Jump to content
A_S

JOYSTICK CURVES ?!?

Recommended Posts

A-S,

 

I understand what you are saying. Unfortunately stability and control is a very complicated subject. I raised this issue in CloD and it was a disaster.

 

Even engineers have trouble with it and a gaming community discussion is just not productive.

 

Unfortunately, it was the one area that makes the most performance and handling difference in World War II era aircraft. Aircraft weight and speeds had outstripped our understanding of stability and control. They struggled to quantify certain behaviors but thanks to the efforts of some brilliant engineers and even more importantly, farsighted test pilots, standards were developed that post war became convention.

 

Amen to that.

 

Also other posts from others are nice and this thread is becoming more interesting now :)

Edited by A-S

Share this post


Link to post
Share on other sites

How do you find the Cirrus sidestick, HeavyCavalrySgt? It seems to have attracted some criticism:

 

From here: http://www.aviationlawmonitor.com/2010/12/articles/cirrus-aircraft/steve-wilson-the-cirrus-airplane-has-serious-problems/

 

If that is anything like correct, maybe we need a few more of Crump's WW2 test pilots and engineers...

Mr. Steve Wilson might just be a teeny bit biased against Cirrus, given that he's in the business of selling Cessna aircraft in Sioux Falls, SD. No doubt he would like to see the Cessna 400/Corvalis TT sell as well as the Cirrus line. Let's hope their recent wing delamination problems were an isolated production hiccup. It looks like a good airplane. If it starts selling enough, though, I fear it will begin to develop the same history of undertrained pilots killing themselves in it.

Steve Wilson's criticism is classic commercial competitor badmouthing, nothing more. This is old news. He has a vested interest in spreading "concern" about Cirrus.

The Cirrus was designed and certified under a set of stability and control standards. Those standards allow for a wide range of aircraft feel,personality and performance.

 

As far as safety goes the standards do a very good job of mitigating the inherent risk of flight.

Share this post


Link to post
Share on other sites

Amen to that.

 

Also other posts from others are nice and this thread is becoming more interesting now :)

If there is anything I can help with, feel free to PM me.

Share this post


Link to post
Share on other sites

If there is anything I can help with, feel free to PM me.

 

With 2GB of PDF files about Modeling and Simulation of Aerospace Vehicles, Flight Test Engineer Handbooks, Aerodynamic Fundamentals, Computational Modeling of Aircraft and Environement, Fundamentals in Fluid Dynamics, Performance and Stability of Aircraft, Pitch Dampening, tons of real performance charts and Handbooks, you name it (because the list of professional documentations i have flying around here is much bigger, its a hobby of mine), i am already "lost" in tons of material no "non-nerd" can process in a life-time - LOL - , but i do appreciate your help and openness. Flight physcis and relatives are just personal interests of mine accompanied by autodidactic studies and an educated share of know-how with real aviators from different areas. Being passionate about this and having shared much time with simulations, i claim to understand a little bit the fundamental implications of flaws in simulations, which can lead to huge differeneces in flight and combat expiriences compared to the real deal. "Feeling" flight and proving flight are two completly different things and this is why i asked for telemetry data in track in this thread: http://forum.il2sturmovik.com/topic/1695-telemetry-data-tracks/

Your seem to understand the problematic described, and YES it IS difficult even for FM engineers. This is why we haven´t seen

proper solutions before.

 

Historical Review:

 

Lock On or Flaming Cliffs 2: The surface deflections were almost instant, regardless of speed or flight regimes, thus a dramatic onset in G-forces with linear inputs. Even formation flight was almost inpossible with linear stick inputs and pilots were "forced" to use curves. The control mechanics were not simulated realistically therefore a "workaround" was invented with an aritificial G-limtter in order to by-pass the problem, meaning the true physical causes/forces were not simulated on the frame, but the end-effects were "chastised" with scripts (modeling the effect). With later generations (DCS A-10) they became more educated in modeling fix-winged airplanes i believe, but i never tried that one tbh.

 

IL-2: Similar problematic. Here the solution was to "slow down" the deflection speed of surfaces like rudders and elevators in order to bypass the same problematic (they move totally different in real). You can see that in comparision with the DCS Mustang or the BOS surface-behaviors (shown in video above) at slow and high speeds and to be honest i am VERY possitively suprised to see it more realistic in BoS, but i haven´t flown it yet, obviously.

 

 

Years later with better know-how and better hardware:

 

ROF: I can´t comment to the flight physcis at all, because WW1 was never my "thing". I have followed the developement externally though and with time passing and Andrew S. leaving EagleDynamics going to ROF and Oleg M. leaving 1C going to DCS (what a dance hah?) all i can say about it is that i really love the athmosperic modeling alot in ROF as "air" and "gravity" didn´t exist as such in other sims before.

 

BMS: My personal favorite at the moment. Flight model was developed by Mav-Jp over 6 years afaik and is my opinion a "master-piece" if it comes to modeling flight on a PC.

Not only is he open to questions and addresses them, but he INSISTS till today NOT TO introduce joystick curves and to be honest it makes absolutly sense, because in his flight-model you don´t need them, regardless what joystick or you have, may it be analog or pressure sensoric. You have the option of dead-zones and saturation (bands) though, but that´s it.

He introduced the Input-Output Module:

 

 

( It handles the link between the action of the pilot and the responses of the corresponding controlling surfaces. This module will be detailed in a later developer’s notes.)

 

I can not cover the extend of his work here as it would explode the limits and would be mostly off-topic, but his documentation is available publically, not to mention, that the flight-computer software used in the simulated F-16 is line-by-line the same code used in the real F-16. For the enthusiasts i recommend reading the rest here: http://www.benchmarksims.org/forum/content.php?121-Flight-Model-%28FM%29-Developer-s-Notes-Part-2

 

BOS: Hope never dies  :biggrin:  and we will see ....

Edited by A-S

Share this post


Link to post
Share on other sites

Answer is simple:

 

Default curves for all if all would have same quality of joystick and those equipment would never loose precision...

 

 

Instead of "Pimp my Plane" abilities curves options are a chance for players with smaller budget to enjoy the flight with medium class/old joysticks.

