It’s a custom weather setting. I have been able to mitigate the issue by ensuring the temperature is low ~ 8-10C/45-50F. It’s when the temperature is higher (around 23C) that the plane is bounced all over the place, which I’m sure is reasonably realistic, but when trim in the sim isn’t very well modelled and you have turbulence set to low I just want it to go away.
Anyway as long as the temperature is kept low it’s fine so that’s a workaround I’m happy with.
Same issue here. Had hoped SU12 would be better but it is downright terrible to fly GA, specially in live weather.
Had flights all over the planet and it’s just downright insane how some planes get thrown around. I understand small aircraft get tossed around a bit but the movements in MFS seem well out of proportions.
Doesn’t help that the planes all have various different quality flight models too, some planes feel like paper while others do feel weighty (while same weights irl).
I have put around 1500 hrs in the game but after SU 8 i believe (?) it frustrates me to the point of just droping it. First the Arrows broke due to the way turbulence was done but now i can hardly fly any small GA plane without just silly movements.
Problem is that the turbulence feels poor. Like some planes just fishtail like mad, while others suddenly pop the nose at 30-45deg up and then to like 20-30deg down in a second.
I’ve got my turbulence setting at Medium as it seems to be closer to the real life flights I remember.
But yesterday, flying the C208B in Brazil I noticed turbulence was still quite pronounced, to the point where the overspeed warning went off a couple of times, following by an “overstressed airframe” crash. This happened twice before I disabled the airframe stress setting.
Real time weather set with varying clouds but generally light winds and no terrain effects - wouldn’t have expected turbulence in real-world flights.
As you mentioned above the aircraft should still maintain some type of stability when encountering turbulence.
Up/down thermal turbulence is generally what you encounter in real life, as the wings fly through the ‘bubbles’ (thermals) of hot air rising, and then exit them as the aircraft moves along. Sometimes you’ll pick up or drop a wing if you catch one on the side - that doesn’t happen often at all though. Generally it’s the whole aircraft jolting up and then more gently letting back down.
The yawing left/right from supposed wind gusts just doesn’t really happen IRL, especially at altitude. Previous sims have all made this error. Hopefully MSFS can do it better in the end.
i just found this thread because i was coming to the forums to see if anybody else was having the same experience i have had with SU12. for me the turbulence has disappeared almost entirely. i am running on “realistic” setting and in the last few weeks i have only seen one or two rough patches of flight.
i just flew out of this huge storm that is over pdx and there was zero visibility until like 25,000 feet but the entire trip was smooth as glass. i’ve been mainly flying the 737 and haven’t done much GA lately so i will try that out and see if i get the effects y’all are describing. the rough patches i have seen were both during departures so it may be related to flying lower altitudes. i’m really only near 5000 feet for a minute or two per flight so that may be why i just never really notice it anymore.
i do like that i don’t get automatic turb just because i’m flying near mountain areas anymore. but i feel like the overall turb experience feels like it got turned down too much …
edit: well i finally got some turbulence on this flight i’m doing from kpdx->pakt . but it started almost exactly when i started overflying some mountains at FL40 right over vancouver island. so maybe that wasn’t fully fixed either i don’t know 
They have reduced the height of thermals to the cloudbase. Mountains that caused turbulence on higher altitudes has also been reduced at higher altitudes.
That means at higher altitudes we have no turbulence at all anymore. Thermals height and mountainwaves height is more realistic now but the missing mixing of air inside clouds and at higher altitudes are more noticable now because of that. I hope they can simulate CAT and turbulence inside clouds soon too. But to simulate mixing of air is much more complex i think. The CFD-simulation they have could do that but it requires much more calculations and stuff. They working on that they said in a q&a a while ago.
thanks, very interesting to hear. i’m glad they are still working to improve it 
CAT is largely caused by the jetstreams which are very hard to calculate/predict, let alone simulate… so that might take a while
Meteoblue has data for CAT light-severe between 150mb-950mb. Then how to implement that data is another question. It needs to feel realistic But the data would make it realistic because if we reduce altitude we may avoid the CAT. The tricky part is to implement that data into the CFD-model. But i hope that will be the future of atmosphere in flightsimulators. That the air is actually simulated like a fluid.
