I need some help identifying the source of my constant backlash issues.
I have had my UM1 for over 2 years now, in that time I have had persistent issues with backlash that results in shells that consistently don't touch infills. I have found various workarounds for this issue to generally get good results out of the machine (hotter extrusion, more material, overlap etc) but these techniques are fiddly and cause other problems like dimensional accuracy and I would like to be done with them.
After a few interactions on the mailing list, and being told it was probably my short belts, I took drastic measures to make sure that it wasn't: I removed them entirely and made my machine direct driven. 2 new shafts, 2 flexible couplers and 2 printed motor mounts, and I have a much more straightforward machine. Even after all of that, I am still to this day seeing backlash issues.
A clue I recently uncovered was that the X motor runs _much_ hotter than the Y motor. For example, 80minutes into the current print job, I am seeing X motor ~= 118F vs Y motor ~= 95F. What would cause one dimension to require significantly more energy to print? Was there a bad batch of UM1s that had axis alignment issues that could be the fundamental problem with this unit? Anyone have any similar experiences with backlash and direct drives?