 

Till the curve option is available for all ( not guys with more money... ) everything is ok.

 

 

 

Example from real ???

You dont find in real 2 same planes of 1 type.

They are almost the same at the begin of duty but after some time they start to be different.

 

Stall speed, max speed, max AoA, etc...

 

Every plane have own soul, problems and need special treatment.

Same like with curves stuff...

Edited by =LG=Blakhart

Share this post


Link to post
Share on other sites

 

Example from real ???

 

You dont find in real 2 same planes of 1 type.

They are almost the same at the begin of duty but after some time they start to be different.

 

Stall speed, max speed, max AoA, etc...

 

Every plane have own soul, problems and need special treatment.

Same like with curves stuff..

 

Philosophically (and partially practically) your approach is reasonable and a "calming" point of view Balkhart and to some extend i even agree, but consider also, that simulating flight on a PC in nothing else than pure mathematics and pure physics from coding perspective.

In simulations everything is "perfect" from mathematical point of view at first (unlike in real) and deviations in order simulate "realistic" random variations for different aspects are applied afterwards.

As comparision - same reason why we have test-pilots and test-programs to validate the engineering theory.

 

"We can break "rules" - and sometimes we should, but we can not break laws "(physics) - this a quote from aerial warfare actually, just out of context.

(Meaning a fighterpilot can break his SOPs and tactics, but no pilot can violate physcis and do more than his plane can - unless its an exploit or flaw in a sim - lol)

 

PS: like "Every plane have own soul" :cool:

Edited by A-S

Share this post


Link to post
Share on other sites

Old logitech 3dPro just need r 10-20-...80-90-100, some dead zone and filtering. In other way plane will be unstable.

Without changing curves option every player would be forced to buy new good stick.

Brand new Saitek X52 Elo Uber Pr0 will give some advantage in 1vs1 with curves set to 100 but it will be unstable at long range deflection shooting in virtual front missions.

Everything have good and bad sides except cancer and impotence...

 

 

"Every plane have own soul"

Those differences are small.

 

I think that real pilots know what i mean but If someone dont belive me just seek in WWII books. You will find a lot of stories that:

"that was a commander plane, it was the fastest plane in squadron"

or

"he had bad luck to take off for interception in worst plane in unit, thats why he had engine failure after take off and died"

 

Same in sims, the curves can be treated like a option for boost plane abilities or some folklore.

Depends on point of view but both answers are correct.

 

Plus

After every  repair, review ,plane is checked with special procedure on ground by enginners and in the air by pilots.

Even when everything is ok with engine, avionics and so on, the pilot can give some istructions how ground crew should change balance of the plane.

So pilot have some influence on plane. Its limited and we cant compare it to changing curves ( too much difference between 1-2-3 and 100-100-100 ) but it exist.

Share this post


Link to post
Share on other sites

Yes all true Blakhart, but it´s much simpler:

 

1) First and foremost, are the surfaces themself just artificial (only graphical movements) and the flightperformance is just an "animated RESULT" as 3D object?

 

2) Is the input to output (stick to surfaces) simulation artificially, meaning no built-in mechanics and transitions of the frames themselfs are simulated and the response characteristics are completly dictated by custom user "curves"?

 

3) Or do real physical forces apply to the shape and individual surfaces as real solid objects resulting in drag, AoA, Lift etc etc?

 

4) And is is there an Input-Output module simulated in the sim, which accuratly reflects the true stick to surfaces behaviors for different models of that area?

 

1) and 2) is the Lockon and IL-2 generation (for that time being still well made).

But 3) and 4) is more like the ROF and BMS generation and more sophisticated in pyhscis and mechanics of simulated flight.

 

IF 1) and 2) applies then you NEED joystick curves, NO doubt whatsoever.

IF 3) and 4) applies then you don´t, not at all. Even if your joystick is a 30$ piece of sh++. Only deadzone maybe, but that´s it.

 

 

Simple as that it is.

 

Even more important is that virt. pilots can not search and find "errors" in the FM, which they can then exploit for tactical reasons.

This kills the nature of a fair real combat simulation as in the example i have shown you privately before.

Edited by A-S

Share this post


Link to post
Share on other sites

Real combat was rarely 'fair' - and if it was, any sensible pilot did his damnedest to rectify the situation.

Share this post


Link to post
Share on other sites

Real combat was rarely 'fair' - and if it was, any sensible pilot did his damnedest to rectify the situation.

 

You are reading stuff into what i write, read again in the right context. "Fair" in terms of unintented usage of errors in the code!

Real pilots also did not have flaws in a FM code, which they could misuse in order to win, did they?

 

I will not give you precise examples in order to prevent the abuse being spread throughout the internet - regardless what sim.

 

Generally speaking you are right though. Their lifes depended on it.

Edited by A-S

Share this post


Link to post
Share on other sites

"Even more important is that virt. pilots can not search and find "errors" in the FM, which they can then exploit for tactical reasons.

This kills the nature of a fair real combat simulation as in the example i have shown you privately before."

 

THIS

 

One of the reasons I am always skeptical of online, not that people are cheating, but that I am flying against an aircraft that will/can be controlled outside of its natural limits. This is available to all in control setup and few people will not modify their "curves" if it makes an aircraft more stable as a firing platform or more responsive than was the case in reality.

 

However to implement this in BOS will be a hard sell as many people are used to being able to customize their aircraft "feel" ;) also reaching a consensus of how an aircraft should "handle" will lead to some interesting debate  :) 

 

From what has been said, with the ROF/BOS flight model this could be achievable and would be a huge leap forward in flight sims 

 

Cheers Dakpilot

  • Upvote 1

Share this post


Link to post
Share on other sites

Also, the argument, that curves are there to stabilize older "spiking" joysticks and loose mechanics and dirty potis.

If that is the case we have "deadzones" and "filtering", which is exactly for that purpose in the fashion how it works.

Cleaning helps aswell.

 

If the joystick is soo broke and out of shape, that one even needs "curves" to compensate, then better buy a new one, seriously.

