The *actual* problem with controlling the helicopter in a hover

[D-Frame]

I already posted something similar in a different thread, but my reply probably went under in the noise. I wanted to put this in a more exposed place because I think it's really important for the dev to read this.

I have some experience myself implementing different helicopter flight models. One physically correct one where actual AoA, lift and drag values are calculated at different portions of the rotor disc and another more arcade-style one where inputs are applied directly to the helicopter's movement vector. I'm not talking out my ass here, is what I'm trying to say.

A lot of people have expressed problems controlling the helicopter in a hover, but have had trouble pointing out what specific issues they encountered. This led to the dev implementing input curves and dead zones, which is a nice addition, but it won't solve the problem.

The problem is that roll rate seems directly linked to cyclic X input. In the sim, the helicopter will continue to increase its roll rate when cyclic input is constant, which means that you have to apply opposite cyclic just to stop the rolling motion. This is what makes it hard for people to control the helicopter, because this is neither the case in the real world nor in other sims.

Since the rotor is what keeps the helicopter afloat, the helicopter's fuselage hangs off of the rotor hub like a pendulum. And just like a pendulum, the fuselage will try to swing back to its neutral position due to gravity. Since the neutral position is underneath the rotor hub, this pendulum effect always acts against the current roll angle.

This is what makes real helicopters (and helicopter in other sims) easier to control in a hover than the one in this sim. Even under constant (but slight!) cyclic X input, the roll rate doesn't increase into infinity, because the pendulum action works against it, stabilizing the helicopter.

This is what needs to be implemented: Gravity applies a downward force on the fuselage, which results in a torque on the rotor hub. If you're not working with actual forces and torques, simply apply an angular acceleration based on the the helicopter's current roll and pitch angles, but inverted.

I hope I've made this as clear as possible, but English isn't my native language. If something is unclear, I'll be happy to elaborate.

[fabrice EYF]

Hello D-Frame, Very interesting what you explain ..... I Hope the dev correct/implement that ASAP ! .... You say you'r not English native ? But the Developper is French ... arn't you more easy with French than English ?? ( just a question to facilitate the conversations ...) ... best regards

[Ironbostik [IT]]

Thanks for this Heli Physics lessons. I am sure Devs will translate the detailed info into a patch !

[Lines Studio]

Hi,

I've made some test to demonstrate the behavior of the Heli
this test are made with constant cyclic forward at 0.1

you can see on the ground the rotor is producing a torque roughly "100" to pitch forward at 0° pitch
and a slight torque to the right because the main rotor is slightly tilted to the left
https://i.gyazo.com/2d24eb976e1938053778b5c6b82635ec.png

when I'm raising the collective, the torque is increase all over the rotor
https://i.gyazo.com/78f50afcb1e8ab25e48ac4458bd9aa9f.png

as soon as I've leave the ground and get some forward speed, the torque value is decreasing (pitch at -20°)
https://i.gyazo.com/4cc6dd30ab01a94a429fecd2ad96f3f8.png

the torque decrease so much that it will become negative (I highly thing that's the pendulum effect you are talking about)
https://i.gyazo.com/3598dd9177de6a3ca7167187102ee916.png
https://i.gyazo.com/def2f1b47b0eb9ff71ceb590af60214e.png

until the point the helicopter is at the opposite pitch (20°) and the torque start to increase again
https://i.gyazo.com/6f4247bc8fe366e507eeec60342cd858.png

remember the cyclic is on a constant value

the effect seems to be less present for cyclic X because of the Inertia momentum which is lower on that axis,
but the effect is still present.

I've try to implement a pendulum effect equation at first, but due to the weight of the heli,
it was totaly stable and uncontrolable (the torque produce by the main rotor cannot counter act a pendulum effect)

the flight model is physically based, that's all maths running in the background, so if you found some ressources
about that effect in deeper mathematical way, please let me know, and I'll check it out

[D-Frame]

Thanks for the insight. It's hard for me to interpret this data without any knowledge of the inner workings, but let me just point out the following:

The pendulum effect I mentioned is not something caused by the rotor, but solely by the fuselage being pulled down by gravity. I've created this terrible little image to illustrate:
https://drive.google.com/file/d/1pA1WcnMcFlDEEDotq5Abx8ZltFktQcq3/view?usp=sharing

The effect is also briefly mentioned on a few sites, for example here:
https://www.danubewings.com/helicopter-pendular-action/

[Stalker]

Clearly if this program is marketed as a simulator then the pendulum effect stated by D-Frame should be built in too. If this is not being simulated and is a real-world phenomenon then we don''t have a true simulation which is what I assume most people thought they were purchasing. As close as possible to flying and controlling a real helicopter should be the main Dev focus - not making it "easier" for people who really shouldn't have bought the software if what they really wanted was an arcade model helicopter and aren't prepared to put the hard work in to learn. Flying a helicopter is not meant to be easy. It is far harder than flying a plane and that very difficulty is what makes them special.