Just an idle observation from an old aviator. In previous flight simulators folk were pleased to have good visuals and nice aeroplanes to fly - my first was all blocks and wire-framed “instruments”.
But now people get uptight when their super-duper study level machine gets thrown about/doesn’t get thrown about enough!
Ah, such is progress and anticipation. Just an observation!
How do you control your trim?
To the OP, yes, what you describe happens in real life. I was flying a Warrior to Nantucket from Nashua a while back and experienced exactly the same conditions you describe. It was so nice to get out over the ocean into the glass-like calm after constantly getting knocked around from Nashua to the coast. The trip back later that afternoon fortunately was a quite a bit smoother of a ride.
At one point, I bent down to pick up my pen and at just that moment I got knocked sideways to a 45 degree bank… That was fun 
Personally I found that setting my elevator trim to my unused throttle axis on my joystick to (trim -100% - 100%) to make flying in MSFS infitinitely better and easier. Trying to trim my plane with joystick buttons up and down was brutal. Using an axis to control trim made MSFS usable.
The physics of using PC controllers to control aircraft without feedback makes the control of planes super difficult. A slight movement of a controller can apply forces to planes in the sims that you’d never be able to do in real life, and I think this is a large reason for people calling planes “twitchy”.
I’m not saying the turbulence in the sim is perfect. Just that there are other factors involved as well.
Well, it depends on the aircraft, but I make gross changes with an electric trim switch on the Alpha and fine changes with the wheel on the Bravo (tuned using SPAD).
One thing about the Alpha is that the elevator axis springs are so stiff that you tend not to make inputs by accident when pressing buttons (confirmed using the indispensable input viewer tool).
I was flying the BB PC-6 yesterday through the Cascades from Hope (CYHE) to some strip (W12) , I found myself observing that the PC-6 is one of the few aircraft I’ve flown in the sim where the trim feels right - it will hold the trimmed speed in a climb or dive without any accelerating deviation from a minor bump. I could even go hands-free at times in the light chop I was getting yesterday. Wouldn’t dare to do that in the Kodiak!
FWIW, I find the realistic turbulence model, well, pretty darn realistic.
As for using an axis…interesting thought. I have noticed on the input viewer, that even using the small inputs from my calibrated trim wheel, that the elevator trim value can jump around when close to zero. Its not as smooth as it is when further away from the zero position. But I’ve never made a close study of this behaviour - just noticed it.
When I think turbulence I think “potholes in the sky” and thats probably due to the fact that nearly all of my air travel has been as a passenger in a 737 or MD-80 or maybe a regional jet. I’ve been passenger in a small Cessna one time and a few times in small ‘puddle jumpers’ as my grandmother always called them. The worst turbulence I ever experienced was in one of those puddle jumpers during summertime rainy conditions between Jackson, Tennessee and St Louis MO. That was bumpy and side to side thrashing (I was towards the tail).
But in the sim I fly a C172 mostly and, no matter what plane I fly, I’ve never encountered ‘potholes in the sky’ but I have encountered a great deal of what is technically turbulence I guess but to me its just windy and gusty conditions. But I am not a real pilot so I cant know what a C172 in turbulence is like. What I have seen through watching videos though is that some planes like a DA42 will yaw quite a bit while it seems others don’t do it much at all.
All of this is to say that I find it better than FSX, not as good as when the sim was released where apparently turbulence was just random air movement and no, not realistic, because ther are no ‘potholes in the sky’ (aka sudden jolts that are noticeable but dont really throw the plane sideways or anything) my.02
The thing is the air has turbulent airflow because of a purpose IRL. It’s not added there randomly. The air is turbulent because of many different things IRL. Most of us knows why there is wake turbulence for example. It’s because the aircraft in front of us causes the air to mix and be turbulent behind it. Turbulence doesn’t mean random. It means unpredictable airflow. Turbulence is chaos and should not be predictable. In low air velocities lower mass should be affected less while in higher air velocities lower mass should be affected by turbulent airflow more. Then also the aircraft model itself is designed to get less affected by the turbulence thats why all of the aircraft should feel different from eachother in the air. I bet it’s really hard to get that part right. It’s not only the atmosphere that needs to be correct designed. It’s also the aircraft model.
