Free Falcon 5.5

[A.S]

@yoda:

 

how much lift you can create on a 90deg banked f-16 (watch the shape of that craft) with altering your AoA and pitching up with rudders?

 

next:

try this...take the F-15 (2.0) fly 550kts . full fuel..no payload...bank 90degs (knife edge pass) ...start at 1500ft ...now hit full the rudders and make a full loop in 90 degs bank finishing it ....realistic? :D

 

(if something is well made, i compliment that !!! if there are flaws, i say it too, i have no contract "to be nice") ;)

Edited May 12, 2010 by A.S

[Case]

To get back on topic...

 

When only tables are used to represent the motion (spring or spring 4.0 :D) it is evident that rules of physics are violated. That is way this argument shouldn't be brought up. I think the expression flight on rails sums it really nice.

 

I can't really see how tables representing the motion can be used at all, as there is the variable user input to deal with. To me the "flying on rails" expression seems more likely due to simplifications of the flight model, where (parts of) physical processes are left out for the sake of computational ease. I understand the need for tables describing drag and left as a function of Mach and AoA, but the fact that, for example, a knife edge pass isn't modelled correctly is not due to the fact that the table is wrong, it is due to the fact that that physical process, i.e. the dependency of drag and lifts on roll, wasn't part of the table in the first place.

[A.S]

your opinion ...fine

my opinion: you havent flown Falcon at all, so i have difficulties listening to that.

 

in respect.

[RvEYoda]

@yoda:

 

how much lift you can create on a 90deg banked f-16 (watch the shape of that craft) with altering your AoA and pitching up with rudders?

 

next:

try this...take the F-15 (2.0) fly 550kts . full fuel..no payload...bank 90degs (knife edge pass) ...start at 1500ft ...now hit full the rudders and make a full loop in 90 degs bank finishing it ....realistic? :D

 

(if something is well made, i compliment that !!! if there are flaws, i say it too, i have no contract "to be nice") ;)

 

we are not discussing particular parts of either implementation. (They both have

strengths and weakneses, depending on what you prefer one has more than the other.)

We are discussing the princple of 2-variable interpolated tables vs dynamic FM.

 

(For F-16 btw at least enough to perform straight flight because it has been done in real airshows)

[GGTharos]

Pointing out a possibly valid flaw in one FM is still not proof that one FM is inferior to another. This is a poor way of doing an actual FM comparison.

 

To compare FM's you need not only the resulting behavior, and data, but you also need to have some understanding of what is done with this data. There are algorithms out there that model certain processes that go in in the real world that do better with less data to start with than with more in some cases; it all depends on what you model and how you model it. If your model is not so good, throwing more data at it will only increase accuracy within that model's limits, but it won't really make the model better.

So what if you can split up your FM or 0.01 Mach segments when all that does is take away some (probably reasonable) interpolation that you had with 0.1 Mach segments?

 

You're not adding new fuel flow calculations, you're not adding new AoA calculations, you're not really adding much of anything. You might end up smoothing out the edges of the envelope, and calculate things more accurately, but your model isn't much more realistic than it was before. It just has more numbers.

 

@yoda:

 

how much lift you can create on a 90deg banked f-16 (watch the shape of that craft) with altering your AoA and pitching up with rudders?

 

next:

try this...take the F-15 (2.0) fly 550kts . full fuel..no payload...bank 90degs (knife edge pass) ...start at 1500ft ...now hit full the rudders and make a full loop in 90 degs bank finishing it ....realistic? :D

 

(if something is well made, i compliment that !!! if there are flaws, i say it too, i have no contract "to be nice") ;)

Edited May 12, 2010 by GGTharos

[GGTharos]

I don't know, somehow I did not pick this up from what you said, but maybe I didn't follow enough of the conversation.

 

Absolutly....that is infact exaclty my point too :D

[RvEYoda]

show me :D ( yoda we know eachother well, and i can understand this perspective...but show me...and we will see why this is not fully right)

 

I'm not going to go search for you - I am@work, you can search for it well yourself.