A very good joystick for WW2 you can get for 30 to 50 euro meanwhile (T1600m ie). Jet simulations and a proper HOTAS systems are more expensive, naturally.

I mean, come on,  one can not model the simulation "wrong", just because some people insist to fly with broken sensors or mechanics in their devices.

Edited by A-S

Share this post


Link to post
Share on other sites

How do you find the Cirrus sidestick, HeavyCavalrySgt? It seems to have attracted some criticism:

From here: http://www.aviationlawmonitor.com/2010/12/articles/cirrus-aircraft/steve-wilson-the-cirrus-airplane-has-serious-problems/

 

If that is anything like correct, maybe we need a few more of Crump's WW2 test pilots and engineers...

 

That was exactly my complaint, actually.  I think I rely a little on the feel of control pressure to get a sense of speed, and in a Cirrus it is artificial which threw me off.    It took maybe 20 or 30 minutes to get used to using just outside references like pitch angle and ignoring control pressure to feel oriented and in touch with what the plane was doing.  

 

I did see something interesting, though.  One gentleman I flew with made a lot of unnecessary roll inputs.  I'm not sure why, it seemed to be unconscious, but almost as soon as the wheels left the runway, the stick was going side to side, sometimes pretty violently.  The stick is pretty sensitive though, like a computer joystick but with a lot more throw and a more natural movement, so I think maybe he was putting in some roll unintentionally, catching himself, over-correcting, etc.

 

The stick is very functional, and fairly light,but it is different and it takes getting used to.  I am glad I don't fly Cirrus much, because transitioning back and forth would be hard.  I would want to fly pretty much exclusively in Cirrus, or very little.

Share this post


Link to post
Share on other sites

Answer is simple:

 

Default curves for all if all would have same quality of joystick and those equipment would never loose precision...

 

 

Instead of "Pimp my Plane" abilities curves options are a chance for players with smaller budget to enjoy the flight with medium class/old joysticks.

 

Till the curve option is available for all ( not guys with more money... ) everything is ok.

 

 

 

You dont find in real 2 same planes of 1 type.

They are almost the same at the begin of duty but after some time they start to be different.

 

Stall speed, max speed, max AoA, etc...

 

Every plane have own soul, problems and need special treatment.

 

 

I would like to see this modeled!  Slightly more or less power, easier or harder to start, slightly faster or slower engine response, or gauges that lag a little, or read off a little on something like MP.Flaps or gear that take longer to transition, 

 

It would probably be pretty hard to track in a campaign game, maybe.... Scratches on the windshield, then the canopy gets replaced and the scratches are gone, or the engine gets damaged or worked on and we have to determine new performance characteristics.

Share this post


Link to post
Share on other sites

RoF does model slight variations in engine performance - perhaps BoS will do the same.

Share this post


Link to post
Share on other sites

Nice thoughts but really just cream on the cake :) I'd rather time spent on getting the bascs right first,

 

A slight random variation in engine power (very small percentage) is great, and as you said i believe this is in ROF , this stops pure performance debates from becoming over the top, in service there would be noticeable differences

 

Cheers Dakpilot

Edited by Dakpilot

Share this post


Link to post
Share on other sites

This is actually quite interesting and important discussion and, unfortunately, a huge hurdle in modeling the real feel of the planes and flight for any simulator.

 

I disagree somewhat with the OP, although I do understand where he's coming from. Mainly two reasons:

 

1) Every stick is different and virtually none of them even closely resemble the real deal. If you take the ability to model curves away you won't be creating a level playing field at a functionally realistic point, but a group of different feel profiles depending on the stick type at arbitrary points regarding realism.

 

I have recently been testing modified Cougars and Warthogs in side stick and extended center stick configurations and although they are similar in many ways (heavy, long throw controllers) each of the configurations gives an entirely different feel to the planes / helicopters. None of them come close to the real thing - I haven't flown real fighters, but it's a safe assumption and I'll go deeper into this at the end of this post.

 

2) Just about every stick can have a curve programmed into it, some even come with built-in curves you have to counter in software if you want to get rid of them. Thus the lack of ability to set curves in-simulator becomes simply a matter of inconvenience.

 

----

 

What makes this discussion important is the fact that there are no easy answers. FFB sticks are the closest thing to the real deal, but I'm not convinced about their capabilities or the FFB modeling in general to produce the real feel of pushing against airflow. Someone mentioned that all simulators / joysticks are FBW and that's mostly correct. Unfortunately we are entirely disconnected from the simulated airflow. I have little experience on real planes, but as someone above mentioned, springs or other artificial centering mechanisms are not the same thing as feeling airflow against the surfaces.

 

F-16 is actually easy to model correctly, since force sensing solutions exist and the plane itself is FBW. It's interesting to note, however, that during the development there were issues with PIO due to the small delay in control response. The modeling is much more complex for planes that have a more direct connection between the controls and the control surfaces, because strength and control inertia come to play. An MS FFB2 is vastly lighter and easier to throw around than for example an Me109 stick at 700kph, so some kind of control inertia / pilot strength simulation needs to be done.

 

There is an unfortunate and serious drawback to this, since it not only disconnects the virtual pilot from the airflow over the control surfaces, but also the actual position of the controls. This makes it very difficult to get a "feel" for the plane and it's something even RC model builders have tried to eliminate as much as possible, let alone those working on real FBW systems. For us there is no good answer, the only proper solution would be to have a control stick that responds exactly as the real thing - which is of course not going to happen.

 

So, although I understand how it would be a benefit to have each plane's control feel modeled (as it is in many cases), I don't see removing the ability to create control curves as a good step forward for the above mentioned reasons. Since the starting point (feel of the actual controller) is not even close to real and its movement is already disconnected from the movement of the simulated plane's controls I think that any tools - be it curves or whatever - that help you get over this disconnect and get a better feel for the plane can only be a good thing.

Share this post


Link to post
Share on other sites

lighter and easier to throw around than for example an Me109 stick at 700kph

The intent of the Bf-109 designer was to meet a stability and control specification the did allow the pilot to kill himself at high speed.

 

Why not approach it from that point of view by limiting the stick force per G?

