RC controlled Reaction wheel flying device.

[MT]

Interesting!

[Brumby]

A great exercise for someone wanting to learn about control systems.  Almost diabolical for a beginner.

[CatalinaWOW]

I believe many control systems of this type control the body by applying a torque to the momentum vector instead of the spin axis to allow a more rapid response.  There is then another loop that over much longer time periods controls the spin rate.  Much thought goes into making sure everything stays within limits and when and how to dump or add momentum to the wheel.

The saving grace is that for many systems the average over time of the control required is zero, so ignoring friction there is no cumulative need to add or subtract momentum.  The problem is that this condition can be violated by initial inbalance, and external forces can cause a non-zero average over periods that can last hours or even days.

[Red Squirrel]

Been following that channel for a bit, he does some pretty neat unusual stuff.   I think he almost had it, probably just the thing of tweaking things.

Wonder if it would be better to spin the wheels at 50% speed initially, that way you just slow down or speed up, but always keep same direction.

[Brumby]

The "50% position" for reaction wheels is stationary.  They then accelerate in one direction or the other, depending on need.  If the flying platform had a naturally stable hovering position, then the reaction wheels would not be rotating at all.

The setup used obviously had some inherent stability issues which were too extreme for the control system to deal with.  Perhaps tweaking might home in on some better results, but from what I saw, it was a long way from critical damping.