Flight Dynamics

[jcomm]

With the graphics already provided by DCS World, and what is comming along with EDGE, and the superlative flight dynamics and overall physics model, FSX/P3D and X-Plane flight dynamics models can't compete with DCS... It's a different league....

 

But this doesn't mean FSX's FDM is crap! It's good in as far as it can go, and still very acceptable for basic flight training for instance.

 

I believe PMDG uses FSX FDM, only A2A making some external calculations for some properties, but still also using the native FDM mostly. At least this is evident when we try to do with their aircraft what FSX can't be forced to model.(*)

 

I once though about the possibility of using FSX / P3D for the visuals of DCS, if some company considered that possibility. This would bring to DCS full World coverage, but then, with the upcoming superb graphics provided by EDGE, I think I don't mind flying in restricted areas....

 

Unfortunately, DCS doesn't model civil aircraft :-/  As a result, if I had to choose only one between P3D v2, FSX and X-Plane 10, I would probably go for P3D v2, and I could use it also as a visual display for yet another simulator - PSX - with it's fully independent and high quality specific FDM for variants of the Boeing 744.

 

(*) Please note I am referring to the FDM *** only *** and not all of the complex systems that are upstream from it, like the complex FBW used for instance in the PMDG 777 or the detailled prop engine modeling of A2A aircraft.

[vololiberista]

I have a really, really expensive book on flight dynamics and it's really a very interesting read (if you love absurdly complex math!).

I have Kuchemann's book. I've had it for years and now see on Amazon it's about £50

[WarpD]

"Aerodynamics and Flight Mechanics" by Barnes W. McCormick. Still around $150 for a new copy.

[ZG862]

Put it this way, DCS actually allows the dev to do what they want. The AFM, SFM, and PFM are only existing models that can be used. The EFM opens the way for other options as well. Technically one could build a model based on the FSX airfile system and plug it in.

 

Thanks for the response Jonathan, though what you say above seems rather at odds with your 1st assertion that DCS requires the developer to develop their own external FDM. I read it that (as you say above) the EFM allows the developer to do so if they choose as an alternative to providing parameters used by the AFM/SFM etc.

 

 

And no, I cant or rather, wont go into the tech details because they are in use where I work. Company secrets at this time. Suffice to say though, such methods of getting around the sim limits as you describe are rather...amateur. Creative yes, and for those who do not have the tech at their disposal to do better its not a bad idea. But there are more effective methods.

 

Fair enough - but in saying "getting around" you do seem to imply the creation of some form of external FDM. For example, I can imagine that if the aircraft designer wanted to add something that simulated the behaviour of a "hyperspace" control (remember "Asteroids"?), they could presumably instruct the aircraft to move to a particular point in space when it was activated, then let the native FDM get on with it from there. For the time that the "hyperspace" code is executing, surely this constitutes an external FDM?

 

Maybe I'm missing your point but it seems to be quite binary to me: either you work within the confines of the FDM (with anything from simplistic to highly sophisticated "fudges" of its variables) or you step outside of it and move the aircraft yourself.

 

Not trying to get the recipe for secret sauce here, just (like the OP) keen to understand what FDE constraints are being imposed and the general approaches that might be used to overcome them in future.

 

All the best  ,

 

Z

[JB3DG]

Last I recall DCS only allowed 3rd party devs access to the SFM or the EFM inputs. This may have changed though I have no way of knowing for certain as I am no longer active in the DCS community so don't take my word alone for it.

[ZG862]

I read something about an aircraft being ported from the SFM to AFM and this was a big task. I understand that there is no public API to the AFM.

 

But all this is going away from the OP's question: What aspects of the FSX/ESP/P3D FDE could be improved?

 

My candidates would be:

1) (in flight) vectored thrust

2) Reaction control systems

3) Asymmetric stalling

 

I guess the lack of explicit biplane support goes to the core of the FDE too. I realise that the FEA X-Plane model is fundamentally different but I wonder if the principle of a user defined number of wings ("Wing.n") would sit well with P3D?

 

Z

PS/ Wouldn't it be handy if you delegate AI generation out to other cores/processors?

[WarpD]

Vectored thrust exists in Prepar3D it's just not a default FDE behavior.

Reaction control systems would require an external FDE process, but Prepar3D does support that.