 

 

It is complicated as even in the Bf-109 or any airplane, the Operating Manual must be followed.

Share this post


Link to post
Share on other sites

After all a good thread indeed with constructive feedbacks from all.

 

My summary:

 

Regardless what joystick one has, may it be long, short, potis, hall-sensors, force-feedback, pressure sonsoric, fat, thin.. you name it ..

the INPUT on the sensors and the range is always mechanically linear from 0% to 100%

OF COURSE every piece of hardware has its own feel, one tighter, one less tight (spring), one smoother the other "more plastic" and worn out, but still, the input range physically is 0% - 100% (considering there are NO devide-interfaces - software - included simulating curves to begin with).

So it will be just natural, that every joystick delievers a different feel, but that is not the point, because all of them will have one thing in common and that is the linear mechanical input band, maybe it be a 1024 or 4096 resolution.

 

IF there is an INPUT-OUTPUT module (stick to surfaces algorithym) built into the aircraft type (the code) BEFORE the actual FM code applies determining the phsysical results of the craft with the environment, it will be possible to deliever the SAME surface behaviour, responses and results for everyone as all raw inputs (regardless joystick type) will be linear from 0% - 100%.

It will not matter if your stick deflection is 10deg, 40deg or 60degrees on a Saitek 52 or a Thrustmaster or anything else as the amount of stick reflection will be taken as the same for all joystick types.

 

The problem is that such a Module is very hard to program and was never done before (not in IL-2, FSX, X-plane, DCS, CloD, FalconAF etc etc) with one exception and that is FalconBMS.

In order to achieve this for WW2 planes we would have to understand and analyze each individual aircraft and what kind of mechanical transitioning (stick> wires> rollers> rods> to surfaces) they have, meaning what kind of (mathmatically) deflection we will get one the surfaces for a specific amount of stick deflection in the cockpit (109, 190, Yak etc etc) - and than for different flight regimes!.

 

This is a "tuff" thing to do as one requires the blue-prints of each plane and the mechanical features have to be translated in "numbers".

Such a Module would simulate the true response nature of a plane ITSELF with linear joystick curves. Surface deflection limitations in angles and move-speeds due to high G-forces and too much stick-forces (in pounds) can be added (simulated) this way too (even you as player don´t feel those forces on the stick) !

IF one still wants to change his joystick-curves afterwards anyways...this option can be given, why not? Also "Deadzones" and "Filters" should be given to players, but the point is, that the TRUE handling- and flight-characteristics are defined and given by the planes own nature to begin with - and not by user settings only !

 

I hope i make more sense now :)

 

Respectfully

 

A.S

Edited by A-S

Share this post


Link to post
Share on other sites

What I don't quite understand is the significance of 100% to 100% mapping especially in games like Il-2 (the old one at least) where you are not even directly connected to the control surfaces but are instead asking the pilot to exert x1 and y1 amounts of force per x2  and y2 amount of stick deflection. In this kind of a system it is often beneficial to have a curved response, because most of the time you are not exerting even close to maximum amount of force. Isn't that btw something like the module you're talking about? You tell the system how much force you're applying to the stick, it determines the corresponding stick deflection, which in turn actuates the control surfaces and flight model takes over from there.

Share this post


Link to post
Share on other sites

 You tell the system how much force you're applying to the stick, it determines the corresponding stick deflection, which in turn actuates the control surfaces and flight model takes over from there.

 

Exactly, "how much force" or in our case "how much stick deflection" is applied (for conventional moving sticks).

 

Whereas the mechanical actuation method* is THE part to understand first, for each model or airframe. This is where the "module" comes in replicating exactly those technical aspects of the transitioning.

 

Once that accomplished, one can (codewise) additionally program force-limitations - In example:

 

On taxi you move your stick 100% and the elevators deflect 100% aswell, but with very high speeds the elevator just deflects - lets say - only 70% (or less as this has to be calucated) EVEN if your joystick on your desk is fully pulled back, because either it would create too much G-forces for the pilot, or the forces on the stick would be so high, no human would be able to pull, or the plane would just loose mechanical integrity and just fail in itself.

 

We can not simulate the true stick pressures (pounds required to pull back on the controler) on our devices, but we can simulate the consequences as mechanical, physical and even humanly limitation within the simulation (RoF did that partially for example).

 

*If i find blueprints of the transition mechanics of some ww2 fighters, i will post them. The "module" (code - interface) would exactly simulate those.

Edited by A-S

Share this post


Link to post
Share on other sites

Actually instead of being truly force based I think Il-2 might be force limited in the sense that you can only pull so much total force, but moving the stick around isn't any slower or faster based on the strength. Also normally the stick positions will probably be basically a one to one mapping. 

Share this post


Link to post
Share on other sites

Ok I think I understand what you're talking about. I see these as the different stages of joystick movement to plane movement:

 

1) Joystick position to virtual controller movement (for example through a force-based system as in Il-2)

2) Virtual controller to control surfaces (pretty straightforward in BoS era planes, may contain hydraulic actuators and FBW systems in newer planes)

3) Control surface to plane movement (airflow modeling, structural bending of the surfaces etc)

 

The thing I'm missing is the significance of removing joystick curve settings. Unless we're talking about a direct mapping in step 1) (which is rare in BoS era simulation) there already exists a disconnect between the joystick movement and virtual controller input, so even with the exact replica controller the response will not be the same as in the actual plane. There is no way to get even close to the real feel of the plane, so it's hard for me to see the value of enforcing linear input when the translation function of joystick to virtual controller is not direct mapping in the first place.

 

For example in RoF the control characteristics (such as ineffective ailerons) are modeled pretty nicely regardless of response curving, since no matter how much force you apply you can only get so much response. One to one mapping is of course the most realistic way, but only if you have an exact replica of the controller and forces. Otherwise you're looking at a compromise in control ability - and curves may get you further or closer from the real deal depending on the case.

Share this post


Link to post
Share on other sites

 

Such a Module would simulate the true response nature of a plane ITSELF with linear joystick curves. Surface deflection limitations in angles and move-speeds due to high G-forces and too much stick-forces (in pounds) can be added (simulated) this way too (even you as player don´t feel those forces on the stick) !

