From:         Martin Fiddler <>
Date:         22 Jun 95 03:07:35 
References:   1 2 3 4
Next article
View raw article
  or MIME structure

>On 21 Jun 95 02:56:54 Jean-Francois Mezei <MEZEI_JF@Eisner.DECUS.Org> wrote:
>The 310 may have the same joystick as the 320 and 340 but there is a major
>difference: FBW. (subject of this thread). In FBW systems, the computer has
>much more intelligence built in to it whereas the more "conventional" system
>have a basic auto-pilot that maintains altitude and course and then have an ILS
>to help position the plane in its landing path. Right ?

No, only the A320, A330 and A340 (and B777) are FBW.  The A310 has a
conventional yoke, not a joystick.  FWB merely means that instead of, say,
the hydraulic actuators on the elevators being signalled MECHANICALLY to
give 'elevators up', a computer in the flight deck ELECTRICALLY signalls
the computer controlling the actuator to give 'elevators up'.  FBY means
that the primary flying control actuators are signalled electrically rather
than mechanically, and just that.

A conventional aircraft can have a complex autopilot which is managed by
the flight management computer, incorporating vertical and horizontal
navigation, fuel management, etc, etc.   Look at the B747-400 systems -
yet that isn't FBW.  Most airliners made in the last 10 years have these
capabilities, but are not FBW.

>In a FBW system, the computer can make lickety split decisions if something
>arises. If the information given to the computer is wrong, then problems occur
>(eg: crash of 320 in France where computer aborted a take-off even though the
>plan was already in air because its altimeter "value" was still at 0).

I thought I was familiar with most of the A320 'crashes' but haven't heard
of this one for some reason...  Which airline, when?    And are you sure?

>In a FBW system, the pilots' commands are interpreted and "validated" before
>being acted upon.  In a non-FBW system, if the auto-pilot is off, the commands
>are just acted upon.
>If the above is not too far off from reality, then I have the following
>When you have a pilot, co-pilot and computer who can "steer" the airplane, in
>the case of the "tactile feedback", how is the following handled:
>Computer wants to veer left real. Pilot wants to go straight. Co-Pilot
>wants to veer right.

The 'computer' will only want to go left if the FMS is in lateral nav
mode, and the pre-programmed course is to turn left at a waypoint.   If
a/p is engaged, it will turn left at the correct time.    If the pilots
do NOT want to turn left because they are routing direct to a future
waypoint, or they want to avoid a storm, say, they can either reprogramme
the FMS using the data input screen, or switch LNav off and dial in
headings for the autopilot.   Or they can disengage the a/p and fly
manually if they really want.   Only the pilot or co-pilot will be flying,
and that one will decide which way to go if they want to re-route.  It is
NOT a long hard battle for the pilots against  the computer, contrary to
one thread running here!

>While the pilot may at first, feel the computer pushing the joystick/whatever
>to the left, he will then feel the joystick going back towards the "middle" as
>the co-pilot is veering to the right. Meanwhile, how will the pilot convey that
>no-movement is the thing to do while the others (computer and co-pilot) are
>"playing with the joysticks " ?

The joysticks have no feedback at all, they are sprung to centre.  Like
I said , it would be unusual to fly manually, but if required it would be
obvious which pilot was to do so, the flying-pilot.   I guess they get
used to roll and pitch rates, and use the primary flying display to get
their 'feedback' when flying an A320.

>Can the "tactile feedback" system also feed to the computer the amount of force
>that the pilots exert ? And does the computer then decide who has the most
>force on the joystick and then let that person't joystick move freely ?

Like others have said, one pilot can lock-out the other's joystick by
pressing and holding a priority button.   There is not normally a conflict
of interest between the pilots.   The one exerting most force does not
'win' - the inputs are summed if one joystick has not locked the other

No-one's mentioned this before, but if the a/p is engaged, there is a
force sensor on the stick.  If a VERY heavy force is applied, the a/p
disengages and control reverts to manual.  That's for emergency (eg
collision avoidance) manouvers where instinct is to bang the stick hard

>I am not a pilot, but I would assume that while the auto-pilot is engaged,
>pilots don't really have their hands on the joysticks/steering devices. Right ?
>Pilots going asleep during very long flights has often been mentioned as a
>problem where it takes a while for pilots to awaken when something does happen.
>Tactile feedback would not be of much help there, but having a smart computer
>able to quickly adjust to a situation such as air pocket would fill that
>"critical" gap between the time the situation arises and the time the pilots
>are "fully awake".

