Introduction: Constructing a Robot Arm

Picture of Constructing a Robot Arm

Hi everybody

For my bachelors assignment for Industrial Design Engineering at the University of Twente, I looked at the construction of robotic arms. More specifically, I looked at the relation between components and their composition on desired performance and cost of pick&place robotic arms.

The paper is basically a large instruction set of information needed to build a robot arm, complete with reasoning to why certain design options are better than others.

To test all this information, I designed, build and programmed the robotic arm, and put it through some tests.
I attached my entire project, complete with solidworks&stl models of the arm.

For this paper I looked at
-how motors work
-comparison between stepper and servo motors
-which motor to choose when
-formulas for sizing the motor torque
-types of bearings
-what bearing is best suited for what condition
-formulas for bearing impact on accuracy
-possible power transmissions
-type of frame
-much more small stuff that is also important

Starting this project I didn't know the first thing about robotics, so don't be afraid to try. You will either build a succesful robot arm or learn a lot from the mistakes you made (or both in my case).

I hope this may help some of you in building your own robotic arms!
If you have any questions or remarks, feel free to post them below.

Step 1:


jbrauer (author)2016-10-28

I like your double sided bearing explanation, and the nearly art noveau style of the sheet metal. But the joint torque calculations assume that the rotating joints at the base and elbow would never be moving at the same time. Is the controller programmed to prevent that?

teun made it! (author)jbrauer2016-10-31

Thanks! I'm happy how the arm optimalisation pattern turned out to be.

Good question, it made my think back to when I was figuring out how to calculate the motor torque. I added an image to help with the following explaination,but in this preview window it looks like a black square... we'll see what happens when I hit the reply button.

Any position of the end effector, except for when both upper and lower arm are in line with each other, can be reached in two ways of positioning the arm. Either by both arms rotating counterclockwise, or one rotating counter and the other rotating regular clockwise (see image).
Often the movement can (and should) be done by rotating the motors opposite to each other.

When however this is not possible, the arms are by definition not fully stretched, so the effective length of the arm (x force) is reduced in calculating required torque. Adding to this, most torque is required when the load is accelerated, when the effective arm is the smallest. As the arm moves closer to where the effective arm length is the maximum arm length it will decelerate the load, adding negative torque to counter the inertia of the mass.

A programmer could program in that when the arms are close to fully stretched out, the acceleration is decreased or that the motors do not move at the same time.

In my mind this explaination makes sense, what do you think?


jbrauer (author)teun2016-10-31

Yup, there are multiple solutions to reach the same position of your end effector. But what happens if both motors are turning at the same time? The elbow motor creates a second order effect that amplifies the joint torque on your base motor beyond what would be applied by only rotating the base motor with the arm fully extended. If you didn't work that out in your kinematic equations, it might make sense to only move one motor at a time.

teun (author)jbrauer2016-11-03

If both arms would move at the same time in the same direction with the same angular acceleration relative to their own center of rotation, let's say 20deg/s^2, that would result in the end effector moving (relative to the base) at 40deg/s^2. I'm not sure, but can these accelerations be added up without creating difficult second order effects?
If so, than the 40deg/s^2 can just be included in the formula for sizing the motor.

However, it doesn't happen offten that the maximum load at maximum reach is accellerated by both motor in the same direction. With programming, the motors could be opperated so that when they are at e.g. 80% max reach, and both accelerating the in the same direction, only one motor is allowed to move at at time.

teun made it! (author)teun2016-10-31

This image should not be completely black... hopefully

Swansong (author)2016-10-28

It looks neat! I hope you get a good grade :)

About This Instructable




More by teun:Constructing a Robot Arm
Add instructable to: