I am converting a 2m DTM to contours, and would like to have the smallest interval possible.  Is there a theoretical limit, or even a rule of thumb for defining contour intervals based on the resolution of the source data?