Have the central core suspended on a rod with some bearings. Put the motor on the outside of the “base” turning it like a big gear. This will give me a lot more support for the weight I think. Maybe look at an adjustable counterbalance, read as the upper arm would have a rail that a weight would run on to offset the force on the other end of the arm. This will probably also help keep the power minimized in the neutral position if I want it to be retracted. If the arm is heavy enough, and with enough dampeners, I could just move the whole your arm on that pivot by adjusting the counterbalance’s position. I need to measure the room to figure out the range I need. I’d like it to be able to reach the far end of the book case to my work table. That might be a stupid want, but I never claimed to think things all the way through before talking a lot of shit. I’m kinda thinking for the elbow to do a spinnig rod driving an offset joint. None of what I’m saying is particularly accurate for serious people terms. I’m not serious.
I did those things I talked about previously. Badly but done. I’ve also started writing up the process for how I’m doing an install by respond the same steps for another SD card.
I also had some plumbing changes today, so I placed an order for some stuff, but I wanted it with free same day, so… I have no self control. I have amongst the things I needed, two stepper motors and five drivers.
This of course means I’m working on things a bit faster than I usually like. You know, doing rather than talking.
My concern is starting to be the weight and torque of what I’m thinking about. Within reason, anything that is on the ceiling mount itself is effectively weightless. Anything on the arm though is going to be multiplied in the bad way by the length of the arm. That could be an issue. I see a couple paths forward. I’m going to remove the possibility of a shorter arm, no retreat, no step back. With that being said, I either need to make the arm stronger or reduce the forces. I’m cheap, no unobtanium for me, so it’s looking like I need to reduce the weight on the end of that arm. I’m thinking Bowden tubes, because they’re my solution to everything. I could also look at using timing belts to the same effect, but if then have to build in a lot of other stuff. Every time an arm moved, it wild change the tension on the other belts, so I’d have to do something like retract motor a if motor b extended. This isn’t a have killer, but something I have to keep in mind. I wonder if this would be a good use case for an FPGA. I could just look at doing this like an adult, with motors all the way down, but even the human body doesn’t do that. Hydraulics? On the ceiling? No thanks.
I suppose the first step, past base software install, is build an arm bracket that I could use to study the arm at a good working height from, and then start building an arm. Without doing a lot of math and research into materials, I won’t know the exact forces at play, so maybe I should just experiment with the forces at play.
I’m also continuing the evolution on my thought on counterbalances and the kinematics. Let’s say I have an arm with a 10′ reach, with 5′ on the upper. This means I gave an arm that’s going to have a hard time ever extending unless it’s doing some work looking double jointed action due to height of the ceiling. This is why I sometimes scoff at people going all biomimicry. I could have the arm, mounted on the central point not by a joint, but a slot. On the back end of the arm, I could throw on some ceiling rolling wheels.
Ceiling wheels – yes; bowden tubes – no, maybe, I’m thinking. I could look at a gravity tensioned cable drive for the arm. The gripper should probably it’s own motors, just the speed factor and grip force. I guess I could look at doing gravity tension all around. I still might look at Bowden tubes just for the cleanliness of the lines. That and so I don’t end up with dirt wearing down the cables. If I go with a cable drive that’s where I’ll want to look at some fat DC motors with a bit of homebrew encoding.
First build a test plate. Then start tests. I don’t need to design a chariot before before building a wheel.
Tesselated set of ceiling rails? Slots that support the weight with legs giving reasonably 2-dimensional freedom, and a crane system for the Z? Ceiling mounted almost polar coordinate system?