DCS vs Other Sims, what makes the difference FM wise?

[BitMaster]

Guys, pleae, calm down :=)

 

I didnt wanna toss this convo into a battle R/C vs the rest of the world LoL

 

I have studied Physics to some degree and I know Galileo and Newton pretty well and Reynolds numbers only play a role at certain attitudes of flight on dont play a significant, next to zero, in others, for example a perpindicular dive where G_effectiv is zero, just to say how much ot differs and fluent it is.

 

I think, a 200kg 7m Wingspan "modell" going 400km/h can very well be compared to a cessna 172, Piper Cup etc..

 

For sure, a Jet Model of 2m Wingspan, 3m length and 25kg going 400km/h is not the same as a F-1x going 1.8 Mach at 48.000ft. I never ment that and would never say so, neither is it my intention to spread this too far.

 

All i wanted to say, there are other good Flying Simulators that let you feel

it pretty good, help to understand what it is about and give you an edge next summer. Mainly AeroFly, RealFlight or Reflex.

 

Mind you, AeroFly has a FullSize Sim with I think 7 AC to start with, more with AddOns, i just never used it and dunno how good it is compared to DCS in Physics alone.

 

DCS has it`s own league cuz it simulates more than just an AC, it simulates a whole Battlefield, now even tanks and soldiers and stuff like that, maybe ships in the future etc..

 

There is a clear point in how much a Home-PC can calculate or do we call a 6144-Xeon Cluster with 4TB RAM our own, no, we dont and we wont in the near future. Even if, calculating aerodynamics is way beyond what man-brain can put into math. Take the tree and all it leaves for just 1 second, every Physics Guy says it is impossible to say when which leave is doing what PERIOD. Nature is far ahead of us and we have to accept an uncertainty whenever we deal with fluids and aerodynamics.

 

It is easier to calculate nuclear explosions than 1 damn tree LoL

 

 

I really thought I was a good Pilot when I started R/C back in 1993, having at least 2000h on EF2000 and all that... I couldnt even land a trainer model

for many many flights. Thats is how much it differs, ok, DCS is way better than EF2000 was with physics, but things in reality have no mercy and everything counts and small counts sum up to a huge difference when you are approaching the strip.

 

Why dont you all get a small R/C airplane and we dogfight a little bit HAHA

 

It is so damn hard to hit each other even if u try but happens easily if you dont want to, same as this convo.

 

No one shall yell about one others point of view, live and let live and be happy that there are others with new oppinions that would have never crossed your mind alone. I am always happy to be tossed to new knopwledge and wisdom, regardless of how hard the toss was, the gain counts.

 

I know ask myself, cuz I just flew DCS for a few hours why does the Ka not slow down when I extend the gear, heck, I remeber times when it speeded up by 2-3 km/h and it also doesnt speed up when I drop the rocket launcher barrels.... maybe just the specific attitude I was in or one of those small things that one cant foresee when doing the FM but then it shows up as not natural. I honestly thought lowering gear causes drag and slows down ) more than a mere 1 km/h and dropping the weapons, the barrel especially, would speed me up.... it didnt, at least not noticable when at 272 km/h straight forward. I dont care about that but it makes me think how complete it can actually be and what we have to accept without starting bitching on the devs.

 

What really makes me think, and it really should you too, the actual numbers of players ONLINE. I counted them all, I keep the number for myself, but it is not what we would all like to see. This is far more important as I do know that number has relevance and even if it is only for someone to look for a nice evening flying. We should get that number WAY UP and not tear ourselves apart.

 

Bit

Edited April 25, 2014 by BitMaster

[BitMaster]

Flagrum,

 

just a bit to the math and physics.

 

NO ONE, absolutely NO ONE, can calculate it all..or "THE WHOLE" as you said. We are still all searching for this 1 equation that explains it all.

 

Unless that has been found you can always ONLY see 1 facet of the diamond at a time, while you do that, all other facettes dance on your nose and change the game.

 

This is not only DCS but symptomatic for all calculations regarding natural behaviour in dynamic environment. The deeper you look, the more you differ.

 

As a guideline, in Physics, there is a law to obey when drawing results, you can only be as precise as your most unprecise calculation is.

 

You can calculate 1 thing down to 0.000000000009 and anothe rone you do only to 0.1. The Preciseness (hope that word exists) of your overall math is ONE DIGIT behind zero = 0.x

 

And if you do not know what all needs to be in consideration, yet some are left out, you can not be precise at all. Back to sqaure 1, do it again Jack !

 

I really dont want to go into Heisenberg now, but from what I understood, we know nada of HOW it actually works in detail, WE DESCRIBE but have no knowledge.

 

The reason why NASA etc. is for sure WAYS ahead of DCS is pure calculating power and super computers with way more RAM. They can DESCRIBE a bit deeper and faster, taking more things into consideration.

This is a pure money and manpower game in this topic.

 

Still, NASA for sure knows that they actually know nothing, otherwise they wouldnt test and test and test and still fail sometimes. That's when nature takes its toll on us idiots trying to be God and think we know it all. We simply dont and never will.

Edited April 25, 2014 by BitMaster

[GGTharos]

I really thought I was a good Pilot when I started R/C back in 1993, having at least 2000h on EF2000 and all that... I couldnt even land a trainer model

for many many flights. Thats is how much it differs, ok, DCS is way better than EF2000 was with physics, but things in reality have no mercy and everything counts and small counts sum up to a huge difference when you are approaching the strip.

 

Um, well ... you also aren't landing from inside the pit ... or with the same controls. I'm pretty sure any competent pilot would have a good chance of crashing an RC on his first go.

 

Why dont you all get a small R/C airplane and we dogfight a little bit HAHA

 

I have one. I let a friend fly it, he let it fly straight 'cause he was afraid to turn it, I ask 'where'd it go?' (I was paying attention to something else for a minute or two), and he points over the river. I can't see anything, so I take the RC box back, and I turn it around. Once I can see the plan view I know where it is, and I get it turned all the way around and coming back at me. It runes out of battery and instead of a nice, smooth landing, it crashes into a wooden fence by like an inch. So it's sitting around with a broken wing right now :)

 

It is so damn hard to hit each other even if u try but happens easily if you dont want to, same as this convo.

 

Give me FPV with moveable camera at least, stick, throttle, rudders, and we'll see :D

 

I know ask myself, cuz I just flew DCS for a few hours why does the Ka not slow down when I extend the gear, heck, I remeber times when it speeded up by 2-3 km/h and it also doesnt speed up when I drop the rocket launcher barrels.... maybe just the specific attitude I was in or one of those small things that one cant foresee when doing the FM but then it shows up as not natural. I honestly thought lowering gear causes drag and slows down ) more than a mere 1 km/h and dropping the weapons, the barrel especially, would speed me up.... it didnt, at least not noticable when at 272 km/h straight forward. I dont care about that but it makes me think how complete it can actually be and what we have to accept without starting bitching on the devs.

 

It's not like there can't be an error in the FM. In the F-15 PFM there's an error where the drag of the landing gear appears to be negative for example. But as for the heli, did it start to climb when you dropped the rockets? Because the heli only has part of it's acceleration vector pointed forward ...

 

What really makes me think, and it really should you too, the actual numbers of players ONLINE. I counted them all, I keep the number for myself, but it is not what we would all like to see. This is far more important as I do know that number has relevance and even if it is only for someone to look for a nice evening flying. We should get that number WAY UP and not tear ourselves apart.

 

Bit

 

The number of online players right now is tiny compared to single play players. There are ideas to enhance the overall MP experience, but I don't know when that'll happen.

[GGTharos]

I really dont want to go into Heisenberg now, but from what I understood, we know nada of HOW it actually works in detail, WE DESCRIBE but have no knowledge.

 

Actually we do have knowledge. We know what it does. Why does it do it? That isn't necessarily something we try to answer.

 

The reason why NASA etc. is for sure WAYS ahead of DCS is pure calculating power and super computers with way more RAM. They can DESCRIBE a bit deeper and faster, taking more things into consideration.

This is a pure money and manpower game in this topic.

 

That's just not really an argument for anything. You don't run a CFD real-time right now so that's just not comparable.

 

Still, NASA for sure knows that they actually know nothing, otherwise they wouldnt test and test and test and still fail sometimes. That's when nature takes its toll on us idiots trying to be God and think we know it all. We simply dont and never will.

 

They actually know loads, and that's why they know what to test.

