garageFab.cc

One problem affecting the accuracy in Contraptor is angular backlash in the motor coupling coming from the rubber spider in between two coupling jaws. While it's not noticeable in leadscrew drive, the error can reach 1/16" in the belt drive because of the low transmission ratio, especially if the stage is heavy. Needless to say, this is way too much. There are several ways to deal with this problem. First, the pulley can be mounted directly on the motor shaft, which conveniently is 1¼" long in Alltronics steppers.

The assembly in the pictures is the Z axis for the contraption aimed to drill bunch of small holes through the cutting mat (the end product being the vacuum table for the paper prototyper). The range of motion is only 0.5", but that's enough for the task. Because there are lots of holes to drill (256 holes per square inch), it has to be as fast as possible, at least faster than 20 IPM that the leadscrew drive can provide. It also needs to be light because Y axis (on which Z stage moves) is belt driven as well. Pulley on the motor shaft eliminates the need for mounting assembly - i.e. 1/4" screw with nuts, shaft mounts, angles and screws to hold it all together, making entire Z stage lighter. While the space is somewhat tight, there is still clearance between the screw under the shaft and the pulley/belt - so long as the screw is in between the pulley flanges. The belt and pulley can be installed/removed by removing the outer screw attaching the motor to the motor mount.

In theory - I need to actually test it - this should also work on an axis with two parallel belts connected via one common shaft at the opposite end from the motor, provided that the belts do not slop and the moving stage is light enough. I'm planning to test this option on the Y axis of drill-the-cutting-mat contraption. 

If the stage is heavy, maybe a pair of motors could be used, one on each belt, effectively doubling the torque, though at the expense of extra motor and stepper driver board ($10+$26=$36). Again, this is in theory.

The option that falls in between the above two, costwise, is the Oldham coupling. The Oldham coupling consists of two hubs and slotted disk in between. It has zero backlash, while allowing some shaft misalignment. It's pricey - one coupling set costs $22 on SDP/SI. I ordered two versions for testing - 3/4"OD and 1/2"OD, with length of under 1" and over 1", respectively; and I'm planning to use it on the X axis of drill-the-cutting-mat contraption.