Introduction to Reaction Wheels

Introduction

Reaction wheels are used to point spacecraft where you want them to point. They don’t provide as much torque as Control Moment Gyroscopes (CMGs) and they require more power for the torque you get, but they’re smaller and lighter than CMGs. In fact, if CubeSat and small satellites have any kind of wheel, it’s a near certainty they’re reaction wheels.

A Bit More Detail

The high-level concept is if your spacecraft isn’t twisting and you want it to twist, then you should twist something. And if your spacecraft is twisting and you don’t want it to twist, then you should twist something else. 

The tool that most spacecraft use to do these things is Wheels. The wheel type that most small to medium spacecraft use are Reaction Wheels. Some people call them Momentum Wheels. And some just call them Wheels. They are just wheels that spin faster or slower or reverse direction to get the balance of twisting in the spacecraft just right.

The other common wheel type is Control Moment Gyros (CMG’s), which are just spinning wheels that get tipped over to create their twisting loads. These tend to be heavy and large and expensive so you usually only see them on the largest and/or most important spacecraft.

There are two main reasons you would use a reaction wheel on your system:

1. You want to point at things, like with a telescope. Pointing in space just means twisting accurately.
   
   
2. You want to keep your spacecraft from tumbling. All sorts of things in space make a spacecraft tumble, especially if it’s flying close to a planet like Earth. There are magnetic fields pulling on any magnets in a spacecraft, atmospheric drag (even though there isn’t much atmosphere), micrometeoroid and orbital debris collisions, solar wind, the oblateness of the earth, gravity gradient forces, and more. 

All of these forces are doing two things to your spacecraft: moving it and twisting it. Wheels in space are one tool you have to combat the twisting part. (Just to be clear, wheels cannot directly help you with the moving part, just the twisting part.)

One of the key things about reaction wheels is that their mass tends to be much, much less than the spacecraft’s mass. (Some CubeSats are an exception.) This means two things:

1. To get the spacecraft twisting (or not twisting) a little, the wheel has to change its speed by a lot. It can spin hundreds or even a thousand times a minute.
   
   
2. You can make your spacecraft point accurately because your wheels can change speed a little bit to change your spacecraft’s twisting a tiny bit.

Another key thing about Wheels is that when they stop working, it’s pretty much the end of your spacecraft’s mission. You can use your thrusters for a while (if you have them), but that usually doesn’t last long and then all that hardware is going to tumble out of control. This means that whoever is paying for your satellite is going to have a lot of questions about how you designed your wheels and why you’re very (very) sure they’re going to last a long time. 

People new to space systems may not be aware of how much attention and time and money will be spent scrutinizing these things. There is no such thing as, “just buy wheels” because you need to know about your mission requirements, how much you think you’ll weigh, what the distribution of that mass is, and so much more. Do not underestimate how much you should budget for these things.

A reasonable assumption is that wheels on medium to large systems will operate at around 2,000 revolutions per minute. They work better when they’re already spinning because of dynamic versus static friction, because they’re through their dynamic modes, and because the system reaches a steady state in terms of lubrication. If we assume 2,000 revolutions per minute and there are about 525,000 minutes in a year, then your wheel will turn over a billion times in a year.

A Bit of Theory

The mathematical term you’ll hear is angular momentum. For the level of discussion we’re having, this is just a fancy way to say how fast the wheel is turning. For Reaction Wheels, it’s the change in speed that matters.

Important Point: It’s the change in the wheel’s speed that matters. A wheel that speeds up from 0 to 10 revolutions a minute will produce precisely the same change in your spacecraft’s twisting as the same wheel speeding up from 2,000 to 2,010 revolutions per minute. Both are changing by 10 revolutions per minute and that’s all that matters. It does not matter how fast your wheel is spinning, only how much (and how quickly) your wheel changes speed.

Design Considerations

Because billion-dollar missions have ended when reaction wheels failed, a lot of lessons have been learned:

1. Lubrication matters. Wheels can spin thousands of times a minute, they need to be able to do this for years, and you can’t change things after you launch. You need a lubricant that doesn’t break down or move even though there are high speeds, heat, and the vacuum of space. Most systems now have a barrier to keep lubricant from creeping out.
   
   
2. Bearings are tricky. You’ve got to get the contact angles, loads, and materials just right. Newer reaction wheels have been moving to ceramic bearings or magnetic bearings to get away from some of the problems with metal bearings. See Reference 2 for an example of unexpected failure.
   
   
3. They jitter. Even precisely tuned and balanced wheels create a little bit of vibration in the spacecraft. This can mess with cameras and other sensors that want things quiet. Some spacecraft are not allowed to change their reaction wheel’s speed when the cameras are taking pictures, for instance.
   
   
4. It’s easy to overshoot your goal. It takes time for wheels to get spacecraft spinning the right way and it takes time for them to get the system to stop. Real spacecraft tend to overshoot their target, then go back and forth around the target dialing things in. A good engineer can help you minimize the “Settling Time”.

