Please fix the mass distribution on fuel tanks!

[Stoney3K]

Since the new 1.0 aerodynamics, we've had to cope with flippity-floppity rockets because the rocket's mass moved backwards a lot due to fuel burn, eventually turning into an unstable craft.

In real rockets, they address this problem by having separate LF and O tanks which are strategically placed to minimize the amount of mass shift due to fuel burn.

I've tried it in KSP: The same craft, with separate LiquidFuel and Oxidizer tanks will follow a perfect gravity turn without any control input, exactly as it should be. However, with a single-tank version holding the same amount of fuel (and thus LF+O combined) the configuration was hardly stable.

That was also visible in the VAB where the center of mass was a lot further back in the single tank version as opposed to the separate tank version. Is there any way that we can fix this on the stock fuel tanks?

[passinglurker]

Since people want LF only tanks anyway for jets and nukes it could make sense to instead of having rocket tanks, and jet fuselages we just have lf tanks and ox tanks they'd have to be a bit more sturdy than they were before in order to be suitable for plane use but the result would be more intuitive.

Or they can give us tweakable fuel tanks as long as pure lf and pure ox tweaked tanks still keep thier 9/11 fuel ratios, or if the 9/11 ratio was abolished for gameplay reasons and it became a straight 50/50.

either way if we get pure oxidizer tanks we would be a step closer to hybrid rocket motors.

The mass of the engines are another thing to look at perhaps they are too heavy? at least the early game ones anyway

[pandaman]

Interesting idea. Would mean more parts for simple rockets as in 2 fuel tanks instead of just 1, but for larger ships that need multiple tanks anyway then it would make little difference.

A 50/50 LF to O by volume mix would mean you could use one identical tank of each and have the right ratio. Different masses for LF and O would not affect this, but would affect CoM depending on placement.

Tweakable tank contents that add a different marking to the tank depending on contents would work well here. As well as reducing the items in the parts list.

[passinglurker]

no matter the solution starting you with the pancake tank and a massive engine in career was a stupid move on squads part. Rockets flipping when they deviate off course may be a real thing but squads broken tree ballance amplifies this several times over, and giving us tiny fins is just a weak bandaid patch.

remember during 1.0 development when harvestor said they were gonna add a fuel flow mode for jets that let them drain from all tanks in a stage evenly but respected decouplers in order to perverse COM balance while still letting you deliver fueled payloads(but then they just used the monoprop fuel flow mode that doesn't respect decouplers instead >=( )? something like that for pretty much all the engines could be another solution (the real one not the monoprop flow mode substitute)

[Jouni]

In real rockets, they address this problem by having separate LF and O tanks which are strategically placed to minimize the amount of mass shift due to fuel burn.

Real rockets have guidance systems that keep the rocket pointing to the right direction. In KSP, we have badly trained monkeys trying to fly a rocket, which often results in the rocket getting out of control.

The rocket flipping "problem" is almost entirely caused by the lack of player experience. Fins can serve as training wheels for new/inexperienced players who are still learning how to fly. Once you've learned the basics, you can make the rocket more efficient by dropping the fins and relying on active guidance. Later, as your skills improve, you may even start to think that flipping a rocket requires deliberate effort, as long as the rocket is halfway reasonable.

[passinglurker]

Real rockets have guidance systems that keep the rocket pointing to the right direction. In KSP, we have badly trained monkeys trying to fly a rocket, which often results in the rocket getting out of control.

The rocket flipping "problem" is almost entirely caused by the lack of player experience. Fins can serve as training wheels for new/inexperienced players who are still learning how to fly. Once you've learned the basics, you can make the rocket more efficient by dropping the fins and relying on active guidance. Later, as your skills improve, you may even start to think that flipping a rocket requires deliberate effort, as long as the rocket is halfway reasonable.

This isn't a case of "GET GUD OR GET REKT NOOB" having to stack small tanks draining from top to bottom in order to get any real use out of an over sized engine made of depleted uranium in the early game makes the problem worse and for many players these problems are unintuitive and invisible so they just blame the new aero instead of the unavoidable flaws in their design. sure these problems can be compensated for with good piloting but not every player is a good pilot there has to be a way to compensate with good building as well.

