What's the current problem with aero, anyway?

[Laie]

I almost feel silly to ask, but I haven't played too much this week and from the thousands of posts I couldn't gleam more than that people are unhappy.

As I understand it, 1.0.0's atmosphere made it very difficult to reenter a spaceplane without burning up. The new 1.0.2 atmo has so much stopping power that in most cases, a heatshield is no longer required.

Question now is, did I get that right?

[Gaarst]

That's basically it, but less exagerated.

1.0.0 atmosphere had a more realistic behaviour than 0.90, and thermal effects were high, ie: easy to heat a craft with engines or reentry; and chutes were OP: you could open them at 2000 m/s, get >30g but be fine overall and fall slowly.

1.0.2 atmosphere has had a lot of drag added, thermal effects reduced (easier to enter atmosphere) and chutes were corrected.

[Laie]

I still don't get what chutes have to do with it. The main problem was that you could deploy them at $RIDICULOUSSPEED, the smaller problem was that their semi-deployed drag seemed to be rather high. But I don't understand how fixing one or the other required changes to the atmo as such.

[Ulzgoroth]

I still don't get what chutes have to do with it. The main problem was that you could deploy them at $RIDICULOUSSPEED, the smaller problem was that their semi-deployed drag seemed to be rather high. But I don't understand how fixing one or the other required changes to the atmo as such.

Is there any reason to think those things are connected in the first place?

The parachute changes are relevant to the other changes from a play perspective, since both are major factors in reentry, but they don't need to have a common cause.

[Aerindel]

The problem was that the air was so thin that the only way to slow down was use parachutes while you where still going insanely fast. If you make the parachutes behave realistically then that means the air has to be thick enough to slow you down to something near terminal velocity before you smash into the ground. Apparently this creates a problem for airplane program enthusiasts who want to fly panel trucks to orbit.

- - - Updated - - -

I think the root issue is that in real life the atmosphere is many times higher, giving a much longer fall time so that it can be both thin and yet still be able to slow spacecraft. Kerbin's air is just thin shell yet it has the same gravity as earth so if the air is not thicker you can't re-enter without invincible parachutes. The very fact that Kerbin is an impossible planet makes realistic flight and realistic re-entry mutually exclusive.

I doubt when squad first came up with this system they ever imagined so many people would want to fly airplanes but now it is what it is.

Maybe Kerbin just needs a twin planet named Plankin or something with 1.0 atmosphere and a SPH only KSP on its surfaces. At least that way we could all share the same basic laws of physics.

Edited May 5, 2015 by Aerindel

[Laie]

Isn't that directly rooted in Kerbin being so small?

I'm reminded of a what if, From what height would you need to drop a steak for it to be cooked when it hit the ground?

Putting it all together, I found that the steak will accelerate quickly until it reaches about an altitude of about 30-50 kilometers, at which point the air gets thick enough to start slowing it back down.

The falling steak’s speed drops steadily as the air gets thicker. No matter how fast it’s going when it reaches the lower layers of the atmosphere, it quickly slows down to terminal velocity. It always takes six or seven minutes to drop from 25 kilometers to the ground.

For much of those 25 kilometers, the air temperature is below freezingâ€â€which means the steak will spend six or seven minutes subjected to a relentless blast of subzero, hurricane-force winds. Even if it is cooked by the fall, you’ll probably have to defrost it when it lands.

I wonder if it would help to extend the atmosphere, to 100 or 120km. The lower 20 or so should scale very much like Earth's, further up it may get thinner more quickly than in reality. I don't think launches would take much longer than they currently do, you still only need to accelerate to a good 2000m/s.

Or would it? Even when returning from Eeloo, we're only coming in half as fast as any orbital vessel on Earth. Squad probably has to crank up the heating or there will be no need to worry about it.

Edit: Maybe it would have helped to change the parachute's default pressure from 0.04. I habitually set them to 0.5, the highest possible value, and in v1.0.0 I was going like 200m/s@5km when they semi-deployed (IIRC). Looked credible to me, at any rate.

Edited May 5, 2015 by Laie

[uglyduckling81]

I tried several times in 1.0 to do emergency separation staging when a rocket was starting to lose control. Only to have my pod slam into the ground despite having a chute deployed.

The chutes seemed to do nothing when being used in relatively low altitude (couple thousand metres).

When I was paragliding my emergency parachute needed 1000ft or 300m to deploy and slow me down enough to be effective according to the manual. So it makes sense a chute shouldn't save you when deployed 10m from the ground but a couple thousand metres should be fine.

[luizopiloto]

Aero is fine... jet engines and MKIII parts not..

[Superluminaut]

I almost feel silly to ask, but I haven't played too much this week and from the thousands of posts I couldn't gleam more than that people are unhappy.

As I understand it, 1.0.0's atmosphere made it very difficult to reenter a spaceplane without burning up. The new 1.0.2 atmo has so much stopping power that in most cases, a heatshield is no longer required.

