Jump to content

Issues with twin booster rockets


Recommended Posts

So I've been trying to get the three engine rocket setup to fly but I'm having problems. The design I use is like the atlas v with two side boosters and a core center engine. The problem I'm having is no matter what I do whenever I launch the rocket flips over when I try and nose down for horizauntal flight. I've seen the real atlas and it doesn't use any wings or stabilizing fins on its center stack. What am I Missing? Do I need to put rcs on the top to keep the vehicle from flipping? 

Link to comment
Share on other sites

Got a pic? 

This has happened to a few people,  as good as the new aerodynamics are I'm not sure the design will work in ksp, there are also some very subtle yet effective design choices on the atlas that make it so strangely stable.

Have your engines got decent gimbal?   

Are you keeping your turn smooth using fine controls?   

Are you keeping the speed up enough /too much?

Is your centre of mass moving undesirably due to fuel use? 

Edited by Palaceviking
Grammar police.
Link to comment
Share on other sites

7 minutes ago, Palaceviking said:

Got a pic? 

This has happened to a few people,  as good as the new aerodynamics are I'm not sure the design will work in ksp, there are also some very subtle yet effective design choices on the atlas that make it so strangely stable.

Have your engines got decent gimbal?   

Are you keeping your turn smooth using fine controls?   

Are you keeping the speed up enough /too much?

Is your centre of mass moving undesirably due to fuel use? 

Yes I have engine gimbal set and try to be as smooth as I can on my turns. The design is just three of the white and black 2.5 tanks together. First stage is all three engines with fuel going from the outer tanks into the center one. Then once two side boosters cut off they are released and center core continues on. I can never get to the stage to eject the two side boosters as every time I nose down the whole thing flips. Won't do it if I've got fins but fins look dumb. 

Link to comment
Share on other sites

Does your engine have gimbal? Aheum, I mean what engine are you using?

Some stock engines like the LV-T30 Reliant for instance have no gimbal. A rocket rotates with either engine gimbal or control fins at the bottom. However, gimbal is the only thing necessary and fins create extra aero drag. But you can use the latter if the former option fails. 
If you have a engine with no gimbal you only have the reaction wheel torque of your pod and/or additional reaction wheels and they can't fight a multi ton rocket on a gravity turn at any xx speed generally, no mind xxx speeds.

Again, got a pic?

Link to comment
Share on other sites

3 hours ago, Dimetime said:

the rocket flips over when I try and nose down for horizauntal flight

Sounds like it is aerodynamically unstable.   RCS, gimbal brute force are not the answer, you need to make it stable.

Is there a lot of drag at the front end of the rocket?  THat could be upsetting things.    If not you'll just have to swallow your pride and put some aero surfaces on the bottom of the rocket.  Passive, non controlling ones are fine since you're probably using an engine with gimbal.     Strakes are nice and subtle and don't stick out much (compared with active fins/canards) but add a lot of lift to the bottom end, making the rocket much more stable.

Aerodynamic parts without a control surface, like strakes , have twice as much lift as an all moving tailplane/canard/elevon of the same area, so they're more discreet.   

Link to comment
Share on other sites

6 hours ago, Dimetime said:

 fins look dumb. 

maybe, but not dumber than a rocket pointing backwards.

You need to either move your aerodinamic center down or you center of mass up. Question is;. do you have something heavy enough on the nose to ballance with 3 engines on the bottom? If not you may try fins or remove nosecones on side-booster (if those are not to high).

Your issue in picture:

Spoiler

UYmzteT.png

 

Link to comment
Share on other sites

4 hours ago, AeroGav said:

Aerodynamic parts without a control surface, like strakes , have twice as much lift as an all moving tailplane/canard/elevon of the same area, so they're more discreet.

also cheaper (granted usually not a big deal)

Link to comment
Share on other sites

12 hours ago, Spricigo said:

also cheaper (granted usually not a big deal)

The problem with fixed tail feathers, however, is that while they contribute to stability, they actually fight your ability to turn.  Thus, you end up having to waste more of your dV to turn than is necessary with movable fins, due to excess gimbal use and/or constantly running RCS or Vernors.  Movable fins are considerably more efficient because they provide both stability and directional control, so you need less gimbal and RCS/Vernor, so that dV can go into lift instead of turning.

