecture but if among flight patterns in FCOM you find "Approach and landing with xxxxxx inoperative", follow it as precisely as you can, if it is there it is there for a very good reason.

[gums]

Some good points by both Clan and Chris.

The advice about supporting elbow seems a great thing to do. We also had a "wrist grip", but most folks stowed it and just used the elbow/forearm support. The good thing about supporting the wrist was for smooth inputs, as our stick moved less than 1/8 of an inch ( force versus displacement as in the 'bus).

Back to the "connected sticks/yokes" once again, huh? I must defend the position of some here that it doesn't make a lot of difference. As Clan has said, if the nose ain't going where it's supposed to be going, then the other guy is doing somethin weird or the jet is. Not only were we not be able to see the other stick ( back seat/front seat), but the doggone thing moved so little it would not have helped. If the jet was not going where it was supposed to, we could "add" input. If that didn't help, we could take control with a paddle switch below the handgrip, and I think a similar feature is implemented in the 'bus.

Some good points for a newbie.

[Chris Scott]

Thanks for that, Clandestino, but I've just realised a careless error in my post: re the examples of "long-lasting commands".

I opined that the "landing flare is comparatively brief", which is true in terms of the time taken to rotate the a/c by only about 3 degrees. However, it overlooks the effect of Landing Mode, which - below a certain height (30ft in my FCOM) - requires the PF to deflect the stick progressively backwards: even to maintain a steady pitch-attitude. This feature goes some way to simulate the stick load on landing with traditional controls. Without Landing Mode, it would be easy to over-flare, as all AI FBW pilots know.

Am amending my post.

Gums,
Good to do business with you again, and I take your point on interconnection. The size and position of the stick really doesn't lend itself to either tactile or visual cues - assuming they were boot-strapped manually or electronically. That's not to say it doesn't matter; only that it's probably not practicable in the way it is with old-fashioned yokes. I remain concerned about what seems to be the persistent pot-stirring of some pilots, but you can only observe it from the P3 seat.

[gums]

Salute!

TNX, Chris.

I'll hang around anytime we talk about FBW concepts and implementation, especially after a mishap. And I'll match my starting date in a full FBW operational plane with anyone here. It's my only credential for joining in the robust discussion amongst the august, knowledgeable "heavy" drivers here.

We really need to define "hand flying", as the thread is titled.

I assume it means no auto anything, including the throttles. And for the "conventional" folks here, there's no such thing as "hand flying" a FBW jet as long as there's a computer between you and the control surfaces. The 'bus "direct" mode seems closest, but you still don't get tactile feedback from any of the control surfaces - kinda like flying the old simulators or most military fighters since the late 50's with only hydraulic fluid to the actuators based on your position of the valves the stick controlled. Doze will be quick to point that out, heh heh.

[CONF iture]

As Clan has said, if the nose ain't going where it's supposed to be going, then the other guy is doing something weird OR the jet is.
That's the valuable advantage of visible control yokes or other linked side sticks, you can forget about that "or" as you're one step ahead, information being readily available, no guessing :

The pilot is doing the correct command but the airplane is not following.
The plane cannot do well as the pilot is obviously not commanding well.

[Stuck_in_an_ATR]

Originally posted Chris Scott
Love it...! (English-style modesty will get you nowhere in the dog-eat-dog world of PPRuNe, but I admire your self-deprecation!)



My post was genuine. After watching the aforementioned video I have even made a rant on a different forum about proper 'Bus flying technique, "making mayonnaise" vs "touching your best friend's @##*" etc. Then I really put a GoPro camera behind the seat and made some interesting observations... :)

In "stable" flight phases, like climb-out, or approach one can really minimize S-S inputs.

However, when things change fast, eg. the landing flare old habits kick in subconsiously and the sidestick tends to be all over the place, without the pilot even realizing it. Thankfully, the airplane doesn't seem to care either...

@A330Zab

Thanks for the C* diagram. Can you post a link to the original article where you took it from?

[A33Zab]

Sure........linked here (http://www.polymtl.ca/biblio/en/epmrt/rapports/rt2005-01.pdf), keep in mind it is not Airbus specific. (AFIK airbus don't use the Kff - feedforward - gain in A32/A33/A34).

[Stuck_in_an_ATR]

Thanks A33Zab! I have something to crunch on for next few days!:ok:

Re. feedforward (or lack of in the 'Bus). I noticed the Kff gain on the diagram you posted and wondered what it was for. Is it some kind of "anticipator", to get the elevators moving bedore the "proper" C* signal is generated?

[anotheruser]

I was always wondering, what's the flare mode good for? Why can't the plane just stay in normal law until touchdown or simply go into direct law at some point instead of going into flare mode?

[A33Zab]

Found this description (Search: FeedForward PID controller) on the net.


In contrast to feedback control, feedforward control acts the moment a disturbance occurs, without having to wait for a deviation in process variable.
This enables a feedforward controller to quickly and directly cancel out the effect of a disturbance.
To do this, a feedforward controller produces its control action based on a measurement of the disturbance.
When used, feedforward control is almost always implemented as an add-on to feedback control.
The feedforward controller takes care of the major disturbance, and the feedback controller takes care of everything else that might cause the process variable to deviate from its set point.

[737Jock]

The video itself is great; handling of the sidestick poor by both PFs. It's a tribute to the FBW system that stirring the pudding to that extent does not produce PIO.

The video's are accelerated from what I can see... So what you see is not in realtime...

[gums]

Guess I didn't read all thru the various laws and such, but found the "flare" mode "upon further review".

Wow! Sucker is basically in "direct" mode with some rate and gee dampening, then requires increasing back stick below 50 feet to maintain or increase pitch. Gotta admit that the mode should make the jet feel just like those of yore.

We didn't have the altitude sensor ( 'bus apparently uses RA), but we had a strong AoA bias with gear down, so you "had to" flare or would just drive onto the rwy at existing pitch attitude.

This should make Lyman and a few others feel all warm, heh heh.

[fantom]

I was always wondering, what's the flare mode good for? Why can't the plane just stay in normal law until touchdown or simply go into direct law at some point instead of going into flare mode?


Because pilots usually want to pull back a little for touch-down.

'Flare mode' applies a little nose-down pitch (which the pilot needs to correct) to make it all feel normal.

[Iver]

Chris Scott,

Thank you for your critique. My point related to the French FMS was that pilots who have made the transition from Boeing to Airbus have TOLD ME that, in many cases, they found the FMS switch/logic confusing. You were not privy to those conversations. Understood? Your opinion might differ from the 5-10 Airbus pilots who have transitioned from Boeing aircraft that I have talked to.

Additionally, I did not provide the "SAIL" Airbus cockpit video as an instructional video. Instead, I, and many others, find it entertaining and interesting - especially for those of us who use traditional yokes. The main pilot in the video clearly does not care about "proper form." And that's fine because the video (especially the descending left turn approach into Rio's Santos Dumont airport) was entertaining.

Cheers

[Fursty Ferret]

My post was genuine. After watching the aforementioned video I have even made a rant on a different forum about proper 'Bus flying technique, "making mayonnaise" vs "touching your best friend's @##*" etc. Then I really put a GoPro camera behind the seat and made some interesting observations...

Another thumbs-up for this post. I was embarrassed after watching my inputs in a GoPro video, and it's something I recommend to every Airbus pilot (doesn't have to be a video, ask to see a flight data playback of your sector).

Despite what I like to think, the Airbus rides turbulence fairly well on approach and I've adopted a much more "hands off" attitude to flying it, preferring to make just one or two corrections to correct trends in the trajectory instead of stirring porridge the whole way down.

One thing you can do, however, is try to make the inputs smooth. It's not immediately obvious at the front but when someone is putting in rapid "squirts" (even small amplitude) it's very pronounced in the cabin and accentuates motion sickness. In calm conditions the Airbus can be flown with "pressure" on the sidestick only.

[Microburst2002]

My personal advise as to how to control the sidestick is that you have to be conscious that you are trimming the airplane whrn you release it to neutral. So, when you are reaching your desired pitch and performance, progressively release it until,you reach zero SS angle exactly at the desired performance.

In final approach, to avoid overuse of the sidestick you have to give a second to the system to counteract lateral gusts itself, then bank to regain loc or runway extended centerline. So it is like this: lateral gust that rolls your wings. Let the system zero the roll rate, then use the sidestick to go back to the desired bank angle to keep or regain the loc. If you act simultaneously with thw system you will begin the pudding thing. Just wait a second then make your input. With pitch is similar but not exactly the same. You have more things affecting, gusts, updrafts, downdrafts, a/thr... You may need to act preemptively

For the flare, I don't recommend the widespread use of repetitive pull- release-pull- release that I have observed many times. I favour a single pull, that you can modulate as required. Ah, and never use the flare pitch bar for landing. That's gross!

