Same issue unfortunately. I guess it's because the two objects don't exactly double in area size and the Texel Density is calculated based on this? The Z is the same for both objects. Here's a test file to illustrate what I'm up against. I expect the big square UV in the second object to be double the size of the first.