[simpits-tech] May the Force Feedback be with you...

[Matt Bailey]

On Wednesday 19 February 2003 06:35, you wrote:
>   Hi Matt,
>
>
> First, let me apologize for so late answer. It happened to me before,
> but this time, I decided to answer back ;)  BTW, when do you sleep?

	Hmmm.....at night, and during class. :)

> The thread is over long time ago, but anyway... here I am
>
> >	I dunno exactly what math is required although from the math you are
> > doing it looks like you're thinking the same method I was, but you know
> > more of the math required. :) This was actually my second choice,
> > however......I was thinking a much simpler way would be to use airfoil
> > lift. Perhaps not as accurate as the previous method but much easier to
> > handle for a non-math dude like me. :)
>
> There's always more than one way of  solving the problem, and at the
> beginning, it not easy to say which is better.  Solution I proposed was
> simplified view of "clasicall solution" used to predict stick force
> while constructing new aircraftes. It involves some math during design
> stage, but the resulting equations are quite simple to implement in
> hardware or in software.

	I naturally am less enthusiastic about the idea since I am not very 
knowledgeable about the math. :) Sounds like it would be reasonably simple 
though for someone who knew what they were doing. Of course, the problem of 
how to implement a fast, powerful hardware system is still there.

> I would like to see more about your proposal,  is it possible?
>
> >	By dynamic pressure, are you referring to the pressure caused by the
> > plabe moving through the air, or the actual pressure on a control
> > surface?
>
> What I had in mind was the pressure that might be presented as
> 0.5*rho*V*V, where "rho" is air density at the altitude of flying, and V
> is airspeed. This pressure is commonly used to connect AD coefficients
> with actual forces. It is not actual pressure on the control surface.

	Sounds cool, I suppose this would be sufficient for most flight conditions? 
What about extreme conditions like high AoA? Something else I thought of is 
weight of control surface (only an issue with elevator, and maybe rudder in a 
few odd situations) but I don't think this would be a big deal, just add some 
down elevator pressure linearly (I think) as speed decreases.

> >	FWIW, X-Plane does not model trim tabs. I've heard that Fly has a really
> >good trim system but I don't know if it models the physical trim tab.
>
> AFAIK it does, but on the other hand I don't own it, so I looked at the
> documentation about the exporting data (in the internet connection
> part), and the position of the trim is listed as variable that can be
> exported.

	Yes X-Plane has a trim system, but it's not actually simulating a trim tab 
and its forces. I believe trim is achieved in X-Plane by simply moving the 
control surface center (like changing the actual linkage between control 
device and control surface). What I am not sure about is if it simply adjusts 
the center (so full control deflection would always result in the same 
control surface deflections, regardless of trim setting) or just always moves 
the control surface a certain amount regardless of control device position 
(for instance, increasing up elevator trim would actually provide a higher 
maximum up elevator travel and less down)

> >	I believe X-Plane already considers downwash of the wing on the
> > horizontal stabilizer, although I don't know how well (it's not doing
> > fluid dynamics, I know that).
>
> Yes it does (and should), and (after bit oc checking) exports it as well :)

	Cooooool. How much can you tell about how it's doing it? I haven't been able 
to run X-Plane for some time so I can't check it myself. I remember seeing 
the values but I never got into the actual calculations it was doing.


> If the nforce on the stick is calculated, why adding aditional input
> directly, why not simply use output of the simulation program?
>  Generally, everything should be calculated within simulation program,
> and results exported. All input needed for stick force calculation
> should be feeded from program (if anybody is interested in theory, I'll
> explain  this in more details).  Doing otherwise almost certanly leads
> into problems .

	I agree we should use as much data from the sim as possible, but X-Plane 
doesn't tell us everything we need. Mainly, control surface force. But, by 
taking control surface dimensions, airspeed, etc we can predict control 
surface force and use that to drive our hardware (as you have suggested in 
your math figures).
	In regard to trim hardware vs software (which I think is what you're 
replying to in this part) as I said before, X-Plane doesn't have a proper 
trim system so we need to do one seperately. My idea is to simply move the 
control device (no software needed) although a more sophisticated setup using 
a software simulation of the trim tab would be more accurate. The problem you 
get into here is that to really do it right, you need to feed some more data 
TO the sim (for extra drag etc from the trim tab).

> Additionally, even if everything goes well, after while, one would like
> to feel (for example) subtle vibrations, or buffeting while approching
> stall AoA, or (for buster comands) hydro failure, and switch to manual
> forces......  If all calculations are made within software, it is not so
> hard to add all these features, hardware remains absolutely the same.

	Well in the case of X-Plane you'd have to talk to its programmer about 
getting these kinds of aero effects into the sim. Since these effects are 
almost exclusively for FF, and X-Plane doesn't support FF, I doubt Austin 
would be enthusiastic about adding them. I agree this stuff is definitely for 
software, personally I'd prefer to just add this stuff to the seperate FF 
software.

> I have to say that I'm bit biased, I preffer software solution 'cause
> I'm software frick (maybe I should introduce myself?),  and think it is
> later much easier to recompile the program, compared to changing the
> hardware. I suppose that my opinion is not spread?

	Good point. I am just leaning towards the moveable control hardware for trim 
because it's so simple yet would match real life so well. As to stall buffet, 
hydraulic forces, etc, I agree this is for software.

>
> at the end ,
> once more I'd realy like to see your "lift connection proposal"...
>
> ciao,
>
> Gordan

	I was thinking that since lift is changed by deflecting a control surface, 
lift would be a natural solution to finding control surface force if you 
don't have actual control surface force coming from the sim. I'm not sure you 
would even need airspeed, since lift is a function of airspeed and AoA. The 
neat thing is the simplicity. If for instance the elevator is producing 
500lbs of lift, you always get a given amount of back pressure on the control 
device, regardless of whether that 500lbs is being achieved by deflecting the 
elevator a small amount at 120 knots, or full deflection at 40 knots, or 
tossing the plane through the air at high AoA while holding the elevator 
neutral. Also, AoA changes would change control device center. For instance, 
if the plane were at high AoA and the elevator were neutral (but at the high 
AoA), the elevator would be producing lift. There would be a tendency for the 
stick to come back.
	(This part applies to other systems as well.) With a little more math you 
could throw control surface weight into the mix. So if you were sitting on 
the ground with the engine off, the weight of the elevator (a fixed force, or 
linked to g force if you want to get real accurate) the elevator would be 
producing no lift and would naturally fall. If the plane started moving 
through the air, the downward deflected elevator would start producing lift. 
Now our lift-linked FF system starts moving the stick back. Force would keep 
increasing as speed (and therefore lift) increased, until the stick/elevator 
is almost neutral, then lift would taper off to whatever is necessary to 
support the weight of the elevator. There's your center position.
	If the plane were tilted in the air (change AoA) the horizontal stabilizer 
lift would change, and therefore force on the stick, until stabilizer AoA was 
once again parallel to airflow. Now you have the new center position.
	This way instead of using airspeed, control surface pitch moment, control 
surface dimensions, AoA, etc, you just need one single value from the sim 
(plus control surface moment, if you don't define a comparable value in your 
FF program). In a way this is perhaps a more efficient use of sim data. The 
sim is already taking control surface dimensions, airspeed, etc to find lift, 
we don't have to recalculate all these values externally........ *if* lift is 
indeed sufficient for determing control force. It would be pretty accurate as 
far as I can tell, but I could be wrong..............................

	-Matt Bailey