There are definitely flight model issues with the Kodiak. The real problem comes down to the core flight model as we’ve written many times. I’ll try to summarise it here. Before I go on a rant though, the plane was tested by Kodiak factory pilots and two Kodiak pilots on our test team and they attested to the FM being very good.
With FSX/P3D we used real dimensions and coefficients for our aircraft, which were usually found on academic or NASA literature. These worked well enough to get you 80-90% close to the real plane. We would then tweak for the sim’s inaccuracies and were able to get to a very accurate level when it came to aerodynamics. I can shamelessly say that the last beta of our T-37 for P3D was able to replicate everything the real plane could do in the air and this was tested by five pilots with lots of hours in the plane, two of them being instructors.
MSFS has a different approach:
You start with putting in the aircraft geometry and basic coefficients
You test-fly and you’ll be 60-70% of the way there in level flight
Stalls will be way off
You adjust the coefficients but it will make the plane twitchy or cumbersome, while improving overall performance
You then make the ailerons, elevators and fuselage bigger than RL if the plane is twitchy, or smaller if it’s sluggish
At that point, these extra square metres will make your plane more floaty and less prone to stall or vice versa
Back to coefficient tuning
Back to resizing the plane and so on
This approach allows us to get the envelope of a plane like the A320 or 172 close enough to be satisfactory. When you go into special aircraft, your edge cases will be off and the more “special” they are, the greater the error allowance. Example:
Kodiak: prop drag has been tuned using factory numbers and pilot feedback and it’s right on. Flaps were tuned with the manual’s climb numbers and are also accurate. Stall speeds are where they are supposed to be. Yet, when trying to maintain altitude at 800ft-lbs you can’t. Why?
Concorde: It is a delta wing aircraft, a wing type that isn’t supported at all by MSFS. How did Dean’s team do it? I had a look at solution and was astouded -I don’t want to describe it. Suffice to say that it works well enough, but from an aerodynamics perspective it will probably be off in many of the edge cases. If you fly it by-the-book it will do what it’s supposed to
Default TBM: It flies by the book, but the P-Factor is well reduced compared to the real plane. There is no torque roll effect in the plane. If you plan on flying it straight and level it’s fine, but if you go beyond that you start to see the shortcomings.
Stalls (all aircraft): You can control the stall AoA of an aircraft without messing up its performance much. There is no way to control the actual stall behavior, though. In FSX we could control what wing would drop, if it would drop and if or how a spin would develop. Gentle stalls like on the 172 were easy to make, as were weird stalls like the F-4’s roll reversal. Planes with a split wing like the Kodiak and Dash 7 can fly at slow speeds because the outer wing stalls after the inner, giving you more time with aileron control. MSFS has taken away these controls and we have to guess what it does when stalling, because its behaviour is scripted. We have some visual tools that show how the wing stalls, but not why and we have no way of accurately defining that.
The bottom line is that all sim aircraft are made with some leeway. MSFS has too much and it shows with special mission aircraft like split wing (Kodi, Caravan, Dash 7), delta wing (Concorde, Mirage, Phantom), taildraggers (bad takeoff handling), floatplanes (bad water handling). It far surpasses other sims on most aspects, but when it comes to FD it needs to give us a lot more contorl.