@Dimetime, the way to minimize the need for, or at least reduce the number and size of, tail feathers is to make the rocket long and skinny.  The longer the rocket, the more lever arm the tail feathers (and gimbal, and any Vernors at the bottom end) have, so the more effective they are.  A lifter stack of 2 or 3 orange tanks needs only a few small elevons/canards (although a few Vernors also help).  The short, wide rockets, especially with lots of asparagus staging, were a Darwinian solution to the state of KSP some years ago.  Back then, aero wasn't a thing so being short and wide didn't matter, and joints were WAY too wobbly to make long, skinny rockets anyway.  But things have changed, so now long and skinny is the way to go.

Edited by Geschosskopf
Link to comment
Share on other sites

It's possible to do a gravity turn with an aerodynamically unstable rocket (no fins) but you have to do a pitch over of a few degrees shortly after launch, around 60 m/s, then set it to a prograde hold for the rest of the launch. That's more or less how real rockets do their gravity turn.

What I do:

  • Have TWR of 1.3 to 1.6 at launch
  • Go straight up until I'm going 60m/s
  • Pitch over 5 to 10 degrees and hold that orientation until the prograde marker matches it, then set it to a prograde hold. (You will slowly pitch further as you follow your trajectory)
  • Ideally, at 10 km I should be pointing at about 45 degrees pitch, and about 20 degrees at 30 km.
  • If Im pitching too fast or too slow, I can make small attitude adjustments ( not more than 5 degrees at a time, very easy to flip), OR I can adjust throttle (pitch rate is proportional to  current speed, so not accelerating as fast will make you pitch faster).

A "perfect" gravity turn has you reach horizontal exactly as you reach orbital velocity at your target altitude, without ever taking it off a prograde heading after the initial pitch maneuver, but in practice you'll usually still have a separate circularization burn, but using this method it shouldn't be more than about 200 m/s.

To summarize, a "proper" gravity turn has 2 advantages:

  1. It's efficient
  2. Once you have a probe core/pilot with prograde hold, your rockets never need fins again.
Edited by BadOaks
Link to comment
Share on other sites

3 hours ago, BadOaks said:

aerodynamically unstable rocket (no fins)

those are not synoms :wink: 

 A barely stable one can go to orbit following a very narrow gravity turn. An almost stable can hold trajectory with a combination of SAS and reaction wheel/RCS

Link to comment
Share on other sites

10 hours ago, Geschosskopf said:

@Dimetime, the way to minimize the need for, or at least reduce the number and size of, tail feathers is to make the rocket long and skinny.  The longer the rocket, the more lever arm the tail feathers (and gimbal, and any Vernors at the bottom end) have, so the more effective they are.  A lifter stack of 2 or 3 orange tanks needs only a few small elevons/canards (although a few Vernors also help).  The short, wide rockets, especially with lots of asparagus staging, were a Darwinian solution to the state of KSP some years ago.  Back then, aero wasn't a thing so being short and wide didn't matter, and joints were WAY too wobbly to make long, skinny rockets anyway.  But things have changed, so now long and skinny is the way to go.

Really? I thought you'd want everything as short as possible to keep the weight evenly spaced. Like the longer the center piece the more likely the top would be to pull it over. All my designs are short and as even as I can make them.

Edited by Dimetime
Link to comment
Share on other sites

18 minutes ago, Dimetime said:

Really? I thought you'd want everything as short as possible to keep the weight evenly spaced. Like the longer the center piece the more likely the top would be to pull it over. All my designs are short and as even as I can make them.

Look at real rockets---they're all long and skinny :wink:   They work like arrows.  The main mass (the arrowhead) at the front keeps the CoM towards that end, and the long shaft behind either itself or with tail feathers provides the stability.