[sal_vager]

If you change these in ResourceGeneric.cfg

RESOURCE_DEFINITION
{
  name = LiquidFuel
  density = 0.005
  unitCost = 0.8
  hsp = 2010
  flowMode = STACK_PRIORITY_SEARCH
  transfer = PUMP
  isTweakable = true
}
RESOURCE_DEFINITION
{
  name = Oxidizer
  density = 0.005
  unitCost = 0.18
  hsp = 1551
  flowMode = STACK_PRIORITY_SEARCH
  transfer = PUMP
  isTweakable = true
}

To these.

RESOURCE_DEFINITION
{
  name = LiquidFuel
  density = 0.005
  unitCost = 0.8
  hsp = 2010
  flowMode = STAGE_PRIORITY_FLOW
  transfer = PUMP
  isTweakable = true
}
RESOURCE_DEFINITION
{
  name = Oxidizer
  density = 0.005
  unitCost = 0.18
  hsp = 1551
  flowMode = STAGE_PRIORITY_FLOW
  transfer = PUMP
  isTweakable = true
}

Your LFO will drain from each tank equally, starting on the set that will be discarded first and working its way up the chain to the root.

Edited August 30, 2015 by sal_vager

[Jouni]

This isn't a case of "GET GUD OR GET REKT NOOB" having to stack small tanks draining from top to bottom in order to get any real use out of an over sized engine made of depleted uranium in the early game makes the problem worse and for many players these problems are unintuitive and invisible so they just blame the new aero instead of the unavoidable flaws in their design. sure these problems can be compensated for with good piloting but not every player is a good pilot there has to be a way to compensate with good building as well.

My point was that this thread is based on a poor understanding of the problem. Real rockets are aerodynamically unstable and rely on active guidance. The reason why many new players have the rocket flipping problem is because their rockets don't have active guidance. The engines you start with in the career mode lack thrust vectoring, while fins with control surfaces are unlocked relatively late in the game. The Swivel is the first engine that's actually useful as a first stage engine, but the game doesn't tell the player that this is the engine they should be using.

Edit: I had to test it, because I'm not too familiar with the aerodynamics of small rockets. I built a simple two-stage rocket using a Swivel as the lower stage engine. SAS was easily able to compensate for a 10-15 degree angle of attack, while breaking the sound barrier at 10 km. The flipping problem doesn't look too bad, as long as you give proper tools to the new player.

Edited August 30, 2015 by Jouni

[passinglurker]

STAGE_PRIORITY_FLOW isn't a solution because it doesn't respect decouplers I already said this when talking about jet's flow mode

My point was that this thread is based on a poor understanding of the problem. Real rockets are aerodynamically unstable and rely on active guidance. The reason why many new players have the rocket flipping problem is because their rockets don't have active guidance. The engines you start with in the career mode lack thrust vectoring, while fins with control surfaces are unlocked relatively late in the game. The Swivel is the first engine that's actually useful as a first stage engine, but the game doesn't tell the player that this is the engine they should be using.

asking for the old aero back is lack of understanding of the problem this is the first other guy I've seen that gets that real rockets and ksp rockets are built differently and as a result of us getting more realistic aero the two don't mesh well. He understands what is involved in flying rockets and he is proposing emulating real world solutions to mitigate the problem. piloting isn't the whole problem its the parts we have to build with as well. With good parts it's easier to learn how to fly instead of giving up blaming bugs and wishing for the old aero back. Therefore parts that mitigate rocket flipping (separate lf and ox tanks, lighter engines, earlier control surfaces, changes to fuel flow etc...) help solve the pilot education problem because it's easier to trust that it's not just the games fault when the parts work with you rather than against you.

Edited August 30, 2015 by passinglurker

[Jouni]

He understands what is involved in flying rockets and he is proposing emulating real world solutions to mitigate the problem. piloting isn't the whole problem its the parts we have to build with as well. With good parts it's easier to learn how to fly instead of giving up blaming bugs and wishing for the old aero back. Therefore parts that mitigate rocket flipping (separate lf and ox tanks, lighter engines, earlier control surfaces, changes to fuel flow etc...) help solve the pilot education problem because it's easier to trust that it's not just the games fault when the parts work with you rather than against you.

