Dear Chris,
I've seen all the incredible work you've put on this project and I'm willing to first thank you for documenting it thoroughly.
However for my application I need speed and looking at your video portions where it was running at 100% it seems way too slow for my needs.
Then I went on documenting my self about the big industrial robot and I've realized that stepper motor are used only for the grip of certain robot "hand" (soft robotic is much better though), but for main axis movement they use brushless motor + encoders. This video seems to confirm my finding
This is 1o1(+1) comparison testing
https://youtu.be/p4ltHDpxrbIcoming
Interesting enough ODrive can also be driven in step mode.
It seems like an ESC with some nice interface/api (in python) to control it and UART/SPI/1wire comunication.
Quite expensive it though... €119.00 to control 2 motor so 357€ for 6 DOF + encoders + motors
but I can avoid arduino entirelly and control it with the rasberry.
So my plan is to buy all the hardware stuff for AR2 and apply it for brushless.
Your thought
--
Eli
That looks very interesting. I will have to spend some time and review the O drive. I think you could certainly adapt the robot and software to use that but you would need to make sure the motors you select will have a gearbox that can bolt on or be adapted easily enough. Im working on the next version robot which I hope will be larger and faster but I have a ways to go.
I think some parts will need a redesign. I'll proceed step by step and find out what is needed for each motor-section. I have no experience in this field (but I can do programming and 3d design), so I'll start with a simple linear rail and build my way up.
I wonder if the planetary gearbox ratio you have chosen should stay the same given that brushless can go faster. I found just the gear-heads but I don't know about the ratio.
I think the ratio can vary - as long as it can provide the torque necessary. If the motors can be driven by steps then you can adjust the calibration on the calibration tab.
another interesting find is the very compact PM6025 with encoder
https://bit.ly/2ABf28v
with PWM and SPI interface
and the bigger PM120
https://bit.ly/2SAZyIC
so many options...
Actually what I said about big robots seems incorrect
https://youtu.be/6YiPrytt_Ss
these looks like just giant stepper motor + encoders
https://www.rls.si/en/products/rotary-magnetic-encoders/absolute-encoders/rm08-super-small-non-contact-rotary-encoder
I think you might be right that brushless is the way to go long term with the motors. I like the idea of a faster top speed and not loosing power with speed.
In addition, if you are using the ODrive, having it open source gives you nice control over power curves and sensitivity which is really nice.
This is an interesting chip along the same lines: https://ams.com/as5048a They have lots of interesting, and rather inexpensive (relatively) prototype boards.
Most industrial robots use ac servo motors, not dc. Odrive has step/dir interface so they're a drop in from the controller side, uart or can would be nice tough.
Then you need to modify the robot for the new bldc motors and add more reduction since the motors are high-speed, low torque. You can only really update joints 2, 3 and 5, joint 4 likely needs more space and joint 6 and 1 are a little unnecessary.
I noticed CUI has an encoder that provides comutation signals for BLDC motors: https://www.cuidevices.com/product/motion/rotary-encoders/commutation/amt33-series