Looking at some red tag machines, we have noticed that there is a variance from parallel between the plane defined by the surface of the bed and the plane that contains the line defined by the x-axis after the automatic "level x axis" gcode is run. There have been some instances where shimming one side of the x-axis during the "level x axis" sequence will bring a machine into compliance with the skirt height requirements outlined in the hibiscus OHAI. The theory is that there is a correlation between the degree to which the x-axis is out of level, and the inconsistent skirt measurements. We would like to know if this is the case, and if so, we would like to know what the tolerance for parallelism between the bed and the x-axis is and how to measure and adjust it so that we can add it to the final assembly instructions and prevent non-conforming printers.
The auto-leveling system will compensate for this to some extent, but I suspect there is still a positive correlation between x-level relative to the bed and skirt height consistency. @marcio do you have an idea of how far out of level the x-axis can go before the G29 z-axis compensation can no longer adequately correct?