Thankfully, the sim provides numerous scalars and tables to adjust in multiple ways the base geometry defined aero results, and more keep getting added. And the wind tunnel already iterates today to get you to your Cl and Cd targets specified.
I’m hearing what you feel is a limitation, but what I’m not hearing is a counter-proposal. If this way of working is, in your estimation, untenable, then I’m certainly interesting in what your design would be that would avoid these limitations. It would certainly be instructive for me to understand. Is the idea here to go back to the single point weight integrated table based approach? The legacy flight model can already do that, so it’s possible I’m a bit lost as to what is being proposed instead.
If you have some concrete algorithms that you feel would be more appropriate for the solution you are talking about, those are the kinds of things we’d definitely want to talk about, and are 100% all ears.
External Flight and Systems modelling, provided the necessary “hooks” are allowed, looks to me like the best future, at least for now, for flight dynamics in MFS, when the users look after close to rw performance data, within the “normal flight envelope”…
Since the first time I oppened the SDK documentation it was evident MFS was designed much more with X-Plane’s approach in mind than any other thing. Unfortunately, for the good and for the not so good, such approaches have their limitations, X-Plane has have them since ever, but with years of development it has somehow been able to offer an overall more sound outcome of it’s BeT-based approach, helped by good 3pd plugins that do their stuff outside of the simulation engine in the most advanced products available for the platform.
MFS still has a long way to go, and I hope they continue to search for the best way to give 3pds the same tools and ease of integration that X-Plane provides.
I believe we will see devs like A2A, PMDG, etc… do their stuff mostly outside of the core “MODERN FM”… Time will tell, and it’s good that they come to play - the sooner the better…
Another big mistake, not brought to this thread because it isn’t so directly connected to the OP is weather. I think closing weather injection / editing / tunning to 3pds is a mistake, and I am really not at all satisfied with what I experienced in MFS, quite on the contrary. The looks, well those can be great ( at times ), but the feels and the consistency leave a lot to be desired…
For me, what comes built in Flight Simulator since Launch is the key to all this, It’s unbelievable it’s not even mentioned on the SDK that the Legacy Page shows the Derivatives for the VirtualWindtunnel Iterations? And this isn’t suggested in any Tutorial as a way for FDE Devs to base the impact of all the attributes in the .cfg on their creations.
It’s the best thing in the Modern Flight Model! And the key point that can push this simulator ahead of all others Civil Sims IMO. If this is already there, ready for use, most of the hard work has been done already. Let me show you some areas of this Study for the USAF on transforming a C172 into a UAV vehicle.
Scripting allows the user to perform repeated testing for a certain set of conditions.
The scripts are written within a JSBSim file in XML format [4]. A series of conditional
statements are performed and when the conditional statement is evaluated to be
true, events take place. When an event is evaluated to true for a given condition,
parameters may take on a new numerical value.
JSBSim can be built using a variety of environments … This approach allows the model to be built directly within a XML file and run the simulation from the command terminal. JSBSim allows for two
methods to interface with the user. The preferred approach is performing simulations
through the software’s Graphical User Interface (GUI), which guides the user to input
the necessary data to run the simulation. The multiplatform GUI interfaces to the
trimming and linearization code. As a result, the GUI makes finding a trim condition
and creating a linear model simple, and helps the user with more intuitive feedback.
For example, the trim conditions for a specified velocity and altitude can be used to
find a nearby trim condition.
"Variation of Stability and Control Derivatives
The longitudinal and lateral motion will be analyzed separately by varying the
stability and control derivatives. The flying qualities will be characterized for the
longitudinal and lateral-directional response modes and compared to the evaluation of
a dedicated set of flight maneuvers. "
The data obtained was for an airspeed of 100 knots (168.78 ft/s) at 4,000 ft
altitude by varying each stability derivative individually and then determining the
flight characteristics such as the damping ratios and natural frequencies. It should be
noted that in general changing one stability derivative usually changes other stability
derivatives. In identifying the key stability derivatives and knowing what quantity
is most affected, the design of the aircraft problem is better understood. The main
purpose is to show the trend because the amount varied will differ for each aircraft.
Since the stability derivatives of an airplane change with aircraft configuration and
flight conditions, some derivatives were in the form of look-up tables. Adjustment
and modification of these stability derivatives allows the modeling of an aircraft in
an altered state.
A lot more can be seen on this document, it’s a very interesting read on how to evaluate flying qualities for an aircraft in a simulator environment. I recommend.
I think you can get the idea. With Scripting, should be easily implemented into the SDK by Asobo I suppose, since it already do some of that when you load an aircraft, you can make available a tool that enables Developers to set the Flight Conditions, run the calculations in real time, and give back all sort of useful informations, trimmed state, basically what JSBSim and freewares such as XFLR5 (xflr5) can already do.
This really opens all kinds of possibilites. The first one that comes to my mind is a script that makes the current Airplane fly through all the Dynamics Mode of Motion. Longitudinal, Directional and Lateral. So, Short-Period, Plughoid, Dutch Roll, etc. This would need no flying by the user, all done internally, and the Output would be the Natural Frequency for the Modes, Damping Ratio, etc. Tweaking an airplane with a tool like that would be a dream.
You can also reverse engineer the process, and instead of the user inputting the values and getting the results, you can set the results you expect, and the Simulator can then run until it finds a close enough solution. Just like the SDK suggests an Empty MOI for the user based on the Weight and Balance data, it can suggests .cfg values, way more detailed then a user would ever be able to enter, based on the WindTunnel Calculations.
This is just a few ideas that came into my mind. I am by no mean an enginner, mathematician or Programmer, but all my reading about Aerodynamics showed some pretty interesting ways to estimate Flying Characteristics in a simulator, even without ever putting your feet on a real plane.
External flight models. Hands down the best solution, fulfils everyone’s wishlist, provides the most flexibility, best solution for the community, 0 downsides.
I think there were two very simple mistakes from the off, one forced the other not.
Two seperate flight models, Modern and Legacy. It causes problems, it causes confusion and I actually don’t know of anyone that uses Legacy. I’m not saying nobody does, just that I’ve never seen anybody use it and the only reference I’ve seen to it is when the sim has reset to it and people are getting problems because of it - “Check your flight model is set to Modern” has to be the single most-used phrase in customer support these days.
The need to ‘not use’ the old FS style flight model. To have something different as the default. It’s almost like creating something entirely new just because an element of the established sim crowd will never accept that the old can give good results if treated correctly. That element will defend ‘their’ sim to the hilt, be it X-Plane, FS9, FSX, P3D. And this is the one that I feel was forced, that led to a massive departure from what went before in the Microsoft Flight Simulator franchise.
So you asked for a counter-proposal, here’s mine for what it’s worth:
Get rid of two seperate flight models.
Go back to basics - the sim has already cornered the market on scenery, weather etc. The one area it is falling down on is the flight model, as evidenced by so many threads about different aspects of it. So go back to the old-style flight model FOR NOW. Not as a long-term solution, but to give the sim a base point, stability that is understood by the various developers (those who are currently ‘in and struggling’, perhaps also those who are ‘outside and waiting’). It will enable people to simply get aircraft built, working within a framework they understand to work with certain limitations. Then, over time (and the ‘ten year project’ keeps being referenced), take from the current ‘Modern’ and add it in bit by bit, possibly adapting to suit the solid base that aerodynamic calculations provide and gradually moving those single point calculations into a much wider, geometry based calculation.
I think the ‘forced mistake’ has thrown the baby out with the bathwater. There was a kernel, admittedly limited in certain respects, but there was also a developed SDK, it could all have been fed and grown and led to a situation where we truly had the best of both worlds.
Except for those freeware developers that do not have the time to learn that. At the moment, the sim is very ‘inclusive’ - by keeping flight models within it, it makes it much more open to everyone to build for it.
I’d agree that it would be useful to allow external FMs for those such as A2A, PMDG that may wish to invest the time, money and effort but do not agree that that is a solution to the problem being discussed here.
Agreed. I mean without allowing external flight models, I’m not understanding how extreme realism to the extent of FSlabs, PMDG etc can be achieved. And MS/Asobo should tell both us normal customers and the third parties that they they would indeed add extreme level of realism to the default flight model. This is important for consumer confidence.
