Squad is slowly reintroducing soup o' sphere with 1.01

[Cerbis]

I did some back-and-forth testing of the aerodynamics and came up with some numbers. Sadly I didn't manage to get a copy of 1.0 saved away to test in.

The rocket was built in Sandbox mode with, top to bottom: Mk16 parachute, Mk1 Command Pod with a Pilot Kerbal, 1.25m separator, FL-T800 fuel tank fully fueled, LV-T45 engine set at 51% max throttle and kept there for the entire duration of the launch. SAS was enabled on the launchpad prior to ignition and the rocket was kept pointing straight up until the fuel ran out, at which point I recorded the following values.

0.9 with latest compatible version of NEAR:

Flame-out at 83,566m. Final apoapsis 351,074m

0.9 stock

Flame-out at 35,434m. Final apoapsis 90-ish km

1.0.1 stock

Flame-out at 45,626m. Final apoapsis 141,060m

As we can see, the new-new aerodynamics are still slightly more lenient than the old stock aerodynamics, probably owing to drag occlusion alongside a more "old-stock" like amount of drag. Having played with FAR/NEAR for a long time, I very much enjoyed the 1.0 aerodynamics and its lower amount of drag, but I can understand these changes if the goal is to have aerodynamics like old-stock with more realistic behaviour.

[Jouni]

The rocket was built in Sandbox mode with, top to bottom: Mk16 parachute, Mk1 Command Pod with a Pilot Kerbal, 1.25m separator, FL-T800 fuel tank fully fueled, LV-T45 engine set at 51% max throttle and kept there for the entire duration of the launch. SAS was enabled on the launchpad prior to ignition and the rocket was kept pointing straight up until the fuel ran out, at which point I recorded the following values.

The problem with this experiment is that you didn't fly the same rocket in 0.90 and 1.01. The 0.90 LV-T45 had a specific impulse of 320/370 s and 200 kN of thrust, while the 1.01 engine has 270/320 s and 168.75/200 kN.

[Frozen_Heart]

Ok I've just been testing the lifters I built in 1.0 in this version and all of them only just managed to space.

I've had to reduce the size of them all due to the engine nerfs in 1.0. Now that the thick atmo is back they don't have enough fuel due to their size, and I can't add more due to the nerfed engines.

Sure I could strap 6 SRBs to the side but I always try to make rockets look like real world ones which is pretty difficult now.

[falconek]

dv to LKO is about 100 m/s higher in 1.0.2, than it was in 1.0.0. It's merely a cosmetic change.

[Cerbis]

The problem with this experiment is that you didn't fly the same rocket in 0.90 and 1.01.

I was aware of that, but I have no idea how to change the modules for a true apples-to-apples comparison. Even then, the power and efficiency goes up quickly with even a little altitude, so the error isn't quite as large as it seems. If you, or someone else, know how to change the values, that would be better than what I did.

What we can assume though is that the new aerodynamics are more lenient overall than 0.9 stock, as flat/improved power and specific impulse will only make the rocket go further still.

[klgraham1013]

I just want realistic aero, but I'm no pilot, so it's hard for me to critique what we have here. What I do know is Squad talked a big game about testing the aero to get it just right. Then they make major changes to it a few days after release. I also remember Mu saying they were contacted by many people offering help on the new aero. Help they turned down. Mu then said this was really out of his comfort zone and one of the harder things he's done in KSP. This was all said during the 1.0 marathon.

My issue is this: If you don't know what you're doing and someone with expertise offers you help, you take that help.

So I'm left wondering if I trust Squad's attempt at realistic aero. All I can really do is listen to forum members I trust. I'm setting the game aside for now, and I'll see what people think in a few days when they've had more time with the changes.

[Jouni]

I was aware of that, but I have no idea how to change the modules for a true apples-to-apples comparison. Even then, the power and efficiency goes up quickly with even a little altitude, so the error isn't quite as large as it seems. If you, or someone else, know how to change the values, that would be better than what I did.

The main difference is the Isp drop from 0.90 to 1.01. Assuming that the amount of fuel remains the same, you get 18.5% more delta-v at sea level and 15.6% more in vacuum, if you use the 0.90 engine instead of the 1.01 engine.

