The Wheels on the Spacecraft Go Round and Round

Spinning a ten kiloton spacecraft around is no easy feat. Even more impressive if you want to be able to turn on a dime. This article covers the issues of Attitude Control of spacecrafts.

A capital ship with four main engines gimbaled off to the side to provide torque.

On land, sea, or in air, rotation is easily done by pushing off the nearest medium, with anything from a tractor tread to an impeller to a rudder. In space, no such thing is possible, which means rotation can only be accomplished in one of two ways: by expelling mass via a rocket engine, or by storing rotational momentum internally. Incidentally, the second technique is only really viable in space due to the lack of any major medium, as friction would quickly degrade stored internal momentum.

The goal of these systems is to rotate enormous spacecrafts at reasonable or high speeds in combat, while being both cheap and non-massive.

The first technique is simple. Firing a thruster off center of your spacecraft will cause it to torque. Since rotation is the goal, and rotational acceleration is not, a second thruster must be fired to decelerate it at the end. And because such a thrust would send the center of mass off center, often two thrusters on opposite sides are used to start rotating, and two different opposing thrusters are fired to stop rotating.

A capital ship fires two vernier thrusters at opposite ends and directions to rotate. The rockets are water resistojets (hence the transparent plumes) and they tap into the main NTR propellant tanks.

These are called Vernier Thrusters, and see heavy use in space travel.

There are significant disadvantages to this method, however, chiefly that additional reaction mass is needed. Not only that, such thrusters usually can’t be Nuclear Thermal Rockets (NTRs), due to radiation concerns (recall that most crew modules are placed as far away from the main engine NTRs as possible). Cold gas thrusters don’t provide near enough thrust to be useful in combat, which means combustion rockets and resistojets are in.

Combustion rockets suffer from the issue that they require propellant(s) that are almost guaranteed to be different than NTR propellants, so additional propellant tanks must be added in, which takes up space and mass. This leaves resistojets as the prime method of providing torque to your spacecraft, since they can use the same propellant as your NTR.

Alternatively, instead of a system of multiple Vernier Thrusters spread out across your hull, simply putting a gimbal on your main thruster will do the trick instead. This allows your main engine to turn, producing off center thrust on your ship, rotating it. In this way, your main engines can serve the dual purpose of getting you places as well as orienting you in combat. This is one of the cheapest forms of thrust vectoring.

A few more complex thrust vectoring designs for achieving reverse thrust.

The primary disadvantage of it is that propellant is still expended when turning, however, no additional propellant tanks are needed as your main thrusters are doing the turning. For capital ships, the amount of delta-v spent turning is negligible, though for smaller crafts like missiles, the delta-v spent can raise a few eyebrows.

A more subtle disadvantage of gimbaled thrusters is that the size of the opening that the engines need to have can balloon significantly. This can yield much larger aft sections of the ship, and increase the targetable cross section of the spacecraft heavily.

Gimbaled thrust tends to be the cheapest and simplest solution to turning in space, and many capital ships and all drones and missiles in Children of a Dead Earth use them. Some capital ships opt for resistojet Vernier Thrusters, primarily for getting a smaller targetable cross section.

Screen Shot 2016-08-18 at 11.05.38 PM.png
A missile’s rocket nozzle is slightly gimbaled to the side, only a few degrees off center, generating tremendous torque. Screen distortion is from the camera being inside the rocket plume.

But suppose you don’t want to spend precious propellant turning? Then you need to invest in the second technique towards turning in space: exploiting Conservation of Angular Momentum.

Given a system without little or no external medium (such as space) and assuming you don’t want to expend propellant, the only way to rotate is by using conservation of angular momentum.

A quick example. Consider a pair of identical masses loosely attached to one another, floating in space. If one of the masses begins to spin in one direction, the other mass must spin in the opposite direction at an equal speed, otherwise it would violate Newton’s Laws of Motion.

This is the basic principle of operation for a Reaction Wheel, which is a flywheel with a motor attached. A flywheel is a mass with a high Moment of Inertia, or ability to resist rotational changes, along a single axis. When you spin a reaction wheel inside your ship using the motor, your spacecraft must spin in the opposite direction. Three reaction wheels must be used to get a full range of motion (one for each axis: pitch, yaw and roll). The Kepler Spacecraft uses reaction wheels.