Good point, but the a/p can handle CAT ('air pockets') OK by itself.   If
they do sleep, only one pilot would be asleep at any time.

>And when the pilot is at the controls, then the FBW system only makes sure that
>the aircrafts' structural integrity is ensured. In the case of Boeing, it would
>SEEM that the pilot can do what he wants with just a warning on a screen if he
>exceeds the aircrafts' capabilities (what ? no tactile feedback?) :-)

In fact Boeing DOES provide tactile feedback, in that a light hand on the
yoke gives feedback regarding control movements, the pitch trim wheel
rotates to show trim changes, and the throttles move when in auto throttle
mode.  That is a pretty good definition of tactile feedback!!

>In the Airbus concept, it would SEEM that the computer won't let the pilot
>exceed the aircraft's capabilities unless a special "override" button is
>engaged. Is that a correct assumption ?

Airbus have full envelope protection for limiting speeds, angle of attack,
roll angle, etc.  I don't think it's possible to get around those
protections...   Maybe your override button is the one on the joystick,
but that just puts the other joystick out of influence!

>Now, it seems that tactile feedback is "required" in order to correctly asses
>who is at the controls (pilot or co-pilot). This is based on many opinions
>posted here. I am not a pilot, but I sure hope that pilots are trained in such
>a way that authority is based on more that just how much resistance a person
>has when try to move the joystick/steering column.

They are, absolutely.  I can't even imagine a case where they're not sure
who is flying.  In an emergency manouvre, the Boeing-style tactile feedback
would let one pilot know that the other is responding to the situation
and he would just monitor (and I guess it's hard to monotor too, on an

>Personally, I would feel more at ease where a system would be on auto by
>default, and if any of the two wish to go "manual", then the first one to push
>the button gets a distinct audio signal that both pilots can hear. Much less
>confusion than trying to figure out where the resistance is coming from in
>tactile feedback.

That's the way it works anyway, once the a/p is engaged.  If you go to
'manual' you get a warning wazz wazz wazz sound, and you assume the other
guy has disengaged the autopilot.  No confusion there.  The other pilot
is most unlikely to try and fight for control!

>As well, if you have feedback coming from multiple sources at the same time,
>you may suffer "information overload" If you not only have to feel the
>directional device to see where the plane is going but also look at the crt to
>see who is steering is (other pilot or computer), then you are wasting even
>more time.

The feedback is subconcious and the same as the pilots have had since they
took their first trial lesson in a cessna.  It won't overload them!

>While many see the lack of tactile feedback in the Airbus as a "fatal flaw", I
>am not willing to dismiss the Airbus concept until I hear some REAL facts about
>how it really works and what sort of computer-pilot communications exist.
>Judjing from some posts and some e-mails, some people seem to beleive that
>Airbus computers do a lot fo stuff and don't tell the pilot about it.
>I do not wish to state that Airbus is better than Boeing. But I think that that
>a lot of Airbus bashing is not warranted because we know so little about the
>man-computer interface in that system. Unfortunatly, Airbus's propaganda
>machine isn't as well organised as Boeing's.

There are, unfortunately, many well researched reports about Airbus
problems.  It is NOT to do with FBW, but partly with the fact that there
are joysticks.  That means that (a) there is no tactile feedback and (b)
artificial control laws have to be devised to translate the joystick input
into elevator or aileron movements.  The artificial laws have to change
depending on the circumstances, meaning many different FMGS/EFCS modes.
When landing, the modes will change from Alt star/aquisition mode, open
descent mode, aplha prot, floor or  max for windshear or go-around, ending
with landing mode - where the nose of the A320 automatcaly de-rotates
downwards, providing the pilot with a need to pull back on the joystick,
and hence to synthesise a 'flare' - which Boeing do not need to provide
an artificial mode for, as their pilots have a yoke which works
conventionally!   Most of these modes are somewhat transparent, and
remember that some modes are hard to get out of, and the throttles may
not respond if you are unknowingly in the wrong mode (see Bangalore).

>Next time I fly on a 320, I'll make a point to ask to go to the cockpit and ask
>the pilots a lot of questions.

I guess they'll be pleased to see you.  If you read up on the FMGS and
EFCS before you go, as well as the 'famous' A320 crashes, you'll make a
good impression.

Sorry this is so long, but the original was very long in the first place.
It's a shame that there is confusion between FBW and Airbus man-machine
interface problems.  The FBW is somewhat of an innocent, being blamed for
all the world's problems!