Edit aircraft flight dynamics in FS2004 safely using aircraft.cfg and AIR files, with the right tools, test method and fixes for ignored changes.
To edit an aircraft’s flight dynamics in FS2004, back up its folder, change the relevant values in aircraft.cfg, and use an FS2004-compatible AIR-file editor for aerodynamic tables stored in the .air file. Adjust one variable at a time, reload the aircraft, and test under identical fuel, weight and weather conditions.
Which files control FS2004 flight dynamics?
FS2004 divides the flight model between the aircraft’s configuration file and a binary AIR file. Most practical tuning starts in aircraft.cfg; deeper changes to lift, drag and moment curves may require editing the AIR file.
| File or section | What it controls | Typical use |
|---|---|---|
aircraft.cfg | Geometry, mass, centre of gravity, control scalars, engines, propellers and landing gear | Routine performance and handling adjustments |
.air file | Aerodynamic tables and coefficients not exposed directly in the configuration file | Stall behaviour, lift and drag curves, pitching moments and detailed stability work |
.mdl file | The visible model and its animations | Not normally used to tune flight dynamics |
The sim= entry under each [fltsim.x] section identifies the AIR file. For example, sim=MyAircraft points to MyAircraft.air in that aircraft’s folder. A mistake we see constantly is editing an AIR file whose name is not the one referenced by sim=.
How do I edit and test the flight model safely?
The safest method is to work on a separate test copy and make one measured change per test flight.
- Copy the complete aircraft folder. FS2004 normally stores aircraft beneath its
Aircraftdirectory. Give everytitle=entry in the copiedaircraft.cfga unique test name so the original and modified versions remain distinguishable. Our backup and validation checklist for aircraft.cfg edits covers the precautions in more detail. - Record a baseline. Use the same fuel, payload, centre of gravity, weather, realism settings and aircraft configuration on every run. Record measurable results such as stall speed, climb time through a fixed altitude band, stabilised cruise speed and fuel flow.
- Identify the active AIR file. Read the relevant
sim=line before opening anything. Do not rename an AIR file unless you also update that line. - Edit aircraft.cfg with a plain-text editor. Preserve section names and parameter spelling, use a full stop for decimal values, and follow the units defined by the FS2004 SDK. Units vary by parameter, so an unchecked conversion can produce a flyable but badly distorted model.
- Use an aircraft editor for the AIR file. Never open a binary
.airfile in a text editor. Microsoft’s FSEdit utility is available in the complete FS2004 SDK bundle containing the official aircraft tools. Keep the untouched AIR file alongside a separate written change log. - Force FS2004 to reload the aircraft. Select another aircraft and then reselect the test model. Restart FS2004 if a change still appears to be cached, especially after AIR-file work.
- Test only the changed behaviour. Revert unsuccessful changes rather than compensating for them with several unrelated scalars. Once multiple variables are altered together, it becomes difficult to tell which one caused the result.
What flight-dynamics values should I change first?
Start with physical data, then tune coefficients and scalars only after the aircraft’s geometry, mass and engine figures are credible.
| Problem | Check first | Common wrong approach |
|---|---|---|
| Controls are too weak or sensitive | Control effectiveness values in [flight_tuning], control geometry and centre of gravity | Changing wing dimensions to alter control response |
| Aircraft oscillates or feels unnaturally stable | Pitch, roll and yaw stability scalars, moments of inertia and CG position | Testing with the autopilot engaged or altering every stability axis together |
| Stall speed or break is wrong | Gross weight, wing area and the AIR-file lift, drag and pitching-moment curves | Changing only [reference speeds]; those entries do not rebuild the aerodynamic stall |
| Cruise or maximum speed is wrong | Actual engine output, propeller data, parasite drag and induced drag | Editing a displayed maximum-speed reference instead of balancing power against drag |
| Climb performance is wrong | Weight, engine or propeller output, drag and the airspeed used during the test | Increasing lift alone, which may damage stall and cruise behaviour |
| Aircraft sits badly or bounces on the runway | [contact_points], static pitch, static height and CG | Using aerodynamic scalars to mask landing-gear geometry; follow our procedure for correcting FS2004 contact points instead |
The main configuration areas are [airplane_geometry], [flight_tuning], [weight_and_balance], [GeneralEngineData], the engine-type-specific sections, [propeller] and [contact_points]. Parameters in related areas can interact, so increasing thrust to fix climb may also create an excessive cruise speed.
For coefficient tables and the reasoning behind their interactions, use our technical tutorial on tuning FS2004 aerodynamic behaviour. AIR-table changes should come after the aircraft’s basic dimensions, weights and powerplant data have been corrected.
Can I copy flight dynamics from another aircraft?
Copying a complete flight model works only when the donor has closely comparable geometry, mass distribution, engine type and performance. Borrowing unrelated AIR files often produces plausible cruise performance but incorrect stalls, trim, inertia or control response.
If you use a similar aircraft as a starting point, transfer and validate one system at a time. Never assume that matching wingspan or engine power makes two flight models interchangeable.
Why are my FS2004 flight-dynamics edits being ignored?
Ignored changes usually mean FS2004 loaded a different file, retained the aircraft in memory or encountered a malformed parameter.
- Wrong AIR file: confirm the filename against the active
sim=entry. - Wrong aircraft copy: duplicate display names can make the original and test versions difficult to distinguish in the aircraft menu.
- Shared repaint: several
[fltsim.x]liveries in one folder normally share the same configuration and AIR file, so a dynamics change affects all of them. - Invalid syntax or units: a misspelt key, misplaced entry or incorrect unit may be ignored or produce misleading results without a clear error message.
- Aircraft still cached: load another model or restart the simulator.
- Custom gauge logic: some complex add-ons use gauges to influence engines or controls, so configuration changes may be partly masked while those systems are operating.
How do I judge whether the edited model is accurate?
Judge the model against repeatable performance points rather than how lively it feels during one flight. Compare stall speed at the same weight and configuration, climb at the same indicated airspeed and power, and cruise after speed and altitude have stabilised.
Keep indicated and true airspeed separate when comparing results. Test handling at several speeds as well: a model tuned to feel correct on final approach can become excessively sensitive in a fast descent. FS2004’s legacy aerodynamic, propeller and ground-friction models also have limits, so prioritise consistent behaviour across the operating envelope rather than forcing one published figure to match at the expense of everything else.