The leg looks good, but the hip bracket looks kind of flimsy to me... if possible, build a cycloidal reducer for the hip motor where the output pins run all the way through to flange outputs on both sides, so you can put a second bracket on the top side. That would give much greater stiffness just by geometry, allowing the brackets to be thinner and lighter, and bearings to be less stressed.

I'm not sure I fully understand why a cycloidal reducer would make this more stable. Can you point me to an example? As for reducers, the motor I'm using already has an 8:1 built-in planetary gearbox.

But you're also right, I'm not 100% happy with it and am still playing with that part of the design. I plan to try making it out of aluminum to see if that improves the stiffness. Also, possibly adding more of a flange and placing a thrust bearing between the motor and the bracket for additional support.

I suppose you could have two output flanges on the carrier of a planetary gearbox too, but cycloidals are easier to make. But if you're using off-the-shelf mini-cheetah actuators, then nevermind the dual flange output. The important thing is to have another bearing above the hip motor so you can have a second bracket. It doesn't have to be actively driven by the motor, just structural support for the leg so the actuator's bearing isn't getting cranked by the moment load. And then both brackets can be simple flat plates rather than the triangular thing you have now.

And of course feel free to ignore my ponderings and continue with it as-is :) The leg motors may not even be strong enough to carry a heavy enough load to stress the hip actuator bearings. And it's probably easier to construct without the second bearing.

I'd love to understand your suggestions better. I'm not happy with the rotator bracket setup either. I'm not sure I follow your description of a bearing above the motor and a second bracket. Could you draw this?  If you're interested, you can view the leg model in greater detail, here: https://a360.co/3Sp2B7d

Here's a really crappy edit to show the general idea. The bracket (dark gray) is screwed to the leg actuator, and turns freely on a bearing (light gray) mounted to the body structure above the hip actuator. https://imgur.com/a/V4rYBor

But the more I think about it, I don't think this second bracket would be enough benefit to bother with. Unless the axial distance between the two bearings is greater than the diameter of the single actuator bearing, it won't have much effect on bearing stress. And it really only provides support for force in the vertical direction. Horizontal force on the foot from walking will cause a twisting force on the hip, and a single bracket that's stiff in all directions will do better at resisting that than two flat brackets, one of which is not actively driven. And if you did actively drive both with a double-output cycloidal like I was imagining at first, the distance between its two output bearings would be even smaller (just the thickness of the discs), since the motor would have to "float" above the top bracket with its shaft sticking down into the cycloidal.

And on top of all that, the bracket you have is probably at least as stiff as the carbon tubes supporting the leg in the first place.

Continue as-is!

Oooooooo! Thank you for the drawing; this makes a lot more sense, and I like it. This would reduce the angled/prying force pulling down against the connection to the motor. This is much better than the other iterations I've been playing with in my head. I'll keep this in mind as I move forward on the project. Thank you!