Short and wide is only desirable in 2 situations:

  1. Landers for hilly terrain
  2. Something you have to dock repeatedly in a variety of fuel conditions.  With this, you've used RCS Build Aid so that no matter how much fuel it has, the CoM won't move, so the RCS is always balanced to make docking easy.  As it turns out, short and wide is the easiest way to accomplish this.
Link to comment
Share on other sites

11 hours ago, Geschosskopf said:

The problem with fixed tail feathers, however, is that while they contribute to stability, they actually fight your ability to turn.  Thus, you end up having to waste more of your dV to turn than is necessary with movable fins, due to excess gimbal use and/or constantly running RCS or Vernors.  Movable fins are considerably more efficient because they provide both stability and directional control, so you need less gimbal and RCS/Vernor, so that dV can go into lift instead of turning.

 

If there is a waste of deltaV is because piloting errors requiring correction and leading to cosine loses. A well executed gravity turn consist of a small nudge/tilt to start it and then keep the rocket pointed prograde with his own stability.  It may the case that with lower control authority the gravity turn need to be initiated a bit early but done correctly it have no perceptive effect on deltaV expenditure. Granted that the piloting errors are expected to happens, it may be the case for a inexperienced player or someone trying a new rocket.

Also there is the question of how much of fins are necessary, with a few basic fin will solve the stability issues of a big rocket the major drawback is really how silly it will looks :D

1 hour ago, Dimetime said:

Really? I thought you'd want everything as short as possible to keep the weight evenly spaced.

As a way to reduce drag loses yes, build the rocket with the least cross-sectional area possible and keep it pointed as close to prograde as possible.

Off course you design need to consider other things too. For example: long rockets are more prone to wobbling.

 

No one mentioned it yet: try disabling gimbals on the side-booster.  It may be the case of control-coupling in those engines throwing you off course.

Link to comment
Share on other sites

2 hours ago, Spricigo said:

If there is a waste of deltaV is because piloting errors requiring correction and leading to cosine loses. A well executed gravity turn consist of a small nudge/tilt to start it and then keep the rocket pointed prograde with his own stability. 

You crack me up, you really do :) 

The reason why it's called a "gravity turn" is because the rocket is NOT directionally stable.  This lack of directional stability means gravity is causing the rocket is falling over like a tree felled by a lumberjack.  If the rocket was directionally stable, it would instead keep going in a straight line.  Because of this, a true "gravity turn" requires no control input at all after the 1st little nudge to start the rocket tipping over slowly.  The rocket is just tuned in the design phase (TWRCoM position/movement, and aerodynamically) to reach the desired altitude and speed by the time it's rotated 90^.  The long, skinny shape of the rocket, plus any tail feathers, acts as a brake on the rotational speed, so the rocket takes longer to reach the horizontal position when moving forward than it would if it just fell over on the pad.  However, the aerodynamics do not, and cannot, stop the rotation caused by gravity.  Ergo, the rocket is NOT directionally stable.

Any ascent that requires control inputs throughout the ascent to make the rocket curve over is NOT, strictly speaking a "gravity turn", even if the trajectory looks exactly the same.  In this case, the rocket IS directionally stable, going in a straight line despite gravity trying to tip it over.  This is especially true if you have SAS on in default mode without picking the prograde lock.  Maintaining directional stability is its whole job in this mode.  The rocket thus periodically needs control inputs that upset its stability briefly in the desired direction, to make it curve on over.  These control inputs use various systems to have their effect on the rocket's orientation.  Torque, control surface deflection, or directional thrust (gimbal, RCS, or Vernor).  In this case, fixed tail fins again act as a brake on rotation, so you need more force to make the same turn than you would with movable fins.  That extra force costs dV, either from gimbal cosine losses, fuel burned by Vernors, or lugging an extra launch-only RCS system.

This has nothing at all do to with "piloting errors".  It's a function of the rocket design.  If you make a stable rocket, you will have to divert force to make it turn, and you are not doing a gravity turn.  If you make an unstable rocket (only with carefully limited instability), then you don't need control inputs and ARE doing a gravity turn.

Link to comment
Share on other sites

49 minutes ago, Geschosskopf said:

This has nothing at all do to with "piloting errors".  It's a function of the rocket design.  If you make a stable rocket, you will have to divert force to make it turn, and you are not doing a gravity turn.  If you make an unstable rocket (only with carefully limited instability), then you don't need control inputs and ARE doing a gravity turn.

