Can somebody explain what is the reason for the channel densities being in powers of 2? Also, from the process point of view, is there any advantage/disadvantage in using a card with higher densities?

 

It is a question of how many points ( read equipment in the larger picture) is affected if a card fails. Traditionally it was 8 or 16 for inputs and 8 for outputs. Nowadays 16 is defacto for inoputs but outputs still remain 8 since most cards cant handle higher currents ( 8 means 4A at 0.5 A per channel)