In my opinion, the Cr-10 suffer of a bad project base about the x and y belts.
All the belts of the machine aren't working in horizontal and parallel to the trolleys movement planes, but are working with inclination.
The inclination generate an error in the positioning of the X trolley (that is not possible to fix completely with the e-step calibration, because is a variable error along the stroke).
from 0 to 150 mm of traveling, the error between 0 and 0.24 mm in the Trolley positioning.
From 150 to 300 mm of traveling, the error is between 0.24 and 0.28 mm (delta is 0.04 mm, due to the belts more horizontal in the second half of the stroke (please see the attached DWG for better explanations....sorry, but for me is a little complicate to explain it in english!).
For avoid what written above, I create a bracket with a double bearing wheel that allow the belt to work in horizontal position, obtaining a significant reduction of the machine traveling error, and making the x e-step calibration basically useless.
No further modifications are necessary for the installation of the system on the machine, only needed 2 bearings MR105 5x10x4mm or MF105ZZ and a screw m5 for fixing the bearings.
Update of 13/04/2019
Major update to match measures of the tensioner on the other end on the X bridge :
Replaced the Bearings with a Idler pulley GT2 16T as per the tensioner.
Needed a M3 screw L= 25 mm
Will also be the last revision for now.
Update of 30/08/2019
Added version with oversized fixing holes, due to a user request.
Holes are dia 6mm against the 5.5 of the "standard" version.
I didn't try this version by myself, so, beware!