[BitMaster]

@GGTharos,

 

I will check the Ka thing over the weekend, will be on an emergency server rescue trip tomorrow and maybe the whole damn weekend will be spoiled but I gotta make a living somehow LoL

 

I didnt notice a lift, but regardless of the lift, drag would habe been lower resulting in a speed up somehow. Since I am curious and that has been on my mind for a bit longer I will make some tests and report.

 

Regarding your plane : OH MAN, never let a novice fly alone and not on the data-link !!! You can call yourself a good pilot if you got it back without visual initially. Happened to me too many many years ago with an Ultra Fast Hotliner, white piant, sky high and going very fast, 1 second not at it and gone it was. I recalled how long and strong I had to roll and pull and thought, that gotta be a 180 turn. seconds later a saw a tiny tiny blink in the sky and back it came to daddy, Sadly it exploded mid air a few years ago sice it was like 15 years old by then and airstream found its way into the wing while DIVING Hotliner like and it literally exploded have way in the dive. Haha I didnt mind to have it ending that way, better than a stupid Pilot error, which still happens when you think it wont. Lazy minds kill planes !

 

Bit

Holm & Rippenbruch Comrade :)

[Sid6dot7]

I will check the Ka thing over the weekend [...]

Increasing drag when extending the gear doesn't necessarily result in a loss of airspeed. Especially not on an 'instable' platform like a helicopter.

The result could also be an attitude change, which lowers the helis nose a bit and you loose altitude, but maintain your speed.

Attitude changes are energetically more efficient, so an unstable platform will likely do this. More stable platforms like airplanes have a higher resistance to such changes and therefore are more likely to just bleed airspeed.

 

But generally parasitic drag (like a retractable gear) has a lower influence on helicopters due to their lower flight speeds and their overall larger drag index compared to airplanes.

[Genchou]

About the Ka, I think you also need to take the autopilot into account. If you previously set the AP on a course with a set speed and attitude, I think it will make corrections when you lower your gears or jettison stores, in order to mitigate the changes in airspeed and/or attitude.

[Bucic]

X-Plane doesn't use simplified CFD but Blade Element Theory just like DCS.

 

http://www.x-plane.com/desktop/how-x-plane-works/

 

DCS uses apparently some more advanced methods for determining propwash, downwash and wingtip effects (and most likely some other effects). X-Plane uses some generic look-up tables for all aircrafts while DCS can have customized tables for each airframe or even some kind of simulation for these effects. Ka-50 is a good example that shows simulation of rotor interaction with each other and itself that is just unbelievable when you start to examine it more thoroughly.

Thanks for bringing X-Plane up. First, you've rectified the common misconception. Second, X-Plane vs DCS shows well how DCS differ from other producst. What I'm getting at is:

Flight Models are... models, i.e. approximations, which inherently require tuning, corrections and verification of results against real data.

It all boils down to:

The difference is in the people who are developing the models. They need to know what they're doing and what to expect.

[Bushmanni]

The thing with CFD is that it also needs to be validated with real data. You can't just put together a 3D model and expect realistic results. If you could build a model with nearly infinite amount of detail and really dense mesh (the volume where the flow flows) you might be able to do that but it would be computationally too ineffective. The 3d model and mesh will be built with detail only where it really is needed and for that you need experience or data how the flow behaves in real world with the shapes, scale and parameters you are dealing with. You might also not calculate all the possible physical effects like temperature changes to simplify the model for better computability.

 

I don't have enough experience myself about CFD to know how much error you would get with certain kind of shortcuts but this is how the pros explain it.

[Sid6dot7]

The thing with CFD is that it also needs to be validated with real data.

Yep, we would really call it 'coloured fluid dynamics' otherwise. You can calculate everything with CFD but that doesn't mean it's (physically) correct, only mathematical.

What hardly is possible are real time solutions out of exact calculations if the mesh exceeds a certain simplicity (dependent on fluid flow model and iteration methods). Complexity would be the wrong word for such a mesh.

Edited April 25, 2014 by Sid6dot7

[Bucic]

The thing with CFD is that it also needs to be validated with real data. You can't just put together a 3D model and expect realistic results. If you could build a model with nearly infinite amount of detail and really dense mesh (the volume where the flow flows) you might be able to do that but it would be computationally too ineffective.

Another good point. Even with non-realtime computations (e.g. hours for a casing stress computations) you need a man who knows what to expect and how to verify the results. Even CFD/finite element method computation packages are no 'Wizzards' that will throw precise or even correct answers at you.