Capacity of broadcast channels in the near-future CATV architecture

Reports fundamental information theoretic results for near-future CATV downstream digital distribution networks. The CATV industry is evolving to this network architecture to quickly replace plant that would otherwise be left with serious reliability problems owing to deployment of cascades of electronic amplifiers. This new architecture is composed of a passive fiber trunk feeding COAX branches altogether serving about 200 homes. Frequency-division-multiplexed digital signals subcarrier intensity modulate a laser illuminating a single-mode fiber. Only one (electronic) amplifier is used and it is located at the COAX feed. Aside from additive white Gaussian noise (AWGN) from shot noise exacerbated by the amplifier, a significant impairment is the clipping stemming from the intensity modulation. A dimensionless parameter identified as key to describing channel capacity is I₀/(qFB) where I₀ (0.01 to 1 mA) is the photocurrent delivered, q is electron charge, F (10 log F=2 to 10 dB) is the electronic amplifier effective noise figure and B (200 MHz to 1 GHz) is the overall system bandwidth. The following are reported: the optimal design of the head end where signals are launched, availability of extraordinarily high capacities, and how a hefty fraction of capacity might be achieved