ROS Resources: Documentation | Support | Discussion Forum | Service Status | Q&A

Building a better TurtleBot3

Hi everyone-

I’ve written a blog post about TurtleBot3 assembly and component placement. If you’re just building your TurtleBot3 take a look and let me know what are some of your own suggestions related to component placement.

Thanks in advance


Hi Dragan,

I’ve been setting up a TB3 burger this week with ROS2. I was wondering if you would want to chat for a minute about how the experience has been for you.


Hi @Katherine_Scott,

can you jump on IRC on freenode? A bunch of us from Canonical are in #ros


Hi @dragan-s
This is a very impressive article. We seem to be able to use ROS more easily in a variety of ways with a little more thought. :slight_smile: Thank you very much and I will consider how to make it easier for users.

Glad more people are using tmux/screen for their bringup sequences :wink: definitely helps.

Hi @Pyo-

thank you and feel free to ping me anytime you want to bounce ideas around.


Hi @smac-

yes it most definitely helps and the tmuxinator scripting is a great addition. There a couple of tmux extensions that also help with session save and restore, so if you modified a session you don’t lose that when you come back… Definitely a great tool.

For interested parties search for tmux-resurrect or tmux-continuum plugins on github

1 Like

A few things that I found missing is about upgrading the default computer which is raspberry pi to a more powerful one such as Intel NUC or NVIDIA Jetson TX. In order to do processing from depth and tracking cameras.

I have myself tried to interface NUC directly via USB to OpenCR. But, it only works after I re-upload the Arduino firmware from the master branch of OpenCR.

But, it behaves slightly different than the latest binary release for raspberry pi. Also, each NUC and OpenCR requires its own power in order to move the dynamixel motors. As the NUC doesn’t seem to supply enough power via the USB.