Tips for resolving y-axis being off about 0.10mm

Hey just looking for the best way to resolve my y-axis printing about 0.10mm ish wider than the x-axis when printing a 20mm by 20mm calibration cube.

This is the cube I have been testing with:
https://www.thingiverse.com/thing:1278865

When I print it I usually see around z-axis=20.06mm x-axis=20.06mm y-axis=20.16mm… ish my digital caliper is cheap and has a little wiggle room but it shows a fairly consistent difference in the y-axis being a bit off.

Should I try to adjust the endstop switch on that axis so that when it homes it is -0.10mm less that it was before?

Changing the endstops will only move the area on the bed, not change the scale. To change the scale you need to change the number of steps. But first you need to know whether the difference is proportional or constant. If you print a 40mm cube and the difference is 0.2mm instead of 0.1mm, and a 60mm cube has a difference of 0.3mm, then it’s a scale error that you could scale out.

But honestly, 0.1mm is well below the nozzle size so it could also be affected by exactly how much filament is extruded or if there’s a little bump where the layer starts or stops.

Printing out 10x10, 20x20, 40x40, and 60x60 cubes as we speak.

If I find the same 0.10mm difference between all the cubes then would I look at the endstop?
If the difference scales depending on the size then I adjust the ysteps found under the configuration menu on the TAZ 6?

Okay had some time and printed all the cubes out and this was my results:

60x60mm
X= 60.25
Y= 60.30
Z= 60.02

40x40mm
X=40.25
Y=40.25
Z=39.98

20x20mm
X=20.20
Y=20.20
Z=20.02

10x10mm
X=20.20
Y=20.20
Z=20.03

I put down an rough average based off three or more measurements on each side. It looks like I really don’t see a difference in Y compared to X till I get to the 60x60mm cube. Z looks to be just about right no matter what size I print at but X and Y look to be off by .2mm or more on all the prints tested.

I also changed out my $10 Harbor Freight Pittsburgh digital caliper for a slightly more upscale $25 one made by iGaging to see if I can get more consistent measurements. I then tightened the belts for both X and Y so that I think helped get both axis a little more consistent.

Not sure what to do here, when I print benchies and other prints that test other aspects of the printer I get great results outside of dimensional accuracy. I know .2mm tolerance is fine for most prints but if I wanted to print single prints that involve interconnected moving parts like the NASA wrench I would assume I would what to try dial the accuracy in a little more right or am I just driving myself crazy?

Any other things I could try/test to see if I can get my X/Y closer to the accuracy of my Z axis?

So the offset seems approximately constant, which makes me think it’s extruding a wider bead, not a scaling issue. What if you print one of them and dial down the flow rate to 90% and see what happens?

Also, is the width constant with z? Cura’s default settings always overextrudes the base layer for me, which makes it “bulge” a bit at the bottom.