The control movement is translated by the hardware into a number. That number indicates only the actual position of the control, not the movement itself.While maybe possible to model in airflow over the control surfaces in determining how much deflection a specific stick position would cause I don't think this is in the capabilities of the MSFS core as it is.It would require a complete rewrite of the core components of the product, and require a completely different dataset for the aircraft as well.Aircraft would have to have data built into them about control surface area, control system enhancements (hydraulic force amplification of control movement, FBW control surface parameters, etc.), in other worse how much does the power the pilot exerts on the control stick/yoke get amplified before it's applied to the control surface itself.For FBW often there is no direct linkage at all, but the change in control stick position will indicate the actual movement of the control surface itself rather than the position of the stick determining the position of the control surface.While all this is possible to do I doubt it's possible in a generic flight sim environment like MSFS, at least not without making the programming interface for addon aircraft so complicated that it's next to impossible to create any at all.

>I and couple of individuals from the MSFS team have a theory as to why the more experienced pilots feel that the MSFS default aircraft are quite accurate in the flight modeling.

>Ron>I would like to ask you the question I placed above and which>Geof another real world pilot touched on too.>>Our Flight sticks whether they be control columns or whatever>only move a small amount regardless of slow flight or high>speed cruise. Yes Peter, I saw your comments.Measurements I took real world in a twin regarding elevator>and rudder confirms that at slow speed ie in the flare the>column will move several inches as you pull back to land the>aircraft and the control forces are not large. Right, lower dynamic pressure on the control surfaces means they have to be deflected more to get the same force/moment at low speeds. In fact, force varies about as IAS^2 for a given deflection. "Steering" a motorcycle has similarities. At high speeds one applies a twist (torque) to the handlebars to turn. In fact, the twist is to the CCW for a RIGHT turn! The handlebar turns very little, it is the applied Torque that counts. At very low speeds one has to really twist the handlebars to maintain stability. Same as with a bicycle. >At high speed and in a trimmed state the column moved 1 cm to>pitch 10 degrees and the forces are huge.>The same goes with rudder.>Many believe this to be a hardware problem I believe it to be>a programming FDE problem. Those effects depend on parameters and tables in the AIR file. For lower speed (prop) AC, control moment tables are typically flat. If the control 'authorities' are set correctly it should take more JS/Yoke movement to get a given pitch or roll rate than at higher speeds. Center of Lift, C.G., and other AC dimensions are also significant. It is difficult to get the correct balance (in all 3 dimensions) unless the Center of View is set at the 'Center of Lift'. Aircraft.cfg often has a comment that tells where that 'Center' is. At the standard 1/4 MAC, Waterline, and lateral offset. Many FS AC are set with the wrong vertical CG and Center of View (Normally Aerodynamic Center), and that affects various things. Including handling on the ground. Note SUV's have a high CG, and tend to roll over more easily than other autos. --- Slower speed AC are generally set to deflect the control surfaces proportional to the JS. However, for jets, with no realistic FFB, one has to reduce the control movement at higher IAS. Otherwise it's too hard to get fine control. If realistic Force rather than deflection was implemented one wouldn't think much of JS deflection, rather he would apply a force to to get the pitch or roll rate he wanted. The deflection would mainly be noticeable only at low speeds. However, the control reduction isn't all that much. Looking at one of my Jets, I see the elevator is set to move 80% as much at 300 kts IAS as it does at low speeds. While the ailerons move 70% the low speed amount. In part, this is because jet transports typically have two sets of ailerons, and only one set is employed at higher speeds. Spoilers also start moving at some aileron deflection, and help to increase the roll rate at lower speeds. The effects are only realistic if properly done! Due to the lack of a decent Yaw Damper in FS, I've set the rudder to move much less above Rotate Speed. In real jet transports the rudder is limited to about +/- 5 degrees above 180 kts or so. >If we take a stick with no spring at all the FDE should>require the stick to move several inches during the flare and>1cm to pitch 10 