A simple reaction wheel.

A Momentum Wheel is a reaction wheel which is constantly spinning at a very high speed. If a momentum wheel is spinning inside a spacecraft, simply braking it can cause a huge change in rotational momentum, causing a significant torque on your ship. Six momentum wheels are needed instead of three, two for each axis in opposite directions.

Momentum Wheels are often used for spin stabilization, as the huge amount of stored rotational energy will resist external torques. The Hubble Space Telescope uses Momentum Wheels.

Control Moment Gyroscope (CMG) is a single momentum wheel on a dual-axis gyroscope. By rotating the momentum wheel about the two gimbal axis, the angular momentum balance of the spacecraft can be altered at a whim. A single CMG can rotate a spacecraft along any axis by simply rotating the gyroscope to be in line with that axis. They are the most expensive and complex of the above systems, and are used on the ISS.

A Control Moment Gyroscope. The two axis of the gimbal allow the wheel to be oriented in any direction.

These systems promise the ability to rotate your spacecraft without expending propellant. Not only that, they require no exposed external systems, so they can’t be damaged unless the main bulkhead armor is penetrated. Sounds like a win-win, right?

Unfortunately, their effectiveness is very poor for combat operations. Modern CMGs are often used to very slowly change orientations over the course of minutes and hours. I originally had never intended to implement Vernier Thrusters or Gimbaled Thrusters in game, and was only going to use Momentum Wheels and CMGs. That rapidly fell apart when I did the math on them.

Consider the above example with the two masses on a string. Because they are identical, the spins will be equal and opposite. But if one mass has twice the moment of inertia, it will spin at half the (reverse) speed as the other. The greater the moment of inertia, the slower the spin.

If one of these masses is the spacecraft and the other is the reaction wheel, you want the spacecraft to have a lower moment of inertia. Thus, if your spacecraft has a moment of inertia 100 times that of your reaction wheel, it will spin 100 times slower than that reaction wheel.

Moment of Inertia is proportional to mass and the square of distance from the rotation axis. Roughly speaking then, very voluminous and very massive objects will have the greatest moment of inertia.

When headspinning, rotational speed can be increased by pulling ones’ legs in. This reduces the distance from the rotation axis and thus reduces the moment of inertia.

Spacecrafts by nature will have a greater volume than your reaction wheels, since they envelope these wheels. And they are guaranteed to have a greater mass than your reaction wheels, unless you are okay with an abysmal mass ratio (less than 2). Thus, you are guaranteed a moment of inertia far tinier than your spacecraft’s moment of inertia. And as I discovered, even spinning your wheels at enormous speeds yields rotations that take minutes or even hours.

In short, these techniques were not viable for combat rotations, barring some sort of future technology.

This leaves thrusters as the only viable method of spinning about in combat. How fast can they spin?

Because thrusters affect acceleration rather than velocity, the answer is that it varies. For instance, the time to spin 90 degrees is not going to be twice the time it takes to spin 45 degrees. And it varies based on which axis of rotation is used.

A simple metric is the Full Turnabout Time, which is the time it takes to spin 180 degrees about the slowest axis. This is essentially the “slowest” possible turning time for the ship, and most turns will be much faster, a fraction of this time.

When the enemy targets parts of your ship away from your center of mass, you can dodge simply by spinning along one axis. This particular ship is spinning with a single gimbaled NTR.

For medium sized capital ships with gimbaled thrusters in game, 20-30 seconds is a common value. Capital ships with vernier thrusters tend in the 10-20 second range, as do small sized capital ships. Very large capital ships can take up to a minute to do a full turnabout. Gimbaled drones and missiles tend to take 5 seconds or less for a full turnabout, with some being able to do a 180 in under a second.

Much faster turnabouts are possible by simply adding more and more vernier thrusters or gimbaled thrusters. However, this is often fast enough to deal with the rapidly changing nature of space combat. It is rare for a capital ships to ever need to flip a 180. Most of their turns are much smaller angle shifts, small dodges and broadsides.


34 thoughts on “The Wheels on the Spacecraft Go Round and Round

  1. What ends up happening in space combat that you need to whip about at all? Without having seen your game in action, I’m imagining a slow approach, Waltz of the Danubes playing, where both warships have their most armored and lowest cross section portion facing the other.


    1. Based on previous articles and the few videos on Youtube so far, typical range is not that long, maybe a few dozens km, while relative speeds are still in km/s. So there is a need for ships to turn relatively fast, particularly when armed drones zip by.
      This comes from the very conservative technologies used, I’m curious how fast a RBOD-armed torchship has to turn in combat.

      (The videos are the trailer: and a timelapse of a mission: )


      1. Yeah, I’ve been assuming relatively straightforward advances that are probably still technically possible. Like aneutronic fusion for the power source, which increases the specific power factor by 1-2 orders of magnitude. (because an aneutronic fusion reactor uses the escaping electrons and helium atoms from the reaction itself to directly drive an electric circuit, this conversion ratio can be 90% efficient). Droplet radiators that use liquid tin or even more exotic eutectic metal compositions. 10-100 times more power to weight, and you could easily pack megawatt class lasers, with 5 meter mirrors, that would burn enemy ships in seconds from >1000 kilometer range. (basically under 1k kilometers is lethal laser range)


  2. Is it possible to use hollow reaction “rings”? You would get more momentum from the same mass, as it would be further from the rotation axis.

    A while ago, I imagined a warship built around a spinal mounted gun stuck between two giant rings used as reaction wheels (plus two medium rings for the second axis). It was for a soft(er)-SF universe, but I wonder how well it would have worked, physics-wise.


    1. I can answer this, the TLDR is that making the rings bigger doesn’t actually help. Angular momentum is still directly proportional to the stress in the material the disk in the rotor is made out of. Whether you make the ring big or small, the stress is the same. That’s the limiting factor – you can spin a disk at a million RPM with electromagnetic bearings, but it will explode unless it is a very small disk made out of carbon fiber.


      1. Ah, I see, thanks. So the best material for a rotor (assuming no volume constraints) would be the strongest one, like diamond or woven carbon fibres.
        Why are today’s flywheels are made of metal? Is it for volume, ease of making it an electric motor? Or are the strongest materials at this scale still metal today?


  3. Have you considered thrust control vanes, a la the V2? A lot of missiles use them, too. That would cut down on your skirt size but probably impact performance.

    Note: before people assume I am an idiot for talking about control vanes in vacuum, I mean control vanes IN the exhaust nozzle.


    1. (Yes, I know, there is a Wikipedia link right there describing it amongst other forms of thrust vectoring – I’m just asking why they’re not included)


      1. Thrust control vanes are problematic since the nozzles are mostly concealed by armor (i.e. the exhaust would be shot back against the inside of the armor). They would have to integrate the thrust vanes directly into the armor, but that becomes a major problem when you have multiple engines inside of a single armor skirt.


        1. I’m sort of wondering, there. If you look at a setup like that used by Copenhgen Suborbitals on their research rockets (in particular the Sapphire and Nexø I rockets), they don’t particularly look like they would need extra armor, or interfere with an armored cowling.

          CopSub’s page on the Sapphire page has some imagery of the configuration. Nexø I used as similar setup.

          Disclaimer: This poster may be an idiot and not realize really obvious reasons why this would never work.


          1. Ah, my mistake, I was thinking about reverse thrust or side thrust exhaust vanes. If the angle of exhaust deflection is kept small, then exhaust vanes can be armored without issue, and you can have torque.

            Most modern rockets use gimbals, because the torque can be made much greater. A gimbaled rocket can rotate all of its exhaust whereas small angle thrust vanes only partially affect exhaust direction.


            1. It would allow the rocket engine to be entirely concealed by armor though, without an open skirt. Also could you have small fairings like on the S-IC extending around the outboard F-1 engines to give room for gimbal, instead of widening the whole opening? Would that be worth it?


              1. That’s true, though allowing a rocket to only gimbal a few degrees keeps the skirt narrow too.

                I don’t think keeping the nozzles outside the armor is a good idea, since they are very fragile. Even if the nozzle is sturdy, damage to it can choke the nozzle and potentially cause a fatal pressure build up. This means most rockets have to be disabled completely when any damage is detected.

                The open armor skirt is still a problem, but it does help some. In particular, in protects very nicely against stray nuclear blasts unless they are directly behind the craft.


