[BUG] Systematic Z levelling problem

[smurfix]

Printer type - XL, multi-head

Printer firmware version - 6.0.0 RC1, but also happens with 5.x

Original or Custom firmware - Original

USB drive or USB/Octoprint - both

Describe the bug
My printer exhibits consistent levelling errors (Z offset too small) on the right side of the print area, but only if the print extends to the left side.

How to reproduce
I created this test print, consisting of a bunch of one-layer rectangles.

When I only print fields 3+6+9+12+14, fields 6 and 9 are basically perfect (left side).

When I print all fields, or only 7+8+9+13+14, fields 6 and 9 are overextruded (right side), obviously due to too little Z offset as determined by auto levelling. Both squares are field 6, center right.

This is 100% reproducible on my printer and does not depend on selected extruder, filament type, dirt on the bed or the nozzle (which IMHO could only result in underextrusion anyway), the relative position of right vs. left Z axis, or any other influence I could discover.

An overextruded right edge might not look like much, but here's a screenshot from a video of a larger print:

Observe the perfectly-flat left side vs. the obvious and pervasive overextrusion on the right side.

Expected behavior
Fields 6 and 9 are always perfectly levelled. My large print doesn't show overextrusion.

First Layer.3mf.zip

[smurfix]

Another test: after printing 1+4+6+9, field 6 is obviously over- and field 1 equally obviously underextruded.

[m-kozlowski]

This looks a lot like a miscalculated compensation problem due to PINDA sampling too close to the right edge of the bed mentioned in issue #3851

[smurfix]

Well except for (a) the XL doesn't have a PINDA and (b) the problem in 3851 is hardware and doesn't fix itself when you limit your print to the area affected by the problem.

I agree that the most likely cause is some sort of miscalculation, however I'm certainly not the only person who covers most of the build plate with a first layer, thus there's bound to be something about my printer that's different enough for the calculations to throw themselves off.

I have no idea what, though, given that prints are perfectly fine when you confine them to the right or the left side of the bed.

[m-kozlowski]

the XL doesn't have a PINDA

my bad. I havent noticed your printer model

the problem in 3851 is hardware and doesn't fix itself when you limit your print to the area affected by the problem.

Oh, but it does:)
Both restoring X_MIN_POS from einsy firmware (effectively centering the bed) or shrinking bed leveling area (MESH_MAX_X) helped in my case.
But i guess neither workaround will help in MK4 case :(