Okay, just a few hints:
1.) the reversers can be built in JSB only as "reversers", means, you can set the angle of the thrust around. You can copy it from the Ruslan, the 767s, the Loreley, it is in so many planes, really.
Code: Select all
togglereverser = func(eng) {
if (getprop("controls/engines/engine["~eng~"]/throttle")<0.1) {
if (getprop("engines/engine["~eng~"]/reverser")) {
setprop("fdm/jsbsim/propulsion/engine["~eng~"]/reverser-angle-rad", 3.14);
setprop("engines/engine["~eng~"]/reverser_pos_norm", 1.00);
} else {
setprop("fdm/jsbsim/propulsion/engine["~eng~"]/reverser-angle-rad", 0.00);
setprop("engines/engine["~eng~"]/reverser_pos_norm", 0.00);
}
}
}
This is a little Nasal to toggle the reversers. eng=engine number starting with 0. Now this one controls if the throttle is down before it allows you to set reversers or go back forward, like in reality, if you don't want that, just cancel the outer "if".
2.) "Bouncy flight controls". This is not FDM, that is AP. Your controllers overreact, either because you use a too coarse granularity on PIDs or some limiters are too wide. Not sure. But if flight controls like the elevator go up and down like a shiver, it is AP, never FDM, the FDM only take the settings the AP makes and simulates the aircraft accordingly.
Having said that, it can be, the AP overreacts on a balance or moment problem. The first thing, one has to realize when going from Yasim to JSB is, that JSB actually simulates the momentums, Yasim does that only rudimentary. So if you start a motion, for example starting to pivot the plane nose up, the momentum doesn't end the same 1/100s you move the elevator, the momentum needs to be stopped and reversed. That is how real world physics works. So, if you AP goes like every 1/60s and checks and finds, "oh my nose is still not down", he sets the elevator more to press the nose down. Means, after some cycles, you have stopped the old up momentum and have now an extreme elevator position and create a gigantonormous down momentum, which you at the end of the motion try to stop again instantaneously. Which of course doesn't work, so your AP does the same overreaction now in the other direction. I am very sorry that real world got in the way of Yasim dreams.
3.) If you still think, it is an FDM problem, for example a balance problem, fly her manually in different load situations. You can't even test for balance problems with an AP on because the AP would either add his own problems to everything and you don't see any problems actually coming from the FDM or in case of my APs (never said, mine are perfect), they would just level the problem out and fly the plane, regardless how lousy the balance is and all you would see would for example some nose-up or down attitude after she levels out. Which makes FDM balance problems also virtually impossible. So, fly her manually, see, if you have to do a lot of correcting and namely look for extreme reactions for example with throttle changes (in many planes, the thrusters are actually lower than the center of gravity, that gives planes a slight up-momentum when the throttle is pushed)
So, bottom line:
@Gabriel: Yes, do systems and model and all the bells and whistles or you end up like I did with the Ruslan, she flies perfect, but model, cockpit, systems ... duuuh ... and as long as Israel is caught up with RL, I don't get model work done.
and as my luck goes, when Israel si back, I am busy with RL. So, do what you can while you can!
@Joshua: This happened now with how many planes, always this term "bouncy" and I still try to understand what you try to tell me ... which makes me think more and more, the whole thing happens because the JSB planes do something the Yasim planes do only very rarely and rudimentary: Simulate momentum. Fly her manually and if that works, try to run your controllers with less ki and adapted ti and td. If you have to fight with reaching the programmed altitude (too much overshooting for example) or course (be it magnetic or true, overshooting usually), take a look at the times of your predictors and make them longer for a test.
@Vincent: Planes like this one are some kind of work of Art. What is not needed is crying about the things you want or don't want but rather a pilots with a very sensible grip on the joystick. So, take her out, fly her manually till you know about the balance, take her through the turns with different loads, provide the devs with data!
J
All with patience we can achieve good changes in the 737 800.If you do not understand me, I'm sorry, I'm trying to write my best.
