Weird offset issues  

Weird offset issues

A friend of mine gave me his old and replaced MK3S. He never had issues - just replaced with a core one.

Happy me did the following:

1. MK3s+ upgrade

2. MMU3 upgrade

I now have very weird issues. On some areas of the heatbed the print is alsmost perfect. Other areas are just DAFUQ.

I created a small item in fusion to systematically check where the errors are. Where quadrant 1 is front right and 20 is back left. This is aligned with the PEI sheet quadrants.

Here you can see a DAFUQ image:

The first should be a nice cone on the top ... and its distorted to the left. The second  should be perfectly aligned and smooth surface with a nice round hole.

Some strings are just like somewhere in the air instead of where they belong to. For example here the bottom of quadrant 12 print.

With another real print i have the following result: It's just fine till the base plate is done ... then the horizontal slices are suddenly off left and right (or maybe front and back - that I don't recall)

With another different print of the same object I have this result:

I tried to reattach the heatbed on the Y-holder - which actually made the gradiant of the Z-Axis more stable - but still with 0.35mm distortion. I also bought a Satin print sheet as I thought that might be the issue. So what you can see here is that with a pretty regular pattern the print head moves a bit out of the original path of flight. This is printed with 0.2 Quality and manual adjustment of extruder speed to even lower speeds. So it's kind of doing this "on purpose" at the same position of a line that should be supposed long and straight it suddenly changes the path. Where the circular holes are supposed to be there is even missing like 3-5mm of filament.

I did one more thing now - I adjusted the pulley of the x-axis belt holder ... It may have been not perfectly aligned with the flat part of the motor axis. I just started a new print - this time with 0.15mm quality and regular speed. The print is much cleaner ... yet one can see that at the exact same area the quality is poorer. Not as bad as with the previous attempt. But also not perfect.

Does anyone has an idea what the heck the problem is? I'm on the newst firmware with a 0.4 nozzle. I obviously did XYZ calibration a couple of time. I also did the first layer calibration repeatedly. And I changed the bed level adjustment (don't recall what the exact name was) from 2x3 to 7x7.