Optimal Trade-Off of Mode-Mixing Optical Filtering and Index Difference in Digital Fiber Optic Communication Systems
01 May 1974
Optimal Trade-Off of Mode-Mixing Optical Filtering and Index Difference in Digital Fiber Optic Communication Systems By S. D. PERSONICK (Manuscript received December 11, 1973) In a digital fiber optical communication system, the optical power required at the receiver input to achieve a desired error rate depends upon the shape of the received pulses. In systems employing multimode fibers and/or broadband sources, we can experience pulse spreading in propagation because of the group velocity differences of different modes or because of dispersion. In an effort to control or compensate for pulse spreading, we can trade off coupling efficiency between the light source and the fiber (by varying the core-cladding index difference or bandlimiting the source), scattering loss in the fiber (by introducing mode coupling), and equalization in the receiver at baseband. This paper investigates the optimal trade-off for various fiber-source combinations. I. INTRODUCTION AND REVIEW OF BACKGROUND MATERIAL In digital fiber optic communication systems, as in other digital systems, the received power required at a repeater to achieve a desired error rate depends upon the shape of the received pulses. A previous paper1 showed that the minimum average power requirement results from a pulse that is sufficiently narrow so that its energy spectrum is almost constant for all frequencies passed by the receiver (ideally, an impulse). For other received pulse shapes, we can define the additional power required, in decibels, as a "power penalty" for not having impulse-shaped pulses.