The detail that surprised me most is how small the error was and how much it mattered. Greylight May found the X axis averaged 149.216 mm and the Y axis 149.535 mm on a requested 150 mm square, a gap of less than a third of a millimetre, yet that tiny mismatch made layers stick out slightly when square components were rotated on the house plaques.
The method is straightforward and well documented. A 150 mm square is cut from thin plywood, the X and Y axes are labelled on the wood before removal, and a pencil grid is drawn across the piece with a ruler. Multiple measurements are taken along the grid lines with a 150 mm digital caliper, the readings are averaged on paper, and the requested and actual distances are entered in LightBurn version 1.2.01 under Edit, Machine Settings, Calibrate Axis before writing the changes to the machine controller. The enclosed cutter with its red-shrouded cutting head appears in the demonstration as well.
The results speak for themselves: after a second cut averaged 150.23 mm on X, the final square landed between 150.00 mm and 150.04 mm on both axes, with the two dimensions matching each other closely. What is still open is the human side, since truncated decimal inputs and calibrating the wrong axis both crept in along the way, and the fix came from re-averaging the measurements and entering full decimal values. In my view, showing those mistakes is the most useful part of the whole thing.
Project by Greylight May. Watch it on their YouTube channel: