However, their algorithim is still a big buggy. The calibration routine seemed to work fine when the head was in the middle (no limit switches activated), but had some issues if it started with one of the X-axis switches acivated:
Since I didn't like the noises the machine was making while this happens, I decided to swap the direction of the Y axis. These are bipolar steppers being used, so you only need to swap the wires on one of the two coils to get the stepper to go in the other direction.
Here's a photo of my Y-axis connector, as it arrived:
Now the y axis moves in the opposite direction, and calibration routine works properly even when starting from the home position.