Question now is, did I get that right?

No, reentry in 1.0.0 was easy. A heat shield was unnecessary in 1.0.0 as well.

1.0.0 had a lot of bugs, and for whatever reason aero was adjusted.

[Windspren]

The problem was that the air was so thin that the only way to slow down was use parachutes while you where still going insanely fast. If you make the parachutes behave realistically then that means the air has to be thick enough to slow you down to something near terminal velocity before you smash into the ground. Apparently this creates a problem for airplane program enthusiasts who want to fly panel trucks to orbit.

- - - Updated - - -

I think the root issue is that in real life the atmosphere is many times higher, giving a much longer fall time so that it can be both thin and yet still be able to slow spacecraft. Kerbin's air is just thin shell yet it has the same gravity as earth so if the air is not thicker you can't re-enter without invincible parachutes. The very fact that Kerbin is an impossible planet makes realistic flight and realistic re-entry mutually exclusive.

I doubt when squad first came up with this system they ever imagined so many people would want to fly airplanes but now it is what it is.

Maybe Kerbin just needs a twin planet named Plankin or something with 1.0 atmosphere and a SPH only KSP on its surfaces. At least that way we could all share the same basic laws of physics.

This is is exactly right, which is why I think that for the next update we need realistically sized planets along with GP2 and fixed water.

[DuoDex]

Punted to SD&D as it's more of a discussion behind the principles of aero.

[Laie]

This is is exactly right, which is why I think that for the next update we need realistically sized planets along with GP2 and fixed water.

God forbid no. Real-sized planets would double or triple the time to orbit. That's great if you like it (in which case: mods), but it shouldn't be forced on the more casual players.

[zilfondel]

Isn't that directly rooted in Kerbin being so small?

I'm reminded of a what if, From what height would you need to drop a steak for it to be cooked when it hit the ground?

I wonder if it would help to extend the atmosphere, to 100 or 120km. The lower 20 or so should scale very much like Earth's, further up it may get thinner more quickly than in reality. I don't think launches would take much longer than they currently do, you still only need to accelerate to a good 2000m/s.

Or would it? Even when returning from Eeloo, we're only coming in half as fast as any orbital vessel on Earth. Squad probably has to crank up the heating or there will be no need to worry about it.

Edit: Maybe it would have helped to change the parachute's default pressure from 0.04. I habitually set them to 0.5, the highest possible value, and in v1.0.0 I was going like 200m/s@5km when they semi-deployed (IIRC). Looked credible to me, at any rate.

Hmm. By comparison, Alan Shepard's Freedom 7 flight during the Mercury-Redstone program - first American in space - achieved speeds up to 8,340 km/h, or 2,316 m/s. The mercury capsule heat shield apparently didn't get very hot (couldn't find temp info on the suborbital flight). By comparison, it would heat up to 3,000 degrees on an orbital flight, while travelling ~8 km/s.

[Aerindel]

This is is exactly right, which is why I think that for the next update we need realistically sized planets along with GP2 and fixed water.

Maybe. But that would create another issue, with a realistically sized planet it would take about eight minutes to get to orbit instead of the two that it takes right now. This is would be great for realism but I could see it turning off a lot of players.

[Giggleplex777]

WinterOwl explained what's wrong with the new atmo here:

I believe everybody who have been griping about the drag in 1.02 are misunderstanding the problem. They see aircraft flying unreasonably slowly, and naturally blame the drag settings. Naturally... but wrongly.

The drag is fine. Capsules slow down in reentry just right. (Heat does need fixed.) Airplanes in lower atmosphere slow down for landing just fine. No problem with the drag, IMO.

The problem in 1.02 is the lift is too high. Not the drag.

In the 1.0 physics.cfg, liftMultiplier = 0.038

In the 1.02 physics.cfg, liftMultiplier = 0.055

That's a pretty big change. The values for liftDragMultiplier and bodyLiftMultiplier both changed as well.

As a very easy experiment, I designed and flew a small, simple airplane in 1.02 aero, noted that it landed at a very slow EAS. Next I changed liftMultiplier to 0.038 and flew that same airplane. It took off and landed at faster airspeeds, subjectively felt like more of hot rod, and was overall more fun to fly. I'll leave the number crunching to smarter heads, and only say that changing that one lift value while leaving the drag at 1.02's setting just plain felt right; it felt like how a little jet airplane is supposed to fly.

[Scoundrel]

The planets are fine.

Personally I don't care about the aero, because the focus of my games are everywhere but Kerbin. But that doesn't mean that how others play the game is any less valid. Spaceplane and SSTO strategies have long been a part of space programs (in terms of goals, as none have yet to actually fly), and with the developers including contracts that require flying about the planet doing surveys, aircraft and glide-recoverable SSTOs have become an essential part of the game, and that in turn means that aerodynamics deserve serious attention.

