On Sun, 16 Nov 2003 21:11:40 -0800, Mary Shafer
wrote:

Every "game" simulator I've ever flown seemed to use the same math
model, one that, as you say, was not dynamically possible. Fun's fun,
but physics is physics.

.... yet any game uses a completely different engine to create the
flight model. The differences are where the game engine does its short
cuts to allow realtime operation.


The sims are too generic, partly
because there just isn't enough time and space for a detailed math
model, because the FCS is proprietary and much too big to be modeled,
because the control surfaces aren't modeled correctly, the mass model
isn't right, and so on.

The main problem of PC flight simulations is that the performance of a
PC is not sufficient to calculate a realtime aerodynamic simulation.
Mass models are ok these days in most flight sims, as well as
performance and envelope data which are in some cases very close to
reality. Only few PC simulations even try to simulate engine torqe
effects on prop aircraft.

The problems start if the simulated aircraft does non-linear maneuvers
(post-stall, spins) - this is when some PC simulations can get very
erratic because their simplified physics model needs to rely on
pre-calculated data (to save computing time). The result is either a
"standard" stall routine (always the same spin, independent on how you
entered it), or erratic movements that does not even look close to
what a real aircraft would do.

So far the only PC simulations that attempt to simulate post-stall
effects are MS Flight Sim 2002 and 2004, MS Combat Flight Sim 2 and 3,
and X-Plane, but the results are not entirely convincing yet.

Any PC simulator is (of course) handicapped most by the input devices
- a PC joystick and a mouse simply cannot give even a similar feeling
to the stick of a real aircraft (or a full cockpit simulator).This is
the cause why the characteristics of a PC simulated aircraft cannot be
even similar to the real thing, even if the performance data
throughout the envelope are very similar.



However, learning to "fly" with a fixed-base, low-fidelity sim game
isn't going to happen. All that will happen is that the student will
pick up responses and habits that will have to be unlearned before the
correct responses and habits can be acquired in the actual airplane.
I've heard flight instructors complaining about how they can always
tell if someone plays with MS Flight Simulator a lot, because it takes
a lot longer to teach them how to fly the actual airplane.

Indeed.
The training effect concerning a PC simulator is that of a procedure
trainer. You can learn to fly standard procedures (even with ATC these
days), learn to program an FMC, to learn where to look at to keep the
plane under control, but the feeling of flight cannot be learned.
Flying a PC simulation too often indeed tends to teach a couple of bad
habits that are hard to train away again (looking a the instruments
too often is one of them).


There are a number of pretty realistic combat flight simulators out
there that simulate aerial combat. If the game engine is good,
real-world combat tactics need to be flown in these games to win a
dogfight. It might be interesting to compare such a game to real-world
dogfighting.

Bye
Andreas