[MOC] 6 Degree of Freedom Robotic Arm

[Glaysche]

2 hours ago, ord said:

@Glaysche Wow, I just read through this thread for the first time and inspected the studio model (thank you!) and this is quite the impressive machine you've made.

Thank you for the kind words.

2 hours ago, ord said:

Using differentials to decouple the joint movements is genius, and I imagine it saves a bit of pain when it comes to programming. What is the backlash like though, considering so many more gears are needed for it?

Backlash was probably the trickiest problem to solve in this whole project. The differentials definitely add some backlash but try to mitigate it with different techniques.  The five gear red differentials are definitely better so I started by using those.  Being 28 tooth gears made it a bit challenging to get the proper gear ratio to cancel out the rotation through the axes but was worth it.  The most effective technique is to gear up before going through the differential and gear down at the end as close to the load as possible. This makes higher speed and lower torque through the differentials.  This works really well with the gripper and sixth axis.  The fifth axis (bending of the wrist) still has too much backlash for my taste.  I couldn’t find a way to gear down enough at the end of that gear train without making the structure too big or too weak.

2 hours ago, ord said:

I read on Rebrickable that you're working on a new version - looking forward to it! Do you plan to omit the 3D printed part and make it pure Lego?

Maybe you were looking at my first version on Rebrickable?  My second version is already up at:

https://rebrickable.com/mocs/MOC-109607/glaysche/6-axis-robotic-arm-mk-2/
 

In that version I re-designed axes 1,2, and 3.

I don’t have plans to eliminate the 3D printed part.  If you look at my very first version I posted in this thread, it was pure Lego.  I had put the motors for the gripper and axes 5 and 6 out in the wrist.  This ended up being really heavy and stressed the axis 4 turntable too much.  Adding the 3D printed part and moving the motors to the elbow dramatically improved axis 4. I haven’t found a way to as good of results with pure Lego.  The 3D printed part is actually quite generic and useful for many designs.  Maybe someday TLG will produce something like it.

One idea that is plausible would but the use the new Spike Essential hub and small angular motors out in the wrist.  This might be light enough to not stress the turntable too much.  Or maybe TLG will come out with a more robust turntable?  The Spike Essential hub has the additional challenge that it only supports “streaming” mode and only the Spike Prime software.  I am currently using Mindstorms software and downloading to the hubs.

[ord]

Thank you for the detailed reply. It's very impressive that you managed to decouple the wrist joints AND the gripper.

12 hours ago, Glaysche said:

Maybe you were looking at my first version on Rebrickable?

Oh my mistake, I was. I saw a comment from ~2 months ago and didn't realise you'd already released a new version.

I see the usefulness of the 3D printed part for infinite gripper rotation and have also thought about mounting a Spike Essential hub and 2 small motors on the wrist of a robot. I think Pybricks might eventually support hub-to-hub communication which would be ideal for this - one Mindstorms/Prime hub on the base + one Essential hub on the wrist = 8 motors... perfect!

Anyway, thank you for sharing detailed updates over time. It's so nice to follow and see the robot get better with every post!

[Glaysche]

2 hours ago, ord said:

I see the usefulness of the 3D printed part for infinite gripper rotation and have also thought about mounting a Spike Essential hub and 2 small motors on the wrist of a robot. I think Pybricks might eventually support hub-to-hub communication which would be ideal for this - one Mindstorms/Prime hub on the base + one Essential hub on the wrist = 8 motors... perfect!

I also have 4 ports used for sensors that are only used during calibration.  Well, actually, the force sensor is also used as a kill switch.  If I make a mistake in the programming, I can hit the kill switch before it damages itself.  The axis 2 in particular is geared such that it could break plastic parts if it runs against the end of travel.  If I eliminated the 3D part, I’d probably end up with two 6 port hubs and one 2 port hub for the full robot.  The 3D printed part is cheaper but it’s not pure Lego.

in fact, with reasonable software, I’d probably use one of two Essentials hubs on the 6 axis remote control as well.  That would make it easier to add gripper control to the remote.

2 hours ago, ord said:

Anyway, thank you for sharing detailed updates over time. It's so nice to follow and see the robot get better with every post!

Thank you.  I’m never sure if people are getting much value out of these posts and my Rebrickable MOCs.  They are complicated and expensive enough that I doubt anyone else will ever build them.  I’m hoping that by providing Studio models, people will be able to borrow some of my ideas and use them in their own MOCs.

[ord]

23 minutes ago, Glaysche said:

I also have 4 ports used for sensors that are only used during calibration.

That makes sense. I'm making my own 6-axis robot and was hoping to just rely on the smart motors to detect end points, like I did with my plotter. Of course, the 3 infinite joints would need some kind of (temporary) stops for calibration.

27 minutes ago, Glaysche said:

I can hit the kill switch before it damages itself

This is a good addition and I wish I had room for one (am using all 6 ports on the hub).

30 minutes ago, Glaysche said:

I’d probably use one of two Essentials hubs on the 6 axis remote control as well.

By the way, this remote is really cool too.

38 minutes ago, Glaysche said:

Thank you.  I’m never sure if people are getting much value out of these posts and my Rebrickable MOCs.  They are complicated and expensive enough that I doubt anyone else will ever build them.  I’m hoping that by providing Studio models, people will be able to borrow some of my ideas and use them in their own MOCs. 

Honestly, I'd glanced over this topic before but am only really appreciating it now that I'm building my own robot. I'm 50/50 whether I'll use your differential method or program out the 'coupling' on my robot. Either way, it's an innovative method and you've documented it (and the rest of the robot) well for people to look at, so I think there are at least a few people out of the 4000 views of this topic (plus those on Rebrickable) who have gotten value from that .