A NZ r/c simulator
first released Dec. 2003, minor update 1.1 Sep. 04
last page update Sep. 04
At last, I have ready for public consumption my most complete attempt at flight simulation (of radio controlled model aircraft), set in a New Zealand style rural area. You can fly in the aircraft , fire missiles, drop bombs and top-dress, but to be honest, the program probably doesn't have wide appeal. I just hope the odd flight enthusiast or perhaps DB programmer will find it as intriguing as I do.
There are various model types, from gliders to jets, to aerobatic to electric powered (as
I used to fly mostly). They fly pretty much like the real thing, ( though a bit too 'well
behaved' and predictable ) complete with gliding/speed pitching and stalling, and use
paramenters for wing loading, centre of gravity position , power , drag, control reaction
.. etc.
They can take off and land properly, with wheels touching down -- 'cartwheeling' or
damage may be added later for fun.
Cursor flight control has been added, but a joystick is vital for the full range of control. When flying ' in the model' you will notice how the smaller ones are quite (realistically) sharp in response to controls. There is a 'real size' light aircraft to give a comparison.
This program is NOT PERFECT ( see below), but I spent so long on some things that I have used up all my enthusiasm for improvement for the time being. In fact, I see that the oldest basic version is dated nearly a year ago. Somewhere I wasted a lot of time and effort.
Here's the link to download the zip file of the program " A New Zealand R/C model
flight simulator."
( Remember again, you need at least Direct X 7 installed.) ( file is 2 megs.) > 
It is now slightly updated to version 1.1, with the rudder action made more obvious and
rudder turn on ground related somewhat to propeller slipstream.
ORIGINAL NOTES .....
Some immediate quick changes were made
Evidently some users were getting unpleasant looking ' Z buffer' effects ... ie 'distance
from camera' calculations that produced wrong visual interaction within objects and in
relation to ground. I have increased the camera view close distance parameter as
suggested, but cannot test whether it has made sufficient difference since the program
always appeared correct on my machine. It can only be taken so far at the moment,
because, as a side effect, the 'in aircraft' view, when on the ground in a small plane,
loses a portion of the ground texture in front of the plane.
At the same time I added some ( indetectable! ) drag for aileron input.
And drag for aircraft yaw. I had gone to some trouble to get a 'yaw' independent of
a delayed turn/ roll reaction, then forgot to add the consequent drag. The sad thing is
you can hardly detect the yaw action, let alone the drag. I can't say either that I noticed
it on my own real flying models .. but well .. it should be there.
As per the note added above, a new version 1.1 now has more obvious rudder action --
added largely because other RC sims seem to show it. I still don't remember any such
obvious effect myself ! I suppose where there is proportional control input this does
not arise, but I have only a simple joystick -- like I imagine, most users.
Some users have worried about the apparent lack of increased drag in a turn. I'm not
sure what turn is meant. It is only when you alter the incidence of the wing with
elevator that drag increases. Otherwise the wing continues to produce the same lift, at
the same drag, even when banked. In that state, without elevator, the lift is diverted
sideways, and the craft loses height . It is true that a fast moving aircaft is 'harder' to
turn, because of the kinetic energy, but the control surface response is also higher. The
two tend to cancel out to the eye, except that there will be a greater angle of attack on the
wing to retain the similar curve of flight path.
Though it doesn't look it, the actual increase in drag in this program is around double
when given full elevator. I once spent many hours with a small home built wind tunnel
( a truly horrrendous noise from two overloaded vacuum cleaner notors driving large
propellers in reverse directions) and the program level is actually slightly higher than the
reality. Though there is room for improvement in my 'incidence effect' and low speed
flight model .. as below.
Extra .. that wind tunnel was fun. (Apart from the noise). It could produce only 43 mph
max, but that was about the flying speed of my electric aircraft. It was used to compare
wing sections and lift and drag at different speeds and angles of attack, and motor
thrust with different propellers as well. You had to measure the amps, volts and rpm,
and ounces thrust over a short period to match up the results. Then there was the
drawing of graphs to dream over. Still .. one learned some interesting things. ( Not
least , that if somone in the house turned on a heater or oven, the speed of the air in the
tunnel changed and ruined all the readings. )
It was quite fascinating to see that the theories in books are ( largely ) correct and see for
yourself ( with threads) where turbulence occurs, and note that some supposedly
inefficient wing sections can under the right conditions be the most efficient.
----------------------------------
Limitations and notes of interest to DB programmers :-
1. Not all models that should have, have proper wheels .. it doesn't worry me for now.
The models are merely what I could use quickly. Yes it would be great to have superbly
realistic models , and more typical model designs .. but ..... I still dislike the way
modelling programs work . A lot !
