As long as your comparator always regenerates from meta-stable point, the time for comparator to solve will 100% depend on CL/gm. However, if you are talking about the logic delay outside the comparator core. It will be another story. If your comparator needs recovery time or your comparator is like OPA-based structure, then please think about that is not the right approach for high speed design.
As long as your comparator always regenerates from meta-stable point, the time for comparator to solve will 100% depend on CL/gm. However, if you are talking about the logic delay outside the comparat ...
yes,you are right.i have solven this problem by taking the postive feedback structure .