How small can I make my robot tree? ...stepper motor vs regular servo

exwhyzed · January 13, 2014, 6:19am

Hi All,

I came up with an idea to make a 40 servo robot tree (or seaweed) that can flail its trunk and branches, lol! The trunk will consist of 8 servos (red), and there will be various other servos scatterd around to provide branch and twig movement (blue)

This site and others have been very helpful just by looking at guides etc. It seems as if all i need to make this robot is servos, a frame, some type of board with the motion program, and power.

From browsing online, it looks like i have 2 choices for a servo, one with a built in potentiometer, and a stepper motor.

It looks like the first type is a bit more expensive, but there are boards out there can handle at least 24 of these .

pololu-mini-maestro-24-channel-usb-servo-controller-pololu.jpg

With regards to the stepper motor, I've only seen a board that can control 6 axis, or six servos.

1. How would I control 32 stepper motors if a board can only control 6? how would i get these boards connected and talk to each other?

It looks like stepper motors are cheaper, but the controller is more expensive. stepper motors can also come in really small sizes. But the super small motors will have very low torque.

So one of my questions is,

2. what is the best type of motor/servo to use to make this robot tree?

For the software, it looks lik Justin Dailey used the 3d animation software bledner 2.49, a 3d mockup of his robot arm, plus some code, to drive an actual robot arm. The best part about that setup, I wouldn't need to program each servo, I can just create a 3D rig, with some type of kinematic spline, and then position parts of the tree where I want to, then the rest of the servos will follow along. I can loop the program and then I'd be done.

3. Does anyone know of any other software where this type of animaiton style programming can be done? the maestro 24 channel board above comes with software that is similar, but I'd have to program each servo, and wouldn't get the fluid movements of a kinematically splined 3d rigged setup.

4. How would i get some joints to rotate 360 degree without the servo wires (ahead of the servo) getting all tangled up?! I guess I would need to use a Slip ring of some type.

300Rpm-12-5mm-6-Wires-6-Conductors-Capsule-font-b-Slip-b-font-font-b-Ring_jpg_220x220.jpg

Can i really slip the wires of 39 servos through a slip ring at the base of the tree?

5.as far as frames for the tree go, I was looking at servo blocks (by Robot Zone)

Are these a suitable choice for this tree application? Will these work with stepper motors? It doesn't look like it...these look to be speciialy built for specific servos. I don' t know how to attach a slip ring to this setup anyway and all help is much appreciated.

Thanks in advance for looking at this post and assisting me with my robotic tree!

lumi · January 13, 2014, 11:24am

I would go with servos. They

I would go with servos. They already have the driver board inside and you control them with only one pin. Also they come with a gearbox.

Ther are 360° servos as well. You just rotate the shaft and not the servo itself, so no problem with twisted wires.

Yep, we need a estimated size and choice of material. If you say the tree willbe 3 meter high then we have to start all over again by recommending a complete new hardware

franciscodr · January 13, 2014, 6:30pm

Take the servos

Hi exwhyzed!

Very nice idea! The robotic tree can be very nice, also if you decorate it properly and make it sensible to air movements

Simplicity: servos 1, steppers 0

From the motor side, I’ll use servos, as they are easy to program and wiring: the more actuators, the complex and more cabeling you’ll need. With the servos you just input the degree and that’s it. With the stepper you have to count steps plus the power driver itself…

Not need acurracy: servos 1, steppers 0

Also you don’t need a very fast nor acurrate control, do you? It’s not a 3D very precise control, so you don’t need a stepper motor.

Not need speed: servos 1, steppers 0

As the servo contains the gear, you sacrifice speed (however they are quite fast) for efficience: with less electrical power, you have same mechanical power than a stepper motor. It’s the same principle as the lever.

Servos is the answer by 3/0

Good luck and keep going!

exwhyzed · January 14, 2014, 3:30am

thanks for all of the great feedback everyone!

I’m in the process of incorporating all of the above ideas into a more detailed drawing. This will hopefully show me the dimensions that I’m looking for,…I really planned to reverse engineer the size from the top down. So i’ll start with small servos and frame parts on the branches, and move onto stronger (and larger) ones as I get closer to the base.

One thing I came across in my findings is this…in the Kuka KR5 robot (and I’m sure many other robots)…

...several of the robot arm servos are grouped at the rear, or far from the end of the tip of the arm. I guess this is to keep the weight off the tip, and to reduce any electro magnetic interference that the servos may cause at the tip, which could interfere with whatever the robot is being used for. The below snapshots are taken from the video Inside Axis 4, 5 & 6 of KUKA KR5 Robot

To get the movement to the joints near the tip, some complicated transmission is being used. it looks like one shaft is coming down the arm...that shaft can rotate the arm...but two gears with belts, that themselves take a twisting journey in the robots chambers to get where they are going, each provide another axis of movement for the arm. I'll do more research and figure out how this feat of mechanics and engineering is put together.

