If you re-focused for a thicker item, that does not affect your scaling - If you move the scan head up 10mm to accommodate a 10mm high item, the projected size on the item will be identical to what it was before. It’s possible that you weren’t properly focused when you did your initial scaling.
The correction files from the CorFile2 utility load properly. If you have that tool, you can use that, or you can just adjust the scale factors for the two galvos to compensate for the scale difference.
If you draw a 50mm square and run it on black construction paper, then measure it, you’ll get a size. The ratio between the 50mm you asked for and the size you actually measured is how much to adjust your scale by.
For example, if you got a 55mm square, it’s 110% of the size it should have been. Take the current galvo scale values and divide them by 1.1 (110% as a scale is 1.1). The result is your new scale value.
The EZCad cor file format is not documented anywhere and has to be reverse engineered. We do not know how the older files work, or what values they contain - they’re not trivial to read, so I’m not expecting to support them any time soon, especially since the CorFile2 versions work as expected.