Creating optical signal-to-noise ratio standards

William B. Gardner

Signal-to-noise ratio, intersymbol interference, and jitter can all affect the performance of receivers. The combined effect of these and other degradations is characterized by the bit-error ratio (BER). Although the BER is the bottom line as to whether a digital channel is functioning properly, it must be supplemented by other parameters to diagnose problems.

The electrical signal-to-noise ratio (known as the Q-factor) has a straightforward effect on the BER, when certain assumptions are made. This relation is documented in Annex A of the International Telecommunication Union`s (ITU`s) Recommendation G.976, "Test Methods Applicable to Optical Fiber Submarine Cable Systems." It also appears in the draft Optical Fiber System Test Procedure OFSTP-5, "Data Analysis of BER vs. Received Power for Digital Fiber Optic Systems," that will soon be balloted in the Telecommunication Industry Association (TIA). Test methods for the Q-factor are being standardized in TIA draft document OFSTP-8, authored by Rajender Razdan of Ciena Corp.

Extinction ratio and eye mask penalties are also measurements with diagnostic value. They are documented in Annex A of the ITU`s draft Recommendation G.691, "Optical Interfaces for Single Channel SDH Systems with Optical Amplifiers, and STM-64 Systems."

Unfortunately, most of these diagnostic measurements are impractical at some interfaces in wavelength-division multiplexed (WDM) systems. To make per-channel electrical measurements at a multichannel interface point would require demultiplexing a potentially large number of channels. This dilemma has shifted attention to optical measurements at such interface points. Appendix I of the ITU`s Recommendation G.692, "Optical Interfaces for Multichannel Systems with Optical Amplifiers," references the optical signal-to-noise ratio (OSNR). This parameter can be measured without demultiplexing, but its complex relationship to the BER is still under study in standards bodies. When optical amplifiers are present, amplified spontaneous emission can have a significant effect on the OSNR (see Lightwave, Aug. 1998, p. 80).

Jack Dupre of Hewlett-Packard has authored the draft OFSTP-6, "OSNR Measurement Procedures for Dense Wavelength-Division Multiplexed Systems," which was discussed at last January`s meeting of TIA`s Joint Subcommittee FO-2.1/6.6, chaired by Allen Cherin of Lucent Technologies. The author offers one measurement option involving a diffraction grating-based optical spectrum analyzer (OSA) and another using a Michelson interferometer-based OSA (see Lightwave, May 1998, p. 70).

The OSA measures the optical power at the peak wavelength of a WDM channel. The optical noise power is then measured on either side of this channel and interpolated to the middle of the channel. A simple calculation using these measurements leads to the OSNR. The dynamic range of the grating-based OSA limits the values of OSNR that it can measure, while the noise floor of the Michelson-based OSA limits its OSNR measurement range.

With either OSA, the resolution bandwidth must be chosen to encompass the power in one (but only one!) WDM channel. OFSTP-6 contains an annex that analyzes the impact of signal spectral width on measurement error, and provides some guidance on optimizing the instrument`s resolution bandwidth for the measurement. q

William B. Gardner represents Lucent Technologies, Norcross, GA, on sev eral fiber standards committees. He can be contacted at tel: (770) 798-2674; fax: (770) 798-4654; e-mail: wbgardner@lucent.com.