Hi Team
I have attached 3 pictures concerning an issue of a spectre transient simulation. The first picture is what my test circuit looks like, which is in differential configuration. Each have 2 tones and a DC value. The DC value are the same and 2 tones are 180 degree out of phase. 
I marked the differential input with RED. then there is a pair of DC blocking caps. The voltage after it is marked as GREEN.
The second picture showed that the the red plot is as what it should be: same DC and out of phase. But the circuit believe that two GREEN wires have different DC value initially and slowly get to its correct value. 
The third picture is the zoom in version of the second picture so it's clear the difference in DC valure
I believe the cause of such plot is due to the initial voltage of the differential path are different. (in my case are RED at the begining ~470mV and ~463.5mV). This makes the circuit after the caps believe the DC voltages are different. I am trying to get rid of the settling time.
I have tried set ic in the cap to be 0V , and the ADEL-simulation-convergence aids-initial condition. Non seems able to get rid of the initial settling time (in this case, more than 420 ns of useless simulation) I could do something wrong. But can anyone help?
Thanks
Allen