YASim vs. JSSBSim

[xcvb]

Yasim models are not as complex as Jsbsim models, but if something doesn't work, I think it is much easier to find the issue in Jsbsim models than in Yasim models.

[Thorsten]

When I say they are both valid, I mean they both help to achieve the goal even if into different ways and with different results.

I'd say they even have somewhat different goals. As I'm fond of saying - they're different tools in a toolkit and you should understand what each can do and select the appropriate tool for your specific task.

As you say - if you do not have data or special knowledge and want something reasonably quickly which flies credibly, you generally should use YaSim. If you have wind tunnel data of your craft, it would be bad to discard that data by using YaSim.

We know YaSim is reasonably good at subsonic speeds within a 'normal' envelope, so I doubt you'd notice much of a difference for, say, airliner operations. We know YaSim can not accurately deal with high AoA situations or spins, so it would be a poor tool for a fighter jet or aerobatics plane which you frequently operate at the edge of the envelope.

If you're interested in detailed failure modeling YaSim would be a dead end, because it does not contain the option to simulate e.g. damaged airfoils (or drastically change aerodynamics runtime) or have systems partially fail - in JSBSim none of that is a big problem.

So in my view the question is not really which is better, but which is better for a particular task. If (and that's a real if) the task is 'to be as realistic as possible' then the answer is JSBSim. If on the other hand the task is to get something flying without detailed data in a normal performance envelope, then the answer is YaSim.

[cobe571]

We know YaSim can not accurately deal with high AoA situations or spins, so it would be a poor tool for a fighter jet or aerobatics plane which you frequently operate at the edge of the envelope.

Thats the true reason cause I added JSBSim to the EXTRA-EA300 aerobatic plane

[Ysop]

From a scientific point of view jbsim is preferable. It works with forces and moments acting on masses so if everything correct all performance and handling items should fall into place and become realistic.

If.

Windtunnel data is often not available.
OpenVSP and Datcom require Information more detailed than used in the 3D-model. This is something very often and understandibly not available by the manufacturer.

Hence there are a lot of unknowns.
Is it possible to use Yasim to close some of these gaps?
For example flight testing a Yasim version and feeding derived coefficients into the Jbsim version.
Or is Aeromatic plus handtuning the way to go if no more detailed methods available?

[dany93]

From a scientific point of view jbsim is preferable. It works with forces and moments acting on masses so if everything correct all performance and handling items should fall into place and become realistic.

YASim and JSBSim work with forces and moments, calculated from tables. The difference is the way these tables are obtained.
1- With YASim, these tables come from the iteration process, starting from pre-defined ones, also taking into account the aircraft geometry and values given in the FDM file. Your control on these tables is very indirect, through the values you enter in the FDM and, at the end, the iteration process. Adjustments are kind of trial and error.
2- With JSBSim, the tables can also start from standard ones (Aeromatic), but very preferably they can be from Wind Tunnel measurements. You can also add external forces, not aerodynamic. You have a total and clear control on these tables.

Is it possible to use Yasim to close some of these gaps?
For example flight testing a Yasim version and feeding derived coefficients into the Jbsim version.

I don't think that you will get better results by taking YASim as a reference.
1- (See this response) With YASim you get what you entered, mainly at the top of the FDM. Although I admit that this can be positive.
2- YASim starts from pre-defined curves, which have no reason to be more accurate than those from Aeromatic or, if you are lucky enough to find such tables, those from some plausible wind-tunnel measurements (e.g. for wings which are close to those of your aircraft). And, of course, pre-defined tables cannot be fit for possibly non-standard wings.

[xcvb]

You can call Yasim in a console and it will give you the moments of inertia. This is all you can get from Yasim except the lift curve, which you also must define in the Yasim model. Many people say that Yasim is better if you don't have wind tunnel data, but I don't agree. In this case you can still tweak the basic lift and the basic drag in Jsbsim until it fits your needs (you can define them as a simple line with only two points). The procedure for Yasim would not be much different.

[Thorsten]

For example flight testing a Yasim version and feeding derived coefficients into the Jbsim version.
Or is Aeromatic plus handtuning the way to go if no more detailed methods available?

If you want to end with JSBSim I'd start with Aeromatic. The rather general possibility of solutions where you fit a given functional form (aka pre-defined curves) with a fixed number of parameters to some key values (approach and cruise speed in YaSim) is that you get a pathological solution, i.e. a solution which accomodates the key values just fine but makes no real sense in-between.

[asr]

Everyone is entitled to its opinion.

I'm no aircraft developer myself. I understand that JSBSim is much more complex and requires a lot of data and knowledge to create a model. Yasim on the other hand only requires a few datapoints and promises to interpolate those automagically. It's tempting to those people that don't have the data/time/knowledge/interest in creating a realistic FDM.

Having said that, as a simulation enthusiast that is always looking for the most realistic experience, I just don't fly Yasim aircraft. Not a single one of them, even those who "seem" to be the most realistic, gets anywhere near what I expect from a flight sim. To me only JSBSim aircraft exist, the rest is just eye-candy from people that are interested in creating a visual model that looks nice and then don't have the time to work on an FDM so they just patch some Yasim in them so that it pretends to fly.

[cobe571]

To me only JSBSim aircraft exist, the rest is just eye-candy from people that are interested in creating a visual model that looks nice and then don't have the time to work on an FDM so they just patch some Yasim in them so that it pretends to fly.

Probably you don't know how much time it requires to make a decent 3D model and how much time it requires to make a good FDM. The time and the nights you've to spent on it. Isn't only matter of code. The blood you've to cry to make it almost pretty good. Likable at least. Said that at the really end it will never be perfect. That's sure you don't know that efforts, and you've not respect for that people that make such kind of job for you. For free! Just remember that without that nice eye candy 3D models you'll go really nowhere or at least you'll never have that pleasure to sit on a plane. Without 3D models you'll have just a mountain of code that will animate nothing. No 3D model? Nothing to animate. So please, respect for those guys that make your dreams come true!

Best regards.

P.

[msalama]

Correct me if I'm wrong, but AFAIK Yasim doesn't model fixed-wing ground effect at all. And that's reason enough for me to prefer JSBsim A/C.

[DFaber]

[

I just don't fly Yasim aircraft. Not a single one of them, even those who "seem" to be the most realistic, gets anywhere near what I expect from a flight sim.

How can you tell, if you don't fly them? And what makes the difference if the pilot experience feels "realistic"?

Everyone is entitled to its opinion.

apparently not ...

. It's tempting to those people that don't have the data/time/knowledge/interest in creating a realistic FDM.

Well, yes. An Aircraft is not only an FDM. Pilot experience is not satisfying, if you sit in mid-air with only a HUD.

... the rest is just eye-candy from people that are interested in creating a visual model that looks nice and then don't have the time to work on an FDM so they just patch some Yasim in them so that it pretends to fly.

Time is indeed a factor for me, since I don't get payed for FG development and have to work for a living. I enjoy modeling, Nasal hacking, and flying, but not being attacked for the choice of my tools. So, if you want better FDMs, stop whining and make one. I would be glad to get a well tuned and realistic JSBSim FDM for the SpitfireIX from you. I could concentrate on other thinghs then. Open Source is about cooperation, you know.
But beware, I've done a lot of reading. A quick Aeromatic FDM won't do.

Greetings

[danielHL]

Correct me if I'm wrong, but AFAIK Yasim doesn't model fixed-wing ground effect at all. And that's reason enough for me to prefer JSBsim A/C.

You're wrong.

Model.cpp, line 357

Code: Select all

// Account for ground effect by multiplying the vertical force
    // component by an amount linear with the fraction of the wingspan
    // above the ground.
    if ((_wingSpan != 0) && (_groundEffect != 0 ))
    {
      // distance between ground and wing ref. point
      float dist = ground[3] - Math::dot3(ground, _geRefPoint);
      float fz = 0;
      float geForce[3] = {0, 0, 0};
      if(dist > 0 && dist < _wingSpan) {
        fz = Math::dot3(faero, ground);
        fz *= (_wingSpan - dist) / _wingSpan;
        fz *= _groundEffect;
        Math::mul3(fz, ground, geForce);
        _body.addForce(geForce);
      }
      if (_modelN != 0) {
        _gefxN->setFloatValue(geForce[0]);
        _gefyN->setFloatValue(geForce[1]);
        _gefzN->setFloatValue(geForce[2]);
        _wgdistN->setFloatValue(dist);
      }
    }

Cheers,
Daniel

[msalama]

Right, so I am. Yet, none of the Yasim A/C I've tried so far seem to have it implemented, or if they have, the effect is really weak compared with, say, the default JSBSim C172. Or am I wrong again?

[enrogue]

It can depend in the aircraft & how it's implemented - it depends on yasim correctly detecting the wingspan, which can be an issue if the FDM creator uses mstabs instead of more recently implemented multi-section wings

I do see it working on more recently updated aircraft

[msalama]

OK, good to know. Gotta check some of those out. Thanks.