Rotor Arm, My current project

[Azraels]

I started this about two weeks ago. Just finished the first arm section today and tested it lifting @3lbs attached at the end of the 8.5" section. I tried 5lbs and it would hold it in place but could not pull it up. Was really hoping for a 5lb lift. I had the stepper speed set in arduino for 100. It is a Shinano Kenshi 8V 1.2A 1.8 step angle which I pulled from a trashed copy machine.
The base is an old brake rotor with a roller blade wheel set in the middle being used as the rotation bearing.
 
The arm is rotated with a cog belt running to a stepper which came from an old heavy duty dot matrix printer. I wasn't  sure of the V so I started it out at 5V and its running great.

The altoids joystick came from my first arm and was really easy just plugging it in to Arduino's analog pins for first run testing.

My little stepper driver circuit uses the 2 I/O pin from Arduino to the ULN2003. I then move that signal on to higher power transistors that are capable of driving these steppers.

The third stepper located on the backside of the turret will drive the second stage (elbow) by cog belts with the first motor to pulley being on the pivot shaft of the first joint and then transferring that up into the arm. Will be staring on that tomorrow.

I know eventually I will have to decide on a better micro-controller than the Uno if I want the arm to do all I want it to but I haven't decided on which one to go with though. Any suggestions and comments are appreciated.

[Azraels]

I forgot to add that all the aluminum was bought at a surplus store. I got 8 (10" x 5.5" x 1/8" ) plates for about $9.
All in, I will only have spent $60 to build this arm and that's including the Arduino. So after replacing the arduino with a BeagleBone or Axon I will probably go over $100. But that is still far less comparatively than what I see being spent on arms that can only lift a few hundred grams! 

[Bciaren]

Might I ask, how did you shape your arm the way you did? 

[MrWizard]

Great project...well thought out.

[Azraels]

Might I ask, how did you shape your arm the way you did?

All of the aluminum was cut using a jigsaw with just a wood cutting blade, and a hacksaw. I have tried the metal cutting blades in the jigsaw but they just don't seem to work well.

I'm bad about not designing things out first because I use mostly scrap and surplus parts so its usually touch and go just to see what is going to work right.

The first section of the arm, shoulder to elbow, I sized the gear at the shoulder end, and pulley at the elbow end then used one of those circle guides you can find at art stores to trace out circles about the size of the gear and pulley and spaced at the distance my cog belt would need to run. You can still see the patterns traced on the aluminum in some of the pictures.

Ive been a bit lazy the last few days but I almost have the elbow joint and cog belt arrangement done and hope to  pictures up tonight or tomorrow.

[Azraels]

Finished the internal belt system that drives the elbow joint. Had to come to a stop today because I'm running out of 6-32 screws which I have been using for most of the construction. As I move closer to the wrist I plan to change to 4-40.
Notice I had to use a set of cog pulleys between shoulder and elbow joint. This was unnecessary and only done because I had no cog belt long enough to reach between joints. Luckily I have plenty of belts at the size I used in case of breakage.

Two of the main things I try to shoot for are the ease of disassemble / reassemble and the strength of the project to not break from being reworked and tested. While I was testing the strength of the shoulder operation the 3 pound weight slipped off the desk jerking the arm back and forth  but the shoulder joint / stepper did not budge and held firm.

The linkage after the elbow has been cut short to allow for the addition of the wrist rotation joint.
Wrist rotation, wrist up/down and gripper left to go!

[Azraels]

Definitely learning some things with this. Since the last post I have worked on quite a few different things.

 The wrist rotation which I'm still not happy with. I will either have to strengthen the mount holding the motor or might change this to another stepper motor. It is currently being actuated by a motor and gearbox pulled from an R/C car  which turns a much larger gear. It has plenty of power but the plastic gear box gives when it turns anything heavy which can cause the gear to slip. The bearings for the larger gear/wrist rotation came from a vcr head and were mounted in plastic sleeves which were cut using a wood spade bit grounded (it was old and rusty anyway) down slightly smaller than the bearings.

The pcb for the stepper driver board. Note to self, always triple check connections in Eagle. After everything was soldered on I hooked up the motors and discovered more than one problem. I forgot to ground the ULN-2003's which was an easy fix with solder. Second was mixed up one connection on the MP4502 (these are great and easy to work with for driving higher power steppers!) That required a bit more innovative fix with some jumpers. But finally got it working and went back changed the problem in Eagle just in case I wanted to make another one. Mounted an arduino UNO and the stepper driver to back of turret.

Wrist up down after the wrist rotation.  Was mounting the Arduino the back of the turret today and had the arm raised in the air. I must have bumped and it slipped, smacking the exposed gear at the end of the wrist against the desk breaking the gear. May still be able to use but not sure.

[jwatte]

I love how you take low-budget, spare, and surplus parts, poke holes in them, and turn them into functioning machinery! It's inspirational.

[Azraels]

Spent the last day and a half working on a basic control GUI in Processing. What I wanted was to be able to hit a button on the screen with my mouse and it move the associated stepper while the mouse button was pressed and stop when mouse button released. Had some issues with the stepper stepping smoothly at first. It sounded pretty rough. I thought this may have been lag time in it reading the serial transmission. After a whole lot of debugging it finally came down to me misplacing a curly brace which solved the whole problem.  Its running smoothly now. I shooting for my final program to be something like flowstone.

During the testing of the program the cog pulley above the shoulder stepper locked onto the shaft for some reason. I wasn't paying attention and the stepper broke the belt in two.  I had left the motor in the hold position for quite a while and it was quite hot. Not sure if that may have caused some problem or not. May have to make some sort of  shield between it and the pulley.

Other than that also worked on securing the wrist rotation motor mount so that it does not give and added a potentiometer to the wrist rotation shaft for some position control. And began experimenting with different types of radio control so that arduino can control its turret rotation wirelessly.