P3D v4 Worst Autopilot Performance

#1
I have been working with my Autopilot, trying to boost its performance. Nothing is working.

I tried doubling elevator trim effectiveness, that did little. Doesnt AP use the elevators though for ALT adjustments?

The main issue is ALT adjustment... This thing is so dang slow to react. When I fast forward, she starts doing the up/down thing like crazy. She cant seem to lock onto the altitude she is set to. I have tried the Cessna 172 AP section, the Beech King Air (yep). The F-16 AP section is my latest (from V4) to try... That too just wont cut the mustard.

The airfiles or flight dynamics files have the 'attitude indicators' and 'directional indicators' in it.

[altimeters]
altimeter.0=1
altimeter.1=1

[airspeed_indicators]
airspeed_indicator.0=1
airspeed_indicator.1=1

[direction_indicators]
//1 Type: 0=None, 1=Vacuum Gyro, 2=Electric Gyro, 3=Electro-Mag Slaved Compass, 4=Slaved to another
//2 Indicator number of slaving indicator if Type = 4
direction_indicator.0=2,0

[attitude_indicators]
//Type: 0=None, 1=Vacuum Gyro, 2=Electric Gyro
attitude_indicator.0 = 2
attitude_indicator.1 = 2

[turn_indicators]
//Type: 0=None, 1=Electric Gyro, 2= Vacuum Gyro
turn_indicator.0=1,1 //Turn AND Bank

[vacuum_system]
max_pressure=5.15 //PSI
vacuum_type=1 //Type: 0=None, 1=Engine pump (default), 2=Pneumatic, 3=Venturi
engine_map=1

[pneumatic_system]
max_pressure=0.000000
bleed_air_scalar=0.000000

[pitot_static]
vertical_speed_time_constant = 1 //Increasing this value will cause a more instantaneous reaction in the VSI
pitot_heat = 1.0 //Scalar on heat effectiveness, 0 = not available

This is the autopilot section; /////////////////////////////////////////////////////////////////////

[autopilot]
autopilot_available= 1
flight_director_available= 1
default_vertical_speed= 1200.0
autothrottle_available= 1
pitch_takeoff_ga=8.0
default_pitch_mode=1 // 0 = None 1 = Pitch Hold current pitch angle 2 = Altitude Hold current altitude
max_pitch=10.0
max_pitch_acceleration=1.0
max_pitch_velocity_lo_alt=2.0
max_pitch_velocity_hi_alt=1.5
max_pitch_velocity_lo_alt_breakpoint=20000.0
max_pitch_velocity_hi_alt_breakpoint=28000.0
max_bank=25.0
max_bank_acceleration=1.8
max_bank_velocity=3.00
max_throttle_rate=0.10
nav_proportional_control=9.00
nav_integrator_control=0.25
nav_derivative_control=0.00
nav_integrator_boundary=2.50
nav_derivative_boundary=0.00
gs_proportional_control=9.52
gs_integrator_control=0.26
gs_derivative_control=0.00
gs_integrator_boundary=0.70
gs_derivative_boundary=0.00
yaw_damper_gain = 1.0
yaw_slip_control_gain=0.000000
yaw_control_uses_trim=0.000000
attitude_indicator =0
default_bank_mode=2 // 0 = None 1 = Wing Level Hold 2 = Heading Hold current heading


Any ideas why my baby is not doing the 'automatics' very well?
 
#2
This is how bad... I turn on ALT Hold and it takes like a solid minute for her to start pulling up... I'll lose a LOT of altitude (getting close to the ground) before she starts pulling up.

Something must be wrong.

I double checked my elevator inputs. Nothing wrong there. Doubled elevator trim effectiveness. No joy.
 
#3
That's difficult to solve since we don't know how your air file and the cfg file look like. The AP uses only trim for pitch control.
You can try to increase, max_pitch_acceleration and max velocity by a factor of 5 for troubleshooting.
What does the elevator trim gauge indicate when you are trying to do an altitude change?
Maybe the trim range is too low and it's already at its limit ? Trying doubling this number [airplane_geometry] elevator_trim_limit
 

Roy Holmes

Resource contributor
#5
Bill,
Which airplane are you talking about here?
Usually the autopilot works just fine.
If it is not working well and there are no obvious mistake in the autopilot section, it is most likely the aerodynamics of the subject airplane.
All things vertical are done by the elevator trim. Against them are a load of flight dynamics like MOI, pitch characteristics etc.
Compare this one to another of your models.If things are different and the autopilot is essentially the same that should point you at where the problem lies.
Roy
 
#6
Bill,
Which airplane are you talking about here?
Usually the autopilot works just fine.
If it is not working well and there are no obvious mistake in the autopilot section, it is most likely the aerodynamics of the subject airplane.
All things vertical are done by the elevator trim. Against them are a load of flight dynamics like MOI, pitch characteristics etc.
Compare this one to another of your models.If things are different and the autopilot is essentially the same that should point you at where the problem lies.
Roy
Hey Roy,

That was my thoughts also. Short tail on this lifting body design, but oddly it was originally set for the distance of the Extra 300. I brought it in 3 feet where it should be, and still didnt work that well.

Thats when I thought it was time to ask questions.
 
