6 DOF robotic arm by Jos

[Skookumjim]

a couple of quick images of the robot arm ...

https://www.dropbox.com/s/dcbc2n7crg0fk8q/Photo on 15-02-2021 at 20.29.jpg?dl=0

https://www.dropbox.com/s/kjjhqybhrnw5a6x/Photo on 15-02-2021 at 20.20 %232.jpg?dl=0

 

 

[Mr Jos]

V2 of the 6DoF. Now more power and teaching with PS4 controller.

[Mr Jos]

After very long programming, thinking and little building the instructions are now available incl program. Please let me know if you try it out and find any problems, it's the first time I upload a MOC.

https://rebrickable.com/mocs/MOC-76666/Mr_Jos/6-degrees-of-freedom-6dof-fully-automatic-robotic-arm-with-inverse-kinematics-programmed/#details

[whitepen]

Sadly I do not have all the pieces to build it, but great MOC!

[Mr Jos]

I found a video of one of my MOC buyers, he added a pneumatic end piece. Hope to see it working soon (or any of the other people that bought the instructions as more people then expected have 2 or more EV3's and like to recreate this)

 

[Mr Jos]

Started with this project again, trying to make it more long range.

Reinforced axis 2, axis 4 and axis 5 axles, also slowed them + axis 3 down for more torque.

Apart from axis 2 all work fine. I will need to have 8gears holding that joint in place as the weight is to much to handle, and it rips apart when it comes from full down.

Would prefer to be able to solve it with some counterbalance system. Will play around with some setups.

[Glaysche]

21 hours ago, Mr Jos said:

Apart from axis 2 all work fine. I will need to have 8gears holding that joint in place as the weight is to much to handle, and it rips apart when it comes from full down.

Would prefer to be able to solve it with some counterbalance system. Will play around with some setups.

Yeah, this was the trickiest axis for me as well.  For quite a while, I used a spring balance to try to help the motor come up from the extreme positions.  It never really worked that well because it needed very stiff springs to have any effect at all.  Once I used stiff springs, friction and bending of the parts became a big problem.

I ended up solving with brute force.  The axis now uses two 60t turntables driven by 4 12t gears each.  There’s a big gear train powered by two motors hard coupled together. It still creaks when extended to extreme positions but seems reliable.

I’ve been thinking about using the new 1:5 scale extra long motorcycle front shocks which are longer and have more travel than any other shock.  You could put these farther out on the arm to get better leverage and lower stress.  Every design I’ve come up, however, with is super bulky so I haven’t pursued it further.  Plus, my B&P order with more of them hasn’t arrived yet.

[Mr Jos]

3 hours ago, Glaysche said:

Yeah, this was the trickiest axis for me as well.  For quite a while, I used a spring balance to try to help the motor come up from the extreme positions.  It never really worked that well because it needed very stiff springs to have any effect at all.  Once I used stiff springs, friction and bending of the parts became a big problem.

I ended up solving with brute force.  The axis now uses two 60t turntables driven by 4 12t gears each.

Same, the gearing I have makes it no problem for the motor, but the bending of the 4axles with 12T gears makes it skip. I already added a working system that allows it to go backwards 15° and forward around 60° without problems, but going the last part to 85° makes it now unable to go up without gearskipping. I'm even afraid having 4gears around instead of 2 will remain having this problem.

Now I do use a counterbalance system using the 2 EV3 bricks as weight to pull it back up, will have to see how I can improve it for more lowering to atleast 80°

Not able to post pictures from phone, will do later.

[Mr Jos]

I have been able to finish the mechanical design of the XL version of my 6DoF.

1st test with counterweight being pulled upwards with chains:

This had only 2 gears under each turntable of joint #2, and worked from 60° forwards max.

 

I modified the counterweight to go backwards (and up) when the 2nd axle bends forwards. It also is loaded by 2 hard long springs to give extra pressure for going back up together with the 2 bricks weight pulling it up. Even when I disconnect this counterweight the 2nd axle works good. No gear skipping going up from 90° forward!

I did add more gears on the turntables (total 4 on each now) as this happened with only 2 each:

Now they don't bend anymore.

Cabling routed so they allow bending at all joints they pass + going to the bricks that move:

Full 90° forwards is no problem for this arm (still need to test max weight in this + less far positions):

 

And the last part, the new homing + small demo moving linear and short roll/pitch/yaw testdemo!:

I only had to modify the gearing variables + homing angle values and it works like this already. Need to do some finetuning to get it working even better and then let it work with the automatic warehouse.

Edited February 2, 2022 by Mr Jos
Wrong photo added

[Mr Jos]

To create stability I had added some ugly bricks on the baseplates, I wanted to eliminate them and have a smaller footprint.

I created this block;

When it's connected to the baseplate, they get only lifted of the table by a very little bit. And it looks way better, more like a concrete base having some steel beams sticking up.

Also been thinking about adding a 7th movement, as there is no space for mechanical axis through the wrist anymore, I came up with this pneumatic hand. I will need some more pneumatic tubes to be able to connect it all the way to the bricks (there I want to place the pump+valve). Will receive 10meter end of next week, I'm using Stihl hose.

For now will continue with the programming to use the fork at the warehouse scissor table.

[Glaysche]

@Mr Jos This inspired me to experiment with using the new large BMW motorcycle shock absorbers as a spring balance on my robotic arm.  This enabled me to change the gearing to be faster.  I think it works much better now.

Here's a 27 second video showing before and after adding the spring balance.

More details here.

I like this solution because it is symmetric.  The shock absorbers push the arm back to center when it goes both forward and backwards.  It is also reasonably compact.  Finally, when using PoweredUp / 51515, you are severely limited by the wire length for the motors and sensors so I couldn't use the hub as a counterweight.  In fact, the exact locations of the hub are pretty much dictated by the lengths of the wires.

[Mr Jos]

1 hour ago, Glaysche said:

@Mr Jos This inspired me to experiment with using the new large BMW motorcycle shock absorbers as a spring balance on my robotic arm.  This enabled me to change the gearing to be faster.  I think it works much better now.

...

I like this solution because it is symmetric.  The shock absorbers push the arm back to center when it goes both forward and backwards.  It is also reasonably compact.  Finally, when using PoweredUp / 51515, you are severely limited by the wire length for the motors and sensors so I couldn't use the hub as a counterweight.  In fact, the exact locations of the hub are pretty much dictated by the lengths of the wires.

The problem with springs is the limited range they have, the BMW does offer a longer + stronger force already, but still only 3studs length, and this limits the angle. If you want to let it go forward 90° like mine can, then the placement + force from these springs is difficult to use.

For my end-effector, I didn't test yet, as I need to have my complete logistics center ready for an event next month, and need to change a few parameters to get it running reliable for 10hours a day.

I'm wondering how the XL 6DoF will be performing and how gears/axles will be looking like after a weekend.

[Glaysche]

7 hours ago, Mr Jos said:

The problem with springs is the limited range they have, the BMW does offer a longer + stronger force already, but still only 3studs length, and this limits the angle. If you want to let it go forward 90° like mine can, then the placement + force from these springs is difficult to use.

You definitely can come up with a geometry with the BMW shocks that would allow a full 90 degrees of travel. It would be hard to predict how effective it would be without trying it. It takes a good amount of engineering to make the anchor points strong enough so they don’t flex too much which takes a lot of work. It looks like the counterweight system you have works pretty well so there’s not a lot of motivation to change it. I am really pleased with how well it worked on mine, however. It’s cool to see the different ways this problem can be solved.