There is an old rule in recording called the 3 to 1 rule and it means the distance between the source and first mic is 1 the next mic should be 3 times that distance if you want less phasing.
To me, this rule (guideline) would factor around the typical fundamental tones of what you are recording, with the fundamentals being much more narrow than the extent of harmonics. With a speaker, and each driver working over a defined range, pointing in a specific direction, with the listener assumed to be at tweeter level, the problem would be more bounded. Remember MTM falls apart in the vertical direction if you are too far off axis.