(NOTE: THIS DESIGN HAS NOW BEEN
SUPERSEDED - FOR MORE RECENT DEVELOPMENTS SEE
HERE)
PERFORMANCE
The single most important feature of a force feedback system
is the quality of the force feedback - I haven't yet figured
out how to communicate this by internet however. So you'll
have to make do with an honest description and some video
clips....
I think the force feedback is reasonably good. The force
application is generally smooth and although there is no sense of
motor cogging torque fluctuations in the feel there is
discernable commutation related ripple. I think that
the elevator forcing is slightly smoother than the aileron
forcing, but am not too sure why.
The performance is sensitive to the set up of the mechanical
transmission also - for example drive belts that are
too slack introduce lag, if they are too tight they
introduce friction loading at the bearings which makes the
movement slightly sticky.
With the aileron axis at higher loads (eg when the ailerons
are deflected into a higher speed airflow) making very small
one-handed position adjustments can bring through
discernable but small force fluctuations in the feel - due
to movement of the motor brushes over adjacent commutator
segments in the motor. Most of the time, with two-handed
operation, lower forces and faster movements the
fluctuations are less noticeable. I think there is a
good chance that I am more sensitive to this than others
might be because I spent a lot of time trying to eliminate
variations in the force feel - so I always notice them. For
a sim pilot otherwise occupied by the tasks of flight and
all the associated cockpit activity these small force
fluctuations might not register. However I also think this
might be quite a personal thing which might be more
noticeable to some people than others.
The elevator feel is good I think, however as with the
aileron axis, motor commutation based force fluctuations are
still discernable, and especially at the higher load levels.
These appear at a higher frequency than the aileron drive
due to the higher transmission ratio used in the elevator
drive. The best way I can describe this is as a slight
rumble in the feel when the column is displaced towards its
end of travel when there is a fast airflow over the elevator
surfaces. Again, these force fluctuations with two-handed
operation at normal force levels might not register, but I
don't like them.
In terms of the overall feel the effect most noticeable in
flight is the way in which the controls stiffen up as the
airspeed increases. The feel is very different cruising at
speed when compared to a much slower landing approach for
example. The way in which the trim position changes with
airspeed is also noticeable. Prop wash effects with the
aircraft on the ground are interesting - the elevator
stiffens as the engines are powered up if prop wash effects
are included in the configuration. I like the difference the
force feedback makes - you sort of get used to the
information flow coming back which tells you almost
subconsciously of the condition of the aircraft. For example
the reducing control stiffness can alert you to dropping
airspeed before you have noticed it on the instruments.
I haven't thought much about incorporating failure
situations that would affect the control forces - any
suggestions?
SOME VIDEO CLIPS
Perhaps a few video clips will give some sense of the flight
column operation. Please note that the programmed behaviour
of the flight column is still preliminary and the clips
illustrate typical behaviours. One of the great benefits of
taking a digital approach to the force feedback is that,
within reason, any loading behaviour can be programmed for
the system.
This first clip below shows the column
being activated with an FSX flight in the air and paused.
When activated the forces bring the controls to their
trimmed balance position and the flight can be resumed. You
can't see the forces off course but note how the controls
return towards center when released. When the force feedback is
disengaged the column returns to its inactive response.
In the second clip below the scenario is similar. Included
in this one is illustration of the effect of adjusting the
trim pots. The trim system is completely separate from FSX's
internal system. Adjusting the trim pots on the column
causes the trim surface force magnitude to be altered as it
is deflected more or less into the airflow - this alters the
force balance on the main control surface and shifts the
control axis zero load position. The trim can be used to
bring the column to balance at any position. However the
trim force magnitude is airspeed dependent, so the balanced
position does not remain in the same place if the airspeed
changes.
For this third clip below the aircraft is on the ground and
stationary. The force feedback software allows the elevator
to be configured as load balanced or unbalanced. In this
case the unbalanced settings are used and the elevator
weight draws the flight column forward when there is no
airflow. On some propeller driven aircraft prop wash
generated on power up can induce aerodynamic loading on some
control surfaces. In this clip powering up the engine causes
the elevator to move to its trimmed balance point.... and to
fall forward again when the power is removed.
And finally the clip below also shows prop wash effects, but
this time using a load balanced elevator configuration - ie
one in which the elevator remains in the position to which
it is moved. Unless of course there is significant air flow
- then the control surface is forced to the load balanced
position.
SOFTWARE
The force feedback model calculates control force from
control surface position, airspeed, trim forces, weight and
aircraft acceleration information. Moderate damping and
friction effects can be added as can some "manufactured"
effects such as engine vibration and stall buffeting. How
good these latter effects are I'm not sure.
The software is still in a preliminary state but it does allow a
fair bit of configuration freedom with settings to be saved
for different aircraft and set ups. All the effects are
tuneable in terms of intensity and there are also a number
of more fundamental parameters available governing the extent of
internal filtering applied to various terms. The software
includes output for the Rudder control loading but I haven't
been able to do any refinement work on this as I haven't
built a set of rudder pedals.
Overall I think I've made progress with the force feedback
system. However it is indeed a tricky thing to get right, the drive
motors are expensive and the mechanical design and build
will be fairly unforgiving of poor design or workmanship.
The performance is fairly good - but not perfect.
In the UK the MSS-4 motors cost roughly UKŁ400 each, the
MSS-2 is slightly cheaper and might be adequate for the
aileron. I wouldn't go smaller than the MSS-4 for the
elevator drive. The remaining electrical and control hardware and
mechanical components might cost around the UKŁ300 mark
which would give an approximate build cost for the force
feedback flight column of UKŁ1100.
Now I need to mull over where to take the project next - a
more refined column design might be worth looking at. Using
better quality ball bearings rather than cheap needle roller
bearings might help further reduce friction levels when the
drive is fully loaded. I would also like to investigate
better location of the aileron drive motor which I have just
stuck on the column at the moment. This works from a
transmission point of view but it does affect the weight
balance of the elevator axis (which is compensated for by
the software) and it doesn't look good. A wider belt on
the final elevator drive stage might be worth trying - the
belt tensions in that stage are quite high. The question is
whether or not a well set up mechanical drive which is stiff
with low friction will help reduce force ripple to the
extent that it can be accepted by the user. I also think it
might be worth experimenting with a much stiffer geared
transmission for the elevator drive - although backlash
control might be a problem there if lower cost gears are
used.