Actually, a stable rocket is EXACTLY what you want for a good gravity turn ... the only time you're not staying on prograde is the pitchover maneuver, the initial turn to set up the rest of the gravity turn.  Once that's complete, you just hold prograde and USE the stability of your rocket to your advantage.

Link to comment
Share on other sites

23 minutes ago, Kryxal said:

Actually, a stable rocket is EXACTLY what you want for a good gravity turn ... the only time you're not staying on prograde is the pitchover maneuver, the initial turn to set up the rest of the gravity turn.  Once that's complete, you just hold prograde and USE the stability of your rocket to your advantage.

Yeah,   aerodynamic stability means aero forces are trying to push the nose back to prograde.   The greater the distance of the nose from prograde , the stronger the aero force pushing it back.      A very stable rocket flies the same with SAS off, as it would with SAS set to hold prograde.

The thing with the gravity turn is that the rocket isn't making any lift ,  so the prograde vector will fall downward  once you start moving away from straight up.    Aerodynamic stability just keeps the rocket pointing at the falling prograde vector.   Depending on the rate of the initial tilt and the TWR,  the rocket may in fact reach orbital velocity before prograde ever sinks below the horizon.

Edit - I prefer active aerodynamic controls on my rockets too,  because it opens up the option of using non-gimbal engines (the Dart is super  efficient late game engine,   the Reliant is a good early game booster).   But the OP wanted something that looks like a delta heavy, so no big sticky out bits !

Edit 2  - I wonder if he's got a buggy fairing creating excess drag at the front.    You can get the game to show drag values on your parts in the right click menus by pressing ALT F12 and going to physics tab.    Nothing inside the fairing should ever show a drag number, and the fairing itself should not have disproportionate drag vs rest of rocket.

edit  3 -   Tall n skinny vs short and fat -  i am pretty sure that 2.5m parts have the best tank capacity to drag ratio of any, by some margin.    Chances are OP is already using them if he wants a delta heavy - orange tank yes?

 

Edited by AeroGav
Link to comment
Share on other sites

2 hours ago, Geschosskopf said:

You crack me up, you really do :) 

Was a teacher. It's a gift and a burden.:rolleyes:

Ok, you have a point. Two actually:

It's a turn. But when we talk about aerodynamic stability we don't use the same frame of reference where this turn happen. For our aerodynamic analyses what we consider is the direction of the airflow. And is relative that variable direction the craft is stable or unstable. Since KSP don't simulate wind or turbulence we can direction of airflow == surface retrograde, which means a aerodynamic stable(see bellow) vessel will maintain the same orientation relative the retrograde/prograde axis.

 

Torque is necessary to change direction. People say stable when the technically correct would be self-stabilizing, like AeroGav explained, the torque comes from the aero forces pushing the vessel. If the vessel is oriented exactly along prograde/retrograde axis all aero forces are in equilibrium and the craft may remain stable, if he deviates from that axis the aero forces will push it back.

 

2 hours ago, Geschosskopf said:

it's called a "gravity turn"

Think about it:  Torque is required to turn. The force of gravity acts in the center of mass of the object, so produce no torque relative to the center of mass(level arm = 0).  Is the name incorrect? Why ?

 

 

 

 

Link to comment
Share on other sites

1 hour ago, Kryxal said:

Actually, a stable rocket is EXACTLY what you want for a good gravity turn ... the only time you're not staying on prograde is the pitchover maneuver, the initial turn to set up the rest of the gravity turn.  Once that's complete, you just hold prograde and USE the stability of your rocket to your advantage.

No, you want at best neutral stability to do a gravity turn.  Stable means that when an object is perturbed by an external force, it returns to its original orientation.  Neutral stability means it stays where the force left it.  Instability means it takes the external force and runs with it (flipping, in the case of a rocket).

If you turn SAS on but don't pick a direction, then you have stability (assuming you've provided the control authority SAS needs).  If you aim the rocket 5^ east just off the pad in this condition, it will remain pointing 5 degrees east even as the prograde marker drifts slowly west as Kerbin rotates under the rocket.  You also have stability if moving the prograde marker east requires control input.  IOW, you're turning the nose farther east than the prograde marker, the letting the rocket sit there stably, and eventually the off-direction thrust moves the prograde marker to match the nose by vector addition.