degrees in high cruise. That would assume correct FFB. I bet many non-pilots aren't aware that when an AC with direct links to the elevator and ailerons is parked, the Yoke is loose. It certainly is not spring loaded. I remember tying the seat belt around the yoke to keep the controls from flapping when parked if a wind came up. Only after the AC picks up some speed on TO does the Yoke start moving to the neutral position. To rotate for TO, one pulls back a bit and feels the opposition of the air flowing over the elevator. How hard one has to pull depends on the elevator trim. If large AC with direct cables to the elevator aren't trimmed for TO they can be near impossible to rotate. At high speeds, the Yoke normally moves very little. Roll rate in WWII fighters is mainly limited by pilot strength. Further, such an AC in a dive can be very difficult to pull out. Partly due to the high pull required on the Yoke, partly due to actual deformations of the control surfaces. If your B-17 went into a dive you can bet both the pilot and copilot were pulling as hard as they could to try to recover. I also suspect the copilot helped hold a B-17's wings level if it was taking off with full power and the torque from four engines needed aileron compensation. >I accept the FDE cannot control trim forces which has to be>simulated through the controls and springs or other forms of>mechanical resistance but the actual control movement should>be dictated by the FDE which it isnt. What is wrong? I said above that in slower (prop) AC, there is normally a one to one relationship between JS movement and control deflection. And, the moments are proportional to control deflection. However, real AC are more complex. Downwash from the wing affects the tail, and Downwash varies with AoA. 'Prop Blast' enhances air flow over the tail, conversely, if the engines are idled, that disappears. So, one would have to move the Yoke farther than if the engine were developing much power. UNFORTUNATELY, prop air flow on the rudder disappeared in FS2K2. Prop air flow on the elevator then disappeared in FS2004. One can't lift the tail of the taildragger AC in 'Century of Flight' with the prop blast. Tail surfaces only 'take hold' when TO speed gets to 30 kts or so. So, the tail control surfaces may be set with excessive 'authority' since there is no prop blast effect to help them. Since idling the engine doesn't decrease the low speed moments as it should, one doesn't have to move the Yoke/JS as much as he would in the real AC. Once below 20 kts or so, the taildraggers tend to become hard to control. I also noted a lesser effect in the C182 at initial TO roll in FS9 compared to FS2K2. The prop blast also has a significant effect at low speeds with tricycle gear, with it gone they also suffer. The FS implementation of FFB is limited. I think they used a FFB approach contracted by NASA and released to the public in 1996. Jerry Beckwidth discovered some the parameters in the AIR file that affect FFB. Setting them more appropriately improved the effect. Regardless, I suspect available FFB JS' are inadequate for giving the smooth forces required. They are more applicable to games that jerk the JS around when you fire a machine gun. ;) However, I have asked a guy who did a DirectX JS interface to FS' XML variables to also look into the FFB variables. Figuring I might eventually calculate realistic forces to send to the JS. Which would include the effect of elevator trim on the JS. After all, that's how pilots trim real AC: by trimming for zero Yoke force. No need to look at the trim scale, unless it is way off from where it should be. I think I can account for wind turbulence and gusts as far as realistic FFB goes. >For me this is one of the biggest hurdles against getting>accurate feel and an accurate landing and slow flight>behaviour (sluggish controls with large movements at slow>speed.>>If there is no option in the FDE for this then to achieve a>realistic flight experience MS need to code it into FSX>Peter Yes, those tables have been in the AIR file for a long time. I hope MS doesn't hard code them in FSX as they have done with other important tables and records in the past. Such as the autopilot record, and to a lesser degree, the Mixture Table. Ron

RonMany thanks for the long description and explanation, lets hope :-)Peter

>Ron>>Many thanks for the long description and explanation, lets>hope :-)>Peter Actually, the effect of missing prop blast on the tail and how it might result in low JS movement at landing speeds just came to me as I wrote that message. ;) Even if others find my comments confusing I sometimes come up with new ideas when I'm babbling. Eventually I find time to check such ideas out in FS flight tests. Ron

Hello, everyone.I see that many opinions exist when it comes to Flight Modeling. Many real world pilots seem to think that it

