TABLE OF CONTENTS The relative-intensity noise (RIN) is an important quantity in determining whether lasers are acceptable for use in optical-communication systems. Its analytic study can require extensive algebra [1], but in this appendix the emphasis is on the physical significance of RIN and simulations using time-domain modelling to estimate its value for DFB lasers.
The discovery of the optical-fibre amplifier, with its relatively low-noise optical amplification, has maintained the interest in amplitude modulation for optical communication. Although both analogue and digital modulation are of interest, the discussion here focuses on digital intensity modulation. The net photon stream is turned from a digital-signal pattern (Figure A8.1 a) imposed on a random photon stream (Figure A8.1 b), the latter being created by (i) the random spontaneous emission and (ii) shot noise in the electron current driving the laser. So far this shot-noise term has not been considered, but it is easily added into the numerical model, as discussed shortly. The signal and noise combine (Figure A8.1 c) and, taking the idealised case, with 100% efficiency for the photon detector, the photon stream is translated directly into electronic current in the load of the photodetector:
| (A8.1) |  |
Shows an ideal digital signal
Photon detection within a decision window varies with a Gaussian distribution
Shows carrier power plus noise power. Decisions are made as to whether one or zero has been transmitted according to a detection of photons above or below the decision level within the decision window of time
Noise peaks give...
