Thanks Chris's great AR3 project, most of us have working and look fancy robot arm. However, I notice there is a lot of room to improve. Since AR2/3 builder a lot, I'm suggest building AR4 which can remain low cost but have more simplified/direct/tough design, and more 'industry grade'. Base on observation in this forum, I propose some of the enhancement (include my wish list):
Use non-contact limit switch to avoid robot arm fall and damage the limit switch
remove too much wiring, use CAN bus (or etc) to reduce wire into minimal (come with encoder)
more flexible robot training method (like hand hold)
Better emergency stop (currently emergency stop will stop power supply and collapse arm position, that look danger and will hurt the arm)
natively support ROS/MOVEIT
tough design, can support >=2kg (in my opinion if possible change the j5 belt to others mechanism)
more fancy and user friendly software design, now program the arm really less user friendly
more up to date computer vision function
Look/Feel compatible to all Mac/Win/Linux
More error feed back/sound feed back
Easier to integrate more sensor/buttons/switches
I'm not good in design hardware and 3D drawing, but I can help a lot in programming.
I still haven't debugged by AR3 yet but count me in for the AR4 effort. I can prototype anything on my CNC and help with the mechanical design and sims
count me in.
I m not good at electronic and programming (basic knowledge) but I m good at 3D modeling, drafting and GDT, manufacturability And precision machining.
cool, i feel we team up can create big thing
Hi Chris,
I suggest we create a new AR4 repository first and put some wish list and draft inside. It is better put under your github, if you feel interested can you create and invite us as member? If no next week I create new one and invite u guys.
i'd make solved the linear kinematics calculation and seems match with provided excel format. For circular movement may not so important at the moment. I will create 1 repository afterward since chris may not feel interested to create it.
Did anything ever come of this? I am defiantly interested in the idea, especially the part of making the wiring less complex.
I started a public git repo here: https://github.com/infinity2o/AR4/blob/main/README.md
I'm planning on building the AR3 as soon as I figure out how to order just AR3 parts.
sorry there is quite sometimes i no continue this topic. i still keen to make new robot arm with new design. currently im writing python web app for ar3 first and once it stable we can use for ar4. if you guys have initiative you guys may help do some homework and we combine here?
@Kstan Simit Thank you for your enthusiasm for an ar4 and your efforts to involve the community in developing requirements of what an ar4 would focus on improving!
Personally, I agree with each of your proposed enhancements. It's a great list! I would like to add:
Haptic feedback: in general, I'm talking about a process when robot arm collides with a person / object, it can 'feel' it and stop.
Limit switches always active.
Look at placing J2 and/or J3 motor in a manner that acts as a counter-weight. Like what https://www.hdrobotic.com have done.
Graceful fail: a pre-emergency-stop solution that stops the robot cold and holds it in-situ to avoid it flopping down and breaking itself.
The ar3 is amazing as it is - completely an incredible job @Chris Annin has done here! But things can always be made better, and now that he has built a fantastic community we can source their experiences and expertise and design THE BEST "democratic robot".
Thank you all for the comments and I agree with each of your suggestions. I am actively working on some cool software updates and working with RoboDK to build a post processor and driver as well as many of the suggestions you have made.
i still havent finished building it but i totally agree with the can bus and wiring
and obviously ROS support to do fancy stuff with it.
ill take a look at the repo but i still need to do more work on the control box.
love the bot btw
thnx chris !
i just signed up, but i have been following Chris for a long time now. i learned a lot here and want to return the favor. i have my company and my goal is to build industrial quality Cycloid reducers,which is used for industrial robotic arms.and i can produce smaller and cheaper versions of them for AR4 if needed.
Im all for it ! the only "backlash" we could encounter from the community ( pardon my humor ) is that i think that one of the main goals was to make the robot accesible with parts that could be sourced easily from anywhere.
Though all ideas to make the AR4 better and more precise are always welcome.
How cheap and small could you make these wonderfull reducers ?
Sorry for taking so long promise to be more active from now on soldering is not my forte.
Happy Automating !
size depands on total weight of robot and reduction rate we gonna choose. im in turkey so price would be nearly free for you guys 😀
