Best Calibration Sticks!

Firstly the problems with the 20mm cube test. The distance measured is somewhat low so small measurement inaccuracy is magnified. Secondly by only measuring the outside dimension of a part your result will be skewed by exactly how wide your extrusions actually are. 

This model fixes these issues very simply. Instead of measuring 20 mm we measure 100, and instead of only measuring an outside dimension we average an outside and an inside dimension to eliminate potential extrusion width error.

Your exact measurements don't matter, the only thing that matters is that the average of the inside and the outside measurements is as close to the target measurement as possible.

If you want you can also paint the seams on to avoid them messing up your measurements.