| In this dissertation the problem of amplifier gain imbalance is investigated, along with several proposed solutions to this problem. First, a simplified general analysis for EDFA's is presented which substantially reduces computation time and software complexity. This analysis can be used in both the small- and large-signal operating regimes and greatly facilitates modeling of systems employing EDFA's. Two general approaches to gain equalization are then studied: end-to-end equalization versus individual (ie., optical filtering) equalization. A point-to-point multichannel lightwave transmission system is simulated and the trade-offs associated with each equalization approach are presented along with specific recommendations for system application.;Next a general multichannel analysis is developed and used to quantify the impact of amplifier gain imbalance and ASE noise in EDFA chains. It is shown that for many practical operating conditions, amplifier gain imbalance can limit the maximum transmission distance. This analysis is then used to investigate, the end-to-end (or pre-emphasis) equalization scheme. Within this investigation, the effectiveness of equalizing the received power (i.e., power equalization) or received signal-to-noise ratio (i.e., SNR equalization) is considered. It is found that the transmission distance allowed by power equalization is limited by SNR, while the transmission distance allowed by SNR equalization is limited by either residual power imbalance or transmitter dynamic range. These two schemes are compared by considering the maximum number of amplifiers allowed by each: the relative performance is studied as a function of amplifier gain imbalance, bit rate, number of channels, transmitter dynamic range, and receiver dynamic range. It is shown that for future ultra-high capacity lightwave systems employing preamplified PIN/FET receivers. SNR equalization is the better choice. However, there are certain situations where power equalization can be effective. Analytic results are in good agreement with results obtained by simulation. |
