View Full Version : [Project] "atron" arm

04-08-2008, 01:22 PM
I've uploaded some images (http://forums.trossenrobotics.com/gallery/files/1/7/0/5/102_0269_img.jpg)of my "atron" robot arm (http://forums.trossenrobotics.com/gallery/files/1/7/0/5/102_0266_img.jpg) I started about 7 years ago. (They're probably too big so I won't include them in the text directly -- sorry.) The dimensions are based on my own arm, and it was intended to be operated using a sensor rig on same.

The circuit in the upper arm is a microcontroller programmed to send signals to the two hand servos and the stepper motor for forearm roll. It was supposed to receive signals on a serial link from a circuit that read voltages from the control potentiometer, but I could never make that work. Thus the mighty atron now gathers dust in the corner of my office.

If Phidgets ever makes a stepper motor controller (please, please, please!) I'll try to bring it back to life.

04-08-2008, 06:16 PM
Awesome! I hope the pictures are older ones. I noticed the 3 1/2 floppies and the laserdisc in the background. :tongue: old skool.
Very cool!

04-09-2008, 01:50 AM
Oh yes -- the photos are from Jan 2001. It's easy to forget how many robot resources didn't exist even a few years ago. We still have laserdiscs but we have been replacing them with DVDs. The 3 1/2 inch floppies however have gone the way of the 6502.

04-09-2008, 09:34 AM
That is one sic arm! Do you have any videos of that in action? I'd love to see this:D

If Phidgets ever makes a stepper motor controller (please, please, please!) I'll try to bring it back to life.

I've been waiting to see this for a looong time! That's the whole reason why I haven't really got into building with steppers. I'm not an EE guy, so all that lower level stuff makes my brain explode;)

04-12-2008, 12:07 PM
No videos unfortunately. I've only done small-scale movement tests, although they did reveal a tendency towards a vibration problem. I was going to work on that once I had the real control rig.

Stepper motor control interfaces would move Phidgets out of the realm of fun hacks and make them a serious prototyping tool. You can't do anything that requires precise motion control without them.

04-30-2008, 03:28 PM
Stepper motor control interfaces would move Phidgets out of the realm of fun hacks and make them a serious prototyping tool. You can't do anything that requires precise motion control without them.Ask, and you shall receive Metaform3d!

Announcing the PhidgetStepper Motor Controller:


ok, this time you really just got lucky that they were in the midst of releasing it:)

Regardless. I cannot wait to see what sort of incredible projects people build with these:D:D:D:D

04-30-2008, 07:56 PM

09-24-2008, 02:06 AM
The last few weeks I have spent doing a mechanical refit of the original atron. Here's a picture of the current arm:

http://forums.trossenrobotics.com/gallery/files/1/7/0/5/atronleft.jpgThe main change was replacing the forearm-roll stepper motor with a servo, which was attached with a standard servo shaft coupler and a custom bracket. The inset shows the motor that was replaced:

http://forums.trossenrobotics.com/gallery/files/1/7/0/5/atronright.jpgMechanical improvements mostly included tightening up some of the original components that were a bit too loose to expect good repeatability. The claw, for example, has a new more rigid pushrod replacing the old bendy wire strut (inset):

http://forums.trossenrobotics.com/gallery/files/1/7/0/5/hand_pushrod.jpgThe three degrees of freedom in the forearm are provided by servos, shown being tested here:

YouTube - atron forearm testThe upper arm (shoulder and elbow) has three more degrees of freedom given by heavily geared motors. Each motor has a corresponding pot to measure orientation. Here are the pots for elbow and shoulder X:

http://forums.trossenrobotics.com/gallery/files/1/7/0/5/other_pots.jpgHere's the pot for the shoulder Z axis motor, using gears made from laser cut acrylic. The gearing in each case tries to map the full range of motion of the joint into a single turn of the pot.

http://forums.trossenrobotics.com/gallery/files/1/7/0/5/shoulder_z-pot.jpgHere's a test of the upper arm. In this case the motors are being driven directly by RC signals:

YouTube - atron upper arm testThe control electronics are next. That will be made from an Arduino Diecimila with 6 PWM outputs and 3 analog inputs. It will be directed by serial commands.

At least that's the plan.

09-24-2008, 02:39 AM
I love how you shot the youtube video chromakeyed so you could shrink the background and make it look huge when it's really only about 5" at the shoulder! Nice key, btw, hardly even notice it. :tongue:

Seriously, could you build it a little smaller - that thing is kinda intimidating...

09-24-2008, 03:07 AM
Metaform3D: great stuff, good to see you have uploaded some video and pics of your project. Gotta love the new stepper controller just released, Ideal for your project meataform3d.

09-30-2008, 02:10 AM
After learning the joys, and tribulations, and joys, of the Arduino I finally have control over all 6 DOF of the atron arm. The three lower degrees are simply servos so it's a matter of sending target angles. The upper three are motors with pots, so I have developed my own very simplistic closed-loop servo feedback algorithm.

Here's a picture of the arm with complete wiring:

http://forums.trossenrobotics.com/gallery/files/1/7/0/5/cimg2046.jpgAnd here's a close-up of the control circuit. Warning: my solder joins are so ugly they can be harmful to look at!

The current system takes seek requests for any motor as serial input, so a pose is just a string. Here's a video of me copying strings describing several different poses for the arm and pasting them into an app that sends the text to the microcontroller.

YouTube - atron pose test
Not super exciting, but it represents a bit of a breakthrough for me. Next step will be improving the motor and servo command algorithms, not to optimize seeking, but to reduce transients that introduce shaking.

09-30-2008, 08:11 AM
Glad to see you got it working - nice job.