Why is drag so high?

[NERVAfan]

I understand that "mass based drag" is there as a simplification... but why does the formula make the drag so high? According to the wiki each ton of mass is treated as 8 square meters of frontal area... I don't see how you could even make a rocket that wide. The FL-T100 tanks are 0.5625 tons and 1.25 meters diameter... that's pi*0.625 squared or about 1.23 square meters.

So even a rocket made totally of FL-T100 tanks, one tank high and many tanks wide (and with no engines, capsules etc. in-line to add mass without increasing area) would have about 2.2 square meters per ton, not 8.

[phoenix_ca]

Stock aerodynamics are dumb. Dumb dumb dumb dumb dumb. The only people that could probably answer that question adequately is Squad.

[r4pt0r]

if they could link drag to the reentry effects, then drag would only effect parts hit by the drag. would that work? (obviously keep calculating drag even when the reentry effects aren't visible)

[phoenix_ca]

FAR does that, more or less (no weird shenanigans with the reentry effects required).

[r4pt0r]

i was just thinking that would be a way to get the game to understand only apply drag to the leading surfaces.

[Geoff-AU]

install FAR

[Tortoise]

FAR should be installed.

[Supernovy]

I'm just guessing here, but it may be to do with balance. Perhaps they wanted to get a different difficulty, or feel, for rocket ascent. Considering the small scale - and thus low orbital velocities - of the Kerbol system, making draggier atmospheres may make sense to compensate for the lower Δv transfers.

Of course, this was all before aerobraking or other planets, so that also could have affected it.

if they could link drag to the reentry effects, then drag would only effect parts hit by the drag. would that work? (obviously keep calculating drag even when the reentry effects aren't visible)

FAR does that, more or less (no weird shenanigans with the reentry effects required).

You may be thinking of Ferram's Raycast Drag Experiment.

[Eric S]

Drag is so high because the current aerodynamics system was put together as a placeholder, just something to work enough that squad could focus on the parts of the game that they considered more important until those more important parts were polished enough that they could come back and do an aerodynamic system that was up to the standards of the rest of the game. This isn't saying that the existing aerodynamics might not become the real one if the improved aerodynamics plans fall through the cracks.

Basically, it's primitive and inaccurate because they didn't want to put effort into trying to make it more accurate when they planned to throw it out and replace it with something better at some point.

[m4v]

Because Kerbin's atmosphere doesn't go up to 100km and neither you need a velocity of 8km/s for reach low orbit. You can't talk about realism in a planet that is artificially scaled down.

This is why I dislike most realism mods that exists for the sake of realism alone, FAR might be realistic, but applied to Kerbin it means that I can do cartwheels with a rocket at 10km and still reach orbit with more fuel to spare than I could get in my best stock launch.

Edited May 20, 2014 by m4v

[Red Iron Crown]

The stock aerodynamic system was developed before we had spaceplane parts, I don't think it was ever intended to be anything but a very rough approximation of atmospheric drag. Once C7's spaceplane parts mod exploited the aero control code to make aircraft possible, the limitations of the primitive model became more apparent. It really should have been replaced by something better when those parts were made stock, IMO, but it didn't play out that way.

[shizophrenic]

/sarcasm start

sometimes i wish they go to a Company with Expertise making extremly accurate air flow and drag modells like Boeing, take those models and put them 1:1 into the game. Just to add realism. i think the FAR planes (never installed so only guess) would fly only like arrows... Well, probably but no one would ever see a plane flying anymore since calculating 10 sec of flying would take a loooooooonnnnnggg time. ... i don't have one of the supercomputer as ring with several 100 cpu and Terabyte of Memory. ah and while talking of realism, get rid of the time warp button, too, /sarcasm end

[phoenix_ca]

Time warp is clearly cheating.

[Comrade Jenkens]

Sadly it seems that a lot of ppl actually prefer the soup drag model to one slightly more realistic. :/ So maybe it won't get updated.

[DaMichel]

... You can't talk about realism in a planet that is artificially scaled down. ...

There is a difference.

KSP orbital mechanics obey real world laws of physics. Granted the density of the planets is unrealistically high. But this is just a parameter and the theory still works. The stock aerodynamics on the other hand obey different laws than IRL. For one the drag force depends on mass instead of the surface area and basically the shape of an object.

Edit: You could say that the mass is a rough approximation of the size of an object. This however doesn't work very well. It has for example the consequence that all objects fall at the same terminal velocity! I.e. a parachute (probably a bad example because it has special case handling in KSP) would fall at the same speed as a steel ball of the same mass if this was IRL! This is also the reason why aerobraking is so easy to predict.

For an object with constant mass density, say a sphere or a cylinder (fuel tank ), its mass scales with its size (diameter etc.) cubed if you scaled it in every direction equally, but its surface area grows with its size squared. Due to this scaling difference KSP overestimates the drag of large objects and underestimates the drag of small objects. An easy fix would be to use a scaled mass Area = mass^(2/3) or an eyeballed coefficient instead of the mass.

There are more issues with how lift works which i'm not qualified to write about though

Edited May 20, 2014 by DaMichel

[phoenix_ca]

Well, no, the orbital mechanics don't obey real-world physics either. It uses patched conics which is a rough approximation of what really happens at best. N-body solutions would be far more realistic, but going with a solution that doesn't require a physics degree to understand certainly makes sense. (And I mean for the devs. Plotting a course with n-body solutions could be easy, but making such a thing work in-game is non-trivial.)

