Turning a Shop Vac into a Super-Roomba
HTML-код
- Опубликовано: 5 окт 2024
- I turned a Bauer 1952E-B6 shop vac into an autonomous robot vacuum. It's a 4 horsepower vacuum driven by 10" wheels to suck up heavy debris that a normal roomba can't handle. The drive motors are Cardone 40-299 automotive windshield wiper motors which turned out to be an awesome low budget motor for a robotics project like this. It's also built out of some old ebike parts and miscellaneous electronics I had laying around. It's powered by an arduino uno and a DF robot two channel brushed DC motor driver (DFR0601). I also used the arduino to read an old PS2 controller for manual control during testing and for transportation. The battery is a sketchy chinese ebike battery, I certainly would not let this thing run unattended. The mechanical bumper switch is somewhat novel, it's a very cheap solution made with a coat hanger, fishing line, and an off the shelf limit switch.
Code here:
github.com/Cla...
Nice project!
The bumper is very clever! Using 2 switches, one on each side, should increase the reliability.
I'm wondering if it could possible to make such autonomous vacuum cleaner as a self-balancing robot, based on hoverboard...
Two switches would probably work better, but I already had the one and didn’t feel like buying another. Self balancing would be cool. It would be hard to keep the actual vacuum nozzle in the right position slightly off the ground. And it may struggle to drive over obstacles.
would love to get the Code and Cad!
Me too!
code used in this video uploaded here: github.com/ClayBuilds/MOAR
A little messy but it runs
@@greatwhiteslark github.com/ClayBuilds/MOAR
Is a board from a ready-made vacuum cleaner robot suitable for localization?
There’s no way a typical robovac board would drive these motors, way too much current draw. Plus the larger wheel diameter would change the linear speed. Perhaps it could be feasible to “intercept” the motor control signals on an original board and use them to inform logic on a separate microcontroller. Assuming the original is closed loop control, you’d need to get creative rigging up or simulating rotary encoders to feed a signal back. Honestly it’s probably easiest to just use a bespoke prototype board and integrate LiDAR or ultrasonic sonar to run SLAM.
@@Clay-Builds There was an idea to make control through a contactor (the wires that turn on the power part are used to close the contactor coils). Perhaps the navigation control does not work on the size of the wheel, namely on the laser rangefinder (the board supplies power to the wheel until it reaches the starting point), if this theory is correct, the implementation of this project is not so difficult