I'll try again.

The center of mass issue is almost completely irrelevant, because aerodynamic stability doesn't matter. While it's technically true that rockets flip because their center of mass is in a wrong place relative to the center of pressure, changing the fuel flow rules solves only a tiny part of the problem. The real solution is starting the game with engines that have thrust vectoring enabled. As soon as you unlock the LV-T45, your rockets can fly with a ridiculously high angle of attack, and the SAS keeps them from flipping over.

[passinglurker]

I'll try again.

The center of mass issue is almost completely irrelevant, because aerodynamic stability doesn't matter. While it's technically true that rockets flip because their center of mass is in a wrong place relative to the center of pressure, changing the fuel flow rules solves only a tiny part of the problem. The real solution is starting the game with engines that have thrust vectoring enabled. As soon as you unlock the LV-T45, your rockets can fly with a ridiculously high angle of attack, and the SAS keeps them from flipping over.

EDIT: rephrase your arguments I'll rephrase mine how about you try compromise and understanding instead of throwing yourself at a brick wall

peoples piloting issues are a symptom of bad parts and mechanics to begin with all you are offering is bandaid measures without addressing the core problem. addressing all of fuel flow, engine weight, and tech tree placement of control surfaces and thrust vectoring instead of just thrust vectoring allows rockets to be as stable as they are in real life (as little as they may be according to you) instead of the badminton flying backwards held straight by unnecessary overkill of control authority like they are now.

[Stoney3K]

EDIT: rephrase your arguments I'll rephrase mine how about you try compromise and understanding instead of throwing yourself at a brick wall

peoples piloting issues are a symptom of bad parts and mechanics to begin with all you are offering is bandaid measures without addressing the core problem. addressing all of fuel flow, engine weight, and tech tree placement of control surfaces and thrust vectoring instead of just thrust vectoring allows rockets to be as stable as they are in real life (as little as they may be according to you) instead of the badminton flying backwards held straight by unnecessary overkill of control authority like they are now.

People that blame the design flaw of liquid fueled rockets in KSP to bad piloting, I challenge you this:

Build a test rocket containing nothing but two fuel tanks, one containing just LiquidFuel and the other only Oxidizer. Launch it.

Do not touch the controls, just dumb fire it and only let aerodynamics and gravity do the work.

With separate LF and O tanks, you should be able to get the rocket into a stable orbit with nothing but a slight nudge towards prograde (east), and after that, keeping hands off the controls until you reach apoapsis. Throttle should be your only control input here. The essence of a 'gravity turn' is that you let GRAVITY do the turning for you, instead of having to actively control it. This is how a real rocket will fly -- guidance is only used to keep it in the right track to its target.

The same will not work with the KSP-style "combined" fuel tanks (check it by building the same rocket, only having the bottom tank completely full and the top tank empty) because the CoM shifts too far back too quickly. The only notable exception are solid fuel rockets, which work as expected because they have only one type of propellant.

[Jouni]

peoples piloting issues are a symptom of bad parts and mechanics to begin with all you are offering is bandaid measures without addressing the core problem. addressing all of fuel flow, engine weight, and tech tree placement of control surfaces and thrust vectoring instead of just thrust vectoring allows rockets to be as stable as they are in real life (as little as they may be according to you) instead of the badminton flying backwards held straight by unnecessary overkill of control authority like they are now.

The core problem is bad advice. Bad advice from the game, from most tutorials, and from the KSP community.

Aerodynamic stability is bad advice, because it makes rocket design harder. In addition to TWR and delta-v, you also have to worry about the center of pressure and the center of mass. (The game also gives very little advice with any of these.) Many people advise beginners to use fins, which basically means adding more complexity to the rocket, having more things to remember, and having more things that can go wrong.