You can also wait for BMS5 with dynamic flight model to be released

which will have the same conclusion (you can watch its early flight model on youtube)

[asparagin]

I can't really see how tables representing the motion can be used at all, as there is the variable user input to deal with.

 

They have this points and use them for interpolation. The points are the rails. (Or at least it is how I see it). And yes, I suppose that the table data is correct, but it's the black spots that worries everyone.

[Pilotasso]

Yup, not possible to hold altitude in knife passes. This has been said by pilots (its not me ) :D

[Pilotasso]

Yup, for the F-16, if it wasnt for the tail would be pretty aerodynamic sideways :lol:

 

As for other aircraft it depends as their fuselage provide different areas when rolled 90ºs off but I realy dont know.

[RvEFuSiOn]

Guys there seems to still be the old Nose bobbing issue with the F-15C.

 

eg. If you let go of the stick during a turn, it immediately unloads pitch and aoa too fast. Im wondering if this small issue can be fixed? The current rate at which the aircraft tries to correct itself is too fast, which makes for an unstable turn. There is far too much jerking. It seems abit too sensitive.

 

I wondering if this can somehow be tweaked by reducing the rate at which the nose of the aircraft goes in to some sort of negative pitch.

Yo-Yo i would love to hear your comments on this, Thanks!

 

Peace

Edited May 12, 2010 by =RvE=FuSiOn

[RvEYoda]

Guys there seems to still be the old Nose bobbing issue with the F-15C.

 

eg. If you let go of the stick during a turn, it immediately unloads pitch and aoa too fast. Im wondering if this small issue can be fixed? The current rate at which the aircraft tries to correct itself is too fast, which makes for an unstable turn. There is far too much jerking. It seems abit too sensitive.

 

I wondering if this can somehow be tweaked by reducing the rate at which the nose of the aircraft goes in to some sort of negative pitch.

Yo-Yo i would love to hear your comments on this, Thanks!

 

Peace

 

Yeah this issue is something that would be good to see attended, but I think it is not possible

with the current SFM, so definitely fixed in DCS. I think it is because in Fc the auto-trim of the

F-15 adapts the flight path instead of pitch, plus some flanker FM code heritage.

[SUBS17]

I do understant, just step onto their forums and say that, they'll swear by it TODAY.

They dont realize or ignore by bias that interpolating flight model from points of constant motion (snapshots of constant motion) into a dynamic motion modeling violates a fundamental law of classic mechanics: F=Mxa. :)

 

Thus why there is a big difference betwee data tables and a dynamic flight model. Its not just because the later is more complicated. ;)

 

What forums? For its day Falcons FM was good compared to some of the other sims released at that time. I'm curious about BMS 5s FM as the videos look quite good. From my own observations F4AFs FM is actually quite good to fly eventhough it contains less detail than other sims at least it doesn't suffer some of the bugs that I have seen in sims in the past. F4s FBW and damage modelling was much more complex than previous sims at the time it was released I doubt there were PCs capable of running anything more complex at the time as F4 pushed PCs to the limit back when it was released. I'm not convinced about your rails theory as F4AFs FM changed significantly when the HFFMs were added, there were definately some big changes with that patch not only in the FM but also damage model and ground handling in later patches. It'll be cool when someone makes a more detailed F-16 flight sim such as DCS or FO but until then F4AF and OF are the best F-16 sims whether you think they are on rails or not.:thumbup:

[SUBS17]

I'll be up for that a DCS F-16 would rock.

[jaeger_101]

Enjoy some released pictures of a real simulator. The Austrian Luftwaffe released pictures from the brand new Eurofighter simulator. One thing about this one can be said; ALL data are 100% correct.

Just compare the computers with the chair and imagine how our UBER mega game computer would compete. Imagine how many cpu´s and ram are inside ;). Take also a look at the F4 style graphics and guess what kind of graphics our cpu´s could handle with this simulator.

