Maximum surface wind

[JeeHell]

Hello,

I have just ran into an issue with wind in P3D v4.5 . I wanted to test my aircraft max demonstrated X-wind (A320 - 38kts).
When I set a surface wind above 31 kts, the resulting wind is almost non existent until liftoff, when it kicks in suddenly.
It seems (not 100% tested) this issue is around since before v4.

Is that a known issue? Is there a work around like a setting in any cfg file?

Thanks
Jean Luc

[JeeHell]

Ok so it affects P3D v5 as well.
The AIRCRAFT WIND X components is "limited" to less than 30kts. If it reaches 30kts, it "overflows" to 0kts. And the aerodynamic engine uses that truly under evaluated component. So actually 50 kts x-wind is much easier than 29kts...
I hope a fix is possible. As soon as the aircraft is in the air, the wind component is correct, the aircraft track is correct etc...

[Martyson]

Hello,

I have just ran into an issue with wind in P3D v4.5 . I wanted to test my aircraft max demonstrated X-wind (A320 - 38kts).
When I set a surface wind above 31 kts, the resulting wind is almost non existent until liftoff, when it kicks in suddenly.
It seems (not 100% tested) this issue is around since before v4.

Is that a known issue? Is there a work around like a setting in any cfg file?

Thanks
Jean Luc

@ JeeHell ,
Jean Luc

How are your testing?
What are you seeing (screenshot, video, etc.) ?

[JeeHell]

Simple enough, take Visual Studio, write a simple simconnect addon and monitor the following variables:
AIRCRAFT WIND X
AIRCRAFT WIND Y
AIRCRAFT WIND Z
Now in P3D (all vanilla) change the surface layer wind (and remove other layers) to anything higher than 30 kts. Turn the aircraft while taxiing on the ground, and the result is obvious.
I do not need to post screenshots, P3D developers know how to reproduce that. I used only vanilla P3D, no addon at all.

[Martyson]

Simple enough, take Visual Studio, write a simple simconnect addon and monitor the following variables:
AIRCRAFT WIND X
AIRCRAFT WIND Y
AIRCRAFT WIND Z
Now in P3D (all vanilla) change the surface layer wind (and remove other layers) to anything higher than 30 kts. Turn the aircraft while taxiing on the ground, and the result is obvious.
I do not need to post screenshots, P3D developers know how to reproduce that. I used only vanilla P3D, no addon at all.

@ JeeHell ,
Jean Luc

How does the aircraft behave on takeoff?

[JeeHell]

The AIRCRAFT WIND X components is "limited" to less than 30kts. If it reaches 30kts, it "overflows" to 0kts. And the aerodynamic engine uses that truly under evaluated component. So actually 50 kts x-wind is much easier than 29kts...

That is on the ground. As soon as the aircraft lifts off, the normal wind returns and there is a sudden yaw moment into the wind.

[Martyson]

The AIRCRAFT WIND X components is "limited" to less than 30kts. If it reaches 30kts, it "overflows" to 0kts. And the aerodynamic engine uses that truly under evaluated component. So actually 50 kts x-wind is much easier than 29kts...

That is on the ground. As soon as the aircraft lifts off, the normal wind returns and there is a sudden yaw moment into the wind.

@ JeeHell ,
Jean Luc

A video of your takeoff showing that would be helpful.

What is the windsock showing?

[YvesG]

I confirm this bug described by JeeHell. It is a very old one dating from FSX or probably much earlier versions. I investigated it many years ago, but unfortunately it was never fixed in P3D. It is related to the general limitiation of surface winds in FSX/P3D.

General limit:
There is no cross wind at all until the aircraft on the ground accelerates to a groundspeed (GS) of 20 ft/s (11.8 kts). Then the cross wind components blends in linearly until the aircraft reaches a GS of 70 ft/s (41.4 kts) (delta 50 ft/s). Only above a GS of 70 ft/s the nominal cross wind is simulated. So you can taxi a Cessna in heavy cross wind without needing any rudder.
This is with the General Realism Slider at maximum. With a mid position it never reaches more than 50% of the nominal cross wind.