Ferram's FAR works because he knows his stuff and has had at least a year and a half to fine tune it for KSP (it is currently 0.14.7 IIRC). Mu basically had 6 months to learn the basics and figure out how to implement that in KSP (along with heating and bug fixing), and like FAR, it will take many iterations of messing around before nuStock gets aero right as well (it's basically FAR 0.2.0 atm, so a ways to go yet).

IMHO Squad could have avoided most of their aero issues (and the resultant growing pains that the community is going through right now) if they had simply contacted all the modders who worked with Ferram on FAR and worked something out so they could have hired Ferram as a consultant or something. Right now we get to go through what the early FAR adopters went through and it will take about 6-8 months before the community has caught up with them in terms of what "normal" is.

[sumghai]

I'll also like to add that the current stock aero system requires special PartModules and handling for special parts such as fairings (top-mounted payload and interstages), cargo bays, boost protective covers etc, which is unfriendly to add-on authors.

The upcoming version of the FAR add-on uses voxel-based aero drag occlusion, and ferram has noted that his new system would automatically handle all possible fairing/cargo bay systems via inherent emergent behaviour.

I suspect that SQUAD didn't simply adopt FAR due to both technical and non-technical considerations, which for certain reasons I am unable to elaborate.

[allmhuran]

With the caveat that I don't think aero is as gamebreaking as some do right now, here are some things that I find a bit strange when playing 1.0.2:

1) Large, heavy craft can land at quite low speeds with rather small wings.

2) (This one's going to be long). Jet engines have a bit of an odd curve with speed and altitude. Takeoff and low altitude is pretty OK in 1.0.2, and the sound barrier does actually have some "barrier" to it, which is good. But as you gain a bit more speed and a bit more altitude, your acceleration skyrockets. You can easily build a craft that has trouble getting through mach 1, but a few seconds later is starting to overheat due to the speed at which it is flying. Then just a few seconds after that (and just a few thousand meters higher) your engines run out of thrust because they have exceeded their peak power output and because of what seems like two rather significant steps in air density that seem to happen at ~20,000m - 27000m. Put all of this together and it makes aircraft a bit weird to fly. Now, there may not actually be two significant steps in air density in that altitude range, but that's kinda the point.... whatever is going on, it just seems odd.

This impression might be exacerbated by the huge step in technology between the basic jet engines and the turbojets/rapiers. Basic jet engines seem pretty OK, but turbojets will easily push you to speeds that will cook your aircraft. Compare this with real jets, where the limiting factor for your engines is almost never that the aircraft is going to overheat... it's that the engines just can't push you any faster. The only aircraft with engines that even come close to the performance of KSP's turboramjets would be the SR71, and yeah, that thing got really hot and had to be made out of titanium because of it. But check out the engines on that thing! Each of the Pratt & Whitney J58 engines puts out 35,000 pounds of thrust, and are very specially designed around their bypasses and intakes for super high speed flight. The F100 engines on an an F15 each put out 14000 pounds of thrust at a TWR of about 7 to 1, and the F15 is itself known as an energy fighter (very high power). The General Electric CJ610 engine (used on, eg, a Lear Jet) puts out about 3000 pounds of thrust.

So if the basic jet engine matches the Lear Jet, and the turboramjet matches an SR71, there's a pretty huge gap in the middle that represents engines for the kinds of planes that many people want to build.

Now, rapiers work a lot like the turboramjets, except that they have a little less thrust at lower speed, max out thrust at higher speed and have a higher max thrust than the turboramjets. I think rapiers are just about right for spaceplanes in terms of their high altitude, high speed performance. But they do still behave like the turboramjets when climbing: An aircraft that can barely push through mach 1 can easily be burning itself up just a few seconds after pushing through. Odd.

3) Apparently re-entry heat is largely irrelevant, but I can't comment on this personally because I haven't played with rockets much yet in 1.0.2.

4) Mach 1 represents a point of very high instability for rockets. 1.0 did this as well, but to a lesser degree which was, in my opinion, right on the money. Whether by design or luck, in 1.0 keeping your navball chevron inside the prograde ring would keep you from flipping. This meant that the interface lined up really nicely with what it meant to fly a craft through a stable launch. In 1.0.2 that's no longer true, and some rockets (not super crazy ones either) will flip even if the chevron seems, visually, to be right in the middle of the prograde ring.

Edited May 6, 2015 by allmhuran

[BmB]

You could re-enter anything safely in 1.0 using airbrakes. Wayyy overpowered.

I did re-enter a spaceplane in 1.0 without too. Trick is to come in shallow and use the wings to brake high before you hit the bulk of the atmosphere, it takes a while of skimming the atmosphere at 45 degrees angle of attack but you come down safely with almost no plasma. Not quite realistic but in the absence of shuttle-like heatshielding for spaceplanes it's better than nothing.