[sal_vager]

I've only flown a plane once (Cessna 150) so it's hard to compare KSP to that, you don't get any real sense of just how much real planes can be like a rollercoaster from playing games (nearly lost my lunch that day).

But I play flight sims and KSP feels a lot closer to things like FS9/FSX, Flight Unlimited, Flightgear and others now than it did previously.

Now, no flight sim is 100% accurate either despite their claims, but they are good enough for entertainment.

KSP may not be absolutely accurate, but really it depends how you're judging it, if you're a real world pilot you'd have more of a sense of things, armchair pilots not so much

[Temeter]

The only thing I don't like about the aero update is that it makes planes less stable due to increased drag/lift. Stable in the sense of wobbly wings/radial sections, especially with SAS/timewarp. On the other hand the construction of an air plane makes more of a difference now, in the original 1.0 almost every plane felt ultra stable (here aerodynamically speaking).

So I'm left wondering if I trust Squad's attempt at realistic aero. All I can really do is listen to forum members I trust. I'm setting the game aside for now, and I'll see what people think in a few days when they've had more time with the changes.

In that case you'd have stock FAR integration. And FAR is imo a fantastic mod, but it's definitively not the right thing for stock KSP.

The new Aero is a compromiss: It has to be restrictive enough to allow and to some degree necessitate depth in plane construction, but it has to be lenient enough so you can still throw bricks into the sky, provided enough fuel and some special lifters. Of course that's a hard thing to do, there is no right balance. A realistic aero on the other hand 'only' needs to worry about being as realistic as possible.

Edited May 2, 2015 by Temeter

[Cerbis]

The main difference is the Isp drop from 0.90 to 1.01.

You're absolutely right, but the rocket still went further in 1.0.1 than 0.9 despite the dV difference! That's the main thing to take away from it.

[problemecium]

Update: I won't retract my earlier statements outright, but I did some calculations and it turned out that, well, I needed moar boosters. My lift rockets weren't able to get to orbit when they could in 1.0.0, but "equivalent" rockets in 0.90 had much smaller carrying capacities, so essentially the effort needed to get to space isn't really more after all. Rather it's simply necessary to build bigger rockets in this patch than the last one.

And I may have discovered the secret to making stable rockets in 1.0.2: it's very, very sensitive about where the center of mass is vs. the center of drag / aerodynamic pressure. You can't see where that is in the editor, but it seems to generally be roughly in the middle of the ship. Thus rockets that drain their fuel from the top are doomed unless they have some pretty powerful fins. But rockets that drain their fuel from the bottom have a much easier time of things, or in other words, you have to block fuel flow to the engines and then run a fuel duct to them from the top of the ship. It's hard to make ships this way that aren't ugly, but once I started getting the hang of it my launches immediately became significantly easier. So, uh, "pro tip" I guess.

Oh, and other pro tip, regarding landing big stuff in water (I may make a proper tutorial later): asymmetrical parachutes. Have most of the parachutes clustered around the empty center of mass (in rockets this is often surprisingly close to the bottom), and one or more pairs of parachutes at the far end, with the pairs on opposite sides of the rocket so it stays aerodynamically balanced, but NOT bound by symmetry. On one side, set them to open with the rest of the parachutes, e.g. at 500m, and on the other, have them wait until 50m. When the parachutes go off, the "center of parachutes" will be off to one side of the rocket due to one side opening first, and the rocket will tip sideways. When the second group opens, the rocket will start to straighten out, but there won't (hopefully) be enough time for it to right itself all the way before landing. Thus on landing it will find itself tilted, and thanks to the improvements in 1.0 parachute behavior, the parachutes near the top of the rocket will stay open and guide it to a gentle repose.

Edited May 2, 2015 by parameciumkid

[Rithsom]

I didn't see anything wrong with the 1.0 aerodynamics. All we needed in the patch was a fix for the drills and LV-Ns overheating, and other minor things like making the heat shields not massless.

We've all experienced enough of the souposphere, why should it return?

[Superfluous J]

Thus rockets that drain their fuel from the top are doomed unless they have some pretty powerful fins. But rockets that drain their fuel from the bottom have a much easier time of things, or in other words, you have to block fuel flow to the engines and then run a fuel duct to them from the top of the ship. It's hard to make ships this way that aren't ugly, but once I started getting the hang of it my launches immediately became significantly easier. So, uh, "pro tip" I guess.