 

I'm always amazed at how people want unusual stall characteristics and other 'outside-the-envelope' behaviors... but few realize that most of that behavior is unknown for a great many aircraft. What a computer does and what the real aircraft does are not ever going to be exactly the same. Theory is just that, theory. It is never going to be an absolute. Multi-million dollar simulators don't do any better than FSX/Prepar3D in this department and in fact a great deal of them simply stop the simulation with a 'crash' if the sim's calculations ever step outside of normal flight.

[JB3DG]

For tubes its not such a big deal. For GA and aerobatics it is to some extent when a person wants to practice unusual attitudes and/or spin entry and recovery. For fighter aircraft it is very important because you do all sorts of things outside of a tubeliner's envelope.

[jcomm]

I'm always amazed at how people want unusual stall characteristics and other 'outside-the-envelope' behaviors...

 

Not necessarily that fancy, but rather simple things like what's happening in the video linked bellow not being possible... ( notice that the climbing turns are performed with the engine quited ... )

 

 

I've replicated it with the best prop add-ons for FSX... 

 

And many other aspects of simple, basic, GA prop aircraft, not to even mention rotary wing...

[ZG862]

Thanks for the info re VTOL. Looks like I need to do more research.

 

As i'm sure you know, in real world GA, one thing fairly likely to kill you is a spin in the circuit; any improvement in sim flight dynamics that injects some much needed realism would be very welcome in my book.

 

We keep straying off the topic though. Is the FDE really all it could be?

 

Z

[JB3DG]

Thanks for the info re VTOL. Looks like I need to do more research. As i'm sure you know, in real world GA, one thing fairly likely to kill you is a spin in the circuit; any improvement in sim flight dynamics that injects some much needed realism would be very welcome in my book. We keep straying off the topic though. Is the FDE really all it could be? Z

 

Bernt Stolle's FDEs spin and stall quite accurately. And that's with the default FSX system. RealAir does the same if I recall correct.

[ZG862]

So Carenado are bypassing the ms engine? Hmmm. Rather supports Erich's initial comments.

 

Z

[JB3DG]

So Carenado are bypassing the ms engine? Hmmm. Rather supports Erich's initial comments.

 

Z

 

Nope. As I said its with the default FSX system. Bernt Stolle just does a darned good job in making it work properly.

[ZG862]

Nope. As I said its with the default FSX system. Bernt Stolle just does a darned good job in making it work properly.

 

Ah. I'm misunderstanding your terminology then. I thought the flight dynamics engine is the piece of code that implements an algorithm to simulate flight. I hadn't considered the .air file and aircraft.cfg to be part of the engine and had based my comments on that preconception. Microsoft define these files respectively as follows:

 

“AIR files are used to provide flight dynamics data, in the form of coefficients and data tables, that determine the flying qualities of an aircraft.”

 

“The aircraft.cfg file specifies the versions of the aircraft included in the aircraft container, as well as the attributes (name, color, sound, panels, gauges, and so on) for each aircraft and where to find the files that define those attributes.”

 

You seem to be suggesting that  Carenado are simply providing 'better' values to the engine (which is great and I applaud their diligence) but this approach can't possibly work if the algorithm takes no account of a particular factor - e.g. dynamic thrust vectoring.

 

(This principle is what is illustrated in Jcomm's video too btw - the aircraft is defying physics either because the parameters are incorrect or as is being implied, the algorithm is flawed.)

 

Cheers,

 

Z

[Gregg_Seipp]

 

 

RealAir does the same if I recall correct.

 

The RealAir C172 FDE spins after a stall very, very nicely.

 

 

 

I'm always amazed at how people want unusual stall characteristics and other 'outside-the-envelope' behaviors... but few realize that most of that behavior is unknown for a great many aircraft. What a computer does and what the real aircraft does are not ever going to be exactly the same. Theory is just that, theory. It is never going to be an absolute. Multi-million dollar simulators don't do any better than FSX/Prepar3D in this department and in fact a great deal of them simply stop the simulation with a 'crash' if the sim's calculations ever step outside of normal flight.

 

Pretty true.  I flew C152s for my flight training and there was one that would always want to wing-over on a power-on stall.  The others were pretty docile.  I remember reading a thread on the C210 that said that every one of them was different.  People also don't take into account that, in the real world, gauges can be off, read high or low.  Things are more pristine in our FSX world than they are in the real world.

 

Gregg