For example, here’s what I got with my FSLabs A320 Update v2.0.2.347:
So, in brief, when I’m flying with some passengers and crew in FSLabs A320, the flight model calculates how they are going to the toilet and dumping their bodily waste, which in turn will have an effect on the centre of gravity and therefore the flight model.
Will Asobo add this particular feature (along with the rest of the features found in FSLabs A320) to the default MSFS 2020 flight model so that both FSLabs as well as anyone else can use this feature when building their Airbus aircraft for MSFS 2020?
I don’t know why everyone seems to assume that you can either have a default flight model or an external flight model, and no other option. Every single simulator up to now has supported both. Despite the incredibly realistic custom flight models of developers like A2A, Carenado have made quite a name for themselves by using the default flight model in FS9, FSX, P3D.
No one said to throw away the default flight model and force everyone to program their own from scratch.
I would argue the opposite. At the moment the sim does not include those that want a better flight model simulation. Or those that want slats. Or swept back wings. By allowing developers to create their custom flight model, the sim will truly be inclusive of all. Everyone gets what they want.
Try quoting my second paragraph - I think you’ll find I actually agree with you but just don’t see it as a solution to the problem being discussed here
Hmm. Chase cutting, Inciteful and thought provoking. Some big hitting going on here!
As a thread, right up there as a thoroughly entertaining - yes, entertaining - read.
So what then do I think I know now that I didn’t know before?
Well, that when it comes to the seemingly insurmountable problem of ‘delivering’ realistic, real time modelling dynamics, no amount of algorithmic cunning - in and of itself - is going to get
the job done. And that perhaps orders of magnitude of greater number crunching throughput will
ultimately be the ‘hammer’ needed to crack this particular ‘nut’!
This is bad news people!
I have already waited the 25 some years it took for the arrival of hardware with enough ‘grunt’ to allow the programmers art to deliver up a reasonable fidelity in visual yumminess. *
Now, I find, that I may well have to wait another 25 some years for realistic flight dynamics!
Ah well, time then to set the 25 year timer and hunker down for the long haul.
Come on MS/Asobo just iron out all those pesky teething troubles and we’re about there!
I don’t think that’s necessarily true… Plenty of very accurate airplanes are available for P3D, and they don’t need a supercomputer to run. True, we have yet to see any of them in FS2020 in order to see actual performance, but i wouldn’t worry too much.
And what you say re P3d is perhaps true [I ‘flew’ in P3d exclusively for many years myself].
But the comparative visual splendour of MSFS 2020 and the hassle of having to do the Orbx, et al, two step has - for me at least - ‘queered the pitch’ such that - and I have tried - I cannot ever go back to P3d.
Well, i didn’t mean that you should go back to P3D. I meant that the addons available for P3D are a good indication that when they do make it into FS2020, they probably won’t be a performance hog.
Gotcha.
But please don’t take me for some kind of misanthrope when I say that I have always very much eschewed addons.
I yearn for properly integrated simulator software.
Thus avoiding the dreaded, aforementioned, two step!
With due respect I highly disagree. I don’t believe that’s a solution. I also don’t see this happening so there is that. What I think is that they should just get completely rid of the legacy flight model as it serves no purpose now. No one is using it anymore and none of the new addons support it.
I have full trust in Seb and his team that they will be able to come up with a better solution for the modern flight model. Just got to give them time. Sim is an year old.
Getting rid of the legacy flight model we both agree on. It makes no sense to have two flight models.
A better solution for the modern flight model? I wouldn’t start from here! Start with aerodynamic calculations - hell, before release there was an interview in which they stated that the base flight model from FSX was fundamentally sound and they hoped to take that and improve on it. That is all I am suggesting they do, but do it in a more structured way. If that means the pain of starting ‘from scratch’, given that we are (as you say) only a year into this sim I think it would genuinely be much better in the long term for every single end user, developer and for the team themselves.
Just watch a few minutes from this point (it should start at 2:42). The talk is about improvement from FSX, that the basic calculations in that sim were correct but simplistic.
That is not what the current Modern flight model provides, in my opinion, and that is a shame - I believe the original aim of the developers was sound, but that the implimentation has failed.