Realism is a poor argument for game design.

[cantab]

I don't know if it's what Ferram uses, but Blade element theory would probably be good for the aerodynamics. It's what X-Plane, probably the most realistic mass-market flight simulator, uses, and that evidently manages good performance on desktop PCs. X-Plane also includes, or did include, the ability to simulate flight on Mars.

[shizophrenic]

(And I mean for the devs. Plotting a course with n-body solutions could be easy, but making such a thing work in-game is non-trivial.)

Realism is a poor argument for game design.

Actually it is mathematicaly proven that for N>3 Body there is no exact solution. for n=3 there is only for 2 Special cases a solution there (one is the Lagrange Points, the second is about three Bodys with the same mass on a Special trajetory). All other solutions use approximations and/or are solved numerical.

Comrade Jenkens

Sadly it seems that a lot of ppl actually prefer the soup drag model to one slightly more realistic. :/ So maybe it won't get updated.

even in case your post doesn't wasn't aimed at my post before. even if my post was exaggerated, it Points out that more realism doesn't equals better. the same applys to less realism. it isn't automatic better either. often it is argumented that more realism is better/more fun, that is just simply wrong.

[Guest]

Actually, more realism equals better in KSP's case. KSP serves not only as a game, but an educational aid. Patched conics is a good approximation, and you can grasp orbital physics with it, even though the nuances differ. Stock aerodynamics, on the other hand, teach some really bad habits and are not an approximation of any kind. Even worse, some people might assume they are, in fact, realistic, much like the rest of the physics in KSP. As such, Squad definitely should correct the aerodynamics system at least to FAR level.

Actually it is mathematicaly proven that for N>3 Body there is no exact solution. for n=3 there is only for 2 Special cases a solution there (one is the Lagrange Points, the second is about three Bodys with the same mass on a Special trajetory). All other solutions use approximations and/or are solved numerical.

Doesn't mean you can't put it into the game. There's a mod in development that does just that. There are numerical integration methods that can work quite fast, even if you do have multiple massive bodies. The downside is, the Jool system would have to be rearranged, since it loses two moons some years into the simulation. This can be done, and I've been suggesting it a few times. Of course, that would require Squad to actually hire an astrophysicist as an advisor/maths dev, since the response I've gotten to my suggestions clearly indicated they don't know jack about actual N-body modelling and numerical integration.

[shizophrenic]

you bring that for educational purposes it is better to add realism. i disagree, you have to consider the audience you are teaching. if you take ppl that are study physics a great amount of realism might be good. if you take a class of 14 years old you want to get started physics it is different, since they probaly could easily overstrained and lose interest forever. even the early lessons in physics study start with simple Problems. more realism doesn't lead atuomactically to better educational value.

and i never stated that i am against N-Body Solutions in game, just stated that it isn't as easy as it seem. One main Problem i see is that the planets/moon orbits are on the rails, and there is a discrepancy to rest of their data.

[phoenix_ca]

As I said, n-body solutions for KSP would be non-trivial.

you bring that for educational purposes it is better to add realism. i disagree, you have to consider the audience you are teaching. if you take ppl that are study physics a great amount of realism might be good. if you take a class of 14 years old you want to get started physics it is different, since they probaly could easily overstrained and lose interest forever. even the early lessons in physics study start with simple Problems. more realism doesn't lead atuomactically to better educational value.

I don't know about you, but I won't forgive the educational system for confusing me by using Bohr diagrams all the way until Grade 12, where they finally ripped all that silliness up and used orbitals to explain things. (I already knew about orbitals before Grade 7 or so so I was even more annoyed.) Orbitals make a lot more sense anyway, so why not teach what's closest to reality first? They aren't all that difficult to understand. The actual math behind them, sure, but the concept is pretty easy.

N-body math though, that **** is hard.

Edited May 20, 2014 by phoenix_ca

[cantab]

even the early lessons in physics study start with simple Problems.

Yes, but then if an experiment is done that experiment is done in the real world, which still has the full complexity of real physics.

Likewise, KSP could model 21st century physics in the most accurate available methods, and it would still be useful for learning about Newtonian mechanics and Keplerian orbits.

[shizophrenic]

this was lost in copy/ paste

I my personal opinion it is better to have a less realistic model for education. The students see something that works generally but they see that it doesn't work like the real world. with asking why they think it isn't like they expect they have to look into the underlying physic. They have to Analyse the physics come up with the improvements by themself. Giving them a premade solution that is quite accurate, there isn't any Need for them to think for themself. Best would it be if they could Experiment with the physics model in the game to test out their solutions

[Dkmdlb]

I don't know about you, but I won't forgive the educational system for confusing me by using Bohr diagrams all the way until Grade 12

How about this one?

"Your lungs are like two balloons."

[DaMichel]

Well, no, the orbital mechanics don't obey real-world physics either. It uses patched conics which is a rough approximation of what really happens at best. Snip.

Right, i see. Ofc. there is no universal theory that describes everything and physical models fail at one point or another. Still you would probably never deny that Newtons Laws describe real-world physics because it is only valid for when you are going much slower than the speed of light.

The thing is that i think that KSP aerodynamics fail way too much at predicting real world stuff that it is almost like some sort of fantasy game mechanic. In a game about a space program going through the atmosphere should work in a believable way.