I will agree that the default flight modeling needs improvement, but since the only aircraft I've flown are the Cessna 152 and 172 (I've only flown a few times in each), I really have no idea how a 747 feels from the control point of view.With that said, flight in my opinion, and I respect everyones opinion in this excellent discussion, should entail more than the handling of the aircraft. To compensate for lack of physical sensation, visual input concerning the simulated world that the aircraft is located should be modeled correctly as well . All pilots look outside the window of the aircraft and visually scan the skys, and many pilots fly because of what they see and feel, the sensation of flight.It is difficult to explain the actual sensory input one recieves when in flight, seeing a falcon soaring above you or thunderstorms forming, or the beauty of the Chesapeake bay (I live near it) in the distance. The more closely all aspects of flight are modelled from the aircraft flight dynamics to weather, from traffic on the roads to geese migrating south, the closer this simulator comes to reaching the full meaning of flight.Ian.

> That depends on available AC information and the ability of>the FM designer. >Wait a sec, isn't that what my whole post was about? > Wiring, etc. is not significant. It is accounted for in the>published CG for the AC. The other things can often be found>in the TCDS, or scaled from accurate drawings. The best>drawings show the Wings L.E., MAC, and other aerodynamic>distances and areas.I would disagree here, it depends on how accurate you want your model to be. You can scale, and generalize, or you can start getting absurd, down to the molecular structure of the materials. Anyhow, generalization leads to inaccuracies in the model, nobody can argue that. A lot of what you were saying has to do with generalizing certain parameters to get things "close enough", and "close enough" leads to strange flight characteristics under certain conditions.>Low production means high cost, regardless of quality. High >production can result in very good products at the lowest possible >cost. Of course, competition helps. ;)>>I used to design RF components for Government Contracts. What cost >us $500 to build would cost a few $ if mass produced. Further, the >mass produced product would have much better performance. >>I expect the same is true for a lot of software. Well, you would certainly think that they would have been able to invest more heavily in flight modeling over the last 20 years or so of this product series. I can understand why they'll never be perfect, but there are definitely things that can be improved. Their turbine model is definitely one, as you mentioned. But, is the payoff for them really that big? Investing that time improving the flight model over adding and improving other things like scenery, ATC etc. when their aircraft fly "good enough" maybe just doesn't make sense to them, I really don't know. Although the hardcore simmers would all certainly like them to devote a little more time to it. :-)