I suppose I could borrow some direct X models from a recently discovered radio
control flying sim by another person ( which actually work with dark basic!!!) , but this
would probably not go down very well ! -- and the download would become too large
for my web site ... anyone offering a free large site ????
2. The landscape is a bit ... sad . Partly this is because improvement would need large
amounts of graphics data in a download, and would demand much from a user's
computer as well. Partly it is due to limitations in dark basic itself. Even my
replacement 1.3 ghz computer would not handle a matrix landscape of sufficent detail
to be smoothly shaped and convincing. The only solution I can see now is to have two
sets of multiple matrices ( one set low detail, the other high detail and smooth) and
have the smooth matrices in the close view area, with the distant area left as low detail
matrices. Assuming this would work, I still don't have currently, enough enthusiasm
left to produce the necessary 'landscape designer.'
I will point out however, that dark basic has the very useful ability to set the matrix
shading separately from the actual ground shape, and hence one can get some effect of
smoothing ... and pleasing "cast shadow" effects.
( Ideally a radical new terrain system is needed in dark basic, creating from source data,
a tiled landscape of graded detail . There would likely be too much data to hold in
graphic memory all the time, so it would have to very quickly 'call up' data and tile
image for only the area in view. Something like this must be used for true flight sims.
And it must have shading which can be controlled. )
3. The landscape objects turned out to be tricky. And they are not interactive.
You can't hit the trees, or blow up the innocent 'sheep' or cloned spectator/ friends.
Nor can you land on the road, as I expect everyone will try !! You will land on the
matrix surface beneath .. not always exactly either. My quick landscape designer was
intended to enable aligning roads to the shape of the land ,.. but it was too hard. I
should have made roads totally horizontal / fixed height and adjusted the land. Obvious
... afterwards .
The tree objects are simple textured plains, which seemed to give a better look than
speed absorbing true 3d objects. They do not face to the camera when view is
airborne for reasons I have forgotten. Lighting I think.
There is only one fence, whereas NZ. landscapes are usually filled with wire fences.
Yes .. the landscape is well short of the objects needed to be fully realistic but I was
doubtful about a large download , or filling memory on users computers. And in the
desire to be finished quickly, I was reluctant to face the bother of adding a system to
select and show only objects ahead in close view.
4. The skybox isn't ! It is simply a cone. This was a quick fix, to avoid the usual DB
problem of visual interaction of sky and matrix landscape and to enable the use of
fogging to get the fading of the sky .. when I started the program it was in 16 bit colour,
and I couldn't get a smooth sky graduation any other way. Also the bitmaps for a
GOOD sky box were thought too large. This may have been a misjudgement. ( one of
many !)
Also, as a result, the clouds were added as a seperate flat matrix layer .
5. The wing tip trail objects do not alway line up perfectly. Any DB programmer who
knows the rotation limitations of DB is welcome to try and do better !! One might add
that the orienting of smoke and bombs and top- dressing dust to the plane and the
camera, so that they acted as required, took some cunning , and once, lengthy
programming. A curse upon DB's two incompatible rotation systems.
Nor do they fade smoothly out of view, simply because, on my computer at least,
fading one object does odd things to the light on every other object .. and the matrices
!! Grrrr
6. My 'home grown' simulation method is not perfect. I would suggest to anyone
wanting to do a simulator that they follow the earlier suggested links to information and
work from there. I will possibly one day correct the things that worry me in my own
system, but as they say, "don't hold your breath! "
( Chiefly I want to ensure the accuracy of vertical position/response to speed and wing
angle and, I suppose ideally, work in 'non linear', individual aircraft responses ... ha ,
and ha . )
7. If I had used only a 'pilot on runway' view, I could have had a very complete looking, smaller landscape with detailed bitmaps and skybox. I hope that people find the extra airborne views worth the loss of that option.
8. Originally I tried placing realistic semi 3D grass near the runway and it would be good to do at some point, IF it could reach a very high standard. It seemed simple at first, but I was just not getting satisfactory results after a lot of time .. and so I just expanded the runway till it was not needed.
9. On appropriate aircraft it would be interesting to land with flaps down and power on, and this may be added. And some slight correction made to incidence when near stall. .. sigh ...
10. The flight character of the planes is not calculated from their shape but comes from
an associated list ( a biplane shape can fly like a jet ) . At some point a "flight test/edit
and save parameters function " would let users add their own aircraft model .
Perhaps in the future .. ... and perhaps not. I tend to move on to new things. I will
probably get so annoyed with the cone sky and landscape I will redo the whole program
.
And I have just realized a way I can 'de-convolute' my flight code with a simple
'dummy ' object. Isn't it wonderful how such ideas strike after you have done so much
work ?
Link back to dark basic page