We evaluate the benefits of using WDM for the dual purpose of increasing data rate and mitigating peak-power nonlinearities in high-power M-PPM optical transmitters used in high-sensitivity applications that incorporate forward error correction.
1Introduction It is well known that optical communications using M-ary pulse position modulation (M-PPM) can provide improved receiver sensitivity [1-4] that is suitable for ultra-long-haul free-space applications. These benefits have been established with demonstrated receiver (RX) sensitivities approaching 4 photons/bit (PPB) in preamplified RXs [5] and ~1 PPB in photon-counting RXs [6][7][8]. The sensitivity benefits continue to increase as the alphabet size (M) increases, and the duty cycle, which varies inversely with M, is reduced [2]. However, practical implementation considerations in the transmitter (TX) such as limited electrical (B E ) bandwidth, modulation extinction ratio (ER), and optical nonlinearities can ultimately limit the range of M for a given data rate (R b ) and thus provide a lower bound on realizable sensitivity [9] for a given average TX power.In order to prevent TX peak power loss, a high ER is required for large M. For ER > M + 15 dB, the reduction in peak power is less than 0.1 dB [5]. Commercially available modulators have demonstrated ER > 50 dB [5,10], which enables M ≤ ~1024-ary PPM with negligible penalty. However, nonlinear effects typically limit M to much lower values. Such effects occur when the peak power exceeds a threshold P th , which is about 250W to 1kW in commercially available EDFAs and YDFAs [9,11,12]. For 10W class TXs, this limits M to < ~128. Fast nonlinearities such as self-phase modulation (SPM), four-wave mixing (FWM), and Stimulated Raman Scattering (SRS) tend to shift usable signal power out of band, an effect that reduces TX efficiency and makes an averagepower-limited amplifier appear peak power limited. Standard steps to increase the nonlinear threshold include using larger core fiber to reduce the power density, minimizing amplifier lengths [13], and distributing amplified signal energy in time to lower the effective peak power through the amplifier. This last approach can be achieved using post-amplified chirped-pulse compression, which is limited by the maximum chirp and compensation that can efficiently be employed. Alternatively, WDM distribution has been suggested as a means of reducing the effective peak power and increasing the data-rate in an electrically bandwidth-limited system [5], as shown in Fig. 1.
MasterLaser Power Amp Power Amp Master Laser Master Laser Master Laser Intensity Modulator Intensity Modulator Intensity Modulator Intensity Modulator WDM λ 1 λ 2 λ 3 λ w P in P out Magnitude Time λ 1 Δt Δt/M Magnitude Time λ 4Δt 4Δt/M λ 1 λ 2 λ 3 λ 4 WDM/PPM Fig. 1 (Left) Block diagram for notional WDM master oscillator power amplifier (MOPA) TX. Single-channel (upper-right) and w-channel M-PPM/WDM waveforms for w = 4 and M = 8 (lower-right)illustrating the WDM method of diminishing peak signal power by d...