Interesting, and it makes logical sense. Do rockets do this in real life, does anybody know?

[sp1989]

My issue with the aero is that if I put a plane into a full throttle dive the dang thing slows down instead of accelerating! So that really unrealistic.

[RocketPilot573]

I don't care that my rockets need slight more delta-v than in 1.0.0, I just care that in 1.0.0 it felt and handled like a rocket, in 1.0.1 it's a brick. No more fun.... Seriously, no one complained about the 1.0.0 aero, maybe the turbojets give a little too much thrust, but increasing the entire drag is really hard on rockets.

What is the point of fairings if the atmosphere is so soupy that having fairings doesn't make you go any faster?

[Frozen_Heart]

The main difference is the Isp drop from 0.90 to 1.01.

The was a dV drop from 0.9 to 1.0, but the engines were nerfed to deal with it. Things stayed pretty balanced.

From 1.0 to 1.01/.02 the atmo was thickened but the engines stayed lower thrust and dV. I'm finding it harder to get into orbit than 0.9 or 1.0.

[klgraham1013]

I don't care that my rockets need slight more delta-v than in 1.0.0, I just care that in 1.0.0 it felt and handled like a rocket, in 1.0.1 it's a brick. No more fun.... Seriously, no one complained about the 1.0.0 aero, maybe the turbojets give a little too much thrust, but increasing the entire drag is really hard on rockets.

What is the point of fairings if the atmosphere is so soupy that having fairings doesn't make you go any faster?

Some people complained. I do remember that. I really liked the 1.0.0 aero.

[JWag]

Interesting, and it makes logical sense. Do rockets do this in real life, does anybody know?

I'm no rocket historian, but I've never heard of a rocket that drains from the top first. You'd need to segment your tanks and put pumps in each one and the weight penalties would be enormous.

In real rockets, fuel and oxidizer are in separate tanks, which each drain from their bottoms. You get a nice advantage with this; you can pick which propellant tank is close to the engines, and which is closer to the nose. The Space Shuttle put the LOX tank at the top of the external tank and the LH2 tank at the bottom. LOX is many times more massive than LH2 at the same volume, so the ET's CG was always nearer the nose.

In KSP your tank just drains top to bottom as if the fuel and oxidizer were pre-mixed.

[magnemoe]

One effect is that heat is less of an issue.

This pod is kind of my Orion, it can reach Mun orbit or land on Minmus, now to protect the landing legs I put the strut system below for thermal protection.

in 1.0 the bottom part always exploded before I could pop parashute at 25-20 km. Now it survives all the way down, yes it would probably blow up if I set Pe lower than 20 km.

[RocketPilot573]

Some people complained. I do remember that. I really liked the 1.0.0 aero.

I suppose that is true, but the vast majority loved the 1.0.0 aero. Now most people seem to hate it after the patch.

EDIT: Also I have seem several reports that flying straight up to 10 km is literally most efficient again. Squad has said that they prefer fun over realism, but when both realism and fun are going down the drain then something is wrong.

Edited May 2, 2015 by RocketPilot573

[georgTF]

On the subject of fun v realism, it may have been a tad unrealistic to rapidly fly a 2-seater plane with a single Basic Jet over to the desert west of KSC, but in 1.0.2 the thing is so slow i got bored just flying over the mountain range. My early career science is invested in aeroplanes and i'm pretty put off by the patch. Honestly "all that was needed" was to make parachutes vulnerable, and i suppose the nukes and whatnot else was fixed. So some players went nuts going way too fast and breaking certain mechanics, but so what? It's their game too. No reason to make it tedious for me. I'll wait for the wahmbulance now...

[SelectHalfling0]

I liked 1.0's original aero system. I tried flying a plane that I developed (looks like real one, flew like a real one). It can't get off the runway. Planes once again need the needlessly massive wings again to get any lift.

Also, the gravity turn at a high altitude was brought back, and I was enjoying starting the turn just after launch.

Can someone tell me where to start the turn now? I'm soooo confused.

