P3D v3 Single Engine Fail - Emulate with what Values?

[ronh]

Has anyone modeled a GA twin and emulate the flight characteristics when a single engine has failed - no feather/windmilling and feathered. The twins I have fly nicely with both engines, but if I shut one down and feather the "real" aircraft has a +100 to 200 ft/per altitude gain (up to 8000 PA) but the ones I have are zero to minus 100 ft per minute decent on one engine feathered, shutdown, aircraft banked 3 degrees to good engine and 1/2 ball.

What air file or cfg file parameters would you change to modify for single engine failure - Can it be done?

 

[Heretic]

Prop feathering was recently discussed here:
http://fsdeveloper.com/forum/threads/prop-feathering.436888/

Since I doubt that LM improved prop handling in P3D's flight dynamics, the findings in that thread wil very likely still apply.

 

[Roy Holmes]

The basic question is whether or not the shut down engine is creating negative prop thrust. If it was just at idle it's prop thrust would be negative. If shut down and feathered it should have zero or close to zero prop thrust.
If it has zero prop thrust and the aircraft is not matching the single engine performance in its Manual, then it is quite likely that it is not matching the all engines performance as well.
Prop thrust is of course A: prop thrust: ,pounds.
I found that prop thrust goes to zero in the sim if you shut off the fuel and feathering makes no difference
Roy

 

[ronh]

Thanks Bjoern, I have been looking at that thread, but the discussion was more on animation.

Thanks Roy, I have to go back and verify that the shutdown engine - windmilling prop is actually causing drag and then I should get a "performance" change/increase when feathered. I will double check if fuel shut off makes any difference to performance. Just wondering why there would be a beta_feather in the aircraft config if it did not have a function when the props are feathered. - Like thrust/drag = 0

It's the DA-42 - So prop feathering is Master Off and shutdown is Master Off, Alternator Off and Fuel Off. I will have to see what combinations do what to the aircraft performance.

 

[ronh]

Ok tested P3D v3. I need to shut off the fuel also to get the props to feather and stop windmilling. Not what I expected. Feather should be at Master Off. Still loosing altitude, so I need to look at things - thanks for the info.

 

[Heretic]

Thanks Bjoern, I have been looking at that thread, but the discussion was more on animation.

Animation driven by the FDE.

Do you have access to the model source or at least the "master feather" L: var used for the button?
If so, you can write your own code that automatically shuts off every other realted system to make the prop feather properly.

 

[ronh]

Animation driven by the FDE.

Do you have access to the model source or at least the "master feather" L: var used for the button?
If so, you can write your own code that automatically shuts off every other realted system to make the prop feather properly.

Unfortunately no. It's a payware aircraft. My only issue is that the aircraft descends at -100 ft/min when it should gain altitude at +100 to 200 ft/min on a single engine. 82 knots. I was hoping to tweak the aircraft.cfg or air file. I have emulated the shutdown/failed engine via simconnect, I get the yaw and roll, but not the proper performance on a single engine. (didn't really expect it, but would be nice since P3D is suppose to be used in training simulations, and that would include engine failure.)

 

[Heretic]

Isn't this more a general engine performance issue if an engine delivers too little horsepower to climb?
Changing engine performance might, however, result in a lot of FDE work to get the plane to fly by the numbers again.


Btw: If the model is FSX native, you can use MCX to generate a Modeldef file for it, which exposes all custom animation, visibility and mouse rectangle tags incuding their L: vars. I've used these funtions quite a bit for aftermarket modifications to third party aircraft.

 

[Roy Holmes]

Ron,
All twins have a single engine ceiling that varies with weight and power setting.
Below the ceiling the aircraft should maintain altitude.
If it loses altitude with full power and the correct RPM and weight and altitude your model is incorrect. Most likely it has too much drag. That is what I meant by "all engine" performance being wrong.
Do you have detailed performance data?
At what weight and altitude can you not maintain height?
Roy

 

[ronh]

The DA-42 has a single engine ceiling of 8000 ft at 1700 kg The sea level performance is 175 ft/min climb at 1700kg and then down to zero at 8000 PA, 1700 kg all at 15 degrees C. I am getting a 100 ft/min decent at 2000 ft MSL, should be 150 ft/min climb. Vyse 82 knots. I'm going to try different weights and see if there is a difference. I need to set up a base line and test more configuration with both engines.

Thanks for your help and suggestions.

 

[jx_]

Roy and Ron,

Don't neglect how common it is for thrust to be off on the good engine in these airfiles!

Try setting power to maintain the target 175fpm climb... notate that higher power setting and do a level flight segment at that higher power setting with both engines on and see if you are faster than you should be.

If drag is too high, you will likely have a higher than normal airspeed with both engines at the higher power setting

If thrust is too low, you will likely have a more accurate airspeed with both engines at the higher power setting

If both are off a little you will have to tweak one of the forces, drag or thrust, in a known wrong direction and map it out the hard way.....unless you can find a drag polar or thrust table!



Another thing to look at is induced drag. Many FS models fly with too much CDi and that gets magnified in a climb on one engine.



and of course give your control drag a once over. The cross control (Rudder opposing thrust & aileron opposing rudder roll) adds drag to the airplane. FS doesn't directly allow editing these values, so you will have to check the slip angle and rudder rolling force. If the hStab isn't producing enough weathervane force (too much yaw angle at a given thrust setting) or the rudder is producing too much roll (the aileron lift needed to counter increases induced drag), the sim will produce too much drag.


and if yaw angle is too high you're losing thrust there as well (thrust is pushing sideways instead of back)

 

[ronh]

Ok, Thanks for all the help and suggestions, I just played with the aircraft.cfg values and have a much better single engine performance and better dual engine performance under different configurations.

I played with:

[flight_tuning]
cruise_lift_scalar = 1.30 // original 1.00
parasite_drag_scalar = 0.90 // original 1.00

[piston_engine]
power_scalar = 1.10 // original 1.00

[propeller]
thrust_scalar = 1.20 // original 1.00

 

[TurboCompound]

If the model is FSX native, you can use MCX to generate a Modeldef file for it, which exposes all custom animation, visibility and mouse rectangle tags incuding their L: vars. I've used these funtions quite a bit for aftermarket modifications to third party aircraft.

I've been trying to find how to do this so that I can figure out which vars are triggered by mouse events and use that info to create 2d pop-ups for certain vc-only aircraft. Unfortunately, I cannot find this information in the manual. What MCX wizard/tool would I use to generate the modeldef?

 

[Heretic]

I've been trying to find how to do this so that I can figure out which vars are triggered by mouse events and use that info to create 2d pop-ups for certain vc-only aircraft. Unfortunately, I cannot find this information in the manual. What MCX wizard/tool would I use to generate the modeldef?

Intuition, logic and willingness to experiment. And clicking the icon with the "Modeldef.xml" tooltip in MCX.

 

[TurboCompound]

Intuition, logic and willingness to experiment. And clicking the icon with the "Modeldef.xml" tooltip in MCX.

Turns out I was using an outdated version. Now I finally got the 2d switches working. Why developers would put system logic into the <mouse> code is beyond me.