Code128 is capable of encoding all 127 characters from 0~127 but Code128 only has 107 unique patterns.
Code128Auto automatically shifts from one character set to another if necessary.
Code128XS, S, M, T, XT corresponds to XSA, SA, MA, TA and XTA versions of the character set A, B, C. The pattern is the same so it is not necessary to make 3 non-human readable font sets.
The human-readable versions must have 3 sets since the human readable is different.
So, if you encode with Code128BMA, the counterpart non-human readable typeface will be Code128M. If you encode with Code128CMA, the non-human readable typeface is also Code128M.
Code128 is a relatively complicated symbology. More information can be found here:
http://www.morovia.com/education/symbology/code128.asp