This post reviews some of the design concepts of a Control Moment Gyro (CMG) for the mech using a real-world example. There are plenty of academic papers looking at certain pieces of CMG technology but few that would be enough to tell you how to build one.
If you want to jump straight to the patent textual description of the parts, go ahead (WARNING: gory patent lingo). But for this post we’ll walk through the primary parts one by one and consider some of the issues for each.
First here is a reminder of what the completed system looks like, including twin CMGs installed in the base of their vehicle:
This annotated diagram shows the six main parts of the system, most of which is hidden deep inside the assembly in the above photo:
The housing is the upper/lower portions of assembly, holding all the guts together inside. Strength of the housing can be an important factor in designing for safety. As the flywheel can potentially break up, it is crucial to have as strong a housing as possible to help mitigate injury from 10K RPM and higher shrapnel.
The housing also serves to keep everything aligned well. While this whole device requires precision machining the alignment pins, axle motors, etc. all have to line up well and most will be connected directly to the housing.
Thickness is an issue as well, especially if the device becomes heated and is tempted to warp or expansion – though one may argue that if it is that susceptible to heat issues then you’ve got other bigger issues to deal with! Thickness comes into play when fastening housings together and connecting other parts to the housing, enough depth is required to provide a solid connection. Keep vibration in mind as well, obviously not something that should occur but at such high rotation speeds, ensuring the housing is as solid as possible is absolutely critical.
As noted further down, air tight seals will likely also be necessary.
Bearings / Gimbal
There are two critical parts for handling rotation. First is the bearing assembly for the flywheel itself. It’s not entirely clear in the diagram, but it is assume that the shaft going through the flywheel is held in place by bearings at the top and bottom of the shaft. Some larger-scale generators that use flywheels use a magnetic bearing to reduce friction as much as possible. For a gyro in a vehicle this would not be possible due to movement, so some sort of solid circular bearing would be used.
It has been discussed historically that the bearings can be the limiting factor in success for gyros that run long term, especially at high speeds or heavy weights. We shall see.
The other area where rotation is at play is at the gimbal point. There are various gimbal designs you should be aware of – see the link to get up to speed. In this design it is quite simple – the housing is suspended on two pivot points and made to be able to swivel by a motor. It still works the same as a more complex gimbal system, but since there is gimbal rotation only on one axis, it’s quite a bit simpler.
For a mech design a two axis gimbal would be desirable. This would help provide a full range of forces for controlling the orientation.
Here’s a video showing off the internals of the C-1 and its CMG – see how the gyros are rotated compared to the above picture? That’s the rotational around the gimbal axis:
The flywheel itself is obviously the most delicate in terms of design choices as it is likely a solid piece of heavy metal and/or carbon fibre. See wikipedia for more examples.
The key variables in how much power is output from the gyro are the mass, radius and angular velocity (RPM). It is said that an increase in rotating speed is 4x more powerful than adding more mass or increasing the size of the flywheel (citation pending, sorry). Once you know these variables you can start to compute the angular momentum and the output torque that can be provided to the CMG overall.
Although the hope is that the mech weight won’t be an issue, it is somewhat easier to increase RPM as you don’t have to go and machine a whole new flywheel to do more tests!
Aside from having adequate bearing support, cooling and/or lubrication, the key safety and performance concern is with the composition of the flywheel. A toy gyro has very little mass and can be stopped with a finger. The industrial sized gyros will take minutes to spin up to cruising speed and would probably love to take your finger off if you go poking its face!
With all the momentum of the flywheel there are clear risks that the components could literally fall apart – spraying a wide field of shrapnel through the behind of the pilot. One way to help mitigate this risk is to use a material like carbon fibre to wrap around the leading edge of the flywheel – helping to encase and strengthen the metal flywheel beneath. As the carbon tube like structure is more fibrous than the metal, it acts much like a cage helping to keep things together.
Note that some of the reasons you don’t have a flywheel already storing energy in your house (yes, they are more efficient than batteries too!) is due to the safety concern. There are historic research reports of fatalities from delaminating flywheels – that’s not something any of us want to replicate!
Velkess is one company that’s hoping to change that, by using stable materials and lower speed to still make an efficient storage system. We’re watching their progress and hope we can also benefit from gains made in this area, though it may not really help in a CMG environment.
Closely related to the flywheel parts are other methods to reduce friction. Many high performance gyro environments manage the air friction within the device. This can be done using a vacuum system to evacuate the air from the housing. Obviously the housing needs to be able to withstand those negative internal forces.
While vacuum control is not shown in the patent diagram we believe that in a larger mech environment it will be necessary. Whether it is part of the C-1 we are not certain at this stage. Though it’s not obvious from any photos we’ve seen we assume it will be at some point.
Two motors are used for each gyro in the CMG. One is a multipurpose motor/generator allowing the system to pull power back out of the gyro as well as put more in.
This primary flywheel motor gets the flywheel up to its cruising speed and then kindly steps out of the way. It is not clear whether power is continuous or intermittent, perhaps a PWM approach keeps it spinning with minimal effort and expensive.
The other motor is called the precession motor – it sets the angle of the gyro, initiating changes to the gyroscopic forces as the gyro turns. These motors must be fast acting, so signals from the control centre can make very fine adjustments very quickly. These may be a sort of servo, though any kind of motor that can track its position may be used.
In a mech environment, the need for very intricate adjustments may be less as the weights and forces involved are so large. A brute force approach that is a little sluggish would likely be manageable.
Which brings us to our final question re: parts of the CMG example we’re looking at. We haven’t confirmed the actual materials used for the C-1 CMG, but in other arrangements we’ve found that stainless steel was used in the flywheel for a large aerospace project (Boeing’s ISS CMG). It is unclear at time of writing what an optimal approach would be for balancing safety, weight and ability for higher speeds.
The same question applies to the housing. If we look at it as a containment system, then apparently stronger materials would be of greater benefit.
More on all these topics as research continues. Comments welcome 🙂
References and papers:
- Wikipedia: Gimbal – Flywheel – Control Moment Gyroscope – Gyroscope
- Videos of C-1 that include some peeks at the CMG in action – __1__ – __2__ – __3__
- Mitsubishi anti-rolling gyro (marine) – Seekeeper design
- Boeing/NASA review of failed ISS gyro – basic ISS CMG info – and more here
- CMG in a backpack experiment
- Academic papers on – CMG control – table top CMG (lotsa math and construction notes) – another source for equations