IF one still wants to change his joystick-curves afterwards and adapt anyways...this option can be given, why not? Also "Deadzones" and "Filters" should be given to players, but the point is, that the TRUE handling- and flight-characteristics are defined and given by the planes own nature to begin with - and not by user settings only !

Edited by A-S

Share this post


Link to post
Share on other sites

I have read that text before, but I still don't really understand how it answers the question. I can see the relevance of one to one mapping in the part where the virtual control stick movement is translated to control surface movement (in direct linkage cases, of course, otherwise it's more complicated), but how do linear joystick curves translate to realistic feel in any other case besides the force-sensing cougar and Falcon?

 

If forces are simulated or there is some other translation function between the stick and virtual controller it won't be a one to one mapping anyway. Then again each controller is different, so even if they were directly connected to the virtual controller's movement their feel and controllability would be different for everyone. For example an UH-1H feels slightly nervous and twitchy in hover with a modified Cougar in sidestick configuration, but with an extended shaft in center stick configuration the chopper becomes far less sensitive, of course requiring much more pronounced control movements when you really need to move around.

 

So to me it feels that of course the control properties of the planes should be modeled, but I still don't see how it would be more realistic to have linear joystick response. That way instead of putting the feel of the planes in the hands of the players it would be decided by joystick manufacturers. Although neither is entirely realistic, I don't really see how the latter would be any better.

Share this post


Link to post
Share on other sites

That is a bold statement. You're by-passing hysteresis, dead zone and centring force, which differ per joystick.

You will have to match the joystick characteristics to the flight model, which is why the tuning capability is there.

Interesting thread, I've been fascinated by open and closed loop flight control systems and how they have changed due to fly by wire technology.

 

Could you explain hysteresis in the context of how it manifests in a simulator control system? It might be interesting to the community here.

Share this post


Link to post
Share on other sites

@ AndyHill.

 

What we have in modern jets nowadays are modern computer-controlled Fly-By-Wire systems. The pilot just inputs commands with the joystick in what direction he wants to fly, roll or pull, but the computer and only the computer commands and controls the actual correct surfaces and decides how much deflections will be applied to each flight-surface over electrical and hydrolic systems, mostly as double- or quad-backup systems. The pilot has no direct influence over the flight-surfaces as such as the computer takes over by gaining data for various flight-regimes with various sensors and computes from that telemetry data the best results in terms of stability, limits and performance etc. etc.

Most Jets (or especially semi-FBW systems such as the Mig-29) provide the ability to override certain aspects of the flight control system allowing the pilot individual manual "override" control options for different situation.

 

So, in this case the previously so called "Module" is the Flight Control System itself, or for better understanding and simply put,  just a software-interface from stick to surfaces. Thus, the "linear" stick inputs by the pilot are computed or translated into "dynamic" (or "curved" as analogy) surface behaviours.

 

 

In WW2 planes we have mechanical flight control systems. Computing interfaces didn´t exist at that time, but mechanical interfaces DID !!!

What i mean by that? Simplified, such a system (a closed linear transmission) would look like this:

 

gk9azzp5kardvsvbmg0l.png

 

 

This example (pic above) shows as principle a linear 1:1 transmission, meaning no forces are mechanically re-distributed nor are any movements (lenght) mechanically converted like a conventional pulley system would do (pic below).

 

 

450px-Four_pulleys.svg.png

 

 

But in most WW2 fighters this mechanical system is much more complex as the 1:1 example above, as the pulleys (disks) are not always "round" and of same size and therfore act as such, counter-springs are mostly added etc. etc, so in other words we have a true polley-like force and lenght distribution system  relative to your linear controller inputs. The pilot still moves his stick "linear", whereas the surfaces respond "non-proportional" (or "curved" as analogy) - as above.

 

This is the important part to understand!

What we understand as "joystick curves" is already built-in as "mechanical response properties" (force and lenght wise) into most airframes of that generation. I wish i could find few detailed pictures from that for the BF109 or the Mustang and it would make much more sense.

 

So, in this case the previously so called "Module" would be the mechanical flight control system itself with its mechanical properties of force and length distributions, or for better understanding and simply put, just an additional ingame "software-interface" simulating the mechanical response- and control-characteristics from stick to surfaces of the plane ITSELF. Thus, the "linear" stick inputs by the pilot are computed or translated into "proper" (or "curved" as analogy) mechanical surface behaviours.

 

 

This way true responses can be simulated with linear (!) joystick inputs (as all joysticks have THAT in common). You will get a "true" response- and control-feeling of the simulated aircraft with LINEAR joystick settings. AFTERWARDS you still can - if you like - change your "joystick curves", "deadzones" and "filtering", if necessary. Now those game-features become secondary game-options (to compensate hardware) and not primary flight- and control-characteristic determinations :)

 

The problem in simulating such a "Module" is, that we need to have or get all technical specification of each simulated aircraft with its manual flight control system and convert those mechanical properties into "numbers", which can afterwards be computed in a INPUT-TO-OUTPUT software module. Once that achieved, the FM (flight physics) Module is added simulating the real physcial forces of the airframe and the flightsurfaces with the environment. This is the only proper way to do it from what i understand.

 

Easier now?

Edited by A-S

Share this post


Link to post
Share on other sites

I mean, come on,  one can not model the simulation "wrong", just because some people insist to fly with broken sensors or mechanics in their devices.

 

Dont forget the "gamepad". :)

 

Now serious:

 

 

Not only is he open to questions and addresses them, but he INSISTS till today NOT TO introduce joystick curves and to be honest it makes absolutly sense, because in his flight-model you don´t need them, regardless what joystick or you have, may it be analog or pressure sensoric. You have the option of dead-zones and saturation (bands) though, but that´s it.

 

Cl+D remove the joystick curves applet of previous il-2'46, has only sensibility and deadzone adjust, you know it are introduced something new in joystick response, or are just a "rushed" :rolleyes:  GUI?

 

Sokol1

Share this post


Link to post
Share on other sites

Yes I understand what you're talking about with the linkages and already agreed that they should be modeled. In addition to linkages you also have the structural flexibility of the control surfaces / wings that has effect on controllability - but I don't really know how much of an issue that was before the transonic jet era. As I mentioned, modeling the controls as they are is a good idea and should be done, no question about that. Actually I think we agree on just about everything else, I just don't quite understand the role of joystick curves in all this.

 

Earlier I divided the control system into three phases:

1) Joystick position to virtual controller movement (for example through a force-based system as in Il-2)

2) Virtual controller to control surfaces (pretty straightforward in BoS era planes, may contain hydraulic actuators and FBW systems in newer planes)

3) Control surface to plane movement (airflow modeling, structural bending of the surfaces etc)

 

The way I see it, the linkages and other controller to control surfaces are stage 2, whereas joystick response curves are in stage one, which is basically all about translating your hand movements to the simulator. I know of exactly one case (there are probably others, though) where 1:1 mapping is actually possible and that's Falcon with force sensing stick (assuming it can provide an operating mode with realistic control forces, which I assume it can). In many simulators there are things such as control inertia (the virtual stick will refuse to move as quickly as you can move your joystick) and force modeling (aerodynamic forces acting on control surfaces will limit how far / fast the virtual control will move).

 

In those cases 1:1 mapping simply does not happen even with linear curves on the joystick. There is another stage of controller modeling between your hand movements and the virtual controller: the virtual pilot. The hand movement will be delivered to the simulator through a modeling layer that will not result in 1:1 mapping to the movement of the virtual stick regardless of the linearity of the control curves. The 1:1 mapping is in almost every case lost before the cables and pulleys get to do their job.

 

So again, I agree with you on the importance of modeling the control systems as a whole, because they bring a lot of character to the planes. In many cases things such as insufficient power in actuators or bending of surfaces resulted in lethal loss of control at the edge of flight envelope - which is a big deal when modeling the aircraft. Many simulators provide a picture of ideal, rigid objects as planes, which is far from the truth especially when approaching the limits of the planes' capabilities.

 

The difference in our views is in my inability to really get a grab on the role linear joystick curves play in this, since to me the joystick response should ideally give the player the best possible approximation of both feel and control abilities of the real deal. The former requirement is in most cases a far away dream and the latter depends largely on the make and model of your joystick, resulting in a number of arbitrary controllability responses with linear mapping - none of which are 1:1 to the virtual stick. Although in some cases linear curve may provide the best approximation, especially with shorter joysticks that have a shorter throw will likely benefit from curving.

Edited by AndyHill

Share this post


Link to post
Share on other sites

Conventional PC Simulations:

 

1st Joystick device inputs

2nd FM module (flight physcis, if at all, because some older FMs are just "scripts" or "result-" and not "cause-" calculations - not true physcial forces)

 

Professional Simulations:

 

1st Joystick device inputs

2nd CONTROL INPUT-OUTPUT MODULE replicating/simulating the real mechanical transmission methods.

3rd FM module (real flight physcis)

 

 

Benefits:

 

This way true responses can be simulated with linear (!) joystick inputs (as all joysticks-devices have THAT in common).

You will get a "true" response- and control-feeling of the simulated aircraft with LINEAR joystick settings.

 

Game-features (curves, deadzones etc) become secondary game-options (in order compensate hardware flaws and to fine-tune), BUT are NOT primary flight- and control-characteristics determining or defining options anymore. Also, virtual pilots don´t have to fiddle and screw around with their joystick curves anymore in order to make it "feel" right*, becuase with an INPUT-OUTPUT MODULE, which simulates the real transmission ratios, the physical limitation etc. etc. and thus the real stick to surfaces controlling- feel,  it WILL BE right from the beginning.

 

* Besides, how many virtual pilots or real aviators, who fly modern props actually do know how a Bf-109 or a Mustang or a P-40 really "feels" ;) ?

 

 

PS: "Linear curves only appy to force sensoring sticks", which is not true btw ..read this  http://www.benchmarksims.org/forum/content.php?167-Flight-Model-%28FM%29-Developer-s-Notes-Part-7

Edited by A-S

Share this post


Link to post
Share on other sites

I think you're missing one stage there, which is quite important as far as joystick linearity goes and that's the pilot. For example in classic Il-2 we're not controlling the virtual stick with the joystick, we're asking the pilot to do one thing or another. Since he is not necessarily able to do what we ask of him (control force, controller inertia modeling or injury), there is no 1:1 connection between the joystick and the plane's control systems at any settings, linear or curved. Also most controllers are not force feedback and even those that are can't replicate the movement range or forces of the real plane's controller.

 

I also don't think that joysticks have linear input in common. Perhaps semantically, but how can it be common when a linear input on five different joysticks will give five different feels and five different control capabilities? Choice between them will probably be largely a matter of opinion, but very likely none of them will be close to real. (I think we can omit some early Saitek models that had exponential curve in hardware from this discussion) Perhaps you could clarify a bit what you mean by "feel being right", because to me it sounds odd, since when flying any simulator with my FFB2 or the extended ES-Cougar the feel will be very different for both even if everything else is the same.

 

I'd like to clarify a bit about the force controller; linearity is not the point there, I just think that currently the combination of force sensing Cougar and Falcon is the only one where the feel will actually be very close to the real thing when connected to realistically functioning control system. The only exceptions would probably be G-LOC or other injury/fatigue situations where the virtual pilot will be exerting less force to his stick than you are capable of providing to the force-Cougar.

Share this post


Link to post
Share on other sites

I fear you are getting lost in "what if" scenarious or peripherial visions of probable variations and thoughts about them AndyHill

 

There is no "grey zone" here. There are a wrong ways (old fashion) or meanwhile better ways how to simulate mechanics and physics (FLIGHT) on a computer, especially with the performances available today compared to Pentium 4 or pre 64bit days.

You need to look a litte bit beyond the horizont of the simulations for the gaming-industry and see how "professionals" solve those challanges.

