Incidental layer shift

#1
gMax 1.5 XT with heated bed; Marlin 1.1.8
I have a layer shift problem that presents as a single Y offset of about 1mm. The shift seems to occur at random and has been observed with several different files, although each file will print with no problem typically. I was lucky to have been watching when one layer shift happened and observed that the table ran all the way to the back limit of travel in the middle of the print and then immediately return to the print location to continue at the offset position, so I assume that there was a skip on the belt during this event that resulted in a layer shift. The problem does seem to be happening with increased frequency now. So now I need to determine how this command to run to the extreme position is becoming injected into the code. Since it is not a file problem, I suspect that a code value is becoming corrupted at some place during downloading or during the execution. Since the error is the identical over multiple files, I'm wondering if the problem is some sort of numeric overflow, but I'm just throwing out wild guesses. Has anyone observed the same problem or have any thoughts on how to catch and correct this issue? Thanks for any assistance!
 
#2
I have had Y axis layer shifting problems on both gmaxes I own. In those cases, it was the driver voltage being set to low. I have also had Z axis problems related to a failing control box fan.

If you are printing from an SD card, you might try a new one. If you are printing from a PC, you might try an SD card or Octoprint and see if the problem persists.