>> That depends on available AC information and the ability of>>the FM designer. >>Wait a sec, isn't that what my whole post was about? Pretty good data turns up many times. People have sent me engineering drawings, graphs, etc. for a variety of AC, from an AC520 to the Concorde. FM data is always available for commercially released Jet Transports since there are pilots, etc. who fly the AC in the development group. It is often available for under $50 otherwise. I found the Rotex torque curves, etc. at the Rotex site. The Rotex powerplant is used in the RQ-1 Predator. Dimensions and some data on the Predator is publicly available on the WEB. So, the Aspect Ratio is known. That leaves only two parameters to set for the top speed, fuel consumption, etc. One is Oswald's Efficiency, and I estimated something like 0.80. It's not that critical. The other is zero lift drag. I set that so the RQ-1 would hit the top speed mentioned at WEB sites. I found I could get 40 hours endurance at 'best endurance' IAS and a lean mixture. Later I ran into a comment on the WEB "The Predator has flown mission up to 28 hours, but can fly as long as 40 hours". Creating a Flight Model is somewhat of a jigsaw puzzle. Some pieces are fit together by matching colors, others by shape. What is left are fit by experiment. Unlike the jigsaw, one doesn't know if his FM is appropriate without a lot of testing. One can fit pieces together in different ways. And, the dynamics are still not completely known. One oddity about the RQ-1 is the inverted tail. That reverses the sign of several 'stability derivatives'. Which makes the AC easier to turn: the tail aids, rather than opposes turning. Something I looked for and noted in flight testing. Since the RQ-1 is remotely piloted, the lack of FFB, a wide view, etc. is potentially well simulated in FS. >> Wiring, etc. is not significant. It is accounted for in>the published CG for the AC. The other things can often be>found in the TCDS, or scaled from accurate drawings. The best>>drawings show the Wings L.E., MAC, and other aerodynamic>>distances and areas.>I would disagree here, it depends on how accurate you want>your model to be. You can scale, and generalize, or you can>start getting absurd, down to the molecular structure of the>materials. Wing area, S, is important, since it is a factor in most of the aerodynamic effects. Distances to the tail surfaces only affect FFB in MSFS. They actually have a direct effect on many of the AIR file parameters. 10% accuracy is generally plenty close for the latter. Center of Lift and CG are important. The CG limits in a light AC aircraft typically +/- three inches. Typically about the MAC, but due to the way it is specified, may be more like 12% to 28%. >Anyhow, generalization leads to inaccuracies in the model,>nobody can argue that. A lot of what you were saying has to>do with generalizing certain parameters to get things "close>enough", and "close enough" leads to strange flight>characteristics under certain conditions. By definition, 'Close Enough' is just what it says. It depends on what final accuracy is desired. If it takes digital test gauges to show a small effect that's not discernible with the 'realistic' panel instruments I'd say the criteria is met. Dynamics are more difficult; however basic aerodynamic theory and a variety of real pilot's reports go a long way. At least for the normal flight envelope. >Well, you would certainly think that they would have been able>to invest more heavily in flight modeling over the last 20>years or so of this product series. I can understand why>they'll never be perfect, but there are definitely things that>can be improved. Their turbine model is definitely one, as>you mentioned. But, is the payoff for them really that big? >Investing that time improving the flight model over adding and>improving other things like scenery, ATC etc. when their>aircraft fly "good enough" maybe just doesn't make sense to>them, I really don't know. Although the hardcore simmers>would all certainly like them to devote a little more time to>it. :-) The MSFS 'Flight Model' coding should be one of the simpler things in MSFS. It shouldn't have to change much in every new version. Some things have been improved, the trouble is that various things have been broken in the process. The 'turbine model' is pretty good. But, took us years to work out. ;) So is the prop engine model, though it could be enhanced in a few ways. So RPM affects SFC (I did that in an XML gauge), CHT depends on Mixture, and the windmilling drag bug is fixed. However, the flight/powerplant code in SIM1.DLL is only part of the problem. Even if it were versatile and perfect, that still leaves the problem of setting up the individual flight models. Regardless, there are people who have some idea of what is involved, and others who get very involved. In the end, an excellent way to learn aerodynamics. Note just Austin develops the core of X-Plane. He is always making changes, and also introduces many bugs. ;) 'Orbiter' is basically a one man project, yet an improved version is released about once a year. Significant bugs in a release are generally fixed in an update a few weeks after the initial release. Ron

>I assume you do not have a real-world pilot You assume many things - all utter rubbish. I do have PPL-IFR.>Hmm.. this is almost what I expected. No proof, no factsExactly what can be said about your "powerfiul" arguments.Michael J.http://www.precisionmanuals.com/images/forum/pmdg_744F.jpghttp://sales.hifisim.com/pub-download/asv6-banner-beta.jpg

Hi Michael,>You assume many things - all utter rubbish. You really seem to see yourself in a very desperate position, otherwise there would be no reason for you to use such weak insults.Come on, get out of your defensive corner, be a man an face the consequences of the things you say! I'm not interested to prove you wrong, but to have a constructive discussion with you.> I do have PPL-IFR.What planes do you have type ratings for exactly? Where did you make your flight license? How long have you been piloting real-world aircraft? What is the identification number on your pilots license?And more important: How comes that a real-world pilot is satisfied with the simplified MSFS flight model?>> Hmm.. this is almost what I expected. No proof, no facts>Exactly what can be said about your "powerfiul" arguments.I told from my personal experience, so it is a bit more difficult for me to prove what I said. I definitely think the postings of other people in this thread arguing similar to my position are proof enough.For you, it would be very easy to prove your position - if there were any statements about X-Plane vs. MSFS from pilots which share your position, you can quote them. RegardsPanther13

