Lack of left turning tendencies
-
@CdnCptMoustache Very interesting. What you're describing with the sim sounds frustrating for developers trying to get a flight model close on the first draft. IMHO, you should be able to simply plug in real-world values to get something good enough to start making minor tweaks to hit published performance tables and real-world flight characteristics.
-
@Buzz Correct. Not on their own they won't create adverse yaw but coupled with other unique sim-isms and some mildly misbehaving numbers in other areas the configs, you can get visible extremes. For example, combine a wing with too much twist with a plane that has the tendency to skid in a turn and you can get either positive or negative yaw/roll coupling depending on the twist number and direction (and the SDK would lead you to believe the twist works backwards from what it does. Did you know that most helis in the sim have their blades twisted the wrong way and have more aoa at the tips? I digress). Or combine that with a plane that has too little fuselage side area and a tendency to skid in level flight. Or too little yaw stability so it wanders. Or.... well the list goes on. And how the sim handles some of these extremes is less-than-ideal. But the good news is, get those in check and make this number mate up with that number and you get something that flies really, really nice.
This is getting a tad off topic now as I'm not really talking about the Caravan at all with all of the above but flightsim aircraft in general. My favourite was one plane a few years back that would have negative roll coupling to yaw, then when you gently corrected it, it would snap roll the wrong direction.
At any rate, when I can get a minute I'll fire up the sim and see if I can come up with specific recommendations for the Caravan.
@Buzz said in Lack of left turning tendencies:
IMHO, you should be able to simply plug in real world values to get something that is good enough to start making minor tweaks to hit published performance tables and real world flight characteristics.
If only... I realize these are my chosen problems, so I don't want to complain too much about the process, but there are so many interacting configurations that it can be nearly impossible to figure out what change will have the desired output, especially when the documentation leaves so much to be desired. The other challenge here is that the flight model you experience is a poorly documented combination of parameters from 1985, and the CFD physics of today. We can usually arrive at something I'm happy with in the end, but it's always a hard fought battle, no matter how how easy you think it's going to be when you start.
@CdnCptMoustache said in Lack of left turning tendencies:
At any rate, when I can get a minute I'll fire up the sim and see if I can come up with specific recommendations for the Caravan.
I would really appreciate this very much. I've found that sim aerodynamics can really benefit from multiple eyes, as it seems that everyone knows a slightly different set of parameters, and how they affect the simulation. I will be working on it too, but I think we can come up with the best solution together, rather than duplicating each other's work. Thanks again for lending your experience and volunteering. It sounds like you know the pain of trying to get what you want out of the simulator, and I think that's half the battle.
-
@Buzz said in Lack of left turning tendencies:
IMHO, you should be able to simply plug in real world values to get something that is good enough to start making minor tweaks to hit published performance tables and real world flight characteristics.
If only... I realize these are my chosen problems, so I don't want to complain too much about the process, but there are so many interacting configurations that it can be nearly impossible to figure out what change will have the desired output, especially when the documentation leaves so much to be desired. The other challenge here is that the flight model you experience is a poorly documented combination of parameters from 1985, and the CFD physics of today. We can usually arrive at something I'm happy with in the end, but it's always a hard fought battle, no matter how how easy you think it's going to be when you start.
@CdnCptMoustache said in Lack of left turning tendencies:
At any rate, when I can get a minute I'll fire up the sim and see if I can come up with specific recommendations for the Caravan.
I would really appreciate this very much. I've found that sim aerodynamics can really benefit from multiple eyes, as it seems that everyone knows a slightly different set of parameters, and how they affect the simulation. I will be working on it too, but I think we can come up with the best solution together, rather than duplicating each other's work. Thanks again for lending your experience and volunteering. It sounds like you know the pain of trying to get what you want out of the simulator, and I think that's half the battle.
@Black-Square I wonder if someone has done a battery of focused simulator flight tests on a "Generic Flight model" to identify and catalog what all of these parameters do and how they interact with the CFD physics or not? Sounds like a fun project where you would get to play simulator test pilot.
-
@Black-Square I wonder if someone has done a battery of focused simulator flight tests on a "Generic Flight model" to identify and catalog what all of these parameters do and how they interact with the CFD physics or not? Sounds like a fun project where you would get to play simulator test pilot.
@Buzz said in Lack of left turning tendencies:
Sounds like a fun project where you would get to play simulator test pilot.
For some of my other work, I have occasionally referenced documents just like this for other parts of the simulator (not aerodynamics) from as far back as FS95. They are all nicely compiled PDF's, or CHM's (if anyone remembers those), clearly created by retired engineers. It seems the mentality that created those isn't as prevalent anymore, or perhaps the systems are just too complex to "make the juice worth the squeeze" now. Be careful what you wish for, haha.
-
@Buzz Correct. Not on their own they won't create adverse yaw but coupled with other unique sim-isms and some mildly misbehaving numbers in other areas the configs, you can get visible extremes. For example, combine a wing with too much twist with a plane that has the tendency to skid in a turn and you can get either positive or negative yaw/roll coupling depending on the twist number and direction (and the SDK would lead you to believe the twist works backwards from what it does. Did you know that most helis in the sim have their blades twisted the wrong way and have more aoa at the tips? I digress). Or combine that with a plane that has too little fuselage side area and a tendency to skid in level flight. Or too little yaw stability so it wanders. Or.... well the list goes on. And how the sim handles some of these extremes is less-than-ideal. But the good news is, get those in check and make this number mate up with that number and you get something that flies really, really nice.
