>>2938234
>It sounds like the issue being talked about is loss of precision because of thermal expansion in linear rails, and/or the aluminum extrusions they are bolted to, right? Basically if some parts get longer the whole structure changes shape and isn't as precise as a result.
Aluminum (what the structural extrusions the printer is made from) and steel (which the linear rails are made from) expand/shrink different amounts from heat (aluminum about twice as much). Therefore, when you have a steel rail bolted to an aluminum extrusion (common on printers), and it's straight at room temperature, increasing the temperature will cause the aluminum to expand more than the steel, making the assembly bend away from the rail side. On, say, an X gantry, a top-mounted rail causes a low spot in the middle, which can manifest over the course of the first layer as the gantry continues heating up after the mesh bed leveling. The distortion is well-defined and repeatable for a given set of parameters, and it leaves the printer as precise as it was before (unless the distortion causes the motion system to bind somewhere), but it produces inaccuracy.

>>2938266
Depends on the modeling tools you have available. You might sweep a profile along a spiral path. You might loft several suitably-arranged cross sections. You might generate a screw that's about right, then extrude a new object from the spiral surface. You might model a small cuboid, distort that into a single segment of the spiral you want, then stack copies until you have your ramp.