Prototypes and work in progress

Sculptors often produce a small model, called a maquette, of the final object. Kinetics is a bit different as you have to start with the means of controlling the thing.

For the latest project the first step was deciding how many motors I would need. I decided on four and built (or more exactly lashed together) a simple device that has two Arduino boards, four drivers and four motors. Here is is:

This is the front view showing the motors and pulleys that pull and release fishing lines that attach to the maquette.


Here we see two Arduino boards mounted on top, one is a Duemilanove and the other a Uno. Each board controls two motors. I marked the USB sockets with the COM port number that the computer uses to program them.


Underneath are the four driver boards and motors.


This worked well, changing the shape of a simple cube made of drinking straws. However I began to have doubts about the lifting power of the motors. I tested it and found that each could only lift about 1N, about 100g. I want to use dowel or tubes made of aluminium or perspex. A cube of one metre side would require each motor to lift at least 200g. I decided to move to NEMA motors.

The NEMA17 size seemed to be the best compromise. It is fairly small but produces plenty of force and is quite cheap at about £10.

The decision about controllers and drivers was more difficult, mostly through ignorance. I started with the 2M542 controller. It is very sophisticated but rather large and expensive. I got the setup to work with the code I had already written. This is how it looks:

The blue and brown wires from the Arduino screwshield carry the DIR (direction) and PLS (step pulse) signals to the controller. The three core flex carries the 35 volts from the power supply.



Now I am experimenting with a four-axis controller which includes four slightly less flexible versions of the 2M542. I did debate going back to four simple drivers and coding to output codes rather than pulses. In the end the £60 cost of the four-axis seemed the best option. It does have an additional output to drive a motor at varying speed in case I ever find a use for it.

Here is the complete system of motors, controller, power supplies and Arduinos undergoing a long-term soak test. The system must work continuously for years. All components are used at well below their maximum ratings. Now I have to fit it all into two boxes that are as small as possible.



The lifting tests have proven very satisfactory. I set the current at 0.7A, which is only 20% of the maximum for the motor. Microstepping was was set at 3200 per revolution, where the motor naturally does 200 full steps. It could then just lift 750g, or 7.5N, force using a pulley of 10mm radius. This gives a torque of 0.075 mN. I shall use smaller pulleys in the final work, so this is more than I need. In any case I can put the current up to a higher value if I need to.

I have gone back to the 2M542 controllers. I want to plan for bigger, and therefore heavier pieces, that might need more lifting power. Using four controllers makes the wiring more complicated but more flexible. I can add more powerful sizes of NEMA motors and use the very fine microstepping to give slower and smoother movement.

Compare this prototype with the above:



(C) Peter Scott 2012

Last edit 5 January 2013