Is this type of thing also done in micro robotics, or just in big robots? I think that this is something I will incorporate into the tree, to give me exposure on how construct these complicated mechanics. Instead of being a robotic tree, the main trunk may be a small scale replica of a robotic arm, like the Kuka Kr5. Now I'm realizing that the size of the robot arm/tree trunk may be dependant on what size of belts are available. I'll have to reverse engineer the whole thing based on that part availability. It looks like the Kuka belts have some type of knobby grips on the inside, so I won't be able to use smooth belts which are widely available.

This way I can use the robot as an arm for different projects or program it to pick up things. But in tree mode, i'll attach small and light tree-branch modules. Now the tree will hopefully be more versatile, and offset the cost of those Dynamixel servos which hoLy mOLy are robustly priced! On ebay, I found another brand that is significantly cheaper, called Spring RC, but I'm sure there is a negative tradeoff going that route.

I'll post a new and more complex drawing as soon as I can get it done. Thanks!

ollyr · January 14, 2014, 12:45pm

The first thing that struk

The first thing that struk me is that you are housing the servos at the pivot! why not house the servos in the base, or bottom of the trunk, and have the movement actuated by cables. A bit like tendons in a hand?

this would also probably cut down how many servos you needed as a curve would be averaged out over a number of joints. Just have one servo pulling the middle of the trunk one way, and another pulling the tip the other. The rest would sort itself out…

Edit: also, cant you get wire which expands and contracts when voltage it put through it?, You could make use this for the upper branches?

Edit 2: I see Osipee beat me to to it!

http://www.jameco.com/Jameco/workshop/ProductNews/musclewire.html

exwhyzed · January 19, 2014, 10:37am

Thanks for all of the

Thanks for all of the feedback once again. Muscle wire, using cables, the pros and cons of servo vs steppers, making it sensible to air movements (is that possible?), and the 32 servo demo?!! All of these ideas have actually got me so much more interested in robotics than I thought would happen. I especially find robots used in manufacturing, and picking/placing very fascinating! It’s encouraged me to get some books on robots and learn about how they are used in the industrial world and for which purposes.

I think I really want to make a small scale replica of a pick and place delta style robot, or just this robot tree project with a gripper on one of the branches, and have it actually pick an place small items…but i won’t have a vision system, i’ll just tell the picker n placer which locations to pick and place from, and then cross my fingers that it doesn’t lose accuracy over time.

Anyhow, moving along to my current dilemma. it looks like the tree will be more like a robo-cactus.

If I use stepper motors, I will be able to get pick 'n' place precision. But I'd like to connect the stepper motors in this configuration so I can use the least amount of materials, and get cactus sections that can rotate independently (but not 360 degress since the wires will get tangled).

I guess i'll have to use lighter step motors as I go up the tree, so as not to strain the lowest motor. And I'm hoping this whole ensemble will be able to fit in a 1 ft x 1ft x 1ft area. If its taller that won't be a problem.

Is there a technical name for this type of motor configuraiton? I've looked online, and can't find the name...its not "2 motors on the same shaft" either.

I will probably have to learn the basics of CAD just to create a hub mount, and send the file somewhere to get a few CNC copies.

And Maxhirez, your comment has led me to believe that there are alternate "animation style" programming options? Please share the details:) I think an animaiton style motion control program might let me do the above "divide movement by 3" scenario.

Anyhow, take care everyone, and I look forward to your suggestions on how to get this cactus online!

exwhyzed · January 19, 2014, 5:09pm

Thanks for the links

Di Duane,

thanks for the link to the Service Droid Wrist…food for thought. And polymorph! I’ve never heard of a plastic that can be manipulated like that, what a great idea. The landing gear you made has clean edges, did you use a mold of some type?

exwhyzed · January 20, 2014, 4:51pm

Servos win! (sorry steppers)

After reviewing all of the above posts, I don’t know what I was thinking regarding steppers, Servos are the way to go!!! I don’t need precision for the tree, I just need the limbs to sway to and fro, and possibly respond to air movements.

The reason I was thinking about steppers was to make a half tree and half robot arm…what a bad idea. In trying to get the best of both worlds, i’d end up with a crappy tree a bulky robot arm. It would be best for me to keep this project strictly as a robot tree/seaweed piece of art, and have a second project to address the robic arm (or functional scale model of a SCARA robot).

I’ll create a new post regarding the robotic arm picker/placer.

I also found, through some old robot technology books, how the Kuka KR-5 arm mentioned above is configured, I’ll post pics as soon as I get those books out of my locker. Lets just say it involves lot of motors, gears and shafts, and possibly even a clutch mechanism like in a car. Kind of over my head for now, but not once I finish reading that chapter.