OTOH, if the prograde marker moves east by itself without control input, then the rocket is NOT stable.  It's being continually perturbed by gravity and making no effort to counter this.  This is at best neutral stability.  It could actually be limited instability that doesn't become excessive before altitude removes the aerodynamic forces.

Link to comment
Share on other sites

11 minutes ago, Spricigo said:

It's a turn. But when we talk about aerodynamic stability we don't use the same frame of reference where this turn happen. For our aerodynamic analyses what we consider is the direction of the airflow. And is relative that variable direction the craft is stable or unstable. Since KSP don't simulate wind or turbulence we can direction of airflow == surface retrograde, which means a aerodynamic stable(see bellow) vessel will maintain the same orientation relative the retrograde/prograde axis.

Not really the correct line of reasoning.  It's the other way around.  The prograde marker is an effect, not a cause.  If the prograde marker is moving, it's because the ship's direction of motion is changing.  IOW, to move the prograde marker, you have to move the rocket's nose OFF the prograde marker and hold it there until vector addition eventually causes the prograde marker to catch up with the nose marker.

So, with a true gravity turn, you're not giving any control inputs and, as you say, the aerodynamics are trying to keep the rocket weathervaned in the direction it's already going.  So why is the prograde marker moving towards the eastern horizon despite its natural tendency to drift slightly west (due to Kerbin's rotation) as the rocket gains altitude?  Because gravity is pulling the rocket over like a falling tree.  As the rocket falls more and more towards the horizontal, its thrust vector also points more and more horizontally.  This causes vector addition to drag the prograde marker in that direction.

If the prograde marker wasn't already moving "under its own power" (as in, being moved by something external to the rocket, like gravity), then SAS prograde lock would NOT make the rocket turn.  The prograde marker would stay in the same place (apart from westward drift due to Kerbin's rotation under the rocket while in Surface mode) and the rocket would keep going in that same direction relative to the ground under it as before.

 

11 minutes ago, Spricigo said:

Torque is necessary to change direction. People say stable when the technically correct would be self-stabilizing, like AeroGav explained, the torque comes from the aero forces pushing the vessel. If the vessel is oriented exactly along prograde/retrograde axis all aero forces are in equilibrium and the craft may remain stable, if he deviates from that axis the aero forces will push it back.

In a gravity turn, the torque comes from gravity.  The aerodynamic forces resist this, trying to keep the rocket going the same direction as before by weathervaning it, but they're not strong enough to stop it.  Thus, they only slow the rotation being caused by gravity, so that the rocket doesn't fall parallel with the ground below until it's at the desired orbital altitude and velocity.

It goes like this....  When the rocket spawns, it's (usually) pointing straight up on the pad.  When you fire the 1st stage, it will thus go straight up, and then the prograde marker appears right where the nose is pointing, straight up.  At the appropriate time, you give your 1 and only control input to kick the nose very slightly east off the prograde marker.  Because the rocket isn't going very fast yet, the prograde marker almost instantly latches back onto the nose marker.  At this point, you set SAS prograde lock.

So why does the prograde marker fall all the way to the eastern horizon?  It's because the rocket is now at an angle to the ground below.  Thus, it's CoM is off to the side as well, so the direction of the pull of gravity no longer coincides with the rocket's axis.  And neither do the aerodynamic forces, which are acting through the CoP some distance behind the the CoM.  So, you have an upwards force acting at the CoP and and dowmwards force acting on the CoM some distance away, resulting in a turning moment (aka torque).  This is what makes the rocket turn without further control input.  But you have to design and tune the rocket so that this resultant moment is small, to keep the rocket's rotational velocity at just the right magnitude.  Too much and the rocket flips.  Too little and it doesn't level off at the right place at Ap.

Link to comment
Share on other sites

14 hours ago, Geschosskopf said:

Not really the correct line of reasoning.  It's the other way around.  The prograde marker is an effect, not a cause.

I didn't said it is cause of nothing.  I said this is the reference used to say the craft is stables. 

Its like if I said "I took my place in the bus and stayed there the whole trip,  enjoy the view" and you argue that I cannot enjoy the view because of how the speedometer works.

There's no "reasoning" to be contested when all I'm saying is from where the observations are made.

