Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[BT~wasted]

I have the Mi-8 DCS module for quite a long time already, but I still didn't manage to learn how to fly it properly. It always tends to go somewhere in all axes. You constantly need to trim it all the time, but trimming mechanism is different then in AC and I assume that it would be quite difficult to model it in IL-2. You can find more info on youtube tutorials about Mi-8 DCS.

As for autopilot Mi-8 has a very simple autopilot, that acts simultaneously in all axes, but you can turn them on and off. And it doesn't flies the machine for you, it only helps you a bit and makes it more stable.

all in all you should be very careful with controls there. Do only small input, no harsh movements etc. Also you need to know and prejudge how the helicopter will act during your next move. Not much of an improvisation inflight, at least for such a noob as I.

Update:
Check this vids and see how unstable it is:

[Koty]

Thanks for the links. I'll go through them. My Joystick has a very stiff spring and is not really accurate so I tend to overcompensate, unless flying at over 160 km/h. Then it kind of stabilizes, but not in roll. It is also horribly oversensitive to any kind of input as I modeled it. With +- 1° deflection, it can do a 360° in something over a second. I will play a bit more with the flight model, but there is only little info I have been able to find on helicopter physics online.

[Koty]

Tried adding more detailed code for the rotor blades to run side-by-side, but it was too much. It was lagging terribly.


EDIT: Turns out, problem was elsewhere.

[Koty]

Code for automatic anti-rotation stabilization done. (Heli puts some force against helicopter rotation, by deflecting blades.)

Next up will be force-trim code (as in real helicopter). Turns out it's actually easy to do. The principle is that when you need to keep your stick forward, you press the button and it relieves the forces on the stick, so that that position acts as new center (new 0-0 position) for it. That is, with forced feedbak (controls in helicopter). Since we mostly use spring-loaded joysticks, it will be done so that you press the button, center the stick - but it will still remember how much you are pushing. So any new deflection of the stick will be done from that remembered position to the front.

(I hope someone can do a clearer explanation.)

The actual trim will be used for autopilot. The autopilot is trying to level the helicopter, where the desired attitude can be changed. Which means that if you let go of controls with autopilot on, it will level the helicopter. Good for hovering.

[Koty]

And here we go,
here is a Kappa version. (In the first post is a Beta version... so to differentiate and since I am Koty - Kappa.)

https://www.mediafire.com/file/2ccb49kdj6zwc6f/%21++000000_Mi-24V_Kappa.rar
(put this on top of your #SAS folder!)

Features:

> New semi-realistic flight model (few features were left out, since they just compensate for each other)
 >>Lift is now independent on game's .emd code!
 > Working force trim
 >> Press Misc 1 to trim the helicopter
 >> Press Misc 2 to reset trim
 > !!! Side feature - overhauled rocket spread
 >> May not be online-compatible
 >> Will cause problems when playing replay recorded after the session end, if you want accurate replays, you have to initiate them during the mission
 > A few more loadouts, will need 3do folder in Su-25 mod (should also be in Vega's weapon mod) to run them
 >> includes B-13 rocket-pods and make-shift GSh-23 gunpods
 >> Yak machinegun now performs like a giantic shotgun, just like the real thing

Flight procedure:

> CHECK THROTTLE AND PROPELER PITCH 0%
 > Start both engines and let them spool up
 > Engage  brakes
 > Throttle to full
 > Start gently adding propeler pitch until the helicopter starts lifting
 > Gently correct for torque or any other kind of movement

 > You're flying

Known bugs and missing flight model features:

> No retreating blade stall (you can go as fast as you can correct with cyclic)
 > The .fmd flight model gets a bit "shaky" when going over 300 km/h - the tail occasionally twitches up and down when flying at negative AoA (nose-down, normal attitude for flight forward)
 > The .fmd flight model reduces area accesible with cyclic at speed (increased forces), you can eliminate that with trim - I am thinking of actually increasing the effect of this, the controls will be stiffer, but will make trim useful - I would like your input here

The helicopter still does not have an autopilot, that will be next step.

[Fresco23]

I don't currently use a setup this will run in... but I am silently watching this thread from the sidelines. Looks like this will be lots of fun soon!

[BT~wasted]

First of all I want to say that the project is not dead!

Second is that I am amazed by the work that Koty did! We now have the most real representation of the helicopter in IL-2. But seems to me that it needs a bit of tweaking. If it is possible, Koty,  can you explain the values in your code, please?

[Koty]

Sure, I'll sit to it in the evening. Most of it are guestimates, and they do need tweaking... (I also did a full fidelity model of main rotor, but scratched that - it is uncontrolable with the delay on controler that I could not remove + it needs the pilot augmentation modeled, what I did is a good aproximation for how we are limited on info/controls.)

[BT~wasted]

What I want to do is to add more sensitivity at high speeds on the "rudd" channel. It loses too much of a control at high speeds. On the contrary I think it would be great to reduce sensitivity on the "aile" channel.

Also I can't understand if the ground effect is modeled?

[Koty]

1) ground effect not modeled
2) rudder sensitivity does not change. The FM takes into account the fact that it has stabilization surfaces, this is realistic behavior. Although I will have a look into my code if I did not do some duplication or something.
3) actually sensitivity should be high, the problem is I cannot transfer control input in real time - the game puts a delay on it, as if there were forces in the controls that pilots has to overcome. (Also one of the reasons why I did it the simplified way.)


EDIT: Also, the cockpit looks good so far

[BT~wasted]

I would me more then glad to see your source code with some at least brief explanations of the lines and formulas. I have been out of IL-2 modding for years and I hardly can read my own code, and when I look at your code my brain starts to melt a bit 

EDIT:

Cockpit is almost ready, some cosmetics left.

Also the new Sturm-V missile guidance logic is introduced to achieve a more realistic behavior. But I need to consult real Mi-24 pilots about the actual guidance system.

[Koty]

The guidance should be either laser guidance or line-of-sight alignment. One of those two. Both using the gunners sight.

I'll look into the file as soon as possible, fixing bombing mode on MiG-21