- - - Updated - - -

I don't know how helpful this is, but here's the old physics.cfg from 1.0:

dragMultiplier = 6.0
dragCubeMultiplier = 0.06
angularDragMultiplier = 2
liftMultiplier = 0.038
liftDragMultiplier = 0.03
bodyLiftMultiplier = 8
aeroFXStartThermalFX = 2
aeroFXFullThermalFX = 3.5
aeroFXExponent = 3
thermalMaxIntegrationWarp = 100
spaceTemperature = 4
solarLuminosityAtHome = 1360
solarInsolationAtHome = 0.15
convectionDensityExponent = 1
convectionVelocityExponent = 3
fullConvectionAreaMin = 0.2
fullToCrossSectionLerpStart = 0.8
fullToCrossSectionLerpEnd = 1.5
machConvectionStart = 2
machConvectionEnd = 3.5
partEmissivityExponent = 4
radiationFactor = 1
convectionFactor = 40
newtonianConvectionFactorBase = 1
newtonianConvectionFactorTotal = 1
conductionFactor = 10
internalHeatProductionFactor = 0.03
aerodynamicHeatProductionFactor = 1
standardSpecificHeatCapacity = 800
DRAG_TIP
{
	key = 0 1 0 0
	key = 25 1 0 0
}
DRAG_SURFACE
{
	key = 0 0.02 0 0
	key = 0.85 0.02 0 0
	key = 0.9 0.0152439 -0.07942077 -0.07942077
	key = 1.1 0.0025 -0.005279571 -0.001936768
	key = 2 0.002083333 -2.314833E-05 -2.314833E-05
	key = 5 0.003333333 -0.000180556 -0.000180556
	key = 25 0.001428571 -7.14286E-05 0
}
DRAG_TAIL
{
	key = 0 1 0 0
	key = 0.85 1 0 0
	key = 1.1 0.25 -0.02215106 -0.02487721
	key = 1.4 0.2287166 -0.01326022 -0.001389867
	key = 2 0.275 0.03981932 0.03981932
	key = 5 0.3333333 -0.003474526 -0.02333333
	key = 25 0.1428571 -0.004285714 0
}
DRAG_MULTIPLIER
{
	key = 0 0.5 0 0
	key = 0.85 0.5 0 0
	key = 1.1 2 0 -0.6
	key = 2 1.2 -0.5444444 -0.5444444
	key = 5 0.6 0 0
	key = 10 0.8 0.06700063 0.06700063
	key = 14 0.93 0.006815632 0.006815632
	key = 25 1 0 0
}
LIFTING_SURFACE_CURVES
{
	LIFTING_SURFACE
	{
		name = Default
		lift
		{
			key = 0 0 0 1.965926
			key = 0.258819 0.5114774 1.990092 1.905806
			key = 0.5 0.9026583 0.7074468 -0.7074468
			key = 0.7071068 0.5926583 -2.087948 -1.990095
			key = 1 0 -2.014386 -2.014386
		}
		liftMach
		{
			key = 0 1 0 0
			key = 0.3 0.5 -1.671345 -0.8273422
			key = 1 0.125 -0.0005291355 -0.02625772
			key = 5 0.0625 0 0
			key = 25 0.05 0 0
		}
		drag
		{
			key = 0 0.01 0 0
			key = 0.3420201 0.1 0.1750731 0.1750731
			key = 0.5 0.4 4.557837 4.557837
			key = 0.7071068 2.828427 4 4
			key = 1 4 4 0
		}
		dragMach
		{
			key = 0 0.25 0 -0.8463008
			key = 0.15 0.125 0 0
			key = 0.9 0.375 0.7227947 0.7227947
			key = 1.1 1 0 0
			key = 1.4 0.65 -1.29191 -1.29191
			key = 1.6 0.5 -0.4376471 -0.4376471
			key = 2 0.42 -0.1475873 -0.1475873
			key = 5 0.275 0 0
			key = 25 0.4 0.0006807274 0
		}
	}
	LIFTING_SURFACE
	{
		name = BodyLift
		lift
		{
			key = 0 0 0 1.975376
			key = 0.309017 0.5877852 1.565065 1.565065
			key = 0.5877852 0.9510565 0.735902 0.735902
			key = 0.7071068 1 0 0
			key = 0.8910065 0.809017 -2.70827 -2.70827
			key = 1 0 -11.06124 0
		}
		liftMach
		{
			key = 0.3 0.167 0 0
			key = 0.8 0.167 0 -0.3904104
			key = 1 0.125 -0.0005291355 -0.02625772
			key = 5 0.0625 0 0
			key = 25 0.05 0 0
		}
		drag
		{
			key = 0 0 0 0
		}
		dragMach
		{
			key = 0 0 0 0
		}
	}
	LIFTING_SURFACE
	{
		name = SpeedBrake
		lift
		{
			key = 0 0 0 0
		}
		liftMach
		{
			key = 0 0 0 0
		}
		drag
		{
			key = 0 0.01 0 0
			key = 0.3420201 0.1 0.1750731 0.1750731
			key = 0.5 0.4 4.557837 4.557837
			key = 0.7071068 2.828427 4 4
			key = 1 4 4 0
		}
		dragMach
		{
			key = 0 0.25 0 -0.8463008
			key = 0.15 0.125 0 0
			key = 0.9 0.375 0.7227947 0.7227947
			key = 1.1 1 0 0
			key = 1.4 0.65 -1.29191 -1.29191
			key = 1.6 0.5 -0.4376471 -0.4376471
			key = 2 0.42 -0.1475873 -0.1475873
			key = 5 0.275 0 0
			key = 25 0.4 0.0006807274 0
		}
	}
}

