Problems continue and don't know what to do

[Doddy]

As way of explanation - it's not normal (never say never) to lose steps etc in the motor as it decelerates towards zero (before accelerating in the other direction) - which is what is happening at your 3/6/9/12 o'clock positions in the X/Y planes. But what is happening is that any slack / backlash between the screw and the nut (or in the motor coupling or mechanical interface to the spindle), at the point of reversing direction the screw has a limited travel (defined by backlash) before it engages on the contra-face of the thread of the nut to allow it to start accelerating in the opposite direction. Whilst the screw face is moving to interface with the thread of the nut, the spindle isn't moving and that would present a flat-spot at the point of change of direction (3/9 o'clock on X, 6/12 o'clock on Y). That's why we've quickly jumped on to that as a first port of call.

You can also lose steps on change of direction, but this would be limited to one per change of direction and unlikely to be substantially visible just scribing a circle.