Why do some planes have "bad/wrong" flightmodel?

[Nova]

I read in different reviews, of different aircrafts, that some testers claim,

that the tested aircraft has bad/wrong/wierd flight dynamics.

But how can that be? Isn't X-Plane calculation the whole airflow around the

aircraft multiple times per second?

If someone has enough data of the aircrafts' shape, then everything should

be right. And with right engine data everything should be 99% to the real

thing, shouldn't it?

So how can a tester state, that things are wrong?

[FlorianR]

Well, the planemaker model and airfoil design may be off. But good question, I'd love to hear some developers answers!

[Goran_M]

I've answered this question to many people who have asked it.  

Here's the honest answer.

Most add ons (payware included) are inaccurate.  Often SEVERELY inaccurate.  I think it's the freeware community that started the story about how easy it is to make a flight model in Planemaker.  Other freeware developers who call themselves professional, claim to be able to make a flight model in a matter of 2-3 days.

The truth is, a LOT of information is needed from the manufacturer to build an accurate flight model.  A person cannot just go to the FAA website, download a TCDS and expect to make an add on for X-Plane.  You need manuals and to get those, you have to spend money.  Not just Aircraft manuals, but also airfoil manuals, FCOMS, etc...and these will usually run into the hundreds of dollars.  

I won't mention any names (although I think we all know who I am talking about...) but the Planemaker tutorial videos made by another developer, showing how to make a flight model in Planemaker is only the tip of the iceberg.  He covers about 10% of what is really involved.  What he has done, I'm sure, has started many people on the path to add on development, and he has provided a stepping stone for would be developers, but there is a lot more to learn.

Planemaker is extremely flexible and accurate.  But it will only make WHAT YOU TELL IT TO MAKE.  It doesn't make flight models on it's own or with just a few numbers thrown in there.  If you're only adding information in the basic areas, you're not doing anywhere near enough.  I spent about a year (roughly 3 -4 days a week) on the Saab flight model and then that was further refined by Lukasz for approximately 2 more months, adjusting certain figures to match the flight envelope (Flight envelope charts are available in the official manuals.), testing, re-tuning, re-testing...several times over until we hit the numbers in all areas.

If you want to make flight models, please, feel free to do so.  But do not expect Planemaker to do most of the work for you.  The more information you give it, the more accurate the flight model.  If you leave default values, you will be disappointed if you are expecting accuracy.

[karingka]

Some manufacturers, like Embraer, dont share airfoil data. I asked, they denied. No embraer ever made for x-plane will have an accurate flight model.

[YYZatcboy]

That is not true. Airfoil data can be interpolated from performance data, but it is very difficult, and it takes a LOT of time and information not only on performance, but also on all of the other systems that affect the flight model.

[Pete_SMS]

Some manufacturers, like Embraer, dont share airfoil data. I asked, they denied. No embraer ever made for x-plane will have an accurate flight model.

I don't agree with that. Not having the airfoil data does not mean that you can not create an accurate flight model.

If you have performance data for the aircraft from a flight manual, information about the flight characteristics, stall speeds, V-n diagrams etc. you can come up with a pretty accurate flight model by reverse engineering. However, you have to have the knowledge and background to be able to do that and it takes a good amount of time.

Even if you have the airfoil data, you have to become creative here and there to get it right, because X-Plane is no virtual wind tunnel and not everything is perfectly OK.

[karingka]

Yeah. It would help to have the data, without it, it would take a long time...

[tkyler]

The shape of the aircraft is only one aspect of what is called "the flight model".  There are many many other fixed parameters like weight, Center of gravity, balance, control sensitivity that can be adjusted to tweak the way the aircraft feels...in addition to the airfoils and angle of incidence etc.  If any one of these are off, then they can be "felt" by the end user.   Say the CG is off by a few inches, the plane my pitch up so that it always wants to climb.   User's would know this was "not right" very quickly.   Another big complaint is that aircraft feel "too twitchy".  Because of the lack of any real muscle effort, the perception of "effort" doesn't really exist.  so then, an author has to make adjustments that compensate for the fact that you're sitting at a computer.  You can do this to "fool the brain" even though the numbers you input may be right.   In addition, many of the numbers used by Austin are input into "approximations" as it is very difficult to be all things to all aircraft.   If you don't understand his approximations or assumptions and blindly trust him, then you will probably get performance that is a little bit off.  