#7
That's difficult to solve since we don't know how your air file and the cfg file look like. The AP uses only trim for pitch control.
You can try to increase, max_pitch_acceleration and max velocity by a factor of 5 for troubleshooting.
What does the elevator trim gauge indicate when you are trying to do an altitude change?
Maybe the trim range is too low and it's already at its limit ? Trying doubling this number [airplane_geometry] elevator_trim_limit

That helped quite a bit, BStolle. A thousand thanks. Going to experiment with it some more. A thousand thanks.
 
#8
Ive been experimenting and experimenting. I have a pitch issue in the planes dynamics. Most of the flight model (air files?) is duplicate of the Extra 300 from FSX. My wingspan is about the same. Elevator (and invisible rudder) are shorter though, 11 feet instead of 14 feet. But.............. I have a lifting body. My wing root is about 14 feet. That might be whats going on.

I remember doing a small wingspan and at one point, a crash occurs in the system and the plane suddenly becomes nearly impossible to land, jumping into the air when it so much as 'touches' the ground. If you put a few inches back on the wingspan, its fine. Take that out, short wing chord, and you crash the flight dynamics for landings.

I try to fly the plane with autopilot on at fast forward (for lack of a better term) x2, x4, x8, and she is diving down and up. At x8 its getting scary. At x16 you can hit the ground in the second hop. I tried flying it manually (over and over) and its nearly impossible. So I opened the Extra 300 (I imported it into V4 from FSX) and its flyable at x8 (fast forward).

I didnt try moving the elevators back to 14 feet in the flight dynamics/geometry. I want 'some' parts to be real world, at least. Not all... This should work. I know on the real F16, when they first flew it, its pitch was CRAZY!!! The test pilot barely lived through that. All you need to do is do real world tuning on the controls, etc. I have tried and tried mixing settings with the 'system' on 'trim effectiveness' and 'elevator effectiveness', also pitch stability. Nothing seems to work to get her stable.

So... I think I just have to deal with the cards I have. The autopilot system is ok.. it works on 'real time' and x2, but in x4 on up, its a roller coaster. I think that FS/P3D just do not like my lifting body with a giant wing root. I like that she has good elevator control. The system perimeter and outline, target design, is for a STOL plane, a futuristic bush plane, something that can land in very short runways, do flat turns, fly slow, and fly fast. I have that. I just have a roller coaster when it comes to cruise and high cruise with pitch instability. I'll live with that. At least for now.

Pitch dampers... I need pitch dampers, dangit! I need six pitch damper control computers with liquid helium cooling systems so they do not overheat!
 
#10
1. I have a lifting body. My wing root is about 14 feet. That might be whats going on.
2. I didnt try moving the elevators back to 14 feet in the flight dynamics/geometry. I want 'some' parts to be real world, at least. Not all... This should work.
3. I know on the real F16, when they first flew it, its pitch was CRAZY!!! The test pilot barely lived through that.
4. All you need to do is do real world tuning on the controls, etc. I have tried and tried mixing settings with the 'system' on 'trim effectiveness' and 'elevator effectiveness', also pitch stability. Nothing seems to work to get her stable.
1. A 14ft wing root most likely creates a problem if you don't adjust other important parameters accordingly.
2. Why do you want 'some' parts real? That's the beauty of the lookup table FDE design ;)
3. It was the way too high roll sensitivity which created the problem, not pitch.
4. Try to (at least) double the Cm_q pitch_moment pitch rate (damping) in section 1101 and/or move the CG forward by 1ft. Maybe this helps. (If you increase pitch damping, you need to increase Cm_de elevator control)
5. Btw, this lifting body does look cool!
 
#11
Make sure you use the program Visualizer to setup the contact points of wheel and wings, ect... This will prevent strange crashing issues. Also the answer was right to adjust the elevator_trim_limit so the autopilot has enough bite for the AP pitch control. Looks Like a nice plane. I can see the short body being an issue. I would make sure you set the root chord extremely wide. So the short wing can still create lift. For the tail section i would still create one that is not realistic to your airplane. Just make it small and near the back.
 

Roy Holmes

Resource contributor
#12
Changing the physical dimensions of an airplane in FS other than wing area and wing span has no effect on how it flies. In the real world it would affect the aerodynamics as measured by the aerodynamic coefficients but these coefficients reside in the .air file and are unaffected by such things as the size or distance of a tailplane from the wing.
You could give it the dimensions of a brick shape and it would not change how it flies.
Roy
 
#14
Making changes to this, matching my 1199 section to the Lear45 (I know, bigger and faster bird, but it adjusts to settings so crisply), and I still get over-shoots in ALT Hold of 150 feet... up and down, up and down... up.. and dang down...
 
#20
Changing the physical dimensions of an airplane in FS other than wing area and wing span has no effect on how it flies. In the real world it would affect the aerodynamics as measured by the aerodynamic coefficients but these coefficients reside in the .air file and are unaffected by such things as the size or distance of a tailplane from the wing.
You could give it the dimensions of a brick shape and it would not change how it flies.
Roy
Increasing the size of the elevator size would have no effect? Also when you adjust htail_incidence this effects the pitch and is used to calibrate the pitch of an aircraft. So to say the size and distance from the wing make no difference is new to me. I will try an test today to see what happens if the tail is too close to the wing.

Update: The distance and size of the tail in fact does not matter. Wow!:eek: Why do they ask for this info? I always created the correct info and will continue to but had not idea this was ignored.
 
Last edited:
Top