Advanced techniques for fast offset cancellation limit amplifiers

Abstract

As the need for higher data rates to end users increases, new technologies are required. Passive optical networks promise a vast improvement over previous methods, but they require amplifiers which can quickly adapt to changing input power levels. This thesis proposes two methods for addressing this problem, which are implementable in pure CMOS. Advanced modeling techniques allow for improvements on previous analog designs, resulting in a 50x improvement in signal acquisition time to 20ns. A novel digital design reduces this acquisition time even further. Results were verified using behavioral and component-level simulators