[Lines Studio]

The maths behind the pendulum effect equation doesn't reflect the reality at all.
please do the maths behind this effect

as I say, I've allready implement this equation into the game

for a R22 it's ultra stable (far too much)
for a CH53, it's so stable that you will only make it move up and down

I deduce it doesn't work like that

[Stalker]

Fair enough and I accept your comment. However it would be great for another experienced pilot to actually comment on the realism of your flight model. Armchair simmers here on Steam only "think" they know how a helicopter flies. It is the "as close to reality" as possible I am looking for so if that is what is what is being reflected in this program here and a User with actual flight experience would care to comment I'd be very interested to hear their thoughts.

[sirtraven]

It has been several years since I flew a R22 but I can tell you that hovering is one of the hardest things to do in a heli as there is nothing naturally stable about it. It is a delicate balance of cyclic, collective, and pedals. For every little move you put in you have to take at least half of that movement back out almost immediately to stop or reduce the motion in that direction. The heli will not stop its motion naturally, at least not on the R22. I think most of the heli simulations, if 100% accurate would be immediately complained about due to the difficulty in trying to land or hover the heli. Without being able to feel the craft move or having peripheral vision makes it very difficult to hover in simulators, and why I suspect most sims choose to have a more stable slightly less accurate heli. I think 3D HMDs may help here as you can turn your head easier and more naturally to help with the directional judgement really needed for proper hovering.

As to the pendulum affect mentioned earlier, this really comes into affect if you start to ham fist the controls as you setup an oscillation that can, if unchecked, crash the aircraft. An unskilled pilot will push the cyclic in a direction a little too hard and then push hard in the opposite direction once they realize their mistake. This causes the body to swing wildly back and forth as the pilot attempts to gain control. Eventually the swinging will go outside the tolerance of the rotor and cause it to dip loosing lift and then down you come. This is why when hovering or taxing it is so important to extremely small inputs to the cyclic. You should be able to hover with only one finger on the cyclic, that is how much press it should take. I know I haven't provided any math equations to describe this but I hope my experience helps to understand the effect a little better.
Just trying to provide some of my personal experiences.

[fabrice EYF]

Hello, I absolutely agree what Stalker and sirtraven says ! before talk of my experience, i must say there is a big think to avoid on a 2 rotor blade, it's Mast Bunping ...it is FATAL, google it for eplanations .... in fact it is due to the pendulum effect described here, but this is is only present in hard manoeuvers like Sirtraven eplained ... just to mention the "Mast Bumping" is only present with 2 blades rotor ...like R22 , R44, B206 and so-on ... so i think the pendulum effect is only a problem on these aircraft and not on the "more than 2 blades" ones .... this , plus the fact there is only present in a very hard manoeuvers and eventually "fatal" ... we may ask us if we really want to spend time to simulate that ? if it's fast and easy to code, OK but ...?
I Must say i want pure realism like Stalker wants ....
But i dont' think ( personally, perhaps i'm wrong ???? ) the pendulum effect a parameter who helps to control a chopper in a hover ....
Like Sirtraven says very well, a chopper is hard to keep stationary, is a constant "balance" of several controls and parameters ....
My experience : very little, but I have flown a Schweizer/HUGUES 300 Cbi for almost 20 hours, it's a 3 blade, but i can tell you the hover is like all the chopers , a constant balance of controls !!
When i can configure my "scale 1" USB controller on this sim, i'll test the flight model and i'll tell you what i think about .... but for instance i must play with a little joystick which is not able to let me "feel" the real reactions of the helo modeled ... Yes "muscle-memory" is important ....
Okay this is a long post , sorry i just want to expose my point of view and ( Little ) expereince of real flight on a chopper ...
This sim is great and if all the guys like you continue to inform the developper regularly, we'll have a very good simulator to work with !! Fly safe, regards ... Fabrice

[fabrice EYF]

Just to mention another 2 things .... if the chopper is hard to control in hover in the sim, may make us think the flight model is accurate .... ( not 100% sure but , at opposite side, a helo "on rails" is purely a bad model !! ) ...
Second and not least , it's well known since several year, in MSFS ( FS9 - FSX etc ) X-Plane DCS and ALL the helo sims ... that the size of the stick is IMPORTANT .... to control a helicopter ( specially in a hover ) you MUST have a stick wich is the SAME size and travel than a real Cyclic for exemple ... With real cyclic in a hover the inpts are very small 1 or 2 centimeters ... it correspond to a very little "angle" to the cyclic gimbal ... imagine the lever arm corresponding .... Now imagine the same angle on a 10 cm long joystick !!!! you'll must imput only a millimeter !!!
So peoples who complains iy's hard to hover is because :
1 Yes it's Hard !
2 It's Accurate !
3 you have probably not the good controls ...

