Ksp players.

Parent: Son, why is the lawn still unmowed?

As professional multicopter flight control engineer who did a lot of such angular rate loop experiments with DIY builds I would say the biggest catch with your quite nice setup is control bandwidth. The delay from the gyro measurement until your reaction wheel is finally counteracting the disturbance is too long to keep the drones attitude stable in the air. It took me a long time to figure out this problem with my own first flight control attempts (all from scratch). I'm guessing the bottleneck might be the motor controller of your brushed motors it's very likely not designed for low latency feedback loop reactions which all good multicopter ESCs are. The key is reducing the angular rate loop delay and then tuning mainly the angular rate proportional gain. You need to be able to feel a virtually instant and rather strong counter reaction from the actuator already in your one axis hanging test when pushing with the finger even only slightly. If you cannot get the gain high enough because it starts oscillating early you either have too much delay or too high vibrations in the gyroscope measurements. Be aware that the way (stiff, soft, vibrating part, ...) your gyroscope is mounted to the drone matters in this regard.

Hey, a retired aerospace engineer here. All I can say is that your videos are a ton of fun to watch, and really inspirational; I should go build something. And I really liked your explanation of why you would try to use reaction wheels to control a drone.(something like "I have no idea") That is completely cool.

Honestly, the fact that you got as far as you did with this project is amazing. I mean genuinely incredibly impressive that you were able to get it flying at all with even a modicum of stability. I would've expected this to never work in a million years just due to the difficulty that rotation stabilizers like that have of doing the fine controls that are needed to fly.

Well colour me impressed.

You got very sidetracked,

Great Experiment. But I feel that you missed a big opportunity here :

I'll be honest this seems like the kind of thing someone in the 1800s would've thought up

Would love to see more reaction wheel stuff

You just recapped 3 months of Physics class in 1 minute and 30 seconds... what have I been doing with my life?

Make sure your gimbals are center of mass and aligned with the other wheels rotational axis. Take into account gimbal procession induced from internal and external forces. Another useful tool would be utilizing the stabilizing effect of a three axis, high rpm, set of gimbals. This would help hold the craft in orientation and reduce wobble.

I did my physics 'O' and 'A' levels over 35 years ago - was good then, would have been a hundred times more fun doing it now with drones and computers! Good video!

i know all about reaction wheels i play kerbal space program

Fantastic original thinking -- your best project since the trebuchet. It wouldn't even have mattered if it hadn't flown at all; the concept itself is just great. Well done!

"Let me build a test rig"

Keep it up Tom, I love these vids! Trying out stuff that is outside the box. Maybe the industry engineers know this stuff, but for you to show the different things you try with your 3D printed bits and pieces (which I love BTW) to show us the inadequacies of various effects, is just brilliant. keep it up.

You madman! That's a great spin on drone control! Now I really, really, really have to make myself a CNC.

24:15

Excellent proof of concept! Your channel is one of my new favorites thanks 😎👍🏼