dirac16
Member level 5
I am trying to use a couple of identical inverters in 0.18um TSMC technology. For the lowest propagation delay I ran a Monte Carlo simulation for a single inverter, where I got ~1ps for the std of the propagation delay, and roughly 13ps for the mean delay. Now if you compute 6*sigma value you get ~6ps variation in the delay, which is not allowed for my application. The best std I wish for is ~200fs. Is there anyway to alleviate PVT variations on the propagation delay?