Did you take the cut width into account? If not, then the calibration will be incorrect because a “150 mm” square will be too small by the width of the cut, so adjusting the step/mm calibration to make it the correct size will make other sizes come out wrong.
A previous discussion explains the situation:
If the machine has mechanical backlash, then all bets on calibration are off. Scale this test pattern to fit the platform and run it as fast as it will go in Line mode with optimizations turned off and power set to mark cardboard:
Any distortion in the results on the cardboard indicates a mechanical problem.