 

Pricetag: 10 million euro, only in package with Eurofighters and endless security checks.

 

I take our available desktop simulators as what they are; immersion and fun. Falcon and now FC2 do a great job in there respective area. Well done ED team.

 

Austrian Eurofighter Sim:

http://www.bmlv.gv.at/download_archiv/photos/flieger/galerie.php?id=1148&currRubrik=55

 

cheers

 

Jaeger

[sobek]

Just compare the computers with the chair and imagine how our UBER mega game computer would compete. Imagine how many cpu´s and ram are inside ;). Take also a look at the F4 style graphics and guess what kind of graphics our cpu´s could handle with this simulator.

 

That's a little besides the point, IMHO. Just because current harware can't handle the most advanced flight model there is ATM doesn't mean you should fundamentalistically stick with an old one. There's room for improvement in between the 'extremes'.

[Pilotasso]

Precisely and there's lot more going on in a simulator than just the physics. You will equaly need more resources simply to display more 3D objects.

[jaeger_101]

That there is room for improvement is true for almost anything and is news worthy like newtons law of gravity.

 

The vast majority of the consumer wants hot GFX and tons of eyecandy and a smaller part wants realism. Its clear that it is a cooler when you see the hardware of a real simulator and compare it with your good PC. :megalol:

 

cheers

 

Jaeger

[rattler]

One must remember a simulator regardless of the cost is still a simulator. When it comes to combat aircraft old or new there is always information not available to public. Combat simulators as well as civil a/c simulators are made under contract and info available to produce such sims. is and will always be top secret. What the public gets is what they (manufactures) are willing to give out. Our PC flight simulators do a great job to provide us with great features, entertainment and friendship, which is what they are meant for. I guess however that there will always be the, which one is better than the other threads. It still boils down to personal preference and enjoyment. Lets enjoy.:)

[Bearitall]

One must remember a simulator regardless of the cost is still a simulator. When it comes to combat aircraft old or new there is always information not available to public. Combat simulators as well as civil a/c simulators are made under contract and info available to produce such sims. is and will always be top secret. What the public gets is what they (manufactures) are willing to give out. Our PC flight simulators do a great job to provide us with great features, entertainment and friendship, which is what they are meant for. I guess however that there will always be the, which one is better than the other threads. It still boils down to personal preference and enjoyment. Lets enjoy.:)

 

Well Said Rattler

[asparagin]

I guess however that there will always be the, which one is better than the other threads. It still boils down to personal preference and enjoyment. Lets enjoy.:)

 

Yes i enjoy those threads also :D

[mav-jp]

no hard feeling , but it seems to me that you have misunderstood what the problem of falcon4 FM is.

 

Let me try to explain.

 

First thing to understand is that a fligth model (and generally any motion model) is

 

1) a model that computes forces and torques applied on the model

2) A routine that solves the 6 dof motion equations taking as input the forces and torques and giving as output the speeds, rot speeds, accel and rot accels.

 

In the specific case of Flight model, the forces and torques are generally dependant on motion itself (i.e Speed, AOA, BEta etc...) which make the differential equation of motions beeing of the second order (or more) and coupled.

 

Aditionnaly, the aero forces are very often dependant on several parameters, for instance , CL (adimensional coefficient for Lift) is dependant on AOA and Speed...(approximation).

 

From that you have two options:

 

1) you decide to have a linearisation of the forces and torques with some of the parameters , CL (mach,alpha) , CD (mach, alpha) etc...which means you have TABLES of coefficients.

 

2) you try to calculate the forces and torques by using complete resolution of Navier stokes equations (not applicable for simulation) or using simplified aero equations (Xplane for instance).

 

 

The solution (1) is in 99% of the case the one chosen in fligth simulation (including pro simulation)because it allows the best precision and is very fast. Aditionnaly in some extreme cases, all the coefficients static and dynamic are available wether from real wind tunnel testing, wether from complete NS calculation (not done in RL time).

 

SO when you have a complete desciption of the forces and torques, then you can solve the equations of motions, which is not a big deal because it is hopefully simplified in coupled equations of the 1st order and can be solves by algo like Runge Kutta...

 

Now what is the problem of falcon4 ? well falcon4 has two major flaws :

 

1) The only aero data that are available for the model are CL/CD and thrust (partial CY as well) ..nothing about Cl,Cd, Cn (torque coefficients) ,and the dozens of other aero coefficients for specific cases (LEF/ TEF, Sb, etc...). Aditionnaly those aero coefficients in F4 are NOt dependant on the position of the controls and they are not dependant either on Beta. TO summarize: the problem of the data does not come from the fact they are tabular , it comes from the fact that they are not exhaustive and not dependant on vital parameters.

 

2) there is not resolution of the equations of motions, in fact the position of the GC is computed by Euler integration, but the rotation in the x,y,z axis are not computed but are generated from a FLCS code routine... It means that the position of the AC is somehow independant from the rotation motion. In fact , AC is calculated as if it was a dot and rotation is calculated by scripts routines corresponding more or less of what the FLCS is supposed to do.

 

As a conclusion:

 

* dont believe using tabular data is wrong, because it is the best thing to do for real time simulation. It is very bad when you have only 1000 values, but when you have 1 000 000 of them (that can come from Wind tunnel testing or NS computation), then it is much better than any approx aero Real Time computation

 

* the "FM on rail sensation" comes from those two flaws

 

* Of course, real equation of motion solver would be the first step to make falcon4 FM react as a FM....but that would demand hell of a work and probably a complete rewrite of the FM code and integration in F4 code, which would be IMHO very diffcult.

 

NOTA : As far as edge pass is concerned for the F16 , it is possible to hold it under very specific circonstances which is a very low fuel amount (<1000 lbs) . Indeed, in that case, most of the "lift" is not produced by the tail itself , but by the Engine thrust that is angled at around 8 / 10 degrees. If the speed is sufficient when entering the manoeuver, the combination of thrust/ low weight and "Lift" generated by tail + fuselage is sugfficient to allow the AC to hold it. In that specific case, the FLCS does not restrain the Yaw imput of the pilot because speed is >250 kt and AOA is low

Edited May 18, 2010 by mav-jp

[Oceandar]

For those who don`t know, Mav-JP was the man behind HI-FI FM for Falcon. Its used in FF5 and Open Falcon. Ask your doubt about Falcon FM to him :D. This should be interesting :)

 

IMHO FF5.5 is a result of FF team great effort but somehow I have loss alot of fun with this release especially after they restricted some ordinances in the campaign.

 

And with the locked DB, I think they should change the name to RF (Restricted Falcon).

Edited May 18, 2010 by Oceandar
change the word "removed" to "restricted"

[RvEYoda]

Great that you wanted to step in here Mav-jp. I have never seen the equations myself, but

made the assumption that the calculation of drag/lift coefficients could not be reasonably

simplified/approximated as two-variable functions (Your 2D tables).

 

Naturally you can make N-D tables and you end up with just as good results as with a solver,

but the problem then is, you have resolution 10.000 pts per dimension and 4

dimensions? You wont even have space on your harddrive for enough points.

 

In order to solve any such equation (assuming it requires 4 or more variables), you need to

solve it, in real time and not use table data.

 

Now assuming that some of these equations can be reasonable simplified to 2-variable

functions (again I have never seen these equations or studied aero/fluid dynamics, so I just

draw conclusions on the two different cases) then of course the tables work :P.

[Case]

Let me try to explain.

Thanks for this thoroug explanation mav-jp!

 

This is precisely what I was looking for when I made this post earlier in this thread.

 

Can you refer to a document describing the equations of motion?