Aren’t ‘random’ and ‘chaos’ sorta the same thing? especially in this case of computer simulation? I believe you think like I do and want proper cloud turbulence generated by warming, cooling, expanding, and contracting air masses all interacting with each other and causing ‘unclean interactions’ (turbulence) where conflicting air masses meet. That would be great but the sim will be limited by the resolution of the sim (the CFD part). The simulation of turbulence is limited by the resolution of the CFD sim. Seb, in one of the Dev Q and A videos, showed that wake turbulence is already a thing in the sim – or at least it can be. I don’t know if its actually modelled to affect other planes. It should be though I would think.
The micro (and some macro) perturbations are random, but overall turbulence is somewhat predictable, given atmospheric soundings, radar data, and other knowns.
Well, to be honest turbulence is one of the most complex thing to simulate 100% realistic. But asobo has this CFD and that they could actually use to implement LES (large eddies simulation) model. We can never expect it in full resolution but any type of resolution would make it as close as we will get turbulent flow. That would actually be a real simulation of a fluid inside the flight simulator.
For the “random” part of turbulence/chaos in flight simulators i think you are correct. But make the turbulence just random makes it feel less like turbulence. IRL it’s not just random. It’s dependant on the initial condition but the final state of the turbulence is hard/impossible to predict because of the complexity. theoretically it should be predictable but no computer in the whole world can predict chaos 100% correct because it’s still unpredictable. Maybe in the future with machine learning and stuff it will get predictable but it will never be random. Thats why forecasts not always matches the real world conditions. Should be same in the simulator. But in flight simulators we want the air to be predictable for example when we check a METAR we want the velocity be fixed at that value near ground where turbulent flow occure the most. In the real world air doesn’t behave like that. We as flight simulator users needs to accept to have the air flow dynamically to have a realistic fluid simulation of air. Same thing with the visuals of weather. Because the weather we see is often turbulent/unpredictable flow. But in flight simulators we want the weather to be fixed in for example a broken or scattered state because the METAR tells us BKN or SCT. That in my opinion is a really simple simulation of weather or airflow we as users want 
But i hope we can get it more advanced/organic in the future
indeed. I want the timing of the weather to be close to real-world timing as possible because it’s impossible to use 90% of useful aviation weather tools to make aeronautical decisions when the timing of the arrival of weather systems is 30-60 minutes or more off.
I did more research to this effect during the outbreak two weeks ago, both away from reporting stations and at reporting stations.
Findings: a) the reporting stations did not seem to affect the clouds at all - only wind and visibility. Iowa city was getting slammed, the METAR updated first and reflected that with wind and visibility, but the clouds did not arrive for 30 minutes after (and they were pretty weak for what was really going on).
And b) the weather outside of the reporting stations was also at least 30 minutes behind.
If we’re going to frame a side discussion, let’s frame it correctly.
As far as turbulence, some of it is synoptic scale, some mesoscale, some storm scale. Predicting that is more of an area thing versus a specific location. So I think it will have to be binary - turbulence on, turbulence off. And just like everything else, as it moves, new coordinates will be injected and it will suddenly shift.
Well, i just tried to explain that it’s impossible to simulate atmosphere when we as users needs the weather to 100% predictable. I’m not 100% sure how MB implemented the METAR data into the MB model in the new MB model they use since su7. Only Asobo and MB knows. In my opinion it feels/looks less like weather after after su7. Did it make it more accurate to METAR, i don’t really know. But that thing we have discussed in other topics here on the forums.
I think and hope Asobo will improve the turbulence model in the future because now we have the option to set it at low if we don’t like it
It is hard to say how it’s implemented. I’ve been trying to find the seams and discover any commonality between discrepancies and I really can’t other than to say when you’re flying around the edges of very inclement, let’s say severe weather, it’s not where it should be and it kind of blinds a pilot as to what they should do next.
If we start implementing storm-level turbulence and we blindly fly into it because the storm is where it should have been 30 minutes ago, then we’re really in for a disaster, haha. Or when storms are popping in the summer, etc. it’s the edge-cases along the weather interfaces where the accuracy is very important.
But it seems these are all separate, siloed issues going on:
The depiction of weather (which is more generic in cloud types, lack of thunderstorms, etc)
The location of mesoscale and storm-scale systems (delayed)
The injection of METAR (also delayed, but hard to pin down exactly what it’s doing/affecting
And the generalization of turbulence, which is probably more of a lack of computational power than anything