Lawnmower project - photos and code

32 views
Skip to first unread message

Sergei Grichine

unread,
Aug 20, 2020, 2:12:20 AM8/20/20
to hbrob...@googlegroups.com
Hi All,

Thanks for giving me an opportunity to talk about my project at ROS SIG meeting tonight.


H-Bridge used in servos:  https://smile.amazon.com/gp/product/B07XXST7K9/  

I will be posting code at GitHub as I go, and generally keep RSSC.org updated through its mailing thread.

Let me know here if you have any questions, I will be happy to respond.

Best Regards,
-- Sergei

carl....@gmail.com

unread,
Aug 20, 2020, 2:20:55 AM8/20/20
to HomeBrew Robotics Club
Very cool & impressive. It seems we are thinking along similar lines for the vision & GPS, although your mechatronics are way more impressive than mine. You have done some great work there. Those servos are absolutely brilliant. I am looking forward to seeing this develop - perhaps we can share code as we get further in to it - even though the platforms are on different scales, the rest is basically the same.

carl....@gmail.com

unread,
Aug 20, 2020, 2:25:24 AM8/20/20
to HomeBrew Robotics Club
....and if anything goes wrong during development, you can always sell the idea to Hollywood as a plot for a horror movie - perhaps "The Alabama Lawnmower Massacre" :)

Alan Federman

unread,
Aug 20, 2020, 11:52:46 AM8/20/20
to vital...@gmail.com, hbrob...@googlegroups.com
You should look into linear actuators instead of servos to control manual levers. They are relatively inexpensive, and only require a 12 volt supply, a relay capable of  reversing polarity, and a potentiometer to control absolute position.

It would require a lot less capital and programing overhead to get working. Linear actuators for your project would be less then $50 each. A linear actuator and an RC electric car throttle controller, snd you in business.

Alan Federman, from my phone.

On Aug 19, 2020 11:12 PM, Sergei Grichine <vital...@gmail.com> wrote:
Hi All,

Thanks for giving me an opportunity to talk about my project at ROS SIG meeting tonight.


H-Bridge used in servos:  https://smile.amazon.com/gp/product/B07XXST7K9/  

I will be posting code at GitHub as I go, and generally keep RSSC.org updated through its mailing thread.

Let me know here if you have any questions, I will be happy to respond.

Best Regards,
-- Sergei

--
You received this message because you are subscribed to the Google Groups "HomeBrew Robotics Club" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hbrobotics+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/hbrobotics/CA%2BKVXVPWv8cn4zpqNoNcsjO1vV6sEj8WHK05nhx1wWqniQSyBQ%40mail.gmail.com.

Chris Albertson

unread,
Aug 20, 2020, 1:01:44 PM8/20/20
to hbrob...@googlegroups.com
I would argue that the linear actuators are roughly equivalent to his servos.   They do the same thing.   If I were to change something I'd lose the PWM interface and use a serial interface where I send the desired set point in radians.      In other words, I'd make the servo work like a Dynamixal servo rather than hobby servo.

You lose precision in converting to and from PWM. Also, it is very hard to control velocity and acceleration with PWM.  But simple to do this over a serial link.

That said, non of this matters to much.  The hard part is autonomous navigation and safety controls.  You want to at least reduce the number of cats and small children that are run over by the rotating blades.  That and the machine still has to know how to drive home even after a wheel hits a rabbit hole and spins the mower 90 degrees.  In the end any mechanics that works is good enough




--

Chris Albertson
Redondo Beach, California

carl....@gmail.com

unread,
Aug 20, 2020, 1:44:46 PM8/20/20
to HomeBrew Robotics Club
Looking at  the video of it driving around, I don't think he should change anything with the servos and control system.

Sergei Grichine

unread,
Aug 20, 2020, 2:22:47 PM8/20/20
to hbrob...@googlegroups.com
Hi All,

thanks for chiming in and for your kind words, I appreciate your comments and your interest in the project.

Here are my thoughts on the servos discussion in this thread:

 - Cheap (~$120) linear servos are slow and generally underpowered. If they have control circuitry, that would be closed source. Alan - I looked and haven't found anything in that space. Let me know if you have something that fits.

 - A typical large hobby servo would have something like "Operating Speed: 0.13sec / 60 degrees (6.0V no load)" in the specs - fast enough to control an airplane model. My servos are fast - about 0.5 second for full stroke. I am in complete control of torque - by choosing the motor, gearbox and pulley dimensions.

- Movement precision in my case isn't a major requirement (as Chris noticed), although my servos are doing fine in that department. R/C PWM so far fits fine (1000us range with 10us noise is still a 1% precision, and RPi Hat or Teensy Pulse Position Library do much better). If my servo doesn't reach its setpoint in 3 seconds, I cut off motor power. This trick sacrifices SOME precision, but saves motors and H-Bridges. No matter how I torture the servos, they run cold.

 - My control circuit is a typical Arduino thing that has many connectivity options (R/C PWM, I2C, Serial) and it is completely open for tinkering. Whole servo is $25 (motor) + $10 (BT-2 H-Bridge) + $3 (Arduino Nano) + $3 (sensor). Note the magnetic rotation sensor for feedback - its stability and precision is amazing, I won't use POTs ever again. Also, those automotive motors are made to last, and under normal load are quite reliable.

Overall, I highly recommend playing with the concept of self-made servos. It isn't hard, and it opens many possibilities. There are now brushless motor based servos, they are the future for robots.

Carl:  our projects are indeed similar, although at the moment I put a lot of emphasis on PH4 autopilot, running on RPi 3 with a custom hat (Navio2 inspired). You can see it in the robot "skull" in the photos. I gave a 30-minute presentation at the RSSC.org meeting about my plans and the basic use of the autopilot and QGroundStation - https://youtu.be/UAVATDUWX7Y?list=PLUEDYYT2aLNMpZigEr-VQspI149PWWbnc&t=3885    Most of what we need can be achieved by autopilot alone, but you know how much fun it is to tinker with Jetson Nano and its two eyeballs. And yes, Hollywood scenario is always on my mind, and I already call my lawnmower "Crazed Rhino", although it doesn't deserve this and is so far well tamed. :-) 

Anyway, thanks once again - and keep making good robots!

Best Regards,
-- Sergei


Reply all
Reply to author
Forward
0 new messages