All said before is of course for "accurate mode" with no assistance !!

Ok it's my " 2 cents" Just to help .....
I promise as soon i can test the model with my controller i tell you what i think ....

[fabrice EYF]

OK Guys i'd success to bind my home cockpit controller into Helicopter simulator ( i'll post another topic to speak about that .... ) .
And I flew a little ....

First thing : Its better than the basic HOTAS joystick !!!!
Second : I put the Assistance of Pedals and Collective OUTS .....
I only use the Cyclic assistance .... I think there is a problem with cyclic is way TOO reponsive !!
Third : I think there are some problems into the fligjht model, but i'll speak about that another day ....

'MY" Conclusions (just my opinion) :
- Except one or two things the flight-model seems to be accurate
- You need REAL controls if you want to fly as close as real chopper, specially if you have a Real flight experience on helos before .... traditional Joysticks, whatever their quality, are playables but not adequate to REAL sensations .....
- You need to calibrate correctly your controls and choose the right curves .... and mostly absolutely NO deadzones , that they are from "Hardware" or"Software"...

Personnally, i control the bird into hover with no problem more than the 30" required ...
With the folowing configuration :

Scale ONE hardware controller (here is DIY but hardware like Proflight trainer is strongly advised) ....
All Axes = NO deadzones ( 0.0 )
All Axes Linear ( or 3rd curve NOT second .... need a try between Linear and 3rd).
Rudder Assistance = NO
Collective Assistance = NO
Cyclic Assistance = Yes

Collective and Power ( gazs ) are "Progressive real axes ".....

Hope this helps !!

Fab "EYF"

[espani0la]

TBH cyclic assistance in this game does all the work for you in hover and you don't even need to touch your joystick

Originally posted by fabrice EYF:

.
I only use the Cyclic assistance .... I think there is a problem with cyclic is way TOO reponsive !!

...

Personnally, i control the bird into hover with no problem more than the 30" required ...

...

[fabrice EYF]

Originally posted by espani0la:

TBH cyclic assistance in this game does all the work for you in hover and you don't even need to touch your joystick

Originally posted by fabrice EYF:

.
I only use the Cyclic assistance .... I think there is a problem with cyclic is way TOO reponsive !!

...

Personnally, i control the bird into hover with no problem more than the 30" required ...

...

Hello I do not agree, but i understand what you explain !
I agree with the fact the cyclic assistance do "too much work" , but in otehr hand, cyclic response without assistance is way too reponsive and irrealistic ....
So if i want to fly in this Sim i must put cyclic assistance ....

What i do not agree is when you say you don't need to touch the stick ....
I think we do not have comparable hardware ..... Can i ask to you wich joystick you use please ??
As it a spring return to center ??

Fab

[espani0la]

Maybe I was slightly exaggerating when I said it but assist does most of the work

I'm rocking Thrustmaster T16000+TWSC Throttle+TFRP and it's not sprtingloaded it's some kind of magnetic ♥♥♥♥ I dont understand

[fabrice EYF]

Hello,
Sorry i'm not English native, I'mfrench, so I may be misunderstood ....
by " springloaded" i mean, the joystick return automatically at centre when released ....
Like most of Joystick ....

If it's the case, it explain a part of the problem .... ( I'm Ok the flight model is not very accurate, and the input system is bad ...i'm also Ok the assistance are too strong, but flight model without assistance is too sensitive too ..!!!! ).

The problem is that we don't need any return to center, and NO deadzone on a helicopter cyclic.

When we moove the cyclic forward for example to go forward ... in real copter, we put the cyclic forward a little and we MAINTAIN the position ... so the Helo tilt forward a certain amount and stay at that amount and we go forward ....then to stop we pull the cyclic backward, the helicopter "tilt back", then we immediately but gently put the cyclic in the middle and the copter stops ...it's always a "search of balance" and this is very difficult to do with a regular joystick ....

If we want to fly "real simulation" we need proper controls ...and of course good flight model....
Other side if we need to fly with a joystick , we need to accept to fly a little "arcade" and that the developper has implemented an adequate mode....

I read in the recent Devlog, the developper plan to implement the 2 modes in the sim ...

The dev also told me he understand there is a problem with the joysticks and the flight model and it will be a lot of work, but he plan to correct all this in a big update .... we don't know when, but it's worth the wait ....

Fabrice