This is getting a tad off topic now as I'm not really talking about the Caravan at all with all of the above but flightsim aircraft in general. My favourite was one plane a few years back that would have negative roll coupling to yaw, then when you gently corrected it, it would snap roll the wrong direction.
At any rate, when I can get a minute I'll fire up the sim and see if I can come up with specific recommendations for the Caravan.
@CdnCptMoustache speaking of that, isnt it amazing how the “coding” of nature has everything in order. Such complex systems functioning in a predictable and quantifiable manner without any “bugs”
-
@CdnCptMoustache speaking of that, isnt it amazing how the “coding” of nature has everything in order. Such complex systems functioning in a predictable and quantifiable manner without any “bugs”
@django1489 i can think of quite a few bugs in nature I'd squash...
-
@CdnCptMoustache any luck on finding a temporary fix?
-
@CdnCptMoustache any luck on finding a temporary fix?
@Buzz No, I'm away from my sim gear for most of the next 6 weeks. If I get a weekend back at home I'll give it a shot
-
For everyone that wants a hefty left yaw on that aircraft could test the following settings in the configs, until there is an official fix from BlackSquare:
engine.cfg
ThrustAnglesPitchHeading.0 = 0, -6Flight_Model.CFG
rudder_engine_wash_on_roll = 0leave all other parameters as is.
Have fun with lots of right rudder.
Cheers
-
For everyone that wants a hefty left yaw on that aircraft could test the following settings in the configs, until there is an official fix from BlackSquare:
engine.cfg
ThrustAnglesPitchHeading.0 = 0, -6Flight_Model.CFG
rudder_engine_wash_on_roll = 0leave all other parameters as is.
Have fun with lots of right rudder.
Cheers
-
@Black-Square is this issue going to be fixed in the next update?
@Buzz Yes. JayDee and I appear to have discovered that the ThrustAnglesPitchHeading parameter is interpreted drastically differently between MSFS 2020 and 2024, which has led to some of the confusion. As nobody else has mentioned this on the developer forums yet, I'm hesitant to declare it yet another undocumented change in MSFS 2024, but most aircraft in MSFS don't make use of this parameter at all. The identical values from JayDee above, which are reasonable in MSFS 2024 produce completely unflyable results in MSFS 2020. Without understanding for this disparity, it appears I will be publishing the update with two different values in the two simulators, which is something I like to avoid whenever possible, because you never know when a sim update might break (or reintroduce) parity.
-
@Buzz Yes. JayDee and I appear to have discovered that the ThrustAnglesPitchHeading parameter is interpreted drastically differently between MSFS 2020 and 2024, which has led to some of the confusion. As nobody else has mentioned this on the developer forums yet, I'm hesitant to declare it yet another undocumented change in MSFS 2024, but most aircraft in MSFS don't make use of this parameter at all. The identical values from JayDee above, which are reasonable in MSFS 2024 produce completely unflyable results in MSFS 2020. Without understanding for this disparity, it appears I will be publishing the update with two different values in the two simulators, which is something I like to avoid whenever possible, because you never know when a sim update might break (or reintroduce) parity.
-
Thank you Jaydee for posting this!
I've been playing around with this, mixing it with the prop_mod_moment_scalar_yaw setting. I'm currently using -4 for the Thrustangle and 2 for the prop moment.
Increasing the prop moment scalar will also increase the "left turn tendency", but will make the aircraft unstable with higher values. By mixing the two, you can tweak how responsive the aircraft is/feels to torque changes.Also worth trying is turning on CFD_ReinjectVTailX and CFD_ReinjectRotors in the flightmodel.cfg. Without this the aircraft will start turning left more, with increasing speed, during the takeoff roll (with the same engine torque). Requiring increasingly more rudder input during the takeoff roll. This seems incorrect, since the rudder effectiveness should increase with increasing airflow, requiring less input.
During rotation it should increase again, because of higher AoA / P-factor increasing.
I haven't had enough time to test this, but it seems to work better with these two settings on. -
Without playing around myself, I definitely don't recommend using raw engine pitch/yaw angle to impact left-turning tendencies. This can have other relatively drastic side effects like how well the plane stalls, handles slow flight, etc. My countdown is a further 4 weeks until I can truly spend some time tweaking the sim, myself, though there's the possibility I could do some playing around this weekend.
Just to clarify, most aircraft IRL would use a positive heading adjustment for an engine with a naturally left-turning tendency to help combat it. Setting the heading to -6 just artificially induces left-turning due to the engine being angled to the left.
-
It also causes you to maintain rudder input at cruising speed, where the plane normally is balanced out.
But if you try to get the same result with the prop_moment scalar, you'll need higher settings of around 15, which makes the plane instabile on the yaw axis, noticable when you turn the AP on.
That's why i was trying a mix of both settings and increased P-factor. -
I still have this problem with the cargo variant. It's like it was never fixed in the last update. Left rudder for takeoff and then a lot of left rudder trim during climb. The passenger version seems to be fixed even if I think the slipstream effect should be much greater.
-
I fly the cargo variant more than any of the others and I do not have that left rudder on takeoff issue. I actually think I'm going to start leaving a few turns of right rudder on the trim wheel instead of centering it before every flight.
@jmarkows said in Lack of left turning tendencies:
I fly the cargo variant more than any of the others and I do not have that left rudder on takeoff issue. I actually think I'm going to start leaving a few turns of right rudder on the trim wheel instead of centering it before every flight.
I might have messed up when I installed that variant (using an old installer etc.) then. I'll check tonight.
