Accuracy of Glitches Observed in Transient Simulation

Hi,

I am performing a transient simulation on standard cell based design in transistor level using spectre (IC6.1.7). At the moment I use APS with ++aps, liberal settings without post Layout-Optimizations in ADE. When I run the simulation for longer time periods (i.e. 3us), I see glitches at certain points and I do not see them when running it for shorter periods (i.e. 100n).

The observation is same for the conservative mode. Is it related to the simulation step size (this is default settings though)? Is there anyway to decouple the simulation time from the accuracy ?

Thanks & Regards

Anuradha