Here is a mod for the flight model of the DA40NG, an aircraft I know quite well.
The mod is based on actual performances filmed in flight purposedly to fine tune the MSFS DA40NG FM (CL vs aoa, CD total at various speeds, flt controls response…)
Fixes are:
accurate W&B model based on the real aircraft manual.
Thank you so much Captain! Great! I stopped flying her because of some lacks that I’ve noticed and you just start to fix them!
Good job! I give a try now!
I was flying the C152X new mod, love it, but it goes slow which I like too but for the world tour I started from LFQT (North France) and now at LIML so it’s good to have a PA which provides the DR400NG!
Now that the weather seems to work accurately (at least on my side in Europe) having the correct metar that strangely now match with Meteoblue, getting different layers of wind, even turb on final… so I can at least make a kind of “world tour in Europe” before the weather gets fixed in the next update to travel in all the world during months and getting back to France …but since the last patch it’s getting better in terms of weather accuracy! … don’t know if it’s a coincidence or the last patch by the way… Anyway great to have the DR400NG enhanced thanks to you!
You have to play with the parameters to get proper lift, drag, pitch moment etc…
If it was just a matter of copy pasting real aircraft values it would be too easy !
WB station were verified several times with lights thanks to sdk tools.
You can check with these tests on the real aircraft
I’m curious to see how you get the engine to not cruise at 2300rpm the whole time… I highly recommend the G1000 mod out there for dimming the G1000 at night too
I’m trying this mod and the page weight & balance says that center of gravity is out of limits. You have fixed W&B model based on the real aircraft manual, so I suppose that we can ignore that page warnings, the plane should have the center of gravity in the correct range. Is the total weight that we can read in that page accurate? Can we change safely change weight & balance of the plane from that page?
Thanks
Limits are shown with a wrong MAC positionning, which is barely controllable. However SDK says MAC position and length are not impacting modern FM, which I verified.
You can visualize force position with debug and CG position is logical.
One thing that I have noticed, I cannot set the flaps to take-off position.
I have flaps increase/decrease mapped to my hotas, and with default flight model works, but not with your mod. The flaps go from up to landing and from landing to up.
I miss something or someone else can confirm?
Flaps are going into TO position physically (you can see it on model and feel it with lift/drag) but it seems the indicator does not follow. Not sure to understand why
Great work. A few things based on comparisons to the real plane:
Flap position indicator shows landing flaps no matter what setting, but I think you know that bug by now
T/O flaps produce too much drag compared to real life (correct amount of lift, but too much speed loss)
Landing flaps do not produce enough drag compared to real life, which makes it feel a touch too floaty and difficult to lose speed during landing, but as a result of speed staying up, not necessarily as a result of too much lift
Flap position indicator shows landing flaps no matter what setting, but I think you know that bug by now ==> Yeah just discover that. I need to investigate if I can fix that
T/O flaps produce too much drag compared to real life (correct amount of lift, but too much speed loss) ==> Compared to flight tests on real one (idle descent in TO flaps config (see here ,flaps TO set @ 1.'24", @1’41"we descend at -7.5deg/84kt/-950 fpm).
Landing flaps do not produce enough drag compared to real life, which makes it feel a touch too floaty and difficult to lose speed during landing, but as a result of speed staying up, not necessarily as a result of too much lift ==> **@2’48 of the same video above we maintain -6.5/73kt/-1050fpm in LDG flaps (set @ 2’20"). In sim I have the same results even a little more drag than in real. ** [/quote]
Assuming lift is correct, AOA are correct. So pitch attitude is a proxy to descent flight path angle. If pitch in descent is correct then FPA is correct, which means drag is correct since thrust is almost 0.
@BusPeeGee Ah, thanks for that. I think you’re getting different numbers since you’re comparing it to a DA40-TDI, not a DA40NG. They have slightly different characteristics, in part due to the CG difference from the Austro AE300, but also a few other slight but noticeable differences.
I can look, but they might already have the CG correct for the NG. It didn’t seem too off to me, but the L&D in each config definitely does, along with the engine performance and fuel burn.
