This project is primarily concerned with making a relay-based controller for a RepRap, but the motors must also be considered, since the type of motor used determines the control signals required.

At the outset of this project I had intended to use stepper motors made from 3D printed parts and other materials. Googling '3D printed stepper motor' brings up a couple of motors constructed in a similar way.

Since I want to avoid using any more materials than necessary, I would like to avoid using permanent magnets. That means making a stepper motor using PLA, nails and copper wire. My guess is that making a stepper motor with a high enough torque using these materials would be challenging, and I suspect that the resulting motor would be very power hungry.

When reading up on existing RepRap mechanical arrangements, I came across this diagram of a mechanism for moving a bed in the Y-direction: http://forums.reprap.org/read.php?2,377858,601938#msg-601938 . By replacing the green fixed point on the right hand side with a second motor, the result is a mechanism that can move back and forth in the Y direction, but where each motor only needs to turn in one direction. This video illustrates the mechanism:

So there is no need to make bidirectional stepper motors. Since the motors only need to rotate in one direction, why not use gravity as the power source rather then electricity? All that is needed, then, is a weight attached to a spool on a shaft, with the rotation of the spool controlled by a stepping device. This video shows a prototype:

The control signals required for a bidirectional linear mechanism driven by two of these gravity powered unidirectional motors (GPUMs) are much simpler than those required for a conventional stepper motor. The simplest relay-based conventional stepper motor driver than I was able to come up with has 14 SPDT relays: This driver has direction and step inputs. A bidirectional mechanism driven by GPUMs requires only 1 SPDT relay to give equivalent functionality - the direction input causes the relay to direct the step input to one GPUM or the other.

N.B. In the video of the GPUM shown earlier, the shaft is powered by a falling weight. This is what I plan to use in all prototypes. The disadvantage of this, of course, is that eventually the weight will hit the floor. Looking further forward, the shaft could instead be powered by an on-demand waterwheel (i.e. a waterwheel where the water only flows when power is taken from the shaft).