I am not talking about subjective opinions here, but about solid solutions such as programing in exchangeable and compatible "modules" from coding perspective, which can take care of specific aspects and tasks individually but can also operate combined (modules).

 

It does absolutly not matter if you fly with a 10.000$ stick or with a curved "banana", because the simulation ITSELF (controling- and response-characteristics AND flight-model) must be approached and solved in "true" and realistically logic ways WITHIN the simulation FIRST.

 

A simulation can only provide a for all device types CONSISTENT correct response-characteristics internally based on their exnternal LINEAR inputs, because this is what every joystick has in common - moving a mechanical "grip" with sensors of different built attached from 0% to 100% - directly proportional.

 

Obviously you will have different "feels" with different joystick models, but as unfortunate this may sound, its YOUR luck or misfortune, what joystick you bought, NOT the fault NOR the responsibility of the simulation producer.

In mathematics and physics (simulations) such a thing like "it feels" doesn´t exist.

 

Respectfully

 

A.S

Edited by A-S

Share this post


Link to post
Share on other sites

 (control force, controller inertia modeling or injury), there is no 1:1 connection between the joystick and the plane's control systems at any settings, linear or curved.

 

THIS (and the transmission ratios) is why a Input-Output module is built (coded) for :salute:

Share this post


Link to post
Share on other sites

A-S, the very simple answer to your question comes down to biology. Human beings are very good at sensing and closely imitating a pressure/force input. We are not so good at sensing or imitating a positional input. In real life, pilots do not fly with controls that require precise positional inputs to achieve the desired flight path, instead pilots learn (either consciously or subconsciously) to apply pressure to the controls in the amount necessary to get the desired result. This is specifically mentioned in almost every modern flight training textbook.

 

Unfortunately, force feedback and other technologies have not become very popular or progressed well in the flight simulation genre. Due to this, a large majority of players use joysticks which have sensors which read the change in position of the stick to determine the joystick/rudder's output. As I said above this is completely wrong in terms of human physiology, pilot training, and actual aircraft design. To compensate, various parameters are built into the joystick or into the games themselves to account for this problem. Joystick curves, dead zones, and other options are available to tweak the input/output relationship in an attempt to overcome these limitations in hardware design. It isn't perfect, but it helps. Certainly, flying with a PC joystick pales in comparison to the ease of the real thing.

 

This won't likely go away until we have hardware peripherals that more accurately simulate the way that real pilots fly. When a large majority of peripherals account for this disparity and are as easy to use as in real life, the need for software tweaks to fix our hardware limitations will become obsolete.

Edited by Crow

Share this post


Link to post
Share on other sites

Again I'm not saying there should be no control system modeling, because there absolutely should and to my knowledge in many simulators there already is. But the pilot needs to be modeled as well (in most cases), which by very definition excludes the possibility of full 1:1 mapping of joystick to virtual controller. As to the difference in joysticks; why would it be better for example for someone to fly a Huey that feels twitchier and is less accurate to steer in hovering than the real deal just so that the steering can be fully linear? Some physics are actually based on subjective things (such as visual brightness classification of stars), but things like control throw, centering force and spring type are easily quantifiable factors and differ greatly among different controllers.

 

THIS (and the transmission ratios) is why a Input-Output module is built (coded) for :salute:

 

I'd rather divide such module into at least two parts, joystick to pilot and pilot to virtual stick. I've only sat in cockpits of relatively old jets (and none of them were moving at the time), but I think their controllers had more in common with prop fighter controllers (which I've only handled when they weren't installed in a plane) than any joystick I've seen. They all were relatively massive and built to take a lot of force - since you had to use a lot to move the plane in certain situations. I couldn't move them very fast even at 1G. Now if there was direct 1:1 mapping between an FFB2 and any kind of a fighter plane, I would be able to do absolutely superhuman thing with it. I could throw the planes around the sky like I do with model planes even at high speeds, of course I would have to be careful about over-G, since at speeds the slightest twitch would cause the plane to disintegrate mid-air. That's why I think you need additional layers to the modeling and 1:1 mapping isn't the answer - unless you are actually operating the real plane.

Share this post


Link to post
Share on other sites

A view from a different example but highlighting one of the issues being discussed

 

"One thing I wish was "corrected", or at least additional guidance provided, notwithstanding the hardware, is that there seems to be too much latitude with regard to the axes in terms of available curvature and sensitivity, so that in effect you can create your own custom flight model depending on how extreme the settings, and again it would be nice if there were stricter limits placed on this functionality."

 

Cheers Dakpilot

Share this post


Link to post
Share on other sites

 

does absolutly not matter if you fly with a 10.000$ stick or with a curved "banana", because the simulation ITSELF (controling- and response-characteristics AND flight-model) must be approached and solved in "true" and realistically logic ways WITHIN the simulation FIRST.

 

 

Absolutely.  I think too much is made of the joystick curves anyway because folks do not understand so much is a characteristic of the aircraft.  I do like the idea of one fixed curve and as limited input as possible to keep the integrity of the model.

 

The basics to simulate unique stability and control are already found in the basic FM I believe.

 

Maybe a method could be devised by treating the FM as an Onion with distinct layers.  Outlined below is a very simple method explaining the basic concepts.  I don't think a discussion on equations of motion will be very productive.  Instead lets concentrate on the basic concepts.

 

First Layer - Aerodynamic Forces at the AC

 

IL2 BOS FM sums the forces at the Aerodynamic Center for every condition of "flight".  It simply does the math, most likely with section theory, and calculates forces at the AC. 

 

The moments about AC and the Moments about CG determine the size and direction our aircraft will move as well as the forces the pilot will experience.

 

Second Layer - Weight and Balance to determine the stability margin.  One could also use MAC here so this is not intended to write the programing only discuss the basic concept to be used to implement stability and control modeling.

 

Simple Moment = weight x arm

 

CG location = Total Moment / Total weight

 

Stability Margin = CG - AC

 

Resolve the pitching moments along the axis.  A positive stability margin means our pitch will move toward stability.  Negative means towards instability and a zero value is neutral.  This gives us the direction the movement of the aircraft will go.

 

Third Layer is pitching moment about the CG.  This tells us the size of the movement in the direction we already determined.