  4. For a more outlandish option, how well would it work to use the propellant itself as angular reaction mass? For longitudinal rotation, we could have the tank itself being spun. there would have to be baffles inside to drag the fluid with it, alternatively only the baffles could be rotated inside the tank, if it is easier engineering-wise. The reaction mass would be lower, but probably not by that much.
    For transverse rotation, it would be a bit more complicated: you would need wheels with baffles rotating in the transverse axis inside the tank.
    Alternatively, the tanks could be divided in multiple chambers, with propellant mass pumped from one chamber to the other.

    I have no idea how practical this is, though it may be recycled as hard-SF plot element when Our Heroes must turn the crippled ship around despite crippled flywheels/verniers/gimbals.


    1. That’s definitely an interesting idea, using your reaction mass as a reaction wheel!

      Some quick notes though. I think the biggest issue is that your maneuvering speed would gradually reduce as you expend delta-v.

      Splitting your tanks up into different “wheels” would be necessary, or maybe only using spherical tanks which can spin in any direction.

      The interaction of the rotation with either the surface tension netting or the diaphragm, whichever is used, would likely cause troubles. Additionally, turbulence and vortexes forming in the fluid would be another major problem, possibly causing cavitation. Cavitation is already an issue when trying to pump the fluid into the engines, as it can shatter the plumbing, so it’ll only get worse when spinning the tanks.

      Lots of potential problems, but it’s definitely a novel idea. I think the idea would work much better for solid propellants (like for an Orion drive).


  5. Somewhat unrelated question, but are solid-fuel rockets modeled ingame?
    A missile designer might opt to use them for last minute high speed course correction during the missile’s final attack phase.


    1. No, the inability to turn them off once active is a major problem. As you mention, it would require having a liquid fuel rocket for the most part, and then a solid fuel rocket for only the terminal phase. This mostly defeats the purpose of a solid fuel rocket, which is that it’s so small and compact.

      For the most part, I’ve found monopropellants like Nitromethane can provide enormous thrust, more than you’ll ever need for a missile.


      1. The idea was more along the lines of mounting it off-center the nose, and using it to execute a snap-turn faster than simply gimbaling the engine would allow.

        There’s a missile IRL that does this to intercept other missiles-I forget the name.

        Would solid fuel rockets be cheaper than nitromethane, deltaV wise? A Spacy strapped for cash might opt for a cheap, short range missile/rocket weapon on temporary external racks to supplement the main guns.


        1. Gotcha. You can mount liquid fuel engines off-center like that (and it costs no additional reaction mass, since it can use the main fuel tank).
          Cost and fuel density are the main reasons for using solid rocket fuel. In terms of cost though, missile propellant generally isn’t the bottleneck. A nuclear or explosive warhead usually is many times more expensive than the propellant and engine itself.


        1. I am aware of pulsed solid rocket motors, as well as hybrid rocket motors, both of which allow turning on and off. However, in both cases, the cost and complexity increases to be much closer to liquid fuel rocket motors (hybrid rockets especially, since they’re similar to bipropellants). Pulsed rocket motors also can not be precisely controlled, thrust is very granular with them.


          1. Yes, you manipulate them by controlling pulse length. Basically modulate in the time instead of amplitude domain. I think they would be extremely useful in space warfare for hybrid railgun/guided ammunition. The solid rocket motors would have an easier time withstanding the thousands of gs of acceleration during a railgun launch, and a ship firing this way would have immensely more range for the same mass of ammunition.


  6. Also, another unrelated question: Is it possible, IRL and ingame, to use a pyrotechnic payload and an engine together to mimic another type of engine for short periods of time?
    It wouldn’t fool anyone strategically, but it might fool a missile in terminal phase. Or point defense during same.


    1. Yes you can, and yes it has been done by some of my testers. I personally avoid it because I find the mass cost a bit too high for my tastes (an engine plus a pyrotechnic payload plus reaction mass is rather expensive in terms of mass for a single decoy).


        1. Using decoys for drones is a little difficult because drones shoot, which means back tracing drone bullets is a guaranteed way to always know which drones are real and which are fake.


          1. Unless the drones are armed with rocket-propelled projectiles, which they could just threw away and ignite with some delay – so the launching point of projectile would not correspond with the drone position.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s