I need this! Does this bring back the heat shield bug or not?

Edited May 2, 2015 by SelectHalfling0

[maccollo]

EDIT: Also I have seem several reports that flying straight up to 10 km is literally most efficient again.

Nope, it's the complete other way around. If you just go straight up to 10 km there's no way you can perform an efficient gravity turn.

1.01/1.02 drag is more unforgiving if you want to perform an efficient gravity turn. While it only takes a couple 100 m/s more to get to space (3200 is my lowest so far), you have to be much more accurate so that the trajectory begins to flatten out just as the atmospheric density begins to drop really fast. If you pitch is a bit to weak you end up flattening out much higher than nessesary and you loose a lot more to gravity losses. Pitch a bit to hard and the trajectory will flatten while you are still flying through the dense atmosphere, and even if you can save it you are punished severely for it in atmospheric drag losses.

I'm no rocket historian, but I've never heard of a rocket that drains from the top first. You'd need to segment your tanks and put pumps in each one and the weight penalties would be enormous.

In real rockets, fuel and oxidizer are in separate tanks, which each drain from their bottoms. You get a nice advantage with this; you can pick which propellant tank is close to the engines, and which is closer to the nose. The Space Shuttle put the LOX tank at the top of the external tank and the LH2 tank at the bottom. LOX is many times more massive than LH2 at the same volume, so the ET's CG was always nearer the nose.

In KSP your tank just drains top to bottom as if the fuel and oxidizer were pre-mixed.

*Ninjad*

When made the core for my SLS mod I split the hydrogen tank from the oxygen tank to make sure this effect was present.... and it's really strong. Try as I might, I could not flip the core stage, even with the 30 degree gimbal range of the RS-25 engines. Anyway, real life rockets never deviate far from the prograde vector during atmospheric flight, because they do not have the structural integrity to survive the aerodynamic stresses this create.

[cantab]

When made the core for my SLS mod I split the hydrogen tank from the oxygen tank to make sure this effect was present.... and it's really strong. Try as I might, I could not flip the core stage, even with the 30 degree gimbal range of the RS-25 engines. Anyway, real life rockets never deviate far from the prograde vector during atmospheric flight, because they do not have the structural integrity to survive the aerodynamic stresses this create.

I was wondering how much impact that would make, using separate fuel and oxidizer tanks. Would the fuel:oxidizer mass ratio also affect it? KSP uses a nearly 1:1 fuel-ox mass ratio, whereas I believe in real rockets the oxidizer is a few times more massive than the fuel.

[Superfluous J]

I just tried a gravity turn that was pretty efficient both in FAR and in 1.0, in 1.0.2 and then launched the same rocket to 10km, cranked it hard 45, and then followed my old 0.90 gravity turn.

The old 0.90 gravity turn saved me about 50m/s

NOTE: This was with as aerodynamic a rocket as you can make in the early tiers. Just a straight stack, 2 stages and a mk1 pod with a parachute on top. I added the new fins too to help with stability.

Edited May 2, 2015 by 5thHorseman