I have been using FS since year 2002. So not that long. Last April I took a flight in a Varga 2150A. My very first time at having any control of an aircraft. I taxiied the plane. I took off in the plane. Flew around, and I landed the plane. He adjusted flaps and fuel, I controlled stick and rudder for the major part of the flight, except when he wanted to show me something. The first time I landed it was a little flat (not enough flare). With some time still left, we took off again and did a traffic pattern back to the runway. This time I stalled the place about 10ft off the ground and the pilot was ecstatic. Yelling Yeeeehaawwwww when the stall sound went off. (He is an ex fighter pilot by the way). I learned this from Rod Machado in the pilot lessons of FS2002. Point being, is that nothing during the flight was unexpected. My time with FS really paid off. The pilot also told me it was one of the most fun times he ever had with a first timer because, and I quote, "You knew what you were doing up there". He hated to hear that I cant get a PPL because of health issues. As I said earlier in the post, my only real beef with the FM in FS is the way the aircraft reacts to turbulence. Would love to see and "feel" the burble affect during high alt stalls as well. Other than that, seems pretty darn good from my brief experience in the sky. At least with little planes. Scott

>>I and couple of individuals from the MSFS team have a theory>as to why the more experienced pilots feel that the MSFS>default aircraft are quite accurate in the flight modeling.<>>I would like to know why? too :-) I have also in the past>placed non simmer 18000 hr pilots in FS with the usual effect>of them overcontrolling and shooting wildly over the sky.>>The usual complaint was no feel. Rob Young used to refer it as>"a puppet on a string " effect.>And that's exactly what it is.Without the feedback of gravity on your butt and the peripheral vision and depth perception offered by the real world (the latter slowly improving as scenery gets better and better) you tend to overcompensate.Therefore even if the control input is exactly on the mark the user tends to deflect his control more than strictly needed and more than he would in the real world.>Modellers have since then damped down the control response>making the models much more handflyable rather than making the>tiniest of tiny movements specific to flying the sim.>See my comment above ;)>Whether there is a half way house way of achieving that or a>cruder form which wouldnt involve the problems he talked>about?>One option might be a active control device with a non-linear response.Whether such a device exists and is reliable for a variety of platforms (maybe an active programmable damper?) I don't know.Would certainly make for an interesting piece of hardware, and likely a pretty expensive piece of hardware as well.I'm not talking FFB here, but a device that would be programmed using the speed and altitude data retrieved from FS (for example, some common interface might be needed to make it applicable to other simulators) combined with aircraft profile data to translate the actual force applied by the user into a desired control input for the simulator.It could then also apply dampening on the control deflection so the control itself would deflect more like that in a real aircraft and require more force to achieve the same amount of deflection at higher speeds (while yielding greater response for that force).That would solve the problem people now have who have flown real aircraft that the control input (in degrees of motion of the yoke) needed in FS is far larger than in the real thing while at the same time the control input (in Newtons of force) is the same or less.But don't look at me to design and program that thing, I don't have the necessary skills in electronics and mechanical engineering ;)

>I and couple of individuals from the MSFS team have a theory as to why the more experienced pilots feel that the MSFS default aircraft are quite accurate in the flight modeling.<Peter,What we speculated on was that the highly experienced pilots, such as the lead test pilots from Boeing, was that the default 777's flight modeling was close enough to the real thing, that the missing gaps (so to speak) were mentally (sub-consciously) filled in by the test pilot's own past experience. Rather the default 737/747/777's flight modeling was perfectly accurate or not was irrelevant, as their own extensive experience with the real thing would compensate for any weakness in the FS9 default jetliners.Bear!

And of course never forgetting the target market for the product, which is decidedly different from the target market for say PMDG or PSS.If the default 777 were to take half an hour of checklists to get the engine started and were to be less forgiving in its flight modelling than it is (not saying the real one isn't, just an example) the average customer for MSFS would be sorely disappointed because he would be quite incapable of enjoying his experience.