 

This comes from derivation using the airfoil data or measured using the whole aircraft.  It is a polar curve defining the relationship of Lift forces to pitching moments and shows us the slope of the pitching moments.

 

http://adg.stanford.edu/aa241/stability/images/image6.gif

 

Along the Axis:

 

Mcg = CMcg * q * S *c

 

Moment about the CG = Coefficient of Moment x dynamic pressure x reference area(wing area) x Mean Aerodynamic Chord

 

 

This pretty much concludes the resolution of the movement of the aircraft.  We have the pitching moment about the CG and the moments about the AC.  With this information, we can determine direction and magnitude of motion.

 

It does not resolve the forces experienced by the pilot nor does it discuss trim.  Those too can be resolved in similar fashion which I will post later.  Not going to spend all day on this.

 

You should get my idea on how a stability model can be built from the information contained above.

  • Upvote 1

Share this post


Link to post
Share on other sites

A-S, the very simple answer to your question comes down to biology. Human beings are very good at sensing and closely imitating a pressure/force input. We are not so good at sensing or imitating a positional input. In real life, pilots do not fly with controls that require precise positional inputs to achieve the desired flight path, instead pilots learn (either consciously or subconsciously) to apply pressure to the controls in the amount necessary to get the desired result. This is specifically mentioned in almost every modern flight training textbook.

 As I said above this is completely wrong in terms of human physiology, pilot training, and actual aircraft design. To compensate, various parameters are built into the joystick or into the games themselves to account for this problem. Joystick curves, dead zones, and other options are available to tweak the input/output relationship in an attempt to overcome these limitations in hardware design. It isn't perfect, but it helps. Certainly, flying with a PC joystick pales in comparison to the ease of the real thing.

 

 

Ablosullty correct!

 

This is why professional simulations have an input-output module besides deadzone, curves and joystick filters. Such modules can be coupled with advanced controls, simulating true stick forces (rare though, as most modern planes fly electronically), but in our case this is not possible as no force force-feedback stick can simulate such a force anyways.

 

It is absolutly clear that a PC flight expirience will never be able to deliever the real flight sensation. About this we don´t even have to discuss.

 

All what such a module does is simulating aspects of the mechanics from the 3D controller to the 3D surfaces such as control forces (defleciton limitation), controller inertia or injury or fatigue and most important the transmission ratio as in real WW2 one does not have a 1:1 controller to sick deflection. Real WW2 figthers (not all) have a "curve like" stick behavior naturally by their own nature, and this it to be simulated the first time aswell! This has not been simulated before, thus the difficulty to fly with linear "curves" previously and the attempts to replicate the "real" feeling with "fictional" curves in the joystick setttings.

Btw, even IL-2 has a simplied Input-Output module - if wou will - as once your cables are destroyed you loose control-abilities over the surfaces. ;)

 

So, what can such a module do then and why we need it?:

 

- It can simulate the true control- and response characteristics of each individual airframe modelled.

- Simulating real transmission ratios (relative 3D stick to 3D surface deflections) - mechanically pre-defined "curve-nature" of the plane itself.

- Control forces (deflection limitation), meaning limitations of deflection due to high forces on the stick in certain flight regimes based on an "average human" (no human (real or virtual) can pull 250kg with one arm). The 3D controller doesn´t travel further as permitted by the "virtual" force limitation even if your hardware device is on full deflection.

- Controller inertia, meaning difference in surface deflection-speeds at 0 and high speeds (shown in BOS video already)

- Injury or fatigue (bent, damaged etc)

 

- what else?  any ideas anyone?

 

So, as you can see we have our hardware device (our joysticks) and we have the actual flight-model itself, which is nothing else than the physical and mathematical computution of the airframe as solid body and its surfaces with the environment (air, airdensity, temperature, forces, lift, drag etc. etc. etc.)

 

But we also NEED something inbetween the input device and the flight-physcis modelling and that is an INPUT-OUTPUT MODULE for the virtual controls.

 

4js5ggltze5u9nu6v7h4.jpg

 

Thanks everyone for the feedbacks S!

Edited by A-S

Share this post


Link to post
Share on other sites

Well I refuse to read this entire thread. But I have voiced my opinion on this elsewhere and will repeat it breifly here. I have flown lots of different planes from the DC-10 down. Some planes were either fully hydraulic or boosted or just simple cables and pushrods. One was FBW. The current is a 750 pound Pitts S1S that I own. I have no engineering knowledge and can't argue math. But I can argue how things are supposed to fly based on 30 years of flying. (Jeez, that's really true. Getting old!)

 

I believe that the game designer needs worry about one rule when it comes to input: Player gets what he calls for. Neutral stick is trimmed control surface. Full stick is full control surface displacement. Curves are the players discresion. Not realistic? I disagree. When I fly the Pitts I have the option of 2-fingering the stick which I do in formation or grabbing the thing with two hands (although I never do that). The controls are so light on the Pitts that I can get full displacement in nearly the time I can do so with my CH. (What happens when I do that is where flight sims show there limitations--not how they handle input devices.) Not allowing the player to set up his stick as he sees fit does little or nothing for realism. But it severely hampers the enjoyment he gets from the game.

Share this post


Link to post
Share on other sites

Not allowing the player to set up his stick as he sees fit does little or nothing for realism. But it severely hampers the enjoyment he gets from the game.

 

Noone says that ;)

 

The user still can (should have the option) to set up curves, deadzones and filtering as device compensation (if his stick is really crappy), BUT NOT  as primary flight-characteristics changing, controlling and manipulating options :) .

Otherwise you (or everyone else) ALWAYS will get "false" and "fictional" repsonses. You need to read the whole thread indeed.

 

PS: The days where one could "fuck around" with his unrealistic "joystick curves" in order gain tactical or performance advantages (exploiting the game) should be OVER !

 

ARCADIAN HIGH >> BYE BYE. (I know that this scares many as they depended on those features).

 

Quote: "This sim does feel strange for you because you never flew a realistic FM, you just believed you do"

Edited by A-S

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×
×
  • Create New...