In just a little bit I’ll PM you one of a few good DA40NG videos I have. It’ll show what I’m talking about, but a few reference points are:
Clean (no flaps) in real life -
Much better glide ratio
50% power, level @ 4.4gph, 100kt
75% power, level @ 6.8gph, and a touch faster than what it shows in game
Best glide is 88KIAS. That’s about -2.5 degrees pitch @ -900fpm
Will come back to this later after I have a chance to make a more direct comparison this evening or over the weekend, but just wanted to mention it –– Note for the below numbers
Full fuel (18+g per tank, 14g indicated), one 150lb person
30.30in pressure, 1400ft MSL pattern altitude, 27c OAT
T/O flaps -
Full power @ +7.5 degrees pitch = 72-74KIAS, roughly +850fpm
Idle @ -5 to -7.5 degrees pitch = ~88 to 90KIAS, -550fpm to -750fpm
Idle @ -10 degrees pitch: will quickly accelerate to over 92KIAS and climbing, plus will increase rate of descent to roughly -1000fpm
Much much more right rudder required when removing flaps at +400AGL
The plane in game stays well-coordinated almost the entire time, which is weird
LDG flaps -
Immediately need to add power (40-50%) as speed quickly drops to ~80KIAS and falling
The method I follow consists in solving the aerodynamic model step by step, by eliminating unknowns.
I did measurements in idle descent especially to get rid of engine unknown informations and to solve the couple Cd0 and Cdi. It’s like solving a mutiple vars equations systems. You do descents at various speeds that will lead to different glide path angles.
There is only 1 single solution (a couple of Cd0 and Oswald efficiency) to match all descent speeds with correct glide path angles. Before that I make sure CL vs AOA is correct by doing measurements in level flights. This also gives information on pitch moment and so center of lift position/elevator lift through trim value. Once drag and lift are correct, you can tweak the engine model to get proper TO, climb and cruise performances.
CG aspects has no effect (or marginal only if you consider trim’s drag) on performance (lift, drag, climb). It will have influence only on controlability, stability and trim values. Just my 2 cents.
After having tried several modifications on the engines.cfg, it turns out that several parameters are not having any impact on the engine performance. Especially prop_efficiency_table and prop_power_cf. After checking with the debug engine pages of the aircraft editor in dev mode, changing these tables that are crucial in the tuning process have no impact on the live values of these very same coefficents displayed (and no impact on performance. You can even delete these tables, it loads anyway.
It seems Asobo uses internal generic tables, proabbly thoese beautifuls curves display in green and red in the debug window (but with no figures). Maybe these entries are reserved for legacy mode, but then it’s very sad because, developpers have very limited options to fine tune an aircraft in the modern flight model.
It’s the same problem with lift_coef_aoa_table that lists CL vs AOA. As sdk mentions, it uses a “subset” of this table. In fact it uses probably CL max and CL0 to scale a generic lift curve profile mathematicaly. You can still adjust height of this pre profiled lift curve with another parameter but that’s all.
As a consequence, post stall lift is not modifiable. In the case of the DA40 model, the post stall dip is too strong, leading to a mush and a spin. In reality it’s just parachuting and remain controllable and stable. It’s a pity that in Legacy mode we can adjust this and in modern we are so limited. I mean if we could have more freedom to adjust the models we could show the modern flight model in all its glory.
Frankly after having spent hours fine tuning the model, I am convinced it gives a huge benefit in terms of realism and handling once properly tuned. It’s very convincing. I hope Asobo will consider this opinion and open up a litlle more modeling capabilities.
Fair enough point on any engine load being a differentiator, but you still have to recognize that you are comparing an NG to a TDI. Modding the NG to be more like a real life TDI isn’t going to be quite right since it is not the same plane in terms of weight or aerodynamics.
Taking the engine out of it, I looked again to make sure where I’m getting these numbers from was exactly 0% load and not even a few percent. I looked again and the T/O flaps numbers I posted before actually had about 5-6% load on the engine, which seems to make a huge difference in terms of glide.
The below are based on 0% load, so should be a better comparison - yet we’re still saying different numbers.