14 hours ago, Geschosskopf said:

  IOW, to move the prograde marker, you have to move the rocket's nose OFF the prograde marker and hold it there until vector addition eventually causes the prograde marker to catch up with the nose marker

Prograde mark is the represent the directhion of movement.  It will change accordingly with the forces acting upon the vessel. 

In any case what I said is that direction of movement (prograde) is  related to  the frame of reference used to evaluation of aerodynamic stability (airflow). If one change, the other change but it don't matter what (if any) is cause of the change.

14 hours ago, Geschosskopf said:

So, with a true gravity turn, you're not giving any control inputs and, as you say, the aerodynamics are trying to keep the rocket weathervaned in the direction it's already going. 

The term to "direction it's already going" is prograde. frame of reference tied to this direction change as the direction change.

14 hours ago, Geschosskopf said:

So why is the prograde marker moving towards the eastern horizon despite its natural tendency to drift slightly west (due to Kerbin's rotation) as the rocket gains altitude

As explained above, it doesn't matter. In any case because of forces acting upon the vessel. 

And I made that question first:

15 hours ago, Spricigo said:

Think about it:  Torque is required to turn. The force of gravity acts in the center of mass of the object, so produce no torque relative to the center of mass(level arm = 0).  Is the name incorrect? Why ?

 

14 hours ago, Geschosskopf said:

Because gravity is pulling the rocket over like a falling tree.

Gravity alone don't explain it. Like the hinges alone don't explain the movement of a door.

14 hours ago, Geschosskopf said:

As the rocket falls more and more towards the horizontal, its thrust vector also points more and more horizontally.  This causes vector addition to drag the prograde marker in that direction.

Vectors are representations of physical entities(e.g. forces). Use the name if the force that cause the change of movement instead of a cryptic expression. 

14 hours ago, Geschosskopf said:

If the prograde marker wasn't already moving "under its own power" (as in, being moved by something external to the rocket, like gravity), then SAS prograde lock would NOT make the rocket turn.  The prograde marker would stay in the same place (apart from westward drift due to Kerbin's rotation under the rocket while in Surface mode) and the rocket would keep going in that same direction relative to the ground under it as before

There are forces acting upon the craft. The resulting torque of those forces (plus the magic torque of reaction wheels) need to be null for the craft maintain a fixed orientation. 

Coincidentally SAS lock to prograde is how I usually make a not turning vessels to start to turn. Much more precise that way.

14 hours ago, Geschosskopf said:

In a gravity turn, the torque comes from gravity

The mathematical definition of torque is: [Torque]=[Force]  X [lever arm]    (those are vectors, X means cross product)

If you say torques comes from Force of Gravity (aka weight) you need to tell us around which axis. That's assumed, when not stated, to be an axis passing by the CoM. But in that case there is no torque because a force applied on the rotation axis produce no torque. (Test it: push the hinges of a door and see if it will open/close)

Notice that it may be valid to consider the rocket turning around any axis. The values will be different but the maths is the same.

14 hours ago, Geschosskopf said:

aerodynamic forces resist this, trying to keep the rocket going the same direction as before by weathervaning it, but they're not strong enough to stop it.  Thus, they only slow the rotation being caused by gravity

A resulting torque in the direction of the rotation cause an acceleration of that rotation. To slowdown the resuting torque needs to have opposite direction.  That is (not a suggestion,  not a good idea) the law of inertia applied to rotation.

14 hours ago, Geschosskopf said:

It goes like this....  When the rocket spawns,

In my game, it don't go like nothing you described.

That should be some indicative since all my rockets but one* (include several available on KerbalX) are designed to do the gravity turn with the piloting limited to 1. Staging 2. Change SAS mode. (No human comnand for pitch/yaw/roll.)

My designs just don't rely on 'fancy stuff' to control heading heading. The aeroforces are enough. 

*made for shuttle challenge,  only 3 flights for testing and demonstration. It just become obsole before finished development. 

However, I'll keep the detailed description and explanation to 'my own' thread. I will just let there this amazing flight of @ManEatingApe :

From launchpad to orbit,  without a control point!!!

 

Link to comment
Share on other sites

This thread is quite old. Please consider starting a new thread rather than reviving this one.

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...