Failure Modes

The most common way they fail is when their bearings stop providing a smooth, low-friction rotation. This forces the motor to work harder, which takes more power and generates more heat. It can also make the wheel vibrate more, which can be bad for sensors. Although there have been stories where a “sticky” wheel gets better, it’s almost always the case that it continues to get worse. Eventually, the cost isn’t worth the benefit and you shut it down.

Because the wheels are in space and cannot be inspected, it’s difficult to know for sure what the root cause of this failure mode has been. Leading theories include:

Lubrication failure. Some of the older lubricants would creep and leave the mechanism “dry”. Others would evaporate in the vacuum of space. And others would get gunked up with metal particles. This is one of those “gotcha” areas that you end up spending significant time and money on. Side note: some missions that suspect lubricant problems due to overuse have been able to turn their wheels off to let the lubricant settle out again and turn them on again successfully. Some systems use an active lubrication system that can add lubricant from a reservoir or a type of sponge.

Bearing materials. Traditionally, they’ve been made of metal. Older bearings couldn’t take the load, though, and they would get very small flat surfaces that caused jitter. Others would get worn down in a groove over time because of bad racetrack design. To combat this, the metals became harder and more durable over time. However, most new models are replacing metal with ceramics that are harder and stiffer. 

Or they’re even going to magnetic “bearings”. Note that magnetic systems will give you a longer life and tend to create much lower vibration levels, but cost you ten times the price so many missions choose to go with the lower-cost option.

Improperly balanced wheels. Even tiny imbalances add up, grinding away on surfaces. Modern wheels are usually balanced on special equipment that can tell you to add/remove fractions of a gram at a given location.

Launch loads. Wheels tend to be massive relative to the spindles/axles they’re mounted on. Your structural analyst needs to make sure the load input to the reaction wheel interface matches what it can handle.

Radiation. Like any electronics in space, they are susceptible to the radiation environment. They don’t tend to carry significant logic chips, though, so tend to recover from those events as long as nothing shorted. Interestingly, a 2017 paper seems to show that solar storms can damage reaction wheels in an unexpected way: they create an electric charge build-up between the bearing’s race tracks. The metal bearings provide just enough of a conduit to release a spark that creates pits in the metal and/or damages the lubrication. See the second reference for more information.

Summary

Reaction wheels spin faster and slower to make your spacecraft spin faster or slower. They’ve been used on hundreds of spacecraft so you benefit from decades of people improving them. You still need to do your homework but it’s likely you can find an “off the shelf” solution that will work great for you.

Glossary

A few terms you might hear when people talk about reaction wheels.

Caged: Set the spin speed back to zero. Equivalently, set the angular velocity to zero (or whatever its happy center is). It usually means the wheel is not turning relative to the spacecraft.

Flywheel: Another name for just the wheel or disk part of the reaction wheel.

Momentum Wheel: Just another name for Reaction Wheel.

Saturated: The wheel can’t spin faster. You've gotten as much useful torque out of it as you can get.

Curated Videos

1. https://www.youtube.com/watch?v=woCdjbsjbPg
   This is great for showing that reaction wheels don’t move things, they help point them in the right direction.
   
   
2. https://www.youtube.com/watch?v=vqV6nazapFQ
   You can build your own test system with the plans provided by James Cochrane.
   
   
3. https://www.youtube.com/watch?v=ZPWiIBcHOh4
   The University of Bristol has created an educational video with hardware demos. It’s short and gets right to the point. 

Curated Links

1. https://cdn.intechopen.com/pdfs/13470/InTech-Lubrication_of_attitude_control_systems.pdf
   This academic paper discusses lubrication failures on expensive systems and how researchers have addressed them. A great read!
   
   
2. https://www.aero.iitb.ac.in/satelliteWiki/index.php/Reaction_Wheels
   This is a wiki-type website for satellites. Their reaction wheel page introduces some of the equations and concepts. As of this writing, it’s better than Wikipedia’s page on reaction wheels.
   
   
3. https://pressbooks-dev.oer.hawaii.edu/epet302/chapter/7-7/
   The University of Hawaii hosts a wonderful satellite education site. The link takes you to one long page that covers various control methods, of which reaction wheels are a section. It’s great for showing how they fit in the ecosystem, though.

References

1. Curtis, Howard D. Orbital Mechanics for Engineering Students. Amsterdam: Elsevier Butterworth Heinemann, 2005. Print.
   
   
2. https://esmats.eu/esmatspapers/pastpapers/pdfs/2017/bialke.pdf
   Bialke, William, and Eric Hansell. "A newly discovered branch of the fault tree explaining systemic reaction wheel failures and anomalies." Proceedings of the European Space Mechanisms and Tribology Symposium. 2017.