Re: Idea: A RepRap Hanging From the Ceiling

The "Finding a reasonable home position"-part turns out to be really easy. I will explain it in the 2d-case first, to avoid unnecessary typing.

The only thing we control in this system, is the length of the three wires. Every possible combination of wire-lengths, (l₀, l₁, l₂), corresponds to a unique point in the avilable print area (a triangle). To map a point (x', y') to the corresponding wire-lengths, all we have to do is place the wires anchor-points in the cartesian coordinate system, and compute the lengths from each anchor point to the point (x', y'). Like this:


When we get 4 anchor points, we get a tetrahedron shaped print volume instead of a triangular print area. The computational work is still only 3 subtractions, 3 squares, 2 additions and 1 square root per wire. I don't expect the atmega to have any troubles with that.

This way of mapping (x',y',z') to (l₀, l₁, l₂, l₃) imposes no constraints on where to put anchor points or home position, which is really nice. Keeping anchor point coordinates (x_a0, y_a0, z_a0), (x_a1, y_a1, z_a1), (x_a2, y_a2, z_a2) and (x_a3, y_a3, z_a3) stored away somewhere will (hopefully) be all the configuration the user will have to do. I guess getting the 3d coordinates of the anchor point in the ceiling won't be trivial to the user, but it shouldn't be horribly impossible either.