The reason for this limitation probably lies in the flight model. The yaw and roll moments caused by the sideslip angle increase linearly in FSX/P3D. This is not realistic. So, when the aircraft is at rest on the ground, the sideslip angle in a full cross wind would be 90°, causing almost uncontrollable high forces (aircraft flips and turns into the wind). With the implemented surface wind limit, FSX/P3D avoids high sideslip angles (it is hard to get above 16° or 17°).

Bug:
The cross wind component on the ground, A:RELATIVE WIND VELOCITY BODY Y (and derived parameters such as AIRCRAFT WIND), cannot exceed 50.00 ft/s (29.6 kts) on the first run. If it is set to a higher value, it starts counting from zero again. The values 20 and 50 are hardcoded in sim1.dll.

50 ft/s (29.6 kts) set = 50 ft/s (29.62 kts) experienced
but
51 ft/s (30.2 kts) set = 0.6 kts (0.3 kts) experienced
100 ft/s (59.2 kts) set = 50 ft/s (29.6 kts) experienced
150 ft/s (88.9 kts) set = 100 ft/s (59.2 kts) experienced --> in the second run it does not fall back to zero

Resulting issues:
Even at a cross wind component of less than 29.6 kts, such as 25 kts, your aircraft can be affected, because you normally will have to do some heading corrections in a strong cross wind. This turn can temporarly result in a relative wind velocity that exceeds the 29.6 kts limit and the cross wind falls back to near zero for a second. The aircraft breaks out and is almost uncontrollable.
The other issue is with the case described by JeeHell when the aircraft becomes airborne and the cross wind limitiation suddenly disappears.

Yves

[JeeHell]

Thank you Yves for the accurate description of the problem.
And also thank you for your huge work on aerodynamics applied to FSX/P3D. I am a big fan as Hervé S. can attest :)

Have you got a memory location in FSX's sim1.dll for the wind limits ? (Or a way to find it?)

Salutations de France
Jean Luc

[YvesG]

Thank you, Jean Luc. When I started to investigate the FS flight model in FS95/98 I never thought that 90% of the model still applies in 2021;-).

I can provide the wind limit offsets in sim1.dll for FSX SP2. I always changed them in P3D too up to v3.1, but then it became more and more difficult to identify them. There are now too many occurencies of these values to test them all. Many of the values stored in the sim1.dll are used by multiple functions. I think this is also true for the two surface wind parameters, so use it with caution.

The values 20 (lower limit) and 50 (delta) must be changed.

FSX SP2 sim1.dll
Surface Wind Lower Limit:
Offset hex A58: 00 00 34 40 (20 ft/s) --> set to 00 00 08 40 (3 ft/s). Do not set it to 0 or the turboprop engines cannot be started (20 seems be used for spool too)
Surface Wind Blend-in Speed Range (delta between 70 and 20 ft/s):
Offset hex 10510: 00 00 49 40 (50 ft/s) --> set to 00 00 00 00

[JeeHell]

I just checked on V4.5 HF2, double precision "50.0" 8-byte aligned and you get only 2 offsets only (there are a couple more if you don't byte align), that's manageable. And changing the first one seemed to do the trick for what it's worth.

However, that is NOT usable in an addon, only for personal use. And the fix is probably a 10s work for LM. I hope they can repair that...

Thanks again Yves,

[YvesG]

In P3D v5.1 HF1 there are 8 instances of '50' and 7 of '20'. It takes a bit more time to find the right one, but still could be done. But I haven't tried it.

I have the strong impression, that the flight model is treated as a Black Box by LM. There were very few improvements since the first version of P3D. The latest changes affected turboprop engines in v4 by moving some hardcoded values to parameters in aircraft.cfg. But they forget to document them in the SDK. So here it is:

[turboprop_engine]
lo_idle = 62 ;minimum %CNg (%CN1) when Condition Lever (mixture axis) is at 0% (default 62)
hi_idle = 70 ;minimum %CNg (%CN1) when Condition Lever (mixture axis) is at 100% (default 70)
starter_n1 = 12 ;max %CNg (%CN1) that the Starter reaches. With Autostart (CTRL+E) a minimum value of 11.1 is required since with Autostart fuel is injectet at 11.0. Default value 12.