There is a simple trick to keep the center of mass roughly in place during the ascent. It's so simple that many (most?) beginners use it without thinking, because that's how real rockets are built. The trick is staging. If you have two vertical stages and a payload, the lower stage starts draining fuel from near the center of mass. As a result, the center of mass moves very little during the ascent, until the engines start draining the final tank and the center of mass starts moving upward. This trick works well in KSP, because the upper stage and the payload take up a large fraction of the launch mass due to high payload fractions, heavy payload parts, heavy engines, and heavy fuel tanks.

KSP rockets actually don't have that much control authority. The difference to real rockets is that KSP rockets can use the control authority to maintain stability with a high angle of attack. If you tried the same with a real rocket, aerodynamic forces would just rip the rocket apart.

- - - Updated - - -

The essence of a 'gravity turn' is that you let GRAVITY do the turning for you, instead of having to actively control it. This is how a real rocket will fly -- guidance is only used to keep it in the right track to its target.

A gravity turn means that the rocket points more or less prograde after the initial pitchover. Gravity turns the trajectory of the rocket, but it has nothing to do with the attitude of the rocket. The attitude can be adjusted both by active and passive guidance.

Real rockets typically use active guidance, because passive guidance is rather imprecise and because not having fins saves mass. Some real rockets only follow a gravity turn during the initial phase of the ascent in order to minimize aerodynamic stress. After leaving the dense lower atmosphere behind, the rocket may switch to a higher angle of attack e.g. to compensate for the low TWR of the upper stage.

[passinglurker]

Again it's more than the advice it's the parts to and it's the parts design flaws that are leading to this bad advice therefore the parts are the core of the problem. The problem is you can't build a rocket that you don't have to cover in fins and fight into orbit or a suborbital trajectory with just fl-t100's, t-30 reliant's, basic fins, srbs, and a decoupler. This is what players essentially start with in career this is their first impression unless they are a sandbox user. If they try to make a liquid fueled rocket with stages big enough to lift their engines and touch space at all it will flip with the drop of a hat. If they try to use srbs the inefficiency of the engine will force the design so large it can't be controlled with just the pod reaction wheel or if they try to use both they just get the worst of both worlds. Early career is broken because of poor part design. People's impressions of early career is where the bad advice comes from. Therefore fix the part design you fix the bad advice

[Jouni]

The problem is you can't build a rocket that you don't have to cover in fins and fight into orbit or a suborbital trajectory with just fl-t100's, t-30 reliant's, basic fins, srbs, and a decoupler.

This is a good example of an advice problem.

KSP is a sandbox game. An essential part of any sandbox game is that you're free to try whatever stupid ideas you have. Of course, the attempt may end in disaster, because the idea was stupid.

In the beginning of a career mode game, you're playing with unguided boosters. Trying to go suborbital with them is obviously a stupid idea. Experienced players see it immediately, but nobody is telling the new player that the idea is stupid.

Even worse, nobody is telling the new player that they should use the new engine they just unlocked for the more ambitious missions, even though it's heavier and less powerful than the engine they already had.

[bv1]

I like this idea. It would be a relatively easy change for Squad to make, but it would improve the flipping situation for single-stage rockets considerably.

[KrazyKrl]

Yes, I agree with OP's suggestion. LFO fuel tanks should inherently be nearly CoM neutral in respect to the tank itself. Currently the LFO tanks act as if it's one big tank, and not 2 tanks in 1 cylinder. Programming this simple thing into the game however, would probably be quite complex. (Calculating CoM for two objects within one rigidbody; which both change according to the LF and O amounts contained.)

As for AoA, if you're running more than about 10 degrees off of prograde in-atmo; your rocket should and does aerodynamically stall... just like real aircraft. Aerodynamic forces end up trying to shove your craft in a different direction, which causes failures of both control, and structure.

You can get suborbital (and even orbital if you're careful.) With no gimbal, and no fins... on SRBs. It's all about aero and your AoA. Your crafts default state should be aerodynamically stable at speed in atmo, using fins (rightfully) should be a "sledgehammer" approach. If you force your craft off of prograde too much, it should act like a brick. Even the Saturn V had tiny fins (for its size), but that's probably due to the F1 engines being so massive and just being such an inherently massive craft.

Edited September 3, 2015 by KrazyKrl