In general...if authors would simply input manual radii of gyration, tweak the CG,  adjust the control falloff speeds as well as "control delay"..then a good 90% of bad models could be made to "feel good"; however, these are not parameters that most people think of when they think of aircraft design, hence they're missed very often.

[Warmbrak]

Very informative feedback in this thread and should belong any X-Plane wiki. To summarize what has been said, developing aircraft for X-Plane is a "garbage in- garbage out" affair. This has given X-Plane a bad reputation in some circles as many people believe that X-Plane just gets it wrong. As one can see from above that is far from the truth.

The smoking gun is Planemaker, and I am not criticizing it. Quite the opposite, Planemaker is relatively easy to use for newcomers and it is not difficult at all to produce an aircraft that looks like the real thing and get it to fly. Having used MS Flightsim for many years up till this day I still don't know how to get a model flying. Back when I bought X-Plane 5.40 it wasn't a week before I started fiddling around in Planemaker and making changes to aircraft to see how they react. However, as some of the great developers on this thread have noted, it takes more to get that aircraft to fly like it does in real life. I have a project that I have been starting and stopping over the past 5 years because I can't make the investment in time to get to learn the intricacies of flight modelling, and a bunch of other things. I take my hat off to those that persevere and make it happen.

There are thousands of Swiss army knives out there, but not that many McGuyvers 

[Pete_SMS]

Fully agree with Tom. Shape and airfoils are only pieces of the puzzle for a good flight model, and more or less only affect aircraft performance. Of course they have an effect on other areas, such as stability as well, but in order to get the "feel" right, one has to do much much more.

Tom mentioned custom radii of gyration or cg location. This is very important to work at. I have seen freeware and payware aircraft, where these areas are left at default or just in the right ballpark, making the aircraft handle like a radio controlled aircraft. High roll rates, almost no lagging motion/intertia in roll, pitch or yaw, killing all the fun.

Not placing and tuning the cg correctly can also create very weird and unrealistic behavior. For example, fly at a low speed and yank back the stick. Of course you wouldn't do that in the real aircraft, but I use this test often in order to see if the control power is realistic. Flying at low speed, close to stall speed, and yanking back the stick often results in AOAs of 40-50 deg or more, which is completely unrealistic. For some aircraft that might be correct, for most it is not. In real live, the aircraft elevator would not be in the position to create such a pitch up motion at low speed, due to low dynamic pressure and loss of control effectiveness at higher AOAs. This is often due to a wrong cg placement, a wrong pitch moment coefficient of the airfoil or a combination of both, but could be easily corrected.

From my point of view, Planemaker is doing a pretty good job, but does not necessarily give you all the control you need for specific parts of the envelope, or if you really want to represent the personality of an aircraft with all its pros and cons. In order to go the extra mile, one needs to take care of it by coding plug-ins, which give much more control. However, you have to know what you are doing and you have to have the knowlege about what needs to be changed, tweaked or controlled dynamically via a plug-in, in order to create a certain effect or a certain behavior . For my aircraft, I am overriding the complete flight controls and big parts of the engine by a plug-in and I can basically re-create any effect or behavior I want to see or I want to be simulated. I can change control surface effectiveness, supercharger behavior, control surface buffeting, engine gyroscopic precession, engine corkscrewing slipstream, transsonic effects and much more on the fly and exactly how I want it. I more or less have full control.

Does it sound cool? Yes!

Is it easy to do? No. It's a lot of extra work, needs proper testing, creative out of the box thinking and hours and hours of research and coding.

...but it is also no black magic. Anybody could learn how to do it.