[gums]

I have finally been rewarded by Fursty and Micro.

Their techniques and observations have answered many of my questions and confirmed my suspicions.

Either of them would adapt to the Viper stick/control laws in about 4 seconds.

The plane's stability when in turbulence became very obvious to us after a mission or two. Our little jet weighed less at takeoff then the weight of the fuel in the 'bus, heh heh. Nevertheless, the reaction by the jet to keep roll rate and trimmed gee per the laws made the thing feel like a much larger/heavier plane in light to moderate turbulence or when zipping along at 540 knots and 200 feet over a desert with all those thermals. So the techniques described and the aircraft's reactions fit my experience perfectly. Same-o, same-o. Control surfaces were constantly moving to achieve the trimmed/commanded gee or roll rate.

Same for the flare technique. Never reverse the stick beyond neutral unless a severe gust, just don't pull as much or maybe relax back stick ever so slightly, but one long, gentle backstick input until the wheels squeak. If you ain't holding a slight back stick at touchdown, you'll probable bounce.

So I finally got to hear how that sucker flies when you are not letting Otto do everything. Attaboy. And thanks, guys.

[DozyWannabe]

Guess I didn't read all thru the various laws and such, but found the "flare" mode "upon further review".

Wow! Sucker is basically in "direct" mode with some rate and gee dampening, then requires increasing back stick below 50 feet to maintain or increase pitch. Gotta admit that the mode should make the jet feel just like those of yore.

Hey gums - to be fair, the discussions we were having before wouldn't have touched on flare mode because we were talking about a cruise scenario - even a non-pilot like myself knows what it is and would have been happy to tell you. The systems were checked out by nuts-and-bolts engineering pilots from the get-go - the best that could be hired, and the truth is that every single mode was designed to be as transparent to the pilot as possible.

Very glad it's clicked for you though! :ok:

[gums]

Salute!

Lemme review the bidding, Doze.

I thot this thread was about "hand-flying" the jet. I realize that much was focused upon the cruise phase, but sooner or later ya gotta land the sucker. Oh yeah, ya gotta take off to get to the cruise phase. Then a climb, and then a descent. I can just see Joe Baggodonuts ( think Bolin, RIP) descending in sub-law 2(b) modified, or "direct". Think you would have a smooth ride back on seat 26B?

Oh yeah, the modes are so damned transparent that many pilots don't even know what mode they are in!! And then we have a crew that turns off the air data in order to wrest control from HAL when the AoA sensors went to hell. Gotta hand it to those guys.

BTW, you may not have read my previous post about landing the jet and flying in turbulence. Sheesh. Your post hours are getting close to my 2,000 hours in single-seat, single engine jets with about 2/3 of those either full FBW or FBW assisted ( go look up the A-7D "control aug" and then think about Viper prototype using the A-7D stick grip because it had all those force transducers that our system used to make things smoother and more precise).

And there' no such thing as a dumb fighter pilot. Out!

[DozyWannabe]

I thot this thread was about "hand-flying" the jet. I realize that much was focused upon the cruise phase, but sooner or later ya gotta land the sucker.

I was referring to the previous threads on AF447 - not this one. And my previous post wasn't meant to be a challenge! :)

Oh yeah, ya gotta take off to get to the cruise phase. Then a climb, and then a descent. I can just see Joe Baggodonuts ( think Bolin, RIP) descending in sub-law 2(b) modified, or "direct".

Doesn't matter what law you're in - slamming the PFC back in cruise is not the done thing. Normal Law will protect you if it is available should you do so, but directly following an abnormal AP disconnect is not a good time to rely on Normal Law being in operation. The history of "startle effect" contains many instances of pilots pulling up instinctively in unprotected aircraft - this cannot help but cast reasonable doubt on the idea that F/O Bonin pulled up because he reasoned he would be protected.

Oh yeah, the modes are so damned transparent that many pilots don't even know what mode they are in!!

I think that most (probably all) type-rated pilots do know, but it's usually a background detail, and that's kind of the point - the system was designed such that in terms of day-to-day operations, the current mode (distinct from flight law) should seem like a seamless operation to the pilot

And then we have a crew that turns off the air data in order to wrest control from HAL when the AoA sensors went to hell. Gotta hand it to those guys.

As their training and systems knowledge would have suggested. It was a good recovery, no doubt - but the systems aren't as unfathomable as some would like to argue.